Annals of the Missouri Botanical amen, umber Volume 89, Number 1 Winter 2002 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 Committee 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 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 2002 is $140 per volume U.S., $150 Canada & Mexico, $175 all other countries. Four issues per vol- ume. The journal Novon is included i in the sub- scription price of the ANNALS. annals@mobot.org (editorial queries) http://www.mobot.org/mbgpres 8. © — Botanical Garden 2002 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. PosrMAsTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897 The mission of the Missouri Botanical Garden is to disc "over and share — about plants and thoi their and enrich life, LS ANSI/ 150 239.48-1992 — — of Paper). Volume 89 Annals Number 1 of the 2002 Missouri Botanical Garden A SYSTEMATIC REVISION OF Sylvain G. Razafimandimbison?? BREONIA (RUBIACEAE- NAUCLEEAE)' ABSTRACT Breonia A. Rich., a Malagasy endemic genus of the tribe Naucleeae (Rubiaceae, Cinchonoideae), is revised here. The morphology of Breonia species is described and compared with that of the other Malagasy Naucleeae, Breonadia Ridsdale. Gyrostipula J.-F. Leroy, and Janotia J.-F. Leroy. Breonia is кизү gu from de 'se genera by having flattened and elongated placentae, imbricate ovules, and indehiscent, multiple fruits r than triangular placentae, ovules attached side by side to the base of the placentae, and free, lo fruits as * these genera. Neobreonia Ridsdale is again included in Breonia. Twenty species are recognized, including eight that are new to science. A full taxonomic treatment, keys, distribution maps of the recognized species are provided. The newly described species are illustrated. Key words: | accrescent disk, Breonia, Cinc egit floral nectary. Mader! ar, Naucleeae, Rubiaceae. Breonia A. Rich. is a Malagasy endemic genus swampy forests; they are absent from the semi-arid of the tribe Naucleeae (Cinchonoideae, Rubiaceae) regions of southern Madagascar. The present revi- and is the most diverse member of Naucleeae there. sion recognizes 20 species within Breonia, of which Species of Breonia are large trees, or rarely shrubs, 8 are newly described herein. Breonia is distin- that occupy habitats ranging from eastern rainfo- — guished from the other Malagasy Naucleeae, Breon- rests to western deciduous dry forests as well as adia Ridsdale, Gyrostipula J.-F. Leroy, and Janotia ! | thank Charlotte Taylor for guiding this revision and generously sharing her knowledge of Rubiaceae with me; my sicher Mick Ric — Pete Lowry, George Schatz, and James Miller for their advice and support кн my stay in St. Louis; Peter Stevens, Toby Kellogg, three anonymous reviewers, and all the members of the Desmond Lee molecular systematics lab of the Biology Department of the University of Missouri-Saint Louis A their invaluable comments; Roy Gereau for his help with the Latin diagnoses; Petra De Block, Diane Bridson, and Aaron Davis for their assistance and advice; Trisha Consiglio for her assistance with ArcView; Kendra Sikes for her assistance with the TROPICOS database; David Frodin for sharing his thoughts on g Cephalanthus dient issues; Raphael Govaerts for sharing his opinions on the Hichard and De Candolle issues; Barbara Alongi for her excellent illustrations; MEF and ANGAP for collecting permits; and the following herbaria and pan staff for providing loans, access to collections, and/or assistance in the field: BR, K, L, MO, P, PRE, TAN, ТЕК, and WAG. This research was supported by the Ire Andrew W. Mellon Foundation, the Liz Claiborne Foundation, the Rockefeller Foundation, the Garden Club of Alle ghe пу County m and the Missouri Botanical Garden. This work was part of my Ph.D. dissertation presented to the Faculty of the Gra rmm School of the University of Mis souri-Suin Louis, U.S.A. ? Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, * Current address for reprint requests: fae of Systematic ем. — kis Centre, University of Uppsala, Norbyv. 18D, SE-752 36, Uppsala, Sweden. sylvain. razafimandimbison@ebe.uu ANN. Missouni Bor. GARD. 89: 1—37. 2002. Annals of the Missouri Botanical Garden J.-F. 1 tae, eroy, by having flattened, elongated placen- imbricate ovules, and indehiscent, multiple fruits rather than triangular placentae, ovules at- tached side by side to the base of the placentae, and free, capsular fruits as in these genera. Evi- dence from recent molecular studies (Razafiman- 2002) showed that the mono- typic genus Neobreonia Ridsdale and Breonia sensu dimbison & Bremer, Ridsdale together formed a monophyletic group. A more recent phylogenetic study by Razafimandim- bison and Bremer (in prep.) based on four data sets (ITS, rbcL, trn T-L-F. Neobreonia within Breonia, making the latter par- and morphological) placed aphyletic. Both genera have multiple fruits and large accrescent disks. In light of this evidence, I here merge Neobreonia with Breonia. Both studies also showed that Breonia is more closely related to Breonadia, Gyrostipula, and Janotia than it is to the rest of Naucleeae. Recognition of Breonia as a separate genus 15 1975). its circumscription and species limits have been 1897; Homolle, 1938: Ridsdale, 1975). Many more specimens were avail- widely accepted (e.g.. Ridsdale, However, controversial. (Haviland, able for the present study than for earlier studies. Ridsdale's specimens of Breonia in the Paris herbarium (P). revision was based primarily on. the However, many of the specimens from the Forestry herbarium in Madagascar (TEF) studied here are not held as duplicates in P. Therefore, Ridsdale did not see them. MATERIALS AND METHODS This revision is based on the examination of over 200 herbarium specimens and on field observations of 9 of the 20 Breonia species. Specimens from the following herbaria were examined and annotated during this study: BR, K, L, MO, P, PRE, TAN, and TEF (see Appendix 1). “S All specimens with “SF” meaning “Station Forestière, апа “RN” meaning “Réserves Naturelles,” were collected by Malagasy technicians under the supervision of the French I used the “SF” and “RN” numbers rather than the collector numbers because botanist Réné Capuron. many specimens of this series do not have collector names on their labels. Also, these are the numbers used in the TROPICOS database of the Missouri Botanical Garden. Fieldwork was conducted in Madagascar in May-July of 1995 and 1996, respectively, and Jan- uary—June of 1998 to collect herbarium specimens and pickled material of Breonia species, as well as to gather ecogeographical data. The pickled mate- rial was preserved in formalin/acetic acid/ethanol, Table Distinctive fertile morphological features separating Breonia and Sarcocephalus. Characters Breonia Sarcocephalus Inflorescence posi- axillary terminal lon Involucral bract always sur- never surround- rounding the ing the young young inflo- inflorescence escence Attachment of pla- to the septal to the middle of cenlae to septa apices the septa Stigma shape globose, clavate, spindle-shaped capitate ca. basal Receptive areas on са. apical the stigmatic obe Accrescent disks present absent Number of ovules ] 10 9 more than 50 per locule FAA (Radford et al., 1973). This allowed detailed studies of inflorescences and infructescences and revealed a number of important characters, which are impossible to observe in dried specimens. TAXONOMIC HISTORY IDENTITY OF BREONIA Breonia was originally described by Achille Ri- chard (De Candolle, Sep. 1830); he named the ge- nus after the early 19th century French botanist Jean Nicolas Bréon, one of the first collectors. In 1879, Baillon placed Breonia in the synonymy of Sarcocephalus Afz. ex Sabine solely because both have multiple fruits, a character that evolved at least three times in Naucleeae (Razafimandimbison & Bremer, 2002). Baillon included two species of Sarcocephalus, S. Rich.) Baill. from Madagascar. Breonia is quite distinct from Sarcocephalus in many aspects (see Table 1). Schumann (1891), Haviland (1897), and Homolle (1938) all endorsed the recognition of Breonia as a separate genus. The madagascariensis (A. and S. richardiana Baill., separation of Breonia and Sarcocephalus and the monophyly of Breonia is strongly supported by ITS and rbcL sequence data (Razafimandimbison & Bremer, 2002). AUTHORSHIP OF BREONIA The authorship of Breonia ч bee n cited previ- Rich. 1975, 1978). I disagree with a pes Pyramus de Candolle asked Achille Richard for a copy of his manuscript of the Mémoire de la Famille Rubiacées ously as Breonia A. . (see Ridsdale, Volume 89, Number 1 2002 Razafimandimbison 3 Revision of Breonia Table 2. Species circumscriptions of Breonia. (Richard, Dec. 1830) to avoid any possible nomen- clatural conflicts between their work. At that time, Richard’s work was already in press for publication. Richard kindly sent De Candolle the manuscript, along with 45 illustrations and later the page proofs, hoping that his work would come out before De Can- dolle's Prodromus Systematis Naturalis (Stearn, 1957). De Candolle was able to cite Richard’s work ahead of its actual issue, since the Prodromus was published in September 1830 ahead of Richard’s book (published in December 1830). Comparison of Richard’s original description of Breonia with that included in De Candolle’s book showed that De Can- dolle used Richard’s description of Breonia, but in a different format. The authorship of the genus should remain with Richard. Therefore, I cite Breon- ia as Breonia A. Rich. in [DC.]. PREVIOUS SPECIES DELIMITATION IN BREONIA The species limits used by previous authors have been controversial. Haviland (1897), in his first worldwide revision of the tribe Naucleeae, recog- nized 10 species of Breonia (Table 2). He used pet- iole length and secondary leaf vein number to di- vide the genus into two major groups: the first group Haviland (1897) Homolle (1938) Ridsdale (1975) Razafimandimbison (2001) boivinii boivinii boivinii boivinii coriacea —! not mentioned —* cuspidata cuspidata — cuspidata longipetiolata dubious species dubious species dubious species madagascariensis madagascariensis madagascariensis madagascariensis mauritiana —! — — membranacea membranacea — membranacea parvifolia parvifolia sphaerantha sphaerantha richardiana richardiana — — stipulata —! stipulata decaryana Neobreonia decaryana decaryana havilandiana — havilandiana keliravina — — louvelii — louvelii macrocarpa — macrocarpa Jerrleri perriert errieri citrifolia chinensis capuronil fragifera owryi richardsonii sambiranensis —3 taolagnaroensis tayloriana Isaratananensis —' merged in B. richardiana; — merged in В. madagascariensis; —? merged in B. citrifolia; — merged in B. chinensis; —^ merged in B. decaryana. comprising Breonia madagascariensis, with subses- sile leaves and 12 secondary veins (there are ac- tually 18 or 19 secondary veins, but the leaf blade of the type specimen is folded, so Haviland must have counted what he could see); and the second group containing the remaining species, which had petiolate leaves with 9 to 11 secondary veins. Hav- iland then used leaf blade length and width, as well as inflorescence width, to delimit the species of his second “petiolate” group. Forty years later, Hom- olle (1938) adopted new species circumscriptions for Breonia. She retained six species from Havi- land's treatment, but also described six new species (see Table 2). Unfortunately, she did not provide a key to the species she recognized. Ridsdale (1975) proposed new circumscriptions for Breonia based on a combination of the following characters: stip- ules of terminal vegetative buds conical and ob- volute; inflorescences always lateral; calyptra-like bracts coherent, surrounding the young inflores- cence, and circumscissile; adjacent calyx tubes free and persistent; calyx lobes densely pubescent; and placentae attached to the upper third of the septum. He used the shape and length of stipules of ter- minal vegetative buds, the mature stipule length, Annals of the Missouri Botanical Garden and the leaf shape and size as primary characters and recognized only five species (Table 2). He placed Breonia decaryana Homolle and B. kelira- vina Homolle in his new genus Neobreonia because of their partly fused calyx tubes and complanate terminal vegetative buds. Capuron (1973a), deposited at P, recognized 13 species using varia- in an unpublished manuscript tion in leaf blade size, stipule shape and length. calyx tube length and indumentum, ovule number, and fruit surface. However, he provided neither de- scriptions of the new species that he recognized nor a key for their identification. ON THE IDENTITY OF CEPHALANTHUS CHINENSIS LAM. AND ANTHOCEPHALUS А. RICH. This section presents the problem of the identity of Cephalanthus chinensis Lam. |= Breonia chinen- sis (Lam.) Capuron] and its historical confusion with Breonia. This issue has caused a long and contro- versial debate during the last 30 years (cf. Bak- huizen van den Brink, 1970; Capuron, 1973b; Ridsdale, 1975; Bosser, 1984, 1999), Lamarck (1785: 678) claimed that a specimen he received from Sonnerat was the type of his Ce- phalanthus chinensis. On the other hand, Richard (Sep. 1830) also stated that he described his An- thocephalus |= Neolamarckia Bosser| as an Asian, and not Malagasy, monotypic genus (Anthocephalus indicus) based on specimens labeled “C. chinensis Lam.” Lamarck and Richard gave no further details about their type specimens. Bakhuizen van den Brink (1970) and Ridsdale (1975) surmized that the collection of Sonnerat that served as the type of C. chinensis was also used by Richard to describe his Anthocephalus. Capuron (1973b) and Bosser (1984, 1999) argued that C. phalus A. Rich. were not necessarily based on the chinensis Lam. and Anthoce- same specimen. In the Lamarck herbarium in Paris (P-LA) there is an authentic specimen of Cephalanthus with five different labels attached to it. The first label is La- о marck’s and says ‚н chinensis enc n The second label is in Commerson's handwriting p Bosser, 1984), with a dcn Latin description and the country of origin mentioned as "Isle de France" (i.e., Mauritius). .. > c'est la Morinda de Chine.” The fourth label The third label is Sonnerat’s and says is also Sonnerat’s, on which he wrote “Morinda de Chine," and below this name Lamarck added “Ce- phalanthus.” The fifth label was written by Roeper "Nauclea orientalis Lam." This specimen sheet was consid- ered by Capuron (1973b) and Bosser (1984, 1999) and says "Nauclea purpurea Roxb.” and to be the type of Lamarck’s Cephalanthus chinensis. However, Bakhuizen van den Brink (1970) and Ridsdale (1975) disagreed because this specimen does not match Richard’s description of Anthoce- phalus. Despite the fact that no one has ever found LA that could have served as the type of both C. chinensis Lam. a collection of Sonnerat s.n. in P or P- and Anthocephalus, Bakhuizen van den Brink and Ridsdale inferred that Lamarck must have based This collec- his C. chinensis on such a collection. tion, they further surmized, must contain two ele- ments, one of Breonia and the other of Anthoce- phalus. They also argued that Richard must have typified his Anthocephalus indicus by only the ele- ment of Anthocephalus of this hypothetical Sonnerat collection. Bakhuizen van den Brink then typified Anthocephalus by only the element of A. indicus of this hypothetical Sonnerat collection, excluding its element of Breonia. Ridsdale (1975) endorsed Bak- huizen van den Brink's arguments and. chose the unverified or non-existent collection as the lecto- type of Anthocephalus. Capuron (1973b), (1984, 999), disagreed with Bakhuizen van den Brink (1970) and Ridsdale (1975). Capuron and Bosser ar- gued that the type specimen of Cephalanthus chi- followed by Bosser nensis Lam. was the specimen with five labels in P- LA because the original description of C. chinensis am. (Lamarck, 1785: 678) — this plant, and amarck himself annotated it. Bosser argued. that amarck was apparently — ed by Sonnerat’s la- bels stating “Morinda de Chine,” and assumed that the plant originally grew in China, the Philippines, 1984: 244). Bosser then inferred. that Sonnerat’s label meant that Sonnerat and the Moluccas (Bosser, had seen the specimen of C. chinensis now in P-LA, and that it reminded him of what he called “Morinda de Chine.” Bosser (1984) also argued that the sec- ond label was Commerson's because there is a sim- ilar specimen in P labeled C. chinensis Lam. and it was collected by Commerson in Mauritius. In addi- tion, the specimen with five labels in P-LA shows axillary inflorescences and adjacent flowers with fused ovaries that are characteristic of Breonia. Ca- puron (1973b) then concluded that C. chinensis Lam. was based on a plant originally from Madagascar rather than Asia, and therefore belongs to Breonia. Consequently, Capuron made the new combination B. chinensis (Lam.) Capuron. l agree with Capuron's and. Bosser's arguments. A large number of living plant collections of early French botanists (e.g., Chevalier Etienne de Fla- Jean Nicolas Bréon, court, Philibert Commerson, André Michaux, Louis Armand Chapelier) were sent to Mauritius for cultivation in the late 18th Volume 89, Number 1 2002 Razafimandimbison 5 Revision of Breonia century (Dorr, 1997). Commerson sent a large col- lection of dried plant specimens to Paris, but these specimens were labeled and distributed to various 1773 (Dorr, 1997). This may well have caused some confusion herbaria only after his death on 13 May between Commerson’s and Sonnerat’s specimens [see. e.g.. Guého (1976) regarding the type of Phy- Heine (1968) for Cordia sinensis Lam.|; the type of C. chinensis is apparently another lica nitida Lam.; one of these cases. Bosser (1984) has pointed out correctly that our problem is to find out whether the specimen of C. chinensis in P-LA conformed to what Lamarck said in his original description of C. chinensis, rather than what Richard said about Anthocephalus. In- deed, Bosser argued that Richard may not have seen the specimen of С. chinensis in P-LA because Lamarck sold his herbarium to Roeper in 1824 and the entire collection was not returned to Paris until 1886, while Richard’s work was in press for pub- lication in 1830. Richard’s original description of Anthocephalus indicates that he used specimens with flowers and fruit, whereas Lamarck had specimen with flowers only; therefore, even if Ri- chard had seen the specimen in P-LA, he would still have needed a specimen with fruit. Finally, Anthocephalus, as Richard noted, has terminal in- florescences, whereas the specimen of C. chinensis in P-LA has axillary inflorescences. Therefore, Bakhuizen van den Brink’s and Ridsdale’s scenar- ios of C. chinensis being based on mixed elements of Breonia of Sonnerat and Anthocephalus is unten- able, in agreement with Bosser (1999). Richard's protologue for Anthocephalus stated: “Species observata: Anthocephalus indicus nob. qpephatanthus chinensis Lam.—Nauclea purpurea * (Richard, 1830: 157). Bosser argued that the type specimen of B. mauritiana [= B. chinensis herein] in P, collected by Commerson (s.n.) i — =) Mauritius, may well have been one of the speci- mens on which Richard based his Anthocephalus because: (1) this specimen has two old labels; the names on the second label are the same names as those in the protologue of Anthocephalus; and (2) there is a packet with pieces of an infructescence attached to this specimen, and the separated indi- vidual fruits of this infructescence match Richard’s description of the infructescence of Anthocephalus. Thus, Richard may well have used the fruiting head attached to the type specimen of B. mauritiana as part of his original description of Anthocephalus. We know only that the element of Anthocephalus was attached to the type sheet of B. mauritiana when Haviland (1897) described this species: “ It is this specimen which has attached to it some heads of Anthocephalus indicus and the curious confusion about Nauclea purpurea and Cephalan- thus chinensis.” | consider this fruiting head at- tached to the type specimen of В. mauritiana as a piece from the type specimen of Richard’s Antho- cephalus indicus, which is now missing. In conclusion, Cephalanthus chinensis Lam. is Breonia, whereas Anthocephalus A. Rich. is Asian. Bosser (1984) created a new generic name Neola- marckia for the Asian taxa that match Richard’s original description of Anthocephalus, and made the new combination Neolamarckia cadamba (Roxb.) Bosser based on Nauclea cadamba Roxb. (1824). | agree with Bosser's combination because N. cadam- ba Roxb. is the older name, and thus it has priority over А. indicus A. Rich. Ridsdale (1998), in his treatment of Rubiaceae for Sri Lanka, continued to use Anthocephalus rather than Neo- 30). However, lamarckia for these plants. SPECIES CIRCUMSCRIPTIONS IN BREONIA SENSU RAZAFIM. Haviland's (1897) circumscription of Breonia spe- cies based on secondary leaf veins, leaf blades, and inflorescences is problematic because these charac- ters overlap for some well-defined species. Using such features as stipules of terminal vegetative bud size, mature stipule length, and leaf shape and size as pri- mary discriminating characters forced Ridsdale to lump taxa and recognize only five species, of which three (Breonia citrifolia, B. madagascariensis, and B. sphaerantha) become rather heterogeneous. Neither Haviland (1897) nor Ridsdale (1975) used floral and fruit characters for their keys for Breonia. However, there are a number of useful taxonomic characters, such as shapes of the stipules of the terminal vege- tative bud and inflorescence axis, corolla lobe indu- mentum, the degree of fusion of adjacent ovaries, ac- and infructescence surface a | abandon Haviland's (1897) and Ridedale’s 1975) species delimitations and adopt here new spe- crescent disks, result, —. cies limits within Breonia. My criterion for recognizing species is the pres- ence of one or more apparently fixed or non-over- lapping morphological differences between putative species. Morphologically distinct groups of speci- mens were identified and delimited by non-over- lapping or fixed diagnostic differences; these di- agnosable morphological units, which I refer to as species, are hypotheses until new data are available to refute them. To relate this functional species con- cept to more theoretical ones, one must test it on the basis of the criteria defining each theoretical species concept. Annals of the Missouri Botanical Garden = С б. А B р Е Е Figure l. Variation in shape and size of the stipules of the vegetative buds in Breonia (A, С-В) апа Janotia. Breonia decaryana.—B. Janotia macrostipula.—C. 'ariensis MORPHOLOGY AND TAXONOMIC CHARACTERS OF BREONIA A detailed overview of the morphological char- acters of Breonia species is presented as well as their taxonomic significance. The allied Malagasy genera Breonadia, Gyrostipula, and Janotia are also discussed for comparison. HABIT Plants of Breonia are medium to large, some- times emergent trees up to 30 m tall, rarely shrubs. n Breonia sphaerantha. Buttresses are reported The bark is smooth to rugose. The trunk is typically straight. The branches are usually plagiotropic. LEAVES Breonia leaves are simple, opposite, and decus- sate, as in most Rubiaceae, and generally coria- ceous to membranaceous. The length of petioles and the size of leaf blades are useful for recognizing some species. The leaves of Breonia are petiolate or rarely subsessile (B. madagascariensis). Most species have small to medium-sized leaves, but two species (B. macrocarpa and В. madagascariensis) the former to 38 The have relatively large leaf blades: X 25 em, and the latter up to 45 X 35 em. leaf blades are mostly glabrous, or rarely pubescent in a few individuals of B. perrieri. Breonia macro- carpa is characterized by having brown, long hairs on the lower surface of the leaf blades. Domatia. | Domatia are cavities located in the ax- ils of veins on the abaxial sides of leaf blades (Ja- cobs, 1966). They typically occur in the axils of sec- ondary veins in most species of Breonia. Mites or mite eggs can be found in the domatia, although the occurrence of domatia is not dependent on the —A B. capuronit.—D. B. chinensis. —E. B. stipulata.—V. B. madagas- arachnid presence; these domatia may provide the host plant protection from herbivores or pathogens (Pemberton & Turner, 1989). The type and location of domatia are useful for species delimitation i Breonia. In B. macrocarpa, they occur in the axils of the secondary and tertiary veins. Both secondary and tertiary domatia have never been seen in other Naucleeae; they are only known in a few members of Rubiaceae (e.g.. Aoranthe, Pleiocoryne, Oligoco- don) (Robbrecht, 1988). In Breonia, domatia can be tufts (covered by dense hairs), or cryptic and. gla- brous, or pits (depressions in the lamina with a broad opening). Venation. The venation of the leaf blades is usually prominulous on the upper sides of the leaf blades, but is prominent in Breonia havilandiana, The secondary venation is usually eucamptodromous Radford et al., 197: Stipules of — vegetative buds. ules of terminal vegetative buds are usually conical and obvolute in most Breonia species (Fig. 1C-F). In these, two stipules overlap in the bud such that В. madagascariensis, and B. tsaratananensis. — The stip- one half of each is external and the other half is internal (Harris € Harris, 1994). lengths vary between species, which is useful for their identification. In Breonia decaryana and Jan- otia macrostipula, the stipules of terminal vegeta- tive buds are complanate (Fig. 1A, B), the two stip- ules being flat and pressed together. In Gyrostipula, Their sizes and they are long. filiform, and convolute, one stipule being rolled inside another (Radford et al., 1973). Mature stipules. The stipules of Breonia and other Malagasy Naucleeae are interpetiolar, boat- shaped (cymbiform), and usually abaxially carinate. The keels are prominulous in many species and prominent in B. tayloriana and В. madagascarien- Volume 89, Number 1 Razafimandimbison 7 Revision of Breonia In Gyrostipula, the stipules are without keels and are twisted when dry. The size and persistence of the stipules vary among species of Breonia and are useful for species distinction. Stipules are typ- ically deciduous in Breonia. In Breonia tayloriana, they are semi-persistent. INFLORESCENCES Like other Naucleeae, Breonia species have glo- bose inflorescences on which numerous small flow- ers (usually more than 50) form dense clusters rath- er than capitate other Rubiaceae (e.g., Schradereae, Psychotrieae). The individual flowers are inconspicuous, and the entire cluster of flowers acts as a single attractive unit. Inflorescence stalk. In Breonia, the inflores- cence stalk is typically articulate and composed of a peduncle, a node, and an inflorescence axis (e.g.. Figs. ӨК, OF, 14C). Ridsdale (1975) used the term “flowering axes” to refer to what I call inflorescence axes; he called “true peduncles” what I refer to as inflorescences found in peduncles. An evident node separates the peduncle from the inflorescence axis. The inflorescence axis is located between the branch bearing the inflores- cence and the node. The peduncle is located be- tween the node and the base of the head; this is hidden by the bracts prior to the separation of the bract lobes. The peduncle can be absent (e.g.. Figs. 15A, 16F); in some species (e.g., Fig. OF, Breonia capuronii; B. decaryana), this continues to elongate after the inflorescence axis has finished growing at the end of anthesis. The length and shape of the inflorescence axes are useful for recognizing spe- cies in Breonia. Inflorescence axis shapes can be terete (e.g... Вгеота саригопи, B. chardsonii, B. sphaerantha) or flattened (e.g.. В. chinensis, B. havilandiana, B. membranacea, B. sti- pulata, and B. tayloriana). The inflorescence axes are usually longer than the peduncles; however, in Breonia decaryana, B. capuronii, and B. sphaer- antha, the two are similar in size and shape. racts. Most Breonia species have bracts that are coherent, but histologically distinct (Radford et al. 1973), valvate, and completely enclose the young inflorescence in a calyptra-like fashion. As the inflorescences develop, the bracts separate lon- gitudinally into two lobes on the nodes lasting for only a few days and usually falling off before an- thesis. They may persist for two or three weeks after anthesis in Breonadia, Gyrostipula, and Breonia de- caryana. In B. richardsonii, the bracts are tubular, appressed to the inflorescence axis, and terminated by three to four broadly triangular lobes. They nev- er enclose the young inflorescence. This is unique in Breonia and the whole of Naucleeae. Finally, the bracts in Breonia are usually glabrous, except in B. macrocarpa and B. tsaratananensis, where they are densely pubescent. The position of the bract varies depending on whether or not the peduncle contin- ues to elongate after the inflorescence axis ends its elongation. When the peduncle does not elongate. the bract lies immediately below the inflorescence, as in B. chinensis, B. cuspidata, B. madagascarien- sis, and В. taolagnaroensis. Ridsdale’s (1975) descriptions of calyptra-like bracts of Breonia being coherent, surrounding the young inflorescence, and circumscissile, appear to be incorrect. I have seen two herbarium specimens of Breonia chinensis showing circumscissile calyp- tra-like bracts. In several species, including B. chi- В. havilandiana, В. macrocarpa, В. branacea, and В. the normally separate longitudinally into two equal hemispheres. This apparent circumscissile ruptur- ing of the bracts evidently happens during speci- men preparation. The illustration of Sarcocephalus richardii Drake [= Breonia chinensis (Lam.) Ca- puron] in Grandidier (1897: 457) shows the circum- scissile calyptra-like rupturing of the bract. It is possible that Ridsdale (1975) was influenced by this illustration. Interfloral bracteoles are absent in Breonia, as well as in Gyrostipula and Janotia. However, they are present in Breonadia. In Breonia richardsonii (Fig. 14F), the ovary bases are surrounded by long hairs. nensis, mem- tsaratananensis, bracts FLOWERS The flowers in Breonia are actinomorphic, her- maphroditic, protandrous, sessile, and usually 5- merous, although 4-merous examples are occasion- ally encountered. The flowers are typically 4-merous in B. capuronii, B. decaryana, B. fragifera, and В. sphaerantha. Calyx. The calyx tubes (i.e., tubular parts of limb above the hypanthium) in most Breonia species are free from each other and clearly visible. In B. fragi- fera and B. decaryana the calyces are barely evident. These calyx tubes of adjacent flowers are only partly fused in B. decaryana; they are completely fused in B. fragifera. Therefore, the degree of calyx fusion is useful for species recognition in Breonia. The shape and surface of the calyx tubes are useful at the species level. In Breonia chinensis, B. lowryi, B. tayloriana, B. stipulata, and B. taolag- naroensis, the calyx tubes are infundibular and smooth. Other species have infundibular and Annals of the Missouri Botanical Garden DENO 00 Ш ге 2. . B. fragifera.—D. В. macrocarpa.—E. B. chinensis.—F. ribbed calyx tubes. The ribs may be prominulous (e.g, B. sambiranensis) or prominent (B. tsaratan- anensis and B. boivinii). In B. tsaratananensis the calyx tubes are constricted above the middle and narrowed toward the base and the lobes; this is unique in Breonia. The shape of the calyx lobes varies from oblong truncate apices. to triangular, sometimes with Breonia sambiranensis has a distinct calyx. lobe shape: short, truncate, with a shallow depression in the center and a short appendage on the edge to- ward the style (Fig. 14D). Broadly triangular and pubescent lobes characterize Gyrostipula, while Breonadia and Janotia have long, linear, pubescent obes. Corolla. Тһе corolla of Breonia is always hy- pocrateriform and vellow-tinged. In B. boivinii, B. sambiranensis, and B. tsaratananensis, some of the corolla lobes have dorsal protuberances (e.g.. see B. sambiranensis, Vig. 14E). Corolla indumentum is useful for species recognition. One species group (B. chinensis, B. cuspidata, В. louvelii, B. mem- branacea, B. stipulata, and B. taolagnaroensis) has glabrous lobes. Breonia boivinii, B. sambiranensis, and B. tsaratananensis all have pubescent or pu- berulous lobes. In B. macrocarpa and B. madagas- cariensis lobes are puberulous and marginally gla- brate. Nectaries. As in other members of Naucleeae sensu Razafimandimbison and Bremer (2002), the floral nectaries (i.e., disks) in Breonia are incon- spicuous, surround the style base, and are embed- ded in the hypanthia between the base of the co- rolla tube and the top of the ovary. For all Breonia, nectaries continue to grow during infructescence development and become hardened and ме. ous. | call such nectaries “accrescent disks.” Ca- puron, in his unpublished treatment for Malagasy Rubiaceae at P, also observed this type of disk in Breonia but did not understand its nature. These accrescent disks appear to be the single synapo- morphy among members of Malagasy Naucleeae. However, they tend to be much larger in Breonia species than in Breonadia, Gyrostipula, and Jano- Different shapes of accrescent disks in Breonia fruits. —A. Breonia decaryana.—B. B. sphaerantha.— B. taolagnaroensis. tia. The shape of these accrescent disks (Fig. 2) varies from obconical to obtriangular, to pentagonal or rounded. Their shape and size are useful for spe- cies distinction. This is the first report of the use of this character in the classification of Breonia as well as in Rubiaceae. Stamens. Breonia filaments are usually very short and flattened; they are inserted in the throat of the corolla tubes and sometimes slightly exsert- ed. The anthers are always bicuspid at the base, basifixed, introrse, dehiscing along longitudinal slits. Gynoecium. In Breonia, ovaries (hypanthia) of individual flowers are always bicarpellate. They are coherent and nonseptate in Breonia capuronii, В. decaryana, B. fragifera, and B. sphaerantha (see Figs. OD, ОР), whereas they are fused (syncarpous) and septate in the remaining species. Adjacent ova- ries are typically fused. However, in B. richardson- п. prior to and during anthesis the ovaries are fused only at the base, but as the development continues, the fusion extends to the mid-point (post-genital fu- sion), producing multiple fruits (Radford et al., 1973). Ovule number varies from 1 to 9; all four nonseptate and three of the septate species (B. cus- B. louvelii, and B. uniovulate; the. remaining 13 septate species are pidata, taolagnaroensis) are multiovulate. P Breonia. acenta size and shape vary between species in The uniovulate species have small and ovate placentae (Fig. 9F), whereas the multiovulate species have large, flattened, and elongated ones (Fig. ЗС. D). The placentae are triangular in Breon- adia, Gyrostipula, and Janotia (Fig. ЗА. B). Pla- centae are usually apically attached to the septum, and adnate to the lateral sides of the nectaries, lo- cated in the distal center portions of the carpels 13H). recognizing Breonia from the other Malagasy genera (see Fig. These two characters are useful for of Naucleeae, but have never been used for clas- sification by earlier authors. Defining Breonia by the placentae attached to the upper third of the septum as originally done by Haviland (1897) in Volume 89, Number 1 2002 Razafimandimbison 9 Revision of Breonia M Figure 3. Different shapes of placentae in гове Breonadia salicina.—C. Breonia sambiranensis 0.1mm his first worldwide revision of Naucleeae, and later by Ridsdale (1975), appears to be incorrect. Ovule attachment to the placenta is another use- ful i that distinguishes Breonia from Breon- adia, Gyrostipula, and Janotia. In Breonia, the vules are flattened and imbricate (Fig. 4D) along the length of the placentae; they are attached side by side at the base of the placenta in Breonadia, Gyrostipula, and Janotia (Fig. 4 Breonia stigmas are typically globose to clavate, or rarely capitate, pollen presenters. The receptive areas are always restricted to the top of the stig- matic lobe. INFRUCTESCENCES In Breonia species the infructescences develop from syncarpous ovaries to form multiple fruits that are fleshy, almost woody, when dry. They are or- namented by accrescent calyces in most species. In B. fragifera and B. decaryana, the calyx remnants are barely evident and their infructescence surfaces are pusticulate (with small, broad, slight elevations A = € с o B ure 4 Ovule 8 salicina. Ra ‚ D) and its allies (А, B). —A. Gyrostipula foveolata.—B. —D. B. macrocarpa. not high or abundant) and rugose, respectively. In Breonia, all the fruits are fused and constitute mul- tiple fruits; the individual fruits are berry-like, bearing 1 to 9 seeds completely enclosed by thin exocarps and thick, hard endocarps. SEEDS In Breonia, the seeds are usually strongly flat- tened ellipsoid. Concavo-convex seeds are found in Breonia chinensis and В. stipulata. Breonia seeds are usually unwinged, although rudimentary wings are sometimes observed at both ends (e.g., B. chi- nensis and B. taolagnaroensis), or only at the base (e.g., B. perrieri). The seeds are possibly released when the corky accrescent disks fall off (see the section on seed dispersal). These disks can be eas- ily removed when the infructescences are still fresh, or after soaking dried ones for several hours. Seed-coat is lineate (e.g., Fig. 6C) in the nonseptate species, and typically reticulate (e.g.. Figs. 9C, 11C, 14H, 16E) in the septate species. 4 NU attac 'hments to d entae in — par its = (A-C). —A. Gyrostipula foveolata.—B. Breonadia Annals of the Missouri Botanical Garden ECOLOGY POLLINATION ECOLOGY How Breonia pollinates remains unknown. In the Ankarana Reserve (northern Madagascar) in 1998 | saw the beginning of anthesis in a young individ- ual plant of B. perrieri. Flowers in each inflores- cence opened simultaneously to emit ап intense odor of nectar, which might attract visitors from considerable distance. The plant was visited by several honeybees for a few hours. DIASPORE DISPERSAL Entire infructescences. In Breonia, the entire in- fructescence can function as a dispersal unit. In 1998, I observed in the Anjanaharibe Sud Reserve (northeastern Madagascar) a number of mature in- fructescences of B. chinensis, a facultative rheo- phyte, floating on the river. The corky accrescent disks may act as flotation devices allowing the in- fructescences to float until they reach the river- bank, where they may be eaten by any frugivores. After flower fall, the nectaries have a new function, as cork-like seed releasers. A different floating device has also been reported on the seed-coat of the African Naucleeae Sarcoce- phalus pobeguinit Pob. ex. Pell., which is a facultative rheophyte (Abbiw, 1985). Abbiw collected mature seeds of Nauclea diderrichii (De Wild. & T. Durand) Merr. and Sarcocephalus latifolius (Smith) Bruce (both terrestrial species — ‚ as well as S. pobeguinii. A simple experiment was performed by soaking the seeds of each species in separate containers of water. After 24 hours, the water was poured off: seeds of Nauclea diderrichii and S. latifolius lay at the bottom of the containers, whereas those of S. pobeguinii were still floating. Nauclea diderrichii and S. latifolius seeds were noted to sink after only a few minutes of im- mersion. The seed-coat apparently acts as a floating device preventing the seeds from sinking. Seeds. — Infructescences of Breonia are used food by all species of Malagasy lemurs. In Lokobe Reserve (Nosy Be District, Antsiranana Province, in northern Madagascar), B. boivinii and B. sambira- nensis were the second most important food item in the diet of the black lemur (Lulemur macaco) in early 1992. Thirteen percent and 25% of its feeding time during January and February, respectively, were spent on these two species of Breonia (Birkinshaw. 1995). Birkinshaw frequently observed intact seeds of B. boivinii and B. sambiranensis in lemur drop- pings. He found that 2 of 7 defecated seeds germi- — nated, compared with 7 =~ - -30 seeds collected from the ripe infructescences used as a control. The ma- ture infructescences were not observed to be eaten by any other frugivores in the Lokobe Reserve. Both species of Breonia also occur in the Ambanja region where Eulemur macaco is also known to be common, suggesting that £. macaco may be an effective agent of seed dispersal for B. boivinii and B. sambiranen- s. Notably, B. sambiranensis is known from only two specimens. outside of the Sambirano regions. Seed dispersal mechanisms for other Breonia species re- main unknown. ECONOMIC USES All Malagasy Naucleeae are known under the lo- “Valotro,” which is reserved for a particular use (Boiteau, 1985). From the 19th century through the mid 20th century, the Malagasy Naucleeae were used for cal name “Valotra” or meaning that posts placed around villages to protect against en- emies. All Breonia have various local uses such as boats, bridges, making handerafts and furniture, and for general construction purposes because they produce high-quality, hard wood. Several species are used medicinally. The infructescences of Breon- ia are eaten by animals and humans. SYSTEMATIC POSITION OF BREONIA SENSI IN NAUCLEEAE SENSU RAZAFIM. RAZAFIM. & BREMER Breonia has been placed in quite different tribes over the last 170 years. Richard (1830) and De Can- dolle (1830) placed Breonia in the tribe Opercular- ieae along with other tribes with pluriovular locules. Endlicher (1841) placed the genus in his subtribe Sarcocephalinae (of the tribe Gardenieae), along with Sarcocephalus, Zuccarinia Blume |= Jackiopsis Ridsdale], Lucinaea DC. |= Schradera Vahl]. and Canephora Juss. Schumann (1891), followed by Hav- iland (1897), Verdcourt (1958), Ridsdale (1975), and Robbrecht (1988, 1994), placed Breonia in the tribe Naucleeae based on its compact spherical inflores- Bremekamp (1966) transferred Breonia, along with Adina Salisb., Mitragyna Korth., cences. However, Uncaria Schreb., and Neonauclea Merrill, from the tribe Naucleeae to the tribe Cinchoneae, arguing that they differ from the other genera of Cinchoneae in the capituliform inflorescence only. All these genera, with the exception of Breonia, have capsular fruits with winged seeds, which are characteristic of Cin- choneae. Bremekamp provisionally placed these genera in a separate subtribe within Cinchoneae based on their capituliform inflorescences. Evidence and rbcL sequence data (Razafimandimbison & Bremer, 2002) strongly suggests that Breonia belongs to Naucleeae from recent molecular studies based on ITS Volume 89, Number 1 2002 Razafimandimbison 11 Revision of Breonia and is closely related to the other Malagasy genera Breonadia, Gyrostipula, and Janotia. It is worth not- ing that one species of African Uncaria, U. africana var. africana, is also present in Madagascar. These five genera can be identified using the key below. Ккү TO THE GENERA OF MALAGASY NAUCLEEAE la. Climbers, ре fang — — — CUncadrid Se eh м species) Ib Trees o or ere dnd fang pac abser N c Hypanthia of the adjacent flowers fused or at least at the base; fruits fused, indehiscent Breonia A. Rich. . Hypanthia of the adjacent flowers free; fruits free, dehiscent. 3a. Leaves verticillate; stipules intrapetio- lar: interfloral bracteoles present; seeds unwingec (20 pee ies) 3 - Breonadia Ridsdale (monotypic ) La eaves в бый stipules interpetiolar; in- terfloral brac teoles 2 seeds winged. 4a. Leaf blades x 8-10 cm; stipules of 8 vegetative buds беде, — — persistent — — Janotia J. о. Leroy (monotypic) Ab. Leaf biudes 8-10 X 3-5 cr i ules of the terminal vegetative buds convolute; stipules deciduous — c n; stip- Gyosipila J- -F. Leroy (2 species, E ۳ ıich is yet to be described) RELATIONSHIPS AMONG SPECIES Species of Breonia can be subdivided into two dis- (1) a group with reduced calyces, coherent carpels, non- tinct groups based on the following characters: septate, and with lineate seed-coat (Breonia саригопи, B. decaryana, B. fragifera, and B. sphaerantha); and (2) a group with well-developed calyces, syncarpous carpels, septate, and with reticulate seed-coat (re- maining species). The monophyly of these two groups needs to be tested using fast-evolving markers. The included key to Breonia species does not re- flect the above relationships; instead, | propose a more practical key that puts more emphasis on the most obvious unique vegetative character states, such as the shape and arrangement of the stipules of the terminal vegetative buds and size of leaf blades. TAXONOMIC TREATMENT Breonia A. Rich., in DC., Prodr. 4: 620. Sep. 1830. T 4 РЕ: Breonia madagascariensis А. Rich БЕ ТЕ" A. Rich.. Mém. Fam. Rubiacées. 210. Dec YPE: Cephalidium citrifolium A. Rich. E * onid — Franchetia Baill., Bull. Mens Linn. Paris 1: 1885. YPE: Pains — Baill. [= B sphaerantha |. Elattospermum Sol., Bull. Herb. Boissier 1: 277. dad TYP " lale Foe Sol. [= | 540. TYPE: sphaerantha майы "Ridsda le, Blumea 22: 1975. Bre reonia decaryana Homolle. Trees or emergent trees, or rarely shrubs. Bark gray. Stipules of terminal vegetative buds conical or rarely ovate to obovate, obvolute or rarely complanate, gla- brous or pubescent. Leaves simple, opposite, decus- sate, persistent or deciduous; domatia present in axils of secondary veins or rarely tertiary veins or absent: stipules interpetiolar, mostly cymbiform or rarely com- planate, deciduous or rarely semi-persistent, entire. Inflorescences usually solitary or sometimes 2 or rare- ly 4 to 8 per axil, axillary, globose; inflorescence axes unbranched or rarely branched, flattened or terete, usually slender or rarely robust and woody, glabrous to pubescent, articulated or not; bracts usually calyp- tra-like, cohering and completely enclosing the young inflorescence, separating longitudinally into two equal hemispherical shells, remaining attached to the node for a few days and then falling off; peduncles elon- gated (as internode above the inflorescence axis) or not. Flowers hermaphroditic, radially symmetrical. mostly 5-merous or sometimes 4-merous, closely con- gested, sessile; calyx tubes infundibular, inside usu- ally densely pubescent, outside glabrous except for a few straight long hairs around the base, free from oth- ers or rarely fused; calyx lobes oblong to truncate, densely pubescent; corolla tubes hypocrateriform, in- side glabrous to puberulous, outside glabrous; corolla lobes oblong, glabrous or puberulous to pubescent, aestivation imbricate in bud; stamens 5 or rarely 4, inserted on the throat of the corolla tubes; anthers introrse, partly exserted, dehiscing by longitudinal slits, basifixed: filaments short, flattened, glabrous: stigmas 1 per flower, clavate to globose or slightly cylindrical, exserted from the corolla tubes, pollen presenters; receptive areas restricted to the top of the stigmatic lobe; ovary of an individual flower always bicarpellate, syncarpous or coherent, adjacent ovaries typically syncarpous or rarely fused at the base: ovules | to 9 per locule, strongly flattened, pendulous, imbricate; placentae apically attached to the septum, usually elongated, flattened, adnate to the septum: flo- ral nectary epigynous, embedded in the hypanthium between the base of the corolla tube and the apex of the ovary. Infructescences formed by multiple fruits, globose, with persistent calyx remnants. Individual fruits berry-like; endocarps hard and glossy or some- times fibrous and soft; disks accrescent, conspicuous, variable in size; seeds usually strongly flattened, sometimes concavo-convex, or plano-convex, ellip- soid, not alate or rarely with rudimentary wings, red: seed-coat reticulate or lineate. Annals of the Missouri Botanical Garden ^ — ~ TO BREONIA SPECIES Stipules of terminal vegetative buds flattened, complanate; inflorescences 4 to 8 per axil ›. В. decaryana Stipules of terminal vegetative buds conical, кане infor 'scences usually ] or rarely 2 per г anil. 2a. Lower surfaces of leaf blades always pubescent, tuft-domatia present in the axils of secondary and tertiary veins О. В. macrocarpa 2b. Lower surfaces of leaf blades always glabrous, tuft- domatia absent in the axils of secondary and tertiary ve за. Leaf blades 45 X 25 em; inflorescence axes 16-21 em long 1. B. madagascariensis ЗЬ, e blades less than 32 X 16 i › em; inflorescence axes less than 13 cm long. la. Stipules semi-persistent 19. B. tayloriana tb а deciduo За af blades with the 3 secondary veins diverging from the base of the midrib, base E late 3. perrieri 5b. Leaf blades without 3 secondary veins diverging from the base of the ШЕК, base noncordate. а. Bracts tubular, terminated by 3 or 4 broadly triangular lobes, never — the young inflorescence, persistent; adjacent ovaries fused at the base 4. B. нш Ob. Bracts e like. ешш; the young inflorescence, deciduous; adjacent ovaries comple le ly fused. Ta — ent c alyx "t completely fused: infructescences with barely evident yx remnants ». В. fragifera Adjac ent calyx tubes free: infructescences with well-developed calyx remnants. a. C silla — pubescent to puberulous, recurved. а. af blades with — c i ic-Lype domatia on lower surface, — n on upper surface; fertile peduncles densely pubescent; calyx pubes unique ch dilated above the middle and constricted at both ends 20. B. tsaratananensis Leaf TM without domatia; fertile peduncles wlabrous: calyx tubes funnel-shape Ob. 10а. Lower surfaces of leaf blades yellow-red-tinged when dry: calyx tubes 1-1.2 mm long; lobes 0.2-0.3 mm long, bearing shallow depressions on the center and a short protuberance n the edge toward the style 5. B. sambiranensis 10b. Lower surfaces brown-tinged when dry; calvx tubes 2.5-3 mm long: lobes 2-2.2 mm long, without alloy 6 'pres- sions, without protuberance 8b. Corolla lobes glabrous and not recurve Ila . B. boivinii nflorescence axes terete; ovaries m rent but not histologically 12a. Petioles I. 52 em long: secondary veins drying red-tingec 16. B. — 12b. — always more than 2 cm long; s sec ondary | veins ng vellow-tinged IIb. Inflore 'scence axes flattened; ovaries — st eie ie 13a. Stipules of terminal vegetative buds 15-22 mm one: ^ Leaf blades obovate, always wavy a п hy . € — В. havilandiana 14b. Leaf blades elliptic to lance latte, not wavy when dry 17. B. stipulata 13b. Stipules of terminal — tative buds 4-10 mm ling: 15a. Ovaries uniovule loa. Leaf blades 9-16 X 3.5-6.5 em long, ovate to oblong in shaj l . taolagnaroensis lob. Leaf blades 4 0 X 1.5-3.5 em Tem oblanceo- late in shape. 17a. Le af blade apices acute, эю cuneate; fertile peduncles 2—4 cm lon cc 3. B. louvelii 17b. Leaf blade apices санай; bases attenu- ale; fertile — les 4.5-6.5 em long . B. cuspidata 15b. Ovaries multiovulate 2 to 9 кита Pe r Р ule >). 18a. Petioles 3-6 mm long 2. B. membranacea 18b. Petioles at least 9 mm long. Volume 89, Number 1 Razafimandimbison Revision of Breonia 1. Hasen boivinii Havil., J. Linn. Soc. Bot. 33: 1897. TYPE: Madagascar. [Antsiranana province], District Nosy be [without exact lo- cality], Boivin s.n. (holotype, K!; isotype, P not seen). Shrubs, ca. 7 m tall. Bark rugose. Leafy stems terete, Stipules of terminal vegetative buds conical, 5— ca. 5 mm, glabrous. Leaves persistent; petioles (10—)15—25 X ca. 4 mm, terete, glabrous; blades 11.5-23 X 4.5-13.5 cm, ovate to broadly obovate, glabrous, coriaceous, glossy. apex base cuneate, lower surfaces glabrous. acute to rounded, brown-tinged when dry; margins glabrous, entire: secondary veins ca. 9 pairs per side, eucampto- dromous: tuft-type domatia in the axils of midribs or absent; stipules 6-9 X ca. 6 mm, cymbiform, 19a. Stipules terminal vegetative buds more u m long; inside of calyx tubes lox ule - ‚ B. chinensis п. ай 2—4 ovules per 19b. Stipules of ый уе — bude З тт long; inside of calyx tubes glabrous; тоге numerous ovules (7—9) per loc Ж . B. lowryi sometimes emarginate, not carinate, glabrous, free at the base, deciduous. Inflorescences solitary; heads 3.5-3.8 cm wide, including stigmas; inflo- rescence axes 2—4(—6.7) cm long, quadrangular to slightly flattened, glabrous; bracts calyptra-like, de- ciduous; peduncles (2.5-5)15-25 mm long. Flowers 5-merous; calyx tubes 2.5—3 mm long, prominently ribbed, glabrous, dilated above the middle, con- stricted at both ends, lobes 2—2.2 mm long, trun- cate, prominently ribbed, inside puberulous to to- mentose, outside glabrous, toward the base of lobes tomentose; corolla tubes 6-7 X ca. 0.5 mm, gla- brous; lobes 3—3.1 mm long, oblong, recurved, cil- iate, inside puberulous with a few scattered long hairs around the apex, outside glabrous; anthers ca. 1.5 mm long; filaments ca. 0.5 mm long, glabrous, 50° N 3c E А boivinii : 14? ` 14° © havilandiana * macrocarpa Ш рете 18° 18° 272° 229 0 200 Kilometers 50° 46° Figure 5. 5. Distribution of Breonia boivinii, B. havilandiana, В. macrocarpa, and B. perrieri. Annals of the Missouri Botanical Garden V b ^ “2 S CP pA UA Tus O Өе — M N Б SINS: D e d X (°С Figure 6. Breonia capuronit.—A. Fertile branch with infructescences. В. Fertile branch with inflorescences. — C. Seed: dorsal view (left); lateral view (right). —D. Two adjacent flowers, showing two adjacent ovaries separated: dissection through corolla and calyx tubes and an ovary of one flower (left); separate flower (right). —E. Median Volume 89, Number 1 2002 Razafimandimbison 15 Revision of Breonia * 14° А decaryana yk fragifera capuronii 200 Kilometers 46° Figure 7 flattened: styles 8-9 mm long, stigmas capitate: ovary 2-carpellate; carpels syncarpous; ovules 6 per locule, pendulous, imbricate; placentae flat- tened. Infructescences 3-3.5 cm diam., with well- developed calyx remnants; fruits with endocarp soft and fibrous: disks accrescent, obconical; seeds 4 to 6 per locule, strongly flattened, ellipsoid, white- tinged; seed-coat reticulate. Habitat and distribution. Low- and mid-alti- tude evergreen rainforests; Districts of Nosy be, Ambanja (Sambirano regions) and Vohémar (Fig. 5). Common name. — Valotro (which is reserved for Flowering November to December; fruiting January to February. Additional specimens — MADAGASCAR. Antsiranana: District Nosy be, Réserve Integrale de Lo- kobe, Antilahimena 19 (MO). үзе шин. 44 (МО 2751 RN (Р), Birkinshaw 67 (К, МО, Р, TAN); АК. T Distribution of Breonia capuronii, B. decaryana, and B. fragifera. 1432 SF (TEF); Doanilahy, 4905 RN (TAN), 6234 RN (ТЕЕ); District Ambanja, Baie d'Ambato, 23424 SF (TEF). Randrianaivo 247 (MO), a s.n. (P): District Vohémar, Canton Antsirabe- — Forêt d'Analamateza, 27593 SF (TEF), 27636 SF (TE 2. Breonia capuronii Razafim., sp. TYPE: Madagascar. | Antsiranana province], Massif de nov. Montagne d'Ambre, around “Station forestière des Roussettes,” ca. 800-1000 m, 15 Feb. (fl, young infr), 22059 SF (holotype. TEF). Figures 1C and 6 Haec species ad congeneros nonnullos loculis uniovu- latis accedit, sed ab eis petiolo longo (semper 2 cm ex- cedente) — foliorum nervis secundariis in sicco flavi- dis distinguitt Trees, 15-30 m tall. Bark white-tinged. rugose. Leafy stems terete, puberulous to pubescent, len- ticellate. Stipules of terminal vegetative buds con- ical, 4-5 X 1.5-2 mm, puberulous. Leaves decid- E dissection of developed post-anthesal mature flower showing calyx remnant, accrescent disk, and the single seed per locule (left); entire fruit (right). —F. Portion of branch apex with a mature infruc tescence, a stipule of terminal vegetative bud, and two petioles. A-D from 22059 SF (TE F) and E, F from 78518 SF (TEF). Annals of the Missouri Botanical Garden uous; petioles 22-35 X ca. 1 mm, terete, glabrous, lenticellate; blades 9.2—11.5 X 5.3-7.8 em, broad- ly elliptic to obovate, clustered near the stem api- ces, glabrous, membranaceous, glossy, apex acute to mucronate, base rounded; margins glabrous, en- tire; secondary veins ca. 8 pairs per side, eucamp- todromous, adaxially conspicuous, yellow-tinged; domatia absent; stipules ca. 4 mm long, cymbiform, not carinate, glabrous, free at the base, deciduous. Inflorescence solitary; heads ca. 2 em wide; inflo- rescence axes ca. 1.7 em long, twisted when dry, terete, densely puberulous; bracts calyptra-like, de- ciduous; peduncles ca. 1 mm long, densely pubes- cent. Flowers typically 4-merous; calyx tubes ca. 1 mm long, green-yellow-tinged, inside velutinous, outside pubescent, lobes 0.7-0.9 mm long, trian- 0.2-0.5 mm, glabrous; lobes ca. 1.5 mm long, oblong, gla- gular, pubescent; corolla tubes 2.8-3.0 X brous, toward the apex gradually puberulous, cili- ate; anthers ca. 0.5-1 mm long; filaments ca. 0.2 7.8-8 X ca. 0.1 mm, glabrous; stigmas globose or clavate; ovary 2- mm long, glabrous, terete; styles carpellate; carpels coherent; ovule 1 per locule, pendulous; placentae small, elongated. Infructes- cences 6-10 mm diam., with persistent calyx rem- nants; individual fruits with endocarp hard, glossy: disks accrescent, obconical; seed ] per locule, strongly flattened, ellipsoid, red; seed-coat lineate. Habitat and distribution. Mid-altitude ever- green rainforests and semi-deciduous forests: Dis- tricts of Antsiranana Il, Ambanja, Sambava, and Tsiroanimandidy (Fig. 7) Common names. Valodrano (Valotra growing in water), Valotro. Phenology. Flowering January-February: fruit- ing March-April. Discussion. This species is different from the other species of Breonia with a single ovule per locule by having long petioles (always more than 2 cm), and secondary leaf veins drying yellow-tinged. It has a large geographical distribution, but appears to be locally rare or perhaps under-collected due to its large size. The type species was collected within the Montagne d'Ambre National Park boundaries. The specific epithet honors Réné Capuron, a French botanist based Madagascar for several years in the mid 20th century who worked mainly on Malagasy woody plant families (including Ru- biaceae). Paratypes. MADAGASCAR. Antananarivo: District Tsiroanimandidy, Forêt d'Ambohijanahary, 18518 SF (Р, TEF ntsiranana: Montagne d’Ambre National Park, Malcomber et al. 1219 (MO); District Ambanja, Canton Ma- > Analamateza, 7906 RN (TEF), 12975 RN (TEF); Manongarivo, Gautier 3772 (G, MO, ТАМ); District Sam- ava, Canton Maroambihy, Andranomadiohely, 9002 RN (TEF). rovato, 3. Breonia chinensis (Lam.) Capuron, Adanso- nia, sér. 2, 13: 473. 1973. Cephalanthus chi- Méth. 1: 678. 1785. ). [Without Commerson s.n. (holotype, P- nensis Lam.. TYPE: exact locality |, Encycl. "Isle de France" (Mauritius). LA photo!). Figures ID and 2E. Méth. 4: 435. . Rich., Mém. шшс citrifolia Poir., in Lam., Encycl. 789. C Н € (Poir.) A * Rub ‚ Dec. Згеота citrifolia (Poir.) Ridsdale, pe 22: Er 1975. TYPE: Madagas- car. [Without exact um collector unknown s.n. (holotype, Р-ГА phot и Бенни Baill. Adansonia 12: 187 о" ric hardiana i l.) Havi Linn. Soc. ey . 1897. TYPE: Madagascar. r [W ithout ct loc — e der s.n. (holotype, P!) — coriacea. Havil., J. Linn. Soc. Bot. 33: A 1897. TY e E: an ipt ar. M ithout exact loo ality], Hum- i. (holotype, К!) Phat. mauritiana Mod. J. Linn. Soc. Bot. 33: 35. 1897. TYPE: Mauritius. [Without exact. loc АШУ]. Cümmarm s.n. (holotype, P!). Sarcocephalus richardii Drake, in A. Grandidier, Hist. Pl. denen ar 36: 457. 1897. prose richardii Pa- lacky, Cat. P nd ar 4: 50. YPE: Gran- didier А. 457 (not sec — Shrubs or trees, 7-25 m tall. Bark rugose. Leafy stems quadrangular, glabrous. Stipules of terminal 5—0 brous. Leaves persistent; petioles 9-15 X ca. vegetative buds conical, 1-1.3 mm, gla- mm, adaxially canaliculate, glabrous; blades 6.4— 2.4-5.5 cm, obovate, glabrous, coriaceous, not glossy, apex broadly to narrowly cuspidate, base cuneate; margins glabrous, entire; secondary veins 8 to 11 pairs per side, eucamptodromous, slightly prominulous; domatia absent; stipules 7-11 X 1.5- 2.5 mm, cymbiform, carinate, glabrous, free at the base, deciduous. Inflorescence solitary; heads 1— 2.5 cm wide; inflorescence axes 3.4—6.5 cm long, glabrous; bracts calyptra-like, deciduous; pedun- cles 1-2 mm long. Flowers 5-merous; calyx tubes ca. 1.5 mm long, inside lanate, outside glabrous, pubescent toward the base of the lobes, lobes ca. 0.5 mm long, truncate, lanate; corolla tubes 2.5—5 X 0.5 mm, glabrous; lobes 1.5-2 mm long, oblong, 8 mm long, glabrous, flattened; styles 7.5—9 abrous; anthers 1—1.5 mm long; filaments ca. 0.1 х 0.3- 0.4 mm, glabrous; stigmas clavate to capitate; ovary 2-carpellate; carpels syncarpous; ovules 2 to 3 per locule, pendulous, imbricate; placentae elongated. Infructescences 0.7-2 cm diam., with well-devel- oped persistent calyx remnants; individual fruits with endocarp hard, glossy; disks accrescent, ob- Volume 89, Number 1 2002 Razafimandimbison 17 Revision of Breonia 50° N 14°] 5 14° Ж chinensis 18° 18° 22° 22° 200 Kilometers 46° 50° Figure 8. Distribution of Breonia chinensis. conical; seeds 2 to 3 per locule, with rudimentary wings at both ends, strongly flattened, concavo-con- vex, ellipsoid, red; seed-coat reticulate. Habitat and distribution. Evergreen rainforests and mid-altitude humid forests, occasionally in dry- land sites and riverbeds; Districts of Antalaha, Sambava, Farafangana, Fort-Carnot, Ifanadiana, Manakara, Mananjary, Vangaindrano, Vohipeno, Brickaville, Mahanoro, Maroantsetra, Soanierana Ivongo, and Vavatenina (Fig. Common names. Molompangady (lips of Valo- tra), Valopangady (spade of Valotra), Valotra, Valo- potsy (white Valotra), Vavalotra, Voakiringy, and Voamalopangady. Phenology. Flowering August to October; fruit- ing November to March Discussion. Breonia chinensis is the most com- mon species of Breonia in Madagascar but grows only in evergreen rainforests. The epithet “chinen- sis” is not the most appropriate choice for Breonia, which is restricted to Madagascar; however, it has priority over the other available basionyms. Can- didate basionyms of B. chinensis were all consid- ered separate species of Breonia until Capuron (1973a), endorsed by Bosser (1984, 1999), clarified the identity of Cephalanthus chinensis Lam. as a separate species of Breonia. Homolle (1938) re- duced Breonia coriacea Havil. to synonymy under B. richardiana Havil. Breonia coriacea was not mentioned by Ridsdale (1975), but both species def- initely belong to Breonia chinensis based on their leaf blade shape and size, inflorescence axis shape, and ovule number. It has been accepted that the original VAN d of Nauclea citrifolia Poir. [= B. chinensis] is in marck's Encyclopédie (1789: 435) and that this spe- cies belongs to Breonia (Bakhuizen van den Brink, 1970; Ridsdale, 1975; Bosser, 1984, 1999). How- ever, Poiret did not specify type specimens of any of his species. This has caused some confusion as to which specimen he based N. citrifolia on and led to disagreement over the identity of the name. Richard (1830: 219) stated that his Cephalidium citrifolium (Poir.) A. Rich. was a new combination based on N. citrifolia Poir.: “Nauclea sp. Poir., En- cycl. méth., Species unica observata: Cephal- idium citrifolium, Nob. (Nauclea citrifolia, Poiret, . C.).” Like Poiret, Richard also did not cite the type specimens of any of his species. There is one specimen (lacking the collector's Annals o Missouri Botanical Garden name and number) with four labels in P-LA that may be the type of N. citrifolia Poir. The first label i Lamarck’s and says: “Nauclea citrifolia dic n° 2^ The second is Roepers and states: “? 8. Nauclea cadamba Roxb.? [and] Nauclea citrifolia Poir.!!!, per- quem affinis orientalis! immonimis affinis." The third states that this specimen returned to P-LA: Herbier de Lamarck, Novembre 1886." The fourth label is that of Capuron and states: “Breonia chinensis (Lamk.) R. Capuron (1973a) considered that this Lamarck specimen is the type of Nauclea citrifolia Poir. and is conspecific with Breonia chinensis (Lam.) Capuron. I concur with him because the protologue of N. citrifolia Poir. in Lamarck (1789: Ridsdale (1975) also endorsed Capuron's view, but made the new combination Breonia citrifolia (Poir.) Ridsdale because he did not recognize Cephalan- thus chinensis Lam. as a Breonia. Noteworthy is that Ridsdale did not cite the type specimen of N. trifolia Poir. Bosser (1984) disagreed with Capuron. arguing that an unnumbered Commerson collection from Madagascar in P-LA is the type specimen of N. citrifolia Poir. It is unclear whether Bosser re- ferred to the Commerson s.n. in P-LA that serves as the type of Cephalanthus chinensis Lam. In any case, this does not correspond to the protologue of N. citrifolia Poir. Despite Richards (Richard, 1830: 210) that his Cephalidium citrifolium |= Breonia chinensis| was based on Nauclea citrifolia Poir., Bosser (1984) argued that the name Cephalidium was not based on the same type as N. citrifolia and therefore should be cited as Cephalidium citrifol- ium A. Rich. Also, type specimen of C. citrifolium is another unnum- “acquis en Capuron.” 135) agrees with this specimen. statement Bosser further argued that the bered collection of Commerson now in P. This Com- merson s.n. at P collection is misidentified as N. citrifolia Poir., but it is actually a completely dif- ferent species of Breonia | describe here as B. tao- lagnaroensis. This P collection of Commerson has fruit only, whereas both Poiret's protologue of N. citrifolia and Richard’s protologue of Cephalidium citrifolium included flowers and fruits. Ridsdale (1975) treated Cephalidium citrifolium (Poir.) А. Richard under his Breonia citrifolia (Poir.) Rids- dale, but I here consider it as a synonym of Breonia — chinensis (Lam.) Capuron. MADAGASCAR “suivant le Ambo- е е alaha, SEES SF (ТЇ Additional m Antsiranana: District hitralanana, 9982 Ri iud TEF); ме ине Forêt de Bero, 12447 S jd (TEF); Am- Vinanivao, Forêt de Sahaen- N); Canton a Vi- D bodiazovola, pipes. jika, Bernard 327 (MO, TA b 238 SF — e Pare Masoala, Bernard 187 (MO, TAN); District Sambava, Canton Analamanara au Sud de Antsirabe-Nord, Vigreuse 15418 de 7 en Sambava and ; [Unknown locality |. Tartana, 27192 SF (TE ТЫ ( Fianarantsoa: District Farafangana, Karianga, Decary 5511 (PRE), Decary 5513 т) Dec ary 5573 (PRE): Efatsy, Forêt d? alia 15381 SF (TEF); Canton vain, SF (TEF), Aedes ^ Dis- Forêt d’ жашашып, M Forét d'Analila, 162 trict Fort-Carnot, ded 15287 SF (TEF); Canton Tolongoina, 5964 SF (TEF); Ambatomalama, 15518 (TEF); Canton Manampatrana, "exploitation forestiere" yit lonila- d FEE); District Ifanadiana, Ranomafana м p ‘tween Morafeno and Sahavanana, Turk " 643 ( TAN); Canton. Kianjavato, entre nalis Ned жу, 23917 SF (TEF); District Manakara, Fort E Andafa. 14728 SF (TEF); Forét de Manakara au d du terrain aviation, 86-R-118 (TEF); District Mus no, Ifanirea, vestiges de forét po 1 Nord de Vohipe no, 23691 (TEF); Tsararano, 6300 (TEF); District Vanno. Са ton Anosimparihy, ab ana, 14437 SF (TEF); Canton Mor- aleno, CES да. 16174 SF (TEF); Marofototra, For- êt de Mananjavara, 14712 SF (TEF); District Vaingain- drano, between Loy and агре 23661 SF TEF). Toamasina: District Brickaville, Managisy, 12352 SF (TEF); I district Mahanoro, Canton d' Ambodinanidilana | km à l'Est d'Ambodiala, Forêt d'Amboagibe, 19658 SF ТЕЕ); remnant native vege — 1-2 km E of eid ambo, Schatz et al. 3858 ‚ TAN); District Maroantse- tra, Canton. Andranofoy, Anata au bord de la ri- vière d'Andranofotsy, 21596 SF (TEF); Baie d'Antongil, bassin de — between TT and Beanana, 9047 SF National Park, Andrombazaha, Raha rion el ) 2410 SF (ТЕЕ); — os Il, Cant asa Mahatsara. 33240 SF (TE F), < А 711 SF (TEF), ` (TEF); Ambodiriana, 17299 SF (TEF): Vohi- marangitra, 3/809 (TE Ja Forêt d eet à l'Ouest de Foulpointe, 28087 SF (ТЕК); District Vavatenina, Canton Sahatavy, Maizinandro, 26703 SF (TEF). 4. Breonia cuspidata (Baker) Haviland, J. Linn. Soc. Bot. 33: 37. 1897. Nauclea cuspidata Bak- er, J. Linn. Soc. Bot. 25: 319. 1890. TYPE: Madagascar. "North West" [without exact local- ity]. Sep. 1887 (fl); Baron 5563 (holotype, K!). Trees. height unknown. Bark with annular fis- sures. Leafy stems rounded, glabrous. ‘Terminal vegetative buds conical, 4-5 X 1-1.2 mm, gla- 10-13 mm long. adaxially canaliculate, glabrous; blades 5.5- 10 х 1.9-3 em, oblanceolate, glabrous, membra- naceous, glossy, apex caudate, base attenuate; mar- brous. Leaves persistent; petioles c: gins glabrous, entire; secondary veins 8 to 9 pairs per side, eucamptodromous, adaxially inconspicu- ous, abaxially conspicuous; domatia absent; stip- ules ca. 5 mm long, cymbiform, not carinate, gla- deciduous. Inflorescence brous, free at the base, solitary, heads ca. 2 cm wide: inflorescence axes 4.5-6.5 cm long, strongly flattened, slender, gla- brous: bracts calyptra-like, deciduous: peduncles Flowers 5-merous: not present. calyx tubes ca. 1 Volume 89, Number 1 20 Razafimandimbison 19 Revision of Breonia mm long, free, inside lanate, outside up to the mid- dle glabrous, above the middle lanate; lobes ca. 0.2 mm long, triangular, lanate; corolla tubes ca. 5 X 0 lobes 1.1-1.3 long, glabrous; anthers 0.7-0.8 mm long, filaments mm; mm long, broadly ob- 0.2-0.3 mm long; styles 8—9 mm long; stigmas cla- vate to capitate; ovary 2-carpellate; carpels syncar- pous; ovule 1 per locule, pendulous; placentae re- duced, pendulous. Infructescences not seen. Distribution and habitat. Northwest Madagas- ar; habitats unknown Common names. Unknown Flowering September; fruiting time Discussion. Breonia cuspidata appears to have a restricted distribution and has not been collected since 1890. This species is distinguished from the other species by its ovaries with one ovule per loc- ule and its leaf blades with caudate apices and at- tenuate bases. Original paratypes. e locality]. Baron 6602 (K!) and Scott Elliott 2214 (K!). 5. Breonia decaryana Homolle, Bull. Soc. Bot. France 84: 460. 1937 [publ. 1938]. Neobreon- ia dec aryana (Homolle) Ridsdale, Blumea 22: 540. 1975 province], ріне Farafangana, Ifandana, De- :: Madagascar. [Fianarantsoa cary 5199 (holotype, P!; isotype, L not seen). Figures 1А and 2A. Breonia keliravina Homolle, Bull. Soc. Bot. — 84: 460. 1937 [publ. 1938]. TYPE: Madagascar. Ana- amazaotra, Thouvenot 91 (lectotype, es by P!). Ridsdale (1975: 546), Trees, 10-30 m tall. Bark rugose with annular fissures, rarely smooth, lenticellate. Leafy stems quadrangular, glabrous, always dichotomously branched. Terminal vegetative buds complanate, 8— 14 X 6-10 mm, glabrous. Leaves persistent; peti- oles 10-20 X 2—3 mm, adaxially canaliculate, gla- leaf blades (4—)6.5-12 X (1.5-)3-7.5 cm, oblanceolate to broadly obovate or broadly elliptic, brous; glabrous, coriaceous, glossy, apex rounded to broadly cuspidate, base cuneate to attenuate, ‹ - r rounded: margins glabrous, entire; secondary veins (6 or 7) to 10 pairs per side, eucamptodromous, adaxially inconspicuous, abaxially conspicuous: domatia absent; stipules 10—14 mm long, ovate to obovate, abaxially carinate, glabrous, free at the deciduous. base, Inflorescences 4 to 8 per axil; heads 1.5-1.8 cm wide; inflorescence axes 2.8—5 cm long, terete or slightly flattened; bracts calyptra- like. persistent; peduncles 0.8—2 cm long, sparsely pubescent. Flowers 4-merous; calyx tubes ca. 1 mm long, completely fused, lobes ca. 1 mm long, inside and outside densely pubescent; corolla tubes ca. 4 X | mm, red, glabrous, lobes ca. 1 mm long, broad- — ly oblong, yellow-tinged, glabrous; anthers ca. mm long. Styles 7-8 X 1-1.5 mm; stigmas globose: ovary 2-carpellate; carpel coherent; ovule ] per loc- Inf, Тах Сс С € ule. pendulous; placentae small. 0.5-1.7 em diam., rugose, with calyx remnants barely evident; individual fruits with endocarp hard, glossy; disks accrescent, rounded, deeply di- vided; seed 1 per locule, strongly flattened, red; seed-coat lineate. Habitat and distribution. | Evergreen rainforests, occasionally in riverbeds; Districts of Andapa, Mor- amanga, and Farafangana (Fig. 7). Common names. Molompangady keliravina. Marotsaka, Valompangady, Valotro, and Valotsy. Phenology. Flowering May to August; fruiting September to February. Discussion. Breonia decaryana can easily be distinguished from the rest of the Breonia species by its complanate terminal vegetative buds and 4 to 8 inflorescences per axil. This species was re- moved by Ridsdale from Вано simply because it has flattened terminal vegetative buds and partly fused corolla tubes. 1 here include it in Breonia because this species shares one morphological syn- apomorphy with Breonia sensu Ridsdale (large ac- crescent disks); additionally, they all have multiple fruits. | endorsed Ridsdale’s decision on sinking B. keliravina in B. decaryana because the former has all the diagnostic features of the latter. The speci- men Louvel 216 (P!) was one of the two cited by Homolle (1938) in the protologue of B. keliravina. Additional specimens — MADAGASCAR. Antsiranana: District Anda Réserve Spéciale d ig eit -Sud, Г Жалы Том SG 257 (MO, TA masina: District Mor ман Canton Anala- ао. s.n. (Р), Louvel 216 (P); Ambodi- voasary, 12-B-R-172 (TEF); he 26804 SF (TEF) 28414 SF (TEF), 28445 SF (TEF). — 6. Breonia fragifera Capuron ex Razafim., sp. nov. TYPE: Madagascar. Antsiranana prov- nod District Antalaha, Canton Ampanavoana, Antsiramoranga, 6 Dec. 1954 (fl), 6809 RN (holotype, TEF). Figures 2C and 9. Haec species a congeneris disco nectarifero accrescente permagno atque fructu pustulato distinguitur. Shrubs, spreading shrubs, or trees, 5-15 m tall. Bark rugose. Leafy stems terete, glabrous. Terminal vegetative buds conical, 3-3.5 X ca. З mm, gla- brous. Leaves persistent; petioles 25-70 X ca. 1.5 mm, terete, glabrous; blades 8-13 х 2.2-A(-6.5) 20 Annals of the Missouri Botanical Garden NYA Pbi! O ДУ Bi ri Figure 9. Breonia fragifera.—A. Fertile branch with mature infructescence. —B. Fertile branch with inflorescence. —C. Seed, showing flattened lateral profile (left); dorsal view (right). —D. Seed-coat texture. —E. Mature infructesc- ence. Е Median dissection through flower, showing velutinous calyx tube and unicarpellate ovary with single pen- Volume 89, Number 1 Razafimandimbison 21 Revision of Breonia cm, oblanceolate to ovate, glabrous, coriaceous or membranaceous, glossy, apex cuspidate to acute, yase cuneate; margin glabrous, entire; secondary veins 7 to 9 pairs per side, eucamptodromous: dom- atia absent; stipules ca. 3 X 2 mm, cymbiform, not carinate, glabrous, free at the base, deciduous. In- florescences solitary; heads 1.5-2 cm wide: inflo- rescence axes 1.2-3.5 cm long, terete; bracts ca- lyptra-like, deciduous; peduncles 2-3 mm long. Flowers mostly 4-merous; calyx tubes ca. 0.5 mm long, velutinous, lobes lanate; corolla tubes ca. 4 X l mm, white-tinged, glabrous; lobes ca. 1 mm long, oblong, ciliate or glabrous; anthers ca. 0.5 mm long: filaments ca. 0.1 mm long, glabrous. terete: styles ca. 7 X 0.2 mm, glabrous; stigmas capitate: ovary 2-carpellate; carpels coherent; ovules 1 per locule, pendulous. Infructescences 1.6-2 cm diam.. pusticulate, with calyx remnant barely evident; in- dividual fruits with endocarp soft, fibrous; disks ac- crescent, massive, obtriangular; seed 1 per locule, strongly flattened, ellipsoid, red; seed-coat lineate. Habitat and distribution. | Low-altitude forests: Districts of Ambanja, Antalaha, Fort-Carnot, Toa- masina IL, Fort-Dauphin (Fig. 7). Common names. Valotra, Valotralahy (male Valotra). Phenology. Flowering November—March; fruit- ing June to March. Discussion. This species is diagnosed by its massive, accrescent disks and pusticulate fruits. It was lumped by Ridsdale (1975) in Breonia sphaer- antha. The species name was taken from the labels of the herbarium specimens of Breonia fragifera re- ceived from TEF. The epithet “fragifera” indicates strawberries-bearing, referring to its inflorescences. Paratypes. MADAGASCAR. Antsiranana: District ABhanj n 251 (MO); Beampangibe, 2961 F (TEF): ara, е Gautier et al, 3288 a Mos а nton — i 7 SF (TEF); Presqu'le d’Ambato, foret classée, — etal. 324 (MO); District Antalaha, Canton Ampahana, An- dranomadio, 2/570 SF (ТЕЕ). Fianarantsoa: District Fort- — Canton lfanirea, 19791 | SF (TEF), — SF (TE — Ambodiriana 9045 RN (ТЕЕ); virons E E Baie d'Antongil, Massif d'Antsirosiro, 87 34 z 3 =a E T SF (TEF); Toamasina Il, Ampasimbe, Andranotsara, SFF Mihatsara. — i. SF (T EF). Comtet 33536 SF (TEF). Com aput pd Comtet 34413 SF (ТЕР), Noyes et al. p . MO, P, TAN); Réserve Naturelle In- tégrale de ия — Sv е el al. 259 (TAN). t Fort-Dauphin, 5 (P. TEF). Урда кыд p (MO); Canton rate sa, Antsako, 10803 SF (ТЕЕ); Isaka-Ivondro au bords de la rivière Kovazaza, 11503 SF (TEF). 7. Breonia havilandiana Homolle, Bull. Soc. Bot. Fr. 84: 464. 1937 [publ. 1938]. TYPE: Madagascar. [Fianarantsoa province]. “bords de l'Anosivola (Mangoro)." 700 m, Sep. 1911 (fl), Perrier de la Báthie 3904 (holotype, P!). Trees, 10-20 m tall. Bark fissured longitudinally. Leafy stems glabrous, lenticellate. Terminal vegeta- 15-19 3.5—4 mm, Leaves deciduous; petioles 15-25 X ca. 2 mm, te- rete, glabrous; blades 8-19(-21) х 5-9(-12) ст, ob- ovate, always rippled or wavy when dry, glabrous, tive buds conical, glabrous. coriaceous, apex obtuse to rounded, base cuneate; margins glabrous, entire; secondary veins 8 to 10 pairs per side, eucamptodromous, abaxially promi- nent; ciliate-type domatia; stipules ca. 20 X 3 mm. cymbiform, abaxially carinate, glabrous, free at the base. deciduous. Inflorescence solitary, rarely 2 per axil, heads ca. 2.3-2.9 em wide; bracts calyptra-like: inflorescence axes 3—4 cm long, glabrous: peduncles ca. 34 mm long, glabrous. Flowers 5-merous; calyx tubes ca. 3.5 mm long, inside at the base velutinous, toward the lobes pubescent, outside glabrous. a few straight long hairs surrounding the base, lobes ob- long. tomentose; corolla tubes ca. 5 X 0.6—0.7 mm, glabrous, lobes 15-2.5 mm long, oblong: anthers 1.1-1.2 mm long; filaments ca. 0.1 mm long. gla- styles 8.5-9 brous; stigmas capitate, shallowly bifid; ovary 2-car- brous, flattened: ca. 0.2 mm, gla- pellate; carpels syncarpous; ovules 2 to 4 per locule, flattened, pendulous, imbricate; placentae flattened, elongated. Infructescence 1.7-2.5 cm diam., with well-developed calyx remnants; individual fruits with endocarp soft, not glossy; disks accrescent, obconi- cal; seeds 2 to 4 per locule, strongly flattened, ellip- sold, red; seed-coat reticulate. Habitat and distribution. Low- and mid-alti- tude evergreen rainforests; Districts of Ifanadiana, Befandriana-Nord, Moramanga, Fort-Carnot, and Anosibe an'Ala (Fig. 5). Common names. Molompangady, Molotr'angady (lips of spade), Mamalifangady. Phenology. Flowering October to November: fruiting September to January. Discussion. This species is distinctive in having leaf blades that are undulate when dry. Additional examined. MADAGASCAR specimens — dulous ovule each (left); entire flower (right). — edian dissection through infructescence revealing — disk and uniovulate carpels (left); entire fruit (right). ^. E F from 6809 RN (TEF) and C-E, G from 79775 SF (TE Annals of the Missouri Botanical Garden = louvelii + richardsonii 14° ж sphaerantha A stipulata Ж taolagnaroensis © tayloriana 50° ү Е 14° $ 18° 22° 200 Kilometers 46° Figure 10. B. tayloriana. ‘ianarantsoa: District Fort-Carnot, Canton Tolongoina, Ankadilalana, 19295 SF (TEF); Forêt d’ Ambutohareaha. 7136 SF (TEF); District Hanadiana, lo — al Park, Randrianasolo Rasabotsy 32 (MO), 9974 SF (TEF), Бакайата Нов SG 389 (MO, TAN). оь ga: District Befane dri iana- Nord, eb Matsondakana, Be- Héserve Spéciale d' Anjanaharibe- Sud, Ravelona- rivo € Rabesaonina 570 ). Toamasina: District Anosibe an'Ala, Tsaratampona, Mangorobilika, 25526 SF (TEF); District Moramanga, Canton Périnet, Sahamaloto, 6105 SF (TEF & lalona, — 8. Breonia louvelii Homolle, Bull. Soc. Bot. France 84: 461. 1937 [publ. 1938]. TYPE: Madagascar. [Toamasina province], Analama- zaotra, Sep. 1925 (fl), Louvel 125 (holotype, P!). Medium-sized trees, 7-10 m tall. Bark with lon- gitudinal fissures. Leafy stems rounded, glabrous, lenticellate. Terminal vegetative buds conical, 5— 1-1.5 mm long, glabrous. Leaves persistent; pet- mm long, adaxially canaliculate, gla- ioles ca. 7—17 ^ brous; blades ca. 4-9 X 1.5-3.5 em, oblanceolate to elliptic, glabrous, coriaceous, glossy, apex acu- minate to obtuse, base attenuate; margins glabrous, entire, slightly involute when dry; secondary veins 10 to 11 pairs per side, eucamptodromous, adaxi- 50° Distribution of Breonia louvelii, B. richardsonii, B. sphaerantha, B. stipulata, B. taolagnaroensis, and ally inconspicuous, abaxially conspicuous; domatia absent; stipules 5-9 X 1.5-1.8 mm, cymbiform, not carinate, glabrous, free at the base, deciduous. In- florescence solitary, heads 1.8-2 cm wide; inflores- cence axes 2—4 cm long, flattened, slender, gla- brous; bracts calyptra-like, deciduous; peduncles ca. 2 mm long, glabrous. Flowers 5-merous or rare- ly 6-merous; calyx tubes ca. 1 mm long, free, lobes ca. 1 mm long. inside lanate, outside glabrous, lobes densely pubescent, truncate to oblong: corolla tubes 56 X long. broadly oblong; anthers 1—1.5 mm long; fila- ments 0.2-0.3 mm long; styles 7-8 X 1-1.5 mm; stigmas capitate (shallowly bifid); ovary 2-carpel- 1-1.2 mm, glabrous, lobes 2-3 mm ate; carpels syncarpous; ovule 1 per locule, pen- dulous; placentae small. Infructescences not seen. Habitat and distribution. | Evergreen rainforests; District of Moramanga (Fig. 10). Common names. | Hazomarotsaka (slender but tall trees), Molompangady. Phenology. Flowering September to October: fruiting December to January. Discussion. This species has a very restricted geographical distribution and appears to be rare. Volume 89, Number 1 Razafimandimbison Revision of Breonia Additional e MADAGASCAR. Toamasina: District Moramanga, Canton Périnet, Antan- on dm 7921 SF (TEF); "ire УЛ 7562 SF (TEF); Ré- rve —— d'Analamazaotra, Razafimandimbison SG 358 (MO, TAN). spec imens 9. Breonia lowryi Razafim., sp. nov. TYPE: Mad- agascar. Fianarantsoa province: Ambalavao, Andringitra Reserve, Lewis et al. 755 (holotype, MO 05066234; isotypes, К, TAN). Figure 11. iec species a congeneris loculis corollinis glabris, ubo calycino intus glabro atque loculis 7- ad 9-ovulatis distinguitur Trees, 10—30 m tall. Bark rugose. Leafy stems quadrangular, glabrous. Terminal vegetative buds ca. 3 X 2 mm, glabrous. Leaves persistent, petioles 10-14 mm long, adaxially canaliculate, glabrous: blades (5.5—)7.8-12 2.4—)3.4—6.4 cm, glabrous, coriaceous, not glossy, apex broadly cus- pidate, base attenuate; margins glabrous, entire: secondary veins 6 to 9 pairs per side, eucampto- dromous, abaxially conspicuous; pit-type domatia in the axils of the secondary veins, pubescent; stip- ules ca. 5 X 1-1.2 mm, cymbiform, not carinate, glabrous, free at the base, deciduous. Inflorescenc- es solitary, heads 2-2.3 cm wide; inflorescence axes ca. 1.8-3 cm long, flattened, glabrous; bracts ca- lyptra-like, deciduous; peduncles ca. 2 mm long, glabrous. Flowers 5-merous; calyx tubes ca. 2 mm lobes obovate, long. inside glabrous, outside puberulent, 0.5-0.8 mm long, oblong, inside glabrous, outside velutinous; corolla tubes ca. 6 X 0.4 mm, reddish, inside puberulous, outside glabrous, lobes 2-2.2 mm long, oblong, puberulous; margins glabrous; anthers 0.5-0.8 mm long; filaments ca. 0.2 mm long, flattened; styles « 0.6 mm; stigmas elongate to capitate; ovary 2-carpellate; syncarpous; ovules 7 to 9 per locule, pendulous; placentae elongated. 1-1.5 ст diam., with well-developed calyx remnants; indi- vidual fruits berry-like; t, fibrous: disks accrescent, rounded; seeds with rudimentary carpels Infructescence endocarp soft, 7 to 9 per locule, strongly flattened, wings, red: seed-coat reticulate. Evergreen forests be- Districts of Fort- Habitat and distribution. tween 900 and 1500 m altitude: Carnot, Moramanga, Ambalavao, and Fianarantsoa (Fig. 12). Common names. Molompangady, Valompangady. Phenology. Flowering time unknown: fruiting May and October. Discussion. This species differs from the other Breonia species that also have 7 to 9 ovules per locule and glabrous corolla lobes because the in- side of the calyx tubes is glabrous. The epithet of this new name honors Porter P. Lowry Il, head of the Africa and Madagascar Department of the Mis- souri Botanical Garden, who has done much to fur- ther studies of the Araliaceae as well as other plant families in Madagascar. Paratypes. MADAGASCAR. енли District Fianarantsoa Il, Canton Ampamaherana, 2078 SF (TEF): District Fort-Carnot, To E aes ee аз SF (TEF), 11594 SF (TEF), 14-B-R-230 (TEF asina: District = йы Forêt d’ Fede Périnet), 24150 SF (TEF). 10. Breonia macrocarpa Homolle, Bull. Soc. Bot. Fr. 84: 461. 1937 [publ. 1938]. TYPE: “Madagascar,” Pérrier de la Bathie 3933 (lec- totype, designated by Ridsdale (1975: 545). P!). Figures 2D and 3 Trees, 10—20 m tall. Bark longitudinally fissured. eafy stems quadrangular, glabrous. Terminal veg- etative buds conical, 2740 X 4-6 mm, puberu- lent. Leaves persistent; petioles 35-60 X ca. 4 mm, terete, lenticellate, puberulous; blades 19.5-38 x 15-25 em, broadly ovate to broadly oblanceolate. brown-tinged when dry, adaxially glabrous, abaxi- ally sericeous, coriaceous, glossy, apex cuspidate to rounded, base rounded; margins glabrous, entire: secondary veins 12 or 13 pairs per side, eucamp- todromous, abaxially densely pubescent, adaxially glabrous: tuft-domatia in the axils of secondary and tertiary veins; stipules 27-34 X 6-9.5 mm, cym- biform, abaxially carinate, puberulous, free at the Inflorescences solitary, heads ca. base, deciduous. 4.7 cm wide; inflorescence axes 8.5-12.5 cm, flat- tened, glabrous to puberulous; bracts calyptra-like. deciduous; peduncles not elongated. Flowers 5- merous or rarely 4-merous; calyx tubes 2.8—3 mm long, green, below the middle velutinous, around the mid-part glabrous, toward the lobes puberulous: lobes ca. 2 mm long, truncate to oblong, velutinous: corolla tubes ca. 11 mm long, glabrous, lobes ca. 3.5 X 0.25 mm long, oblong, glabrous. alado toward the apices; margins ciliate; anthers ca. mm long; filaments 0,5—1 mm long, flattened. Styles 21 X 0.5 mm, glabrous; stigmas clavate; ovary 2-carpellate; carpels syncarpous: ovules 4 to 6 per locule, flattened, pendulous, imbricate; placentae strongly flattened, elongated, adnate to the septum. Infructescences 3—4 cm diam., woody when dry, with well-developed calyx remnants; fruits with endocarp hard, glossy; disks accrescent, pentagonal; seeds 4 to 6 per locule, strongly flat- tened, ellipsoid, red, reticulate. individual Phenology. Flowering March to May, August to 24 Annals of the Missouri Botanical Garden September; fruiting May to July, December to Jan- uary. Habitat and distribution. tude of eastern evergreen rainforests; Districts of Brickaville, Moramanga, and Toamasina П (Fig. 5). Common names. рар заду апа Valotra. Discussion. Ridsdale (1975: 545) included Breonia macrocarpa as a synonym of В. madagas- Low- and mid-alti- cariensis, dismissing the former as "a hairy-leaved form" of the latter. However, these two are quite different morphologically, as is clear from Table 3. Despite the striking morphological differences be- tween the two species, B. macrocarpa has not been recognized since its original description, partly be- cause Homolle (1938) did not provide kevs for the species of Breonia she recognized and partly be- cause no collections of B. madagascariensis have been made for almost 200 vears. While describing this species, Homolle cited three specimens, Thouvenot 117 (P!), Perrier de Báthie 3933 (P!), and Perrier de la Báthie 5270 (P!), without selecting the type species. Additional examined. MADAGASCAR. Toamasina: District Brickaville, Fetromby, Ambalakon- 3265 SF (TEF); District Moramanga, Perinét, 26906 F (TAN), Noe el i Mg K, MO, P, TAN), Razafiman- pd төз SG 352 Р, TAN): District Tenia П. Canton Ambodiriana, аад, 9052 RN (Р), 91: RN (Р); Canton Ambodiriana, SF (P, TEF), 21160 SF (TEF), 28136 SF (TEF), 32344 SF (TEF). specimens m N © 11. Breonia madagascariensis A. Rich., in DC., Prodr. 4: 620. Sep. 1830. Sarcocephalus madagascariensis (A. Rich.) Baill., Adansonia 12: 311. 1879. TYPE: “Madagascar.” [Without exact locality], Chapelier s.n. (lectotype, des- ignated by Ridsdale (1975: 545), K!; totype, Р!). Figure ТЕ. isolec- Trees, height unknown. Bark rugose. Leafy stems quadrangular, glabrous. Terminal vegetative buds conical, ca. 35 X 6 mm, glabrous. Leaves persistent, x 35 glabrous, subsessile; blades ca. 45 cm, broadly obovate to broadly spathulate, coriaceous, not glossy, apex caudate, base attenuate; midrib angular, lenticellate, prominent; margins glabrous, entire; secondary veins 18 or 19 pairs per side, eucamp- todromous, abaxially prominent; domatia absent; stipules ca. 35 mm long, cymbiform, abaxially car- inate, glabrous, free at the base, deciduous. Inflo- rescences solitary, heads ca. 6 ст wide; inflores- cence axes 16-21 cm long, flattened, puberulous; bracts calyptra-like, deciduous; peduncles not elon- gated. Flowers 5-merous; calyx tubes ca. З mm long, with prominulous ribs, inside lanate, outside gla- brous except on upper parts of ribs, lobes ca. 2 mm long. oblong to truncate, pubescent; corolla tubes c: З х 1 mm, oblong, inside glabrous, outside puberulous in 1.5 mm, ribbed, glabrous, lobes са. upper third, margins ciliate; anthers ca. 2 mm long; filaments 0.25—0.5 mm long, flattened; styles 17-18 X са. 0.25 mm, glabrous; stigmas clavate; ovary 2- carpellate; carpels syncarpous; ovules 8 per locule, pendulous, imbricate; placentae strongly flattened, elongated. Infructescences not seen. Habitat and distribution, Habitat unknown: Madagascar. Phenology. Unknown. Discussion. Breonia madagascariensis is known only from the six authentic herbarium specimens collected by Commerson, Chapelier, and Bréon during the early botanical explorations they under- took separately in 1770-1771, 1794-1808, and 5 September-5 December 1818, respectively (Dorr, 1997). No collecting localities were mentioned on their labels. This species has not been collected since 18 Chapelier is known to have resided on the East Coast at Foulpointe for a dozen years, and he es- tablished an experimental garden in a place known locally as Isatrano (Tamatave or Toamasina). Also, he made botanical collections around Ivoloina, Foulpointe, and on the island of St. Marie (all with- in Toamasina province). Hence, it is possible that Breonia madagascariensis was collected from the above-mentioned areas of the east coast of Mada- gascar. Most of the eastern coast of the country has been deforested for agricultural purposes and re- placed by secondary vegetation. characterized. by the dominance of Ravenala madagascariensis Sonn. (Strelitziaceae). However, B. madagascariensis may still be extant inside protected forests (Betampona Reserve, Zahamena Reserve, Analamazaotra Re- serve) in this area, so efforts to rediscover В. mad- agascariensis should focus on these three protected areas and their surroundings. => Figure 11. Breonia lowryi.—A. Fertile brane i with TUE infructescence. —B. Mature infructescence. С. Seed, showing tate ne E — ral profils (left); dorsal view (right). —D. Median dissection through fruit, showing calyx tube and bicarpellat ith 7 to 9 pendulous ovules th r locule ic lo entire fruit (right). —E. Two adjacent flowers, showing two dH. ovaries separated: median dissection through flower, showing glabrous corolla amd pubescent calyx tubes and bicarpellate ovary with ovules (left); entire flower, showing only part of corolla tube (right). ). Lewis et al. 755 (MO) Volume 89, Number 1 2002 Razafimandimbison Revision of Breonia Annals of the Missouri Botanical Garden Table 3. separating Breonia macrocarpa and В. madagascariensis. Distinctive vegetative morphological features Breonia В. madaga 5- Characters macrocarpa cariensis Petioles well developed subsessile Leaf blade size 19-38 X 15-25 cm 45 X 35 cm Abaxial indumentum densely pubescent glabrous on leaf blade Midrib shape rounded angular Domatia present absent Peduncle length 8.5-12.5 cm 16-21 cm The lectotype Chapelier s.n. selected by Ridsdale (1975) was from the six known collections. Additional sj ns examined. “Madagascar, Herbier Drake, P- 001. 3910 (Py: "Madagascar, herbarium Ri- d P-001 а (P). Bréon s.n. (P). Commerson s.n. (P not seen). J. Linn. Soc. "Madagascar. 12. Breonia membranacea Havil.. Bot. 33: 36. 1897. TYPE: [Without exact locality], Perottet s.n. (lecto- type, designated by Ridsdale (1975: 545), Р!). Trees, 12-18 m tall. Bark rugose. Leafy stems te- rete, glabrous. Terminal vegetative buds conical, 10— 20 X ca. 1.5 mm, puberulent at the base. Leaves persistent; petioles 3-6 mm long, adaxially canalicu- 16, puberulous when young, glabrescent; blades 8.5— { 4.5-5.4 ст. oblong, glabrous, membrana- ceous, adaxially glossy, apex broadly cuspidate to acute, base rounded; margins glabrous, entire; sec- ondary veins ca. 8 pairs per side, eucamptodromous, adaxially prominulous; domatia hairy in the axils of the secondary veins; stipules 12-21 mm long. cym- biform, uous. Inflorescences solitary, heads ca. 2.2 cm wide: not carinate, glabrous, free at the base, decid- inflorescence axes 4.5—5.5 em long, strongly flattened. glabrous; bracts calyptra-like, deciduous; peduncles 3 mm long. glabrous. Flowers 5-merous: calyx tubes 1—1.2 mm long, inside lanate, outside glabrous. lobes 0.5-0.7 mm long. oblong, velutinous: corolla tubes ca. 5 X 0.50.7 mm, inside puberulous, outside glabrous, lobes ca. 2 mm long, obtuse to truncate, puberulous, margins ciliate: anthers ca. 1 mm long: filaments ca. 2 mm long, flattened, glabrous; styles 7— 8 X ca. 0.5 mm, glabrous; stigmas capitate; ovary 2- carpellate; carpels syncarpous; ovules З per locule. flattened, pendulous, imbricate; placentae small, elon- gated. Infructescences 1.5-1.8 cm diam., with well- developed calyx remnants; individual fruits with en- docarp soft, fibrous; disks accrescent, heart-shaped: seeds 1 to 3 per locule, flattened, concavo-convex, ellipsoid, red; seed-coat reticulate. Habitat and distribution. Evergreen lowland rainforests, littoral forests; Districts of Moramanga and Fénérive Est (Fig. 12). Common names. Molompangady, Valotra. Phenology. Flowering January February; fruiting March to April. Discussion. This species is distinct in its leaves with relatively short petioles. The lectotypification made by Ridsdale is questionable because the lec- totype, Perottet s.n., does not match with Haviland's (1897) remarks concerning the terminal vegetative buds of B. membranacea: “| have seen none [spec- imens] with the apex uninjured." Perrottet s.n. has an intact and uninjured apex and was annotated by Haviland as Breonia membranacea. Haviland clear- ly stated that the type specimen of B. membranacea was in the Paris herbarium. Haviland mentioned that the description of Breonia mauritiana was based on a single specimen: he did not mention that this was also the case for B. membranacea. Hence, he could possibly have seen more than one specimen of B. membranacea, although he did not cite any other specimens in the original description. Haviland is known to have visited the herbaria at the British Museum (BM), Leiden (L), and Paris (P) (Haviland, 1897). However, from Kew (K). L. and P, I received only the specimen of Perrottet s.n. from P. It is possible that the "injured" specimens seen by Haviland are at BM. Ridsdale visited all three herbaria while preparing his revision of Af- rican and Madagascar Naucleeae (Ridsdale, 1975); he did not cite any syntype for B. membranacea. This implies that he saw only the specimen of Per- rottet s.n. at Paris. Before one can attempt to resolve the nomenclatural problem of B. membranacea, one must visit BM, L, and P and look for possible spec- imens of B. membranacea annotated by Haviland. Additional specimens о ed. MADAGASCAR Toamasina: District. Fénérive-Est — Tampolo, 13094 SF (TEF), 16496 SF (TEF). 16621 SF (P); District Moramanga, Forêt d'Analamazaotra, Louvel 186 (P). 13. Breonia perrieri Homolle, Bull. Soc. Bot. Fr. 84: 461. 1937 [publ. 1938]. TYPE: Mad- agascar. [Toliara province]: Morondava, Forét 3513 designated by Ridsdale (1975: de Marofandilia, Perrier de la Báthie (lectotype, ) 546), Medium-sized trees, ca. 8 to 15 m tall. Bark with longitudinal fissures. Leafy stems flattened, lenticel- ate, glabrous. Terminal vegetative buds conical, 9— 11 X 34 mm, glabrous. Leaves deciduous: petioles 33-60 X 1.5-1.8 mm, terete, glabrous or pubescent toward the apex, lenticellate; blades 9-15 X 6-1: Volume 89, Number 1 2002 Razafimandimbison 27 Revision of Breonia 50° N [=] lowryi He: 14° Ш membranacea d 14? @ sambiranensis * tsaratananensis 18° 18° 22? 229 200 Kilometers 46^ Figure 12. cm, broadly obovate to orbicular, pubescent or gla- brous, coriaceous or rarely membranaceous, apex cuspidate to rounded, base cordate; margins gla- brous, entire; secondary veins ca. 9 pairs per side, the first 3. diverging from the base of the midrib, prominulous; domatia absent; stipules 8-14. X 34.2 mm, cymbiform, abaxially carinate, glabrous, free at the base, deciduous. Inflorescences solitary, heads 2.5-2.0 cm wide; 2.4-3.5 cm long, terete, pubescent or glabrous; bracts calyptra- inflorescence axes ike, deciduous; peduncles 2—3 mm long. Flowers 5- merous; calyx tubes ca. 2.5 mm long, green, inside below the middle glabrous, velutinous from the mid- dle toward the base of lobes, outside glabrous, lobes 1—1.2 mm long, oblong, puberulous; corolla tubes 4— 5 X ca. 1 mm, yellow-tinged, glabrous, lobes ca. 1.5 mm long, oblong, glabrous; anthers 1-1.1 mm long; filaments ca. 1 mm long, flattened, glabrous; styles х 0.2 vate; ovary 2-carpellate; carpels syncarpous; ovules mm, glabrous; stigmas globose to cla- 2 to 3 per locule, strongly flattened, pendulous, im- bricate; placentae small, elongated. Infructescences 8-2.2 cm diam., with well-developed calyx rem- nants; individual fruits with endocarp hard, glossy; disks accrescent, heart-shaped; seeds 1 to 2 per loc- 50° Distribution of Breonia lowryi, B. membranacea, B. sambiranensis, and B. tsaratananensis. ule, with rudimentary wings at the base, strongly flat- tened, ellipsoid, white-tinged; seed-coat reticulate. Phenology. Flowering November January; fruiting January to March. Habitat and distribution. Western and. north- western deciduous dry forests; from Districts of Antsiranana П and Ambato-Boeni, Befandriana- Nord, tirano, Marovoay, Antsalova, Belo Tsiribihina, Main- Ambilobe, and Vohémar (Fig. 5). otra. Soalala, Common name. Val Phenology. fruiting January—February. Flowering November—December; Discussion. The following were among the four specimens cited by Homolle in the protologue of Breonia perrieri: Perrier de la Báthie 825 (Р not Perrier de la Báthie 17852 (P not seen Greve s.n. (P not seen). . and м seen), Additional specimens examined. MADAGASCAR, An- tsiranana: District Antsiranana 5 Ankarana, Ambiloman- godro, Forêt de — 29653 SF (TEF); eta Ré- serve Spéciale d'Ankarana, Malcomber et al. 1833 (K, MO, P, TAN), Rasafimandirdicon SG 273 QB 2s Forét a Ambis, 6181 SF (TEF), 6616 SF (TE Е); Plateau d'Ankarana à l'ouest d'Ambondromifehy, 3026 SF (K, TEF); District Vohémar, Forét d’Analafiana au nord de la base Annals of the Missouri Botanical Garden Manambery, 27503 SF (TEF); 14 km E of Vohémar, near Analafiana, 15673 SF (TEF); Ankara, 6696 SF (TEF). Ma- hajanga: District Ambato-Boeni, Ankirihitra, Forêt d'Anatialabe, 19379 SF (ТЕЕ); Di: MN Befandriana-Nord, Ambahivahy, Forêt d'Andembikely, 79061 SF (TEF); кон êt de Marohogo à l'ouest de — 18450 SF (TEF): Di trict — a 3 km à FF Bevitika, Labat & Conté 2674 (K, › TAN), — et * 2261 (К. MO, P. TAN): Plateau de zh ко aux environs de — 6754 5 (TEF), Jongkind et a 3412 (K, MO. P, ). Villiers et al. 4846 (MO. P, TAN); District Soalala, че, Andrano- 2 SF PUN Andranomavo, 17565 SF (TEF). To- B — Belo d Tsimafana, Ankirijifotsy, 7987 SF (TEF), Noyes et al. 1028 (K, MO, P, TAN); District те lava, Tsimembo-limite concession de M. Barthe, 8245 SF (TEF); Ankilatsy, 4677 SF (TEF) 14. Breonia richardsonii Razafim., sp. nov. TYPE: Madagascar. [Toamasina province]: Ma- roantsetra, Jardin botanique de Farankaraina, 16 June 1965 (fl), 14.359 SF (holotype, TEF). Figure 13. — me aec species a congeneris inflorescentiae bracteis tub- ulosis adpressis in | triangulares tres quatuorve desinen- tibus atque ovariis contiguis post anthesin distinguitur. Trees, 15-30 m tall. Bark fissured longitudinally. Leafy stems terete, glabrous. Terminal vegetative buds conical, 34 X 1-1.5 mm, glabrous. Leaves persistent; petioles 7-10 X ca. | mm, terete, gla- brous: blades 5-7.7 X 2.5-3.7 em, oblanceolate to ovate, glabrous, membranaceous, apex cuspidate to acute, base acute; margins glabrous, entire; second- ary veins 5 to 7 pairs per side, eucamptodromous, inconspicuous; domatia cryptic-type in the second- ary veins, glabrous, adaxially prominent; stipules ca. 5 X 1-1.5 mm, eymbiform, not carinate, free at the base, deciduous. Inflorescence solitary, heads ca. 2 cm wide; inflorescence axes 2.5—3 cm long, terete, glabrous: peduncles ca. 5 mm long; bracts tubular with 3—4 broadly triangular lobes, not enclosing the young inflorescence, persistent. Flowers 5-merous or rarely 4-merous; calyx tubes ca. 1 mm long, inside glabrous, except at the base velutinous, outside at the base velutinous, in the mid-parts glabrous, to- ward the lobes puberulous, lobes broadly triangular, 0.3-0.5 mm, gla- 1.5 mm long, oblong, glabrous: an- puberulous; corolla tubes са. 4 X brous, lobes Ca. thers 2-2.5 mm long: filaments ca. 0.1 mm long. flattened: styles ca. 7 X 0.5 mm, glabrous; stigmas globose to clavate: ovary 2-carpellate, adjacent ova- ries fused only at the base; carpels syncarpous; ovules 5 per locule, pendulous, imbricate: placentae large, elongated, occupying the upper third of the locule. Infructescences 1—1.5 cm diam., with well- developed calyx remnants, adjacent ovaries fused up to the mid-point; individual fruits with endocarp soft, fibrous; disks accrescent, obconical; seeds 5 per loc- ule, strongly flattened, ellipsoid, angular, red; seed- coat reticulate. Habitat and distribution. Lowland rainforests; District of Maroantsetra (Fig. 10 Common names. — Valotrafotsy. Phenology. Flowering April; fruiting November. Discussion. This species is very different from other species of Breonia because of its tubular, ap- pressed inflorescence bracts with 3 to 4 broadly triangular lobes and post-anthesis fusion of adja- The specific. epithet honors. Mick cent ovaries. Richardson, who was my Ph.D. advisor as well as the Manager of the Graduate Program at the Mis- souri Botanical Garden. This species is extremely rare and has not been collected since 1965. Paratype. MADAGASCAR. Toamasina: District Ma- roantsetra, Jardin botanique de Farankaraina, 18311 SF (TEF) 15. Breonia sambiranensis Razafim., sp. nov. TYPE: Madagascar. [Antsiranana province]: Nosy be, Réserve Naturelle Integrale de Lo- kobe, 13°24'S, 48?20'E, 350 m. 7 July 1995 (fl), Antilahimena 237 (holotype, TAN; isotype, MO). Figures 3C, 4D, and 14. species a congeneris calycis tubo brevi аши costato et lobulis brevibus in centro non profunde foveo latis, directas gerentibus distinguitur ad margines protuberationes breves stylum versus (6-10-20 m tall. Bark smooth, glabrous. Leafy stems quadrangular, gla- Trees, rarely shrubs, brous, lenticellate. Terminal vegetative buds coni- cal, ca. 5 X 2-3 mm, glabrous. Leaves persistent; үк 15-30 mm long, adaxially canaliculate, glabrous; blades (8.5-)]2-17.5 X 4.7-9.5 cm, broadly ovate to broadly obovate, or broadly ovate to narrowly ovate, glabrous, coriaceous, glossy, apex mucronulate to obtuse, mucronulate to acute or - base cuneate, lower surfaces vellow-red-tinged when dry; margins entire, glabrous; secondary veins 5 to 9 per side, eucamptodromous, dark red when dry, abaxially prominulous; domatia absent; stip- ules 6-7 X 4-5 mm, cymbiform, not carinate, gla- the base, deciduous. Inflorescence brous, free at solitary, heads 3-3.7 em wide: inflorescence axes 2.5-3 cm long, flattened, glabrous: bracts calyptra- like, deciduous; peduncles 1.5-2.2 mm long, gla- brous. Flowers 5-merous; calyx tubes 1-1.2 mm long, infundibular, prominulously ribbed, glabrous; lobes 0.2-0.3 mm long, truncate, bearing a shallow depression in the center, a short protuberance on the edge toward the style (Fig. 14D). tomentose; corolla tubes 7-8 X 0.8-0.9 mm, inside puberu- lous, outside glabrous, lobes ca. 3 mm long, oblong, — Volume 89, Number 1 2002 Razafimandimbison 29 Revision of Breonia recurved, with protuberance abaxially (Fig. 14E), glabrous, with a few scattered hairs at the apex, ciliate; anthers ca. 1.2 mm long; filaments 0.1—0.2 mm long; styles ca. 13 X 0.1 mm; stigmas clavate: ovary 2-carpellate; carpels syncarpous; ovules 4 to 5 per locule, flattened, imbricate: placentae elon- gated, strongly flattened. Infructescences 1.2-1.6 cm diam., with accrescent calyx remnants; individ- ual fruits with endocarps soft, fibrous; disks ac- crescent, napiform; seeds 4 to 5 per locule, con- cavo-convex, ellipsoid, red; seed-coat reticulate. Habitat and distribution. This species occurs in low- and mid-altitude evergreen rainforests; Dis- tricts of Nosy Be, Ambanja, and Maroantsetra (Fig. 12) Common name. Valotra Phenology. Flowering March to July; fruiting October to November. This lyx tubes aceite ribbed, short calyx lobes with Discussion. species is distinct in its short ca- shallow depressions in the center, and short protu- berances on the edge toward the style (Fig. 14D). The specific epithet refers to the Sambirano regions. Paratypes. MADAGASCAR. Antananarivo: District Ankazobe, Forét d'Ambohitantely, 84/3 SF (TEF). Antsi- ranana: District Ambanja, Manongarivo Special Reserve, * Bekolosy, 71430 ; District Nosy be, SF (ТЕР), E 407 SF (TEF). 24753 SF (TEF), 5524 RN (TEF), Birkinshaw 59 (K, MO); Bemarivo, Androranga, ravin be Betsomanga, 841 SF (K, MO, ТЕР). Toamasina: District Maroantsetra, Masoala Peninsula, S of the village of pear pesmi Behasy & Vasey 203 (MO), Rabe 115 (MO) 16. Breonia sphaerantha (Baill.) Homolle ex Ridsdale, Blumea 22: 546. 1975. Franchetia sphaerantha Baill., Bull. Soc. Paris 60: 477. 1885. TYPE: Madagascar. [Without exact locality], Hildebrandt 3309 (holotype, P!; iso- types, K!, 0). Linn. "р fongipetiovaiism Soler., Bull. Herb. Boissier . 1893. TYPE: Madagascar. [Without exact lo- ics Hildebrandt 3309 (holotype, L!; isotypes, K! P!). Figure Breonia — Вай, J. Linn. Soc. Bot. 33: 37. 1897. TYPE: “Madagascar.” [Without exact loc v Hil- debrandt pere (holotype, K*; isotypes, L!, — Shrubs to large trees, 7-18 m tall. Bark white- tinged, rugose. Leafy stems terete, puberulous to pubescent, lenticellate. Terminal vegetative buds, 2-3 persistent; petioles 15-20 X 0.9-1 mm, terete, pu- 26.7 X 2.5-3.9 cm, elliptic to oblanceolate (oblong to oblanceolate), adaxially gla- conical, 1.2-1.5 mm, puberulous. Leaves berulous: blades 5 brous, abaxially puberulous, membranaceous, glossy, apex acuminate, base rounded to truncate; margins glabrous, entire; secondary veins ca. 7 pairs per side, eucamptodromous, barely evident on upper surfaces, red-tinged when dry; domatia pock- et-type in the axils of secondary veins, glabrous: stipules ca. 5 X 2 mm, cymbiform, not carinate, puberulous, free at the base, deciduous. Inflores- cence solitary, heads 1—1.2 cm wide; inflorescence axes 2.4-3.8 cm long, usually bent and twisted when dry, terete, pubescent to velutinous; bracts calyptra-like, deciduous; peduncles ca. 1 mm long. Flowers mostly 4-merous or rarely 5-merous; calyx tubes ca. 1 mm long, green-yellow-tinged, inside velutinous, outside pubescent, lobes ca. 0.5 mm long, triangular, pubescent; corolla tubes 4.2—4.3 ca. 0.2 mm, yellow-tinged, glabrous, lobes ca. 1.2 mm long, elliptic, lower parts glabrous, toward the apex puberulous, ciliate; anthers ca. 1 mm long: filaments ca. 0.2 mm long, terete, glabrous: styles 0-7 X ca. 0.8 mm, glabrous; stigmas globose to clavate: ovary 2-carpellate; carpels coherent; ovule | per locule, pendulous; placentae small, ovate. In- fructescences 4-10 mm diam., densely pubescent, with well-developed calyx remnants: individual fruits with endocarp hard, glossy: disks accrescent, obconical: seeds 1 per locule, with one side convex, the other flattened, ellipsoid, red; seed-coat lineate. Lowland evergreen Nosy Be, Ambanja, and Befandriana-Nord (Fig. 10). Habitat and distribution. rainforests; Districts of Antsiranana II, Phenology. Flowering January—July; fruiting March—December. Baillon (1885), Solereder (1893), and Haviland (1897) all used different sheets of the same collection, Hildebrandt 3309, as the basis for Franchetia sphaerantha, Discussion. Elattospermum longipe- tiolatum, and Breonia parvifolia, respectively. Nei- ther Solereder nor Haviland knew about Franchetia sphaerantha, and Haviland did not cite Elattosper- mum longipetiolatum either. Idditional specimens examined. MADAGASCAR. Antsiranana: Dist dg Antsiranana Il, Canton Anamakia, Ankotikona, 15273 SF (TEF); bords de rivière Kongony, Ambaliha. 2950 pes (T " TEF): District Nosy be, Lokobe Reserve, Antilahimena 62 (K, MO); Lokobe, 7822 RN (TAN). 0229 RN (TAN), 9453 RN (TEF), 11417 SF (TEF), 24764 SF (TEF); District Ambanja, 5б уйрет 255 (МО); — Marovato, Beangona, Rakoto 9 (TAN). Ma- — District Befandriana-Nord, Commune Tsaraho- папа, Forêt domaniale d'Antetezana, 15947 SF (TEF); Se ee locality], 6257 RN (ТЕЕ). = 17. Breonia stipulata Havil., J. Linn. Soc. Bot. 33: 35. 1897. TYPE: “Madagascar, Northwest” [without exact locality], 1841 (fr), Pervillé s.n. (holotype, P!). 30 Annals of the Missouri Botanical Garden F uim m H E E i | E "| ARE 2 1 (0 = o f Figure 13. Breonia richardsonii. —A. Fertile branch with mature infructescences. —B. Fertile branch with inflo- ature infructescence, bearing a tubular bract. —D. Seed surface. —E. Seed, lateral view (left); dorsal rescence. (С. Volume 89, Number 1 Razafimandimbison 31 2002 Revision of Breonia Trees or shrubs, 15-25 m tall. Bark rugose. Leafy 18. Breonia taolagnaroensis Razafim., sp. nov. stems quadrangular, glabrous. Terminal vegetative 19-22 х 2-3 mn, glabrous. Leaves deciduous; petioles 12-22 X 1.5-2 mm, adaxially canaliculate, glabrous; blades 10—18.5 (224.5) X 4.5— 6(-8) cm, elliptic to lanceolate, glabrous, coriaceous, buds conical, not glossy, both apex and base acute; margins gla- brous, entire; secondary veins (9—)11 or 12 pairs per side, eucamptodromous, abaxially prominulous; dom- atia absent (hairy or crypt-type domatia in the axils of secondary veins); stipules 15-18 x 2.2-2.5(-3) mm, cymbiform, narrowly carinate abaxially, glabrous, ree at the base, deciduous. Inflorescences solitary, heads 2.2-2.5 cm wide; inflorescence axes 4—6 cm long, strongly flattened, glabrous: bracts calyptra-like. deciduous; peduncles 1 mm long. Flowers 5-merous or rarely 4-merous; calyx tubes ca. 2 mm long, inside lanate, outside glabrous, toward the lobes puberulous, lobes ca. 1 mm long, oblong, tomentose; corolla tubes ca. 5 X 0.9-1 mm, glabrous, lobes 1.5-2 mm long, oblong, glabrous; anthers ca. 1.5 mm long; filaments 0.2-0.3 mm long; styles ca. 10 X 1 mm, stigmas clavate to slightly cylindrical: ovary 2-carpel- glabrous; late; carpels syncarpous; ovules 1 to 2 per locule, pendulous, imbricate; placentae ovate, small. Infruc- tescences 1.5-1.7 cm diam., with well-developed ca- lyx remnants; individual fruits with endocarp hard, glossy; disks accrescent, heart-shaped; seeds 1 per locule, pendulous, concavo-convex, ellipsoid, both ends acute, red; seed-coat reticulate. Habitat and distribution. Districts of Maintirano, Sambava, Antsiranana II, and Antsalova (Fig. 1C Deciduous dry for- ests: Andapa, Common names. — Valitsy and ИОНИ НОРРРРЕ И Phenology. Flowering October to December: fruiting December to February. Additional specimens examined. MADAGASCAR. An- tsiranana: District Ds Mandena, Forét de — jy — ¢ паа 2 9 (TAN): lisière de forêt, otozafy & Ral — 22 a т AN Marojejy Est. f Mandena, TAN); District Asana II, tamisakana, 15057 SF lova, | km à l'Est f P); Botomena, 11089 RN (TEF); Tsiandro, Forêt d'Antsingy aux environs de la clairière de piste d’Ambodiriana, 6791 ); trail to the summit | of Miller 3609 AS MO, P, 1 ‚ Ambarara- TYPE: Madagascar. [Without exact locality]. Commerson s.n. (holotype, P!). Figure 2F. Haec species a congeneris disco nectarifero pentagono accrescente permagno distinguitur; ex sylvis littoralibus prope Taolagnaro tantum cognita. Shrubs, 4—6 m tall. Bark rugose with annular fissures. Leafy stems quadrangular, glabrous. Ter- ) X minal vegetative buds conical, 5 а. 3 mm, ^ glabrous. Leaves persistent; petioles 13-20 X ca. 2.5 mm, adaxially canaliculate, glabrous; blades 9— 11(-16) X 3.5-6.5 em, ovate to oblong, glabrous, coriaceous, glossy, apex rounded to broadly cuspi- date, base cuneate; margins glabrous, entire; sec- ondary veins (7—)9 to 10 pairs per side. eucamp- todromous; domatia absent; stipules 6-10 mm long. cymbiform, not carinate, free at the base, glabrous, deciduous. inflorescence axes (4—)6—8.5 cm long, strongly flattened: bracts not seen; peduncles absent. Flowers not seen. In- Inflorescence nol seen; fructescences 1.5-2 em diam., with well-developed calyx remnants; individual fruits with endocarp hard, glossy; disks accrescent, pentagonal; seed 1 per locule, concavo-convex, with rudimentary wings at both ends, red; seed-coat reticulate. and evergreen rainforests; Districts of Fort-Dauphin and Farafangana (Fig. 10). Common names. and Marotsaka. Habitat and distribution. Littoral forests Valotr'angady, Molompangady, Phenology. Flowering unknown: fruiting Octo- ber to December. Bosser (1984) argued that the type specimen of Cephalidium citrifolium (Poir.) A. Rich. is the Commerson collection on which I base Discussion. Breonia taolagnaroensis. | disagree with Bosser be- cause Commerson s.n. now in P has fruits only, but Richard used a specimen with both flower and fruit on which to base his Cephalidium. Commerson s.n. has a flower only and matches Poiret’s original de- scriptions of his Nauclea citrifolia. Breonia taolagnaroensis is distinguished from the other species of Breonia by its massive, accrescent pentagonal disks, and is known only from the littoral SF (TEF), 12023 SF (TEF), т? з G EF): Forêt de Va- forests of the Taolagnaro (Fort-Dauphin) regions, loala, [collector unknown], 2200 (T: thus the specific epithet. The type specimen of this < view (right). —F. Median dissection — flower, showing bicarpellate ovary with numerous ovules (left): entire flower, showing part of corolla tube entire individual fruit (right). 3s d carpel, showing . Dissection through a cay x remnant and a carp el of a separate fruit (left): la ent disk adnate to the septum and 5 ovules БЕ). = = attached to the placentae. A-F бой 14359 SF (TEF) and G from 18311 SF (TE 32 Annals of the Missouri Botanical Garden \ en KM MP = 00). „ и IFA \. Fertile branch with infructescences. —B. Fertile branch with inflorescences —К. A corolla lobe bearing a protuberance. —F. Two adjacent flowers, showing Breonia s Figure 14. Mature infructescence. —D. \ two adjacent ovaries separated: median D. Calyx. dissection through flower, showing a calyx and corolla tubes, and ovary (left); —C. Volume 89, Number 1 2002 Razafimandimbison Revision of Breonia species was identified by Capuron (1973b) as Breon- ia chinensis (Lam.) Capuron; however, it is distin- guished from B. chinensis by having bark with an- nular fissures and by a single ovule per locule. Paratypes. MADAGASCAR. Fianarantsoa: Dist pura em 4855 SF (TEF); Forét de (ioe 23641 SF Toliara: District Fort-Dauphin, Forét d'Analalava, cds 1367 (К, MO); Forêt d'llandy au Nord de Fort-Dauphin, 20558 SF (TEF), McPherson 14391 (K, MO, TEF); Manantenina, Beharena, 10925 SF (TEF). 19. Breonia tayloriana Razafim., sp. nov. Madagascar. Toamasina: Fénérive Est, Tampelo 3 Jan. 1956 (fl), 15703 SF (holotype, TEF). Figure 15. зесіеѕ a congeneris stipulis oe арз ашаа carinatis perfacile distinguitu Medium-sized trees, 9-10 m tall. Bark longitu- dinally fissured. Leafy stems quadrangular, lenticel- late, glabrous. Terminal vegetative buds conical, 13— 30 X 7-10 mm, glabrous. Leaves persistent, petioles 22-35 X 24.5 mm, adaxially canaliculate, gla- brous; blades 18-32 х 10-15.5 cm, broadly obovate to broadly ovate, reddish, glabrous, coriaceous, glossy, apex obtuse to rounded, base cuneate; mar- gins glabrous, entire; secondary veins 7 to 9 pairs per side, eucamptodromous, prominulous; without domatia; stipules 19-31 X mm, cymbiform, united at the base, semi-persistent. Inflorescences solitary (2 per axil), heads 3.2-3.5 cm wide; inflorescence axes 3.24.2 cm long, strongly flattened, glabrous; bracts calyptra- like, deciduous; peduncles 1—2 mm long. Flowers 5- merous; calyx tubes ca. 2 mm long, green, inside lanate, outside glabrous, lobes ca. 1.7 mm long, ob- long to truncate, lanate; corolla tubes 8-9 X ca. 1 mm, light beige to cream, inside puberulous, outside glabrous, lobes ca. abaxially carinate, glabrous, .8 mm long, oblong, at the lower parts glabrous, puberulous toward the apex; anthers ca. 1 mm long; filaments ca. 0.1 mm long, glabrous. terete; styles ca. 13 X 0.2 mm, glabrous: stigmas clavate; ovary 2-carpellate; carpels syncarpous: ovules 2 to 4 per locule, strongly flattened, pendu- lous, imbricate; placentae elongated. Infructescences 2-2.4 cm diam., nants; individual fruits with endocarp soft, fibrous; with well-developed calyx rem- disks accrescent, obconical; seeds 1 to 2 per locule strongly flattened, ellipsoid, red; seed-coat reticulate. Habitat and distribution. Evergreen rainforests and littoral forests; Districts of Maroantsetra, Brick- aville, and Fénérive Est (Fig. Common names. Molompangady mena (red lips of spade). Phenology. Flowering December to February: fruiting February to April. Discussion. The specific epithet honors Char- lotte Taylor, a Rubiaceae specialist who has studied mostly the Neotropical Rubiaceae. Charlotte has taught me various aspects of taxonomy, including nomenclature, describing new species, writing a monograph, herbarium curation, and specimen an- notation. The present revision was carried out un- der her supervision. District anton Ampasina, А Zavah Paul TEF); District ae :'kaville, Ambila Lemaitso, Ampanotoamaizina, 8317 SF (ТЕЕ); — jer nns "tra. Masoala peninsula, coastal sand forest just N of Antalavia (K, MO, P, TAN); Jardin жае de Farankaraina, 5652 SF (TEF), 63- R-199 (TEF); [Locality unknown], 32928 SF (TEF). Paratypes. MADAGASCAR. Toamasina: 20. Breonia tsaratananensis hazafim., sp. nov. ГҮРЕ: Madagascar. [Antsiranana province]: Massif de Tsaratanana, haut bassin de la Bean- drarezina (Andranomena), affluent rive gauche de la Mahavavy, 2000-2300 m, 11 Nov. 1966 (fr), 27049 SF (holotype, TEF). Figure 16. Haec species a congeneris tubo calycino supra medium dilatato apicem basemque constricto facile distinguitur. Trees, 15-30 m tall. Leafy stems quadrangular, glabrous. Bark rugose. Terminal vegetative buds conical, 8-9 X ca. 3 mm, glabrous. Leaves persis- tent, petioles 15—20 mm long, adaxially canaliculate, glabrous; blades 9.5-11.3 х 5.2-6.5 cm. elliptic to obovate, brittle when dry, glossy, apex mucronulate, base attenuate; margins glabrous, glabrous, coriaceous, revolute, entire; secondary veins ca. 7 pairs per side, orange to yellow-tinged, eucampto- dromous, prominulous; cryptic-type domatia at the base of the secondary veins, evident, ovate, glabrous: stipules ca. 10 X 3 mm, cymbiform, not carinate. glabrous, deciduous. Inflorescences solitary, heads ca 1.7 em wide; inflorescence axes 2.5-3.3 cm long, slightly flattened, woody, densely pubescent: bracts 1-2 long, densely pubescent. Flowers 5-merous or rarely calyptra-like, deciduous; peduncles ca. mm ft entire flower (right). —G. bu adjacent 2 — two adjacent carpels separated: median dissection of mature infructescence, showin escent disk and s lahimena 237 (MO) and G ae 11407 SF (TEF). Seed, dorsal view (left); lateral view (right). A—F from Anti- 34 Annals of the Missouri Botanical Garden nt кү AV / Vy \ } UN Jl ur M DM Ñ {л ya ГІ May Lu V Eun En NA ly ШУ Ч ҮЛ ( anrs Ee — mm Figure 15. Breonia tayloriana.—A. Fertile branch with inflorescence and semi-persistent stipules. —B. Two ad- jacent flowers, showing tw о adjacent ovaries separated: median dissection through flower, showing lanate calyx and ovaries (left); entire flower (right). —C. Median dissection through a separate fruit, showing accrescent disk and ovaries with young seeds. A-C from 15703 SF (TEF). Volume 89, Number 1 Razafimandimbison 35 2002 Revision of Breonia Figure 16. Breonia tsaratananensis.—A. Fertile branch with a young inflorescence and a mature infructescence. —B. Two adjacent flowers, showing two adjacent ovaries separated: median dissection of calyx with adaxial view (left): abaxial view (right). —C. Median dissection through fruit, showing accrescent disk and carpels with seeds (left); entire fruit (right). —D. Dissected unilocular ovary. —E. Mature seed: dorsal (left); lateral (right). —F. Mature infructescence. A, B from 973 SF (TEF) and C-F from 27049 SF (ТЕЕ). Annals of the Missouri Botanical Garden 4-merous; calyx tubes ca. 2.5 mm long, dilated above the middle, constricted at both ends, promi- nently ribbed, inside glabrous, outside puberulous, lobes truncate, ribbed, quadrangular, tomentose; co- rolla tubes 5-6 mm long, inside pubescent, outside puberulous; lobes ca. 2 mm long, recurved, puber- ulous; anthers not seen; style 7-8 mm long, glabrous: stigmas clavate; ovary 2-carpellate, rarely 1-carpel- late; carpels syncarpous; ovules 3 to 5 per locule, flattened, elongated. Infructescences 2-2.5 cm diam., with ac- pendulous, imbricate; placentae large. crescent calyx remnants, persistent on the branches until the next flowering season; individual fruits with endocarp soft, fibrous; disks accrescent, triangular, seeds 3 to 4 per locule, concavo-convex, ellipsoid, dark red; seed-coat reticulate. Mid- and high-alti- District Ambanja (including Habitat and distribution. tude humid forests; Tsaratanana Mountain) (Fig. 12). Common name. Valotro. Phenology. Flowering December to February: fruiting March to November. This species is known from five col- lections made from three individual trees; therefore, it is considered to be rare. Also, this is one of the two species of Breonia found only in high-altitude (above 1500 m) humid forests. The specific epithet refers to the highest Mala- Tsaratanana (2880 m), where the type species was collectec . gasy mountain, aratypes. MADAGASCAR. Antsiranana: District Ambanja, Massif de ar vallée de la Mahavavy. Morat 2306 (MO, TAN). hajanga: Mangindrano, Am- bohimirahavavy, 97.3 SF (Ko MO, TEF). DUBIOUS SPECIES Breonia longipetiolata Havil.. J. Linn. Soc. Bot. 33: 36. 1897 definitely belonging to Breonia, TYPE: Leprieur s.n. (P!). This specimen, was reported to have been collected by Leprieur from French Gui- ana. However, the locality must be considered doubtful because Breonia is restricted to Madagas- car and has never been reported from South Amer- - ^ Also, Leprieur never collected plants in Mad- The of B. longipetiolata does not match any of the species agascar. type and only specimen recognized here. This species is distinguished by its leaf blades with acute to rounded apices and long (ca. 9.6 em) and strongly flattened peduncles. EXCLUDED SPECIES Setch. = (Setch.) Setch. & Christoph. This species was orig- Breonia mayortt Sarcopygme тауоги inally placed in Breonia simply because of its mul- tiple fruits. It has been transferred to the genus Sarcopygme of the tribe Morindeae based on the presence of raphides and bifid stigmas. Literature Cited Abbiw, D. K. 1985. The site struc ture of Nauclea кж ji eti . 497—498 in P. Goldblat P. P. Lowry H (editors). um Systematic ga e Апо an Botany. Monogr. Syst. I Missouri Bot. Ww = = rd, 2 — ” 1879. Observations sur les Nauclées. Adan- sonia i 311-312. . 1885. Liste des Se de M ull. Soc. Linn. Paris 60: 17 Bakhuizen van den Brink, К. € 1970. Nomenclature and — of the genera of йн 'eae-Naucleeae and 'oposal to conserve the generic name Nauclea L. Taxon 19: 4608—44 d . R. 1995. Тһе Irporisnoe of the Black Le- nur o m macaco, Lemuridae—Primates), for Seed Dispecs al in Lokobe Forest. Madagascar. Doctoral The- sis, University College, London, England. Boiteau, P. 1985. Dictionnaire de noms Malgaches de vé- ве taux, Microédition no. 85 04 10, Archive des docu- ‚ Muséum а d — ire Natur ене; Гап. 9 — ^^ Madagascar (suite). > ат. Neclamafokia, un nouveau nom pour ени phalus auct. non A. Rich. (Rubiaceae). Bull. Mus. Natl. ist. Nat., Sect. В, Adansonia, sér. 4, 6: 243-248. 1999, A propos d'Anthocephalus A. Rich. et de Cephalanthus chinensis Lam. (Rubiaceae). Adansonia, sér. 3, 21: 93-95. т аи C. E. B. 1966. Remarks on the position, the poe d and the subdivision of the Rubiaceae. Acta . 15: 1 A. De. [Sep. E 1830. Prodromus Systematis Na- Iv: Paris. . Sur l'identité du ne chi- аА sér. 2, 13: 73b. Revision des Rubiacées de Madagase ar Z a nensis pu 71—47: et dis м omores. Unpublished manuscript: Notes regrou- I dacty — s de F. Chauvet. Laboratoire Phanérogamie, P 1997. Plant Collectors in Mac — ar ane the Co omoro Islands. ie Botanic Gardens, Kew. Endlicher, S. F. L. . Enchiridion bu ‘um. Lipsiae Sumptibus ( Gull. tim — Viennae. Grandidiet y 1897. Histoire physique, — el Маанаке аг Vol. 36: plate 457. Publié Г Imprimerie Nationale, Pa Guého, J. 1976. Sur l'identité de Phylica — ées) des Masc are Ignes. — sér. 2, 15: 509-513. Harris, J. G. € . 1994. Plant ш 'ation Terminology. Bine la Ut tah. Haviland, G. D. mee A revision of the tribe Naucleeae. et mises en forme par J. Bos ser, itique de . Linn. Soc. . : 1-94 Heine, H. 1968. A propos de la nomenclature e ч Sé- : — de l'ancien monde. Adansonia, sér. 2, 8: 181— А А. М. 1938. Le genre Breonia de Madagascar. Bull. Soc. 3o Fr, 84: 457—402 ds M. s. Proc. . On үш = newpoints — some e Ned. Akad. W ser. С, 69: 275-316. Volume 89, Number 1 2002 Razafimandimbison Revision of Breonia 37 Lamarck, A. 1785. a Méthodique Bota- sos Panc — ke, Paris. 789. о е Méthodique Botanique 4. GA Paris. 435. q E. Turner. redator and A mites in leaf domatia. Amer. J. Bot. 76: 105-112. Radford, A. E., W. C. Dickison, J. R. Massey & R. Bell. ^a Vascular Plant Systematics. Harper & Row, New Yor байый. 5. G. & В. Bremer. 2002. Tribal de- limitation of Naucleeae (Cinchonoideae-Rubiaceae): Inference from melee ub and torba icd data. Syst. Geogr. Pl. (in Richard, A. [Dec. ] 1830. €— ү Rubiac: ‘Сев. 270 p.. reimpr. In Mém. Soc. . Paris 5: 81-304 1834). ( Ridsdale, C. 1975. A synopsis of the African and Mada- gascan د‎ a ai eeae. Blumea 22: 541—553 1978. A revision of the tribe Naucleeae s. s. (Ru- biac — 24: 307-3 8. Rubiaceae. Pp. п М. D. Dassan- ayke hs — A Revised Тт... to de Flora of Cey- lon, Vol. XIL A. A. Balkema, Rotterdam, Netherlands. Robbrecht, 5 ee — woody Rubiaceae. Opera Belg. 1 Г}. A 1994. $ Saleen u to S 1988 outline of the clas- sification of the Rubia . Index genera. Opera Bot. Belg. 3(1993): 173- 196. Schumann, К. 1891. Rubiaceae. In A. Engler & K. Prantl, : 55-00. 893. Anatomischen C hora и und zur Не. Boissier 1:277 éd. “Mémoire sur la famille des Rubiacées.” Pos O: 186 187. Verdcourt, B. 1958. Remarks on the cla ssification of the l. Rubiaceae. Bull. Jard. Bot. Bruxelles 28: 209-28 APPENDIX 1. T OF SPECIES . Breonia boivinii Hav 1 Breonia capuronii Razafin m. 3. Breonia chinensis (Lam.) Capuron : Breonia cuspidata (Baker) Havil. . Breonia decaryana Homolle 6 Breonia fragifera С ;apuron ех ее ‹ 3 © =; & 8 =. 8 E FP & 3 & a» 7 16. Breonia ik (Baill.) оке ex Ridsdale 17. Breonia stipulata Havi 18. Breonia taolagnaroensis Rasen: 19. Вгеота tayloriana Razafim. . Breonia tsaratananensis Razafim. INDEX TO EXSICCATAE "xamined specimens are listed alphabetically by col- lector, followed by collection numbers; the species is in- dicated by a number in par tenthe ‘ses corresponding to the number in the List of Species 1989. Occurrence of Andrianarisata et al. 259 (6); Antilahimena 19 (1), . 237 (15); Antilahimena et al. 324 (6). n 5563 (4), 6602 (4); Behasy & Vasey 203 (15); Bernard 187 (3), 327 (3); Bernardi 11899 (1: 3); Birkin- shaw 59(15), 67 (1); Barin s.n. (1). P 3). s.n. (11); Commerson s.n. (3), s.n. n. (18). Decary 5199 (5), 5511 (3), 551 1367 — Gautier 3772 (2); Gautier et al. 3288 (6). Hildebrandt er (е; Humblot s.n. (3). Jongkind et al. 3 13). Labat et al. 2761 id 2263 (17). Labat & Conté 2674 (13); Lewis et al. 755 (9); Louvel 125 (8), E (12), 216 (5). Malcomber 1219 (2), Malcomber et al. 1833 (13); Mc- Pherson 14391 (18); Miller 3609 (17); Mora 2306 (20). Noyes et al. 961 (6), 974 (10), 1028 (13). 12); Perrier de la Bâthie 351: 3 (13), 3904 1). 3933 (10). 5270 SF (10); eps s.n. (17). Rabe 115 (15); Rahajasoa et al. 747 (3); posse : (16); Rakotozafy & Raharilala 2189 (17), 2271 (17); Randri- Hey" 241 (1), 251 (6), 255 (16); pape capi: ; Randrianasolo & Rasabotsy 32 (7 ); Ra 3 (3). 5573 (3); Dumetz — velonarivo 13 (13), SG 352 (10), SG 358 (7). SG 389 (8); Ré- serves Naturelles: 2751 RN (1), 4905 RN (1), 5152 RN ‚ 5524 RN (15), 6234 RN (1), 6257 RN (16), 6809 RN ‚ 7822 RN (16), 7906 RN (2), 9002 RN (2), 9045 RN (6), 9052 RN (10), 9129 RN (10), 9229 RN (16), 9453 RN (16), 9982 RN (3), 11089 RN (17), 12975 RN (2). Schatz et al. 1912 Ue 3858 (3); Scott Elliott 2214 (4); Station Forestiere: 841 SF (15), 973 SF (20), 1432 SF (1). 2078 SF (9), 2410 SF D 2950 SF (16), 2961 SF (6), 3026 E (13). 3265 SF (10), 3882 SF (13), 4677 SF (13), 4855 SF (18), 5652 SF (19), E. SF (3). 6105 y (7), 6181 SF (13). 6360 SF (3), 6616 SF (13), 6696 У 13). — SF (13), 6791 SF (17), 7136 SF (7), 7 7562 SF (8), 7921 SF (8), 7987 SF (1: H, 8245 p (13). 8317 SF (19), 8413 SF (15), 8734 SF (6), 9047 SF (3). озы (15), 9238 SF (3), 9711 SF (9), 9974 SF (7), 10803 К (6), 10925 SF (18), 11407 SF (15), 11417 SF (16) Ша SF (15), 11503 SF (6), 11594 SF (9), 12023 SF (17), 12352 SF (3), 12447 SF (3), 13094 SF (12). 14999 SF (14), 14437 SF (3), 14712 SF (3), 14728 SF (3), 15057 SF (17), 15273 SF (16), 15287 SF ( (3), (3), 15495 SF (3), 15518 SF (3), 15673 SF (13), 15947 SF ‚ 16174 кый eb 16217 SF (3), § F (3), 17565 SF (13), 18116 SF ty les SF (14), 18450 SF (13), 18518 SF SF (13), 19295 SF (7). 19379 SF 5 SF Ay 19791 SF (6), 20558 SF (18), : SF ( Т), 26103 SF | 3), 26804 SF (5 E н SF (20), 27192 SF (3), 27503 SF (13), 27593 SF SF (1), 28081 SF (3), 28136 SF (10), 28414 SF (5). : SF (5). 29653 SF (13). 31809 SF (3), 32928 SF (19). : SF (10), 33237 SF (6), 33240 SF (3), 33536 SF (6). 3 SF (6). 34413 SF (6), 34711 SF (3), 34719 SF (3). S/N SF (3). Тюшген s.n. (о), 91 (5), 117 (10); Turk et al. 643 (3) Villiers et al. 4846 (13); Vigreuse l 5418 (3). y Е 524 I avah 321-R-107 (19); 5524 RN (+ Unknown collectors: 12-R-B-172 (5): 14-B-R-230 (9); 86-R-118 (3); 63-R-199 (19); Herbier de la station agri- cole de l'Alaotra, 2200 (TAN). CLASSIFICATION, ORIGIN, AND DIVERSIFICATION OF THE NEW ZEALAND HEBES (SCROPHULARIACEAE)' Steven J. Wagstaff,? Michael J. Bayly,? Philip J. Garnock-Jones,' and Dirk C. Albach? ABSTRACT The New Zealand hebes үө rophulariaceae) are members of a large Southern Не — clade nested within Ve- ronica. Analysis of ITS and rbcL sequences suggest s that ancestor that arrived via long-distance dispersal. Zealand, the hebes have unde "gone at degree of morphological diversity in the New + divergence. New Zealand w high mountains or forest margins. Our oil sugges was a source of ne to South America, ~ supported by the sequence data. Key w ords: the N After the establishment of this initial founder population in New t least two major episodes of diversification, | hebes contrasts with a at least one instance from New Zealand to Australia, anc ew Zealand species are derived from a single common six clades. The great агаш ыа low level of sequence South Paci giving rise to s uinea have occurred relatively recently. Shorter hops to the Chatham Islands and the subantarctic islands are also Hebe, ITS, New Zealand, phylogenetic analysis, rbcL, Serophulariaceae, Veronica. Long-distance dispersal has a profound influence on the evolution of insular floras (Carlquist, 1974), and there is substantial evidence suggesting that it occurs relatively frequently (Godley, 1967; Pole, 1994). One of the most remarkable examples of dis- persal followed by adaptive evolution on islands is the New Zealand hebes (Serophulariaceae). Wags- taff and Garnock-Jones (1998, 2000) suggested that the New Zealand hebes are the descendants of a small founder population that may have been de- rived from a single seed. They proposed that com- bined influences of inbreeding, genetic drift, and strong selection acting upon small populations have probably played a major role in the rapid diversi- fication of the group. The hebes are one of the largest and most eco- logically diverse plant groups in New Zealand, in- cluding over 120 species, with outlier populations in eastern Australia, Tasmania, New Guinea, Rapa Island, and South America. They range from al- pine cushion-forming plants (Fig. 5D) to lowland woody shrubs or small trees (Fig. 5M, R, S), and in New Zealand are conspicuous elements in most terrestrial ecosystems except forests and wetlands. Species such as Hebe armstrongii, H. cupressoides, and H. speciosa have patchy or localized distri- butions and are considered rare or endangered; about 70% of the species are confined to small regions within New Zealand. The New Zealand hebes were formerly included in a broadly defined circumscription of the genus Veronica (Wettstein, 1891; Cheeseman, 1925), but recent flora and taxonomic treatments (Ashwin & 1961; 1993a, b; Heads, 1994a, b) recognize less inclusive groups Moore in Allan, Garnock-Jones, (see Table 1), usually accepting four genera in New Hebe, Heliohebe, and Para- 1993a, b). Heads (1987) de- scribed an additional genus, Leonohebe. Although Zealand: Chionohebe, hebe (Garnock-Jones, we do not accept his wide circumscription of that genus, the name Leonohebe could be applied to a small clade of four or five species that is supported by the analyses of Wagstaff and Garnock-Jones (1998, 2000). ! The authors gratefully acknowledge curators of the experimental gardens and CHR ae TT Research, Lincoln, sequencing. Dick Olmstead contributed s tics Ltd.) ا‎ many of the photographs. a ind Alison iefited gn — тот the ins . This research was funde along with the assistance of Anita Thorne, Andy Gla azie г, and Ё izabeth MacAvoy w ightful comments of Christine Bezar, the NA Kellow assisted um preparation of figures. Earlie versions Ilse Breitwieser, Mar k ( ew Zealand by the Foundation for Research, Science thase, d in Ne and Tec :hnology (contrac : : 09618 dnd MNZOOI), 34 he Marsden Fund (е ontract LANOOI 2 Landcare Research, Box 69, Lincoln 8152 ! Museum of New — Te Papa Tongarewa “PO. Box v Zeal: ind. 107. .CrI. nz. New Zealand. арза Welling glon, ! School of Biological Sciences 2d Island Biolog Research Programme: Victoria University of Wellington, P.O. Box 600, Wellington, New Zealand. ` Botanisches Institut der Universität Wien, Rennweg 14, ANN. Missouni Bor. GARD. 89: 38-63. 2002. 1030 Wien, Austria. Volume 89, Number 1 Wagstaff et al. Diversification of the New Zealand Hebes el. Classifications of New Zealand hebes. The New Zealand species were placed in three sections of Wettstein’s akon treatment of Veronica, and Cheeseman (1925) placed them in two divisions of Veronica. win & Moor Ashw e Wettstein (1891) Cheeseman (1925) in Allan (1961)! Heads (1987, 1994b) Garnock-Jones (1993a, b) Veronica Veronica sect. Pygmea Division Pygmea Pygmea Chionohebe Chionohebe Inc luding Parahebe “Group B”) sect. Hebe Le onohebe sect. Densifoliae Division Hebe Leonohebe Connatae ebe Heb “Semiflagriformes” sect. “Semiflagriformes” “Connatae” sect. “Connatae” S Ф A sect. Apiti “Flagriformes” Flagriformes “Flagriformes” sect. Aromaticae Salicornioides hee n sect. “Buxifoliatae” sect. “Buxifoliatae” “Subdistichae” sect. Sind hae ^| sect. Subdistichae “Subcarnosae” sect. Glaucae sect. Glaucae Heb sect. Hebe sect pertae” ser. Hebe ser. Hebe “Occlusae” ser. Occlusae ser. Occlusae Parahebe “Grandiflorae” “Grandiflorae” “Paniculatae” sect. Paniculatae Heliohebe sect. Chamaedrys Division Euveronica Parahebe arahebe Group A, B, C” “Group A, C” sect. Labiatoides — — Derwentia sect. Paederota =P == — — sect. Paederotoides — — — — sect. Pseudolysimachia — — — — sect. Veronicastrum — — — — sect. Omphalospora — => — = sect. Beccabunga — — = = ! [n the Flora of New Zealand. — l erir 1961), "mes and the informal grouping "Flag was prepared by L. B. Moc This research contributes to ongoing efforts to create a phylogenetic classification of Scrophulari- aceae. Olmstead and Reeves (1995) and Olmstead et al. (2001) showed that the Scrophulariaceae, as traditionally circumscribed, are not monophyletic. They identify clades from a dismembered Scrophu- lariaceae s.l. that could merit formal recognition. In their studies Veronica was nested within a large clade they called ue prop. (Reveal et al., his large clade was recognized by — i Reeves (1995) and in- cludes part or all of Bentham's (1876) tribes Digi- taleae, Antirrhineae, Cheloneae, and Gratioleae, the small tribe Angelonieae, and the small families b ard nom. cons. Callitrichaceae, Globulariaceae (excluding Selagi- naceae), Hippuridaceae, and Plantaginaceae. The aim of this research is to identify well-sup- M. B. Ashwin prepared the treatment of Parahebe, Pygmea, ` of Hebe. The remainder of the Hebe treatment, including the informal synopsis, ported monophyletic groups among the New Zea- land hebes, to improve their classification, infer their origin, and explore underlying processes of diversification. We propose that diversification in the group reflects transoceanic dispersal and adap- tive radiation. The hebes have successfully exploit- ed a diversity of ecological niches that were prob- ably created during the recent uplift and glaciation of the mountains of New Zealand. MATERIALS AND METHODS Our sampling strategy capitalized on the unique characteristics of rbcL and ITS sequences. The plastid encoded gene rbcL has relatively few vari- able sites, which allowed sequence comparisons among distantly related outgroups, and placement Annals of the Missouri Botanical Garden of the hebes within Serophulariaceae. It is also use- ful in that a large number of published rbcL se- quences are available for comparison (see Chase et al., 1993; Küllersjó et al., 1998; Olmstead et al., 2001, and references therein). Finally, Albert et al. (1994) and Bremer and Gustafsson (1997) suggest- ed that the gene rbcL approaches clock-like behav- ior in its evolution, and hence the amount of se- quence divergence could be used to estimate divergence times. By comparison, the nuclear en- coded ITS-region has many more variable sites than rbcL, to resolve е at lower taxonomic levels which provides more informative characters (Baldwin et al., STUDY GROUP The rbcL study group consisted of 33 species including 12 of the New Zealand hebes with at least l representative from each of the currently recog- nized genera, 5 species of Veronica, and 1 species Nineteen rbcL ~ =a Veronicastrum. sequences were newly published herein along with 13 published sequences of Antirrhinaceae from Olmstead et al. (2001), and Nicotiana tabacum (Solanaceae) was designated as the outgroup (Lin et al., 1986). Seven sequences were considered redundant; even though they were not identical, the resolution of missing data could potentially make them identical. We therefore excluded Derwentia derwentiana and D. perfoliata, and the New Zealand accessions of Hebe elliptica and H. salicifolia, from subsequent analy- ses. Nineteen of the 37 species included in the rbcL analysis were also included in the ITS survey. The ITS study group included 78 sequences, 19 of which were newly published. Among these are 58 representatives of Chionohebe, Derwentia, He- liohebe, and Parahebe, including conspecific acces- sions of Chionohebe ciliolata and С. densifolia from Australia and New Zealand, Parahebe lithophila from Australia, and P. randewateri from New Guin- ea. Thirty-five species of Hebe were also included, and among these were: at least one representative from each of Moore's (in Allan, 1961) informal groups; H. formosa from Tasmania; H. benthamii from the New Zealand subantarctic islands; H. bar- keri, H. chathamica, and H. dieffenbachii from the Chatham Islands (east of the main islands of New Zealand); accessions of H. elliptica from both New Zealand and the Falkland Islands; and H. folia from both New Zealand and Chile. Pseudoly- Veronicastrum, and Wulfenia salici- simachion, Veronica, emerged as potential sister groups of the hebes in the analysis of Hong (1984) and Albach and Chase (2001); therefore a total of 19 species representing these genera were included in our analysis. The Asiatic species Veronicastrum sibiricum was desig- nated as the outgroup for the analysis of ITS se- quences. Voucher specimens are listed in Appendix 1, along with collection information, literature cita- tions, and GenBank () accession numbers. The complete data sets are available upon — from the first author, and TreeBASE ( 50% are given above each node. Volume 89, Number 1 Wagstaff et al. 2002 Diversification of the New Zealand Hebes Hebe benthamii Hebe salicornioides Hebe odora Hebe elliptica 3.9 Hebe salicifolia mybp Derwentia nivea Veronica arguta H Parahebe catarractae Parahebe vandewateri — Hebe cupressoides [ Hebe macrantha Chionohebe densifolia Veronica catenata Veronica persica Hebe formosa — Hebe cheesemanii - 9.9 Heliohebe raoulii mybp Veronica anagallis-aquatica V eronica officinalis Veronicastrum sibiricum "n Digitalis purpurea Plantago lanceolata Callitriche heterophylla Hippuris vulgaris Globularia cordifolia Chelone obliqua Collinsia grandiflora Antirrhinum majus [ — — Gratiola pilosa l Amphianthus pusillus Bacopa caroliniana Angelonia pubescens Nicotiana tabacum — 5 changes иге 2. One of the maximum parsimony trees recovered from a —— Кар of rbcL sequences. The branch lengths are proportional to the number of changes along eac E — 1. See scale at bottom. Fossils of Veronicastrum sibiricum are reported from the mid Miocene about 15 m bê. ence RE are xq at the ancestral node TB of Australasian species (V. persica is Eurasian) and the anc sl udi of the Hebe clade Annals of the Missouri Botanical Garden € Table 2 Insertions and deletions inferred from ITS sequence comparison. Insertion/ Taxon etion Size Position Sequence Veronicastrum sibiricum Deletion 1 41 Uc Veronicastrum sibiricum, Wulfenia Insertion | 54 с carinthiaca Veronica glandulosa Deletion | 67 Ug Hebe vernicosa Insertion 3 68-79 gla Veronica calycina Insertion | 77 t Veronica chamaedrys Deletion l 78 c/g Veronicastrum sibiricum Deletion 1 89 Wulfenia carinthiaca Insertion 44 90-134 aatctagglgtgcaagcccctttgttgagag- lccegegectgetc Pseudolysimachion spicata, Veronica Insertion 27 108-134 gaclagicgagtececcgcictegete anc Moris ipd V. bellidioides, V glandulosa, V. glauca, V. officinalis, \ serpyllifolia, V. urticifolia Veronica bellidioides, V. officinalis, \ Insertion l 162 a urticifolia Derwentia derwentiana, D. nivea, D Insertion l 207 a perfoliata, Parahe * ое. Veronica arguta Veronica macrostachya Deletion 3 208-209 ce Hebe salicifolia Insertion l 17 c Veronica bellidioides, V. glandulosa, Insertion l 234 c/g V. officinalis, V. urticifolia, Wulfenia carinthiaca Veronica persica Deletion l 248 a — ed ‚ anagallis- Insertion l 260 a aque P. — hion spicata, Veronica Insertion l 442 a chamaedrys, V. macrostachya, V. oltensis Veronica urticifolia Insertion l 454 с Veronica anagallis-aquatica Insertion l 467 t Pseudolysimachion spicata, Veroni icc Deletion l 474 g anagallis-aquatica, V. belli divides, fruticulosa, V. glandulosa, y ow V. officinalis, V. saturejoides serpyllifolia, V. urticifolia, Veronicastrum sibiricum, Wulfenia carinthiaca Parahebe canescens Insertion 1 47: g Veronica bellidioides Insertion 2 501 сс ulfenia carinthiaca Deletion l 503 | Heliohebe hulkeana, Н. lavaudiana, Н. Insertion l 508 [ raoulii Veronica persica Deletion l 513 t pendens H. elliptica, Parahebe Deletion 8 583-590 tetegtge ‚ P. brevistylis, P. decora, P. lyallii, rathulata, p vandewateri Parahebe planopetiolata Deletion 7 590—596 саісісе Parahebe canescens Deletion 8 591—598 atctecge Veronica persica Deletion 3 592—594 Heliohebe hulkeana, H. lavaudiana, H. Deletion l 606 g raoulii Parahebe vandewateri Deletion l 609 a Veronica chamaedrys Deletion 2 609-610 Volume 89, Number 1 2002 Wagstaff et al. 45 Diversification of the New Zealand Hebes Table 2. Continued. Insertion/ Taxon deletion Size Position Sequence Parahebe canescens Insertic 3 615—616 cat Veronica chamaedrys Deletion 2 620-621 te Pseudolysimachion spicata Jeletion 2 625-626 ac Veronica chamaedrys, V. Insertion ] 631 a Veronica austriaca, V. olter Insertion 2 638—639 Veronica arguta Insertion 1 640 c one-base insertion and a one-base deletion. Both accessions of Hebe elliptica and nine species of Parahebe have an eight-base deletion that appears to have evolved independently at least three times (Table 2). This deletion is lacking in Parahebe lin- ifolia and in P. catarractae subsp. catarractae and subspecies martinii. The South American accession of Hebe salicifolia has a unique one-base insertion that is lacking in the accession of Hebe salicifolia from New Zealand. The ITS a faster rate than rbcL. The average rate of change per variable site for rbcL was 1.7 (tree length of 532/number of variable sites 144 + 155). The aligned ITS sequences were shorter than rbcL, sequences in our study were evolving at there were more variable sites (331), and the av- erage rate of change per variable site was 3.6 for the ITS region. Parsimony analysis of the ITS-region recovered 6931 maximum parsimony trees distributed in at least two islands of 1213 steps (consistency index — 0.41 excluding uninformative characters, reten- tion index = 0.73); a strict consensus tree is shown in Figure 3 and one of the maximum parsimony trees in Figure 4. Based upon the results from anal- ysis of rbcL sequences (Figs. 1, 2). Veronicastrum sibiricum was designated as the outgroup. The ear- liest divergence within the ingroup is between Wul- fenia carinthiaca and all other taxa. The Northern Hemisphere species of Veronica are found in five clades that form a grade basal to a Southern Hemi- sphere clade comprising the New Zealand hebes and their relatives. A heterogeneous Australian clade comprised of Derwentia, Hebe formosa, Par- ahebe lithophila, Veronica arguta, and V. calycina (8296 jackknife; 18 synapomorphies) is sister to the New Zealand hebes, though there is relatively little support for this relationship (70% jackknife; 10 synapomorphies) (Figs. 3. Six well-supported clades are identified among the New Zealand hebes, but the relationships among these clades are unclear (Figs. 3. 4). The first is a clade that comprises Leonohebe s. str. in- cluding a well-supported group. Hebe tetrasticha, H. cheesemanii, and Н. ciliolata (99% jackknife: 6 synapomorphies) with Hebe cupressoides weakly supported as their sister (61% jackknife; З syna- he Chionohebe A clade (98% jack- knife; 5 synapomorphies) includes Parahebe plan- pomorphies). opetiolata and the cushion-forming species of Chionohebe with both the New Zealand and Aus- tralian accessions of C. ciliolata. The Chionohebe B clade consists of Parahebe trifida and both the New Zealand and Australian accessions of Chio- nohebe densifolia (91% jackknife: 4 phies). The fourth clade includes 6 species of Par- ahebe (93% jackknife; accommodates the informal “Groups Ashwin (in Allan, 1961; Table 1) and P. spathulata. The fifth clade includes all the species of Heliohebe in our analysis (100% jackknife; 13 synapomor- synapomor- 7 synapomorphies) and & C" of phies). The sixth includes the remaining species of Hebe with both New Zealand and South American accessions of Hebe salicifolia and Hebe elliptica (100% jackknife; 11 synapomorphies). Hebe ma- crantha is weakly supported as the sister to the rest of this clade (50% jackknife; 4 synapomorphies) (Figs. 3, 4). DISCUSSION arge, unwieldy genera with a cosmopolitan dis- tribution such as Veronica pose among the most dif- ficult taxonomic problems for plant systematists, whose opinions are often strongly held. One of the most vexing of these problems is the inconsistent means by which taxonomists define generic bound- aries and the recognition of rank within a hierar- chical classification scheme. Recent taxonomic treatments in the Southern Hemisphere have fa- vored narrow circumscriptions, and several new genera have been segregated from Veronica (see Ta- ble 1), whereas taxonomists in Europe and North America have traditionally embraced a broad ge- 46 Annals of the Missouri Botanical Garden ebe elliptica 10 = Дере die bachi ebe chüthamica | ep € ockayneana I Pee Salic gonis Hehe pimeleoiqes townsonii ebe odora Де е pauciramosa ebe epacridea ере ramosissima ebe elliptica Hebe arkert 2 e saticornioides ae armstrongi Heliohebe Parahebe VG cheesemanii be 2 be densifolia ] Chionohebe B di MA al 2 Es i АШ irane ebe ir Chionohebe A Leonohebe E] | aden е tana perfoliata je PW e 20d Ауга 'arahebe oe Hebe e formosa 9 eronica austri 86 Veronica o a mac rostac hya d * ^ 100 а алак», aids a Veronica serpyi — о та carinthiaca onicastrum sibiricum Stric Figure 3. Notable clades are ‚ identifie d with brackets. Jackk nife v ractae subsp. martinii, subsp. hectoril. Wettstein (1891). This discrepancy of opinion contributes to taxo- neric definition of Veronica, e.g., nomic ambiguity and instability. Our results support those of Albach and Chase (2001). implying that the genus Veronica is at best paraphyletic by exclusion of the Southern Hemi- sphere genera Chionohebe, Derwentia, Hebe, He- liohebe, Leonohebe, and Parahebe (Figs. 1. 3). as 'onsensus of 6931 minimal length trees produc ed by parsimony analysis of the entire ITS values > “Parahebe catarractae subsp. rias: ‘Hebe hectorii subsp. subsimilis, ‘Hebe hectorii -region. 90% are given above each node. 'Parahebe catar- well as the Eurasian genera Paederota and Pseu- dolysimachion, and the North American genera Synthyris and Besseya. One possible solution is to lump them all in a broad circumscription of Ve- ronica. This move, however, would create a cas- cade of nomenclatural changes requiring the rec- ognition of many new combinations and the adoption of old combinations within Veronica. We Volume 89, Number 1 Wagstaff et al. 2002 Diversification of the New Zealand Hebes Hebe . Mieres Chile | be pimeleoides ё fownsonu € ana Uo ера i yin ы : en ы! е elliptica slands Бес dt лас E Heb be phonic Chatham v propinqua Islands Hebe сорда 4 eb opodigides he п И be ar ib on i Su bantarctic e * niha Islands rantha ma Lei ШОТ y 1 zhe 'aoulii ebe pe hookeriana 4 Parahebe c atarractae rc arahebe canescens nee ln li nifolia |. р ara New Zealand hebes AAA ter Я Parahebe "che seman a New Guinea f Chie — н Ар Australia trifida инн) e tham uon 5h sbe thomsonii Chionohe be pulvinaris . С Tuonohefe ciliolata Australia Chionohebe ciliolata Par ahebe planopetiolata (13 tetrasticha ciliolata ‘heesemanii ebe с “upressoides a Ver onica arguta Derwentia derwentiana А De De. лла perfoliata Austral 1a nivea Parahebe lithophila Hebe (pi ^ronica austriaca = eronic a mac Tasia hya Veronic гон а ре r SIC a ronica chamaedrys [ Veronica fruticulosa 'eronica edo стая мө spicata E " onicq glau urasi Ver — а be vate 1014 es asia nica officinalis ía т а glandulosa eronica anagallis- їїса —— lo la Wulfenia carinthiaca Veronicastrum sibiricu Ver Ve ronjcan a ur бас — 5changes One of the maximum parsimony trees recovered from a parsimony analysis of the ITS- region. The New nrbs KR comprise a well-supported monophyletic group with outliers on the offshore islands and in Australia, New Guinea, and South America. Dispersal away from the main islands of New Zealand is inferred in nine species. Branch lengths are proportional to the number of changes along each branch. See scale at bottom. 'Parahebe catarractae subsp. martinii, *"Parahebe catarractae subsp. catarractae, *Hebe hectorii subsp. subsimilis, ‘Hebe hectorii subsp. hectorii. sequences and describe patterns of diversification accept that retaining a paraphyletic Veronica ob- in the New Zealand hebes. scures phylogenetic relationships: however, an al- ternative approach is to recognize smaller, less in- clusive clades as generic segregates of Veronica. This approach was adopted by Hong (1984). Here A heterogeneous clade composed of Derwentia, Hebe formosa, Parahebe lithophila, Veronica arguta. MAJOR CLADES OF NEW ZEALAND HEBES we identify major clades supported by the DNA Annals of the Missouri Botanical Garden Figure 5. Plate illustrating some of the morphologic al diversity В * New Zealand hebes. А. Flowering shoot of Derwentia pe rfoliata with toothed leaves obscure in this pict —B. Flowers of Parahebe catarractae. —C. Panic- ulate and terminal inflorescence of Heliohebe raoulii subsp. — P. J. Garnock-Jones 2123. —D. Cushion habit Volume 89, Number 1 2002 Wagstaff et a 49 |. Diversification of the New Zealand Hebes and V. calycina is sister to the New Zealand hebes (Figs. 3, 4). The genus Derwentia includes nine currently accepted species that are endemic to Australia, from southeastern Queensland to Tas- mania and west to Kangaroo Island in South Aus- tralia, where they are found mostly in tableland or cool temperate regions (Briggs & Ehrendorfer, 1992). Toothed leaves [only on the lower branches of H. formosa and often obscure in D. perfoliata (Fig. 5A)], and a dense ring of hairs in the corolla throat (glabrous in H. formosa) are possible synap- omorphies that unite the clade. Derwentia and Hebe formosa also have similar growth forms; their new shoots are initiated at the base of the plant, over- topping older, short-lived branches. Hebe formosa is distinguished from Derwentia by the occurrence of a one-base insertion (Table 2 number of Hebe formosa is n = 21, this possibly being the ancestral state within the Derwentia clade; chromosome numbers of n = 19 or 20 are published for Derwentia (Briggs & Ehrendorfer, 1992) . The chromosome We identify six major clades within the New Zea- land hebes that are supported by the sequence data, which we refer to as the Leonohebe clade, the Chionohebe A clade, the Chionohebe B clade, the Parahebe clade, the Heliohebe clade, and the Hebe clade (Fig. 3). ITS sequences provide strong sup- port for the Leonohebe clade (Figs. 3, 4), comprising Hebe cheesemanii (Fig. 5K), H. ciliolata (Fig. 5J). and H. tetrasticha. Hebe cupressoides (Vig. 5L) is weakly supported as sister to this clade. These spe- cies are endemic to the South Island of New Zea- land. They have traditionally been included їп Hebe, but here, as in previous analyses (Wagstaff & Garnock-Jones, 1998, 2000), are far removed from the other species that Moore (in Allan, 1961) in- cluded in that genus. Hebe cupressoides and mem- bers of the Leonohebe clade lack the dorsal capsule compression typically found in members of Небе, and a potential synapomorphy for the group is pos- session of a distinctive cupressoid growth habit, which has apparently evolved independently from that in the whipcord hebes [e.g.. Н. annulata, Н. armstrongii, H. hectorii, H. lycopodioides, H. pro- pinqua, and H. salicornioides, which have a similar growth form (Figs. 5T-W)]. Hebe cheesemanii, H. ciliolata, and H. tetrasticha, along with H. tumida, comprise the informal group “Semiflagriformes” « Moore (in Allan, 1961). These similar species are subshrubs that occur in rocky areas at high altitude and are characterized by possession of lateral in- florescences (Fig. 5K), dioecious breeding system. and sour-scented flowers. We refer to this group as E the Leonohebe clade, because Leonohebe ciliolata (H. ciliolata) was designated as the nomenclatural type for that genus by Heads (1987). If the group. including H. cupressoides, is treated as generically distinct from Hebe (and other genera of the Hebe complex; see Garnock-Jones, 1993a), then the use of the name Leonohebe seems warranted, though this is a much more restricted use of the name than the clearly polyphyletic circumscription originally employed by Heads (1987). The Chionohebe A clade encompasses the cushion- forming species of Chionohebe. All of these species occur in the South Island where they are high-alpine plants of rock and scree. The cushion-forming species tft of Chionohebe thomsonii, male plant, Eyre Mts., South Island, P. J. Garnock- Jones "d —E. Lateral view of erect tubular flower of че erus pulvinaris, from Takitimu Range, South Island. v а: var. brachyphylla. —F. Habit of a plant ca. 15 em tall from same population as M. Bay » 560. rthur, South Island. —G. Apex of v — shoot from Mt. Arthur, WELT 82554. H, I. He he фай п —Н. үч. of a plant ca. 10 cm tall, same population as M. J. Bayly 795, Mt. St. Patrick, South Island. —I. Brane hlet of ү ae unknown locality. —J. Hebe — branchlet showing ciliolate leaf margins, plant from Mt. Arthur, WEL ES ebe a plant, from same аи as M. J. Bayly 756-757, Black Birch Ra., des Ed —L. Hebe cupressoides, ui anchlet, bern ated , Nelson, South Island, WELT 82553. M, N. Hebe parviflora. —M. Habit of a plant ca. 2.5 Р) J. — d 2257. O, P. F / vu bud, from Mt. J. Bayly S-31. speciosa, — from a plant cultivatec S. Hebe salicifolia, shoot of plant from Upper Wairau a Hebe albicans. —O. subsp. lycopodioides. — qos portion of vegetative branchlet, cultivé rndale, South Island, WELT 82551. Үү 2T. ‚ Lake Tennyson, South Island. —W. vated i a Landcare Research Gardens, Lincoln, үм к T lon Scale bars: A, B, m; C, G, I, tall from same Со as Y J. Garnock-Jones 2258, Hauhangaroa Ra., North Islan Shoot apex — young latera Arthur, South Island, WELT 82555. ‚ Hebe O shoot of plant from М. | b South Island, at Otari-Wiltons Bush, V Habit of plant — same popula d. —N. Inflorescence, inflorescences and large apical Shoot of plant from Cobb ү South Island, М. M. J. B iyly 1478. —R. Hebe Wellington, — — Maunganui Bluff, North Island. У. Н Valley, South Island, M. J. Bayly Hebe ү ition as M. J. y O, Ms HEN, South Island. —U. care ا‎ Garde ns, Lincoln, originally from Fish Lake, nce of a plant from same population as M. J. Bayly )ranc na ч Ice a showing terminal infructescences, cu ы - 1 Nigger Stream, Canterbury, South Island, WELT 8255 mall, light seeds of Hebe elliptica (by pic 'al o PN of most Hase] cultivated, Otari-Wiltons Bush, ier n N = 5 mm; E, V. Q, К, 5 = 2mm: J, К, U; X = I mm O, Р, Annals of the Missouri Botanical Garden of Chionohebe (e.g., Fig. 5D) are united by several synapomorphies, including cushion habit with the de- cussate leaf pairs slightly offset to form a pseudo- spiral (Heads, 1994b), thick-walled bristle-like eglan- dular hairs, solitary flowers with erect. long corolla tubes (e.g., Fig. SE), and corolla veins branching dis- tally in the tube (Garnock-Jones, 1993a). The Chio- nohebe A clade is sister to Parahebe planopetiolata, “Group ol As win (in Allan, 1961). With Parahebe planopetiolata they share a five-lobed corolla and hygrochastic — >) one of the species of Parahebe capsule dehiscence, but the species in the Chio- nohebe B clade also have these features. The Chionohebe B clade includes С. densifolia (Australian and New Zealand populations), plus Parahebe trifida from the informal Parahebe "Group B” of Ashwin (in Allan, 1961) (Figs. 3, 4). Parahebe birleyi is not included in this clade in all trees, and in the consensus tree forms a polytomy with it and several other clades in the complex. The species of the Chionohebe B clade are all alpine plants found in the southern region of the South Island. Parahebe trifida occurs in alpine flushes and snowbanks, P birleyi is a plant of nival rock ledges, and C. densifolia is found in a range of stony alpine habitats. Chionohebe densifolia is also found in the Kosciusco National Park in Australia. Chionohebe densifolia, P. trifida, and P. birleyi are all similar in appearance. Several of their shared characters are likely to be synapomorphies, includ- ing few-flowered inflorescences, large flowers with purple anthers, and presence of long glandular hairs on leaves. Other shared characters are also shared with the Chionohebe A clade, including old leaves withering and fading but retained on stems. All three species of the Chionohebe B clade are thought to form hybrids with cushion-forming spe- Wagstaff & Garnock- There is little evidence from mor- cies of Chionohebe A clade ( 2000). phology to separate the Chionohebe A and B clades. Garnock-Jones (1993a), from a cladistic analysis of morphological and flavonoid data, proposed a more Jones, inclusive “Chionohebe” clade incorporating all the species of Chionohebe and Parahebe “( — Group B” « Ashwin (in Allan, 1961). Such a grouping appears paraphyletic at best in this and earlier ITS studies (Wagstaff & Garnock-Jones, 2000). The Parahebe clade includes all the represen- tatives of the informal Parahebe “Group A" of Ash- win (in Allan, 1961) plus P. spathulata. Parahebe spathulata is anomalous in this clade, which oth- erwise has morphological support from inflores- cence, floral, and flavonoid characters (Garnock- Jones, 19934). clade and P canescens, the sole species of Ashwin's A sister relationship between this "Group C." has 9396 jackknife ‘Support. The infor- mal “Parahebe Groups A and C” share several flo- ral apomorphies (Garnoc igne 1993a). Parahebe brevistylis and P. linifolia (from Ashwin's “Group В”) form a weakly supported small clade, which is sister to the “Group A and С” The clade is represented in both the North and South Islands f New Zealand, species. ^ © where species occur in well- drained soils associated with river banks, cliffs, and screes. Parahebe canescens is a creeping diminutive herb of South Island lake shores; its reduced fea- tures match convergent similarities seen in other plants associated with this habitat. The entire clade except for P. brevistylis and P. spathulata is united by floral features such as short corolla tubes, col- ored nectar guides (Fig. 5B), and stamen filaments narrowed at the base. In the case of P. brevistylis, the differences can be explained as losses of ad- aptations for insect pollination (Garnock-Jones, 1976b). Parahebe spathulata shares some features of habit and flower morphology with P. cheesemanii and might have an allopolyploid origin involving species from the Parahebe clade and the P. chee- semanii lineage. ıe Heliohebe clade (Figs. 3. 4) was formerly recognized as Hebe "Paniculatae" in the informal in Allan, 1961), and later segregated as a distinct genus by Garnock-Jones classification of Moore — (1993b) (Table 1). It was also previously recognized as a distinct group in the key of Cheeseman (1925). Heliohebe includes five species that are found in northeastern parts of the South Island on rock out- crops, cliffs, and sometimes in grassland. Mono- phyly is well supported by two unique indels and several possible morphological apomorphies in- cluding an inflorescence that is a terminal, com- pound raceme or spike (Fig. 5С), protogyny (also evident in Fig. 5C, where styles are protruding from buds on the lowest inflorescence branches), sta- mens erect, anthers cream or yellow, seeds fusiform to irregular in shape and winged, and hemitropous ovules (Garnock-Jones, 1993b). The Hebe clade corresponds to Hebe sensu Moore (in Allan, L961) with the exclusion of Heliohebe and members of the Leonohebe clade. The majority of species within the Hebe clade form a well-support- ed group (100% jackknife value), with weaker sup- port for Hebe macrantha (Fig. 5F, G) and H. petriei as sisters to this group (Figs. 3, 4). The Hebe clade is largely endemic to New Zealand, including many of its surrounding islands, with two species also extending to South America, and one species (H. rapensis, not included in this analysis) endemic to Rapa Island (Fig. 6). The clade is large and mor- phologically diverse; unambiguous morphological Volume 89, Number 1 2002 Wa gstaff et al. 51 Diversification of the New Zealand Hebes + س‎ Y Mas Le 80° э ^ e PP a | 2 Ри ae e^ е y \ \ 4 Kermadec ' Islands MEE i = \ \ * Chatham Islands Figure 6. synapomorphies are difficult to identify, with ab- sences or reversals in some taxa. Wagstaff and Gar- nock-Jones (1998) suggest that synapomorphies may include: a shrubby or arborescent habit (e.g.. Fig. 5M, R, S), large leaf bud (e.g.. leaf margins, protandrous flowers, peltate placentas, Fig. 50). entire acute capsule apices, 3/5 or 5/8 inflorescence phyl- lotaxis. Within the Hebe clade there is little reso- lution, but several relationships are worthy of note. Firstly, the “Connatae” of Moore [in Allan, 1961, represented by H. benthamii, H. petriet, H. epacri- dea, (Fig. 5H. 1) and H. ramosissima] are polyphy- letic; with some members closely related to “Bux- ifoliateae” (H. odora and Н. pauciramosa). These relationships were first suggested by Wagstaff and Garnock-Jones (1998), taxa in this study continues to support this earlier and the addition of further assessment. Secondly, as found by Wagstaff and Wardle (1999), three of the cupressoid species, Н, salicornioides, Н. armstrongii (Fig. 5W), and H. an- nulata (which share possession of fused anterior 21), Thirdly, two of the calyx lobes and chromosome number of n = form a well-supported clade. Chatham Island endemics, H. chathamica and H. 3, 4). with dieffenbachit, are sister species (Figs. Map showing postulated dispersal of hebes from the main islands of New Zealand. H. barkeri, ing placed in a large polytomy that includes this the third Chatham Island endemic, be- grouping. Hebe macrantha, placed with weak jack- acks many of the previously mentioned synapomorphies knife support at the base of the Hebe clade. for the group. This species has long held an isolated or ambiguous position, being placed by Moore (in "Grandiflo- by Heads (1987, 1994b). For the present we suggest its re- tention in Hebe, Allan, 1961) in its own grouping, Hebe rae," and was included in Parahebe with 1993b) and the Leonohebe clade, monophyletic (see which still leaves the genus, the exclusion of Heliohebe (Garnock-Jones. Table TAXA OF UNCERTAIN AFFINITIES Many of the genera and subgeneric groupings historically recognized in the Hebe complex (e.g Heliohebe, cushion-forming species of Chionohebe, and some Hebe groups) are shown to be monophy- letic in this study. Significant — are Par- 1 Allan, 1961) and species groups within Hebe les ‚ sects. Hebe, Sub- “( ahebe “Group В” (of Ashwin Annals of the Missouri Botanical Garden distichae, Glaucae, and the informal group “Con- natae" of Moore (in Allan, 1961)] (see Table 1). The scattered, forming a grade among the New Zealand * species of Parahebe “Group B” are widely hebes, and few show any close relationship to other members within their informal group. Previous au- thors (Ashwin in Allan, 1961; Garnock-Jones, 1976b, 1993a) have suggested the Parahebe “Group B” species exhibit considerable morpholog- ical similarity, but now it seems this. similarity might derive from plesiomorphic character states (relative to the Hebe—Heliohebe clade), from con- vergent evolution of floral features associated with 1976b), and per- haps from reticulate evolution in P. spathulata (dis- self-pollination (Garnock-Jones, cussed above). Garnock-Jones (1993a) suggested that Parahebe “Group B” and Chionohebe should be united, but this view receives no support from the ITS analyses (Wagstaff & Garnock-Jones, 2000, and herein). Wagstaff and Garnock-Jones (2000) in- ferred that the ancestors of Parahebe and Chionoh- — ebe initially evolved in a montane or alpine envi- ronment, then subsequently radiated into lowland environments during episodes of Pleistocene gla- ciation, Extinction probably had a more profound effect on the basal lineages of Parahebe and Chio- nohebe, and this process further confounds our ef- forts to resolve relationships. The relationships among species of Hebe remain ter ITS lengths are too short for us to confidently derive an uncertain. after analyses (Figs. 3, 4). Branch infrageneric classification from this study. It may be that speciation and diversification in the Hebe clade is too recent for ITS divergence to reveal its phylogenetic pattern. Reticulate evolution might also have clouded the molecular signal either through diploid hybrid speciation or allopolyploidy in this group where about 32% of the species are polyploid. ORIGIN, DIVERSIFICATION, AND DISPERSAL Two widely differing opinions have been pre- sented regarding the age and origins of the New Zealand hebes. Some authors (e.g., Skipworth, 1973; Heads, 1994a) have proposed a Gondwanan origin to account for the present distribution of the , Raven, 1973; Garnock- 1993a; Wagstaff & Garnock-Jones, 1998) have suggested the group has arrived more recently group, whereas others (e.g. Jones, in the Southern Hemisphere and that dispersal has played an important role in shaping its distribution. The data presented here lend support to the latter proposition. A primary contribution of the present work is its assessment of relationships between Veronica and Al- though a close relationship between these two Australasian members of the Hebe complex. groups has long been assumed on morphological grounds (e.g.. Wettstein, 1891; Cheeseman, 1925; Raven, 1973), some authors (e.g., Hong, 1984) have directly opposed the notion that Australasian taxa are derived directly from within Veronica, Our anal- ysis of rbcL clearly supports a close relationship of Veronica to the Australasian genera, with the strict consensus including a well-supported (100% jack- knife) Hebe—Veronica clade (Fig. 3). Analysis of ITS sequences shows the Australasian members to form a clade nested within a paraphyletic Veronica. This pattern of relationships is congruent with the notion that the Hebe complex is an Australasian radiation of Veronica. Our assessment of the time frame for the origin and radiation of the Australasian genera, and the New Zealand hebes in particular, relies on infer- ences from the fossil record, the geological and cli- matic history of New Zealand, and the distributions and ecological tolerances of extant species. For Hebe, the earliest appearance in the fossil record is in the Pliocene (Mildenhall, 1980), for Scrophular- laceae it is in the mid Miocene (Tiffney, 1985), and for the whole of the Lamiales (sensu APG, 1998) it is in the mid Eocene (Muller, 1981). Although there is always the possibility that older fossils will be found, it would be inconsistent with this record to assume that divergence, either within the Austra- lasian Hebe complex, or between members of that group and Veronica, occurred in Gondwanan (Cre- taceous) or earlier times. We acknowledge that the divergence estimates presented in Figure 2 are crude and await further refinement. The paucity of the fossil record and the occurrence of undetected multiple substitutions on Figure 7. of trees produced by analysis of ITS sequences. Species are regarded as alpine if > Details of the natural elevational range of species of New Zealand hebes, overlaid on the strict consensus Sp they occur predominantly in areas above the natural tree line. The — two elevational zones are less precisely defined with, on the two main islands of New Zealand, lowland corresponding to those areas below ca. 500 m above sea level, and montane being those between ca. 500 m and up to ca. 1000 occur in alpine or montane environments. 1 (but below natural tree line). Members of the basal grade of New Zealand hebes all Wagstaff et al. 53 Volume 89, Number 1 2002 Diversification of the New Zealand Hebes sapiossa4dno әфән E пиршәхәәцә әфән m 01р[0112 ӘДӘ та DYINSDAJA] әфән m p1pjoi12doup|d әфәцрлра ка 01]12 24240101417 m 3 D]DJO1]12 242940101417 ша : siupuiajnd aqayouly) m MUOSIUOY] 2q210u011/7) та 1491414 AGAYDAD та Юр1141 AGAYDAD m рцојхиәр aqayouolyy) m pijofisuop aqayouoly) m пиршәхәәуцә AGAYDAD та мәрмәрира IQOYDAD E SIJAISIADAO әфәурлра ES pijofiui] 2QIYDADA Montane unordered С] Lowland ШШ Alpine EA uncertain E equivocal Elevational Range SUPISAUDI AGAYVDAVE mE ADIIDAADIDI IQIYDAD f E 11104] AGAYDAD ЕЗ pipjnyinds AqayvAdd ra ا‎ 2424P4D A we 1nopa 24210112] кау pupipnp. * әфәцоцән ка DUDIAMY әдәцо1әН{ E DIUDAODIU m = — 228 та d әдә m ТШ 2q2]] m ay aqai m уйно ^4 2q2H m [1401221| 2q2]] m pnbuido4d әфән m Ds021u424 29H NSUOAISIUAD 2q2]] m юрурупиир aqa = SIPIOVMAODM]DS 2q2H 14244Dq әфән са pDUDAUADYIOI әфән та DINUDYIDYO әфән a ijo»quo[[oip әфән E DLUISSISOLUDA 242 H на papiuoDdo әдә] та DSOUDAL. — әдә m DAO nups1440: ?^2qoH кз Annals of the Missouri Botanical Garden long branches are two potentially significant sourc- es of error, both of which could lead to inaccurate estimates of divergence times. Though crude, the estimates nonetheless provide intriguing compari- sons with patterns of ecological diversification and aspects of the geological history of New Zealand. The geological and climatic history of New Zea- land suggests that differentiation of the New Zea- land hebes is likely to have occurred in the late Tertiary. Members of the basal grade of New Zea- land hebes, in particular the Leonohebe clade, and the Chionohebe A and B clades (Fig. 3), are all alpine or montane plants, most occurring in areas above the natural tree line (Fig. 7). If the present ecological requirements of these groups are indic- ative of those of their past (i.e., assuming that each lineage has not independently and recently adapted to alpine habitats, or that each has not seen selec- tive extinction of lowland members), it can be in- ferred that hebes occurred in alpine environments, with colo- early differentiation of New Zealand nization of the lowlands being a secondary event. The evidence is that alpine environments have only existed in New Zealand since the Pliocene or latest Miocene, subsequent to the onset. of mountain building, in what was previously relatively low-ly- 1979; Ollier, 1986). Prerequisite in any hypothesis of a late Tertiary ing land (Flemming, origin for the New Zealand hebes is colonization of New Zealand by long-distance dispersal of ances- tral form(s). Assuming a minimum number of dis- persals or extinctions, the topology of cladograms derived from ITS sequences (Figs. 3, 4) suggests that differentiation of New Zealand hebes followed a single colonization from either Australia or Eur- asia. This differentiation was succeeded by second- ary dispersal from New Zealand. As illustrated here and elsewhere (Wagstaff & Garnock-Jones, 1998, ITS sequences (Fig. 3) of extant species with trans- 2000) the morphology and oceanic distributions provide evidence of the ca- pacity of New Zealand hebes for long-distance dis- persal. Such transoceanic distributions are seen in four species, all of which are included in our ITS study group, and all of PEO h are nested within the well-supported clade of New Zealand hebes. Of these species, Hebe — and H. salicifolia nat- urally occur both in South America and southern New Zealand, whereas Chionohebe ciliolata and C. densifolia occur both in the South Island of New Zealand and southeastern Australia. Within each of these four species, populations separated by ocean gaps show no apparent morphological differentia- tion and only limited sequence divergence (Fig. 4). This suggests that the transoceanic disjunctions in the distributions of these species are relatively re- cent phenomena, and given that these species and all of their closest relatives occur within New Zea- land, that these disjunctions are products of long- distance dispersal of propagules from New Zealand. Godley (1967) suggested oceanic birds as likely vectors for the dispersal of seeds of H. elliptica (Fig. 5X) and Н. salicifolia from New Zealand to South America. Trans-Tasman dispersal of the two alpine Chionohebe species from New Zealand to Australia is less intuitively explained owing to the sexual di- morphism of one species (Delph, 1988, 1990), and their splash cup method of seed dispersal (Gar- nock-Jones, 1993a). The implied direction of dis- persal from New Zealand to Australia is also against the prevailing westerly winds but, as noted by Wardle (1978), weather conditions sometimes occur in which the usual direction of winds across the independent dispersals of Chionohebe from New Tasman Sea is reversed. The occurrence of two Zealand to Australia may seem unlikely, but the alternative explanations are either an extended pe- riod of stasis in both morphology and ITS sequenc- es (assuming distributions produced by fragmenta- tion of Gondwana), or widespread extinction in Australia (assuming dispersal in the opposite di- rection) Apart from the dispersal prerequisite to explain transoceanic species distributions, another six dis- persal events from the main islands of New Zealand are required to explain the current distribution of Hebe (Vig. 6). Most of the postulated dispersals are to New Zealand's the Pleistocene-age (Sykes, 197 outlying islands, including one to 7) Kermadec lIs- — ands (where Hebe breviracemosa is endemic), and three to islands of the New Zealand subantarctic (where H. elliptica and Н. odora have populations disjunct from those on the main islands of New Zealand, and Н. benthamii is endemic). One dis- persal to the Chatham Islands has been postulated on morphological grounds (Moore in Allan, 1961; Garnock-Jones, 1976a; Wagstaff & Garnock-Jones, 1998). This is partially supported here by analysis of ITS endemic species, H. chathamica and Н. dieffenba- sequences (Fig. 4), which places two of the chii, as sister taxa, and the third, A. barkeri, in the polytomy that includes the branch uniting the other two. A final dispersal, probably from the Chatham Islands (Garnock-Jones, 1976a, 1993a), is also pos- tulated to account for the distribution of H. rapensis (not included in our analysis), which is endemic on Rapa Island in French Polynesia. The presence of Parahebe in New Guinea is dif- ficult to explain. Here, as in the analysis of Wagstaff and Garnock-Jones (2000), ITS sequence data for Volume 89, Number 1 2002 Wagstaff et al. E aa of the New Zealand Hebes only one New Guinean species of Parahebe were included. That species, Р. vanderwateri, is nested within the New Zealand hebes (Fig. 3) with 99% jackknife support. The most parsimonious interpre- tation of the present data (assuming a minimum number of dispersals or extinctions) is long-dis- tance dispersal from New Zealand to New Guinea, as proposed by Wagstaff and Garnock-Jones (2000). Further sequence data for New Guinean Parahebe (of which 12 species are described) and Detzneria (monotypic and endemic) might provide a clearer ationships between taxa from the two — picture of re areas. Literature Cited Albach, D. C. ‚ W. Chase. 2001. Paraphyly of Veron- ica (Veroniceae: i. rophulariaceae): Evidence from in- ternal transe — — d мер nces of nuclear ri- us 11 iig й ВІ; 9-18. Albert, Bac luni K. генно М. W. Chase. J. R. — D. Mishler & К. C. Nixon. 1994. Fuss - tional constraints bs rbcL — nce for land plant phy- logen . Missouri Bot. Gard. 81: 534—567. Allan, H н. 1961. Flora of New Ze d Vol. 1. Govern- ment Printer. Wellin ngto APG. 1998 [1999]. An r classification for the fam- ilies of flowering plants. Ann, Missouri Bot. Gard. 85: —-553. eem B. $ M. J. — M. n M. F. Woj- iechowski, C. S. Campbell & J. e 1995. The ITS region of nuclear — DNA: A valuable source of evide — on — phylogeny. Ann. Mis- souri Bot. Gard. & 27 Bentham, G. 1876. ^ vini Vr MN Pp. 913-980 in G. Bentham & J. D. Hooker (editors), Genera Plantarum, 7-277. Vol. 2. Reeve, London. Bremer, К. & M. Н. С. Gustafsson. 1997. East Gondwana ancestry of the sunflower alliance a families. Proc Natl. Ac p Sci. U.S.A. 94: 9188-919 Briggs, В. G. & К. Ehrendorfer. 1992. 4 revision of the Austr hs species of а and Parahebe (Scro- phulariaceae). Telopea 5: 241—2t Carlquist, S. 1974. Island Biology. ИР" Univ. Press, Y Ne rk. Chase, M. W., D Soltis. R. G. Olmstead, D. Morgan, D. H. Les, B. D Mishler. M. R. Duvall, R Price, H. G. Hills, Y.-L. Qiu. K. A. Kron, J. H. Rettig, E Conti, J. e nns J. n. Manhart, K. J. Sytsma, H. J. Michaels. W. J. Kress, К. G. Karol, W. D. Clark, М. Hedrén, A S. Gaut, R. K. Jansen, K.-J. Kim, C. F. s, Q.- Wimpee, J. F. Smith, G. R. Furnier, S. H. Straus Xiang, G. M. Plunkett, P. S. Soltis, S S. К. Williams, P. A. oo С. t Quinn, L. E. Golenberg, G. H. Learn, ge Graham, 5. С. Barrett, S. E dee no UR 1993. Phylo- genelies of seed. plants: An ees of nucleotide se- quences — the plastid к gene rbcL. Ann. Missouri Bot. 520-50 . M. Swensen, E. Eg не Н. Gard. 80: 528- Cheeseman, T F. pel — of the New Zealand Flora, p 5 rnment Printer, Wellington. Deg 1988. The Evolution and Maintenance of Gender еей in New Zealand Hebe (Scrophular- iaceae). Unpublished Ph.D. Thesis, University of Can- — ыр һигеһ The А of gender dimorphism in New а Hebe (Scrophulariaceae). Evol. Trends PI. 4: 85-97, Doyle, J. J. € J. L. Doyle. 1987. A rapid DNA isolation procedure for sal pousse of fresh m tissue. Phy- tochem. Bull. Soc. Ames 19: 11-15 "ama, J. S., V. E Albert t. M. Källersjö, D. Lipscomb & . Kluge. 1996. Parsimony nk outperforms s Il: — — of New Auckland. Brown) а! neighbor-joining 2 — stic Flemming, C. A. The кы and Its Ж. Pu Капа Univ. Press. Garnock-Jones, P. 976a. Hebe rapensis (F. Garn-Jones — nov. and its relationships. New Zea- land J. Bot. e 79-83. . 1976b. Breeding systems and po " New Zealand Paral (Se rophulariac eae). New Zealand J. Bot. 14: 291— em За. D idee ny of the Hebe e — ET roniceae). Austra „ Syet, Bot J3b. Heliohebe (S (Ser a new genus í segregated from Hebe. New Ze b J. Bot. (Se — : 457-479 roniceae \. lariaceae: 31: 323-3: Godley, К. ñ 1967. Widely i ta г species, land bridges Үк continental drift. ге 214: 74—75. Heads, М. J. P New names in — Zealand Scrophu- Bot. . Otago News 5: 4—11. ende 'eae. 1994a. E Ау are iter и and taxonomy in the Hebe sum on pee à eae). ull. Natl. Hist. Nat., B, Adansonia 16: 994b. biogeographic ae). Bol. J FE | \ t Hershkovitz, M. A. & E. A. 4 теме w of Parahebe . 115: 65-89. . Conservation patterns in angiosperm rDNA sequences. Nucl. Acids Res. 24: qe d чер P. K.. М. H. Holmgren & L. C. Barnett. 1990. Index РЕ ЯВ: The dh of the World, 8th ed. iq Hong, D. 1984. Taxonomy and evolution of the Veroniceae (Sc — eae) with s special reference to palynology. Opera Bot. 75: 5—00. oe M., J. S. Farris, M. W. Chase, ay, C. J. Humphries, G. Petersen, ы 1998. Анесов parsimony jackknife i ysis of 2538 rbcL DNA es reveals support for major clades of green — land — seed plants and — plants. Pl. Syst. Evol. 213: 259-287. Lin, C. M., Z. Q. Liu & 5. D "King. 1986. P deii roplast genome: Correlation between the DNA sequenc- es RÀ the isoelectric focusing pattern of the LS of rub- lec. Biol. 6: 81—87. Liu Tih- im & C. L. Schardl. 1994. quence in inte mal transcribed spacer l of plant nue * ar rRNA genes. PI. . Biol. 26: 775-778. Mildenhall, D. С. ‘oa. Ne aw Zealand Late Cretaceous and Cenozoic plant biogeography: A $e Palaeo- Part 1, О. Veg. B. Bremer, M. F. K Seberg & sequence - A conserved s :eogr., Palaeoclimatol., Palaeoecol. 31: 197-23: Muller, J. 19 Fossil pollen rec wu at extant angio- sperms. Bot. Kev. (London) 47: 1— Ollier, C. D. 1986. F origin of id xoc in Aus- tralasia. Pp. 3-25 in B. A. Barlow (editor). Flora and Fauna of Alpine ba. Ages and um CSIRO 1 Melbourne. Olmstead, Н. G. & P. A. Reeves. 1995. Polyphyletic origin of the Sc lara eae: Evidence from je y ndhk Ann. Missouri Bot. Gard. 82: —193. Australia, sequences. Annals of the Missouri Botanical Garden „Н. J. Michaels, К. M. Scott & J. D. Palmer. 1992, Monophyly of the Asteridae and identification of their major — inferred from DNA sequences of rbcL. Ann. Missouri Bot. Gard. 79: 249-265. —- .. M. Scott & J. D. Palmer. 1993, А parsimony analysis of the Asteridae sensu lato based on rbcL sequences. Ann. Missouri a — 80: 700—722. — ———-, C. W. DePamphilis, A. D. . N. D. Young & Р. A. Reeves. 2001. Disinte ie of dhe Scrophulari- aceae. — i — 88: 163-212. Pole, M. 19 > New Zealand flora—Entirely long- distance ай Г . Biogeogr. 21: 625-635. Haven, P. H. 1973. Ev olution of the subalpine and e ри groups in New Zealand. New Zealand J. I { —2( Re a J. L., W. S. Judd € R. G. Olmstead. 1999, Pro- posal to conserve the name Antirrhinaceae and with : superconservation proposal to — Анпесе over Plantaginaceae (Magnoliophyta). Taxon 48: 182. Continental drift and Ne ^w A 'ü- . New с and J. Geogr. 57: 1-1: 1998. PAUP*: Phyloge — — Us- ing Seed ( кыз d methods). Vers. Sunderland, Ma assachus Sykes, W. 7. Kermadec Island Flora, an Annotated Checklist. Ne. W Ze aland Department of Scientific and 4. Sinauer, Industrial Research, Bull. Wellington. Eve J. D. . Gibson, F. Plewniak, F. Jeanmougin & D. нон 1997. The CLUSTAL_X windows in- s F lexible strategies for multiple sequence align- ment aided by quality analysis tools. Nucl. Acids Res. 25: 4870-4882. Tiffney, B. H. 1985. Perspectives on the origin of the flo- ristic similarity between Eastern Asia and Americ а, J. Arnold Arbor. 66: 73-94. Wagstaff, S. J. & P. J. — Jones. 1998. Evolution and bioge pla of the Hebe complex (Se WAV e x i inferred from Is. sequences. New Zealand J. Bot. ime 219. Government Printer, eastern North . 2000. Patterns of diversification in Chionohebe and ‘Porahabe (Se ра inferred from 1 F s — ices. New Zealand J. Bo 389-407. — Wardle. 1999, Whipcord he be — mat⸗ ics, dian ecology and evolution. New Zealand J. Bot. : 17-3 Wardle, ÉS 1978. — of the New Zealand mountain flo- ra, with special reference to trans-Tasman relationships. New Zealand J. Bot. 16: 535—550. wes in, R. von. 1891. Sc — eae. Pp. 39-107 in . Engler & K. Prantl (editors), Die natürlichen Pflan- ze ue т, Vol. 4, part За. W. Engelmann, Leipzig. Wagstaff et al. Volume 89, Number 1 2002 tion of the New Zealand Hebes IVersifica D OLOPEOAV €c89c0AV LOOPEOAV Fe89c0AV 1891 UI 029€c LIV 889111 c29tc LIV €£29t£cLTIV OVOOZZA V T8ELEOAV I9€2£04V 6t06cc V 8t06cc1V CLELEOAV OVSYEOAY LEOGECAYV LEOOZZA V (0007) souof -yoouresy) Y JLISSEN (8661) *»uof -YOOUIRL) Y JJEISBE үү (8661) sauof -YOOUIE у jjersaew (0007) Te 19 peojsui[() (000) seuof -yoourey) Y Jase (000) səuof -ҳоошех) Y jeissew (8661) sauof -ҳоошвх) Y JJPISSPNM 1oded Бї] (000Z) səuof -JJoue Y Jessen, (000) sauof -yoouresy) Y JJPISBE ү (0007) “Je 19 peaswjo — 2661) [E 19 peaisui[() (0002) “Te 19 peairsuio — (£661) ‘Je 19 peoisuio (0007) [€ 19 peaisumo (0002) “18 19 Рәш) corels HHD 98tcl€ HH? 69tcl€ YH C8PelS ЧНО 9c696t HH) PpovelS HHO CBL9Eo HHO 8esoor YH) 189806 YH) 69619 ео 985819 €6/TSS 9ECOLD OF tO окар uapae) [p1uourLrodx*g цәлвәв -ƏH atRopuey рәли) Сү [VHLISOV SOI T6 fpisany пәрлесу үешәшиәахУ цәлвәѕ -ƏY әвәриет рәли) "VFTVHLISOV 08122 uosqo(q iL "y FIUEWSE] "VIVA LSOV 6821'8 sau0f-190U.1D1) f a ‘upusaH а ‘Yed 904 '(INV'IVAZ WAN £6/[€€ SIYÁS Y M *jurod 3117 juuns чпцну “YA “UOSTON (IN VIVHZ MAN 0666 “Чоон d Jseurdy tmuoyoH, CUNVIVAZ WAN FUOLSUAY сри OOLE ye кәү әшчүү OISNLISON әзиру шер “SSP FIOOUWON CVETIVHISTV £6 “dy zz “ploy у шох] `Ч 2P4 d "9sueM 319qp "puepsow "(CN VTVHZ MAN 0029 uosduig Y TN 8304) uo nue H ‘puowog пә ^erueuisep үүнү (ASA) Апола veugrpu] “HOY xo (CCR иәрдорү M 7) SISS sipiduing op Mo (NNOD) ZSIZ YM (OVA) ZEr06 sipiydumqgop M 7) (ASA) URSIYOI N Jo AJISIOA чий “HOY хә P0706 YA WD LP M 7 (DVd) £o ro sipidumgop M 7) (эма) ££706 sipidungop M 7) yey (14 Y) pinyofied niuomao(q р а y s33LIg 9 “Y (Три) veu ю1уиәлмә(] puaiq Ņ s33ug су ч (swoipuy) рирциәлмәр DYUINLA] Три p1oyfipupas 018и11]07) puaiq4 y sssug с) 79 (ueueqong) znuosuiojy] 242401017) pusiqq Y s3311g 2) ч UJ "100H) siubura nd aqayouoly’) "pualyg Ņ Sodlig 299 Сөп У) pijofisuop aqayouoly’) “‘puslyy y БИСЕ 9 Ч (TENA У) pijofisuap aqayouory’) ‘puslyy ow 833119 “үчү? чоон) 01010119 aqeyouoly’) puaiq y s33LIg °“) 9 CJ 00H) 217701712 aqauouonj?) "ү pnbijgo 2u0]217) qeq ‘puss using pypAydosajay 2421111107) "Чоң "p ^g pupiuoadD2 ndoong “т snou шпицітиү “‘yWueg suaosaqnd piuojasuy зло], snjpsnd snyjupiyduy “1904 Sil UOISSIII® que gus) UOH BIO ainje19jr] UOISSIDIB UnLreq. 9H UOISSIDIR U3p.re4) a21n0s/411[P20] [еши sato9da "sip шол I[QPAVLI] ol UOT]EULIOJUT UOot]29[[09 3]3]duioo чим (ишпүоә ципо 395) 19qUunNUu UOISSIDIP anbrun В ӘАТӘдӘ1 WALI GAY HH) ә ul suauttoedg “(0661 “Те 19 US UOH) wundoq H хәри] моо) 'sasaujuaed ut "SUOT]EIASIQ(QE Untcrec.r9 H ‘Арп INO Ul p2pn[our әвәәвиеүпцЧоләс jo soroads ou 10] sSIoquinu UODISSIDIR yurquos) pue UOTPULIOJUT 1IYINOA ‘| xtpueddy Annals of the Missouri Botanical Garden LLOPEOAV VIOTCOAV 800V E0AV ELOPEOAV LE89COAV ScctcIAV 606 LOT COELEOAV CCOTCOAV BLELEOAV P8ELEOAV 60€ LE0AV [S8LEDAV LLELEOAV L8ELEOAV 106774V VLELEOAV £91690.1V 1916904 V ELELEDAV (8661) seuof -Pouw у Jese 1aded siy} (8661) souo[ -ADOUIBO Y JJe1se M (8661) souof - oou.) » jessew (8661) səuof -ҳэошв‹) A еар taded siy (8661) souof -AYOUIBL) Y jersey (8661 ) səuof -ҳәоше‹) Y JPEISBPAY (0007) S9UOf-yooure+) y резем ded siy} (8661 ) səuof -JPOURA Y рер (6661) 3[pI£A Y JJesae ү (6661) PILA Y Jesse yy (8661) səuof -YOOUIBL) Y рер (0007) 7° 19 peaisumo (0007) [e 19 peaijsuio) (€661) "|? 19 pearsum() 26891 ALIJA COPIE HHO OotbtcI€ HHO 8€tcI€ YHD trbcIC HHO pLvVclS HHO cLicI€ YHD bevels HHO OSO89P HHO 9бћсІс ННЭ ESPECIES HHO LOVISES YHD ВРІ HHO ££6€1:) 2£9819 460/19 Калы L6/cóc 16/60 SLOFID 190919 900919 v6/99€ PELIS APAPHPW `N “PI SSEJIE') ‘ривер puPpi[e 4 ‘VNILNAOHV SIOFI uou) ^M q eoiexureN ‘ривер шефецо “INVIVAZ WAN 6861 ?9(qp SI ‘uong `y ‘3109 Pee y Anduaau "(IN VTIVAZ MAN P86] uossav]) “Y g seuof-3o0u40*) ^f. q “ред suoi] “OW PUE Pny quinos “INVIVAZ MAN £881 souof-yoouane) “fy KA әрпинәсу "puepqinos “INVIVAZ MAN 66'p'€] Louooy “Gq ۸2| [PA APOJIABH ‘UuoS[3N INV IVAZ MAN 6661 кәиоГ-уәоили) `f J “UDUIIH 4 “URIS 921027) “YSNOLOGHEW “INVIVAZ MAN swni 'V d Ot£gc[ чәл) y q eoresureN ‘Spue[S] шефеҷо (NVITVMZ WAN 066! AON LE ?2uDM “Y “Y “UH Yew -994 ^pue[s] [[oqdure?) INV IVAZ WAN €g-[c[ uou) Y dg pue[sp uei) “UNWIVAZ AIN CBEOZ хорор og muas “oy лих AMG RO INV IVAZ MAN v661 “eq, “мохаш A PIDA q “ndyeyey oye] Ieu 1941. 3AM “PUBTYIMOS "(IN VIVAZ WAN gg TE A9U00Y (| Je] лзәрүпоң "uos[oN ISIMULION. (IN V'TVHZ WAN F£'06 sipjdumgop MD Jayonoa ou тоту "I uwouxun [[puuaq (718104 *£)) nondijJo әфән uelly y әчАрхәогу (yug) ry2pquaffoip əqəH wy Y JUABADO?) (J *NooH) sapiossaudno əqəH ƏS) NUDÍLLIO) 292 ue[[V w eu&e 1907) (чешәзәәцг)) DUDEUADYI0I aqap uelly Y әчАвңәогу (J ‘YooH) DIDIONII әфән uelly Y JUÁBADO?) (ueueyong) 1uDuas2242 әдә "IV JUÁLADO?) (ueueqong) DINUDYIDYI 29H uP[[V y QUARYIO') (у *400H) 11upquoq әрән JUABADO?) (QUARYIOD) 14240Q ƏQƏH uP||V Y әцАрР 4907) (ASU cg 7f) 272uo0418uun 2q3H ue] -|y y эч $207) (910194) njnjnuun aqa JUABADO?) (21124) SUDIIG]D әфән “XY, »sopid jonas) "] DIJOfIp10) Diunqnqoj^) “| paindind sippnziqq 71291 SLI UOISS32)P yuequo«) uot 1815 21nje13]r] UOISSad0R unuıeq12H UOISSIIIE пәрлес) әәлпозулүүрәо] euB saraadç ‘penunuoyy “р xipuaddy 59 Wagstaff et al. Volume 89, Number 1 2002 tion of the New Zealand Hebes iversitica D — . SS8Tt0AV cV 06ccAV 9910690.1 V — . vy€8TV€£0AV LIOVEOAY — 88€L€01V OIOTEOAV —-.][ L6£2 £04 V — . €£€9FCOAV 9€T0690,1 V 00€ 2 €04V 0916904 V C[OTCOAV —. €E8ELEOAV 698€2£04V — . EO€LEDAV 1oded si (0007) souof -POUA у JSB (6661) [DIEM 5 PRISA 1oded siy} (8661) seuof -4ooude4) Y увів (8661) s9uof -YyooulR4) Y етае 1oded sip (6661) эрюм Y jisse ^ (8661) s2uof -YOOUIPL) Y JLS. (6661) PEN Y уез ү (9661) sauof -yoourrey Y JJPISBE ү (8661) Souof -YDOUIP<) Y це yy (8661) souof - YOUR) y PRISA. 86719 YH) I9tcI€ HHO LVVGIS HHO IprelS ЧНО Реге ЧНО 8971 YHD O8belS ЧНО 18966 HH) O0€tcI€ HHO Oorels ЧНО 62YcI€ ЧНЭ T8 => TIE ЧНО 906219 4802819 881219 08219 218219 0L£819 188819 16/981 68 661 Чош) Г `хә Suape puswuodxy qirase] 3IPopue'[ PARADA 'ANVIVAZ MAN c06 1 SUOS -Y30U104) T d Yoo) OLAS, "pue[uinos "(IN VIVAZ WAN Q9'CC AJUOOY “(| "S19 -шо 1N CAInqaoque?) IN VIVAZ WAN 6861 ириәә} q AIQS1NN [994 “HA шол) «¡eu -IBLIO *uopae«) puswnodxy цолгәвә aleopuey P3۸11") “Q(NWIVAZ WAN 0661 ?9(] t sauof -20u4D*) T d “Stesiny [994 WW ONVTIVHZ WAN 681121 uDuaapg d UP -189 OURO Y9INYOISLIYA) шолу A[[eu -13110 *uapaes үрішәшиәахә yo.eosoy элворив” рәли) (TINV'[VMZ WAN 60'€6 fisa ^f су uuo “Y nejd WORT “UOSTON CUNVIVAZ WAN £0'€o Ліха f s “OEL VW | апав») XINVIIVUWZ WAN xel ady nq d V (eBuey әшцепу jo ands) osury OYOWRDY “ABE SMEH “UNWIVAZ WAN cg61 $9UO[-Y90U1D2) "f. 4| SUY nuire pueipnnos (INVTVMAZ MAN OLEVOLB вшәдивә) *suopaee) oruejog “VIV MLS IV 10°66 ffoiszog ^f ^s `оҹеләр 1N AMARO "(IN V'TVHZ MAN 68'6c[ YP 1994 “HA Чо «¡eu -15110 "uapaez [Piuour1adxa цәолвәвә} элроривт POP ALN “(NV IWVAZ WAN -uo9H d *хләвп!\ ug[[V y JUÁBADO”) (учоон) sapioajound әдә} черү у eudRyoory (ивирцәп ) mud 997 9100] ч “I (uejjy Y ou&eo07)) vsowvnonod oqop] ue] -[V Y vudeyoor (үү) p40jfri40d әдәр QUARYIOT) (у 00H) D40po aqoH] ueT|V y ousroo?) (J 00H) DYJUDLIDUL әдә ue[[Y Y OUARYOOT) ("у YOOH) DYJUDIIDUL 99H] sopioipodoo4] ‘dsqns uelly Y JUABADOS) (J Чооң) sopiorpodoot] əqəH PEN y reise (OSUa]O")) хушхат "dsqns 11401994 99H 1101994 ‘dsqns ue] -[V Y 2useNo07) (р 00H) 11103994 әдә м DSOUIOS әфән перу y QUAPY¬207) ("у Чооң) рәрмәраә 991 ouARyoory (14 "M Ццәччә,] (18104 *£)) Didi] 99H qoe SLI UOISSIIIR quegquo: ) поро әлпрлә1”] UOTSSIIIR WNLIEGLO LY UOISSIIIB Uo pe) QIINOS/AT [RIO] JEU) sotoada ‘panuyuoyy `] xipuadd y Annals of the 60 Missouri Botanical Garden (0007) ѕэчо[ OI6I sauof -YIOULDY "f q 72247) 214] 19dd 7] *urej SpeəH АЧ -Pour Y geiszew c€tcI€ HHD Z68819 -uno[y 2144 pue[qinog “INVIVAZ MAN (S9UO[-"UIBL)) $17Á 151421 9QIYDID шоу "Ww M 6802 sauof-320u404) "f а ‘ш OPEL “Yea (0007) sauof IƏMOJ, JO Y Ənbar ш ue] әлоде *ogue “ALLO PVOOGCAV -Y 20U14) у Jese TOIOZT HHO nuper "puepqnos "(NVTVAZ WAN CHOW CA (19 74 N) 26949 292190404 (S661) єє 894929M Y peoajsui() 261-66 $2122 YP projsui) "I suungqma sumddig (8661) seuof | 9g'c'cc ÁNO] "7 *ә} 600PEDAY — О8Є/ЄО4ЯУ -YOOUIRL) Y Jesse 6StelS HHO C8291 -J4 UN "usnodoqpre w "INVIVAZ MAN Sauof-‘ures) (J ооң) mona 29/01/27 (0007) seuof 286'8 tono ^g “peoy unns *e[nsur sauof €r06ccAV YOURS) IY greissew 28VcI€ HHO 90819 — -uaq sxueg “Amquajuer) UNV'IVAZ MAN -URS (moe) nupipnpap] aqayorjay (8661) sauof 996T TLE 4249 q “a souof 6LELEOAV -YOOUIRL) x JBE O09tcI€ HHO €LIZI) -MY ә ‘Ҷапоо рер “UNVIVAZ MAN UIE) (PNW 74) PuUDIIINY 29240112 8981 sauof-20u40*) `f q “yung ue[[V 8S8PEOAV 1oded sip €€9CO9p HH) 4265/19 ndejesune|y "uospN “INVIVAZ MAN y AUÁRADO?) (J *400H) рѕоәтиләл ән "88 6% ?!pinq "A “Y “15207 uelly Y LC8TPEOAV 1oded siu O8ÞPZSI HHO 111/19 `по5]әМ ISOMYHON "(IN V. TVAZ MAN = 9u&exoo7) (ueuasa24)) HUOSUNO] әдән LOBPEOAV 680102 aon] а "y `чешәѕәәц) | uwy 998V E0 AV jaded sip TEVZIS HHO 911819 YA подив) “INVIVAZ MAN Y SUABADO)) (Y *400H) 24211801121 әдән [ZII 112] COBLEDAV -Dg `[ “Usa "M туо Muesuneja UBITV POSVEDAV toded sry} LEPZIS HHO 180719 pueppony YON “INVIVAZ MAN Y eudeyoory (uung *y) Dso12ads 9дәН (6661) 990196 лон Cg `[ “owuey | uelly Y 6LOVEOAV = СОТ6ООЯУ эрге Y JISIM S'vtcI€ ЧНЭ 20999 '"sseq sxpe[ “AINGIONUPD (QN V'TVAZ MAN 9 9useoo») (y ^xooH) 59р101и1021]05 әдә} (8661) səuof OT'S6 fisan f S S8ELEOAV -YOOUIR) Y Jesse 99PEIS HH) "Ned ерү Cuos[oN. "(IN V'TVAZ MAN [[puuaq (715104 4) n1Jofiorpos aqa (8661) səuof SLOFEOAV 98BELEOAV -Ҹэоше‹) зу geiszew 86891 ПІЛУ М 6661 uoixng ^M PEPY ues “A TIHD [[puuaq ("15104 *£)) DIJOfIDDS əqəH 0616 чош”) ‘f *nqnuoen uosulou [ 9C8T€0AV taded sip 8c9cI€ ЧНЭ 699819 -deL W `Ҷапогодеу “UNVIVAZ MAN 'S [y човЧш “9 DUISSISOUIDA 99H (6661) COE OE ?]p4v4 d “9APT F10۸ | uelly Y 8€T6904V PIE y Jessen 8LPETS ЧНО H8€6/L0I -EW unos “PURTYINOS “(INVIVAZ MAN — 9useoo7) (ueuasaay)) DNbuidosd 99H "|2q4 SLI поцеџә UOISS322P UOISS322P 321nos/A]t[eoo[ теш) sardadg UOISS9200 эшо) ənə r] шпиефәң uapaez) 'penuguo? “| xipuaddy Wagstaff et al. Volume 89, Number 1 2002 Diversification of the New Zealand Hebes OcOT £O AV I€06cc AV 198, €0 AV 0€06cc.4V €6t7€04V 8r06cc4AV LVO6CCAV O98TEOAV LTOOCCV 9F06ECAV 6S8VEDAY POELEOAV 9662 €04 V (0007) зәчо[ -yooures) у JJPISSEN ләава siy} (0007) sauof -ҳәоошве) Y руеіѕаем (8661) seuof -yoouie4) у JASSEN (0007) Sauof -A90UIBA) Y jpessew (0007) sauof -YOOUIR) Y JLISSEN 1oded si (0007) seuof -yvouresy » jesse (0007) seuof -ҳоошвх) y Jase Joded sup (8661) sauof -YOOUIR) Y JISE (8661) sauof -YOOUIR) Y JJEISSB AY 09891 ALIAM 9c9cI€ YHD OPPIE ЧНО EOS HHO LpCOLT HH) LOTECIC ЧНО clc60c HHO 9CcpcIS HHO OtreIc HHO SO/LYI #66219 16/28 1 688119 06/T9c 077 Sauof-yoou -1D4) T d WU 009] “peor Jo do] 1e saul -p[Inq oprsaq “ple Bis вол], *nqodeny "aw ‘nejd NULA CIN VTVHIZ MAN POPE sauof -320u4D-) "D d *AYISIDATU(] eop рәли (INV'IVHZ MAN c6 Sl uosdung ‘N ‘ndneyey eye] API MA ‘uoyssury “PUB]YIMOS "(IN VTVHAZ MAN 609€ [T ?manqd M Сү 1819294 "uA Anque YHON (NVTVAIZ WAN COST uosuwuof “[ 422 - cg SAA 9 9 neg PEN sure еи ənjg ‘speq YHOMJUI AY Jo S Wy ) jnoxoor| suoye pop MOTE xoed uy o uy | SILA unos MON "VITTVHLIS OV 1661 `Чә4 cc ??"(1 d V лоор T dog ang») SAQUIRT) “A9][LA xpo[oaep] nqog) "(IN VIVAZ AIN QI pg souo[-joouan*) ‘f 44 nu -odeny jy "uoisur]ey CINVIVAZ MAN 90°66 кар fF су Kuua Y тро HPe[ CUOS[T2N. (IN VTVHZ. MAN TOO] “Чү gZ vouno] 7] '22uDM '(] ^g "p|[ouy«ous MOQUIBY Jo Ay urseq ‘SULY pneury S ‘UOS[9N “G(NWIVAZ MAN EOPE souof-Yyooulny `f а "pM sseq 10u “TEM unodeuepy “03810 (INVTVAZ WAN SOLOL PSL “PUM cy d FAY BURA, “Ҷапоо ер "(IN V'TV4IZ MAN 0661 ny “s9UOf[ -90UIDO ‘f d *uopu&] axe Jo pu? S ‘mque "(INVTVAZ WAN MIO “Y YH cM Счїшәя) vi nyiods oqoupang ALO "a MOM Cug) орту oqojupang] » 'd AIO “A "Y № (иоѕшоці, ‘S "f uosdurig :0)) njnjoijadounjd aq. png "MIO HOHEM (F ооң) 117794] 290040] ‘pus 1 y s3ug +9 ^g opydoyn) oqoupang AMO “Y HUM (J YooH) ofi] eqoyoing A10 ‘YM CAJA) DUDILIYOOY 242410 UIMYSY юлодәр 2424010 АЦО YOM Счшәң) пиршәѕәәцә 9QIYDAD Y IDIIDLIDIDI “dsqns АЦО "d Y MÁ 3510 4 5) 9D]2DA44D]D2 9Q9YDID] sauo0f-* UIE“) пили ‘dsqns IDIIDLIDIDO 9Q9YDAD] АЦО 'd "d Eu SUIISIUDI 9094010] PY SLI UOISS322E YUE ua) uone 31n]e19]rT UOISSIIOR ш neq H UOISSIIIE пәр1їєс) әздповулуүвдо] TRULSLI() satoadg 'penunuo? сү xipuadd y Annals of the 62 Missouri Botanical Garden €66cItAV TcO€I€AV PCOPEOAV 66661 EAV 900ETEAV 900€ LEAV FO0tICAV £00€ LE AV COBPEOAV OTIO£ICAV 000€ LEAVY CZOVEDAV COBVEDAV IcOTEOAV — LOELEOAV CCOELEAV — 157981 ECOPEOAV c€06cc.1V OLELEOAV (1007) эеэ Y Yeqry (1002) 9Sty) Y ҷову 1aded su (100g) esky Y uoeqry (1007) эеэ) Y uoeqiy (1007) 9s249 Y uoeqry (1007) 29884) Y uoeqiy (1002) sey 9 Чорду 1oded siy} (1007) эѕецэ Y ҷову (1007) 25U) Y qoeqiy taded sy} (8661) səuof -J POUL) Y JJPISEN (1007) ase Y quoeqry (£661) зәләә} Y peasujo (0007) səuof -YOOUIR) y резе (8661) səuof -YOOUIR’) у jesze [ЄВС HHO “моя “u's 22971S YHO NNO us €c9cI€ HHO S9bZIS MHD NNOg “u's c000€€ HH) crvelS HHO 94AN cVEl-BLol MAM L1€0€-696[ MAY PSSPl-EZ6L MAM uuog uəpes) "og 98€T1-026T. MAM Р6/281 192-9661 MAM 6/981 (IM) SOFI NNAS 7] “OMAN “XAMYNL (М) S8€2 ОМИ имо Zune 4 'NIVLIN LV AMO F861 AMI £z sqoo4g q є "]ZIuas07) orq uqe) “Boy “LTV LI (М) ZIZ OMN “мэм ОЯН “UMouyuy (М) Z768 ЭМИ ™aY OAH “AOAAHD 86/Є12 194951] “VANAM uuog чэрв) "log "uMou»xur (М) OIL JAN AIAHPH "SO :pue[sioH "AVAHON 8107098 BLI9{UPT) *suapae) oruejog шолу A[[PUISLIO “USpIeo) [ерушәшиәаху qoıeas -ƏY әлеәрицет PARADA ‘VIIVHLSNY (М) OIL DAN 297242) q f MALE A 1900H 23219q0x20N, suey "VIHISTV uuog UJpJe«) "log "uMouyxur OLEPOLB "wLHOque-) *suopae«) oruejog шолу AT[PUIBLIO “Uapiey [еұшәшиәаху qoaeas -9H әлгәрцет PARANA "VI'TVHISTV 6076 MISA ^f "s 4991) suey cnqiojue?) “(IN VIVAZ MOAN uuog чэрве) "lod "uMouxur "HS p»ojsumn() 8661 “des ZI "uop]xng ^M L] рәлојоә-івәрә лвәи ‘u OSIP “EW ‘IEA u212N 'VANINS MAN [9[c Seuof-yoouswy "f 4 “uu Оф] `ишшп» 1e2u ado[s y "uossuua[ IW SIW эглте‹) “PURTYINOG “(INVIVAZ MAN "uOJOM SISUd}]0 DITUOLIA “I sijpuraffo DITUOLA "p spumo varuosa [PA DAYIDISOLIDUL DIVUOLIA "UIS Y "и 291078 юәтиолә\ ‘YOO psompup]2 3140424 "| Psopmonnaf 02100424 S. (рә DUD Yo 0.1 0194 ug "d DUIJA]DI 0.110494 "[ SOP101P1)J9) n21u042 T Saproipijjeq Da 1 "ү DIDUISND n21400494 ‘Ig y 077540 DJ1U0424 "1 оәотрр-ғ1одтир DJ1u0424 zid() што uonjopunstjopnosqq "| P1DJoo2un] оошту пә -AO7Y d (Пэччәд) MaJDNIPUDA 9(Q91[D4Dd MIO 7H H MA PPH oqoupapg PH SLI UOTSSIIIR YUE f] ud‘) uone 31nje1o1rT UOISS9)2)8 uin LIEq.9H UOISS32)P U3p.e*) 321nos/Aj[eoo] [еши sateds 'penunuo) `p xipuaddy 63 Wagstaff et al. Volume 89, Number 1 2002 Diversification of the New Zealand Hebes €cOt L€AV GIOPEOAV — 80ELOEAV LLO€ Leavy 618980 AV COOE Lt. V CSOT ECOPEOAVY 8L89€0OAV (1002) e8eyo 29 HOPqIV (8661) Sauof -ҳоошвх) Y jpeiszew (100c) 9sty) Y Чәчү 1oded Бї] (1007) 2884) Y uoeqiv (S661) 89429291] Y ркәвщ() iaded sup NNO Us сурст HHO NNO “U's £€£881€ HHO cerels ЧНО uuot] U9p.re*) "lod |9joxrounie?) *uojuewN “VIH IS AV 60'P6 S2UO[ -Yy90U1D%) Г J око f `$ ‘uepıes) [RJUGULIEd XY Youresey a129pue] ppAnIn) “ONVIVAZ MAN uuog uapır<) лоя "uwouxur cce ffs f s ious IAB] BUTE әлциәлрр ‘ALSO әлпүрү SOUAOI) “AINGIOIURT) (IN V'IVAIZ. WAN (NMA) 80pT nnas "Т SAN unouxur PP Ito proisu) TO'T6 [pisso T CS сәличәлре U9") YURASAY 39U31 pue [PAN] noL y "upoour] MG RO INV IVAZ WAN "boef popup DIUI/MAL auuag (VJ) tunonaqis um4]spoiuo4o II d CI 1110 140124 "boef 0170/121111 02140421 "| viJof ¡Adios no1u0424 “SIA sopioloanips D21U0424 “1104 DJIs49d. DIVUOLOA но, D3is4od DJ1U0494 1294 SLI U0ISSIIIE ЧЕ] пә“) uon 31n]e191r] UOTSSIDIIR шивә p UOISSIIIR Uope«) 321nos/A]i[poo| [еши әтәч ‘ponuyuoy | xipuodd y PHYLOGENETIC RECONSTRUCTION OF THE NEOTROPICAL FAMILY QUIINACEAE (MALPIGHIALES) BASED ON MORPHOLOGY WITH REMARKS ON THE EVOLUTION OF AN ANDRODIOECIOUS SEX DISTRIBUTION! Julio V. Schneider, Georg Zizka? Ulf Swenson,? and ABSTRACT Based on morphology, a cladistic analysis of the Neotropical family Quiinaceae (Malpighiales) was performed to generate a hypothesis of the ر‎ netic relationships within the family. 'hnaceae were used as all genera, apart from Lacunaria, species of Oc are monophyletic The monotypic Medusagynaceae and four teroup. Sing equal weights, ins Quiinac eae find strong jackknife support and . Lacunaria receive support only after successive weighting. The aberrant species Lacunaria opposifolia ан Daa pteridophylla are a within their respective genera, — separation of the former as monotypic 1 be discardec is rejected. Froesia is the REDE CUR ally i most distinguished genus and sister to i amazonica, once suggested to be recognized a С once phic ME +; һа iracte bus sized. Lacunaria, ey Ut K ‘rnin ds . t the generic guianensis. и is күп most diverse and derived genus with hig eduction toward smaller нА ‘es and f with certainty. Proposed close relationship of the tw al evel, forms a small but closely related clade with 7: ghly unresolved relationships and numerous — flow wers, fewer stamens, and carpels can be g the evolution of sex distribution, androdioecy was fixed ea b in a common ancestor of — and Touroulia, and subsequently dioecy evolved in Lact ‘unaria. Androdioecy, Froesia, Lacunaria, morphology, phylogenetic analysis, Quiina, Quiinaceae, Touroulia. 8 8 The Quiinaceae are a Neotropical dicotyledonous family of 51 species, including several taxa not yet described. It presently comprises Froesia Pires (5 species), Lacunaria Ducke (10 species), Quiina Aubl. (34 species). and Touroulia Aubl. (2 species). The Quiinaceae occur principally in primary low- land rainforests, with a few species found in pre- — montane and cloud forests, reaching an elevation of about 1500 m a.s.l. The family is distributed from Belize and Jamaica to southern Brazil and Bo- livia with a center of diversity in the Amazon low- and forests (Fig. 1), and it comprises shrubs or medium-sized trees. While the systematic position of the family within the Malpighiales seems to be comparatively clear, recent studies have proposed embedding Quiinaceae in Ochnaceae on the basis of a single gene, the rbcL (Chase et al., 2000; Sa- volainen et al., 2000). this systematic amalgamation may not find support if additional molecular markers ered (e.g., Soltis et However, and/or morphology аге consid- al., 2000; Jansen et al., 2001). The infrafamilial systematics, despite the important contributions of Pires (1948, 1950, 1953, 1960), still raises many questions due to our incomplete knowledge of the family, an issue we wish to ad- dress in this paper. Cireumscription and mutual relationships of the genera of Quiinaceae are issues not hitherto inves- F, G, GH, IAN . LPB, M, MO, NY, by dec бош. — he For: vchungsge meinsc cn (Zi 557 from Helge Ax:son Johnsons К und and the ? Botanik/Paliiobotanik, age 25, 0-6) — Forschungsinstitut Senckenberg yartment of Syste ANN. MISSOURI Bor. GARD. l We 2 ~ pus ‘tors of * peas s — for the га of mate 2 A. AAU, B, BM, BR, CAY, CEN, COL, Г, ОСА, R, RB, S, SP, U, VEN, W, WAG. The study was supported /31), “Se ne — — the Naturforschende Gesellschaft, and » Swedish Natural Science Research Council. thanke d for their input ar ч constructive criticism оп the manuscript. Two anonymous re viewers are and Johann Wolfgang Goethe-Universitüt, Sencke — Frankfurt am Main, Germany. jschne отт жа uni-frankfurt. ге gzizka@sngkw.uni-frankfurt.c Jepi matic Botany, Lund University, Östra Vallgatan 18, 22: Department of Hon Stockholm University, 106 91 Stockholm, Sweden. ulf. а ѕи.ѕе 89: 64—76. 2002. 61 Lund, Sweden. Present address: Volume 89, Number 1 2002 Schneider et al. Quiinaceae EN د شلک "e‏ Lacunaria, Quiina Froesia Figure |. species, respectively: Froesia with 5 tigated in a cladistic framework. Based on anatom- ical data, Froesia was postulated to have a some- what isolated position in the family (Gottwald & Parameswaran, 1967). While Froesia forms a ho- mogeneous, easily identified group, the two Tour- oulia species differ from each other, especially in anatomical and the whether they form a monophyletic group. Delimi- tation of species and generic boundaries are espe- cially difficult to discern in Quiina and Lacunarta. The latter is comprised of a “core group” of similar species plus a few species more or less aberrant. The systematic position of Lacunarta oppositifolia Pires is uncertain, and generic rank for this taxon has been proposed (Pires, in sched.). Similarly, ge- neric rank was proposed for Touroulia amazonica Pires & A. 5. Foster (Pires, in sched.). Quiina pter- idophylla (Radlk.) Pires is another re spe- cies, being intermediate between Quiina and Lac- characters, question arises unaria in several morphological characters. In flowering plants, sex distribution is an impor- tant character for understanding the evolution. of androdioecy and dioecy. Evolutionary hypotheses concerning mating systems have been postulated and reviewed in the light of population. genetics sex allocation theory) (Ross, 1980; Barrett, 1984; (sexual selection, 1978, 1982; Bawa, 1981; Charlesworth, Thomson & Thomson & Brunet, Distribution of the neotropic ‘al family Quiinaceae: species; and Touroulia. Touroulia Lacunaria and Quiina, circumscribing 10 and 35 1990: Richards, 1997; Pannell, 1997) or phyloge- netic 1998). denced by the existence of androdioecy (and di- oecy) in different, rather divergent major taxa, il most lis originated independently and several times from hermaphroditism (Swensen et al., 1998). Careful examination shows that most of these spe- cies are only morphologically androdioecious, but constraints (Swensen et al., As evi- м functionally dioecious because of inviable pollen or indehiscent anthers. Functional androdioecy is very rare in seed plants. According to Swensen et al. (1998) it is known only from four species: Datisca glomerata (C. Presl) Baill. (Datiscaceae), Mercuri- alis annua L. (Euphorbiaceae), Saxifraga cernua А. Gray (Saxifragaceae), and Phillyrea angustifolia L. (Oleaceae). An additional Schizopepon bryoniifolius Maxim. (Cucurbitaceae), was reported by Akimoto et al. (1999). In Quiinaceae, Froesia has perfect flowers, Touroulia and Quiina are an- The mor- species, drodioecious. and Lacunarta is dioecious. phologically androdioecious genus Quiina is probably functionally dioecious too, producing inaperturate pol- len in morphologically hermaphroditic plants (Pires, cited after Kubitzki, 1995 pers. comm.; Schneider, 1998). With respect to the monophyly of Quiinaceae, all species share a characteristic combination of ana- tomical characters of wood and bark (Gottwald & 66 Annals of the Missouri Botanical Garden Table 1. Characters and character states for the Quiinaceae and the outgroups Elvasia, Ouratea (Ochnaceae), and Medusagyne (Medusagynaceae). 8. B) PWNS an -l ہے م للا МЈ‏ 10. 11. 13. 14. 15. 16. ae 18. 19. 1 вә 21. . Stipules free (0); partly fused with deep, subulate lobes (1); completely fused (2). 3. 24. 25. . Stipule margins entire (0); serrulate (1) 27. N N 2: NN N © ډ- ټم ف‎ o ы — . Sepals 2 or 3 (0); 4 (1): 5 to 8 (2). 7. Sepal texture membranous (0); coriaceous (1 Plants not caulirosulate (0); caulirosulate ( Terminal internodes terete (0); laterally conspicuously compressed (1). Leaves alternate (0); opposite (1); verticillate (2). Adult leaves simple (0); compound (1). Leaf rachis absent (0); not alate (1); alate (2) Leaf blade or leaflet ovate (0); elliptic (1); obovate (2 Leaf blade or leaflet glabrous or principal vein pubescent (0); principal vein and leaf surface along the veins м2 T pubescent (1); abaxial surface pubescent (2). Leaf or leaflet apex acuminate to acute (0); obtuse (1); retuse (2). Leaf or leaflet margin revolute (0); flat (1) Leaf or leaflet margin entire (0); minutely serrulate (1); serrate (2). Leaf or pis vein apices not protruding or reaching the margin (0); conspicuously protruding from lamina margin into teeth ( Number of an less than secondary veins (0); more teeth than secondary veins (1); teeth equal in number to secondary veins (2); teeth absent (3). Venation cras} (0); camptodromous (1) Prominence of leaf venation: abaxially more prominent (0); abaxially and adaxially + equally prominent (1); adaxially more prominent (i Intersecondary veins not deve — (О); conspicuously developed (1); present but not conspicuously developed (2). Tertiary veins scalariform (0); parallel (1); plumose-reticulate (2); reticulate (3 Stomata paracytic (0); anomocytic (1). Stomata positioned on a flat abaxial surface (0); in small depressions (1). Petioles canaliculate (0); terete ( Petiole base not conspic ous broadened (0): — broadened, not distinguishable from upper part (1); pul- vinoid broadened, distinguishable from upper part (2). Stipules not interpetiolar (due to alternate phyllotaxis) (0); interpetiolar (1); absent (2). Stipules persistent (О); caducous (1). Stipules glabrous (0); pubescent (1). Stipules sessile (0); stipitate (1). Stipule lobe(s) cuspidate (0); acuminate to acute (1): obtuse (2). Plants —— (O); andromonoecious (1); androdioecious (2); dioecious (3). — — — Inflorescence terminal (0); axillar Inflorescence glabrous (О); inc onspic — pubescent with trichomes mostly —0.2 mm long (1); conspicuously pubescent (velvety, tomentose) with trichomes mostly >0.3 mm long (: Inflorescence branched with one main axis (О); brang hed with several axes, fasciculate (1); unbranched (2). 2. Staminate flowers absent (О); in groups of 1 to 3 on terminal branches (1); in groups of 4 to 12 on terminal branches Pedicel i in upper part slightly widened to cylindrical (0); obconical ( Pedicel articulation between base and 1/3 of the length (0); clearly more than 1/3 of length (1). Pedicel pud part) glabrous (0); pubescent и (1). VI — . Sepals glabrous (О); margin pubescent (1); margin and abaxial surface pubescent (2). . Sepals monomorphic, size equal (0); heteromorphic, outer smaller than inner (1). ) Petal aestivation imbricate (О); contorted ( Petal margin (best observed in bud) glabrous (0); pubescent (1). Petals obovate (О); elliptical ( Petal apices rounded (О); notched (1) Functional stamens in hermaphroditic or staminate flowers 10 to 25 (0); 30 to 80 (1); > 100 (2). Filaments basally free from petals (0); adnate (1). — Anthers (in outline) elongate, linear-oblong (0); narrowly elliptic (1); + round (2 Anthe rs opening by apical pores (0); longitudinal slits (1). . Gynoecium synearpous (0): apocarpous (1). Carpels 2 (0); 3 (1); 4 to 8 (2); 10 to 14 (3); >16 (4). Volume 89, Number 1 2002 Schneider et al. 67 Quiinaceae Table 1l. Continued. О. Ovules per locule 1 (0): 2 (to 4) (1). О); all ovules develop to seeds (1). — E Some OV ules abort Fruit a capsule (0); berry (1): follicle Fruit apex rounded, centrally + concave (0): conical. Fruits not spotted i with yellowish spots D. Fruit pericarp in tra сл ел сл єл єл сл yu Ve wn 56. Fruits glabrous (О); e nt ( (2); schizoc meric — (3 erse section without lacunae (0): inconspicuous lacunae (1); conspicuous lacunae (2 1). ); dry, indehiscent (4). e (1). ) 51. Exocarp of mature fruits smooth (0): longitudinally furrowed by underlying lacunae (1): fruit appearing conspicu- ously ribbed ( 58. Seeds к. p pubescent (1). 59. Seeds longer than 60. Crystalliferous cells abundant, in long rows along the veins (О); wide, ellipsoid (О); as long as wide, globose (1). — not abundant, sometimes in groups near the veins Ol. Crystalliferous cells without special thickening of cell wall (0): always present, with u-shaped thickening (1). 1967). ous leaf venation with very closely spaced tertiary Parameswaran, An unusual craspedodrom- veins is found in Touroulia, Froesia, and Lacunaria (a slightly different pattern is observed in Quiina), a set of characters also regarded as unique to the 1950a. b, 1951: Roth, 1996; Zizka 1999). Touroulia display a characteristic u-shaped family (Foster, & Schneider, Aubl. thickening of the cell wall (cristarque cells) (un- Froesia and gui- anensis published data). Another common character is the presence of mucilaginous cavities, particularly in petioles (Schofield, 1968). and is frequently ob- served in all genera. Quiinaceae have formerly been regarded as part 1981). More re- cently, phylogenetic studies based on DNA se- quence data (rbcL, atpB, and 185) place the family in Malpighiales (Fay et al., 1997; Källersjö et al.. 1998; Litt € Chase, 1998; Nandi et al., 1998: Chase, 1996, This position was ac- cepted in the ordinal classification of the flowering of the order Theales (Cronquist, — pers. comm.). plants recently published by the Angiosperm Phy- logeny Group (APG, 1998). These studies suggest that the closest relatives of Quiinaceae are Och- naceae and/or Medusagynaceae. Despite this affin- ity, the relationship between each of these families is weak. In the study by Nandi et al. (1998), Quiin- aceae are sister to Ochnaceae when based solely on weighted rbcL parsimony. When combined with non-molecular data, Quiinaceae shift and lie basal to the sister pair of Ochnaceae and Medusagyna- ceae. The positioning of Medusagynaceae, Ochna- ceae, and Quiinaceae within the Malpighiales was unstable depending on character sets used. For the forthcoming treatment of Quiinaceae for the Flora Neotropica (Schneider € Zizka, in prep.). the present study based on morphology attempts to resolve questions of generic circumscription, in- cluding morphological synapomorphies. We also in- vestigate the affinities of the aberrant species Lac- and unaria oppositifolia, Quiina pteridophylla, Touroulia amazonica. Finally, we provide a familial interpretation of gynoecium morphology and sex distribution in a phylogenetic perspective. DATA AND METHODS TAXA Ingroup monophyly is necessary for correct root- ing of the tree (Nixon & Carpenter, 1993). The out- the Ochnaceae and Medusa- group was therefore chosen [топ two most closely related families, eynaceae (Medusagyne oppositifolia Baker). Gott- wald and Parameswaran (1967) already pointed out that the most closely related family to Quiinaceae is Ochnaceae, and herein the tribes Elvasieae E nel. — and Ourateeae Baill. of the former E sensu Engler. Hence, two species each of the gen- era Elvasia DC. Aubl. Because information on infrageneric relationships and Ouratea were selected. within the outgroup is lacking, the species included in the analysis were selected according to the cri- teria of (1) their distribution within the geographi- cal range of Quiinaceae, (2) their previous use in cladistic analyses (Amaral, 1991), and (3) the num- ber of specimens available for detailed morpholog- ical and anatomical studies. Within the ingroup, a total of 22 species were Nx as terminal taxa. Since the principal in- tention of the study was the resolution of generic relationships within the family, it was considered redundant to include all species. The present se- lection comprises representatives of all genera and, more importantly, all somewhat aberrant species as well as core representatives from Lacunaria and Quiina. Froesia is, on one hand, a comparatively 68 Annals of the Missouri Botanical Garden homogeneous group; on the other hand, it is the most distinct genus within the family. Thus, three species were chosen and regarded as sufficient be- cause the inclusion of all its species would not change the principal topology of the cladogram. Quiina is represented by 12 out of its 34 species, many of which are difficult to diagnose and circum- scribe because of remaining taxonomic and nomen- clatural problems. CHARACTERS AND CHARACTER STATES Characters were extracted from morphological studies undertaken (see Table 1); the final data ma- trix appears in Table 2. Multistate characters are generally unordered in phylogenetic reconstruction, but if there is reason to believe a character state is intermediate between two other states, such a char- acter may be ordered (Wilkinson, 1992, 1995). In our analysis, only two characters (6 and 8) could be treated as ordered. These refer to leaf outline and leaf apices where an intermediate transition state can be perceived. Unknown character states are coded with a question mark (7) and inapplica- ble states with a dash (-). Character evolution was traced by using the software MacClade (Maddison & Maddison, 1992). Autapomorphies can either be included or ex- cluded in phylogenetic reconstruction. Yeates (1992) argued that removal may also remove infor- mation from a data matrix, but Bryant (1995) dis- agreed, saying that they should be removed. Ar- guments against autapomorphies rest on the fact that they are unique, uninformative, and inflate the index (CI). however, consistency Phylogenetic reconstruc- tion, is much more than consistency in- dices. This study, for example, aims to reconstruct the major evolutionary lineages within Quiinaceae with approximately a third of the species of Quiina sampled. To facilitate future studies using a wider sampling, autapomorphies were therefore included to avoid excluding potential synapomorphies. Polymorphic characters are abundant in Quiin- aceae and can be treated in different ways. For example, they could be coded as missing entries, which then introduce erroneous consistency indices i 1991). phic characters could also be included and scored and tree length (Nixon & Davis, Polymor- for the observed intraspecific variation, albeit a low 1995). Kornet Turner (1999) recommended that polymorphic phylogenetic signal (Wiens, and characters should be coded as plesiomorphic in fa- vor of the observed intraspecific variation unless the ancestral state is unknown. Assessment of the ancestral state, at least in Quiinaceae, is a critical point and generally lacking. Since we believe that polymorphic characters do provide a phylogenetic signal and resolution, they are scored with the ob- served states. Habit. which means that leaves are crowded at the stem Plants are caulirosulate in Froesia, or branch apex (char. 1). Leaves. The presence of compound versus sim- ple leaves (char. 4) is problematic with respect to homology and comparability of character states. For the current analysis the single leaflets of species with compound leaves are regarded as homologous t equivalent units (Raunkiaer, 1934). Furthermore, ~ » simple leaves because they are functionally this reflects the equivalent venation pattern of leaf- lets and simple leaves. Additional support is given by the presence of simple leaves in seedlings of pinnate-leaved Froesia venezuelensis Steyerm. & С. Bunting and pinnatifid leaves in seedlings of simple-leaved Quiina pteridophylla. In Quiina and Lacunaria the leaf margin (char. 10) can be incon- spicuously serrulate, only seen with a strong lens (magnification > 20X). In this case minute inci- sions or papillae are seen. The pattern of leaf ve- nation is a peculiar and particularly important character in Quiinaceae (Foster, 1950a, b, 1951; Roth, 1996). All Quiinaceae exhibit a craspedod- (char. 13: 1979). The presence of conspicuous romous type Hickey, 1973, intersecondary veins (char. classification following 15) discriminates Qui- ina and is regarded as a synapomorphy for the ge- nus. Intersecondary veins have a diameter inter- mediate between secondary and tertiary veins; they originate from the primary vein and do not reach the leaf margins. In camptodromous Medusagyne it is hard to judge the intermediate veins as distinct intersecondaries. For the analysis they are consid- ered as inconspicuously developed intersecondary veins. The tertiary venation pattern (char. 16) is unique to the Quiinaceae. In Froesia, Lacunaria, Touroulia, the tertiary and veins are densely spaced, strongly parallel, percurrent or, more fre- quently, anastomosing at different distances from their origin (e.g.. Zizka & Schneider, 1999). In Qui- ina they are less parallel and more conspicuously branched, anastomosing with intersecondary or ter- tiary veins, This pattern is called plumose-reticu- late according to Foster (1950a, b). A scalariform (ladder-like) pattern is observed in Elvasia. Stipules. Stipules are always interpetiolar in Quiinaceae (char. 21), but the number of stipules or stipular lobes may differ among the genera (char. 22). Stipule number obviously varies with the phyl- lotaxis. In verticillate Lacunaria there is one stipule between neighboring petioles (as in verticillate 0. Schneider et al. Quiinaceae Jaquinu JdJOR.IBY") T OOTTOCOOTO TSOTCOTOO0 театеттоао TOCTOOPTOT OTOOTOOOZT CEOOPCTIOO 'Iqny sisuoupinz Di fmomor 0 OOTTOCOOTO TZOTZOTOOO ELELE OOO TOETOOTTOL OLOO0LOOO0C? q*9000crLO00 SV Y 9S21lq DITUOZDUID DINOMO] 0 OOTTOTOOTO TOOTZTPOOL OPTTPOO0AZ TTCTOOTTOT TTOOCTOOEO 0000TOOTTO еа. тој типп) 0 OOTTOTOTTO TOOTZTPOOT OTATTOOOT? CECLOOLLOL TOTOCTOOEO 0000800110 [ny mdop puim() 0 OOTTOTOOTO TOOTZTOOOT OTTOCTOOTS ea T аео ШЕШ ШШ атообоосоо sang (APLY) Р уЧормә типп, 0 OOTOOTTOTO TAOTZTTIOOL OTCTITOOTT ATCCOOTTOTL TOOOZTOOEO OTOOTOOTTO вәол] Y SAMGg sisuavspd vunne) 0 TOTTTTOTTO TOOTCETOET OECECLOOCC CTCAOTTOOT CLOOCTOOTO ?0*0c00IIO0 sou szsuaoodpio vunne) 0 OPTTOTOOTO TOOTZTPOOT O LEEQUOELL TTCTOOTOOT CLOOCICOIO "0207200190 [AL 2701000 pump) 0 OBYTTOTOOTO TOOTZTPOOT OPTTPOOOTZ ССОО ЕО TOOOZTOOEO 0000100110 [п], »gj&do400ui bum) 0 OOTTOTOTTO TA. DIBIb OD: E OL Ede 0l Tbeb АШ TTOOCTTOCO 0000100110 PUBLI] Y cuoue[q xo nads )]0fi2uo] vunne) 0 OFTTOTOOTO TOOTZTPOOT OTTTTORTTLT PTZTOOPTOL TTOOZTTOTO TTOOADOTTO quy sisuaupmz nunne) 0 TOTOOTTOTO T90ICIOOO0I OTATTEEOCC TICTOPTEOT TODOZTOOLO 0000TOOTTO т, Ppu04 vunne) 0 OTTTOTOOTO TOOTCTPOPT OTTTCOOOTT CLCTOOTTOT TTOOZTOOTO ETOCTOOTTO "qostie) nupidazonao uin) 0 OOTTOTOFETLO TOOTZTOOOT OTTTTOOOTZ TIZTOOTTOT 20002 TTOOTOOTTO "ls `7) "Y DIUOZDUD vunne) 0 TOTOOCTOTO TZOTTOOOBT TIZIATOSTO CEETOOPTOT TLOOT( 0900100190 solid (S2114) prjofiisoddo niDunovT 0 OOTTOCOOTO TEOTZOTOOL OICIIIOOCO CLETOOIICI [9001000€0 000400 “US усу CL) DAYIDISOLIDUL DUDUNIVT 0 OOTTOCOETO TEOTZOPOTE OTETTTRPOZO LEETOOQETE I PEOOTOOOED eooeqoozoo АО) HUB рытрипәр] 0 LOLLOCOOLO ELOTCOTOOO OICIII*0cO LI£IOO?ICI I*"001000cI c00*I00c00 IPN CY[PPY) »40sp2ap nipunov] 0 TOTTOCTOTO LcOLCOLOOOQ OILCIIIOPIO qecroorrct qeoo0roooce c000800200 "US 7) cV (Чї) PIPU Dipunor] I OPOTOTOOZO TTTTTOCTOO TTCTeTOO00 ADOTOOTOTT 00001000c0 COOOTTITOT 'S `9) Y ишәләр sisuajanzauaa 018201] E OPOTOTOOZO LIIIIOCIOO LICICLIOO00 q00LOOLOIL OLOOLOOOZ? C*00€*LLIOI зәл] Dd4DJ14] 015904 T OPOTOTORZO TTTTTOCTOO ТТС Те, ОШОО COOTOOTOTT 0000T000Z0 eegQeeTTIOT zonbse 4 "pou Y neam?) psnffip VISIO] AVAOVNITAO 0002000000 TVOTTOTOOO 0001200020 DU Due С OOOTECTIOO CTLCOTOOTOO 1oxeg piyofipsoddo au{@Dsnpa J AVAIVNADVSAGAN 0 00000000€I 0c00000000 1009200000 09010000-0 0000711000 5000900000 Te Coq) »aograapd pompano T 00000000€T 0200000000 т002с00000 0Р0Тт0000-0 POOOCTIOLL c000100000 Puy (Puny) $US] Danan) 0 0600000070 0000000090 0000P00000 0*000000-0 00000000Е0 0000100000 SI) ('Qoue[q) saproispa]a DISDA]H $ 0600000070 0c00000090 0000P00000 09000000-0 *?00^000000 1000100000 "GOV pajpdoppo DISD] AVADVNHDO O68L9STECT O68L9SPECT 068L9STtCI 068L9STtcCl O68L9STECL 068L9STvtcl 9999959955 SyvVYVVVYVY VEECECEEEE COGCCECGEGG GLELECETEL T Volume 89, Number 1 2002 ‘BPP SUISSIWL = "xugeui əy} UT ‘(QRAORUABRSNPa|A ) au&Spsnpapy — (әвәәвицә()) DANA *mispa]y *dnougjno E/E = 90/10 = P ‘2/0 = 2 7 “1/0 = F S199] Ҷим exe] orydaousjod pue (-) usep e цим papoo әле s9]e]s a[qeorddeur əy} pue ara2eurın() dnoasur əy} Jo siojoe1euo [eorgo[oud.our [9 jo xuqeu eC "c APL 70 Annals of the Missouri Botanical Garden pteridophylla). In opposite-leaved Quiina and Lac- unaria oppositifolia there are four per node (paired between petioles), while in opposite-leaved Tour- oulia only two per node are observed. In the latter, this apparently single stipule is interpreted as a product of paired fusion similar to cases observed in Rubiaceae (Goebel, 1932). In opposite-leaved Froesia, the two stipules per node are deeply di- vided into setose lobes. In Quiinaceae, stipules are present on the terminal node at least. If stipules are generally lacking on the more basal nodes they are coded as caducous (char. 23); if generally pre- sent on the three uppermost nodes they are treated Sex distribution (char. 28) is hetero- geneous in Quiinaceae, with Froesia being bisex- ual, Lacunaria unisexual and dioecious, and Qui- ina and Touroulia morphologically androdioecious. with male and hermaphroditic flowers on different plants. The number of sepals (char. 36) is variable in Quiinaceae. Four sepals are common and usually constant for a species, while species with five or more sepals display a higher variability in number. Less than four sepals is an exceptional condition and therefore given its own state. The hermaphroditic Froesia flower bears numer- ous stamens, frequently more than LOO (char. 44). In the androdioecious genera, the staminate flowers generally produce 30 to 80 stamens, whereas the hermaphroditic flowers normally produce fewer. For coding, only the flowers with functional stamens are included in the analysis. Character 45, filaments free or adnate to the petals, only refers to the her- maphroditic flowers of species of Quiina because in that genus staminate flowers do not show this 48), Froesia is clearly apocarpous while Ouratea is syncarpous, trait. Concerning the gynoecium (char. because apocarpy is only gained during fruit de- velopment (Amaral, 1991). In Lacunaria and Tour- 49) is variable and relatively high (4 to 14), while in Quiina there oulia the number of carpels (char. are commonly 2 carpels. In the latter, only a few species (e.g.. Quiina florida Tul., Q. paraensis Pires & Fróes) show transitional states with up to 5 car- The exocarp is usually longitudinally striate to furrowed by underlying resiniferous la- cunae that are more or less conspicuous in tran- section (char. 57). In Quiina florida and Q. paraen- sis, the exocarp appears quite smooth and is not striated by the lacunae. Additionally, their fruits are spotted by a substance that looks like dried resin- ous exudate (char. 54). Crystalliferous cells. One characteristic feature is the presence of the cristarque cells (char. 61), a term introduced by van Tieghem (1902). These are crystal-bearing cells with a conspicuous u-shaped wall thickening. They are often cited in familial descriptions (Cronquist, 1981, as solitary crystals; Amaral, 1991; Bhattacharyya & Johri, 1998; Ku- bitzki, 1995 pers. comm.) and hence could be er- roneously assumed to be characteristic for the en- tire family. So far we were able to confirm them only for the genera Froesia (F. tricarpa Pires, E venezuelensis) and Touroulia (T. guianesis), in ac- cordance with the findings of Foster (1950b). In addition to these specialized cells, crystal druses can be regularly observed in the leaf tissue (char. 60). PHYLOGENETIC ANALYSIS r € The data matrix in Table 2, containing 27 taxa and 61 characters, was analyzed with PAUP* 4.0 for Macintosh (Swofford, 1998), using the branch and bound algorithm under the assumption of Fitch parsimony (Fitch, 1971). Multiple character states were interpreted as uncertain. An initial search was undertaken with equal weights saving all optimal trees, To evaluate characters with the strongest phy- logenetic signal and to choose among equally par- 1988, 4). succes- 1969) The settings simonious trees. (Carpenter, sive weighting analysis (Farris, was undertaken after the initial search. used the rescaled consistency or RC index (Farris, 1989), set to 1000. The process was reiterated until the A similar and to avoid fractions the base weight was — same tree length was obtained twice. analysis was carried out where characters 6 and 8 were ordered. Branch stability was estimated with Bremer sup- port and jackknife analysis for both the equally weighted and weighted characters. Bremer support is defined as the number of extra steps necessary › lose a clade in the consensus tree, using the converse constraints approach (ат 1994; Kiillersjó et al., 1992; Farris, 1996). it weighted and rescaled branch support values cal- culate the robustness for each branch in the weight- 1994; Gustafsson & To ease the construction of all nec- ed consensus tree (Bremer, 1995). essary constraints, the computer program Auto- Each con- Bremer, Decay 4.0 was used (Eriksson, 1° straint was estimated with a heuristic search of 100 replicates of random additions of the taxa (10 rep- etitions of each replicate), tree bisection-reconnec- tion (TBR) branch swapping, holding five trees at each step, and saving all equally parsimonious trees. Jackknife (Farris & al., 1996) investigates the structure, or phylogenetic signal, in a matrix with- Volume 89, Number 1 Schneider et al. 71 Quiinaceae Felsenstein, — out permutation, contrary to bootstrap 1985). but excludes an assigned fraction of char- acters, here set to 35%. The search strategy was set as for the Bremer support, but with 1000 rep- licates and saving no more than 100 trees. RESULTS The analyses using morphology to infer the phy- logeny of Quiinaceae yielded 1643 most-parsimo- nious trees using all characters as unordered as well as when characters 6 and 8 were ordered. The trees are 163 steps long (when ordered 164 steps) with an RI of 0.775 and a CI of 0.561, or 0.529 when uninformative characters were excluded (Fig. 2). Counting steps within the polymorphic taxa, the trees are 282 steps long, indicating numerous poly- morphic characters. Successive weighting of the characters yielded a stable result of 2 trees after three iterations, 73.206 steps long using unordered characters. This tree is a subset of the initial 1643, resulting in a consen- sus with a much better resolution of the otherwise collapsed genera Lacunaria and Quiina. Consensus of the primary 1643 trees and the 2 trees obtained after successive weighting together with jackknife fractions, Bremer support values, and weighted and rescaled support values, are shown in Figure 2. One of the most-parsimonious trees with characters optimized on the branches is shown in Figure 3. Referring to the unweighted analysis, Quiinaceae form a well-supported monophyletic group, a grade from Froesia at the base to the most derived genus Quiina. Within the family, all genera are monophy- letic except for Lacunaria, a genus completely col- lapsed to a comb. Based on equally weighted char- acters, no resolution within Quiina can be retrieved. The small genus Touroulia forms the sis- ter to the Lacunaria—Quiina complex. Support for Froesia and Quiina must be regarded as strong, while support for Touroulia is moderate. No clear signal for a monophyletic Lacunaria can be re- trieved, and L. oppositifolia attaches as sister to Quiina, a position with low jackknife support of 54%. but not found in the consensus (Fig. 2). As to the weighted analysis, resolution is im- proved and only one trichotomy remains, the one in Froesia. All clades found in the equally weighted analysis are retrieved and, as a general trend, mod- erately or strongly supported groups are often better supported after successive weighting. Froesia, sup- ported by a maximum jackknife value in the equal- ly weighted analysis, needs five extra steps to be lost when unweighted, a value twice as strong after successive weighting. A similar situation is ob- served for Touroulia, and the support for Quiina increases almost three times. Using the weighting approach, Lacunaria is now found to be monophy- letic, although with a Bremer support of only 0.9. DISCUSSION Because a cladistic analysis can only elucidate ingroup relationships, we cannot put forward a hy- pothesis of relationships within the order Malpigh- iales, or determine whether Quiinaceae are sister r sister to a paired Medusagynaceae—Ochnaceae, « to either one separately. Support for Quiinaceae 15 strong, with little difference in support values be- tween equally weighted and weighted characters. This implies that several characters are initially strong and basal on the tree; these non-homopla- sious synapomorphies include pubescent stipules (char, 24 eromorphic sepals (char. 39: 1), and a fruit exocarp with lacunae (char. 55: 1; Fig. 3). Other synapo- morphies for Quiinaceae, addressed in the intro- 1), pubescent sepals (char. 38: 2). het- duction, are the unique leaf venation pattern with densely spaced parallel or plumose-reticulate ter- tiary veins (char. 16) and the interpetiolar stipules char. 21: 1). In a molecular study of the relationships of Me- dusagyne oppositifolia by Fay et al. (1997), Quiin- aceae were represented by Quiina pteridophylla, — Touroulia guianensis, and Lacunaria jenmanit (Oliv.) Ducke, all included in our study. Contrary to our results, Quiina was found to be the most basal taxon in the family. Quiina pteridophylla is a morphologically aberrant species within the genus with some characters resembling species of Lacu- naria, such as the phyllotaxis and the stipule num- ber. Therefore it may be an inappropriate represen- tative of Quiina and could attach as sister to all other genera due to long-branch attraction. Froesia is indicated by our cladistic analysis to be the most basal genus in Quiinaceae, followed by a grade of Touroulia, Lacunaria, and Quiina. Par- ticularly noteworthy is the apocarpous gynoecium (char. 48) of Froesia, a character state known nei- ther from other Quiinaceae nor from Ochnaceae or Medusagynaceae. However, some Ochnaceae ex- hibit a secondary—otherwise called “ecological” Baum, 1951) (Amaral, 1991). In her cladistic analysis, Amaral (1991) interpreted this secondary apocarpy as a de- rived state differing from the states observed in the Quiinaceae and Scytopetalaceae (both families apocarpy during fruit development — therein used as outgroups). Whether apocarpy in Froesia evolved secondarily is difficult to infer. The existence of a compitum, providing support for the 72 Annals of the Missouri Botanical Garden Medusagyne oppositio. Elvasia calophyllea Elvasia elvasioides Ouratea lucens Ouratea parviflora Froesia diffusa апо то - Froesia tricarpa 100/100 5/10.8 [100/100 mM — Touroulia amazonica Froesia venezuelensis 69/87 4/5.5 1 12.6 ~ = Touroulia guianensis -161 | Lacunaria oppositifolia Lacunaria macrostachya – 10.9 78/88 2/3.5 Lacunaria jenmanii Lacunaria crenata Lacunaria decastyla Quiina paraensis Quiina macrophylla [ — Quiina longifolia —— Quiina rh ytidopus Quiina florida Quiina tinifolia г Quiina cruegeriana L— Quiina guianensis Figure 2. morphologic ‘al characters (Tables | ane 2), analyses yielded 1643 trees (163 s other, Dotted — indicate coll; — Г anc ches i in “the prime weights (left) a ranches are Bret mer support values, i. and weighted and rescaled characters (right) view that apocarpy is derived (see Endress, 1982; Kubitzki, 1995 pers. comm.), could not be ob- served. Nevertheless, according to the present anal- ysis, this character state is considered a synapo- morphy for Froesia. The cladistic analysis indicates an isolated po- sition for Froesia in terms of morphology. Support well as weighted — ‘ters (right) from the primary ind .. additional steps needed to c dines. a node for unweighted characters (left). Quiina amazonica Quiina obovata Quiina oiapocensis Quiina pteridophylla Strict consensus tree of Quiinaceae retrieved from a Волаи cladistic analysis applying к qual we ights to econdary analysis using s successive The = two most-parsimonious the гу analysis — a ше fractions above 50% using equal are snc for the genus is strong with no less than eight syn- apomorphies (Fig. 3). These results correspond to the findings of Gottwald and Parameswaran (1967), who even proposed, based on anatomical studies, the establishment of a separate subfamily for Froe- sia. Besides the apocarpous gynoecium, deeply di- vided stipules with setaceous lobes (in all but one 73 Volume 89, Number 1 Schneider et al. 02 Quiinaceae Ф зай 5 qc S © £ y DES TS SG Ф Ф © o Q D EE AS оо о ® кх у jk yO SP лт о ол д сєєк CG CG CG О 90 2:0 0 17:1 — SE 38:2 14:1 as 44:1 „22:2 Td 531 ma 26:1 56:1 4-370 9.0 E691 14:2 25:1 31:1 351 33 36:2 K19:0 6:2 ا‎ — 23:0 20:2 1:1 9:1 5:1 10:1 22:1 12:1 + 40:1 ЩИ 43:1 "2442 щи 48:1 m" 49:1 ща 522 ааз 30:1 "m 283 312 35:0 07 38:1 49:0 21 15:1 со 4:1 16:2 12:2 19:0 16:1 40:0 45:1 R 81 30:2 а, 16:0 = 12:3 pos 20:1 : 29:1 85:1 CJ 36:1 57:1 CH 41:1 Ed Legend 23:1 on ши no homoplasy 46:2 52:1 55:2 58:1 >< reversals f the two most-parsimonious trees from the successive weighting analysis of the Quiinaceae with Figure 3. One o characters optimized on the branches (number left of colon = character: right = character state, cf. Table 1). groups (char. 32: 1). a berry-like fruit (char. 52: 1), and pubescent seeds (char. 58: 1). Touroulia forms a monophyletic group and is recognized by only one synapomorphy, the compound leaves with an alate species), glabrous seeds, and hermaphroditic flow- 2) As previously mentioned, ge- ers are important characters to distinguish the ge- nus. This example addresses the importance of in- cluding autapomorphies in phylogenetic. analyses. following Yeates (1992). If only one Froesia species rachis (char. 5: nerie rank was proposed for T. amazonica (Pires, in sched.). a view also advocated by Gottwald and Parameswaran (1967) based on their anatomical had been included, a number of characters would studies. Nevertheless, from a morphological point have been regarded as autapomorphies. uninfor- of view, the two Touroulia species are very similar mative, excluded, and hence probably overlooked as potential synapomorphic characters for the ge- nus. \ Apart from Froesia, the remaining species of (Zizka & Schneider, 1999), confirmed by the pre- Quiinaceae are distinguished by several synapo- sent analysis, and, hence, Pires’s proposal is re- morphies, including the evolution of staminate flow- jected. A number of characters unite the genera Quiina ers (char. 28: 2). their position in small terminal 74 Annals of the Missouri Botanical Garden and Lacunaria. Polymorphic characters such as the position of the inflorescence (char. 29) in Lacunaria are interpreted as inconclusive of relationship (soft 1999). Regarding Lacunaria, conflicting data fail to conclusively re- starts sensu Kornet and Turner, solve the taxa in the analysis using equal weights, but support for Quiina is strong. However, a phy- logenetic signal for Lacunaria is picked up if the approach to character weighting is used, gaining a fully resolved tree, but with very low support val- ues. A jackknife value of 61% and a rescaled Bre- mer support value of 0.9 cannot be considered as much support for a monophyletic genus. Based on the weighted analysis, a possible single synapo- — morphy is identified for Lacunaria (char. 28: 3). Most other characters are homoplastic within the genus, and some are also observed in, for example, Quiina pteridophylla. Our knowledge of interrela- tionships within Lacunaria is not satisfactory, but pending future analysis we refrain here from pro- posing any systematic rearrangement and treat the genus in its traditional circumscription, 3: 2) and completely fused stipules (char. 22: 2), have Two characters, verticillate leaves (char. evolved in the clades of Lacunaria and Quiina. Based on our morphological studies, Оила pteri- dophylla was expected to be closely related to Lac- unaria. The present analysis indicates that Q. pter- idophylla shares almost all other generic characters and that it is deeply nested within Quiina rather than having a basal position in the genus. Forcing a sister relationship of Q. pteridophylla with L. op- positifolia, which it superficially resembles, costs no less than 10 extra steps. Thus, there is no doubt that these two species are not closely related and cannot form a sister relationship. Reasons for previous hypotheses of recognizing Lacunaria oppositifolia at generic rank can here be understood. Following the weighted analysis and D tracing. character. evolution. (Fig. 3), the species forms a clearly distinct branch with no less than five homoplastic characters. The remaining lineage of Lacunaria has, interpreted from the tree in Fig- ure 3, gained the same number of homoplastic characters, although they are different. Although a there 1 Л monophyletic genus сап be envisioned, also weak jackknife support for L. oppositifolia forming a separate lineage nested between Lacu- naria s. str. and Quiina if equal weights are used. One option is to include £. oppositifolia in Quiina, but this solution is not attractive because a mor- phologically well circumscribed group would be lost. Leaf venation pattern, the absence of interse- condary veins, and the number of carpels do not coincide with the generic concept of Quiina. More- over, cladistic support of the genus Quiina is almost four times as strong as the next internal node. A second solution is to describe a monotypic taxon, but due to the weak support, and not appearing in the consensus, the issue of the rank and position of L. oppositifolia has to be postponed until other data, particularly from DNA analyses, are obtained. Taxa sampled of Quiina for this analysis include about one third of the total number of the species. As a genus, Quiina is rather well supported, but support clearly increases after successive weight- ing. A cursory inspection of Figure 3 also reveals a number of reversals and parallelisms on branches leading to the terminals, especially in the most ad- vanced part of the tree. We interpret this, in con- junction with a high ratio of polymorphic charac- ters, to indicate that novelties are allowed to rise and spread within the lineage. They represent soft reversals without hard starts of character states (Kornet & Turner, 1999). Two reasons, and possibly more, explain this observation. First, the lineage might be fairly young and presently going through numerous speciation events without character state fixation. Second, interpretations of our character states may be incorrect, i.e., misinterpretation of homologies. EVOLUTION OF ANDRODIOECY Different modes of sex distribution in the Quiin- aceae—androdioecy in Quiina and Touroulia, di- oecy in Lacunaria—seem to have evolved from a hermaphroditic ancestor. Hermaphroditism is the plesiomorphic state, found in the basal genus Froe- sia and in the outgroup Ochnaceae. Andromonoecy is an autapomorphy for the outgroup taxon Medus- agyne and is not further considered herein. Andro- 28: 2) is : Quiinaceae that evolved, dioecy (char. synapomorphy within m-— "* - — phylogenetic per- spective, in the following pathway: androdioecy is the plesiomorphic state and dioecy independently evolved in Lacunaria. Concerning the evolution of androdioecy and di- oecy, several models have been formulated (Ross, 1978, 1982; 1981; Charles- worth, 1984; 1990; Richards, 1997; Swensen et al., 1998). The majority regards dioecy as derived from hermaphroditism via andro- Thomson & Barrett, Thomson & Brunet, dioecy or gynodioecy (Darwin, 1877; Westergaard, 1958; Вама. 1980: Richards, 1997). Another model 1982) is from hermaphroditic flowers via andromonoecy, of androdioecy evolution proposed by Ross ( as observed in some Solanum species. However, andromonoecy is only known from the outgroup. Medusagyne. In contrast, Swensen et al. (1998) ar- Volume 89, Number 1 2002 Schneider et al. 75 Quiinaceae gued that [functional] androdioecy evolved from di- оесу in Datiscaceae. This path is improbable for Quiinaceae since androdioecy evolved once in the ancestor of Touroulia, Lacunaria, and Quiina, with a further reduction to dioecious flowers in Lacu- пага Mechanisms that favor the evolution of andro- dioecy and dioeey are explained by models of sex- ual selection and sex allocation, and are discussed by several authors (Ross, 1978, 1982: Bawa, 1980: Barrett, 1981; Charlesworth, 1984: 1990; Richards, 1997; Pannell, 1997). One hypothesis argues that in female-sterile Thomson & Thomson & Brunet, plants the loss of seed production is more than compensated for by the reallocation of resources to increased pollen production (Thomson & Brunet. 1990). This may be achieved through an increased number of staminate flowers or stamens, or through larger anthers. Flower numbers are more or less equal among the Quiinaceae or even lower in the dioecious taxa. — (andro- Anthers are roughly of the same size in the family, hence not serving as a use- ful argument. The number of stamens, indeed, generally increased in staminate flowers of Quiina and Touroulia. Nevertheless, comparing the genera, the stamen numbers do not provide an explanation for the evolution of the observed pattern of sex dis- tribution since the supposedly ancestral hermaph- roditic flowers of Froesia bear the highest number of stamens within the family. If we trace an evolu- tionary pathway from hermaphroditism through an- drodioecy to dioecy, as postulated above, we see a reduction of stamen numbers from more than. 100 in Froesia, ЗО to 80 in Touroulia, most Lacunaria, and some of the Quiina species, to approximately 10 to 25 in Quiina and in a few Lacunaria species. including £L. oppositifolia. Thus, other mechanisms such as the avoidance of inbreeding may provide a better explanation for the evolution of dioecy in the present case. Literature. Cited Г. Fukuhara < K. Kikuzawa. variatii chizopepon bryoniaefolius (Cucurbitaceae). Amer. J. Bot. 86: — ). Amaral, M. C. К. 1991. Phylogenetise he — matik der Ochnaceae. Bot. Ji ene Syst. 113: 105-190. APG. 1998. An ordinal classification for flowering plants. Ann. Missouri Bot. 253. Baum, H. 1951. Die Frucht von Ochna — ein Fall ökologischer Apokat „= К. 5. 1980. Evolution of diseno d in c ants. . Rev. Ecol. Syst. Blas tharyya, B. & B. Akimoto, J.. and ¿ genetic 1999. Sex ratios on in a functionally androdioecious — s of Gard. тре. Oesterr. Bot. Z. 383-39. 15-39, M. Johri. 1998. Flowering Plants: Taxonomy and Phylogeny. р r- Verlag. Naruso Pub- 9 lishing Hones New Bremer, К. 1988 in angiospe rm phyloge netic reconstruction. Evolution 12: 795-803. 12 Se a h support and tree stability. Cladis- lies 10: Bryant, Н. С po Why ol uelis — 2t re- moved: A reply “ ha o Cladistics 11: 381— Carpenter. J. M. . The limits of amino acid sequence data Ж ES: among "ийн ET peu С Ang € adis stic s4: — ———,. 1994. Successive weighing re UE pa evi- den Сыйга 10: 215—220. Androdioces y and the evolution of . ya к 23: 333—348. Fay Chase „М ау & У. Savolainen. 2000. Higher- level classification. in ka angiosperms: New insights from the perspective of DNA sequence data. Taxon 19: Cronquist, А. 1981. An Integrated System of Classification J £ of Flowering z P ш Columbia Univ. Press. . Different E Fonos of Flowers on P n ШЕ vl the Same Spe cies. Murray, Londor К. 1982. -» arpy and ils alternative modes or Darwin. C. Endress. P. of escaping disadvantages of apocarpy in primitive an- Taxon : 8—52. 1998. AutoDecay, version 4.0%. Stockholm. "n. Farris, J. S. 1969. A successive ии approac h to character weighting. Syst. Zool. 4-385. . 1989. The re tention” index бе m rescaled con- sistency ‘ude x. Cladistics 5: 417—419 1996. Names and origins. Cladiaiños 12: 263— 2 —, V. A. Albert, M. Kallersjo, D. Lipscomb & А. Kluge. 1000. Parsimol tgs — outperforms ne "i —124. bor- joining: Cladistics 11: \ Fav. M. ‚ Swensen ^ 3 W. Chase. 1997. Taxo- nomic aliinitie s of d uis oppositifolia (Me dusa- evnaceae). Kew Bı S : 111-120. Felsenstein, J. 1985. E e limits on phylogenies: An approach using the bootstrap. Evolution 39: 783 91. Fite h, W. M. 1971. lution: rue — 106-416. Towards defining the course of evo- ora seoil tree topology. ysl. Zool. ALS. T n in 1 Quiina ae n Ducke. — and venation г tha leaf Amer. J. Bot. 37: 159- Foster, 1950b. Touroulia guianensis Aubl. \ Bot. 37 Ve — and histology of the leaflets in and Froesia tricarpa Pires. mel Яир M. > — 1951. hes Heterophyly and foliar venation in Lacu- — Bot. Club 78: 382—400. gmap der Pflanzen. Part 3, 3rd naria. Goe be ed. — pri Gottwald, H. & N. Par atomie und Systematik der dent d eae. Bot. Jahrb. Svst. 81. H. G. « K. Bremer. 1995. Киро ову апа ıylogenelic interre els of the Asteraceae, C: lyceraceae, Campanulaceae, Goodeniaceae, andı re late rales). Amer. J. Bot. 82: 250-205. Mickey, L. 1973. Classification. of pr architecture. of to ns leaves. Amer. J. Bot. 60: 17—33. A . A revised classification P E architecture of dos EMI leaves. Pp. 25-39 in C. R. Metcalfe erlag. Jena. rameswaran. . Beitrüge zur An- ) ) O TNR M. — — (Aste 76 Annals of the Missouri Botanical Garden & L. Chalk (editors), Clarendon Press, Oxford. Jansen, S., P. Baas & E. Smets. occurrence and systematic importance in eudicots. Tax- Anatomy of Dicotyledons ‚ Vol. 1 on 50: 135-167. Kiillersjé, Farris, A. G. Kluge & C. Bull. 1992. 15-2 Skewness and pe е Cladistics 8: 27! ——, M. Chase, B. Bremer, M. oF Humphries, G. Petersen, O. Seberg & K. Bremer. 1998, Simultaneous parsimony jackknife analysis of 2538 rbcL DN! reve M support for major clades of green gem land plants, seed plants and flowering plants. PI. — 213: 259-287. — D. Turner. 1999, Coding ee — — reconstruction. Syst. ‘Biol. 48: 365 sequences 79, Litt, А. & М. W. Chase. 1998. The systematic position of Fuphronia, with comments on the position of e An m based on rbcL sequence data. Syst. Bo 01—40 Maddison, y P & D. К. ا‎ Analysis of Phylogeny Chara 3.0. Sinauer, Sunderland, Massac Nandi, O. L, M. W. Chase 8 P. K Endres. 1998. A combined cladistic analysis of angiosperms using rbcL and non-molecular data sets. Ann. Missouri Bot. 85: 137-212 Nixon, K. C. & J. M. ا‎ 1993. On outgroups. Cla- distics 9: 413-420 & Davis. laxa, ‚ missing values and e lacis 7: 233-241. Pannell, J. R. . The maintenance of la and a in a me tapopulation. E volution 51: 10-20. Pires, J. M. 1948. Quiinaceae. Pp. 22-32 in G. A. Black & J. M. Pires (editors), Deia géneros novos Curupira e Froesia, cinco espécies novas, uma nova combinag йо, 1992. MacClade: r E SIMI. vers. Gard. 1991, фо * analysis. Cladist с — e observações sobre ги is . Bol. Técn. ed região amazón- Inst. Agron. N. 15: 1-32 950. Contribuia io Ares a Quiinaceae, Bol. Téen. Ins 19 isis Amazónica. . Agron, N. 20: 41—52 2. lanis novas i Amazónia. Quiinaceae. Bol. Teen, * Agron. М. 28. 45—17. . 1960. Plantas novas da Amazonia, Quiinaceae. Bol. Técn. Inst. — N. 38: 23— Raunkiaer, C. 1 The Life Forms of Plants and Statis- tical Plant Анан C isl ndon Press, Oxford. Richards, A. J. 1997. Plar маны Site ms, Chapman and S London Ross, M. D. 19 The evolution of gynodioecy and sub- а Pus 32: 174-188. 2nd ed. 2001. Vestured pits: Their . 1982. Five evolutionary иши to subdioecy. Amer. ote 119: 297-318 Roth, I. 1996. Microscopic venation patterns of leaves and ч ir ните іп the distinction of (tropical) species. T -19 — al. (editors), Encyc — {ан Anatomy Vol. 4. Gebr. Borntraeger, Ber lan . E Fay, г a Albach, A. Backlund, м. van der Ban > M. Cameron, S. A. Johnson, M. Lledó, J.-C. Pintaud, M. Powell, M. _ Sheahan, D. E. Soltis, P. S. e Р. Weston, W Whitten, K. J. Wurdack & M. W. Chase. 2000. ias ny of the eu- dicots: A dic — le on — based on rbcL gene se Ve * es. Kew Bull. 257-309 Se acta ar J. V. 3. El género Quiina (Quiinaceae) con especial referenc ia a las especies de Venezuela. Acta Bot. 1— Schofield, — — А 1968, Petiole anatomy of the — and re — yd Mem. New York. Bot. Gard. —55. Soltis, D. E., P. S. Soltis, M. W. Chase, M. E. Mort, D. C. Albach, M. Zanis, V. а ОЕ n, "X H. Hahn, S. E Swensen, L. M. Prince, . Farris. 2000. Angio- — phylogeny infe med from 128 aper rbcL, and tpB se a nces. js ot. J. . Soc. 133: 461. Swe nsen, 5. M., a «T bis & uas 8 rg. 1998. site Monophyly and e se ey nce of breeding s vem evolution, Syst. Bot. 23: 157-169, Swofford, D. L. . PAUP*, Fine p Us- ing тее — other methods). У Sunderland, Massac jor s. Thomson, J. & 5 H. Barrett. 1981. Selection for oute rossing, se xual di and i evolution of oecy in plants. Amer. Naturalist 118: 443—449. —— & J. Brunet. 1990. Hy; f dioecy in — plants. TREE Kress, tig М.К. Кау, М. de 1, S W. J 1. Nix Datiscaceae rev . 4. Sinauer, ` di- vhs for the evolution : 11-10. Tie gu чп, P ~ 102. Sur les Oc ino ées. Ann. Sci. Nat. Bot., de l — Weste — 195 Тһе mechanism of sex determi- nation in s ‘ious owe ‘ring plants. Advances Genet. 9: 217-28 Wiens, J. J. 1995 5. Polymorphic characters in phylogenetic systematics. Syst. Biol. 44: 482-500. W — M. 1902. Ordered versus unordered charac- ane stics 8; 375-385. ‚А comparison of two methods of character — ов. Cladistics 11: 297-308. Yeates, D. 1992. : 387-389, Zizka, G. & J. V. Schneider. 1999, 5 Aubl. (Quiinaceae). Willdenowia 29: Why remove autapomorphies? Cladistics genus Touroulia A REVIEW OF THE GENERA Warren D. Hauk’ ROENTGENIA AND POTAMOGANOS (BIGNONIACEAE)!'? ABSTRACT Roentgenia and Potamoganos (Bignoniaceae) are small genera of lianas with “Cydista-type” corollas that are closely related to larger and n ore well known genera such as C — ‚ Clytostoma, and perhaps Phryganocydia. Roentgenta is ditypic, containing R. bracteomana and " sonda whereas Por tamoganos is monotypic, containing only P. microcalyx. The two species of Roe nigenia have white to lavender, campanulate-funnelform, Cydista-type e stems with eight phloem arms, bi-trifid tendrils, 3(—4)- ме. ollen with reticulate exine, 2-seriate ovule organization, nectariferous disk lacking, linear-oblong fruit, and winged : um Roentgenia a is distributed from southern Colombia lo northern Bolivia and north-central B scil. whereas A. sordida is restricted to eastern. Venezuela, Guyana, Suriname, French Guyana, and portions of northern Brazil. Potamoganos нир, ias lavender to magenta, testis funnelform, Cydista-type corollas, with four phloem arms, trifid tendrils, 3-colpate pollen with retic wikis xine, d-seriate ovule organization, and a nectariferous disk. Potamoganos microcalyx is known from southern Venezuela, Guyana, Suriname, and northernmost Brazil. A key to species, maps of species distributions, reproductive phenology. and illustrations of R. bracteomana and p microcalyx are provided. Key words: Bignoniaceae, Bignonieae, Cydista, Potamoganos, Roentgenia. Roentgenia (K. Schum. ex Sprague) Urb. and Po- distinguished by a combination of characters: stems tamoganos Sandw. (Bignoniaceae) belong to tribe with four phloem arms, trifid tendrils, 3-colpate Bignonieae, which is composed almost exclusively pollen with reticulate exine, 4-seriate ovule orga- of lianas and which has been demonstrated recently nization, and a nectariferous disk (Gentry, 1977, to be monophyletic using rbcL and ndhF gene se- 1978: Gentry & Tomb, 1979; Tomb & Gentry, un- quences (Spangler & Olmstead, 1999). Roentgenia published; Hauk, 1997). Potamoganos has a “Cy- is a ditypic genus of primarily lowland lianas with dista-type” corolla (Gentry, 1978) and is closely re- white to lavender, campanulate-funnelform, “Cydis- lated to several larger genera (Table 1) such as ta-type” corollas (Fig. 1), stems with eight phloem Cydista, Clytostoma, Roentgenta. p — y- arms, bi-trifid tendrils, 3(—4)-colpate pollen with dia (Gentry & Tomb, 1979; Hauk, reticulate exine, 2-seriate ovule organization, nec- The late Alwyn Н. Gentry — a tariferous disk lacking, linear-oblong fruit, and and/or Potamoganos in several regional floras winged seeds (Gentry, 1977, 1978, 1997; Gentry & (Gentry. 1977, 1978, 1983, 1993, 1997. in press). Tomb, 1979; Tomb & Gentry, unpublished). Roent- but his monographie work did not encompass вета is closely related to such genera (Table 1) as Roentgenia, Potamoganos, or other genera of the Cydista Miers (6 species), Clytostoma Miers ex Bu- tribe Bignonieae. This paper utilizes the vast reau (12 species), Potamoganos (l species), and amount of information chronicled by Gentry in his Phryganocydia Mart. ex Bureau (4 species) (Gentry, studies of Bignoniaceae and attempts to compile 1977, 1997: Gentry & Tomb, 1979; Tomb & Gentry, these sources on Roentgenia and Potamoganos into unpublished: Hauk, 1997). a single treatment. The database established bs ~ Potamoganos is a monotypic genus of primarily entry's investigations has allowed production of E yple 8 | \ B lowland lianas with lavender to magenta, campan- detailed maps of geographic distribution (Figs. 3. ulate-funnelform corollas (Fig. 2). Among genera of — 4). Distribution data presented here provide an im- tribe Bignonieae with “Cydista-type” flowers, it is portant foundation for future investigations of the ! This paper is number 11 of the о INVITATIONAL SERIES, in acknowledgment of contributions to the study of the Bignoniaceae made by Alwyn Н. Gentry ? | thank Peter Raven and the Missouri Botanical "Cards n for the — to conduct this research. In particular, | am grateful to the late William G. D'Arcy, whose advice and guidance were an invaluable contribution to this work. Susan A. Moore provided the illustrations. Financial support was — by the National Science Foundation (grant DE В- 950927 О). ! Department of Biology, Denison University, Granville. Ohio 43023. U.S.A. ANN. MISSOURI Bor. GARD. 89: 77-87. 2002. Annals of the Missouri Botanical Garden exine Pollen 1997; Gentry 44. Pollen aperture Seed Fruit surface Ovule organization Corolla pubescence Nectar disk arms Phloem Morphological/anatomical characters and character states for Roentgenia and Potamoganos and three genera putatively closely related (after Gentry, 19 Tendrils & Tomb, 1979; Tomb & Gentry, unpublished). Table 1. taxonomy and systematics of Roentgenia and Po- ate ate ale ate ate tamoganos. reticu reticul: HISTORY AND SYSTEMATICS In 1916, Urban described Roentgenia and trans- ferred Cydista bracteomana К. Schum. ex Sprague into the genus as its sole species. Urban considered Roentgenia distinct from Cydista based on the trifid tendrils and plurisuleate pollen of the former (Sprague & Sandwith, 1932). Sprague and Sandwith transferred Arrabidaea sordida Bureau & K. inaperturate/pericolpate 3(—4)-colpate Inaperturate inaperturate Schum. to Roentgenia in 1932, after examining ad- G 3-colpate ditional collections more complete than the holo- type. Macbride (1961) questioned the use of pollen characters alone as generic markers, and noted that м. other characters, for example simple vs. trifid ten- bialate bialate corky drils, that separate Roentgenia from Cydista are not stable in Bignoniaceae. Gentry (1974a) acknowl- edged that the primary features distinguishing the two genera, tendril and pollen, might not be suffi- Y cient to retain Cydista and Roentgenia as distinct. Gentry (1978: 262) wrote that Roentgenia is “barely separable from Cydista,” although he treated them non-echinate non-echinate non-echinate echinate as distinct in his floristic works (Gentry, 1977, 1978, 1983, 1993, 1997). Gentry (1974a: 880) reported that the distinction between the trifid tendril of Roentgenia and the 2-seriate 2(—4)-seriate 2(4)-seriate }-seriate simple tendrils of Cydista “breaks down” in a col- lection of R. bracteomana (Seibert 2146; not ex- amined by this author), However, the putative pres- ence of anomalous tendrils in a single specimen does not provide a strong argument for treating Roentgenia and Cydista as congeneric. From a cla- distic perspective, the perisyncolpate pollen. linear linear-trisect. bracts and bracteoles, and trifid glabrous/lepidote lepidote lepidote lepidote tendrils of R. bracteomana and R. sordida provide synapomorphies that indicate they are sister spe- cies. Because there is no evidence, either morpho- logical or molecular, that provides a sound basis for absent absent present absent absent combining the two genera, this treatment. retains Roentgenia as distinct. from Cydista. Whether Roentgenia arose from within a paraphyletic Cydis- di со pu е ta (or conversely), or as sister to Cydista or another closely related genus, is not clear at this time. Only future, detailed morphological and/or molecular in- vestigations will be able to address rigorously the monophyly of Roentgenia and Cydista, and the re- bi-trifid simple simple trifid lationship of these two genera to the presumably closely related Clytostoma, Phryganocydia, and Po- — . F. W. Meyer described Bignonia microcalyx in 1818, and Sandwith (1937) transferred it to Po- tamoganos, noting its affinities to Cydista and lytostoma Roentgenia Cydista Phryganocydia Potamoganos 4 Roentgenia. However, Sandwith (1937) asserted C Volume 89, Number 1 002 auk 79 Roentgenia and Potamoganos that the presence of a nectariferous disk distin- guished the taxon from both Cydista and Roentgen- The cupular calyx and “Cydista-type” Potamoganos (Fig. 2A) led Gentry (1978) to con- clude that Potamoganos is allied to Cydista and Roentgenia. However, Potamoganos differs. from these genera in having a well-developed nectari- ferous disk (Fig. 2C). and lacking glandular fields 1978). other genera, Clytostoma, Roentgenia, and Phry- [presumably on the calyx] (Gentry, Three “Cydistatype” corolla but 1978). However. Phryganocydia has a long, spathaceous calyx, not ganocydia, possess the * ack a nectariferous disk (Gentry, the cupular calyx characteristic of Roentgenia and Cydista. The 3-colpate pollen. of Potamoganos differs from the inaperturate pollen of Clytostoma. Phry- ganocydia, and most species of Cydista (Gentry & 1979). with Cydista-type corollas that has aperturate pol- Tomb, Roentgenia is the only other genus len (Le., colpate), although perisyncolpate pollen is known in four of the six species of Cydista (C. de- a] cora, С. y heterophylla, and C. ae- 1979). Roentgenia pollen is verrucate to scabrate. and un- diversifolia, quinoctialis; Gentry & Tomb, like the medium-reticulate surface found in Pota- moganos pollen (Gentry & Tomb, 1979). Phrygan- ocydia has finely scabrate pollen that differs from the medium-reticulate type of Potamoganos, Cly- tostoma, — most species of Cydista (Gentry & Tomb, The — numbers of Aoentgenia and Po- Goldblatt and Gentry (1979) reported chromosome numbers of 2n 40 in 21 of the 23 genera of tribe Bignonieae including Cydista and Clytostoma. tamoganos are not known. However. they surveyed, and it is likely that Roentgenia and Potamoganos have a similar chromosome number. DISTRIBUTION Collections of R. bracteomana extend from south- ern Colombia to northern Bolivia, with some exten- sion into north-central and western Brazil (Fig. 3). Collections of R. Guyana, sordida occur from eastern. Ven- ezuela, Suriname, French Guiana. and northern Brazil. with a single collection from Ama- zonas. Brazil (Fig. 3) Collections of Potamoganos microcalyx occur from south-central Venezuela, Guyana, Suriname. and the northernmost portions of Brazil (Fig. 4). western Suriname, Р, microcalyx grows in lowland forests (0-60 m) and is present in many moist for- ests and riverine habitats (pers. obs.). flower of The surface of PHENOLOGY Roentgenia and Potamoganos are closely related to a group of genera (Cydista, Clytostoma. and Phryganocydia) that Gentry and Tomb (1979) sug- gested constitute a natural group because they share similar pollen morphology, a tendency toward “multiple bang" flowering phenology, and the ab- sence of a nectariferous disk. "Multiple bang" spe- cies have numerous, synchronized, short flowering days) that may occur at any time of 1974b). The absence of a nectar- iferous disk (and presumably nectar), coupled with periods (ca. 3 the year (Gentry, conspicuous visual and olfactory attractants. indi- cates that pollinator deceit may be the ultimate pol- lination strategy (Gentry, 1974b). The short, re- peated floral bursts may serve lo lure novice pollinators that effect pollination through visits to only a few flowers, after which they seek a more ample nectar source (Gentry, 1974b). Visits by po- tential pollinators are infrequent. presumably be- cause pollinators learn quickly that the flowers offer no nectar reward (Gentry, 1974). Fertile collections of Roentgenia bracteomana and R. sordida are known from throughout the year and are consistent with a "multiple bang" pollina- tion syndrome. However, little is known about the pollinators of these two species, and more detailed examinations of flower production are necessary to document reproductive mode. The reproductive biology of Potamoganos is not documented, but the close relationship of Pota- moganos to “multiple bang” genera suggests a pre- disposition to a "multiple bang" strategy. However. the nectariferous disk in Potamoganos microcalyx (Fig. 2С) could signal a departure from the “mul- tiple bang” strategy. The flowering specimens ex- amined for this treatment document that PF micro- calyx flowers in February, April, May, and October. Thus. the phenology data available are consistent with a “multiple bang” syndrome for P. microcalyx. MATERIALS AND METHODS Gentry compiled a private database of label in- formation from herbarium specimens he collected and from specimens at other herbaria that he ex- amined personally. Gentry’s database has been in- corporated into the Missouri Botanical Garden da- TROPICOS, also contains label information for all Roentgenia tabase-management system, which and Potamoganos specimens housed at MO. All types were assumed to have been seen by Gentry unless otherwise noted. Gentry did not always des- ы 99 99 66+ ee ignate types as holotype, isotype, or “syntype, and the designations presented here are based upon Annals of the Missouri Botanical Garden inferences drawn from Gentry’s work and the orig- inal literature; these type designations were not based on personal verification of specimens at the various herbaria. An index to numbered exsiccatae is provided in Appendix 1. Data used for mapping and phenology were down- loaded from TROPICOS, . For records with no latitude/ longitude coordinates in TROPICOS, approximate co- ordinates were obtained from gazetteers produced by the U.S. Board on Geographic Names, Office of Ge- ography, Dept. of the Interior. Distribution — were produced using the computer program VE MAP 1.51 (C. H. Culberson, Newark, DE, оо TAXONOMIC TREATMENT Ber. Deutsch. Bot. Ges. 34: 747 Roentgenia bracteomana (K. Roentgenia Urb., 916. TYPE: Schum. ex Sprague) Urb. Lianas; stems terete to subtetragonal with 8 phlo- em arms in cross section; branchlets terete to sub- tetragonal with interpetiolar glandular fields lack- ing and transverse interpetiolar ridges present or absent, the younger branchlets striate, glabrate to lepidote; pseudostipules caducous to somewhat persistent, foliaceous, chordate-orbicular, glandu- lar, glabrate to lepidote. Leaves opposite, estipulate, petiolate, bifoliolate with the terminal segment of- ten modified into a trifid tendril; petioles and pet- iolules terete to slightly flattened, glabrate to pu- berulent or sparsely lepidote; ultimate segments entire, chartaceous, marginally plane; venation bro- chidodromous; midrib prominent with secondary veins pinnate; axils of secondary veins lacking glandular fields. Inflorescences elongate, axillary or terminal racemes, branched or unbranched, sever- al- to many-flowered; rachis and peduncles terete to subtetragonal, puberulent to lepidote, and brac- teate, the bracts linear to linear-trisect, each seg- ment linear, persistent to caducous, glandular or eglandular, puberulent to caducous; pedicels terete, puberulent to lepidote; bracteoles linear to linear- trisect, persistent to caducous, glandular or eglan- dular, puberulent to lepidote. Flowers ovoid in bud: ~ calyx campanulate-funnelform, glandular or eglandular, costate or ecostate, glabrate to lepidote or puberulent, the calyx margin usually split, api- cally truncate except for five minute teeth; corolla zygomorphic, tubular-funnelform, white to magenta with purple markings and the tube yellowish, gla- brate to lepidote externally, pilose pubescence at the base of the filaments; corolla lobes 5 (2 upper and 3 lower), short-orbicular, the inner surface gla- brate to lepidote, the outer surface lepidote: sta- mens 5, fertile stamens didynamous with the fifth stamen modified into a staminode, all adnate to the corolla; fertile anthers with two spreading thecae, included, glabrate; disk absent; ovary cylindrical, lepidote, the ovules biseriate in each locule; stigma included, bipartite, the divisions laterally flattened. Fruit a septicidal capsule dehiscing parallel to the septum, linear-oblong, the valves compressed and not conspicuously thickened, the margins not se- rially constricted, drying brown, the midline not ev- ident, the surface rough, plane and glabrescent or sparsely lepidote to puberulous; many-seeded; seeds flattened, oblong, bialate, the body distinct and bipartite. KEY 1. Bracts 10-15 mm long; bracteoles 5-10 mm long, extending beyond the base of the calyx R. bracteomana I’. Bracts 1-5 mm long; bracteoles 1-3 mm long, not extending to the base of the calyx R. sordida TO FLOWERING SPECIMENS 1. Roentgenia bracteomana (K. Schum. ex Sprague) Urb., Ber. Deutsch. Bot. Ges. 34: 747. 1916. Cydista bracteomana K. Schum. ex Sprague, Verh. Bot. Vereins Prov. Brandenburg 1908: 121. 1909. TYPE: Brazil. Amazonas: Victoria, Rio Jura, May 1901, Ule 5497 (ho- lotype. B not seen; isotypes, HB not seen, 4 not seen, MG not seen). Figure 1. Lianas; branchlets drying light brown to gray, te- rete to subtetragonal, with interpetiolar transverse ridge evident but inconspicuous, the lenticels not evident or inconspicuous; pseudostipules caducous, 8-11 X 15-18 mm. Leaves 10-37 ст long; petioles 0.5-6.0 cm, terete to slightly flattened, glabrate to sparsely lepidote; petiolules 1.0—5.0, terete, glabra- te to lepidote; ultimate segments (7)10.5-28 X 4.5 X 14 cm elliptic to ovate-elliptic, densely lepidote when young, in older leaves glabrescent to sparsely and inconspicuously lepidote, especially along the basal portions of the veins beneath, apically acute, basally acute to obtuse, with 5-6 principal vein pairs. Inflorescences axillary racemes to 30 em long, unbranched, several- to many-flowered; rachis and peduncles terete to subtetragonal, lepidote, and bracteate, the bracts trisect, each segment linear, 10—15 X pedicels 5— 2—4 mm, persistent, glandular, lepidote; 7 mm long, lepidote; bracteoles 5-10 X 1-3 mm, persistent, glandular or eglandular, lepi- dote, extending beyond the base of the calyx. Flow- ers: calyx 5-10 X 4—7 mm, glandular or eglandular, costate, the outer calyx surface sparsely lepidote, the inner surface glabrate, the margin intact or split 1/4—1/2 the length of the tube, the valves apically Volume 89, Number 1 2002 Hauk 81 Roentgenia and Potamoganos Figur ~ after Plowman «€ Schunke V. 75( acute to truncate except for 5 minute teeth: corolla exserted 30-35 mm beyond the calyx lip (35—40 mm total length), 2-3 mm wide at the calyx lip; corolla lobes 15 X 10 mm; fertile stamens 9 or 10 mm long. inserted ca. 10 mm beyond the corolla tube base. the staminode 2 mm long, inserted 2 mm proximal to the stamens; ovary 2-3 mm long: style ca. ЗО mm long. Capsule 27-60 X 2.0-2.5 mm, I. Roentgenia bracteomana.—A. Leaves with inflorescence, after R. Burnham 1563.—B. Fruit with seed. )3.—€C. Seed, after Plowman & Schunke V. 7503. glabrescent; seeds flattened, 2.0-2.3 X 5.5-0.2 cm. 1-2 mm thick, oblong, bialate, the body ovoid and bipartite. Figure: Gentry (1993, fig. 71—06). Distribution, elevation, and habitat. Known from southern Colombia to northern Bolivia with a few collections from western and north-central Bra- zil (Fig. 3). Collections are reported from 150 to Annals of the Missouri Botanical Garden 870 m, in premontane wet and tropical wet forests, on terra firme forests or in riverine habitats. Phenology. | Flowering collections of Roentgenia bracteomana are from: January (10), February (4), March (4), April (1), May (3), July (1), August (3), September (1), October (7), November (1), and De- cember (1). Fruiting collections are from: February (1), March (1), April (1), May (3), June (4), July (3), September (2), and October (3). Representative үөү ОМВ1А. Caqueta: 41 km N of Florencia, 1060 m, e 9169 (MO). Putu- mayo: Selva higrofila del Río Putumayo en las márgenes del afluente izquierda La Concepción, 225 m, Cuatrecasas 10834 (COL). ECUADOR. Napo: Armenia Vieja at Río Napo, 12 km SW of Coca, Lugo 2703 (GB, MO). Morona- Santiago: Taisha, Río Guaguaine, 500 m, Cazalet & Pen- nington 7558 (K, IS). Pastaza: Río Bobonaza, Quil- loallpa below Montalvo, 02°10'S, 76?53'W, 300 m. Øllgaard et al. 34586 (AAU, MO). Santiago-Zam« banks of Río — mi, 500 m, Cazalet & PDA NY, US). Sucumbios: without exact locality, 240 m, 00°08'S, 76°22'W, Carlos E. Cerón & Judith Ayala 9469 фе ГУ A a 'hipe: without exact loc — ty, 930 m, 04°18'S W, W. Pe l. Varg Freire 8717 (МО). И ч ›пав: ONF : Woytkowski 5637 (MO). Huánuco: Codo de Pozuzo, al- luvial fan floodplain of Río Pozuzo, 10°15'S, 74°55’ W, 300 m, Foster 938 0). Loreto: Alto Amazonas, Río Pas- taza, Lago Rimachi, 04°20'S, 76°35'W, 200 m, Díaz S. & Ruiz 924 (AAU, MO). Madre de Dios: 39 km SW of Pto. Maldonado, Laguna Cocacocha, d 69*20' W, Smith & Shuhler 400 (MO). Puno: „пове Бе Насі LOS, ween ps ү andamo and Río Guacamayo, 13°30'S, 69°50’ Е 400—6 1, Gen- try et al. 76941 (MO). San Non Fundo Pts Pa Mariscal Caceres, Tocache Nuevo, 450 m, Even 11788 (F). Ucayali: vicinity of LSU base c amp. x — Shesha, ca. 65 km NE of Puc PE 08°02'S, 73°55'W, 25 m, Gentry & Diaz 58505 (MO). BOLIVIA. Beni: Km | carretera е ae a Colegio Técnico Agro- pecuario Río Colorado, 14°50’'S, 67°05'W, 230 m, Smith et al. 14115 (MO). La Paz: Abel Iturralde Province, Alto Madidi across from mouth of Río Enlatagua, 13?35'S, 68°46'W, 280 m, Gentry & Estensoro 70344 (MO). BRA- ZIL. Amazonas: 54. Maranhao: Rio Xingu, 04°49'S, 52°31 W, Balée 1958 ( NY). Flowering material of Roentgenia bracteomana is distinctive because of the prominent bracts and bracteoles; no species of Cydista or Clytostoma have similar-sized bracts or bracteoles. The lack of conspicuous bracts and bracteoles in R. sordida, coupled with a largely allopatric distribution, makes confusion of the two species unlikely. Veg- etatively, trifid tendrils distinguish Roentgenia from Cydista and Clytostoma, but sterile material lacking tendrils may be easily mistaken for species of these closely related genera. 2. Roentgenia sordida (Bureau & K. Se pes Sprague & Sandw., Bull. Misc. Inform. Kew 1932: 91-92 1932. — sordida Bu- reau & K. Schum., in Mar s. 8 pt. 2, fasc. 118: 30-31. 1896. TY PE: Guyana. Upper Rupunini River, Schomburgk 1296 (holotype, B not seen, presumed destroyed). Arrabidaea — Sprague, Bull. Herb. — 28 5): 373-374. TYPE: Suriname. Sipaliwini: — Saramacca River, Pulle 170, 495 liane U not seen). Lianas; branchlets drying tan to gray, terete, gla- brate to puberulent or lepidote, with transverse in- terpetiolar ridge evident but inconspicuous, the lenticels not evident; pseudostipules + persistent, 5-8 X 6-7 mm. Leaves 11-27 cm long; petioles 1.2-5.0 cm, terete, puberulent; petiolules 0.6-3.0 cm, terete, puberulent; ultimate segments 8-22 X 3-15 cm, elliptic to ovate, glabrate to lep- idote (especially on young material) or puberulent, especially on basal portions of the main veins, api- foliaceous, cally acute to obtuse, basally obtuse to rounded or inequilateral, with 5 to 6 principal vein pairs. /n- florescences axillary or terminal racemes to 20 cm long, unbranched, many-flowered; rachis and pe- duncles terete to subtetragonal, puberulent to lep- idote, bracteate, bracts 1—5 X 0.5-1 mm, cadu- cous, eglandular, puberulent to lepidote; pedicels 3-6 mm long, puberulent to lepidote; bracteoles linear (to linear-trisect), 1-3 X 0.5-1.0 mm, ca- ducous, eglandular, puberulent to lepidote, not ex- tending to the base of the calyx. Flowers: calyx 4— 7 X 3-6 mm, glandular, ecostate, the outer calyx surface puberulent to lepidote, the inner surface glabrate, the margin intact or split 1/8-1/2 the length of the calyx, the valves apically obtuse with 5 minute teeth; corolla exserted (25)32—42 mm be- yond the calyx lip ((29)36—49 mm total length), 1.5-3 mm wide at the calyx lip; corolla lobes 12 X 12 mm: fertile stamens 8 or 16 mm long, inserted 9 mm beyond the corolla tube base, the staminode 2 mm long, inserted 1-2 mm proximal to the in- sertion of the fertile stamens; ovary 4 mm long; style ca. 25 mm long. Capsule 27-34 X 2 mm, sparsely lepidote to puberulous; seeds not ob- served. Figures: Gentry (1983, fig. 39; 1997, fig. Distribution, elevation, and habitat. Known from eastern Venezuela, Guyana, Suriname, French Guiana, and northern Brazil, with a few, scattered collections from western Brazil (Fig. 3). Collections are reported from 350 to 2000 m, in non-inundated moist forests, swampy mature forests, or white sand areas. Volume 89, Number 1 2002 Hauk 83 Roentgenia and Potamoganos Phenology. Flowering collections of Roentgenta — 5 sordida. are from: January (5), March (1), April (1 May (3), July (1), September (1). October (6). No- vember (7). and December (4). specimen examined was collected in March. The sole fruiting Representative spec m BRAZIL. Amapa: Mpio. de 49'N. 51?23'W, S. Mori & К. Souza 17263 . Amazonas: São Pa ulo de Olivenga, basin creek of Sm Krukoff 8623 (BM, F, MO. NY, U). Maranhão: Monção. P.I. Guaja, Rio Turiaqú, 03°07'S, 4670' W, Balée 3517 (MO). Para: Campus of IPEAN, Belém, Gentry 13077 (MO). FRENCH GUIANA. Approuague, en amont de Crique — Oldeman 2303 (САҮ). Cayenne: Mon- tagne de Kaw, Piste Roura/Kaw, 04°33'N, 52°09" W, 350 m, PX 2926 (MO). GUYANA. East Berbice: margins of Berbice pes S of New — 06?00'N, 57°43'W Maas et al. 5573 (U). Mazaruni: Upper Mazaruni jd Mapa 170 m. Tillett & Tillett 45685 (MO). S . Nickerie: right bank of Corantijn River, N P Kaboe rie € E. ‚ Heyde 449 (MO). Para : Jodensavanne- Ma- J 'oene, 02? Ne ri-Weg and de Crane- Weg. W of Le pee in one sec, an on white sand, Heyde 596 (MO). VENEZUELA. Bolivar: Rio Paragua, hasta 12 vueltas т la boca del Rio Топого, 06%03'N O3°4AT'W, 17 — 10320 (MO). Delta Amacuro: E * Rio сш El Palmar, Gentry & Berry 14938 О). Mérida: Marcano Berti, Luis 199 (MO, VEN). When intact tendrils are not present on a spec- imen, the short bracts and bracteoles of Roentgenia sordida make it particularly difficult to distinguish from the sympatric Cydista aequinoctialis (L.) Miers, although the latter can usually be distin- guished by dark glandular fields (when dried) in the axils of the secondary veins beneath; these elandular fields were not observed in Roentgenta. Gentry (1978: 27 paniculate inflorescence and overall leaf appear- 19) suggested that the racemose- ance of Cydista lilacina A. H. Gentry “links” the genera Cydista, Roentgenia, and Clytostoma. How- ever, C. lilacina is distinct from other Cydista spe- cies in its curved bud apices and 4-seriate ovules (Hauk, 1997), and these two features serve to dis- lilacina from К. persistent pseudostipules and echinate y tinguish C. liad-like,” fruit of Clytostoma are distinct from the foliaceous, sordida. The “brome- caducous pseudostipules and rough (but plane) fruit of R. Mey.) Sandw. is partially sympatric with R. sordida sordida. Potamoganos microcalyx (С. and has trifid tendrils, but P. microcalyx differs from R. sordida in having a nectariferous disk. —— Sandw., Recueil Trav. Bot. Néerl. 220-221. 1937. TYPE: Potamoganos mi- Mene (G. Mey.) Sandw. Lianas; stems woody with 4 phloem arms in cross section, branchlets terete to subtetragonal with in- terpetiolar glandular fields present and distal to the node, the transverse interpetiolar ridges not evi- dent, the younger branchlets striate and glabrate to lepidote; pseudostipules caducous, subfoliaceous. triangular, eglandular, lepidote. Leaves opposite, es- tipulate, petiolate, 2- or 3-foliolate with the termi- nal leaflet often modified into a trifid tendril; peti- oles terete, glabrate; petiolules terete to sulcate, elabrate to lepidote; leaflets entire, marginally plane; axils of secondary veins lacking glandular fields. Inflorescences elongate, axillary racemes, un- branched, several-flowered; rachis and peduncles flattened to nearly terete, glabrate to lepidote, brac- teate, the bracts minute, caducous, lacking circular glandular fields, glandular-lepidote; pedicels flat- tened, glabrate to lepidote; bracteoles minute, ca- glandular-lepidote. Flowers ducous, eglandular, ovoid bud; calyx cupular-campanulate, eglan- dular, lepidote, the calyx margin intact, the apex truncate except for five minute teeth: corolla zygo- morphic, campanulate-funnelform, lavender to ma- genta, glabrate to sparsely lepidote, with an inner ring of pubescence proximal to the insertion of the stamens; corolla lobes 5 (2 upper and 3 lower), short-orbicular, the inner surface glabrate, the outer surface glabrate to lepidote; stamens 5, fertile sta- mens didynamous with the fifth stamen modified into a staminode, all adnate to the corolla: fertile anthers with two spreading thecae, included, gla- brous: nectariferous disk present; ovary cylindrical, densely lepidote, the ovules 4-seriate in each loc- ule; stigma included, bipartite, the divisions later- ally flattened. Fruit not seen. 1. Potamoganos microcalyx (G. Mey.) Sandw.. Néerl. 34: 220-221. 1937. Bignonia microcalyx G. Mey., Prim. Fl. Esseq. 211. 1818. TYPE: type, GOET not seen), not seen by Gentry. Fig- Recueil Trav. Bot. Guyana, Meyer s.n. (holo- ure 2. Bignonia microc Ж var. acuminata Miq., in Flora 25(2): 427. 1942. TYPE: Suriname, (holotype?. U not seen). — — — Bureau & K. Schum., in Mart.. Fl. Bras. 8 pt. 2, fasc. 121: 146. 1897. TYPE: Su- riname. prn exact locality, Wullschlaegel 1034 (BR not seen). Lianas; branchlets drying brown to red-brown, striate, glabrate, lenticels not evident: pseudosti- pules triangular, 1 X 1 mm, eglandular, lepidote. Leaves 15-25 cm long, once-pinnate, bifoliolate with a terminal leaflet often modified into a trifid tendril: petioles 3.5-11.0 ст long. glabrate to puberulent or lepidote; petiolules 0.5-3.5 em long. glabrate to lep- 84 Annals of the Missouri Botanical Garden gure 2. a microc alyx.— P r Miller & Hauk 9418.— Tendril. after Miller & Pn 205 idote; leaflets 6.5-13 X 4.0-8.5 cm, elliptic to sub- orbicular, plane, chartaceous, glabrate with a few mi- nute glands scattered over the surface, apically acute to obtuse, basally broadly acute to rounded, with 4 to 5 principal vein pairs, marginally plane. /nflorescen- ces to 20 em long; bracts linear, 1.0 .5 mm; bracteoles linear, 0.5-1.0 X 0.5 mm; flowers ovoid in bud; calyx 3-5 X 5-6 mm, glabrate to sparsely lepidote; corolla exserted 35-55 mm beyond the . Inflorescence and leaves, after Miller & Hauk 9418.—B. Young — Jetail — calyx, ovary, and nectariferous disk, after Miller & Hauk 9418.— level of the calyx lip (40-60 mm total length), 2.5— 3.0 mm wide at the calyx lip, the corolla tube lep- idote inwardly and glabrate outwardly; corolla lobes 18 X 15 mm; fertile stamens 12 or 18 mm long, inserted ca. 5 mm beyond the corolla tube base, the staminode З mm long, inserted З mm beyond the corolla tube base; disk 1 mm tall; ovary 2-3 mm long; style 15-20 mm long. Fruit not seen. Figures: Gentry (1983, fig. 37; 1997, fig. 407). Volume 89, Number 1 002 Hauk Roentgenia and Potamoganos eJ 10N 0 > 10S 20S 30S 90W 80W ow 40W Figure 3. g Geographic South America. Distribution, elevation, Known from south-central Venezuela, and portions of northern Brazil (Fig. 4). Collections are reported from O to 300 m in mixed lowland and habitat. Guyana, Suriname, forest, often on lateritic soils. Four flowering collections were ex- amined: February (1), April (1), May (1), and Oc- No fruiting solite ‘lions were seen or have Р) е H henology. — tober (1). been documented. — = ns. BRAZIL. Amapa: without ex- act locality, 02703'N, 50°48'W, Ducke 1973 (MG zonas: Boa Vista Road, 64 km N 60?08' W, Lowe 4190 (MO). GUYANA. Vicinity of ys 1 of Mazaruni and Cuyuni Rivers, 00730' М, Station, junction : Rupununi: Feat: of TM 50 40'W, Graham 307 (СМ). Kanuku Mts., Moco-Moco village, 03° ‚ 59°35 W, 100 m, Maas К Westra 3906 (MO). йай ы Malali, Demerara River, 05°50'N, 58?15'W, de la Cruz 2726 (CM, US); — Trail, 2: . Don Linden, 06?00'N, 5815'W, 20 m, Grewal 16 (U). SURINAME. Voltzberg: without exact locality, Pulle 307 (U). Saramacea: Cop- pename River, 3 km below base camp, 05730' М, 55°50 W. Florschutz & Maas 2734 (О). Sipaliwini: Mamadam, O04?40'N, 55?40'W, Florschutz & Florschutz 1155 (U); Montibus р fluv. Kabalebo & Coppename Sin- istera, Kabalebo River 1-5 km below airstrip, 03°50'N, 56°50'W, Florschutz & Maas 2340 (U); area of vidt bo — distribution of collections of Roentgenia bracteomana (circles) and R. sordida (triangles) in O4°10'N, 57210'W, Heyde & Lindeman 192 — proje üt MOX vicinity 2 Blanche Marie falls on the Nickerie Riv- ~ O4°45'N, 56°52'W, 60 m, Miller & Hauk 9418 (BBS, vi o Fluv. Sarama ca, — 04^20' N, l ои р 25'W, Baud & Huber 15362 (MO): Río Hennes be S San Fernando de — Cano Morocoto, 1 hour below San Fernando, 04°10'N, 67°42'W, Gentry i al. 10947 Rio Cuao, Río ае ‘0, — 617740'W, 125 m, guire & Politi 28436 (NY, US); Cerro Huachamacari. Río Cunucunuma, 03748' N, da °46'W. 400 m. Maguire et al. 29973 (VEN). The shape and color of the Cydista-type corollas of P. microcalyx are easily confused with those of the sympatric Cydista aequinoctialis (L.) Miers and C. lilacina A. H. the axes of P microcalyx are longer and more Gentry. However, inflores- cence flattened than either Cydista species, and the calyx is shorter and more cupular. The nectariferous disk and trifid tendrils of P. microcalyx should readily distinguish it from all Cydista Gentry (1997) noted that the leaves of Potamoganos mi- crocalyx are similar to those of Distictella Kuntze but that the inflorescences Mansoa A. DC. species. and Ceratophytum Pitt., and flowers more closely resemble Annals of the Missouri Botanical Garden 70W 60W 50W 40W Figure 4. Geographic distribution of collections of Potamoganos in northern South America. and Roentgenia. Like Potamoganos, Roentgenia ‚ In press. Bignoniaceae. In: z rra de Colombia. but the nectariferous disk is absent, and the bracts and bracteoles of P. micro- has a trifid tendril, calyx are not trisect, as is found in Roentgenia. Literature Cited Gentry, A. H. 1974a. Studies in Bignoniaceae 12: New — species of South — an Bons eae. — Bot. Gard. 61: 872—885. )74b. Flowering phe — and diversity in tropic a Bina eae. Е a 6(1): 6 ita — ———. 1977. Fam. 178. Bignoniaceae. гъ Harling B. — (editors), Flora of Ecuador. o ra Bot. 173. T . 1978. Bignoniaceae. /n: The Botany of the Guy- ana Highland: А Mem. New York Bot. Gard. 2 xu аний eae. In: Z. Luces de Febres & T A. Ste — (editors), Flora de Venezuela. 8(4): 7 433. Fundación iie Botánico de Venezuela, Ca- racas. —————. 1993. A Field Guide to the Families and Genera of Woody Plants of Northwest South America (Colombia, Ecuador, Peru) with Supplementary Notes on Herba- ceous Taxa. Conservation. International, „ишн D.C. [Reprinted by Univ. Chicago — )o.] : Bignoniaceae. Pp. 403—491 in: J. A. ermark, А E. Berry & B. K. Holst tt € Flora of the Venezuelan — Vol. 3. Missouri Bo- tanical Garden Press, St. Lou & A. S. Tomb. 1979 [1980]. Taxonomic implica- = ons of Bignoniac eae palynology. Ann. Missouri Bot. . 66: 750-777 Goldblatt P. & A. H. Gentry. i Cytology of Bignoni- . Not. 132: 47: 1997, A review » th. > genus — (Big: 4 Ann. Missouri Bot. Gard 84: J. F. 1961. A. eae. In: — i я ru. is. e Hist., Bot. 3-103. 1818. кле э (a Fl. Esseq 211. Sandwith, N. Y. 1932. Notes on — American Big- noniace a я ueil Trav. Bot. Néerl. 34: 220-223. ——— 1937. Notes оп yo val jua 'an Bignoniaceae. Reci isi] аң Bot. — 1: 220-223. Spangler, R. E. € R. a 1999, е analysis of Rims eae based on the cpDNA gene se- quences rbcL and ndhF. Ann. Missouri Bot. Card. 86: 33-46. Sprague, T. A. & N. Y. Sandwith. 1932. оог to the flora of — al America: IX. Kew Bull. 1932: 18— = © = = = = ^ e e = b (chan, 1. 1916. Uber Ranken n: — de Bignoni- aceen. Ber. Deutsch. Bot. Ges. 34: APPENDIX ] INDEX TO NUMBERED EXSICC, Collections are listed КИ ally by the princ ipal author, followed by collection number, and a boldface number (1—3) that indicates the collected. All specimens entered into TROPICOS were assumed to have spec cies Volume 89, Number 1 2002 Hauk Roentgenia and Potamoganos been examined by A. Н. Gentry. Specimens examined by the author were limited to duplic ates housed at MO, and are indicated by a 7". The coding of the species number : Roentgenia aracteamana = — l, Roe — - 2. еми microcalyx Archer, há A. 8005 2: Aronson, J. 1039! 1 ; Ayala, E et al. 3927 Balée, W. L. 1958 l, 1968 E 3517! 2: Beck. H. T. et al мо! 2 — R. 897! 2; Black, G. A. 47-819 2: aye. J. «€ Asanza. E. 30067! ura, Д Ј. T. zalet, Ww — 7558! 1; rá E. & Ayala, J. — l; Cremers, G. "Croat, T. B. ч 21259! 1: Cruz, J. S. de la 2726! 3; Petite J. 10834. 1 у, ~ Davidse, G. & O. Huber apne З; Díaz S.C. & J. Ruiz 924! 1; Díaz S., C. et al. 1297! 1; dos Santos, G. el Fi ‚1 ж 2 Ducke 1973 3. !1. Florschulz 1155 3: Foster, R. B ; Florse * Р. & Maas 2340, 5 billet, c ; Florschutz, „б 1: Foster, R. B. & В. “Ас д 12004! I: Froes. R. L. 31509 2. ntry, A. H. 13077! 2, 49133! 2: Gentry, A. H. & P. Berry 14938! 2, 14970! 2: Gentry H. & Díaz. C. 58505! 1; Gentry, A. H. & L. Emmons 39392! l: Gentry, A. H. € S. Estensoro 70344! 1: Ge . H. & R. Foster 70842! 1: Gentry, A. H. & N. MN 1. 57083! l; Gentry, A. H. € P. Núñez 59559 1, 69435! 1, 69559! 1: Gentry, A. H. & J. Revilla p l; Gentry, A. H. & K. Young 31820! 1: . H. et al. 9169! С 10917! . 220604! 1, — ‚ 25849! 1, 26787! 1, 27287! 1, re 1, 46076! 1, 68670! 1, 76904! 1, 76941! 1: Graham, E. ч 3: Grewal 16 3: Grández. C. & А. Chigi puma * Heyde, M. — n s ashikat, V. 215€ раш, J. A k “Coello 1433! 1. „Р, — 2: Killip, E. P. & A. C. Smith 28169 g, G. 2043! 2170! l; Kramer & Hekking 7450 2: Krukoff, B. A. A eds 8847 2; Kujikat 357 1. Lindeman, J. C. 4050 2, 5008 2, 5197 2: Linder, D. H. 80 2; Lowe, J. 410! 3; Londoño, C. et al. 124! 3; Lugo. H. 2557! 1, 2570! 1, 2703! 1. Maas, P. J. M. & L. Y. Th. Westra 3906! 3; Maas, P. J. Gentry, A )! "4 * 2: — Isen, ж N. M. & J. С. „ et al. 21432 1: Hu- et al. 5573 2: Maguire, B. & L. Politi 28436 3; Maguire, B. et al. 29973 3; Marcano-Berti, L. 199! 2; McDaniel, S. € B. Marcos 11000! I: Mexía, Y. 6320! 1: Miller, J. S. & W. D. Hauk 9418! 3; Mori, S. et al. 17251! 2: Mori, S. & R. ups 17263! 2. . et al. 6283! 1; Núñez. P. 6149! S Timaná 12149A! Oldeman. R. A. 2303 2: Øllgaard, B. et a ‚ W. et al. 8717! 1; Phillips, O. et al. Pires, J. M. & N. T. Silva 11042 2, 2: Plowman, T. & т chun ke 7503! 1: Prance. G. T. et i Е ; Pulle, A. 170 2, 307 3, s 3. 11852! 1, 345% 6! 1. 329! 1; ‚М. 577 2; Schulz, J. 8371 2; Se Lus Е : 5696! 1, 11788 E; Seibert, R. , > MG8824 2; Smith, А. С. 2847! А с ‚ 14115! 1: Smith. S. F. & prin et al. 620! 1, 1454! 1; Steege, Н. e ! 3; Stergios, B. 10320! 2; Stergios, B. & L. nir e 12998! 2. Tessmann, G. 3245 1; Tillett, 5. S. & Tillett 45085! 2: 1 M. & P. и 1 426! 1; Timaná, M. & O. Phillips 881! 1. "s E. — l. | wed й Young. K. 25 I: 331! p H. J. & D. A. Stratton 3! 1. APPENDIN 2 INDEX TO SCIENTIFIC NAMES {nemopaegma cupulatum 83 Arrabidaea yullet 82 sordida 78. 82 Bignonia mi rocalyx 78, 83 microcalyx var. acuminata 83 Ceratophytum 85 Clytostoma 77, 78, 79, 82, 83 Cydista 77, 78, 79, 82, 85 aequinoctialis 79, 83, 85 bracteomana 78, 80 ecora 79 diversifolia 79 heterophylla 79 lilacina 83. 85 Distictella 85 Mansoa 85 Phryganocydia 77, 78, 79 Potamoganos 77, 78, 79, 83, 86 nicrocal yx 77, 79, 83, 84, 85, 86 Roentgenia 77, 78, 79, 80, 82, 83, 86 bracteomana 77, 78, 79, 80, 81, 82, 85 sordida 77, 78, 79, 80, 82, 83, 85 MOLECULAR DATA INDICATE Marcus Koch? and Ihsan A. Al-Shehbaz* COMPLEX INTRA- AND INTERCONTINENTAL DIFFERENTIATION OF AMERICAN DRABA (BRASSICACEAE)! ABSTRACT The genus Draba — about 350 species — ү in the Northern Hemisphere, with some 80 species described morphologic rally, the existing sectional iong species of in South and Central America. Although spec ies of Dr are well classification is — controversial. American taxa e hib it enormous morphological differences even an d the hypothesis that variation accum ul; ated during migration and i ien of American "the ITS (internal transcribed spacer) regions of the 72 nien; 'an gue taxa and 6 Mi primarily to pru n TI the same section. We teste Draba. The present “eme netic study i nuclear ribosomal DNA and we chloroplast trn L-intron and spacer sequences froi European Draba species. Resul some intrageneric groupings corre and that only to a small degree do — findings agree with previous sec — classific ‘ation. Differences betw ‘S- ed between some of the groups or sections ITS sequence zu lastid DNA y the American Draba ^ 2 — is based on analyses ts suggest that s and trnL-derived phylogenies suggest extensive genetic contact may have exis and that this disjunction = tween К uropean and American — is de ш ы һу sequences suggest | that “European” plastome types be more widely distributed among species, perhaps et — transmissions of Eurasian n chloroplast types d American Draba. Additional sys- tematic analysis demonstrates that the genus Erophila has to be integrated into Analysis on the tribal level — the entire Draba complex to be close to European Arabis and Aubrieta. The data provide additional support for evious assumptions that the existing tribal classifications of the Brassicaceae are mostly artificial and that the seg- re sharin of Draba and Arabis into separate tribes or subtribes does not accurately reflect their phylogenetic relationship. Draba. Key words: Brassicaceae, Draba, Erophila, molecular systematics, reticulate evolution. The wide occurrence of polyploidy in vascular plants reflects its importance to plant speciation. The origin of polyploids and the mechanisms be- hind the establishment of newly evolved popula- tions and taxa are among the challenging questions & Schemske, 1998; 1999; Sol- 1999), New combinations of favorable in. plant sciences (Ramsey Thompson € Lumaret, 1992; Petit et al., tis & Soltis, genes may be more easily stabilized in polyploid taxa, and the permanent coexistence of favorable traits and characters from different parental lines may be effectively preserved as fixed heterozygosity (Soltis & Soltis, 1993). These speciation processes are integral to the genesis and maintenance of plant biodiversity. Recent studies of the overall genome structure of hybrid and polyploid taxa provide new insights about the dynamic nature of complete ge- nomes either analyzed on the basis of artificial hy- brids or by comparative mapping (Rieseberg et al.. Linder, 1999; 1994). In their comparative 1999; Rieseberg & 1998; Kowalski et al., genome analysis of some Brassicaceae, Acarkan et al. (2000) showed that structural rearrangements occurred with a significantly higher frequeney in Lagercrantz, polyploid Brassica L. than in diploid Arabidopsis thaliana (L.) Heynh. or Capsella rubella Reuter. This suggests that at the structural level a more dynamic nature of the genome might be sustained in polyploid taxa. than diploid ones, and this might be a remarkable source for new genetic var- iation upon which selective pressures can then act. Draba 1.. and includes ca. 350 species, or about 10% of the family total (Al-Shehbaz, 1987). tributed primarily in the Northern Hemisphere, es- is the largest genus of the Brassicaceae The genus is dis- pecially in the subarctic to arctic regions and al- mountainous portions of the temperate Nearly half of the pine or regions. taxa are found in the ! We are grateful to je Mitchell-Olds, Max-Planck-Insitute for Chemical Ecology, Jena, the curators of GH, . Harriman, and Victoria Hollowell for critically reading the manuscript and offering the facilities to conduct th arch, and we thank thanks go to Donovan Baile ry, Nei cil. numerous suggestions for its improvemer ? Institute of Botany, ' Missouri Botanical Garden, Р.О. Box 299, St. Louis. ANN. Missourt Bor, Germany, for providing О, and US for the loan of specimens. Special University for hen ultural Sciences Vienna, Gregor-Me "s Str. 33, A-1180 Vienna, Austria. Missouri 63166-0299, A. GARD. 89: 88-109. 2002. Volume 89, Number 1 2002 Koch & Al-Shehbaz 89 American Draba New World and about 80 species are distributed in Central and South America (Al-Shehbaz, 1987). Schulz (1927, 1936) divided the genus into 17 sec- tions (Table 1), but Fernald (1934) criticized this artificial sectional classification largely on its mis- leading and impractical keys. Further, Tolmachev (1939) totally ignored Schulz's sections and recog- nized 29 series for the 91 species occurring in the former Soviet Union. No subsequent botanists have attempted to subdivide Draba into infrageneric taxa. For European Draba, most authors have fol- lowed Walters (1964) who correctly used section Draba instead of Schulz’s section Leucodraba DC... because section Draba includes the generic type. A few case studies have investigated the phylo- genetic relationships among some Arctic Scandi- navian (Brochmann et al., 1992a-d, 1993) and al- pine European (Widmer & Baltisberger, 1999a, b) species of Draba. Some general aspects of these studies have been previously outlined (Brochmann, 1992) and can be summarized as: (1) allopolyploidy is common, with a polyphyletic and polytopic origin for some taxa; (2) gene flow across different ploidy levels is possible and probably occurs in natural populations; (3) complex evolutionary phylogenetic networks have been demonstrated for section Draba and some species of section Chrysodraba:; (4) re- peated migration and colonization contribute to — л — complex distribution patterns; and not unex- pectedly, items 1 through 4 are substantiated by molecular studies with complex patterns of intra- and interspecific variation observed (Brochmann et al., 1992a—-d: Widmer € Baltisberger, 1990a, 1999b). Support also exists at morphological, cy- tological, and ecological levels (Brochmann, 1993: Brochmann et al., 19 Isozyme elec се and DNA analysis of the nuclear and plastid genome have greatly increased the possibility of detecting and distinguishing allo- and autopolyploids, to trace paternal and maternal genome lineages, and to document hybridization, introgression, and reticulate evolution within poly- ploid complexes. Several polyploid complexes in the Brassicaceae have been characterized, includ- ing those of the genera Microthlaspi К. К. (Koch et al.. 1998b; Koch € Hurka, 1999). Draba (Brochmann & — 1992; Brochmann, 1993: Brochmann et al., 2a—d, 1993), Cochlearia L. (Koch et al., 1998a, — Yinshania Ү. С. Y. Z. Zhao (Koch & Al-Shehbaz, 2000), Cardamine L. (Franzke et al., 1998; Urbanska et al., 1997), Biscutella L. (König, 1998), Brassica L. and related genera (Anderson & Warwick, 1999), and Nastur- tium R. Br. (Bleeker et al., 2000). In order to gain a better understanding of the [= phylogenetic relationships within a large polyploid complex, we examined the sequence variation from the internal transcribed spacer regions (ITS] and ITS2) of nrDNA for American species of Draba 1995: Campbell et al., 1995), the 1991: Gielly & Ta- erlet, 1994; van Ham et al., 1994; Koch & Al- Shehbaz, 2000), and the cp trnL-trnF spacer (Ta- berlet et al., 1991). The derived phylogenies were then compared with traditional (Baldwin et al., p trnL intron (Taberlet et al., molecular concepts based on morphological data. MATERIALS AND METHODS PLANT PCR- AMPLIFICATION, MATERIAL, DNA EXTRACTION, AND SEQUENCING Leaf material for DNA extraction was obtained from herbarium specimens (Table 1). 1 was iso- lated from 25—50 mg of dried leaf material using the NucleonPhytoPure Kit (Amersham Lifescience) following the instructions of the supplier. DNA was stored in 10 mM Tris-EDTA buffer pH 8.0 at —20°С. Double-stranded DNA of the complete ITS re- gion, including the 5.8 S rDNA region, was ampli- fied. by PCR using ITS primers initially designed by White et al. (1990) and modified by Mummenhoff et al. (1997a) for ITS4. The PCR profile used to amplify the ITS re- gion followed the following profile: a hot start with min. at 94°C, and 35 cycles of amplification (1 min. 94°C, + 38°C, 45 sec. 72°C), final elon- gation step for 10 min. at 2 C, and — at 4°C. The 18F primer (5'-GGAAGGAGAAGTCGTAA- CAAGG-3') is located at m 3'-end of the 18 S rDNA gene and primer 25R (5'-TCCTCCGCTTAT- TGATATGC-3’) is located at the 5'-end of the 25 S rDNA. PCR products were purified. using the Boehringer PCR product purification. kit (Roche Molecular Biochemicals). PCR products spanned the entire ITSI, 5.8 S rDNA, and ITS2 region and vere cycle-sequenced directly without cloning us- 35 cycles of symmetric л 5 sec. <= ing the Taq DyeDeoxy Terminator Сус ‘le Sequenc- ing Kit (ABI Applied Biosystems, Inc.) and the 18F and 25R primers. Products of the cycle sequencing reactions were run on an ABI 377XL automated sequencer. The trnL. (UAA) intron was amplified and se- quenced by using the universal primers “e” B49317, 5'-CGAAATCGGTAGACGCTACG-3') located at the 3'-end of the trnL(U A A)5'-exon and "d" (A49855, 5'-GGGGATAGAGGGACTTGAAC- 3’) located at the 5'-end of the trnl(UAA)3'-exon (Taberlet et al., 1991). The PCR profile used to amplify the trnL intron followed Koch and Al-Sheh- — Missouri Botanical Garden Annals of the 90 €lOLPIAV — 8069F 149 IOTOLLAV 041роит] єє (OIN) 9222 Diueapipy зу ouvsig *eunuazay J "чоон vsoynoiunf (q 9969141۷ — €€69tLIV 80S9PIAV ә}ХЎио8әршоюцул) 18 pal GOLF saut) *e[onzouaA Чәче Y uopurT pupryounf q PIOLPIAV — 21069PLIV — OOTtOTIIJV — 2]£2uoonum) c£ (SA) GOLF saui) *e[onzouaA "Чәче Xy Uepury puniyounf q SIOLLIAV — 9069tLIV — OSPOPLAV Dqp4po]p-) 1 (OW) £90] sonang- zado] Y UDLA[ZMY *ve[onzouoA "Ydue]g Y uapurT saprojasapf q 8969t[4V €S69PIAV 9OCOLIAV DQDIPO]A cg (SA) €6€8 punjdsy *10peno'q ppom Dsuojxo “(q 8€£40c LIV ZZLOZIAV Dqpapoono] VU q6661 "9219qsn[eg Y 1ouipra, *pue[iozits ләп DIQNP "(T OTOLLIAY — SO69FTAV 8SHObTAV Dqp4poj«] 0£ (OW) 69c] Paoyoarsoy 29 люиә]ц “sopensy J "чоон vssaıdap ‘q 8869t14V — £€69PLIV — S9tOor [AV DQUDApOs A7) c9 (OW) 82382 ито Y што "vs SEI) "Y 5 шор хә ‘HNN DIJOfiSUIP ‘q OLSOPLAV DqD1posA147) ZOI (1880) ZE8Z руту “Y SN nny DYofisuap ‘q 69869t 14V — c£69tIJV — TOTOTLAV Dunjsouio[ 19 (OW) 2729 suijoy 9 ио ұт ABI "Y Y шор хә ‘HNN юуоЙәипә ‘q 6cO0LTLIV — 7689 149 1894V I (SA) 8992 Ј9ә]7) *eiquio[o7) euejues Y [o2ue| DUDSDIALIDNI "(T LCV9T LAV Dqpapo]T 8с (OW) 52901 1m 12 упш; *n1ag J YOoH numpid&ao q 9cO0LVLIV = S689pPLIV — PPPOPIAV © (SA) II9FG О4әшоуу эр spsp2o4]0n7) *erquio[o7) əng ppnjdokao ‘q 0669VI4V — I£69PLIV — €8hOPTAV DQDApOSs A417) 09 (OW) 8£££ unusapus “VSN uos&eq DIMNA (T I669p 149. O£69PLIV — Z8bObIAV D]]aquaq(q 6€ (OW) 1881 7? 12 шә “VS шецеле) Dijofisspao (T IIO!tLIV — OL69FTAV — £OtOTF [AV Dqpapo]p:) ce (OW) 962€ әт “opens zin prjofindofuoo "qq PZOLLIAY — L699PLIV — LbrOTIJV — 2]uoZopum) 8 (OW) EZZ} Lag `т|әпләцәд PIE `4 `$ Pprydouorys ‘q LELOCIAV DQDIPOINY] SOL 46661 uodzioqsnpeg » J9UIptA "PUB LIZ IMG addoy DIDIYJULIDI `1 PLOOVLAV — LPOOPTAV ОО©ОР1АУ 6L (М) £0£ `P 12 qqnaz) “eIquiojor) zequaus-|V DUDIDAO (ү OPPOVLAV a]&z uod onum) L (OIN) 69POÍ[ sonang 4 - zado] y up4o[ -2MY “B[ONZIUIA OTP el © 117127724٩ `1 €ZOLVIAV — 8689PIAV 6bt9blAV II (HO) 86£2 «оттү y Among шо? ZEQYPYS-[V PUDAD]IIDg q COO9FIAV = 9S69FTAV — OOSOPTAV DGDAPONAY A 101 (QASO) 681€ PHNH "ven pjoonn “(J 69690t[4V — ZS69FIAV — SOSOPTAV DUNISOULO | r8 (SA) 6€6 4vao[ ‘nad "PPM s2proqpanp ‘q LIOLTIJ4V — TO6OTPIJV — OS FOFTAV £c (OW) 291 DaouoojsoN 9 1Du2]4ç лорепә МЯН sapiosstqp ‘q €£969t[IV — 8€69tIJV — ZISOPTAV sisdoziy FOI (661€ 188S0) ‘U's 22ffnos; *eujsny "| Seplozip ‘q TELOZIAV sisdoziy 601 96661 "129219qsnpeg Y 19urpra, “рившәдим "т Seploz ‘q DELOZIAV sisdoziy 801 46661 "9519qsn[eg Y 1oupra, “риерәдтм "| sapiozip (T LZLOZIAV ICLOCLAV sisdoziy 101 96661 u2x19qsn[eg Y зәшргдү “риерәлим "| ѕәрто210 `q T969tl1V — 2€69tIJV — IICOTLAV sisdoziy £01 (818£& ASO) "ws 4affnayy *Хиршләс) “| sepiozi ратл I969t [41V — 0969tL4V PICOLIAV LL (OW) 6262 Sump $ uno "yes sui[[oy (SUIT[OY) »20750214ppnb °”) c969tl41V = OS69FTAV — SICOFTAV L (OW) 8662 suvi]ng $ uno “Y SN SUI[OY (4&2) ү) nspjznop pjJarsoisn?) 19a0eds que поли! “que ¿SLI С 9199], dos ou “DY (шАполәр UN Lreq.2u) 10]99[[0» ‘nuno ПОХЕ |, pue “YN Qu (2Z61) 41992 Ses ‘ISLI 0} BUIPIOIOB UOTBOYISSE[O [BPUOTVIIE “HSH 3[qe1 әш uo p231eorpur SI иоцеив[ахә [euonippy "Apnis 19pun [еиәјеш əy} jo elep UOISSIIIY "| AFL Koch & Al-Shehbaz American Draba Volume 89, Number 1 2002 €169T LIV OT69T [LAN I0€9T LAV $81] tsoro 08 (3) [S69] UPN Binden рушла 'dsqns жәл], njpu1ajnd 1 €OTOT [AV DQDA4posA4u7) OL (ON) ZIES UNIDA y YIO: MH "VG UOSAR puiiadsoisajd q OCPOPLAY e (OW) ££49 *oiqifonf ap «png “Жыз zwquaus-[y sapioujupamauoopnosd (| 1809P LAV OFO69T LAV COT9T LAV punisouioj 69 (ON) 12126 ?pr] $ 19497) Y SN ABI) “Y Y шор рока “q €002T LAV 9169T LAV 69V9T LAV 09701рордоцу Vy (OW) 6% ££ uouo]og ‘nag ABI "y nzuodoqoid “(| 900LP LAV CLO9P LAV 89v9V LAV DQUJpO]V) ©? (OW) ОРРР 2?puqopngy niod ZINYS 74 70 (COQ) оиро (| LCOLTIAV TOGO LAV €tV9T LAV a|tAsuoziapuim) $ (OW) PEGE 1m 19 012104 *etquio[o?) доц 7 70) Muezpy-jjouued “(1 0869F LAV 169 LAV Tor9t LAV DQDADOS KAi[;) 12 (OW) 90669 1” 19 иом — apuqoe|y nuosivd (T c969t LAV 6€69T LAV LOPOPLAV DQDAPOSAIY’) 99 (OW) ££Z62 SUjOY Y suiJoy "Vs "NooH puuadsozijo ‘q c669T LAV 6c69T LAV L89 LIV. Dqu.pouapy 8S (OW) 22 70 19 007 45-9 ISO розни "(Jy €869P LAV 8€69T LAV DOVODIAV DQDIPO iyd L9 (М) 161 ЧОН “led z|nuos 74 70) 21107070 (T [L6O9TLAV 0€69T [AV EOSOPIAV Se ct Мей a ca (М) 161 ЧОН “led ZING ^q 'O %9ртоЈотупош q ( san] L00LT LIV p169 I4V -og ) nqpapoono] It (OW) 2127 Diuoanap) XY oupsig "eunuasay "шв" DAUD) IDU '(] 8002 LAV €I69TP IAV LOVOTLAV DQDIPOPQDYY Ot (OW) F290] 70 12 ш Mig ‘J оон nuda ona ‘q QOPOPLAV 6€ (OIN) £€9 P 19 ddna) "exquioor) aqu oun] t] 6002 T LJAV. cI69T 14V COPOVIAV siptsoio( 8€ (OW) 6ZOF Sog `®|әпләцәх Your] (00H) nupun (T SLOLPIAV €069F LAV SCTOT [AV [с (OIN) 6/ZE] Vowopog *eriT[og zequoys-|y pupizpdp] '( PpZLOZIAV DPVAGOINS ] 911 6661 OBIS у PUp "pue[rozis 19-79 pulpy] Y OV Z0c LAV €cZ0cL AV DQDLPOINA] SLI 1666 | PASNE Y WUP "ривроәдим ‘Td- 4g Pulpo] Y CoOo9T LAV 8c69T LAV O8POP LAV Dquiapouopy LS (OIN) 0c29c ?10]paoo4g `оәтхәр MAH sisuojnaof “q T669T [AV LcO9T LAV Dquipouopy 9c (OIN) 9c [c 4210251 7021xX9]A MAH sisuoppaof ‘q T9869F LAV LCo9Or LAV o8T9T LAV 99 (OIN) 6F62 197019 “Y 71 uosugof Y zun[y 1493901 q €869T LAV 9€69T LAV 88F9T LAV DQDAPOING] c9 (OWN) €&2€E8 о 12 su]oy VS uossegq pj422u1 '(] 9869F LAV €€69T LAV L8P9P LAV Т9 (OIN) £9c€8 717 19 sui]oy VS supy 122020274 "(T C669T LAV 9c69t LAV OLPOVLAV сс (OIN) А0692 MSNOpezy *021X9]N чоләрүегу szsuaz]ppni “(| OLO9TLAV LS69T LAV TO0SOT LAV. DqU4pO])) €8 (OW) Fc Чїтш$-толләрү $ ADSUDY “opens Z[NYOS “A 70) DuDAajsuaY (T 9669T LAV CCOOTIAV 8L2v9v IAV Dqp4po]]&ugd ye (OW) Слер Орус *оогхәр 9U991 DUDA PY `Q 8cOLTLAV £689V LAV СҮҮҮОТЯҮ с (S) 0206 /22]) “erquojor) PID Y oen) пиәшшту (f OLOZT LAV TL I69T LAV POPOVIAV Dqu1popp?) 9c (OW) 0££9 sojquing| 10penoj] 3 004 "gy “Y L869bL1WV T€69T IdV 98v9v LAV 0901роэтәт £9 (OW) 269 vopzpog “WSN Чета 0/9409 ‘A CIOLVIAV 6069T LAV COPOVIAV DQDAPO]D/) ve (OW) 878801 40140] Y 40]4n] “ao шү Y ооң «әт ү дәве ча uout que ES LI c 31481 IIS оп "09V (WAUOLOR UID LIEU) 40]29[[09 "Anuno?) UOXBL pue "YN QA (L261) Z[NYos ос ‘ISLI 0] BUIPIOIOR UOTJBOYISSe]O [ечоцоэ< "panurmuo?) `1 3IqeL Annals of the 92 Missouri Botanical Garden [€622€4V OLI 86/1122 Yosyaug *eujsny ләцо (T) Pua "3 ZS6LLEAV Ill 86/SS9% yosysug *euysny due] 4 "y vibpmyiods оруолу L6690Yl4V — tc6OtLld4V — LLTOTLAV cc (OW) 6FIF 9394 zeyoupg ‘nəq zequeaus-[y :oopana (T 1960t14V — TC69tl4V — Z20C9TPLAV Dqp4popp:) 99 (OZOPECHSA) "ws uossumf '10peno'] MAH 22207014 `q 8669tLIV EZ69Þ1AV 9LPOLIAV [€ (OW) #261 uosuozuof “eunuad1y z[nuos ^q 70) SSUIUDUNIN] q OZLOZIAV DQqnD4poorio'] ELI 46661 193195099 » J9UIptAA "pue[rozit ^o aey) DSOJUIWO] `1 [PLOZLAV CZLOZIAV Dqpapoone] ZU 46661 "9219qsnpeg P 1oupra ‘риерәдим Алгу DSOJUIUO] ‘q 9:69t14V — SPOOPIAV — SOFOFTAV GL (OW) ZIZP ssoy Y pjoyosoy “y Sn VU Y suroy »7DjJaqumqns q ZLO9VPIAY — 61694 149 COCOTLAV [9 (М) OS] uosauf ‘1openoy ‘ZOIN], DSOJÁIS "(T 97021 AV Dqpapoonog 901 96661 19319450 [89 » тэшр yooy SLIDJAIS `1 82.69tL1V — €P69FTAV 9619 14V Dqpapo]]&uq Ji (OIN) 99 4494 “W'S хеле) "V Ddap2oido4js (T LLOOVIAV — PROOPIAV — Z26T19V LAV bl (OW) 8911 120212ng| X uvuipoos) үсү uos&eq Y opuqoe|y 2£2]pupis (T SPPOrlAV pqvipouspy 6 (HO) 997€ aanidg “1openoy ppam pupaonads (T 6669t[4V ZZ69WIAY — €ZTOTLAV D(DIPO]D/) 0€ (OIN) 08001 2975/99 Y p103]10 чорепәҷ Зүг) ѕиәриә ‘q 0002tL4V — IZ69PIAY — FZ2TOPLAV pqvipopqnuy 6b (OW) 9ETET vowojog ‘eog ррәдү 814270105 "(T SZOLPIAV — Ө6Ө8Ө1ЇЯУ — SPPOPTAV D(DIPO]D/) 9 (SA) SEZZ 4onvqaoqag ‘nəd z[nuos ^q “O P1111108 ‘q LOOLPIAW — Oc69tLIV — £L TOP LAV Sb (OW) 192 /22]) ‘eıquo[0) josury x PuPjues Damas '( CO0LtI4V — 6169tLJV — cLTOFLAV pqvipopqouy 2% (OIN) 2206 uouojos ‘ern!{og `ррәдү umaopndoos (T OZOLFIAV — LO69PIAV ES POPTAV Dqp4popquiy cl (HD) ££09 4әпюфләфәдү “Ned z[nuos “4 O Maisnyos (T [ZOLPIAV — 0069 149 cCTOFLAV bl (H9) SOLE ¡22 uUDY “erquiojor) [93uey у PURJURS 401101 ‘q ZZOLPIAY 6689bTAWY — ICTOTLAV el (HO) V-£c£z &oiifonf 2 Хороод *eiquio[o-) zequaus-[y pupanoppa ү 6LO9FIAV — ZPOOPIAV — S6POPTAV pursouio[ 4j (OW) 86r£ Yooquazyanyy “WSN р[ечлә] Cure) sunidas q €269t[1V — 9rt69tL1V — O6FOFTAV Dqpapo]]^uqd 97 (OW) 8696F 4Y “VSN xnPAso(] DUISSISOUIDA (ү ЄОО/Р1ЛУ — 8169FIAV IZPOPIAV D(DIPO]D/) 9t (OW) 0168 1” 12 илә “1openoy ‘zounj, DA ydouodd `q POOLPIAV — ZI69PIAY OLPOPTAV ТЕТО] cp (OW) 9202 zmsumy “AO аца 4 орта q 6I0LÞIĄV ZO69PIAV — FCTOTLAV si&isojo( gI (OW) 0ЕОР әд `т|әпзәцәд —— zequaus-y aq 'dsqns "лолу руоштарта ‘q Jaoeds "quad поди ил] ZSLI с 9149], 92s “OU "23V (шАполәр WINLIEGIoy) 101931105 ^ципог) UOXP], рче ‘ұма (2261) 211925 єс "ISLI 0] BUIPIOIIR UOTJBIYISSP] 9) [EUon 'penunuo?) '[ [qe] Volume 89, Number 1 2002 Koch & Al-Shehbaz American Draba 93 baz (2000): a hot start with 5 min. at 94°C, and 35 cycles of amplification (1 min. 94°C, 45 sec. 50°C, 45 sec. 72°C), a final ср ов step for 10 пип. at 72°C, апа storage at ¿R products were purified and eyela-caquended as described for ITS analysis using the amplification primer c and d. The trnL(UAA)-trnF(GAA) intergenic spacer was amplified using the following primers: the universal primer “e” designed by Taberlet et al. (1991) (5'- GGTTCAAGTCCCTCTATCCC-3') and a newly de- signed primer trnF-IGS-rev (5'-AGGATTTTCAGT- CCTCTGCTC-3'). Amplification, purification, and sequencing were performed as described for the trnl, intron. DATA ANALYSIS Outgroup selection. Тһе genus Draba has been variously placed into the tribes Alysseae or Dra- beae (Hayek, 1911; Schulz, 1936; Janchen, 1942; Al-Shehbaz, 1987). However, as pointed out by Al- Shehbaz (1987), the Alysseae are a poorly defined tribe with about 40 genera (15 monotypic) and some 650 species distributed primarily in the Ira- no-Turanian regions. Recent molecular studies (Price et al., 1994; Koch et al., 1999a, 2000, 1999b) clearly indicate that nearly all tribal subdivisions of the Brassicaceae are high- and Mediterranean ly artificial. Therefore, we used the aligned ITS data set from Koch et al. (1999a) in which tribal struc- tures were analyzed, and taxa from the tribes Lep- idieae, Sisymbrieae, and Arabideae were consid- ered to infer relative phylogenetic position of taxa under investigation. We added sequence data from Erophila verna (L.) Chev., D. araboides Wedd., D. aizoides l., and Cusickiella — (Rollins) 110, spectively, Table 1) to the ae described by Koch et al. (1999a). We included C. because it was originally described in — and only recently it and D. douglassú А. transferred to the newly established — Rollins (Rollins, 1988). Data analysis, distance and parsimony methods, were performed as described in Koch et al. (19993) to get comparable results to this. analysis. Rollins (accession nos. ‚ 103, and 77, re- quadricostata were The results were used to choose an appropriate outgroup. Our analysis herein demon- strated that the European Arabis L. and Aubrieta deltoidea (L.) DC. iella is. Therefore, European Arabis (A. are closer to Draba than Cusick- alpina, A. bryoides, A. pumila, A. blepharophylla) and Aubrieta deltoidea were selected as outgroups for the further ITS study of Draba. Cusickiella was excluded from all further analysis. ITS data. DNA sequences were aligned. by hand. In addition to sequences from American Dra- за. we included 6 ITS sequences from European Draba from Widmer and Baltisberger (1999a. Table 1) and 6 sequences from Arabis/Aubrieta serving as outgroups (Koch et al., 1999a), resulting in a total matrix of 80 ITS sequences. Distance analyses were performed with Kimura-2-parameter distances us- ing PHYLIP software package version 3.57c (Fel- 1995). The neighbor-joining algorithm was used for tree construction. Bootstrap analysis (Felsenstein, 1985) was performed using 1000 rep- licates. For all analyses gaps were coded as missing senstein, characters. The parsimony analysis began with a subset of taxa comprising representatives from all putative clades included in the overall analysis. This analysis was performed to demonstrate relative branch length and to estimate additional confidence parameters. The second data matrix used was iden- lical alignment to the first data matrix. Parsi- mony analysis was performed with unordered Fitch parsimony using PAUP* 4.0b2 (Swofford, 1999). The branch-and-bound algorithm was used to find maximally parsimonious trees. Bootstrap analysis (Felsenstein, 1985) was performed using 1000 rep- licates with the bootstrap search algorithm. Decay analysis (Bremer, 1988) was performed in addition to the bootstrap approach, in order to assess the confidence that could be placed in the monophyly of elades. Decay indices (DI) were estimated ac- cording to Baum et al. (1994). For all parsimony analyses gaps were coded as missing characters. DNA sequences were aligned by hand using the boundaries of cod- trnL intron and spacer data. ing regions as anchor points with which to begin the We included 8 trnL spacer sequences from European Draba (Widmer & Baltisberger, 19991, Table 1) and 67 American Draba for a total of 75 trnL sequences. Draba cu- netfolia (acc. no. 61). D. platycarpa (асс. no. 69), D. reptans (ace 84). ingroup taxa (results from the ITS analysis), were alignments. intron and . no. 72), and D. araboides (acc. no. which are only distantly related to remaining used as outgroups. The ancestral position of these four taxa with respect to the remaining Draba an- alyzed herein was also confirmed by an analysis using tral, intron and trnl-trnF spacer regions from and Aubrieta deltoidea L. (GenBank AY034180 and AY034181, Distance analyses were performed for the Arabis alpina L. accession numbers re- spectively). total data set as deseribed for the ITS data. In order to obtain a relative support for branching patterns of representatives from all putative clades, a subset of Draba taxa was investigated with Fitch parsi- mony as described for ITS analysis. For this anal- ysis gaps were coded as additional binary (0/1) Annals of the Missouri Botanical Garden characters. All gaps differing in length and — were coded as separate binary characters. The dis- tribution of gaps and its coding can be — al http://homepage.boku.ac.at/koch/. Amplification failed to yield ITS 1) and D. jorullensis (acc. no. 56), and trnL intron and spacer 7), D. Missing data. data for Draba magellanica (acc. no. 4 data was missing for D. bellardii (acc. no. spruceana (acc. no. 9), D. pseudocheiranthoides (acc. no. 12), D. cryptantha (acc. no. 28), D. litamo (acc. no. 39), D. pterosperma (acc. no. 70), and D. densifolia (acc. no. 102). For Draba carinthiaca 105) and D. stylaris (acc. no. 106) only ті, intron and spacer data were available, and for (acc. no. the two Erophila species no plastidic data were available. Data matrices are available upon request or can be viewed at http://homepage.boku.ac.at/ koch/. RESULTS OUTGROUP SELECTION AND TRIBAL RELATIONSHIPS Our reexamination and analysis combining the ITS data matrix (a sampling of 36 ITS sequences from species across Arabideae, but excluding Dra- ba) from Koch et al. (1999a) with sequences added from Erophila verna, Draba araboides, D. aizoides, and Cusickiella quadricostata (in capital letters, Fig. 1) resulted in trees identical in topology to the phylogenetic hypothesis presented in that paper. The data matrix of 603 bp contained 342 invariable characters. From the remaining 261 variable nu- cleotide positions 163 sites were potentially infor- resulted in 12 most parsimonious trees with a length of 612 steps, mative. Fitch parsimony analysis a consistency index (CI) of 48.9% TO s excluded), and a retention index (RI) of 71.8%. The strict consensus tree (not shown) out of Hid 12 most parsimonious trees is identical (except for Er- ophila verna, Draba aizoides and D. araboides, and Cusickiella quadricostata that are added herein) to the previously published phylogenetic network (Koch et al., 1999a). to that presented by Koch et al. (19993) (Fig. 1). Draba is closely related to the European Arabis L. The distance tree is identical and Aubrieta Adans., and it is positioned within an (Fig. 1). to the western United "Arabideae core group" The genus Cus- ickiella, States, molobus Tausch and North American Arabis. As shown by Koch et al. (1999a), the North American blepharophylla Hook. & Arn. group and a few other species, are better treat- which is endemic was resolved with North American Hali- Arabis, excluding the A. ed in the genus Boechera A. Live & D. Live (Live & Live, 1975). For subsequent ITS analysis of all Draba taxa, European Arabis and Aubrieta were used as outgroups, and Cusickiella has been ex- cluded from all subsequent analysis. ITS DATA The analysis of phylogenetic relationships among ingroup taxa comprises 80 sequences from Draba and Erophila of which 74 sequences are novel. Six ITS sequences from Arabis and Aubrieta were used to root phylogenetic trees as outlined above. The total length of the aligned data matrix is 626 bp. From 283 bp within the ITS] region, 109 sites were variable (including 37 autapomorph- ic changes); from 166 bp within the 5.8 S rDNA, 11 sites were variable (all autapomorphic changes): and from 177 bp within the ITS2 region, 20 sites were variable (all autapomorphic changes). This re- sults in a total of 68 potentially informative sites. Uncorrected pairwise sequence distances ranged up to 10% within outgroup taxa (59 nucleotide po- sitions), up to 9.39€ within ingroup taxa (57 nucle- otide positions), and up to 12.3% among the whole data set. (acc. no. 33) was excluded from these because of an unusual ITS tvpe that differed mark- edly from the remaining Draba taxa (13.1 to 16.6% Argentinian Draba funiculosa Hook. f. calculations sequence divergence or 82 to 104 nucleotide po- sitions). However, in the phylogenetic analysis D. funiculosa clustered among European Draba (Fig. 2). The whole alignment required 44 gap positions of 1 nucleotide in length, and 2 gap positions re- quired 4 (Arabis alpina, A. blepharophylla) and 14 (A. alpina L., Aubrieta deltoidea) base pairs, re- spectively. Identical ITS sequences were detected in D. lindenii and D. pulvinata subsp. pulvinata (acc. nos. 38 and 80, respectively) and in D. sericea (acc. no. 48), D. splendens (acc. no. 50), D. hem- sleyana (acc. no. 83), and D. extensa (acc. no. 85). The results of the distance analysis are shown in Figure 2. Six major clades (I-VI as indicated on Fig. 2) could be distinguished among the genus Draba with relatively robust support for clades I- Ш (Fig. 2) We also indicated Schulz’s (1927) sectional clas- sification (Table 2) in Roman numerals in the phy- logenetic ITS tree (Fig. 2) Fitch parsimony analysis with a subset of Draba taxa from all major groups recognized in the overall analysis resulted in 12 most ай وی‎ trees (MPTs) 208 steps in length, CI of 70% (autapo- morphies lt ed RI of 69.99. of the bootstrap and decay analysis are added to The results the strict consensus tree shown in Figure 3. The overall analysis (Fig. 2), as well as the anal- Volume 89, Number 1 2002 Koch & Al-Shehbaz 95 American Draba Arabis pauciflora EUROPE Arabis turrita EUROPE pe; scabra DRABA ARABOIDES USA EROPHILA VERNA EUROPE DRABA AIZOIDES EUROPE Aubrieta deltoides EUROPE rabis j va ed — a EU Arabis — USA 100 Arabis glabra U Arabis lyalli USA Arabis lignifera USA Arabis divaricarpa USA П 4 0.1 substitutions/site Figure |. Neighbor-joining distance tree based on ITS of Draba to other cruciferous taxa. added to a previous analysis of tribe Sequences. from Arabideae (Koch « "Arabideae core group? previously recogn (L.) Bernh. herein we used the taxonomically more appropriate synonym Turritis glabra L. below the branches from 1000 replicates. ysis of a subset of taxa (Fig. 3), recognized the same six major ITS clades among European and Ameri- can Draba, including the two accessions of Ero- European. Draba are confined to clade HI 2 and 3): this clade is characterized by long internal genetic phila. together with European. Erophila (Figs. 2 distances separating European taxa from the re- maining American Draba. The American species of Draba are found in all remaining clades I. H, IV. V, and \ The most basal clade, clade I, combines se- Оа! аса vulgar IS Cardamine flexuosa EUROPE Р CUSIC Halimolobus perplexa var. lemhiens alimolobus perplexa var. ae pa four taxa analyzed 1 ized by Koch et al (1999a). Distributional area is indicated. Arabideae core group EUROPE Capsella rubella EUROPE | Arabidopsis griffithiana ASIA L Arabidopsi abi — himalaica abidopsis viret ASIA Arabis glabra — KIELLA ат Е сар USA Arabis parishii USA 95- Arabis drummondii USA sequence data to demonstrate the systematic relationships ı this study (shown in capital letters) were 19994). Draba and Erophila were integrated into an For Arabis glabra Lal. Bootstrap support is given quences from mostly North American Draba taxa D. cuneifolia, ace. no. 61; D. platycarpa, acc. no. — no. 72), but also one accession no. 84). Clades I-III are well supported by bootstrap values of 100%. 88%, and 93%, respectively. In contrast, clades IV. V, and VI (combining the majority of taxa) are only 69: D. reptans, acc. from Peru (D. araboides, acc. weakly supported by bootstrap analysis (Fig. 2). However, all clades were recognized in the parsi- mony analysis with a subset of taxa from all six clades. The only exception is D. scopulorum (acc. Annals of the Missouri Botanical Garden 67 Draba violacea (VIII) 86 Draba stylosa (?) 81 Draba Sedi n (VIII) 35 Draba depressa (V) 30 Draba pycnophylla (VIII) 46 Draba hammenii (?) 2 praos А. (XI) 32 raba funckiana a 87 denii (1X) 38 — lin Draba sericea (? Draba hemsleyana (VIII) 83 Draba macleanii (IV) 40 Draba splendens (VIII) 50 D soratensis (IV) 4 — Draba chionophila (ХІ) 8 — Draba nivicola (X) 58 - Draba alyssoides ( ?) 23 — Draba pennell-hazenii (ХІ) З <@— Draba extensa (V) 85 D nA na (X) 9 1 [—— Dra атса (ХІ) 73 ч — 80 Daba rositae (?) 14. аф Draba зааг e 74 ӨГ Draba boyacana (?) 79 Draba arin F ree (?) 11 Draba pus Draba чин М (8) 102 "T — rV) 1 —— jaegeri Draba pie Lim 34 Draba scopulorum (IV) 47 -«&— Draba tomentosa (XIV) EU 113 raba tome Draba ladina (XIV) 115 EU Draba dubia seh ramoisissima (XII) 76 6 ж clade VI clade V clade IV ntosa (XIV) EU 112 +- (XIV) 1 114 EU - 4- — Draba funiculosa (11) 33 —[- Р 59 Draba bellardii (ХІ) 7 —a L Praba cryophila (?) 5* 95 Draba cuatrecasana (?)1 ptans ?) 72 Draba ге 100 Draba platycarpa (XVI) 69 9 Draba cuneifolia (XVI) 61 94 Draba abor (XVI) 84 T^ uil av — — (I) 10 94! Draba aizoides (1) 107 EU | clade Il | clade | A:thriata rlalfnid, cil | 7 Arabis alpina AFR в 100 L Arabis alpina EU pee Arabis bryoides EU Arabis blepharophylla | 0.02 substitutions/site 100 Arabis pumila EU pe cias 111 ЕО clade Ill 110 EU Volume 89, Number 1 2002 Koch & Al-Shehbaz 97 American Draba Table 2. parentheses in Figure 2. Sectional classification system according to Schulz (1927 ). Roman numerals correspond to those used in Sectional classification according to Schulz (1927) Number of species and distribution 1. Aizopsis DC. II. Linodraba O. E. IH. Chrysodraba D( IV. Rhabdodraba o. E. Schulz . Schulz 28 spp.. Schulz | sp., 57 spp., 10 spp., 8 spp.. MI ne. | sp., VII. Helicodraba O. E. 2 spp.. VIII. Cladodraba O. E. Schul 14 spp., IX. Dolichostylis (Turcz.) O.E. Schulz З spp. X. Adenodraba O. E. Schulz Mediterranean and rest of Europ South American (Falkland e m m Patagonia) Asia and North America South America South America (Ecuador and Peru) NW Africa (Morocco) NW Africa (Morocco) South. America (Colombia south into Patagonia) South America (Venezuela) , Central America (Mexico and Guatemala) and South America (Colombia, Ecuador, and Bolivia) XI. Chamaegongyle O. E. Schulz 0 spp.. XIL Phyllodraba O. E. Schulz 29 spp.. ХШ. Nesodraba Greene | sp XIV. Le е DC. 59 spp.. V. XV. Drabella 23 spp.. XVI. Schulz XVII. Schulz ipie (Ral) O. E Abdra (Greene) O. E. © spp.. l sp., Europe, South America (Colombia and Venezuela) North \me rica, К. Asia, and Himalayas Asia. North America, and South America (2 spp.) Asia, North America, and South pues (2 spp.) North America and South. America (2 spp. North America no. 47). This species was positioned between clades IV and II by distance analysis (Fig. 2) with higher sequence similarity to taxa from clade IV, but par- simony analysis positioned it in clade IV with high confidence (Fig. 3). We obtained no significant phy- logenetic structuring among these six major clades, and the trees remained unresolved in this respect. TRNL DATA We considered 76 sequences of which 69 are novel reports herein. In total we considered 72 dif- ferent taxa. The alignment of the tral. intron is 382 bp in length and. contained 6 bp of the first exon of the trnL(UAA) gene at the 5'-end and 30 bp of the second exon at its 3'-end. The alignment of the trnL-F-intergenic spacer region is 404 bp in length and contained 5 bp of the second exon of the trnl(U AA) gene at its 5'-end and 15 bp of the trnF gene at the 3'-end. Both alignments were combined to a final alignment 786 bp in length consisting of the entire tral. intron as well as the second exon of the trnl. gene and the complete tri L-F-intergenic spacer. A region of six base pairs between the in- tron and the spacer that showed no variation among Draba accessions sequenced by Widmer and Bal- tisberger (1999p) is missing. The whole data matrix is interspersed with 24 gaps (7 within the tral, in- tron, including 4 autapomorphies, and 17 within the spacer region, including 8 autapomorphies, re- spectively). Within the tral. intron 47 variable nu- cleotide positions were detected (including 20 au- tapomorphies); within the trnL-F spacer 52 out of 87 variable nucleotide positions were uninforma- tive. The distance tree is shown Figure 4 with support indicated by bootstrap greater than 50%. Fitch parsimony analysis with a subset of taxa similar to the ITS analysis resulted 82 MPTs with 112 steps CI of 72%. and RI 79%. This subset comprises the same set of taxa | » analysis with additional Draba rositae length, used for ITS (acc. no. 14), because this taxon has been separated from other Draba taxa by a relatively high bootstrap support of 63% (Fig. 4). bootstrap and Significant results of the decay analysis are shown along ¢ Figure 2. \rrows indicate positions of taxa that differ from the plastid analys sligale concerte d evolution (for de ail. refer to text). Branch length of S used for triplet comparisons of ITS types to inv Draba funiculosa (ace. no. ЗЭ. clade HH) has Р ш correspond to those in were not recognized by Se 'hulz (1927) are marked w not been drawn to s ible 1. In parentheses the есп — according to Schulz (19 ith ° Neighbor-joining distance tree of ITS sequence Наш: Bootstrap support is given from 1000 replicates. ) | PI sis (see Fig. |). terisks mark accessions that were accession са TS that a ces of outgroup taxa а from irabis and brisa HOR e (refer to text). Taxonomic 27) is given. Ta ITS sequer have been published previously by Koch et al. (19994). Ael an Draba are indictaed by 98 Annals of the Missouri Botanical Garden 99% == DRABA PULVINATA 80 =" clade VI +4 ا‎ DRABA BOYACANA 79 DRABA RAMOSISSIMA 76 83% +2 clade V DRABA HITCHCOCKII 64 53% +1 DRABA OLIGOSPERMA 68 0, ere DRABA SCOPULORUM 47 clade IV DRABA PAYSONII 71 DRABA AIZOIDES 103 829 n EROPHILA VERNA 110 clade Ill 68% DRABA TOMENTOSA 112 +1 DRABA CUATRECASANA 1 63% +1 clade Il DRABA CRYOPHILA 5 : DRABA ARABOIDES 84 >5 clade DRABA CUNEIFOLIA 61 ARABIS PUMILA Figure 3. Strict consensus tree from o h parsimony using a subset of ITS sequences. Bootstrap support from 1000 replicates is seen above branches: dex alues are indicated below branches. Accession numbers following taxa комны to “Table 1. Clades I-V Dee is to ITS clades l-V from Figure 2 branches in the strict consensus tree (Fig. 5). The results of this analysis are congruent with the find- ings of the overall distance analysis (Fig. 4). Five major IN could be distinguished (clades A—E, Figs. 4 and 5) with bootstrap support greater than 50%. There is no support to combine D. oligosper- = ma (acc. no. 68) and D. scopulorum (acc. no. 47 to any of these clades. However, D. oligosperma is more closely related to cp genome types € and D (Figs. 4 simony analysis. The occurrence of these particular cp genome and 5) as revealed from distance and par- types according E their geographic origin is sum- marized in Table 3 4) rec- The more inclusive distance analysis (Fig. ognized five major clades among Draba, with D. cuneifolia (acc. no. 61), D. platycarpa (acc. no. 69), D. reptans (acc. no. 72), and D. araboides (acc. no. 84) in clade A only distantly related to the remain- ing clades. Plastome types from European Draba were found mostly in clades В and D, but one ad- ditional plastome type from Draba aizoides (acc. no. 108) (Widmer & Baltisberger, 1999b) showed an intermediate position between clades E and C/D. In the present analysis, three plastome types of D. aizoides from Widmer and Baltisberger (1999b) were included and all three were diverse. Draba aizoides herein investigated from Germany (acc. no. 103) is more closely related to D. aizoides (acc. no. 107 from Widmer & Baltisberger, 1999b); it assorts — Volume 89, Number 1 2002 Koch & Al-Shehbaz 99 American Draba raba dr. кш — Draba | саа pulvin I Dra — Draba solitara 6 Draba barclayana11 ana13 aa — berryi 18 [Draba schusten 15 ri, de uds ТРЕТИ 33 — Draba cruciata a. 89 Draba — 45 Draba boyacana 7 J —- Draba soratensis 49 Draba standley Draba hidalgensis АЫ Нот чш clade Е Draba araboides 8 L .. ceca АП Draba nivicola 58 —— Draba aureola 1 52 — Draba pulvinata 80 Es Draba aizoides 108 EU RA al aca 1105 EU 1 — aizoides 109 EU a tomentosa 112 EU мі 114 EU clade D 51 74 Draba ladina 115 EU raba stylaris 106 EU Draba scopulorum 47 raba rositae 14 «GH = 63 | Draba phoney piel 73 — Draba subumbellat rta lol ensis 56 Draba magellanica 41 Draba jaegen D aba incerta 65 clade C ramosissima 76 Draba gilliesii 34 Dr. — helleriana 54 — Draba paysonii 71 Draba rei NR 51 57| Draba hitchcockii 6 Draba glabell Draba — 31 —— Draba cryophi D aizoides 107 EU Draba pennell-hazenii З «8&— clade B b sana 1 Draba aizoides 103 EU Draba aizoides 104 EU [^ Draba reptans 72 n h р, 3 Ras clade A 100 5 Draba araboides 84 Draba platycarpa 69 | 0.05 substitutions/site | Figure 4. Neighbor-joining distance tree of trnL sequence data. Bootstrap support is given from 1000 replicates. Arrows indicate positions of taxa that differ fi "ig. Accession numbers of taxa correspond to Table 1. European Draba are indictaed by rom n the ITS analysis (see 100 Annals of the Missouri Botanical Garden г DRABA PULVINATA 80 0, 5 clade Е lo DRABA BOYACANA 79 DRABA OLIGOSPERMA 68 76% DRABA TOMENTOSA 112 — $—1l1 dau DRABA STYLARIS 106 0 52% +1 o DRABA MAGELLANICA 41 13 h | DRABA JORULLENSIS 56 0 100% 66% DRABA ROSITAE 14 >5 == clade С + DRABA HITCHCOCKII 64 Р DRABA RAMOSISSIMA 76 af / | DRABA PAYSONI 71 DRABA SCOPULORUM 47 DRABA CUATRECASANA 1 72% ا‎ DRABA AIZOIDES 103 clade B +2 DRABA CRYOPHILA 5 os DRABA CUNEIFOLIA 61 clade A Ll DRABA ARABOIDES 84 Figure 5. Strict consensus tree from Fitch parsimony using a subset of trnl sequences. Bootstrap support from 1000 replic ales is seen above branches; dec cay values are indicated below branches . Accession numbers following taxa correspond to Table 1. Clades A-E correspond to trnL clades A-E from Figure 4 within clade B (Fig. 4). aizoides are separated by seven — Some 7. D. oli- gosperma, асс. no. 68, and D. aizoides, ace. no. 107 from Widmer & Baltisberger, 1999b) are not sup- ported significantly within any of the five clades A to E (Figs. 4 oligosperma is positioned basal to clades C and D These two plastomes of D. plastome types (D. scopulorum, acc. no. 4 and 5). The plastome type of Draba in distance and parsimony analyses (Figs. 4 and 5). The plastome type of D. scopulorum is supported within clade D by distance analysis (Fig. 4); how- ever, this is not supported by a bootstrap value greater than 50%, and parsimony analysis with a reduced data set (Fig. at all. cpDNA sequence data from Erophila were 5) provided no information not available, and research of the authors focusing on annual Eurasian Draba is in progress. ITS VERSUS frnl DATA There is some congruence between phylogenetic trees from both data sets (Figs. 2 and 4): (1) Amer- ican D. cuneifolia (acc. no. 61). D. platycarpa (acc. no. 69), D. reptans (acc. no. 72), and D. araboides (acc. no. 84) are separated from all remaining Dra- ba taxa, American as well as European material, grouped in ITS clade I or corresponding trnL clade and are A; (2) the majority of American Draba is combined to ITS clade VI, which largely corre- sponds to trnL clade E; (3) taxa that were integrated Volume 89, Number 1 2002 Koch & Al-Shehbaz American Draba Table 3. Distribution of plastome types A-E among Draba accessions analyzed herein according to their sam- pling area. Unic “ambiguous” jue or plastome type of D. scopulorum, D. oligosperma, and D. aizoides (acc. nos. 47. 68. and 108, respective bers of accessions showing a particular plastome type are Чу) have been excluded here. Num- indicated. = : Plastome type Origin of y 1.9 (according to Fig. 2) accessions A B C D E Europe 3 0 LS 3 8 5 Peru 1 8 Venezuela l 6 Colombia 3 l 5 one 2 l hile 1 l ae 2 3 Bolivia 2 Ecuador into ITS clades IV and V possess a plastome type typical for taxa from clade C; and (4) European Draba is divided into two groups among both data sets (clade B and D in the plastid data set: and significant subgrouping within clade HI in the ITS data set including Erophila), but the separation of taxa forming clades B and D does not correspond to that in the ITS clade HI. However. eight accessions (D. pennell-hazenit, no. 14: D. farsetioides, 33; D. helleriana, acc. no. acc. no. 3: D. rositae, acc. D. funiculosa, acc. no. 47; D. oligosperma, acc. no. 73) when both phylogenies (Figs. 2 and 4) were com- acc. no. 31; scopu- 54 D. no. 68: and D. streptocarpa, acc. lorum. acc. no. from America assign to different clades pared with each other. In addition. plastome types from Draba aizoides are found at three different po- sitions (Fig. 4). Most American Draba species with ITS type VI (Fig. 2) possess a plastome type group- E (Fig. 4). However, American Draba 14). no. 54). and D. streptocarpa (acc. no. 73) from ITS clade VI ing into clade rositae (acc. no. D. helleriana (acc. do have the plastome type C and not type К, as in the majority of taxa investigated, and within D. pen- nell-hazenii (ace. no. 3) and D. farsetioides (acc. no. 31) from ITS from clade B. Draba oligosperma (acc. no. 68) from clade VI we found a plastome type ITS clade IV (Fig. 2) and D. scopulorum (acc. no. 47). which is also grouped to ITS clade IV in par- simony analysis (Fig. 3). do not have the plastome type € (Fig. clade IV. However, both plastomes of these taxa, D. 4) as do the remaining taxa of ITS scopulorum and D. oligosperma, are found periph- eral to plastome clade C (Fig. 4). Interestingly, American D. funiculosa (ace. no. 33), with an ITS type closer to European Draba, showed a plastome type grouping to clade E (Fig. 4). DISCUSSION STATUS OF TRIBE ALYSSEAE PHYLOGENETIC POSITION OF AND THE GENUS DRABA Preliminary molecular phylogenetic analysis — Koch, unpublished) suggests that the morphologi- cally poorly defined tribe Alysseae is not. mono- Two of Schulzs (1936) tribes. Lunaricae and Drabeae. Janchen (1942) and Hayek (1911) as subtribes of the Alysseae. The Al-Shehbaz (1987) in the Alysseae, was shown by Zunk et al. phyletic. were recognized by genus Camelina, which was treated by (1999) to be very close to genus Capsella, a genus traditionally placed in the tribe Lepidieae. A ma- jority of the 650 species of the Alysseae (sensu Hayek, and genera Alyssum (170 spp.) and Draba (350 spp.). — Janchen. Al-Shehbaz) belong to the Fhe present analysis supports a closer relationship among Draba, Eurasian Arabis. and the European Aubrieta. Eurasian Arabis and Aubrieta have been shown to represent a core group of the tribe Ara- bideae (Koch et al.. 1999a). However, as tradition- ally delimited by Janchen (1942). Hayek (19101). Schulz (1936), the tribe Arabideae breaks down as 1999a) (Fig. 1). Similarily. the tribe Lepideae is not monophyletic. an unnatural group (Koch et al.. and several of its constituent genera are sister laxa in other tribes of the Brassicaceae (Koch et al.. 2000, 2001). from all molecular studies of the that the tribal relationships of Schulz (1936). Janchen (1942). Hayek (1911). based solely on morphological characters, are high- А | 8 g The consistent conclusion reache d Brassicaceae S and which are ly artificial and do not reflect phylogenetic rela- tionships. Such artificiality was also elucidated at the generic level, especially by molecular studies on Cochlearia (Koch et al., 1999b), Thlaspi (Mum- menhoff et al.. 1997a, b), and Arabis (Koch et al.. 1999a, 2000, 2001). Our findings should encourage other investigators of cruciferous genera to test phy- logenetic assumptions at a higher taxonomic level. HYBRID ITS TYPES AND RETICULATION We found no ambiguous nucleotide positions when sequencing Draba ITS regions, although hy- brid origin for several taxa is likely because of some characteristics frequently found in species groups showing hybridization and introgression: (1) incongruencies between nuclear- and plastid-de- 102 Annals of the Missouri Botanical Garden Table 4 References for chromosome numbers among Draba taxa under study. For taxa that were recognized by Schulz (1927) information about his sectional treatment is given (see also Table 2). Sectional classification according to Schulz (1927) Taxon Chromosome no. and reference Ploidy level Not mentioned D. eryophila 2n = 48 (Galland & Pfitsch, 1986b) hexaploid by Schulz Adenodraba D. jorullensis n — 12 (Beaman et al., 1962) triploid Aizopsis D. aizoides 2n — l6 (Hess et al., 1977) diploid Calodraba D. gilliesii n — 24 (Boecher, 1966) hexaploid Chamaegongyle D. bellardii 2n = 48 (Galland & Pfitsch, 1986a) hexaploid D. chionophila 2n = 48 (Galland & Pfitsch, 1986a) hexaploic Chrysodraba D. oligosperma 2n = 32, 64 oe 1972) tetraploid, octoploid D. paysonii 2n — 42 (Mulligan, 197 — *2 Doliostylis D. pulvinata 2n = 48 (Galland & Pfitsch, 1986a) hexaploic Drabella D. crassifolia 2n = 40 (Price, 1979 НИР Leucodraba D. carinthiaca 2n = 16 (Hess et al., 1977) diploic D. dubia 2n = 16 (Hess et al., 1977) — D. fladniziensis, 2n = | (Hess et al., 1977) dipl D. glabella n= 40 (Mulligan, 1970) oc E decaploid D. incerta п = B (Mulligan, 1966) 14-ploi D. ladina 2n = (Hess et al., 1977) tetraploid D. magellanica n = 32 (Heilborn, 1941) octoploid D. stylaris 2n = M (Hess et al., s 7) tetraploid D. tomentosa 2n = 5, (Hess et al. 7) iploid Phyllodraba D. helleriana n = 8 (Ward, 1983), n = 9 (Ward & Spellenberg. 1988) diploid D. mogollonica n = 16 (Ward, 1983) tetraploid D. ramosissima n = 8 (N 1969) diploid D. streptoc arpa n — 20 (Price, 1979) pentaploid Rhabdodraba D. pickerengi 2n = 24 (Favarger & Huynh, 1965) triploid Tomostima D. cuneifolia 2n = A (Rollins & Ruedenberg, 1971) diploid D. platycarpa 2n = 32 (Rollins « a 1971) tetraploid D. reptans 2n = © (Live & Love 2) diploid, tetraploid 2n = 32 (Smith, 1965) rived phylogenetic hypotheses, (2) the occurrence of polyploids (Table 2). Additionally, for European Draba extensive hybridization and reticulation have been shown previously (Widmer & Baltisberger, 1999a, b). Therefore, we have to assume that con- certed evolution of putative ancestral ITS types re- sulted in one dominant ITS copy. Different modes of concerted evolution of ITS regions have been discussed, e.g., for Gossypium (Wendel et al., 1995a, b), roses (Vissemann, 2000), Quercus (Muir et al., 2001), and other plants (Buckler et al., 1997), and for some cruciferous taxa (Hilliella, Cochleariella in Koch & Al-Shehbaz, 2000; Arad- idopsis in O'Kane et al., 1996; Cardamine in Franz- 1998; Microthlaspi in Mummenhoff et al., Sequence divergence values of ITS types ke et al., 1997b). from putative parents, which produced hybrids in which concerted evolution has been documented, 7.8% (Koch «€ Al-Shehbaz, 2000: 270). Sequence divergence values for Draba ITS range from 3.1 to types exceeded 9%, and these might indicate a rel- atively old age for the genus. Divergence time es- timates of cruciferous plants are found in Koch et al. (2000) for chalcone synthase and alcoholdehy- drogenase and in Koch et al. (2001) for chalcone synthase and maturase К. A comparison of these time estimates with an ITS-derived phylogeny com- 19994) shows that 1% ITS sequence divergence corre- prising a similar set of species (Koch et al., spond to approximately 0.5 to 1.0 million years. American Draba occur at different phylogenetic positions, sometimes together with European Draba (Fig. 4). polyploids (Table 4), and it is likely that hybridiza- plastome types They include numerous tion and polyploidization played a major role in their evolution. In order to test this assumption and show interrelationships between different ITS types, several comparisons of three ITS sequences from three different ITS clades (Fig. 2) alyze the distribution of variable nucleotides. These were done to an- Volume 89, Number 1 2002 Koch & Al-Shehbaz 103 American Draba American taxa compared as such are represented in Figure 2. First, American Draba cryophila (acc. no. 5, clade H) was compared with D. hitchcockii (acc. no. 64) and D. V and VI, respectively. The by asterisks — tucumanensis (acc. no. 51) of clades distribution of nucleotide diversity revealed that D. hitchcockii had 21 characters in common with D. tucumanensis and only 5 characters with D. ery- ophila (acc. no. 5). Of the eight nucleotide positions where D. hitchcockii differed from D. eryophila and D. tucumanensis, three were identical to positions from D. jaegeri (acc. no. 66, also from clade V). Synapomorphic characters were found within the TS2 regions of D. tucumanensis (acc. no. 51) and D. hitchcockii (ace. no. 64), but not within D. ery- no. 5) and D. hitchcockii (acc. no. 64) no. 66). Different results were obtained upon comparison 57) of clade VI (Fig. 2) 51) and D. hitech- In this ophila (acc. or D. jaegeri (acc. — of D. jorullensis (acc. no. with D. cockii/ D. jaegeri (accession nos. 04 and 66). case. both D. hitchcockii and D. jaegeri shared five characters with D. jorullensis (acc. no. 57) and sev- en with D. tucumanensis (acc. no. 51). Three of eight characters of D. hitchcockii (ace. no. 04). which were not found in D. tucumanensts (acc. no. tucumanensis (ace. no. 51) and D. jorullensis (ace. no. 57), were synapo- morphic with those of D. jaegeri (acc. no. 66). Fur- thermore, D. tucumanensis (acc. no. 51) contributed to all variable nucleotide positions in the l1TS2 re- no. 64) whereas D. jorullensis gion shared with either D. hitchcockii (acc. r D. jaegeri (acc. no. 66). fico: no. 57) contributed nucleotide positions only within the ITS] preted as the result of hybridization between taxa region. These findings are inter- of clades IV and VI, as represented by D. tucu- manensis and D. jorullensis, respectively, which produced the hybrid ITS types found in clade V as represented by D. hitchcockii and D. jaegeri (Fig. Four additional comparisons are made to test the possible role of hybridization in the formation of ITS types of clade VI (Fig. 2) that includes no Eu- ropean taxa: (a) D. eryophila (асс. no. 5. clade I1). D. jorullensis (acc. no. 57, clade VI), and D. tucu- clade IV): (b) D. eryophila nos. 5 Э manensis (acc. no. OL, and D. helleriana (acc. and basal in clade VD. and D. tucumanensis (acc. no. 51. IV): (e) D. tucumanensis and D. jorullensis (acc. nos. 51 and 57) and D. helleriana (ace. no. 54): as well as (d) D. eryophila, D. helleriana. and D. jo- 5. 54. and 57, respectively). clade rullensis (acc. nos. The first two comparisons (first, among ace. nos. 5. 51. thal accessions 54 and 57 had only 6 characters in 57. and second, among 5, 51, 54) revealed common with accession number 5 of clade H but 23 characters with accession number 51 of clade IV. The third comparison (acc. nos. 51, 51. and 57) revealed that all synapomorphic characters com- bined exclusively accession numbers 54 and 57. This might indicate close relationships of taxa from clade VI, which includes accession numbers 54 and 57, and a monophyletic origin. The fourth com- parison (acc. nos. 5, 54, and 57) provides no ad- ditional data toward the hybridization of taxa from either clade Hl or IV as a source for ITS types р in clade VI. characters with accession number 5. but accession Accession number 54 shared only : number 54 shared 21 characters with accession number 57. Therefore, based on ITS sequences, it was not possible to verify the hypothesis that ITS types from clade VI evolved via concerted evolution after hybridization. Comparison of the maternal plastid phylogeny Fig. i the ITS-derived phylogeny indicates that plastome see taxa indicated by black arrows. ) with types found in D. pennell-hazenii (acc. no. З) and D. farsetioides (acc. no. 31) similar to those found in European taxa (plastid clade B) are distributed among accessions with an ITS type of clade VI. By contrast. D. funiculosa (acc. no. 33) has a plastome type found in clade E, but its ITS copy is related to European ITS types found in ITS clade HI. Plas- tome types from plastidic clade € (D. rositae. D. helleriana, and D. streptocarpa or acc. 13. respectively) contributed to the genetic consti- 'S clade VI. plastome type E, which is distributed among Amer- ican Draba of the ITS clade VI. plastome type from the European D. aizoides (acc. no. 108, Fig. of D. aizoides (acc. nos. 14, 54. tution of taxa found in TI Interestingly. is similar to a 1) but not from the other accessions 107, 109, 103), which as- sort to plastome clades B and D. Extensive intra- nos. specific cpDNA haplotype variation alpine D. aizoides was demonstrated by Widmer and Baltis- berger (1999b). Therefore, one could speculate that ancient gene flow and chloroplast capture might have resulted in the similar transmission of Euro- pean plastome types into American Draba and in the constitution of the whole complex of taxa with an ITS type from clade VI. Plastome types found in taxa from this group mostly belong to plastome type E (only 5 out of 51 accessions with ITS type VI have been shown to possess other. plastome types than plastome type E) and are only weakly differentiated from each other. There is no resolu- tion within clade E in the trnb-derived phylogeny. Taking both approaches into account, the com- parison of ITS types and the distribution of plas- tome types among different clades reveals strong 104 Annals of the Missouri Botanical Garden evidence for extensive reticulation during the evo- lution of the American complex of Draba. However, concerted evolution of ITS types has resulted in homogeneous ITS copies within single taxa: no dif- ficulties were encountered while obtaining the se- quences by direct cycle sequencing of the purified PCR products. The genetic source of plastome and ITS type variation probably originated from Eur- asian taxa, for which very complex evolutionary scenarios have already been described. Subsequent hybridization probably took place among different ITS resulted їп groups of American Draba (e.g., taxa from clade H with taxa from ITS clade IV taxa from ITS clade V), and some taxa from the ITS clade VI crossed back with taxa from clades V or IV possessing plastomes of type С. Only taxa from ITS clade I (corresponding to trnL clade which are distributed in America, represent a ge- netically separated lineage (Figs. 2, 4). These hypotheses might be simplified; however, they will become more complex when additional individuals from a single taxon or one population are analyzed, For example, extensive ongoing gene flow has been reported within Draba from the Alps and Scandinavia (Widmer & Baltisberger, 1999a, 1992a-d; 1993). Although the study of herbarium material is b; Brochmann et al., Brochmann et al., greatly affected by undersampling. it is rather re- markable that in 8 out of the 74 samples of Amer- ican Draba (> 10%) analyzed, incongruencies be- tween ITS- and trnl-derived phylogenies were detected. One has to assume that much higher lev- els of incongruencies will be observed in genetic analyses at the populational level. SECTIONAL CLASSIFICATION OF DRABA Draba is generally recognized as a natural genus of about 350 species distributed primarily in the Arctic, subarctic, and alpine regions of the North- ern Hemisphere, with about 65 species in South America along the Andes from Colombia to Pata- gonia (Al-Shehbaz, 1987). In the present study we focused on American Draba, including a limited number of species from Europe (8 species from 10 accessions). Eleven species grow in Mexico and 1984). More than 100 grow in North America and Greenland (Rollins, Central America (Rollins. 1993), with the ranges of about 20 of these extend- ing into the Arctic and subarctic Eurasia Schulz (1927, 1936) divided Draba (excluding Erophila) into 17 sections (Table 2) that have been Fernald, 1934; Al-Sheh- baz, 1987) to be highly artificial. Schulz defined his considered by some (e.g., sections primarily on the presence or absence of A), all of stem leaves, flower color (white vs. yellow), the presence or absence of median nectaries, and style length. Four of Schulz’s sections are no longer rec- ognized in Draba. Section Helicodraba O. E. Schulz was shown by Hyam and Jury (1990) to belong to the Southwest Asian Graellsia Boiss. Section Ne- sodraba (Greene) N. Busch was reduced by Ber- kutenko (1995) to Schivereckia Andrzeiowski ex Raf.) О. E. Schulz and Abdra (Greene) O. E. Schulz are believed to form DC. Sections Tomostima — the well-defined genus Tomostima Raf. (Price & Al- Shehbaz, unpublished). Species as assigned to the remaining sections are still maintained in the genus and are not currently questioned (Table 2). Section Acrodraba O. E. Schulz (1 species, Draba oreadum Maire) is restricted to northern Africa, whereas sec- tions Rhabdodraba О. E. Schulz (10 species), Ad- enodraba O. E. Schulz (9 species). Tylodraba O. E. Dolichostylis (Turez.) О. E. species), Calodraba O. E. Schulz (14 spe- Schulz (8 species). Schulz (3 cies), and Chamaegongyle O. E. Schulz (6 species) are distributed in Central and South America. With the exception of section Aizopsis DC. (28 species), which is almost exclusively European, larger sec- tions include the North American and Eurasian Chrysodraba DC. (57 species), the North American and East Asian Phyllodraba O. E. Schulz (29 spe- cies), European and central Asian Leucodraba DC. (59 spp.). and the Asian Drabella DC. (23 spp.). The last included two South American species that doubtfully belong there. Four out of the 17 sections sensu Schulz (1927), which comprise only five species (1.5% of the total in Draba), were not included in the present study. Sections Helicodraba and Nesodraba (4 species) have already been combined under different genera as outlined above. Material of the North American section Abdra (1 species) and African section Ac- rodraba (l species) were not available. Schulz's sectional classification is given in Roman numerals in parentheses in the phylogenetic ITS tree (Fig. 2). The phylogenetic analysis of the ITS sequence data (Fig. 2) shows that all sections exclusively dis- tributed in South and Central America (Table 2, Rhabodraba, Doli- chostylis, Adenodraba, Chamaegongyle) fall within clade VI. Draba aureola (ace. no, 101) of section Phyllodraba (North America, Asia), D. crassifolia (ace. no. 59) North America), and D. cruciata (ace. no. 60) of section Chrysodraba sects. Tylodraba, Calodraba, This clade also includes such taxa as from section Drabella (Eurasia, (Eurasia, North America). This suggests that South and Central American Draba probably evolved from other sections distributed in the Northern Hemisphere. With our molecular data it is difficult Volume 89, Number 1 2002 Koch & Al-Shehbaz American Draba South- North migration events (from South America to the to distinguish possible, relatively recent North) from the earlier North-South migration route. However, in this case some morphologically defined wider distribution species should have a much range, which is not the case. A few species from exclusively South and Cen- tral American sections fall within other clades. For example. Draba bellardii (acc. no. 7) of section Chamaegongyle belongs to the ITS clade Il. where- as both D. Rhabdodraba and D. gilliesii (ace. no. 47) of section 34) of sec- scopulorum (acc. no. tion Calodraba belong to clade IV. Some level of congruence was found when com- paring the ITS clustering of European Draba with the sectional classification of Schulz (1927. 1936). Taxa comprising the ITS clade HI were assigned to sections Alzopsis, e.g., D. aizoides, and Leucodraba. e.g.. D. tomentosa. These sections are primarily Eurasian. Interestingly, D. funiculosa (ace. no. 33) falls within this ITS clade, branch (more than twice longer than the maximum but the extremely long distance found. within the total ingroup) that sets it off from the remaining European taxa indicates that its ITS type has considerably diverged. This partic- ular subantarctic species comprises the monotypic section Linodraba. A comparison of Schulz’s (1927, 1936) sectional classification with the present molecular study re- veals that section Zomostima (XVI, Table 2). which corresponds to the ITS clade | (Fig. 2) and irni clade A (Fig. 4), is the most clearly supported. Schulz (1927, 1936) and all European taxonomic treatments maintained Erophila as a distinct genus that is closely related to Draba. However. the pre- sent ITS data show that it is better integrated into Draba. and these results support the view held by most North American. botanists (e.g... Fernald. 1934; Al-Shehbaz. 1987: Rollins, 1993). The only morphological difference between Erophila and Draba is the presence of bifid petals in Erophila rather than entire to deeply lobed ones in Draba (Al-Shehbaz, 1987). In conclusion, our findings do not provide much support for Schulz’s (1927, fication, although they suggest some species ten- 1936) sectional classi- tatively correspond to a few of his sections. How- ever, several taxa (D. pennell-hazenii, D. rositae. D. farsetioides, D. funiculosa, D. scopulorum, D. hel- leriana. D. oligosperma, and D. streptocarpa, acc. 4, 31. 33. 47, 54, 08, which fall into different clades when the ITS- and — nos. 3, 73, respectively), пі -derived phylogenies were compared, indicate extensive cross relationships not only among taxa of primarily South and Central American sections, folia (diploid, acc. but also among sections with other geographical centers of distribution, e.g., North America, Europe Eurasia), and the subantarctic. A case in point is — the South American Draba magellanica from Ar- gentina (sect. Leucodraba, XIV). which has a chlo- roplast type (clade С, Fig. 4) not found within the remaining species of this section (e.g... D. dubia within tral. clade D). Two other species from Ar- 33. and D. magel- no. 41) have chloroplast types E and gentina (D. funiculosa, acc. no. lanica, acc. C. respectively Although one might argue that in order to better within achieve a understanding of the sectional Draba, needed, especially from the sections not included classification more analyses are in the present study, it is safe to conclude that our molecular data clearly show that the conventional sectional classification of Draba as conceived by Schulz (1927. 1936) is an artificial one. CHROMOSOME AND APOMIXIS NUMBER, POLYPLOIDY. Because the present study is based on herbarium material, no chromosome counts were made from the plants investigated. However, several chromo- some counts are available from the literature (Table base chro- 1). From this, it seems likely that the mosome number for Draba is x = 8. This is also the base chromosome number for sister European genera Arabis and Aubrieta (Fig. 1). Chromosome data demonstrate the existence of diploid, triploid. hexaploid, octoploid, and = 8 (Table 4). Higher chromosome numbers (2n = 112) have been re- tetraploid, pentaploid, decaploid taxa based on x ported for D. incerta (acc. no. 65) from North Amer- ica (sect. Leucodraba). The well-defined clade I (ITS sponding clade A (trnL data), comprising D. cunei- data). or corre- no. 61). D. platycarpa (tetra- 69), D. no. 72), and D. araboides (no count ploid, ace. no. reptans (diploid and tetraploid, ace or кинин number, асс. no. 84), included dip- loid and tetraploid species (Table 4). In summary. XVI) comprises mostly diploid species (as known so far). this clade (well defined as sect. Tomostima, and also represents an ancestral clade to all maining Draba. This finding will need further in- vestigation because section Tomostima is restricted to North and South taxa served as ancestors for Eur- \merica, and it is hard to be- lieve that these asian Draba. The consequence is the recognition of genus Tomostima (Price & Al-Shehbaz, unpub- lished). The ITS data indicate a hybrid origin for Draba 106 Annals of the Missouri Botanical Garden ramosissima (clade V, acc. no. 76, Fig. 2), a species reported to be diploid (Table 4). As outlined before, direct ITS sequence comparisons demonstrated that ITS types from clade V types of clades VI and IV. Therefore, most likely evolved from ancestral ITS in D. ramosissima speciation via hybridization must have taken place at the diploid level and only dip- loid parental taxa could have served as progenitors. Another diploid, D. helleriana (acc. no. 54) has an — ITS type corresponding to clade Ш (Figs. 2, 3. 6 C (Figs. 4—6). incongrueney could be best explained by gene flow and a plastome type to clade This between different groups of taxa, thus indic ating another hybrid speciation at the diploid level. cept for D. cuneifolia and D. reptans from section Tomostima and D. helleriana and D. ramosissima from section Phyllodraba, counts for the remaining American Draba species showed polyploidy (Table 4). By contrast, triploids (D. jorullensis, acc. no. 56; D. pickerengii, acc. no. 44), pentaploids (D. cras- 59: D. and an aneuploid (D. paysonil, streplocarpa, acc. no. 13). pentaploid + 2 no. 71) have been reported (Ta- sifolia, acc. no. chromosomes, acc. ble 4), and these taxa grouped E plastidic clades C and D or ITS Apomixis has been bibi: in a few species clades IV and V such as Draba verna or D. oligosperma (Mulligan & Findlay, 1970; Mulligan, 1971, 1972, 1976: Price, 1980), and it is likely that it might be more widespread among American Draba. Therefore, in Draba apomixis might be an additional mechanism that led to the recognition of several hundred, mor- phologically defined taxa. The occurrence of apo- mixis has also been documented for other Brassi- caceae, and North American Arabis parallels the 1966; Rollins, 1983). North American Arabis, from which many species situation in Draba (Bócher, were more appropriately transferred to Boechera (Live & Lóve, 1975; Weber, 1982, 1989), is a young taxon, presumably of Pleistocene origin. This genus exhibits remarkable morphological. ecologi- cal, and physiological diversification leading to the recognition of more than 50 species, and most of them show the trait apomixis. As suggested by the low sequence divergence 'S clade VI (Fig. 2) or the corresponding trnL clade E (Fig. 4) values, American Draba either of the I'l probably are relatively young taxa similar to Boech- era. Comparable low levels of molecular variation have been found separating the North and South Moench, sensu Meyer (1973, r American Thlaspi L. sensu lato from their 1993: Koch et al., Europe several plas- American Noccaea 1979), о Eurasian relatives (Koch et al.. 1998c). Contrary to this, in tome types from D. aizoides were found by Widmer and Baltisberger (1999b), some of them very close to D. aizoides, accession number 107 (DA2, DA3, DA9, DA6, DA8, Widmer & Baltisberger, 1999b), separated by a maximum of seven mutations (in this analysis corresponding to clade B, Fig. 5), and one plastome type (ПАТ) was separated from Draba ai- zoides, acc. no. 108, by two mutations. All samples analyzed by Widmer and Baltisberger (1999b) orig- inated from the Swiss Alps and demonstrate the enormous plastome variation even within one taxon (Draba aizoides) from a restricted area. ITS se- quence variation among D. aizoides accessions (Fig. 2) is much lower, and all D. aizoides accessions analyzed herein confined to one group within ITS clade HI. In this group we also found опе D. ladina (acc. no. 116); this is not unexpected, because this taxon evolved by hybrid speciation involving D. ai- zoides as one parental taxon (Widmer & Baltisber- ger, 1999a) The close relationships between the floras of Central Asia and western North America are well documented (Parks & Wendel, es therein). The putative Bering bridge connected 1993, and referenc- Asia and. North. America several times throughout 1987). and this land bridge probably served as an immi- the late Tertiary and Pleistocene (Parrish, gration route for several Brassicaceae, including Thlaspi (Payson, 1926) and Stroganowia Karelin & Kirilov (Rollins, 1982). Palynological analyses from the Venezuelan Andes document the occurrence of Draba in Holocene deposits after late Pleistocene deglaciation events (Salgado Labouriau et al., 1988 Literature Cited Acarkan, A., M. RoBberg, M. Koch & R. 'hmidt. 2000. Comparative genome analysis reveals extensive conser- vation of genome organisation for ibi thaliana and Capsella ө = Plant J. 32: 55-62 Al-Shehbaz, I. A. . The genera " Ah ysseae (Cruci- ferae; Brassicac ui in the Ead United States. J. Arnold — — 185-244 Anderson, J. K . Wa 1999, Chromosome num- bere — in ги tribe Brassiceae — aceae): Ev- ‚ Evol. 215: 255- 5 arwic ide e" nce from isozyme number. Pl. Sy — B. G., M. J. Sanderson, J. M. Po ^. Woj ciechowski, C. S. Campbell & M. J. Donoghue: 1995. The ITS region of nuclear ribosomal | valuable ource of evidence — i ee аш eae e Mis- souri Bot. Gard. 82: 247-277 Baum, D. A., K. J. ae sma & P och. 1994, A phy- logenetic analysis of [NA — eae) a on nuclear ribosomal DNA se quences. Syst. Bot. 19: 363— 388. Beaman, J. H., ;. De Jong & W. P. Stoutamire. 1962. C hromosome — in the alpine and subalpine baia ` Mexico е mala. Amer. J. Bot. 49: 41-5 ней nko, . 1995. Detective story about one Ls Volume 89, Number 1 2002 Koch & Al-Shehbaz American Draba 107 naean species of Cruciferae. Linzer Biol. Beitr. 27: 1115-1122. йске W.. M. Huthmann & Н. Hurka. 2000. Evolution of крз ЖЕ" {аха їп Е К. Вг. (Brassi Den Nijs, К. Marhold & Н. (e ditors), — RS in P s Habitats. Spec ial e en s in d s and Biodiversity : Bócher, T. W. on — species. 2 Some Biol. Skr. 14(7): 1-74. Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phyloge netic reconstruction. Evolution 12: 7 2. d сеае). A — .E un rimental and cytologic E studies Arctic and montane cruci- fers. der Broc — 1902. Polyploid evolution in arc — alpine Drab К aceae). Sommerfeltia Suppl. 34. ——. 1993 Reproductive strategies. of bio and ipli pecu of arctic Drab (Brassicaceae). P ist. E 185 ү R. Elven. 1992. g‏ ہے sequences of poly — in arctic Draba (Brassicaceae).‏ Evol. Trends Pl. 6(2 —‏ Borgen & I 1993.‏ ships aid chromosome — of N‏ of Draba (Brassicace‏ l. vol. 185: 55- E cological and genetic con- aba Stedje. Crossing relation- ordic populations ith e — on the D. alpina — Nordic J. E 13 : 121- —— ——,. P S. Soltis & D. | Sola, 1992a. gins of octoploid Se — endemic Draba cac- uminum: Electrophoretic and morphological evidence Nordic J. Bot. 12: 257-272. ———— & ——. |992b. Recurrent formation and polyphyly ‘of Nordic polyploids in Draba (Brassi- caceae). Amer. J. Bot. 79: 673-688. YE "Soltis & P. S. Soltis. 1992c. P tic аи and payee ny - —— polyploids Draba (Brassicaceae). Pl. ‚ Ev t» 35-70. . B. Stedje € L. D n. "1992 2d. Gene po acros levels in Draba (Brassicaceae). Evol. ' 0(2): ae E 4 Buckler. и. of Multiple ori- e Trends P. Holtsford. 1997. The — ка s and phyloge netic ا‎ ations. Genetics 145: 621-832. М. Baldwin & M. Ippolito & ' Босана] DNA: . Donoghue, B. € ternal transcribed. spacers m js congruence with morphology. Amer. J. Bot. 62: 903-918 Lo T, < K. L. Huynh. 1965. In: A. Live & О. Solbrig (e г), IOPB chromosome number reports IV. Taxon 14: 86-92. Felsenstein, J. 1985. Confidence limits on phylogenies: i approach using the bootstrap. Evolution 39: 78: 91. 1995. version — Fernald, L. 1934. Draba in te mperate — astern America ` Rhodora 36: 241—261. 285-305, 314-344, —371, 392—404. pls. 290-310. Bleeker, R. Kohrt «€ H. A.. 1998. Molecular systematics at Cardamine and PHYLIP. Package University of Washington, Seattle. Phylogeny Inference Franz * Hurka. allied genera (Brassicaceae): ITS and non-coding chlo- roplast E A. Folia Geobot. Phytotax. 33: 2252 10. «Ж. Phitseh. 1986a. In: A. Live € D. буе (е rs C tibiis number reports. Taxon 35: 195-19 * ————. 1986h. In: A. Live «€ D. tors), C "ocn cm number reports. Taxon 35: 610-013. Galland, Love (edi- Gielly, L. € P. Taberlet. 1994. The use aaah da d "e to resolve plant phylogenies: — oding versus ‚ве- п es. Molec. Biol. Evol. 160-77 77. Нат. Н. H. Har T H. Mes & J. M. E 1994. Melee ular cd — noncoding re- gions of the chloroplast ee in the ( related species. Curr. Genet. 558-500. Havek, A. von. 1911. Enturf eines SC rucifer phyloge netischer Grundlage. Већ. Bot. Ce ntralbl. 127-335. ssulaceae and en Systems — ты. О. ve isk. Bot. E; — р, ^ Ex Some осоо numbers in Draba. 35: 141-142. sdb & Se — mer angrenzender Cds e. Hirzel. 1977. Flora der Birkhauser, Basel. . Jury. 1990. A review of the genus ¡Gala e me — Draba L. section Не odraba Schulz Brassicaceae. Bot. J. Linn. Soc. 102: 9-22. 94 Das System der Cruciferen. Oesterr. -Shehbaz. 2000. айс» a ihe C hine "se AM Wins aceae): Evidence from plastid tral. intron and nuclear ITS D data. Ann. Missouri Bot. Gard. 87: 246—272. & H. Hurka. 1999, Isozvme ме in the poly- ploid complex Mic — perfoliatum (L. lever: Morphology а and онн ior Flo- ra 194 — K ы nhoff € К. Zunk. 1993. Isoelektrische Fokussi rung der Untereinheiten der Rubisco in 7hlas- i (Brassicaceae): Weitere pore auf а 4 rae s Керегі. 104: 371-381. — Huthmann & Н. MER a. 1998a. speciation and evolution in the PHI с a x Coch- learia L. 51—400. Mummenhoff & H. Hurka. The b. Molecular biogeography * evolution of the о perfol- Molecular system- \ sequence eine. Formen- Ixozv mes, (Brassicaceae). Bot. Acta 111: iatum s. |. polyploid complex — aceae): Chloro- de DN A and nuclear ribosomal DNA restriction site varialion. Canad. J. Bot. m — v — & —— . Systematics and e — Ki of i tolerant Thlaspi caeru- lescens in торе: Evide nce from genetic stud- ies based on isozyme analysis. Biochem. Syst. Ecol. 26: 823-838 . Bishop & T. Mitchell-Olds. 1999a. Molecular sy semi s and « e ut is of Arabis and Arabidopsis. Pl. Bi ol. Western Et Д — —————,. 1999. Molecular phyloge- netics of Cochlearia L. and allied genera based on nu- clear ribosomal ITS DNA sequence analysis contradict — concepts of e ir жаны) relationships. РІ. Evol. 216: 207 4 T. pow; Il-Olds. 2000. Compar- ative evolutionary analysis of the chalconesynthase and alcohol dehydrogenase from e and Arabis rel- atives. Molec. Biol. Evol. 17: | 183—149& ; 2001. A systematics of the | hum ae: Evi * from coding plastome matk and nuclear Chs sequences. Amer. J. Bot. 88: 534—5 Кош C. 1998. Infraspecific morphometric and nuclear genome size — in polyploid sa utella ени (Brassica aceae а), Amer. J. Bot. 85: 14 Kowalski. 5. P, T. Im Lan, K. A. I oon & A. H. Pat- erson. и c omparative mapping of Arabidopsis thal- iana and Brassica oleracea chromosomes re “als 108 Annals of the Missouri Botanical Garden lands of conserved organization. Genetics 138: 499— 510. agercrantz, U. 1998. Comparative mapping between Ar- abidopsis thaliana and Brassica nigra indicates that rassica genomes have evolved through extensive ge- nome replication accompanied by chromosome fusions не frequent rearrangements. Genetics 150: 1217— Live, E < D. Lóve. 1975. Nome nclatural notes on Arctic plants. Bot. Not. mus > -523. — Ф — fas nmm B chromosome number re- ports LXXVI. Tax : m 974—598. Меуег, Е. К. 1973. C :onspectus der Thlaspi-Arten Europas, Afrikas — Vorderasiens. Feddes Repert. 84: 449—470. 979. kritische. Revision der Thlaspi- Arten Eu- ropas, Afrikas und Vorderasiens. |. Geschichte, Mór- phologie y C horologie. Feddes Re pert. 90: 129-15 Muir, G., leming & €. Schlótterer. 2001. Three divergent DN A clusters predate the species divergence in Quercus pus a (Matt.) i ie 2 and — из rahui L. Molec. Biol. es 18: 112- Mulligan, G. ^" 19 0. Chromosome numbers of the family J. Bot. 44: 309-319. Cruciferae III. E 1970. NAT studies and i бше allies in Canada and Alaska. Bot. — 1437. I71. Cytotaxonomic studies of Draba species of pay- of — glabe lla Canad. Canada e Alaska: Т ventosa, D. ruaxes, and D. Canad. J. Bot. 49: 1455-1460 19 12. Cytotaxonomic studies of Draba species in хопи. Canada g5 Alaska: oligosperma and D. incerta. ت‎ г. 1763-17 . The genus hr in ( anada Alaska: Key summary. Canad. J. Bot. 5 393 & J. NF indlay. 197 0, Sexual re — tion and ея in the genus Draba. Canad. J. Bot. 48: 209- — M., A. Franzke & М. Koch. 1997а. Molec- ular data reveal convergence in fruit characters, tradi- tionally used in the classifi ation of Thlaspi s.l. (Bras- sicaceae)—Evidence from ITS-DNA sequences. Bot. J. Linn. Soc. 125: 183-199. — — UN 1997b. Molecular phyloge- netics of Thlaspi s.l. ad on — plast DNA restriction site q internal transcribed spacers of nuclear hanima] A. Canad. J. Bot. 75: 469—482. LG 19 ). (Brassicaceae) base variation and sequences of * X Note 's on the infraspecific taxonomy and le af — of Draba ramosissima Desv. Castanea 34: 210— O'Kane, s. s в. Schaal & 1. А. Al-Shehbaz. The origin of ен suecica (Brassicaceae dicated by nuclear rDNA sequences. Syst. Bot 1996. as in- Parks, ( & J. F. Wendel. 1993. Molecular divergence between Asian and North American Liriodendron (Mag- noliaceae) with — ations for interpretation of fossil flora. Amer. J. Bot. 77: 1243-156. dd Parrish, J. T 1987. G lobal pn and prse 0- climate of the late Cre E us and Early Tertiary. Pp. 51-73 in E. M. Friis, JC — & P. R. Crane (editors), The Origins of е and their Biolog- ical Consequences. Cambridge Univ. Press, Cambridge. 19 Payson, E. B. 1926. The genus > Thlaspi i in North America. Iniv. . Wyoming Publ. Bol. |: Petit, C., F. Bretagnolle € E dd 1995 Evolutionary consequences of E hybrid zones in wild spec ies TI E E Price, R. A. : Systematics and Geography. — of Wisconsin, li . 1980. Draba streptobrachia (Brassicaceae), a new species pA Colorado. : — 32: 160-169 ——. | Palmer & Al-Shehbaz. 1994. System- atic ba cae ra of Arabidopsis : — ular and mor- Meyerowitz Cold Spring om — a crassa Complex (Brassica- M.Sc. thesis, Uni- adison. — al perspectiv 7-19 & C. R. Somerville editors), Arabidopsis. oe bor Laboratory | Ramsey, J. & D. W. $ Уе ый 1998; Pathways, mecha- nisms, and rates > aie formation in flowering lants. Ann. iun ‘ol. Syst. 29: 467-501. Rie — L. к. Doe. 1999, Hybrid classifi- ation: — ге genetic — based studies of ex- perimental hybrids. Ecology 80: 361—370. — ———, J. Whitton € — 1999. Hybrid zones and the genetic architecture of a barrier to gene flow )etween A — r species. Genetics 152: 713-727. Rollins. R. . A new species of the Asiatic — ae ia a ruc ice ») from North America and i bioge — implications. Syst. Bot. 7: 214—220. ——— 1983. Interspecific hybridization and taxon uni- Amer. J. Bot. 70: 625- — formity in кї (Cruciferae . 1984. Draba (Cruc — in Mexico and Guate- mala. — Gray Herb. 213: 1-10. ————. 1988. — — new genus of the Cruciferae. J. Jap. Bor 63: € — The [^ ruc ae rae of Continental North Amer- ica. Sono Univ. Press, Stanfore . Rüdenberg. 1971. С hromosome numbers of Cruc ш . II. Contr. Gray Неф: 17-133. — Lauran M. L., V. Rull, S hubert & S. Va- lastro. 1988. The establishment r vegetation after late Pleistocene de ‘glaciation in the paramo de miranda Ve- — Andes. Rev. Palaeobot. Palynol. 55: 5-18 Schulz, ‚ 1927. Cruciferae—Draba et — In: A E SON [ died, Pflanzenr. IV. 105 (Heft 89): 1-396. Verlag von Wilhelm ке Leipzig. ‚ Engle a K. Prantl (ed- . 17В: 227-658. Verlag eipzig. охоте numbers v — өз vo Smith, E. B. 1 Chrom ering plants. I. Trans. Kansas Acad. Sci. Soltis, D. E. & P. S. Soltis. 1993. Molec es des и те dynamic nature * polyploidy. €. R. C. Crit. Rev. Pl. Sci. 3-27: é — t Recurrent formation and genome evolution TREE 14: 348—352. Swofford, D. L. 9. PAUP*. Phylogenetic Analysis Us- ing Parsimony — Other Methods). Version 4.0b2. Sinauer, Sunderland, Massachusetts Taberlet, ielly, G. Pautou & J. — 1991. Uni- versal primers "for amplification of three — re- pm of chloroplast DNA. Pl. Molec. Biol. 10 1109. Thompson, | R. Lumaret. 1992. The evolutionary dynamics of pied — oe establishment EE 7: 3 nd persistence. us R )2—3( Tolmac hev, 939. Draba. In: d a Komarov & N. A. Busch itor ш URSS 8: d 649, 650 Urbanska, . Hurka, Ё ‚В. Ne uffe r & K. alo. or гин and evolution in Cardamine (Brassicaceae) at Urnerboden, Central Swit- Volume 89, Number 1 2002 Koch & Al-Shehbaz American Draba 109 zerland: Biosystematic and molecular evidence. Pl. Syst. . 204: 233-256. Vissemann, V. 2000. Molekulargenetische und morpholo- gisch-anatomische Untersuc hungen zur E volutio: Genomzusammensetzung von Caninae (Dc.) Ser. Bot. Jahrb. Syst. 12 n und ¡ES N t | Walters, 1964. Draba and — = G. Tutin el xu — Fl. Europaea 1: Ward. 03. Chromosome counts fom New Mexico and ec rn Colorado. Phytologia 54: —30 & R. Spelle — p 88. Chromosome counts of anglospe m е New Mexico and adjacent areas. Phy- hup 64: —:398. Weber, A. po New names and Fog e eM in the Rocky Mountain flora 3 prin- Phytologia 51: аљ" 7 ‚ Additions to the flora of Colorado ХИ. Phy- REN өт. d 20— We nisl. J. F к & T. Seelanan. 1995a. Bidi- rectional inte nine ‘us concerted evolution following allo- ploid spec iation in colton, онур) 84. Acad. Sei. U.S.A. 92: 280-2 Proc. Natl. — & ——. 1995b. DN. i sequence from gossypioides reveals an- Molec. Phy- An unusual ribosomal cient, cryptic : inlerge nomic introgression, logenet. Evol. 4: 298-313. Vhite. T. J.. T. Burns, S. Lee & J. Taylor. 1990. Ampli- fication and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315-322 in M. A. Innis, ) кы & T. J. White (editors). Academic Press, New Widmer, A. & M. Ba 1990. idence for allopolyploid speciation and a single origin of the narrow endemic Draba ladina (Brassicaceae). : 1282-1289. Baltisberger. 1999}. cpDNA variation in alpine А ss aceae): — re oe and population structure. Molec. Ecol. 1415 Zunk. K. K. ba 7 P ni Шогы 1999, netic relationships in tribe Lepidieae (Brassicaceae) ^ Yor iltisberger. ITS molecula ev- Extensive intra- Draba aizoides | Phyloge- based on chloroplast DNA restriction site variation. Canad. J. Bot. 77: 1504-1512. EVIDENCE FOR MOTH AND Peter Goldblatt? and John C. Manning? BUTTERFLY POLLINATION IN GLADIOLUS (IRIDACEAE- CROCOIDEAE)' ABSTRACT Pollination strategies of Gladiolus, one of the largest genera of the monocot family Iridaceae, are unusually diverse and include various bee species, foraging either for nectar or for pollen, passerine birds, long-proboscid flies a for nectar, hopliine beetles that use the flowers primarily as sites fo r assembly, and Lepidoptera. Pollination by insects f this order comprises two cse different sets of pollinators, night-flying moths (Noctuidae or Sphingidae) and butterflies (evidently se one species of Satyridae). These lepidopteran-pollinated flowers have quite different floral adaptations and both types are specialist systems, although moth flowers may be pollinated by a range of different moths. In Gladiolus moth- -pollinated flowers are usually large, long- ا‎ and pale- — or mottled dull — to brown, are usually richly scented, often open fully only at night, and produce relatively concentrated nectar that is sucrose-rich. Butte ily flowers, in contrast, are fully open during the day, close ДА ог ере at night, are often bright crimson to scarlet, usually with prominent white splashes on the lower tepals, but are also large, have a long tube, and produce quantities of relatively dilute nectar, either sucrose-rich or hexose-rich. Comparing the polli- nation systems of related species, we infer that night-flying moth pollination arose 6 times in the genus, whereas butterfly pollination arose 3 times in the 165 species of southern Africa, and that the two lepidopteran pollination systems in Gladiolus arose quite independently of one another. Key words: butterflies, floral ecology, Gladiolus, Iridaceae, moths, nectar, pollination systems. = Pollination in the approximately 35 genera of the (Goldblatt et al., 1998a). We have also presented Iridaceae of sub-Saharan Africa, where over 1000 evidence for passerine bird pollination in Gladiolus species of the family occur, is remarkably diverse. and reviewed pollination systems for the entire ge- Across the continent pollination systems include nus (Goldblatt et al., 1999, 2001). By comparing four orders of insects (Coleoptera, Diptera, Hyme- and contrasting the evidence for moth versus but- noptera, Lepidoptera) and passerine birds, as well — terfly pollination in Gladiolus, we show just how as facultative, and possibly obligate autogamy. Pol- different the two systems are that use lepidopteran lination by insects of the orders Diptera, Hymenop- pollen vectors. tera, and Lepidoptera is itself diverse, involving a Pollination by large butterflies, so far in Gladi- range of families of each order and different sets of olus known to involve only the satyrid, Aeropetes floral adaptations and associated rewards to polli- tulbaghia, has already been demonstrated for two nators. Here we present evidence for two different species of Gladiolus and was inferred for several pollination strategies utilizing Lepidoptera in the more because of the distinctive floral features pre- large, predominantly African genus Gladiolus. sent in species pollinated by this butterfly (Johnson Elsewhere we have documented passive pollination & Bond, 1994). These features include a large, un- in the genus by large-bodied apid-anthophorine scented flower, bright red perianth often with white bees (Goldblatt et al., 1998a), hopliine beetles splashes on the lower tepals, a narrow but fairly (Goldblatt et al., 1998b), long-proboscid flies in the long perianth tube, and the presence of large quan- families Nemestrinidae and Tabanidae with elon- tities of nectar (Johnson & Bond, 1994; Goldblatt gate mouth parts exceeding 15 mm (Goldblatt € & Manning, 1998). Night-flying moth pollination Manning, 1999, 2000), and an active pollination has been inferred for nine species of Gladiolus system involving pollen-collecting female bees based on the presence of features commonly asso- ! Support for this study by grants 5408-95 and 5994-97 from the National Geographic Society is е, ас — edged. We thank B.-E. van Wyk, Rand Afrikaans University, Johannesburg, for the nectar analyses: . Kaiser, Givaudar Roure Research Ltd., Switze vl ind, for scent chemistry analy ses; and Colin Paterson-Jones for the uin of a moth — — maculatus 3. rukoff Curator af African Botany, Missouri Botanical Garden, Р.О. Box 299, St. Louis, Missouri 63166, U.S.A. ба йн tet org *Compton Herbarium, National Botanical Institute, P. Bag X7, Claremont 7 ac.za ANN. MISSOURI Bor. GARD. 89: 110—124. 2002. TT: 5, South Africa. manning(e?nbict.nbi. Volume 89, Number 1 2002 Goldblatt & Manni ing 111 Gladiolus Lepidoptera Pollination ciated with moth pollination, including a pale-col- ored perianth, presence of fairly concentrated nec- tar, and a strong, often distinctive scent with a clove component, and in some species floral odor pro- duced only at night (Goldblatt € Manning. 1998). Field observations and investigation of floral fea- tures allow us to expand our understanding of these pollination systems. Moreover, by using information about phylogenetic relationships in the genus and comparing the pollination systems of related spe- cies, we can infer the number of times that each system evolved within the genus and suggest an- cestral species, thus developing hypotheses about the evolution of the floral traits associated with pol- lination by different Lepidoptera. MATERIALS AND METHODS SPECIES EXAMINED Observations on the floral and pollination biology of Gladiolus were made during the years 1993 to 1999 in the field in southern Africa and in living collections at the Missouri Botanical Garden, St. Louis. and Kirstenbosch Botanic Gardens, Cape Town. Together with our research on the systemat- ies of the genus (Goldblatt & Manning. 1998; Man- ning et al., 1999), we have identified 20 species that have flowers likely to be pollinated by Lepi- doptera from a total of 165 species of Gladiolus that occur south of the Limpopo-Cunene River axis. At least an additional 15 species of tropical Africa (Goldblatt, 1996) have similar flowers and may be assumed to be pollinated by Lepidoptera. Southern African. species fall in five of the seven sections recognized in the genus by Goldblatt and Manning (1998: Blandus, nearifolius, and Ophiolyza). Species in two addi- sects. Hebea, Homoglossum, Li- tional sections of the genus, sections Acidanthera and Decoratus, in tropical Africa also exhibit fea- tures consistent with pollination by Lepidoptera. Candidates for Lepidopteran pollination were recognized by two separate suites of characters. The first is the classic syndrome for moth pollination (Faegri & van der Pijl, 1979) in which pale flower color and production of a strong. rich odor are as- sociated with some features only exhibited at night. In Gladiolus such flowers are usually large. have an elongate floral tube, shortly exserted unilateral stamens, and the strong floral odor is often pro- duced or intensified at night. Pigmentation may be cream or whitish, but several species have darkly mottled to nearly uniformly dark brown perianths. A second syndrome of characters, described in de- tail by Johnson and Bond (1994), combines bright red floral pigmentation, a large flag-like presenta- tion, absence of floral odor, and often lower tepals marked with white splashes (we use the term tepal for perianth lobe here). SEASONALITY, AND FLORAL FLORAL PHENOLOGY, PRESENTATION LONGEVITY, Direct phenological observations are presented on 18 southern African species of Gladiolus made during the years 1993 to 1998 in the field (Table ) and in living collections at Kirstenbosch Botanic — Gardens. Cape Town. Observations include mode and timing of anthesis (i.e., opening of individual buds), anther dehiscence, expansion of stigmatic lobes. followed by withering of the perianth. Data on seasonality are taken from Goldblatt and Man- ning (1998). Compatibility relationships were not examined for the study. Plant vouchers (Table are deposited at the Missouri Botanical Garden Herbarium, St. Louis (MO), and/or the Compton Cape Town (NBG). Herbarium. NECTAR ANALYSIS Nectar volume measurements were taken from unbagged flowers in the field, reflecting both rates of secretion and depletion, and from plants main- tained in the laboratory, and not visited by insects — effectively representing bagged flowers). Experi- ence with Iridaceae has shown that nectar charac- teristics gradually change in species retained 1 water for periods greater than 24 hours, the nectar usually becoming more diluted (Goldblatt et al., 1995). Nectar from flowers on cut stems maintained in water in the laboratory was therefore sampled 18 hours. were picked and nectar was withdrawn from the within To collect. nectar whole flowers base of the perianth tube with 3 wl capillary tubes after separating the ovary from the perianth base (perianth tubes are too elongate and curved to allow nectar removal directly via the mouth of the tube with a capillary tube). Detaching the perianth from the top of the ovary causes minimal damage. and contamination of nectar by fluid from broken tissue is not significant given the large volumes of nectar produced by flowers of species under investigation. The percentage of sucrose equivalents in fresh nec- laboratory on a tar was measured in the field or Bellingham and Stanley hand-held refractometer (0-50%) from five or more individuals per popu- lation, unless fewer individuals were available. Ad- ditional nectar samples were dried on Whatman fil- van Wyk, Afrikaans University, Johannesburg. ter paper no. Rand HPLC sugar analysis. 1 and analyzed by B.-E. using 112 Annals of the Missouri Botanical Garden Table | (other collectors). All study sites are in South Africa. Study sites and voucher information for species studied. Vouchers are housed at MO (Goldbatt) or at NBG fri Species Study site and voucher GLADIOLUS SECTION BLANDUS series Blandus G. cardinalis Curt. H. Wright G. insolens Goldblatt & J. С. Manning : — uae Oberm. G. carmineus C. empervir 7. J. Lewis PRIUS SEC ‘TION LINEARIFOLIUS series Linearifolius G. emiliae L. Bolus W Cape, near Swellendam, Mar., W Cape, near Bain’s Kloof, Jan., Goldblatt & — s.n. no voucher W Cape, cliffs at Hermanus, Feb., 292 W Cape, Piketberg, Jar W Cape, Potberg. Mar “Goldblatt 4 Manning 10176 E Cape, Tsitsikamma Mts. Goldblatt 1 tt & poe 10166 ‚ May. Manning s.n. no — Manning 1057 W Cape, near Riviersonerend, Mar., Stayner s.n. G. guthriei F. Bolus G. J. G. nerineoides Lewis GLADIOLUS SECTION HEBEA GLADIOLUS SECTION HOMOGLOSSUM s Gra С. ee. Coldbisi & J. €. M G. maculatus Sweet anning W Cape, Bain's Kloof, W Cape, Helderberg Reserve, Somerset West, Jan., Runnals 463 W Cape, Jonkershoek Mts., Jan., Mpumalanga, near Morgenzon, Oct., May, Manning 1080 Esterhuysen 32847 Goldblatt & Manning 10071 E Cape, near Grahamstown, Mar., Dold & Weeks s.n. W Cape, Simonstown, June, Manning s.n. no voucher Devil's Peak, May, Manning s.n. no voucher G. recurvus L. series Tristis G. hyalinus Jacq. W Cape. G. liliaceus Houtt. W Cape, С. longicollis Baker ». tristis GLADIOLUS SECTION OPHIOLYZA sitiflor T. Мыш G. saundersii J. D. Hook. Fairfield, Caledon, Sep., — Long Tom Pass, Feb., — Mad 9822 Cape, near Bredasdorp, Aug., Barker 284. W Cape. Helderberg Reserve, Nov., Runnals 513 W Cape, Gydo Pass, Goldblatt E la 974ЗА Lion’s Hea , Barker — s.n. no vouche KwaZulu-Natal, Hillcrest, Jan., Goldblatt & Manning 9854 E Cape, Naude's Nek, Feb., Goldblatt & Manning 9550 Additional species with similar flowers, assumed to be adapted for — by Aeropetes or night-flying moths, i G. J. Le include G. acuminatus F. Bolus (putatively moths) and js — field observations of the floral ecology of these species. but we have no insect visitors for G. albens, G. is (putatively butterflies), also lack ask uin of ir carmineus, G. cruentus, б. hyalinus, С. insolens, and б. E я but those species are included above because we => iave nectar and other data from study populations. FRAGRANCE Floral scent was noted with the human nose in the field and in cultivated plants. Presence of scents too weak to be discerned in the open air was recorded after individual flowers were picked and placed in clean, lidded glass jars and stored in a warm place. The contents of each jar was sniffed after a minimum of 60 minutes (Buchmann, 1983). The site of scent production was examined by im- mersing flowers in aqueous neutral red stain. Scent chemistry was examined by R. Kaiser, Givaudan- Roure Research Ltd., Switzerland, by gas chroma- tography using a DB-Wax Capillary column (Kaiser, )3). Scents were captured in capsules through — which air was drawn by a vacuum pump from a small, lidded chamber containing open flowers. POLLINATION MECHANISMS AND POLLEN LOAD ANALYSES Observations of insects on Gladiolus flowers in- volved at least 5 hours total per species and in some cases up to 20 hours total per species. For species inferred to be pollinated by moths, an ad- ditional 4 hours of observation was made per spe- cies during daylight hours to determine whether di- urnal might also occur. Observations included mode of foraging and wheth- er insects contacted anthers and stigmas during vis- visits by insects Volume 89, Number 1 002 Goldblatt & Mannin 113 9 Gladiolus Lepidoptera Pollination its to flowers. Study populations always included at least 20 individuals at evidently undisturbed field sites. Insects observed to probe the floral tube or to brush the anthers or stigmas were netted when possible and, in the case of moths, then immobi- lized in a jar using ethyl acetate fumes. To prevent contamination of the body of an insect with pollen carried by another in the same jar, the bodies of insect specimens were isolated from each other by wrapping them in tissue prior to pinning. Individual butterflies were netted for measuring and observa- tion of sites of pollen deposition and then released. Body length and proboscis length of insects was Body length was measured from the base of the labrum to the recorded from captured specimens. tip of the abdomen. Mouth-part length was mea- sured from the base of the labrum to the tip of the proboscis. Night-flying moths are not easily cap- tured simply because darkness makes them diffi- cult to locate. Use of flashlights covered with trans- lucent. red. cellophane paper for illumination = significantly assisted observation and netting of moths. Captured moths were identified by Douglas Kroon (Sasolburg, South Africa), and both moth and butterfly voucher specimens were deposited with the South African Museum, Cape Town. Identification of pollen on insect bodies was done by gently removing grains from the body surface with a dissecting needle. The residue from needle probes was collected on glass slides and mounted in 1-2 drops of Calberla’s fluid (Ogden et a 1974). Pollen grains were identified microscopical- ly by comparison with reference to pollen grain preparations made from plants flowering at study sites. Gladiolus pollen grains are recognized. by their large size, monosulcate aperture with promi- nent 2-banded operculum. and perforate-scabrate exine (Goldblatt et al.. 1991). RESULTS SEASONALITY, FLORAL PHENOLOGY AND LONGEVITY, AND FLORAL PRESENTATION (TABLE 2) Flowering times in the Gladiolus species of southern Africa that are pollinated by night-flying moths show no obvious association with their geo- graphic ranges in the summer- or winter-rainfall zones of southern Africa (Goldblatt & Manning, 998). Flowering of the flora peaks in late spring (September to November) in the winter-rainfall zone ut in summer and early autumn (December to March) in the summer-rainfall zone. This coincides with the middle or end of the period of optimal plant growth, during or soon after the main rainy periods. In the summer-rainfall zone. G. longicollis flowers early in the season, mostly October and No- vember, sometimes producing a second flowering flush in December-January. The only other species of the region showing adaptation for moth pollina- tion blooms in spring before the first rainfall of the wet season. This is out of phase with the main flow- ering period for Gladiolus in the summer-rainfall zone (Goldblatt & Manning, 1998 Gladiolus hyalinus. — In the winter-rainfall zone, G. liliaceus, and G. tristis follow the main flowering pattern for the region, but G. albens and G. macu- latus flower early in the season, in autumn or win- ter, and G. recurvus flowers in late winter or early spring. Flowering out of phase with the flowering peak is also characteristic of the two moth polli- nated species of section Linearifolius. G. emiliae and G. guthriet, which flower mainly in April and May, although vegetative growth in these two spe- cies is delayed until the winter and. spring when leaves are produced. Flowering of butterfly-pollinated species of Glad- iolus is always from late December to April. excep- tionally May, and this is when the only butterfly so far recorded on Gladiolus species, Aeropetes, is on the wing (Johnson & Bond, 1994). This means that for the butterfly-pollinated species of the winter- rainfall zone, flowering occurs several months out of phase with the main growth and flowering peak for the region. As a consequence these species must also have special ecological adaptations to support the unusual flowering patterns. These may be a specialized habitat, such as cliffs. stream banks, or waterfalls (G. cardinalis. G. insolens. G. sempervirens), or a growth pattern in which leaf pro- duction occurs when conditions are suitable, later in the season (G. carmineus, G. stefaniae), or both (С. nerineoides, G. stokoet). Population density appears to be moderately dif- fuse, and plants form extended populations with flowering individuals standing 1-3 m apart. Species of specialized habitats such as Gladiolus tristis (moist to marshy sites). G. cardinalis (waterfalls), G. carmineus (coastal sandstone cliffs), G. sempervirens (seeps and wet forest margins) may be locally com- mon and grow in dense stands. The pattern of flower buds opening on an inflo- rescence is acropetal. In all species, a mature bud expands in the early to mid morning. and the open flower typically lasts four days in Gladiolus species (Goldblatt et al., 1998a; Goldblatt & Manning. 1999). In moth-pollinated Gladiolus species flow- ering lasts longer. Flowers of a cultivated sample of С. tristis lasted five or six days. and flowers of G. recurvus lasted five to nine days. Flowers usually open one to two days apart, and hence there are Annals of the Missouri Botanical Garden Table 2. Floral and phenological data for southern African Gladiolus species with flowers adapted for pollination by Lepidoptera. Species are arranged taxonomically according to Goldblatt and Manning (1998). Perianth tube length was recorded at study sites and may not represent the range for out of phase with the peak flowering time for the rainfall zone. the species. An asterisk (*) indicates species flowering Flower Perianth Main flowering Rainfall Species color tube (mm) Scent lime zone GLADIOLUS SECTION BLANDUS series Blandus G. cardinalis red, white streaks 32-40 попе late Dec.-Feb. — winter* n lower tepals G. carmineus crimson-pink, white 30-35 none Feb.—Mar. winter* streaks on lower tepals G. insolens scarlet-red ca. 38 none Jan.—Feb. winter* G. stefaniae red, white streaks 35-45 none Mar.—Apr. winter* on lower tepals G. sempervirens red, white splashes 25-42 none Mar.-May winter* on lower tepals GLADIOLUS SECTION LINEARIFOLIUS ries Linearifolius А emiliae densely brown speckled 32-45 strong fruity Mar.- Apr. winter* >. guthriei dull purple-brown 20-27 sweet-clove Mar.—June winter* G. nerineoides scarlet 25-31 none Jan.—Mar. winter* G. stokoei rmine-red 30-35 none Mar. winter* GLADIOLUS SEC TION HEBEA series Permeabilis G. acuminatus cream 16-22 sweet-floral Aug.—Sep. winter G. robertso whit 28-44 sweet-clove Sep.-Oct. summer* GLADIOL Us SECT ION HOMOGLOSSUM series Gracilis G. albens white-cream 45-60 acrid-metallic Mar.-May winter* G. maculatus cream, heavily speckled 23-35 sweet-floral May-July winter* brown to dull purple G. recurvus cream to pale pink 27-36 sweet-clove July-Sep. winter series Tristis G. hyalinus cream speckled brown, 25-26 usually odorless Sep.-Nov. winter purple or green weet-clove G. liliaceus beige. buff or pale orange. 40—45 strong sweet- Aug.—Nov. winter light mauve at night clove at night ". longic 'ollis subsp. cream, lightly speckled 100—110 strong sweet- Sep.-Nov. summer* p> collis clove at night — white to cream 85-110 strong sweet- Oct.—Dec. summer* clove at night G. tristis white to cream 40-60 strong sweet Sep.—Nov. winter clove at night pred US SECTION OPHIOLYZA entus red, white on lower tepals ca. none Jan.-Feb. summer A — red, white on lower tepals ~ none Feb.—Mar. summer often two or more flowers open at any time on an inflorescence of three or more flowers. Flowers of moth-pollinated species are usually partly closed or at least the tepals are flaccid during the day, and The reverse pattern holds for butterfly-pollinated species. At sunset, the tepals of most species partly or fully close, the tepals then loosely enclosing the exserted anthers and stigmas. During the day the open fully at sunset when the tepals become firm tepals become fully expanded again. Flowers of many Gladiolus species have been found to exhibit mechanical protandry (Scott Elliot, 1891; Goldblatt et al., 1998a; Goldblatt & Man- ning, 1999), and the species studied here conform and fully extended. In species of Gladiolus series Tristis of section Homoglossum scent is released at the same time, although a faint odor may be de- tected during the day. Volume 89, Number 1 2002 Goldblatt & Mannin 115 g Gladiolus Lepidoptera Pollination to this pattern. The anthers dehisce longitudinally one to four hours after the tepals first unfold. This depends to some extent on ambient temperature and humidity, and anthers dehisce later in wet, cool conditions. Pollen grains are clumped together and pollen remains in the anther thecae until removed by an insect. The three style branches, the distal adaxial surfaces of which comprise the stigmas, are loosely held together for the first three (to five) days that the flower is open and lie laxly over on the dorsal surface of the anthers. On the last day of anthesis, the style elongates and the style branches diverge, arching outward beyond the anthers. At the same time, the conduplicate margins of the distal half of each style branch unfold, moist, sticky stigmatic surfaces of the now spath- exposing the ulate style branches. Only then are the stigmas of a flower accessible to pollen deposition, and pollen adheres to these areas following hand-pollination. Thus of the four to six days that a flower is open, it typically has three to five days in an exclusively male phase during which time pollen is usually re- moved from the anthers by insects. Anthers can be seen with the naked eye to lack pollen after three or four days if flowers were actively visited. By the time the stigma lobes unfold the flower is then in an exclusively female phase that lasts for the final one (or two) days that a flower is open. Mechanical self-pollination cannot readily occur, even if pollen remains in the anthers by the time the receptive stigmatic areas are exposed because of the spatial separation of the pollen-bearing anthers and the stigmatic surfaces. Gladiolus carmineus, which we include here as possibly pollinated by large but- terflies although we lack pollinator observations, 15 an exception. The style divides opposite the base of the anthers and the style branches are tangled in the dehisced anthers; thus, selfing could easily occur if there were no incompatibility system (un- known at present). M the G. carmineus study site pollen had not been removed from the anthers, and the stigmatic surfaces become dusted with pollen from anthers of the same flower. Species of Gladiolus are medium-sized, corm- bearing geophytes, typically 45-120 cm high (Fig. LA—F). produce a single, unbranched flowering stem an- Species pollinated by Lepidoptera typically nuallv. Flowering population is synchronous and lasts two to four weeks, and inflorescences are typically secund spikes with the flowers facing to the side and with the floral tube in an ascending position. In G. nerineoides the flowers are crowded at the apex of the flowering stem in more or less spiral arrangement and it is the entire inflores- cence, rather than individual flowers, that makes for a conspicuous display (Fig. 1B). Flowers pollinated by moths are moderate in size to large, depending on sectional affinity (Fig. 1C— = `). In species of Gladiolus sect. Homoglossum the perianth tube is mostly 23—60 mm long. but excep- tionally 85-110 mm in G. longicollis subsp. platyp- etalus. The dorsal tepal, usually slightly larger than the other five tepals, is 32—45 mm long, thus short- er to about as long as the tube (Table 2). In Glad- iolus sects. Hebea and Linearifolius flowers are somewhat smaller, and G. acuminatus has a floral tube 16-22 mm long and tepals 15-21 mm long. Irrespective of sectional affinity, the perianth tube is more or less cylindric but slightly wider toward the apex, with the slender lower portion 1.5-2 mm in diameter, and fairly straight or gently curved. Flowers are zygomorphic and the slightly larger dorsal tepal is typically inclined while the upper lateral tepals are spread outward. The lower tepals. usually slightly smaller than the upper three, are The style and stamens are unilateral and arch to lie held loosely together and directed forward. close to and just beneath the dorsal tepal and are thus contained within the enclosed space formed by the ascending tepals. The filaments are shortly exserled from the tube, or in G. longicollis not or barely exserted, and the anthers lie parallel to one another with the lines of dehiscence facing toward the. center of the flower and the lower tepals. In many moth-pollinated Gladiolus flowers the tepals are somewhat attenuate with the tips recurved, a feature most exaggerated in G. acuminatus and б. recurvus. Moth-pollinated flowers are either shades of white to cream or are lightly to densely mottled with dull purple to brown (Table 2: Fig. IC—F). Each of the three lower tepals may have a weakly contrast- ing. darker median band, often collectively referred lo as a nectar guide (Goldblatt & Manning, 1998). The bases of the tepals and the distal part of the tube together form a wide throat leading to the nar- row, proximal part of the tube. In the sense of Fae- eri and van der Pijl (1979), these are gullet flowers but often with a particularly elongated floral tube. A remarkable feature of Gladiolus liliaceus is the color shift that occurs at sunset. As the tepals be- come more fully extended the normally beige. light brown or rusty colored tepals take on a yellowish background hue with the brown pigment changing to light blue-mauve. First reported by Henry An- drews in 1798, this color change has intrigued bot- anists ever since and is thought to be a direct ad- aptation to moth pollination (Goldblatt & Manning. 1998). A less pronounced color change al sunset 116 Annals of the Missouri Botanical Garden — Figure 1. Comparison of the flowers of southern African Gladiolus pollinated by the satyrid butterfly Aeropetes tulbaghia (A, B) versus night-flying moths mostly of the families Noctuidae and Sphingidae (C-F), with longitudinal ections of flowers of some species. —A. G. sempervirens (sect. Blandus). —B. G. nerineoides (sect. Linearifolius). — С. G. emiliae (sect. Linearifolius). —D. G. maculatus (sect. Homoglossum). —E. G. liliaceus (sect. Homoglossum). —F. G. guthriei (sect. Linearifolius). Scale bar 10 mm. Volume 89, Number 1 2002 Goldblatt & Manning 117 Gladiolus Lepidoptera Pollination also occurs in the related G. hyalinus, the tepals of which become paler and more translucent as day- light fades. The darkly mottled tepal coloration may be an adaptation for camouflage. rendering the flowers less visible to nectar or pollen thieves dur- ing the day. Johnson (1995) has suggested that the maroon pigmentation of the moth-pollinated flowers of the orchid Monadenia ophrydea Lindl. likewise represents camouflage. Moths are believed to locate flowers of this species solely by their scent. Flowers pollinated by the Aeropetes butterfly are fairly large (Table 2), with the exception of Gladi- olus nerineoides, and have a perianth tube mainly 30—45 slightly larger than the other five tepals, is mostly mm long. and the dorsal tepal. usually to about 35-65 mm long, thus somewhat longer In G. somewhat smaller flowers have a floral tube 25—31 twice as long as the tube. nerinevides the mm long and tepals 19-22 mm long, thus shorter than the tube. The perianth tube is more or less cylindric but slightly wider toward the apex. with the slender lower portion 1.5-2 mm in diameter. and fairly straight or gently curved. Flowers are zy- gomorphic (barely so in G. nerineoides) and the larger dorsal tepal either more or less erect, or lightly inclined (G. cardinalis). while the upper lat- eral tepals spread outward. The lower tepals are usually slightly smaller than the upper three and The style and stamens are unilateral and extend out- are held loosely together and directed forward. ward. and well exserted from the tube. Exceptional G. nerineoides has subequal tepals spreading more or less at right angles to the tube and the filaments are included, while the anthers may be partly ex- serted or entirely included. Butterfly-pollinated flowers are shades of red. oi exceptionally deep pink (Gladiolus carmineus). and the lower tepals are often streaked with white. In G. saundersii the lower tepals may be almost en- tirely white with irregular reddish speckling (Table 2). The flowers of all the butterfly-pollinated spe- cies lack detectable odor (Table 2). In the sense of Faegri and van der Pijl (1979), these are flag flow- ers but with a particularly elongated floral tube. G. nerineoides, and б. The flowers of G. insolens. stokoet are uniformly colored. NECTAR ANALYSIS (TABLE 3) Nectar produced by moth-pollinated species of Gladiolus is sucrose-dominant and ranges in con- centration from a low of 19.5% (С. emiliae) to 36.4% (G. tristis) sucrose equivalents. Nectar quan- tities, measured from unbagged flowers, range from 2.2 to 12.4 pl. Nectar of butterfly flowers is markedly variable in character, and ranges from sucrose-dominant (6. cardinalis, G. stefaniae) to sucrose-rich (G. cruen- tus, G. saundersii) or hexose-rich (G. insolens. G. nerineoides) according to the defintion of Baker and Baker (1983). low concentrations, ranging from a low of 18-21% These nectars also have relatively sucrose equivalents in G. stefaniae to 26.8% in one of two populations of G. nerineoides examined. FRAGRANCE Moth-pollinated species of Gladiolus. with one exception, produce strong, and often rich sweet odors (Table 4). In species of Gladiolus series Tristis fragrance is produced at nightfall and is weak or evidently absent during the day. Fragrances vary considerably as perceived by the human nose. though often appear to have a strong clove com- ponent and thus resemble the scent of stocks (Ma- thiola) or carnations (Dianthus). The odor produced by G. albens (sect. Homoglossum) is strikingly dif- ferent. and is somewhat acrid and metallic, while the odor produced by G. emiliae (sect. Linearifolius) is fruity, with elements of coconut and pineapple. Scent production in б. hyalinus is evidently un- common. In four populations we examined (Die Gale. Gydo Pass, Lions Head, Nieuwoudtville). all in the west of its geographic range, flowers pro- duced no detectable odor, but collection notes with some herbarium records from the eastern half of its range indicate the presence of a strong, sweet scent. Absence of odor is combined with mottled, brown- ish or purplish coloration in these populations. Scent compounds mostly belong to different chemical classes of scents from those of bee-polli- flowers (Goldblatt et al., 1998a), for linalool, which is present in some bee-pollinated nated excepting species, notably G. alatus, while benzyl acetate is also present in G. jonquiliodorus. Scent chemical profiles differ considerably among moth-pollinated species, even though linalool is a common compo- nent (Table 4) and is the predominant compound in G. maculatus. As in bee-pollinated species of the genus. numerous compounds combine to produce the characteristic scent of each species, and many as 39 compounds were identified in G. re- currus (R. Kaiser. pers. comm.). Surprisingly, G. ac- uminatus. the only species of Gladiolus sect. Hebea examined, shares no compound with the moth flow- ers of species of section Homoglossum to which be- long the other species analyzed. The spicy-clove type scents of G. liliaceus and G. tristis appear to derive from different compounds, eugenol in the former and eucalyptol in the latter. 118 Annals of the Missouri Botanical Garden Table 3. analyses were provided by B.-E. Wyk, Ra Available nectar characteristics of species of southern African Gladiolus pollinated by Lepidoptera. Nectar nd Afrikaans University, Johannesburg, South Africa. Number of samples (n) is the same for volume and concentration columns. Data marked with an asterisk (*) are from Johnson and Bond (1994 ў % Nectar % Range of sugars Sugar ratio Gladiolus volume pl concentration S/F + G i (n) +SD) Fructose Glucose Sucrose (n) GLADIOLUS SECTION BLANDUS series Blandus G. cardinalis *9.4 (12) *24.8 (n/a) 9 16 75 3.0 (1)* G. carmineus 1.74.8 (10) 26.1 (3.6) 18 30 52 1.08 (1)* G. insolens 4.8-7.6 (3) — 39 38 23 0.26 (1) G. stefaniae 8.6-11.0 (2) 18-21 10-16 18-25 59-72 1.89 (2) *10 12 78 3.5 (1) GLADIOLUS SECTION HEBEA series Permeabilis G. robertsoniae (2) — — — — — GLADIOLUS SECTION иш series Linearifolius G. emiliae 3.5-6.6 (3) 22.3 (1.5) — — — 5.1-6.4 (2) 19.5-20.5 8-15 5-9 76-87 4.4 (2) G. guthriei 4.44.8 (5) 31.4 (3.0) — EM — G. nerineoides Helderberg 1.84.2 (6) 26.8 (2.3) 12-26 23-36 37-63 1.85 (3) ershoe 2.9-4.6 (3) 23.7 (2.6) 8 46 26 0.35 (1) GLADIOLUS SECTION HOMOGLOSSUM series Gracilis G. albens 3.8-5.6 (4) 27.0 (2.2) — — — — G. maculatus 4.8-6.0 (2) 28-30 — =e == r TECUTVUS 4.1-13.7 (6) 31.9 (2.0) 6-7 11-14 80-82 4.3 (2) series Tristis G. hyalinus Gydo Pass 2.2-4.7 (3) 35.8 (4.8) 2-3 7-8 90 9.0 (2) Lion’s Head — — 0-4 2-7 89-98 13.8 (3) G. liliaceus 3.5-6.4 (5) 35.2 (1.2) 1 2-3 36—97 26.0 (3) G. longicollis 2.7-3.7 (3) 24.7 (2.5) 6-27 9-23 50-85 2.2 (2) subsp. platypetalus 5.1-7.9 (3) 28.3 (1.5) 6-11 6-11 78-88 4.3 (2) . tristis 8.5-12.4 (5) 36.4 (2.1) — — — — GLADIOLUS SECTION OPHIOLYZA G. cruentus 4.8-5 : 20.7 (2.5) 26. 39 35.5 0.55 (2) G. saundersii 14.7-20.1 (5) 24. 15 2 43 0.75 (1) POLLINATION MECHANISMS AND POLLEN LOAD ANALYSES Moth pollination. Observations of pollinators on 7 species of the 11 putatively moth-pollinated southern African Gladiolus species showed all of them to be visited by night-flying moths (Table 5). Visitors included sphinx moths alone (G. longicol- lis), sphinx and noctuid moths (G. emiliae), or a range of small and larger moths including Noctui- dae and other families. No other animals were not- ed visiting these species either during the day or at night excepting for a single male Anthophora div- ersipes (Apidae: Anthophorinae) captured while at- tempting to forage for nectar on the long-tubed flowers of С. contact of both anthers and stigmatic surfaces, and recurvus. Its activities did result in Gladiolus pollen was recovered from its body. This bee should probably not be considered a regular pollinator of the species since the reward offered is not accessible to the bee, which has a tongue up to 12 mm long. while the tube of G. recurvus is at least 27 mm long. Details of moth visits are limited because of the difficulty of observing their activity in the dark or under low intensity red light. Except for species of Sphingidae, moths settled on flowers, grasping the lower tepals, before inserting their probosces into the floral tube (Fig. 2). Too few visits were noted on any species for us to make observations of the duration of visits—we were determined first to cap- ture visiting moths for identification of species and location. of sites of pollen deposition. Captured Volume 89, Number 1 2002 Goldblatt & Manning 119 Gladiolus Lepidoptera Pollination ladiolus with flowers adapted for moth pollination (R. Kaiser, pers. comm.). J 7 Scent characteristics of selected species of southern African € Table 4. Scent composition (% constituents above 2%) )- (E, E)-alpha y (E)- ocimene epoxide poxy-3, + F Benzyl cinnamic alcohol farnesene and farnesol 7-dimentyl (E)- Phenylacet- aldehyde Methyl J ocimene | 6-octadiene Eucalyptol benzoate Linolool or benzoate Eugenol \ alcohol dehyde description Species 6.2 31.0 14.1 floral G. acuminatus 4.0 6.9 lily G. maculatus 9.5 69.5 12.6 gardenia recurvus G. 10.5 3 61.5 5.8 clove G. tristis moths were found to carry Gladiolus pollen on the proboscis, but no pollen was recovered from moth bodies. After being netted and transferred to a kill- ing bottle, moths lose many of their body scales and may have also lost any pollen they might have car- ried. Nevertheless, those moths with elongate pro- bosces, including the sphingids Hippotion and Agrius and the noctuid Cucullia, must be regarded as legitimate pollinators of the species on which they were captured. Moths in the families Archiidae and Geometridae with shorter probosces, 10 mm long or less. that were also captured, cannot reach nectar in the floral tubes and must be regarded as accidental visitors, perhaps attracted by the strong floral odors. Butterfly pollination. were seen to be visited by Aeropetes tulbaghia and Five species (Table 5) another four species (Gladiolus carmineus, б. cruentus, G. insolens, and G. stokoei) have similar flowers and are inferred to share this pollination strategy. This large butterfly has a wingspan of ca. 80 mm. a large body ca. 20 mm long, and a pro- boscis 28-34 mm long (Johnson & Bond, 1994; Johnson. 1994). Aeropetes was seen fluttering around. flowers of G. cardinalis, G. nerineoides, G. saundersii. G. sempervirens, and G. stefaniae (span- ning three sections, Blandus, Linearifolius, and Ophiolyza. Table 5). sometimes settling for 25 to 130 seconds, and then to fly to other flowers of the same species. As Johnson and Bond (1994) have noted, the behavior of butterflies included inspec- tion visits when individuals flutter above flowers without settling or feeding. During such visits, the ventral part of the thorax, abdomen, and wings may brush against the well-exserted anthers or stigmatic surfaces of species like G. cardinalis, G. sempervi- rens, and G. stefaniae (sect. Blandus). Pollen and stigmatic surfaces of G. nerineoides (sect. Lineari- folius) and G. saundersii (sect. Ophiolyza) are only contacted during feeding visits when insects settle. grasping the tepals and inserting their proboscis into the perianth tube. In G. nerineoides pollen is deposited only on the upper proboscis because the anthers are not or are only shortly exserted from the floral tube. Pollen is deposited on the head. antennae, and dorsal part of the thorax of G. saun- dersii because the hooded dorsal tepal, under which the anthers lie, forces a feeding insect to orient its body nearly vertically as it grasps the upper lateral or lower tepals. No other animals were ever seen visiting those Gladiolus species that attracted Aero- petes, and this single insect must be assumed to be their sole pollinator. The innate attraction that Aeropetes exhibits for red flowers, particularly those of large size, or col- 120 Annals of the Missouri Botanical Garden e 5. Length of perianth tube of Gladiolus speci for the species, measured from all known herbarium collections were captured because their identity is not in question; deposition were noted because Aeropetes is a legally protected species. Geometridae: Macaria: Noctuidae: Cucullia, Syngrapha, potion. Note: the ° been A. convoluli. 'spurge hawkmoth" reported on G. longicollis by es (from Goldblatt & Manning, 1998, representing the ranges ) and mouthparts of captured moths. Few butterflies the few individuals captured were released after sites of pollen ‘amily affiliations: Archiidae: Hypagoptera; Tychopmoptes; Satyridae: Aeropetes; бй Agrius, Н! J. M. Wood (in Scott Elliot, ip- 1891) may also have Perianth epidopteran tube (mm) 1) Plant species species (r Position of pollen Mouthpart ) on insect body (mm CLADIOLUS SE CTION от saunder. Aeropetes tulbaghia GLADIC OLUS SE CTION | BLANDO S G. cardinali ' G. 3 Aeropetes tulbaghia not measured frons, dorsal thorax not captured not available G. е Фе — es tulbaghia not captured not available G. stefanic -55 — tulbaghia not captured not available GLADIOLUS — LINEARIFO LIUS G. guthrie -18 Cucullia inaequalis (1) ca. 16 distal 12 mm of proboscis G. ile s Hippotion celerio ( 30 distal 25 mm of proboscis Cucullia inaequalis (1) са. 28 distal 25 mm of proboscis Cucullia extricata (1) ca. 20 distal 15 mm of proboscis Macaria simplicilinea (1) <10 no pollen Hypagoptera sp. <10 no pollen ( G. nerineoides — les er GLADIOLUS четох каш not available ». liliacei JAS Сис us extricata (1) A. í distal 25 mm of proboscis РА longicollis $ 5-110 Agrius convolvuli (1) 110—115 25 mm from base of proboscis G. maculatus 23-35 Cucullia terensis (1) 25-30 distal 25 mm of proboscis Cucullia sp. (1) 5-20 from base of proboscis G. recurvus 27-36 Hippotion eson (1 10 distal 25 mm of proboscis z. tristis 10—60 Syngrapha circumflexa (3) 16-17 dorsal thorax and proboscis Tychomoptes inferior (1) ca. 10 no pollen lectively making a prominent display, as in G. ner- ineoides, is well known and has been documented where (Johnson & Bond, 1994; Goldblatt et al.. 1999). The wavelength range of red color exhibited by Aeropetes flowers is virtually the same as that in red-flowered species pollinated by sunbirds. There is no evidence of effective pollination of the but- terfly-pollinated species of Gladiolus by sunbirds as there is, for example, in red, tubular-flowered spe- cies of Tritoniopsis and Watsonia (Iridaceae), Kni- phofia (Asphodel Erica (Ericaceae), which may use both sunbirds and Aeropetes equally as pollinators (Johnson & Bond, 1994; Goldblatt et al., 1999). i the inclined, and somewhat flaccid stem and difficulty in reaching the perianth else aceae), and Possibly tube because of the large, widespread tepals dis- courage sunbird visits. In connection with Aeropetes pollination, the record of this butterfly on G. cruen- tus (Johnson & Bond, 1994) must be mentioned. From the locality (J. J. Vlok, pers. comm.) this is clearly for the high-altitude Drakensberg species, G. flanaganii, and probably reflects the innate at- traction to red color by Aeropetes. The very fleshy, stiff cupped tepals, and of this species (Goldblatt & suited to sunbird pollination sturdy flowering stem, hooded dorsal tepal Manning, 1998) are but seem unsuited to butterfly pollination, as does the nectar, which has a concentration of 35% su- crose equivalents (Goldblatt et al., 1999). The re- port of the Nectarinia famosa visiting G. flanaganii is consistent with the floral adaptations of the spe- cies, and Aeropetes may simply have been a casual visitor. Gladiolus cruentus is restricted to lower el- evations inland from the central KwaZulu-Natal coast (Goldblatt & Manning, 1998) where Aeropetes is rare or normally absent, and this red-flowered species may be pollinated by alternative large but- terflies, such as Papilio demodocus and Р, nereis, as has been found for lower-elevation populations of another red-flowered species of Iridaceae, Hes- perantha coccinea (Backh. & Harv.) Goldblatt & J. C. Manning (Goldblatt et al., in press). DISCUSSION Although aspects of pollination of Gladiolus spe- cies by moths and the satyrid butterfly Aeropetes Goldblatt & Manning 121 Volume 89, Number 1 2002 Gladiolus Lepidoptera Pollination Noctuid moths photographed while visiting flowers of Gladiolus. А. Cucullia extricata on G. liliaceus.— g Fi | 2. В. Cue "Ша cf. terensis on G. maculatus (this particular moth was not captured). closest relatives with different pollination systems makes it possible for us to draw reasonable conclu- sions about the evolution of both pollination sys- tems in the genus, for which a hypothetical phy- logeny already exists (Goldblatt & Manning, 1998 Moth pollination (Table 6) has evolved in three sec- have already been documented (Johnson & Bond, more general terms (Gold- 1994) or discussed 1 [sen Goldblatt et al., 2001). blatt & Manning, these two quite different pollination systems are more fully documented here. Moreover, comparison of moth- or butterfly-pollinated species with their = Analysis of the frequency and taxonomic distribution of night-flying moth and butterfly pollination systems 2 to 5 represent number of species Table in E African. species of Gladiolus. Numbers in parentheses in columns inferred to have the pollination system. Total Taxonomic tal Section pec T sections confirmed Homoglossum bed (series) and inferred Pollination tem Ophiolyza | Blandus Linearifolius Hebea Moth 0 0 2 (0) 0 (2) 5 (2) 7 З (5) 11 (7%) Large butterfly 1 (1) З (2) 1 (1) 0 0 5 З (3) 9 (5%) Annals of the Missouri Botanical Garden tions of Gladiolus, Hebea, Homoglossum, and Li- nearifolius, while butterfly pollination evolved Gladiolus sects. Blandus, Ophiolyza, and Lineari- folius. Within section Hebea the two moth-pollinat- ed species, G. acuminatus and С. robertsoniae, have such different flowers that we do not believe they are immediately related. More likely they evolved from different, bee-pollinated ancestors. Similarly, the moth-pollinated species of section Homoglos- sum belong in different series (Tables 1, 2). In ad- dition, in series Gracilis both floral and leaf mor- phological differences between G. recurvus, which has apomorphic leaves with winged margins, like albens and C. those of G. gracilis Jacq., and G. maculatus, which have leathery leaves without prominent margins or midribs, suggest that G. re- curvus is not immediately related to the latter spe- cies pair. Lastly, in Gladiolus sect. Linearifolius, the only section with both moth- and butterfly-polli- nated species, species with the two pollination sys- tems fall in different lineages (Goldblatt & Man- ning, 1998). These considerations led to the hypothesis that moth pollination arose six times within the southern African species of Gladiolus while butterfly polli- nation arose at least three times (Table 6). Com- parison of the pollination systems of immediate an- cestors of moth- or butterfly-pollinated species makes it likely that moth pollination arose from bee-pollinated ancestors in Gladiolus sects. Hebea and Linearifolius, and in series Gracilis of section Homoglossum. The sister clade of series Tristis, se- ries Homoglossum, comprises only bird-pollinated species, but we cannot say whether the predomi- nantly moth-pollinated series Tristis evolved from bird-pollinated ancestors or the putative common ancestor of the two series, series Teretifolius, which comprises mostly bee-pollinated species (Goldblatt & Manning, 1998; Goldblatt et al., 1998a). In trop- ical Africa, Gladiolus sect. Acidanthera, with seven species, also shows all of the classic adaptations for moth pollination. The perianth is white and the perianth tube 60-150 mm long depending on the species (Goldblatt, 1996), and flowers produce a strong odor that, at least for the widely cultivated G. murielae Kelway (= G. abyssinicus (Brongn. ex Lemaire) Goldblatt), is strongest at night (unpub- lished obs.). The flowers of two species of the sec- tion, б. aequinoctialis Herbert and G. murielae, are unusual for moth pollination in having dark blotch- es on the lower tepals, and the entire section Aci- danthera has an unusual specialization, anthers with | stiff, 1996). Because of the length of the perianth tube in species of section Acidanthera, mostly 90-140 acute apical appendages (Goldblatt, mm long, pollination by sphinx moths (Sphingidae), which have probosces of comparable length, is in- ferred but there are no pollination studies for the Gladiolus species of tropical Africa. One notable aspect of the evolution of moth pol- lination in the three southern African sections of the genus in which it occurs is that these sections (Gladiolus sects. Hebea, Homoglossum, and Linear- ifolius) are the only ones in which floral fragrance is well developed (and is associated primarily with bee pollination). As we have noted elsewhere (Goldblatt et al., 19982), floral fragrance is common in the southern African species of the winter-rain- fall zone where sections Hebea and Homoglossum are best represented. It is also present in a few summer-rainfall zone species of the above sections that have their closest allies in the winter-rainfall zone. The scarcity of moth-pollinated species in summer-rainfall parts of southern Africa may be the result of the absence of scent in sections of the genus centered there. In tropical Africa scent in moth-pollinated species of section Acidanthera ap- pears to have evolved independently within the sec- tion (the section as currently defined does not occur in southern Africa). Floral odor in Gladiolus species (Goldblatt et al., 1998a) is the result of the com- bination of many substances and in moth flowers perceived odor represents different mixes and pro- portions of these same series of compounds. Apart from fragrance, moth flowers differ from their bee-pollinated ancestors in perianth tube length, color, and size of the tepals. Large size is probably important for display purposes at night as is the pale coloration. An unusual aspect of moth- pollinated flowers in Gladiolus is the frequency of dull patterning or mottling on a pale background, most notable in G. hyalinus, G. liliaceus, G. ma- culatus, and even some forms of G. longicollis (all sect. Homoglossum) and in G. emiliae and G. guth- riel (sect. Linearifolius). Dark mottling is less pro- nounced in G. hyalinus and G. liliaceus at night, as the pigments either change in intensity or color or both. The mottling may be camouflage, protecting flowers from predation during the day, as Johnson (1994) has suggested for the maroon-flowered or- chid, Monadenia ophyridea, which is presumably located by its settling moth pollinators by scent alone. Butterfly-pollinated flowers in Gladiolus sect. Blandus appear have evolved from ancestors adapted for pollination by long-proboscid flies, the predominant pollination system in the section, and the one present in species most closely allied to the clade of butterfly-pollinated species in the section Goldblatt & Manning, 1998, 1999). Long-probos- — Volume 89, Number 1 2002 Goldblatt & Mannin g 123 Gladiolus Lepidoptera Pollination cid fly flowers have similarly long floral tubes and differ cream or pink with red nectar guides often outlin- most significantly in floral pigmentation, ing a white zone. In Gladiolus sect. Ophiolyza the ancestors of the two butterfly-pollinated species have flowers adapted for pollination by sunbirds 1999). In series Linearifolius butterfly-pollinated б. nerineo- (the С. dalenii complex) (Goldblatt et al.. ides and G. stokoei (putatively butterfly pollinated) fall within a clade with the bee-pollinated G. brev- ifolius and long-proboscid fly pollinated G. monti- cola (Goldblatt & Manning. 1998, At first it might appear that a simple shift from ancestral pink or cream flowers to red ones may effect a shift to butterfly pollination. In Gladiolus increase in flower size appears equally important. Moreover, other more subtle changes suggest that more complex genetic adaptations are necessary. including a shift in the position of the anthers and stigmatic surfaces, either prominently exserted or included in the floral tube. and a shift in flowering time (butterflies are on the wing only in late sum- mer). vegelative and physiological adaptations would also be required to produce flowers in an area of sum- mer heat, drought, and low atmospheric humidity. a time that necessitates the separation of vegetative growth from the flowering phase of the life cycle. Johnson and Bond (1994) have pointed out a fea- ture of butterfly pollination in the southern African that question. This is that so many butterfly-pollinated winter-rainfall zone relates. directly to this species of the area are rare, or are highly local endemics. We suspect this has nothing to do with the intrinsic effects of having a specialist pollina- tion system and is correlated instead with the fact that species flowering in the late summer and au- tumn in an area of mostly winter precipitation re- quire highly specialized habitats that are by their Thus. б. vives summer drought by growing along watercours- very nature highly local. cardinalis sur- es in cool mountain habitats. often in the spray of perennial waterfalls. Gladiolus sempervirens, G. ste- faniae, and G. stokoei also grow in montane habitats and are restricted to slopes with poor drainage where the summer southeast winds bring moisture in the form of fog that keeps the ground moist al this otherwise dry season. Gladiolus nerineoides is restricted to shaded south-facing cliffs where mois- ture percolating from the slopes above keeps their substrate relatively damp. Even in these special- ized habitats vegetative specialization makes sur- vival for these summer-flowering species possible. Gladiolus nerineoides, G. stefaniae, and G. stokoet produce foliage leaves in the winter, after the wet In the winter-rainfall zone a whole series of season has begun. Their flowering stems bear short, often entirely sheathing, and sometimes dry leaves while the foliage leaves are dry and often lost by the time flowers are produced. An aspect of evolution of specialized pollination systems is a change in the nature of the reward. Bee-pollinated species of Gladiolus mostly produce sucrose-rich or sucrose-dominant nectar of relative- 1998a). The in moth-pollinated flowers is similar, al- This pattern is one that is common in the Iridaceae. In (Goldblatt et al., in press) nectar sugars usually have concen- ly high concentration (Goldblatt et al., pattern though sugar concentration is often higher. moth-pollinated species of Hesperantha (€ trations of 45-50% sucrose equivalents. The rela- tively high volume of nectar in most moth flowers, especially compared to their presumed bee-polli- nated ancestors, may relate to the higher nutrient needs of active, fast moving moths with relatively arge bodies, especially in the case of Sphingidae. which hover while feeding. Unlike nectar of moth flowers, the nectar of but- terfly-pollinated Gladiolus species shows a lower sugar concentration than their presumed long-pro- boscid fly or bee-pollinated ancestors (Goldblatt et al.. 1998a: Goldblatt & Manning, 1999), but a rel- atively high volume. Exactly why the lower nectar sugar concentration is characteristic of large but- terfly pollination is uncertain. The higher viscosity of concentrated nectar may make it difficult for an insect to draw through the proboscis, but this is evidently not a consideration for small arge moths. which have similarly constructed and some- times equally long and narrow mouthparts. Another feature of butterfly nectars in Gladiolus is the shift to higher proportions of hexose sugars. While most species of section Blandus have the sucrose-dom- inant type nectar found in related long-proboscid fly pollinated species (Goldblatt & Manning, 1999), G. insolens has hexose-rich nectar. A similar but less pronounced shift is evident in the two butterfly- pollinated species of section Ophiolyza, which have sucrose-rich nectar with nearly equal proportions of sucrose and hexose sugars, and in G. nerineoides. which shows the same shift, while related species = have nectar with higher proportions of (Goldblatt € Manning, 1999; Goldblatt et al.. 1999). Such changes in nectar sugar chemistry are sucrose rare in the Iridaceae, especially in the subfamily Crocoideae (syn. Ixioideae) to which Gladiolus be- longs. and are presumably pollinator driven, re- flecting a preference of Aeropetes for nectar with elevated levels of hexose sugars. rather than a ran- dom change. Shifts in pollinators are common within the larg- 124 Annals of the Missouri Botanical Garden er genera of the African Iridaceae and are an im- portant factor in the radiation of the family. Else- where we have speculated that as many as 32 shifts in pollination system have occurred just in Gladi- olus in southern Africa (Goldblatt et al., 2001). This represents an average of one shift for every five species. The convergent evolution of moth polli- nation at least six times and butterfly pollination three times reflects the evolutionarily labile floral and vegetative morphology of this highly successful genus. As outlined above, shifts that may appear quite simple, appearing to represent changes main- ly in pigmentation and floral tube length are, in fact, complex and represent a closely integrated se- ries of adaptations that include, in addition to floral patterning, shifts in fragrance characteristics, tim- ing of the opening and closing of the tepals and sometimes scent release, nectar physiology, season- al phenology, and associated vegetative adaptations for those species of the southern African winter- rainfall zone that flower in summer. The develop- mental complexity associated with these shifts makes their high frequency seem even more im- pressive. Literature Cited Baker, H. G. . Baker. 1983. Floral nectar me stituents in re — to pollinator type. Pp. 117—141 in C. E. Jones & R. J. Little (editors), Такі of Ex- perime а) seme ate Biology. Scientific and Academic Editions, New Bue mad S.L. 1983. Buzz pollination in е rms. 73-113 in C. E. Jones & R. J. Li (editors). Handbook of Experimental. Pollination. E Nostrand Reinhold, New York Faegri, K. & L. van der Pijl. 1979. The Principles of ation Ke sls 3rd ed. Pergamon Press, New Yo Goldblatt, Р. 19 "ail in Tropical Africa. Timber Press, pora. Oe on. & J. C. Man 1998. Gladiolus in Southern Af- rica. Fernwood Press, Town. ——.. 1999. The long-proboscid fly pollina- . Ann. Missouri Bot. æ = lion system in E (Iridaceae). Gard. 86: 758-774 Volume 89, Number ‚ 2000. The long-proboscid fly pollina- tion system in southern Africa. Ann. Missouri Bot. Gard. 87: 146-170. Bari & J. C. Manning. 1991. Sulcus variability in че pollen grains of pow ‘eae subfamily Ixioideae. Ann. Missouri pot — 950-961. & | — 1995. Pollination in Lap- etrousia subgenus — (Iridac eae: Ixioideae). in. Missouri Bot. Gard. 82: 517-5: › Bernhardt & J. C. Manning. 1998a. Adaptive аса of bee pollinated Gladiolus species (Iridaceae) in southern Africa. Ann. Missouri Bot. Gard, 85; 492 517. E — ——. 1998b. Pollination by mon- key Бева (5с — Hopliini) in peta taloid geo- phytes i in southern Africa. Ann, Missouri Bot. Gard. 85: 215-230. 1999, Evidence of bird pol- lination in "s iva 'eae of southern Africa. Adansonia er. oO, . 2: ‚ 2001. Radiation of pollina- tion — in КҮ (Iridaceae: Crocoideae) in southern — Ann. Missouri Bot. Gard. 88: 713—734 „Р. Bernha rdt & J. C. Manning. In press. Floral biology P сони (Iridaceae: Crocoideae): How minor spud 1 floral presentation change the pol- lination pA Ann. Missouri Bot. Gard. Johnson, S. D. Evidence for Batesian mimicry in a butterfly- e orchid. Biol. J. Linn. Soc. 53: 91— 104 T — © . Moth pollination of the cryptic — or- chid, Monadenia ophrydea. Flora 190: 105-10 —— & V. J. Bond. 1994. Red flowers and Iter pollination i in the fynbos of South Africa. Pp. 137-148 . Arianoutsou & R. Grooves (editors), do Animal ner ‘tions in oe Type Ecosystems. Kluwer — — Jordrecht. "M user, R. ne Scent of Orchids. Elsevier, Amster- dar dan — J. C., P. Goldblatt & P. D. J. Winter. 1999, Two new species of Gladiolus (lridaceae: Ixioideae) from South Africa and notes on pi probosc ‘id fly pollination in the genus. Bothalia 29: —223. Ogden, E. C., G. S. Raynor, J. x Hayers & D. M. Lewis. 1974, e of Sampling Airborne Pollen. Hafner 2 Pre Scott — "1891. Notes on the fertilisation e South Afric 'an — E an flowering plants. Ann. Bot. 5 05 l, pp. 1-124 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on April 15, 2002. а ра 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 Soper American Journal of Botany : International Journal of Plant Sciences | American Midland Naturalist | Journal of Animal Ecology e American Naturalist | ` . Journal of Applied Ecology Annals of the Missouri Botanical — Journal of Biogeography Annual Review of Ecology and Systematics — Journal of Ecology ee Biotropica Journal of Tropical Ecology Brittonia Journal of The Torrey Botanical — Conservation Biology Limnology and Oceanography Diversity and Distributions Missouri Botanical Garden Jab дерек Ecological Applications _ New Phytologist - Ecological Monographs - : . Paleobiology . — Ecology ets as . Quarterly Bien iof Biology. Evolution : ; Systematic Biology ' Functional Ecology Үе tur улты puer Global Ecology and 1 Biogeography 1 Letters m { —— fe PE Ж, 1 a Ж ISTOR 1s available at nup./ г Y А ишо A oJ o 3 : _ F J 120 Fifth Avenue, New York, NY 10011. — — — CONTENTS A Systematic Revision of Breonia (Rubiaceae—Naucleeae) Sylvain G. Razafimandimbison Classification, Origin, and Diversification of the New Zealand Hebes (Scrophulariaceae) Steven J. Wagstaff, Michael J. Bayly, Philip J. Garnock-Jones & Dirk С. Albach Phylogenetic Reconstruction of the Neotropical Family Quiinaceae (Malpighiales) Based on Morphology with Remarks on the Evolution of an Androdioecious Sex Distribution Julio V. Schneider, Ulf Swenson & Georg Zizka A Review of the Genera Roentgenia and Potamoganos (Bignoniaceae) — Warren D. Hauk Molecular Data Indicate Complex Intra- and Intercontinental Differentiation of American Draba (Brassicaceae) Marcus Koch & Ihsan A. Al-Shehbaz Evidence for MN and Butterfly Pollination i in Gladiolus (lridaceae—Crocoideae) ..... Peter Goldblatt & John C. Manning Annals ' of the Missouri Dotanical "mo" umber Volume 89, Number 2 Spring 2002 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 Committee 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 2002 is $140 per volume U.S., $150 Canada & Mexico, $175 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 © Missouri Botanical Garden 2002 +h 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, Р.О. Box 1897, Lawrence, KS — 66044-8897, The mission of the Missouri Botanical Gadi is to discover end share knowledge about plants and their e "nt, in order to preserve and enrich life. 7 us | @ mis paper meets the requirements of ANSI/NISO — (Permanence of Paper). olume 89 umber 2 002 nnals of the ISSOUTI otanica arden INTRODUCTION: CONSERVATION AND THE FUTURE OF LIFE! George E. Schatz? and P. Mick Richardson? Upon initial thought, conservation might seem an anomalous toplc for a systematics symposium, However, although systematics is often perceived only in its arcane, classificatory function, and con- f dwindling populations of charismatic large animals, servation has primarily had a publie persona o they are in fact two sides of the same coin, Or said another way, they concern themselves with a com- the diversity of life on Earth. Annual Sys- tematics Symposium was to highlight the intimate mon Currency, I... Indeed. one of the goals of the 47th connection between systematics and conservation, Vs scientific disciplines, both encompass broad ap- proaches, from the “top-down” landscape and ecoregion paradigms of conservation, and the “Tree of Life” phylogenetic perspective in systematics, to “bottom-up” Population Viability Analyses and ex situ interventions in conservation, and analyses of genetic variability in systematics. At all levels, sys- tematies directly informs conservation. and all per- spectives ultimately converge upon species, the most recognizable unit of biodiversity. Among the countless species in the cornucopia of biodiversity, one species—our own—has distin- guished itself from all others. Human beings have the dubious distinction of having become the most success. we have «on: the first species to alter ı achieving that — successful species of all time. hi basic biosphere processes since the original pho- l sphere and sent evolution down an aerobic path. tosynthetic bacteria added oxygen to the atmo- Although the rise of one species has undoubtedly often contributed to the decline of another species during that long evolutionary continuum, our suc- cess now directly or indirectly negatively impacts (or eventually will impact) nearly all of the other 10-100 million species with whom we share our home. We are both the problem. and the only po- tential solution. But as К. O. Wilson (2002) so eloquently artic- ulates in his most recent book, The Future of Life. the The following Sympo- there is much about which to be hopeful in urgent search for solutions. sium contributions all embrace that optimism. il- » "bottom- Joel syste matics lustrating the full range of “top-down” ti ultimately human-centered approaches. the during the past several decades and its ensuing di- up.” Crac тай invokes re naissance e in ! This Missouri Botanic ‘cal Garden. The symposium was held i > National Science Foundation unc »eaker ect a fine 5 Diana Gunter for ? Missouri Botanical Garden. P.O. their editing and political skills, and th Box 299, St. Louis. ANN. MISSOURI Bor. and the six articles that follow it constitute ys i Kathy Hurlbert anc d |" Myers for his ett used on the cover of the symposium brochure, Victoria Hollowell. Amy ceedings of и | nh no ıl Systematics Symposium of the ouis “Miss t October 2000. The symposium was grant DEB- ‘998 1642 2. PMR thanks Peter Raven for helping xpert staff for organization and уто sate skills. IcPherson. and St. | ler — e spec akers for bei "ng suc E an inte resting group of scientists Missouri 63166-0299, U.S.A. GARD. 89: 125-126. 2002. 126 Annals of the Missouri Botanical Garden rect application to societal problems as evidence of its relevance in conserving and sustainably using biodiversity. Both he and George Schatz emphasize the fundamental importance of the natural history collections that underpin all our knowledge of bio- diversity, and underline their critical role in di- rectly informing conservation planning using pow- erful new informatics tools. The mapping of biodiversity shows us that most species are not ran- domly distributed over the face of the earth, and that collectively they form distinct associations, of- ten highly restricted in area, that correlate with the variety of life-supporting milieu and historical phe- nomena of isolation and dispersal. Thus, David Ol- son and Eric Dinerstein advocate the landscape paradigm to ensure functional representatives of each of the world’s distinct ecoregions, i.e., those that satisfy the minimum size requirements of all their constituent species, and retain the potential for migratory processes. In virtually all of those ecoregions, we are one of the constituent species, often the dominant опе; for better or, now, mostly worse, all of the rest of the species are consigned to live alongside us. Again, we are the only potential solution. Gary Nabhan and his colleagues paint a portrait of the Colorado Plateau in which indigenous cultural di- versity and biological diversity reciprocally sustain one another, serving as an example for the local, community-based management of biodiversity. De- bra Moskovits and her colleagues present an in- spiring vision of Chicago Wilderness that serves to resurrect our innate capacity for stewardship amidst even our most dense, urban environments. In con- trast, Stuart Pimm validates the alternative scenar- io, dispelling once and for all any lingering myths that we are not the current agents of extinction, which indeed we regrettably are. The consequences of our success, and the ulti- mately self-detrimental manner in which we cur- rently live, are becoming increasingly clear. As we strive to disseminate our emerging knowledge, we must nurture the belief that it will soon foster the ethical principles and moral standards by which we will fully embrace not only each other, but all of our fel low species. Literature. Cited Wilson, E. O. 2002. The Future of Life. Alfred A. Knopf, New York. THE SEVEN GREAT Joel Cracraft? QUESTIONS OF SYSTEMATIC BIOLOGY: AN ESSENTIAL FOUNDATION FOR CONSERVATION AND THE SUSTAINABLE USE OF BIODIVERSITY" ABSTRACT The three missions of Systematics Agenda 2000 (SA2000)—to inventory Earth's species, to understand their rela- tionships, and to use the latter to create predictive information systet ns—define an agenda of research for systematic biology. The recognition that systematic knowledge underpins biologie ‘al knowledge in general, and applied biology i in a | 8 | е particular, has resulted in an amazing growth in systematics over the past decade. Increasingly, systematics is being used to solve societal problems. This paper describes seven great questions within systematics and — 's thelr ‘ir relevance for, and contribution to, conserving and sustainably using biodive sily. These questions fall into four broad categories: Questions about diversity: What is a species? and How many species are there? Questions about phylogeny: What is he Tree of Life? and What has been the history of character transforma in? Questions about biogeograph y: | geograph Where are Earth's species distributed? and How have species! distributions changed over time? and Questions about I 8 phyloinformatics: How is phylogenetic history predictive? Key words; biogeography. diversity, inventory, phylogenetics. phyloinformatics, systematics, Systematics Agenda 2000 Systematic biology occupies the central core of tion of the importance of systematics has never biodiversity science. The four great themes of sys- been stronger. tematics—diversity, phylogeny, biogeography, and There is a sense, however, in which some of classification (to be subsumed here into a broader these discussions linking systematics and conser- context termed. phyloinformatics)—form a critical vation can be said to be too narrow, both from the — foundation for all other disciplines of biology. The perspective of syste matic s and from conservation. discovery and naming of life's diversity. its evolu- Thus, "conservation," a east in the expansive tionary history, its distribution across Earth to form sense it will be used here, is more than just efforts habitats and ecosystems, as well as how compara- to save endangered species and ecosystems. or to tive information about organisms can be organized create protected areas. While certainly embracing in a predictive manner, underpin, to one degree or these objectives, conservation in the real world cov- another, all biological knowledge. ers much more ground philosophically, as well as Beginning in the early 1990s systematists be- in practice. In a world losing its biodiversity at an came increasingly engaged in conservation and the accelerating rate, systematics needs to be seen as role played by systematics in solving societal prob- a crucially relevant and important science in meet- lems (Systematics Agenda 2000, 1994a, 1994). ing the challenges of global environmental change Since then, systematists have published numerous (sensu lato), at the same time contributing strongly papers that address the contributions of systematics to traditional conservation practice. to conservation biology, focusing especially on such It will therefore be a theme of this paper that topics as diversity patterns, species concepts, geo- systematics and systematists should be approaching graphie distribution, and endemism. The recogni- conservation as a subset of solutions that must be ! I thank Peter Raven and P. Mick Ric ен for asking me to participate in the 47th Annual Systematics 5 at the Missouri Botanical Garden. Meredith Lane ea many helpful pea on the manuscript that greatly — its clarity; I am most grateful to her. Many | of the intellectual seeds of this article derive from the hard work and brilliance of the höndreda of systematists who ud to Systematics Age — 2000. which started in the United States but has now spread у vorldwide. The initial effort of Systematics Agenda 2000 was supported by a grant from the U. S. National Science Foundation (DEB-9396035). Consider this yet another contribution from that investment. ? Department of Ornithology, American Museum of Natural History. New York, New York 10024, U.S.A. jle@amnh.org ANN. Missouni Bor. GARD. 89: 127-144. 2002. 128 Annals of the Missouri Botanical Garden implemented to mitigate the adverse effects of en- vironmental impacts in general, all the time real- izing that the contributions undertaken to sustain and improve human well-being can themselves be seen as a conservation initiative. Stemming the loss of biodiversity is essential, but the factors leading to that loss are imbedded in a complex causal nex- us that encompasses all aspects of society. People across the globe use tens of thousands of species to sustain their lives in one way or the other, and the causal complexity of that use is what makes saving biodiversity so difficult—it cannot be sepa- rated from how societies function. This reality is encapsulated in the activities of the Convention on Biological Diversity (CBD) and many nongovern- mental organizations. Effective conservation efforts cannot be separated from the elimination of poverty, which cannot be separated from the improvement of women's health, education, and economic en- franchisement, which cannot be disentangled from governmental policies of many kinds, and on and on. New necessary to meet the many environmental chal- nowledge about the world (science) is lenges created by human activities, and thus sys- tematics can help in all sorts of ways not generally realized by most practicing systematists or other biodiversity scientists. This is what I mean by the need for systematists to have a more expansive vi- sion for their discipline and for its importance. This paper explores the contributions of system- atics to science and soc let y, firstly, by summarizing some of the literature published since the mid- 1990s, when participants in Systematics Agenda 2000 (842000) released many of their assessments, and secondly, by pointing to new developments that will become increasingly important in the future. This will be done within the framework of what might be called the seven great questions of sys- tematic biology. which to me at least, encapsulate most of the research agenda, and importance, of systematics to society. Not all of these questions will be treated equally here. inasmuch as some have been the subject of a substantial recent lit- erature (see citations below). SYSTEMATICS AGENDA 2000: BRINGING SYSTEMATICS TO SOCIETY In the early 1990s the Society of Systematic Bi- ologists, the American Society of Plant Taxono- mists, and the Willi Hennig Society, in cooperation with the Association of Systematics Collections, launched an effort to document the importance of systematics to society (Anonymous, 1991). Through its many publications, Systematics Agenda 2000 transformed the way systematists view their own discipline, and it helped create an atmosphere in many countries throughout the world in which sys- tematics gained in stature and importance as ап integral component of biodiversity science. Jecause many systematists, both young and old, might be unfamiliar with the rich panoply of pub- lished results of this effort, these are summarized 1 Table 1. These papers cover a broad range of subjects that articulate the importance of system- atics. Collectively, they have reached out to system- atists and biodiversity science policymakers around the world and have been responsible, in varying degrees, to promoting and building systematics. Thus, to take one example, following SA2000 sym- posia at the Royal Society in London (Claridge, 1995) and the French Academy of Sciences in Par- is, new initiatives were formed to promote the ob- jectives of systematics across Europe (Blackmore & Cutler, 1996 Systematics Agenda 2000, although begun in North America, was developed as a global effort. Thus, the core documents of SA2000 were inten- tionally not copyrighted so that they could be taken, and used, by systematists to promote systematics research and capacity building anywhere. Transla- tions were encouraged, and that strategy worked as numerous individuals and groups adopted the lan- guage and content of SA2000 in their efforts (Table ). Today, SA2000 is truly international. System- atics Agenda 2000 International (SA2000I) is a program of the International Union of Biological Sciences (IUBS) and functions as the systematics core element of the international biodiversity sci- ence program DIVERSITAS | [DIVERSITAS|. Through DIVERSITAS, SA20001 has been active in providing advice on systematics science and ca- pacity building to the Convention on Biological Di- versity (CBD) and its Global Taxonomy Initiative [GTI]. Some of the SA2000I/DIVERSITAS docu- ments in support of the GTI are available on the CBD website; these documents contain recommen- dations that have been adopted by the CBD Con- ference of the Parties (the 181 countries that have ratified the Convention). (For the record, as of 12 September 2001, only six signatory countries to the CBD have failed to ratify: Afghanistan, Yugoslavia, Thailand, Tuvalu, Kuwait, and the United States of America.) THE SEVEN GREAT QUESTIONS OF SYSTEMATICS The scientific agenda of systematics and its rel- evance for society occupy four great themes: diver- sity, phylogenetic history, biogeography, and clas- Volume 89, Number 2 Cracraft 129 2002 Systematic Biology An overview of the papers produced by participants in Systematics Agenda 2000 over the last decade. Table ‘uss aspects of building systematics These publications treat the role and importance of systematics to society, disc science capacity, or address policy issues involving systematics and biodiversity. Publication Theme General SA2000 Documents Anonymous (1991) Announced formation of SA2000 Systematics Agenda 2000 (1994a) Color brochure disc — — e of systematics and describing SA2000's three missions Systematics Agenda 2000 (1994p) Technical report hw ail for 1994a BioScience 1995 (vol. 45, no. 10) Simpson & Cracraft (1995) Savage (1995) Miller & ер (1995) Role of systematics in agriculture Brooks et al. )5) Role of systematies in ecology and behavior Lauder et al. Lo Overview of systematics and papers in BioScience Role of systematics in biodiversity science Role of systematics in comparative morphology and physiology Davis (1995) Systematics and public health Biodiversity pie Conservation 1995 (vol. 4, no. 5) Claridge (1995 Introduction to SA2000 symposium held at the Royal Society, 12 April 1994 Overview and history of SA2000 Eshbaugh (1995) Building systematics and biodiversity science capacity Cracraft (1995) Wheeler (19054) Prance (1995 Biodiversity and systematics inventorie Systematics, conservation, and DAS develop- теп! Jones (1995) BioNet-International and capacity building McNeely (1995) Annals of the Missouri Botanical Garden 1996 (vol. 83, no. 1) Richardson (1996) Systematics and conservation Introduction to SA2000 Symposium, 41st Annual Sys- tematics Symposium, MOBOT Monson (1996) Phylogenetics and comparative plant physiology and development Rossman & Miller (1996) Role of systematics in agriculture and forestry Vecchione & Collette (1996) Role of systematics in fisheries and marine biodiversity Oliver (1996) Importance of systematics for public health ane-Wright (1996) Role of systematics in conservation Balick (1996) Systematics and ethnobotany Other publications Sy — Age nda 2000 (1994) Wheeler (19 French translation of SA2000 (1994a) pe and biodiversity policy Agenda — 2000 (1996) n translation and expansion of SA2000 (1994) E. — & Cutler (1996) dr SA2000 in — r & Cracraft (1996) Building systematics capac Workshop report, cosponsore sd by SA2000 Internation- . on Convention of Biological Diversity Global lI IL — Museum of Natural History (1999‏ ا р; Taxonomy Initiativ Baas (1999) SA20001, DIVERSIT AS, and the CBD Cracraft (2000) Building systematics capacity sification. Each of these four themes leads to a THE FIRST GREAT QUESTION: WHAT IS A SPECIES? series of fundamental questions—the seven great questions of systematics. These questions broadly cover what systematists do, and are why, in my opinion, systematics is the central, undeniable core No question, probably, has generated more con- troversy, been so opaque to solution, and yet re- mains as crucial and important today as it ever has. of biodiversity science. than “What is a species?" In systematics, which is 130 Annals of the Missouri Botanical Garden a taxon-based science, it is unquestionably the most fundamental question that can be asked. As- tonishingly, many biologists profess to be tired of the controversies over this question, yet instantly remain willing to engage in the debate, whether or not they themselves are professionally concerned with individuating species limits, i.e., whether they are systematists or not. Everyone, it seems, knows what a species is, or more likely, what is not! The reason for the dispute is fairly obvious: at some level, notions of species are at the very heart of biological understanding and so there is a lot at stake. Species concepts intersect with a whole se- ries of conceptual and disciplinary agendas, from those of systematists who discover and describe Earth's diversity, to those reconstructing phylogeny and biogeography, to those attempting to under- stand the origins of species from a genetic or eco- logical viewpoint, to those interested in conserva- tion, ecology, and applied biology. Species are the basic units of systematics, evolutionary biology, and diversity. Thus, species concepts touch upon ele- mental philosophical arguments about the reality of the units of nature as perceived by biologists of varying disciplines as well as scientific. opinion about how life evolves. Given this crucial importance, it is ironic that there is so much divisiveness over “What is a spe- cies?” You would think biologists could have solved this issue by now. Yet each year brings forth nu- merous papers on the subject, and books keep flow- ing (most recently, for example, Ereshefsky, 1992; Kimbel & Martin, 1993; Claridge et al., 1997; Howard & Berlocher, 1998; Wilson, 1999; Wheeler & Meier, 2000), all with no diminution of differ- ences of opinion. Nevertheless, it can be argued that over the last decade, practicing systematists—those biodiversity scientists whose research most directly bears on this issue—are gradually converging upon a com- mon solution to this question, either as a result of ` both. These systematists see species as basal elici practical or theoretical considerations, clusters of diagnosably distinct populations (groups of individual organisms). Most of these systematists do not endorse a particular concept of species; they go about their work inventorying, describing, and monographing without a heavy burden of theorizing. But if one had to put a name on their concept, it would appear to be most similar to that advocated by supporters of a phylogenetic species concept (Nelson & Platnick, 1981; Cracraft, 1983, 1989a; Nixon & Wheeler, 1990; Wheeler & Platnick, 2000 The reason why this is the most common ap- proach to species is very practical, as well as sim- ply a matter of numbers: the majority of systema- tists working on invertebrates, especially insects, many plant groups, and various vertebrate taxa view species as basal diagnosable units because this best fits the way they partition the taxonomic variation they have observed. How could this not be the case when the large majority of described species are known from a handful of specimens, often single individuals? Or when nothing is known about their biology or patterns of genetic variation? Thus, the debate over species, in a practical sense, comes down to a choice between a phylo- genetic species concept (PSC) and a biological spe- cies concept (BSC). The latter, surprisingly, is ap- plied by very few practicing systematists who inventory and describe species, but has strong sup- port from advocates of the evolutionary systematics of Mayr (1942, 1963, 1982, 2000) and by those whose interests lie with population genetics (Coyne 1988; Avise & Ball, 1990) or evolutionary biology in general (e.g., Bock, 1987; Futuyma, 998). The interchanges among advocates of the PSC and BSC have been incessant. It is not my purpose to review these debates as the central ar- guments and positions can be found in Wheeler and Meier (2000). Instead, pursuant to the theme of this paper, | comment briefly on the relevance et al., of species concepts to conservation and applied bi- ology. Species concepts are important because they al- low us to propose hypotheses about the ontology of nature: different species concepts generally imply a different ontology (Cracraft, 1987, 1989b). This is not just philosophical mumbo-jumbo. One needs a clear idea of the entities of nature so that one can count and describe patterns of diversity, as well as understand how entities behave (i.e., how they par- ticipate in processes). Thus, there are very practical consequences stemming from the adoption of a par- ticular species concept, especially in the descrip- tion and enumeration of diversity. Conservation biology, although inherently cross- disciplinary, emerged primarily from an ecological tradition, and ecologists and other non-systematist biodiversity scientists have come to their under- standing of species and speciation through their formal university training in ecology, genetics, and evolution. Most of that training has accepted the BSC because it has been the canonical view of spe- cies since the early influential work of Ernst Mayr 1942, 1963) and because it has been followed in most contemporary textbooks (e.g., Futuyma, 1998). However, many conservation biologists began to see a problem with applying the BSC because of its — Volume 89, Number 2 2002 Cracraft 131 Systematic Biology ambiguous treatment of discrete taxonomic varia- tion. Thus, under the BSC, diagnosable populations might be ranked either as a species or subspecies, or subspecific rank itself might be applied to diag- nosably distinct forms as well as to arbitrary sub- divisions of clinal variation, In an influential paper, Ryder (1986: 9710) sum- marized the opinions of conservation biologists working within the zoo community: “Out of a sense of frustration with the limitations of current mam- malian taxonomy [broadly using the BSC] in deter- mining which named subspecies actually represent significant adaptive variation, those assembled at the Philadelphia conference [of zoo biologists | will- ingly discarded the concept that all subspecies are equal. Rather, it emerged that zoos ought properly to address the conservation of evolutionary signifi- cant units (ESUs within species).” They went fur- ther to suggest that identification of ESUs be made on the basis of concordance across multiple kinds of data and that “when geographic distribution data indicate the existence of discrete [italics added | populations within the range of a species. an esti- male of genetic distance, for example, should be made to determine whether the populations have ESL valionists were getting at: the traditional BSC ap- status.” It is fairly clear what these conser- proach to individuating units of nature was not working for their purposes. While the determination of whether a population could be judged to have "significant adaptive variation" verges on nonoper- ationalism. the key recommendation of their pro- identifying discrete populations—does not. posal They grasped the reasonable idea of trying to con- serve and manage (in situ and ex situ) diagnosable and distinct populations, ESUs as they were called. A very large literature has since developed with- in conservation biology on the concepts of species. ESUs, and other recently introduced terms such as "management units." Systematists have pointed out that ESUs have broad equivalence to the species units identified by the phylogenetic species concept and that use of the PSC conservation community not currently satisfied by the BSC (Cracraft, 1991, 1997; Vogler & Desalle. 1994; Barrowclough & Flesness, 1996). Other con- servation biologists, notably those having a popu- meels the needs of the lation genetic approach to the problem, have con- tinued to support the BSC and seek ways to refine theoretical and practical ue Be to the ESU concept (e.g.. Moritz, 1994a, b. 1995) There is, however, a — ii crucial argu- ment often left out of these debates: if conservation biology is to be a science that relies on a rigorous description of diversity, then it should adopt the anguage and conventions of systematic biology and (Cracraft. 1997: Wheeler. Although systematists may argue vehement- taxonomic practice 1997). ly over which species concept is best. they agree on many issues of formal taxonomy—that species- that names are tied to type specimens, that there are level taxa have formal Latin names, those standard rules of nomenclature (the international codes) so that scientific names can be organized and managed over time, and that there must be voucher specimens to document taxonomic deci- sions and descriptions. None of this is found concepts such as ESUs or management units. Conservation biology should therefore abandon such concepts as evolutionarily significant units. ESUs are not a substitute for formal taxonomy. Un- like formally described taxa with their types and historical continuity in rules of nomenclature, ESUs cannot have “legal standing.” ESUs cannot, and should not. be the units we regulate in trade. pro- tect with legal instruments, or expect to be used by applied biology for biotechnology. biodiversity in- formation systems, and many other uses. This is not the case with taxonomic units (Geist, 1992). Thus. the power of systematics and taxonomy: despite ar- guments over the most fundamental units of na- i agonists share disciplinary standards that prevent names, and thus the iden- tified taxa themselves, from devolving into chaos over long periods of time. THE SECOND GREAT QUESTION: HOW MANY SPECIES ARE THERE? Discovering and describing Earth’s taxonomic di- versity is the starting point for all biological knowl- edge. Because of its scope and complexity, however, knowing how many species inhabit Earth’s ecosys- tems is one of the megascience questions of biology. While it is generally accepted that around 1.5-1.7 million species have been discovered and de- scribed, estimates of unknown diversity range any- where from 10 to 100 million species, with 13—20 million being the most frequently seen number (e.g.. Stork, 1999). So the answer to this megasci- ence question is: We don't know! But we must find out. When one realizes that the use of biodiversity drives the world economy and this has come from knowledge of about 1.5-1.7 mil- lion species—probably less than 10% of all species on planet Earth—it is clear that abundant new ben- efits will flow from newly documented. diversity. Among these benefits of inventories identified by SA2000. one could include: 132 Annals of the Missouri Botanical Garden * they document patterns of diversity across habi- tats and ecosystems * they provide baseline data for monitoring activi- ties they identify areas of endemism and regions in need of conservation and protection * they discover new species having economic and social value to societies (pharmacological, agri- cultural, fisheries, biotechnological) they provide the baseline data for implementing an ecosystem approach to conservation and sus- tainable development * they support and promote tourism. We are disastrously ignorant of the natural world. That this is so was highlighted in the 45th Annual Systematics Symposium of the Missouri Botanical Garden (Richardson, 2000), which summarized the need for, and challenges to, systematic inventory ( 2000; Lundberg et al., 2000) and also highlighted that we are in a new age of discovery (Donoghue & Alverson, 2000). We know so little that even in the most well-known groups .g., Prance et al., ~ such as birds and mammals, new species are being .. Giao et al., 1998; Be- 1999; MacKinnon, 2000). The most outrageous and spectacular example of this is probably the description of six new species of le- murs (in three separate genera) from Madagascar, all published in the year 2000 (250 years post- Linnaeus) in a discovered each year (e.g resford & Cracraft, of the International Journal of Primatology (Thalmann & Geissmann, 2000; Groves, 2000; Rasoloarison et al., 2000). single issue THE THIRD GREAT QUESTION: WHAT OF LIFE? IS THE TREE Reconstructing the Tree of Life (TOL) is the third great question of systematic biology. This is also a megascience question for systematics for several well-known reasons. First, given N taxa, there are The con- sequence is that as the solution to the second great — ] nodes that need to be resolved. question. of systematics—How many species are there?—plays out, the domain of this third question gets larger and larger. A second major reason re- lates to the first: as the number of taxa in the tree increases, it gets more and more difficult to obtain sufficient data on these taxa, and the computational complexity of finding an objective answer also be- comes astronomical, Compared to deciphering the POL, the determination of the sequence of the hu- man genome, which for all intents and purposes was solved by a single, relatively small corporation, turned out to be a rather simple problem (it just required a little money and coordination among sci- entists). In contrast, as a scientific problem, resolv- ing the TOL is much more comparable in complex- ity, say, genome, knowing all the genes in the human how they function, and how their gene products interact to form a blueprint for develop- ment. So the answer to the third question is, of course: we don't know. As an index to the vastness of the problem, we can estimate that there are right now about 1.7 million nodes on the TOL, reflecting the number of species that have been discovered and described. Yet, where do we stand presently? No one knows for sure, but roughly—very roughly— perhaps 50,000 to 60,000 species are represented on one kind of tree or another. That is a pure guess inasmuch as there is no comprehensive database of trees. The only repository approaching what is needed is TreeBASE [TreeBASE], which has ap- proximately 12.000 taxa, but the sample is highly biased in being mostly botanical. Assuming 50,000 taxa already placed on one or more trees, it is fair to say the position of most of those is poorly supported by character evidence. While it is perfectly accurate to say our knowledge of the TOL is growing very rapidly, as measured by the numbers of phylogenetic papers being pub- ished, it is equally accurate to say a large per- centage of the nodes on those trees have relatively little support. this, among the most important being poor taxon and character sampling, poor choice of character sys- There are many reasons for tem, and ambiguities in the methods used to ana- lyze the data. Moreover, because taxon sampling is generally poor in most published phylogenetic studies, it is not at all clear how the different results can be linked with one another to form a general view of the TOL, a “supertree of life ” if you will. The most remarkable observation is that our un- derstanding of the TOL—at least in terms of the 50,000 taxa just mentioned is a product mostly of the last decade. Modern phylogenetics is only about 30—40 years old, and serious much older. “tree thinking” not The rise of “evolutionary systematics” in the 1930s and 1940s, with its emphasis on a population biology/genetic approach toward the his- tory of life, slowed the discovery of the TOL be- cause it was largely assumed that if ancestors could not be found in the fossil record, there was little hope of understanding phylogeny. If that seems a misrepresentation of history, one only has to examine the content of the major sys- tematics journals (for example, Systematic Zoology) prior to 1960 to see that depicting relationships as trees was not of major importance. There was re- markably little “tree thinking" prior to the intro- Volume 89, Number 2 Cracraft 2002 Systematic Biology duction of numerical taxonomy, a discipline that created trees, but was ambivalent in its interpre- tation of them. Many proponents of this approach saw their trees as purely representational of phe- netic similarity, not history; others hoped these di- agrams might reflect some trace of history. The im- portant point here is that, compared to evolutionary systematics, numerical taxonomy developed repeat- able methods that produced trees. Both evolutionary systematics and numerical tax- onomy were eclipsed by Willi Hennig’s conceptual and methodological development of phylogenetic systematics. or cladistics. The broad adoption of cladisties formalized tree-thinking in terms of phy- logenetic relationships and history. Also, the nu- merical methods that were rapidly introduced brought a much needed objectivity, both philosoph- ical and analytical, to the study of phylogeny. The explosion of phylogenetic knowledge over the last decade has resulted just as much from the concep- tual and analytical revolution of the 1960s to 1980s as it has by the introduction of efficient methods to gather new kinds of data, especially those from mo- lecular sequences. Why phylogenetics matters These advances in phylogenetic theory and have revolutionized systematic and methodology comparative biology, and the transformation of sys- tematies into a truly historical science could not have come at a better time. Society is desperate for knowledge about phylogeny. While many systema- tists still see an understanding of phylogeny as a goal in itself, numerous segments of society are looking to phylogenetics to solve entirely new kinds of problems. Consider the following examples: l. Tracing disease transmission The first application of phylogenetic analysis to examine disease transmission employed parsimony analysis to investigate whether a Florida dentist. discovered to be HIV-positive and who had con- tracted AIDS, had transmitted the infection to any of his patients (Ou et al., 1992). These investigators chose the HIV envelope (env) gene because of its high variability and compared sequences from the dentist to those of HIV-positive patients and HIV- positive nonpatients as controls. Phylogenetic anal- ysis showed clear patterns of genetic relationships between the dentist and at least five of his patients who had no identifiable behavioral risks to contract HIV infection (Fig. 1; of phylogenetic methods in analyzing HIV evolu- 1999). tion. see Holmes et al., 1996, and Crandall, for discussion of other uses — Dentist-x Patient D-x 4% Patient О-у Figure l. P орава netic tree of HIV-1 епу V3 sequenc- es from a HIV posi Florida dentist dee his patients A—G (x and y ce efe r to divergent clones, LC refers Ju et al., 1992, for details). These with the hypothesis that the dentist section. This was to local controls: see ( results were consiste nt was the source of the patients’ iw infe the first use of phylogenetic analysis to examine о Reprinted with permission from Ou et al. 65, figure 1. Copyright 1992 Advancement of Science; disease ~ (19 Science 256: 11 — in Association for the http://www.sciencemag.org. Tracking the spread of “emergent” diseases 2 | g Phylogenetics is playing an increasing role in the medical sciences, especially in identifying disease agents that spread from one region of the globe to another. DNA sequences from disease entities can be rapidly obtained and compared to sequences housed in databases such as [GenBank]. The sum- mer of 1999 in the New York City area brought a strange confluence of events. A number of people were stricken with an encephalitis that had the eti- ology of a flavivirus. At the same time large birds, including American crows (Corvus brachyrhynchos). were turning up dead in unusual numbers in the wild and in local zoos. Viral particles were even- tually isolated and their polyprotein nucleotide se- quence determined. Phylogenetic comparison with other sequences identified the newly emergent dis- ease as being related to West Nile Virus (Fig. 2; Lanciotti et al.. 1999), which circulates between birds and mosquitos and from the latter into hu- mans. Similar strains in the Mediterranean region and Middle East were also associated with in- creased avian mortality. The virus has now spread 134 Annals of the Missouri Botanical Garden WN-Romania 1996 H distance 0.045 Lineage 1 N-M —— 1988 N-Kenya un cagas 1986 -Uga Lineage 2 Figure 2. origin of the West Nile- New York City region in 19 These results indicate the New York stra lated to strains ina central and northern Africa, the Mid- dle East, and eastern Europe and was derived from that region. Reprinted with. — from Lanciotti et al. 19 Science 286: 2333, figure 2. Copyright 1999 American Association з the Advane етеп of Science; http://www.sciencemag.org. Phylogene i analysis was used to trace the like virus that broke out in о J9 (Lanciotti et al., 199 un is closely re- well beyond the New York region and also into non- human mammalian hosts. emergent diseases Discovery of “new” In addition to tracking diseases from one region to another, phylogenetic analysis is being used to discover new disease entities. In late 1998 and ear- ly 1999 a new mosquito-borne virus, called Nipah, emerged in Malaysia (Chua et al., 2000). Using pigs as a vertebrate host the virus jumped to humans, causing symptoms that first suggested Japanese en- cephalitis. Eventually 265 cases were reported and 105 people died from severe nervous system pa- thology. To control the epidemic, over a million pigs had to be slaughtered. role in Phylogenetic analysis played a major helping to characterize Nipah virus (Chua et al.. 2000). Comparative sequences were obtained from the nucleoprotein (N) gene and compared to other members of the subfamily Paramyxovirinae. The re- sulting tree demonstrated the relationship of Nipah to another recently discovered virus, Hendra virus, and the sequence differences indicated they were distinct (Fig. : 4. Monitoring and predicting viral host switching Karposi's sarcoma virus is endemic to central Af- rica and has associated with ita rhadinovirus, Kar- pos's-sarcoma-associated herpesvirus (KSHV), Un- til recently rhadinoviruses (y,-herpesviruses) were found in various Old and New World monkeys but not human’s closest relatives, the great apes. After determining sequences of herpesvirus DNA poly- merase taken from wild Pan troglodytes and Gorilla gorilla from Cameroon and Gabon, Lacoste et al. (2000) reported the discovery of new strains of these viruses. When those sequences were com- pared to others already known using a phylogenetic analysis, Lacoste et al. (2000) showed that these new viruses are closely related to KSHV (Fig. 4). The phylogenetic closeness of these new herpesvi- ruses and KSHV raises the potential for host switching into humans as they hunt and consume great apes for food. Phylogenetic analysis contrib- utes importantly to identifying and monitoring this new health threat. 5. Genomics, development, gene expression, and disease Phylogenetic thinking and methodologies аге taking hold in the fields of genomics and molecular medicine (e.g., Eisen, 1998; Pollock et al., 2000) and promise not only to increase our knowledge of the relationships of organisms but also surely will lead to insights into understanding and predicting gene structure and function. Developmental biolo- gists have long acknowledged the predictive and explanatory power of phylogenetic relationships in reconstructing the historical pathways of develop- ment (reviewed in Raff, 1996), and the rapidly ex- panding field of evolutionary development (“evo- devo”) will, reciprocally, result in major new advances in understanding developmental mecha- nisms and will inject new character systems into systematics that will inform phylogenetic relation- ships of major organismal groups. 6. Identification of invasive species The transport of alien species is a major global environmental problem. The United States alone Volume 89, Number 2 2002 Cracraft Systematic Biology 135 — 100 nt substitutions Genus Respirovirus Genus Morbillivirus Genus Rubulavirus Figure 3. Phylogenetic analysis of the nue ‘leoprotein (N) gene of the so-c called de virus that broke out in Malaysia in 1998 “and 1999 (Chua et al.. 2000) showed this paramvxov irus, 2000 American Association for the Advancement of Science: harbors about 50.000 invasive species at a loss of nearly $137 billion annually to control and mitigate their effects (Pimentel et al., 2000: Wolfenbarger & Phifer, 2000). Identifying potential exotic species is a major priority and a first-line of defense against them. In 1984 a tropical marine green alga (Cau- lerpa taxifolia) escaped from an aquarium and in- vaded the Mediterranean Sea. This particular strain proved remarkably hardy and competitive, and spread rapidly to devastate populations of native species (Jousson et al., 2000). The species was re- cently discovered several locations along the California coast. Jousson et al. (2000) posed the question whether the California populations of Cau- lerpa could be identified as an invasion of the Med- iterranean strain, representing a potentially serious threat to coastal ecosystems. Comparing DNA se- quences from the internal transcribed spacer of ri- bosomal DNA from multiple populations, phyloge- from netic analysis united 11 of 12 samples in California with those from the Mediter- sequences ranean: one sequence clustered with a Red Sea/ IndoPacific clade (Fig. 5). It was concluded that an immediate eradication program was warranted. 7. Discovery of microbial diversity Phylogenetic analysis of DNA sequences has be- come a major tool in the discovery of new micro- organisms, especially bacteria. Because most « these organisms cannot be cultured, microbiologists have turned to molecular probes for inventory and s new emergent virus Hendra virus. Reprinted with permission from € hua etal. ( 2000). — 288: ated to another recently discovered 1432. figure 4. Copyright Was с lose ly : http://www.sciencemag.org. identification (Pace, 1997). Typically using probes for rRNA genes, the sequences are compared those databases by various phylogenetic tech- niques. Phylogenetic methods have thus opened up entirely new approaches to understanding the mi- erobial diversity of extreme environments (Horiko- shi & Tsujii, 1999) and have led to a greater un- derstanding of the distribution of microbial life forms. It is now appreciated, for example, that ar- chaebacteria are not only found in extreme envi- ronments such as hydrothermal vents and hot springs but are much more widespread than pre- viously thought, including a variety of coastal and open ocean habitats (DeLong, 1992). THE FOURTH GREAT QUESTION: WHAT HAS BEEN ГНЕ HISTORY OF CHARACTER TRANSFORMATION? The proposition that the history of character transformation might be considered a great ques- tion of systematics may strike some as a bil strange. but reflection will confirm that all we know about the evolution of form and function derives from how character change is interpreted relative to a given which танк have been two main ways in sformations are studied. First, in gen- tree. character tra eraling esa hypotheses under maximum parsimony, characters used build the tree are optimized on it thus allowing inferences about their transformation across the tree. A second method has been to take some tree as given and then op- timize. or plot, characters on it. There is a growing 136 Annals of the Missouri Botanical Garden 100 75 TM. mulatta ow 100 М. жассан 00 М. nemestrina o6 ChRV2 EBV-B958 « y1 J MCMV 89 HCMV 80 p-Herpesvirus an o ives угу 7] GHV2 Figure 4. A phylogenetic tree based on sequences of the DNA polymerase gene for several herpesviruses newly discovered in the chimpanzee (Pan troglodytes) and lowland gorilla (Gorilla gorilla) in Cameroon and Gabon (Lacoste et al., 2000). These results indicated these new viruses were closely related to the Karposi's-sarcoma-associated her- pesvirus (KSHV) found in humans. This suggests the possibility that the new herpesviruses might be transmitted to humans since chimps and gorillas are frequently used for food; thus, phylogenetic analysis can be used to predict possible outbreaks as well as help establish a monitoring program for new infection. Tree modified and reprinted with permission from Lacoste et al. (2000), Nature 407: 151—152, figure 1. Copyright 2000 Nature; http://www.nature. сот/. literature arguing and demonstrating empirically no phylogenetically relevant information, which is that the first approach is to be preferred, primarily seldom true. because the characters of interest are often deter- Nevertheless, the second approach of plotting or- minative with respect to choice of most parsimo- — ganismal attributes on a phylogenetic tree will re- nious tree. The second approach tacitly presumes main popular and exceedingly important. Indeed, that the characters being examined have little or the main benefit that nonsystematists gain from a Volume 89, Number 2 2002 Cracraft 137 Systematic Biology Mediterranean California Australia i a Australia Indonesia New Caledonia Red Sea Australia Caribbean Red Sea Figure 5. the place of origin of a recently discovered invasive alga ( Caulerpa taxifolia) along the California coast (Jousson et al., 2000). the internal tran- scribed spacer of ribosomal I l strains, it was found that the Pa ar strain clustered with those from the highly invasive strain th considerable damage in the Mediterranean. This suggest- ed the need for an immediate eradication program. Re- ке with жн from Jousson et al. (2000). Nature —158. figure Ib. Copyright 2000 Nature: http:// a el Phylogene ‘tic analysis was used to ascertain Comparing se — es from > £e = > = ys — — — fuller understanding of the Tree of Life is that it helps them understand the history of characters and make predictions about taxa for which those char- acters are yet unknown. Significantly, many of the users of phylogenetic information are themselves contributing data about character systems that will potentially inform our understanding of relation- ships. Literally hundreds of papers have used phylog- enies to interpret nonsystematic data, and there is no question that this has led to numerous insights into the history of many character systems in be- havior, ecology, physiology, and other sciences (e.g.. Harvey et al.. 1996). More important. perhaps. will be the power of character transformation analysis in applied biology. It has been used. for example. to guide the search for new pharmaceuticals or bio- chemical products (see Monson, 1996, for review). Peterson — understand the history of gene regulation & Davidson, 2000) and body complexity (Graham et al.. 2000: Cameron et al., 2000). among others. Answers to the fourth great question will take on more importance as phylogenetic knowledge and comparative databases expand and become more available. EARTHS THE FIFTH GREAT QUESTION: WHERE ARE SPECIES DISTRIBUTED? This is the most fundamental question of bioge- ography and any answer will only be as good as the inventories on which it is based. Thus, knowledge of distributions relies on the presence of georefer- enced specimens housed in the world’s natural his- tory collections. Ultimately, the characterization of Earth's habitats and ecosystems depends on these data, as does the ability to manage and conserve biodiversity. The practical importance of having knowledge of species’ distributions is acknowledged by the desire of nations, and intergovernmental organizations, nongovernmental conservation organizations 10 have specimens in natural history collections da- tabased and made freely available on the internet. The development and availability of geographical information systems (GIS) and other software pro- erams for mapping diversity are also increasing the value of digitally captured specimen data (e.g... Funk. 1997). Although such data are increasingly coming online, most of the world’s collections are not databased. This has led some governments, most notably Mexico, to accelerate data acquisition on their own. The value of their efforts has been well documented (e.g... Bojorquez-Tapia et al., 1994: Soberón et al., 2000) and recognized the as evidenced by the formation of the world over, Global Biodiversity Information Facility [GBIF]. Distributional information for individual species leads to the search for patterns of diversity at dif- ferent scales. Most importantly, this builds knowl- edge about areas of endemism, and discovering ar- eas with high numbers of endemic taxa (sometimes referred to as “hotspots”) is widely considered crit- ical for se tting Cc — к 5 (F orey e ta 1994: Nielsen & : Nielsen, 1999). But distributional data * a E larger significance for society than simply that associated with conser- vation. Distributional information tied to specimens underlies drug discovery, ecotourism, trade in nat- ural resources. pest control, control of invasive spe- cies, crop improvement using the genetic diversity of wild relatives. and analysis of global change. among many other applications. THE SIXTH GREAT QUESTION: HOW HAVE SPECIES DISTRIBUTIONS CHANGED OVER TIME? This great question of systematics, and the sec- ond pertaining to biogeography. can be looked at in two ways. First, in the here and now: The vast ma- jority of research and activity in in ecological biogeography, in which ecologists at- “biogeography” is tempt to understand and explain why organisms are distributed where they are, how those distributions are tied to autecologies, and so forth. Yet, without accurate taxonomic descriptions and georeferenced 138 Annals of the Missouri Botanical Garden data vouchered by specimens, the quality of the ecology itself will suffer. A component of this re- search also looks to the future. Human activities are transforming the biosphere and there is interest in predicting how this impact will affect the distri- butions of organisms and, downstream, societal well-being, especially for critical activities such as agriculture. The second way of thinking about changes distributions is through the eyes of the historical biogeographer. Following the introduction and ap- plication of cladistic methods to distributional problems (Rosen, 1978; Nelson & Platnick, 1981), interest in the biogeographic history of taxa and areas of endemism expanded significantly. As was realized early on, the key to understanding the his- tory of biotas is through an analysis of the history 1986). Areas of endemism are evidence that components of bi- of areas of endemism (e.g.. Cracraft, olas (species) have become isolated and differen- tiated, and nested areas of endemism are evidence that biotas have expanded through dispersion to be- come more or less cosmopolitan, only to divide again. Yet, reconstructing this history has not been easy. It has become clear that most of these histor- ical patterns of distribution are so complex that no single method of analysis—at least not one that is currently known—is capable of giving a completely satisfying resolution. Each method of biogeographic analysis, it appears, has various shortcomings in how it handles widespread taxa, redundant distri- butions, and missing taxa and areas, in addition to the fact that the history of areas itself may not al- ways reflect a hierarchical (branching) pattern. THE SEVENTH GREAT QUESTION: HOW IS PHYLOGENETIC HISTORY PREDICTIVE? The third mission of systematics, as identified by SA2000, was to create an efficient, and predictive, systematic information system. This included da- tabasing specimens and making the information widely available, linking to other biodiversity and biological databases, and building informatics ca- pacity to utilize biological—and systematics— knowledge globally. The predictive element was en- visioned as coming from using phylogenetic classifications to guide searches for information. and thus reflect the hierarchical relationships of ife. The expectation that closely related taxa. share similarities not shared with more distant taxa is the foundation for comparative biology. An information system that is queried using the hierarchical rela- tionships of life can be termed phyloinformatics (the power of “phyloinformatics” was noted in an NSF- sponsored workshop [Tree of Life| and as well as by Edwards et al. (2000). Thus, the ability to search multiple databases using the nodes of a phyloge- netic tree may be the single most important contri- bution of systematics to conservation and sustain- able use of biodiversity. Searches that query across databases of various kinds from the perspective of phylogenetic groupings would therefore have im- mense predictive power because the resulting data can be expected to reflect attributes shared by, or potentially shared by, those groups. DISCUSSION: THE FUTURE OF SYSTEMATICS AND Irs RELEVANCE DIVERSITY A safe prediction is that debates over the species question will continue. However, I think there is far more actual agreement among practicing sys- tematists than could be concluded by recent papers supporting the BSC. This comment is not to dis- parage these viewpoints. Rather, it is to reaffirm the aput of others (Nixon & Wheeler, 1990; Wheel- „ 1997) that taxa, and taxic limits, as well as the — governing names, are the primary domain of practicing systematists. And adjudic ating taxic lim- its, moreover, is very much a practical process: as taxa are analyzed with the purpose of understand- ing patterns of character variation, geographic dis- tributions and endemism, relationships, and bio- geographic history, we will more and more need a species concept that looks at basal taxonomic units. That is the trajectory of research by young system- atists doing these kinds of studies, and that trend will certainly continue. But fighting over species limits is not the frontier of the theme of diversity—rather it is discovering and describing the other 95+% of life. Society needs to know what species share the planet with us, and the urgency has never been greater. Inven- torying life's diversity is a megascience question because of its enormous intellectual and technical challenges. Even understanding how to inventory a single taxonomic group in a circumscribed region is a difficult problem (e.g., Coddington et al., 1991), let alone thinking about what approaches and de- signs of inventories might be appropriate for inves- tigations that are global in scope (Wheeler, 1995a; Wheeler & Cracraft, 1996). A global inventory of life will therefore require meticulous long-term planning and capacity building in infrastructure and human resources, and it will be expensive. But it should be a shared expense of governments, in- tergovernmental institutions, and the private sector. Volume 89, Number 2 Cracraft 139 2002 Systematic Biology Indeed, the private sector is getting involved, and molecules alone are yielding the "true" tree of life. with a larger imagination than currently witnessed within most. governmental and. intergovernmental circles. Thus, a consortium of technology leaders with a track record in futurist thinking have rec- ognized the need for a global inventory of species and are beginning to organize a long-term effort to see the job accomplished [АП Species Inventory]. Because of this brilliant and far-sighted effort. in- ventory should take on new life! PHYLOGENETIC HISTORY In the previous section so much space was de- voted to phylogenetics and the Tree of Life because it is often not widely realized, even in the svstem- atic community, how crucially important that branch of systematics is becoming. Phylogenetic knowledge is exploding and the rate at which this is happening will not diminish for quite some time. Problems that undermine the quality of phyloge- netic research are commonplace and unfortunately will probably continue to persist. Many non-system- atically trained biologists, almost all using molec- ular data, have paid scant attention to taxonomic documentation, nomenclatural issues, or lo proper vouchering of sequences by reference to specimen data. Failure to heed these problems will lead to errors and even to potentially dangerous conclu- sions in fields like human health (Ruedas et al.. 2000). Systematists need to work with editors and editorial boards of journals, and with data deposi- tories like GenBank, to improve this situation. There is a tacit assumption on the part of many researchers building trees from molecular data that their results are inherently superior to those relying on morphological data. Results are presented, with little or no discussion, and often outright dismissal or total ignorance, of prior morphological studies. This attitude developed soon after molecular stud- ies were introduced (the “molecules versus mor- phology” debate; e.g.. Patterson, 1987). and as | said, it continues. This conflict is likely to increase as high through-put methods of genomics find. their way into laboratories doing comparative phylogenetics. The vast majority of current studies utilize small taxon samples and molecular data samples. leading to a significant accumulation of questionable results because of sampling artifacts. But this situation ts changing. As methods improve and more resources are allocated to molecular work, more studies are significantly expanding the sizes of matrices i (>) terms of both taxa and characters. Thus, | predict there will be an increasing tendency to think that But this would be a big mistake. All major groups of living taxa have many long-branched lineages that are monotypic or have relatively few closely related species, and these will in all likelihood con- tinue to confound analysis of deep-branch relation- ships. To resolve those relationships satisfactorily, it will almost certainly take the addition of mor- phological characters, especially those from fossil taxa. The relationships of many higher taxa, such as mammals and birds and many others, have been exceedingly difficult to resolve because of the prob- lems just mentioned, and our future hope of un- derstanding the history of life will lie in using all available data. Applied phylogenetics, as it might be termed, is an easily identifiable wave of the future. One of the most remarkable signs of the vitality of phyloge- netics is the expansion of its use into human health, developmental biology, forensics, natural resource management, and other areas. A large number of phylogenetically oriented young systematists are seeking careers in the biotechnology and genomics industries where understanding of phylogenetic methods, and a comparative approach to problem- solving, are needed. This trend will continue for a long time to come. BIOGEOGRAPHY Given the current state of knowledge of global biodiversity (5-10% known), and given that most of the species (insects) already described are known from only a handful of localities, it is fair to conclude that we have very imprecise knowledge of the distributions of Earth’s species (although sin- ele localities could be interpreted as being fairly precise). The task ahead is daunting because vir- tually all of Earth's habitats and ecosystems have been incompletely inventoried, even for well-known eroups such as birds (e.g.. Peterson et al.. 1998 and mammals (Patterson, 2000). Although an in- crease in our knowledge about distributions will ul- timately be linked to the intensity of inventory ef- forts, electronically capturing and correlating what we have already collected will surely increase our understanding of distributions and patterns of en- demism. Biodiversity cannot be managed or conserved without distributional information, thus the unde- niable importance of databasing the world’s natural history collections. Moreover, as global climate change accelerates, and as the anthropogenic con- version of habitats continues, predictions about the 140 Annals of the Missouri Botanical Garden Phyloinformatics: A Conceptual Framework Databases Specimens/Collections | Genetic/Genomic| | Ecological | | Geological Associated biological data f, n А dat Geo UCO | Developmental | | Land use | | | Geographic | Associated literature, etc | Human health | | Economic/trade | | Agricultural | N А Node-based Search D «“— Species/Taxon Nodes/hierarchy —» 2 Fig re ‘Lationships to drive queries among multiple database ure 6 d query for information for all species/taxa (/ on the [Tree of Life | held at Yale University in July 2 consequences of these changes become more and more important. Many of those predictions will flow from the use of historical distributional information linked to specimens. Biogeography is one of the great frontiers of systematic research. Factors such as widespread, redundant, and missing distributional data hin- der understanding of history, but these problems themselves suggest that the field is still wide open to theoretical, biotic methodological, and empirical research. Knowledge of biogeographic history is so central to understanding patterns and processes of biological diversification, cluding speciation, in- as well to how biotas evolve over time, that it will continue to be a core as The concept of a phyloinformatic search strategy uses the hierarchy implied by knowledge of phylogenetic tead of unde rtaking searches one species at a time, nodes earches for ялы тее by related species. Here, searches on node (or line 'age) с ) — that node. This figure was first used in a NSF works shop area of research for many years to come. There have been very few studies of biotic history that integrate patterns of relationships among areas of endemism with information from paleontology, paleoclimatology, and paleogeography. There is an unfortunate disjunct, on the one hand, be- tween paleontologists who study diversification over time and rarely concern themselves with present-day patterns and processes such as spe- ciation, and those neontologists who look at pat- terns of historical biogeography using distribu- tions of the Recent but who rarely incorporate paleontological data on changing di- versity. A bridge will have to be built before we have a satisfactory picture of biotic history. biota Volume 89, Number 2 2002 Cracraft Systematic Biology PHY LOINFORMATICS | predict that phylogenetics will have its greatest societal impact by empowering and enriching the search for information and data associations across many different kinds of biological and systematic databases. Information flow will make peoples’ lives better. That is what phyloinformatics can, and will, do. While queries of biodiversity databases will al- ways make use of species” names as pointers lo information (Bisby, 2000). queries can expand and integrate searches and in- the use of node-based formation to another order of magnitude (Fig. 6). classifications will facilitate. а Phylogenetic new way of gathering biological information and linking it to other nonbiological databases. The establish- ment of the Global Biodiversity Information Facility (СВІ: see 2000) on 1 March 2001 holds the key to making biodiversity infor- Edwards et al.. mation readily accessible to all. Phyloinformatic queries will expand the potential of GBIF in ways not previously imagined. The predictive power of phyloinformatics. based as it is on an understanding of the relationships among Earth’s species, argues persuasively for dis- covering all branches of the Tree of Life as rapidly as possible. Literature Cited Agenda Systematik 2000. 1996. Agenda Sy stematik 2000 Erschlie ssung der Biosphire. Kleine Senckenberg- Reihe Nr. 22: 1-55. American Museum of Natural History. Taxonomy Initia Using Jeet Country iad Renomi Nee 1999, The Global stematic Inventories to . Center for Biodi- cuin an Мил seum of Natural ps e also (http://researc ^. amnh.org/ i.html)] ve rsity and C — \ Anonymous. 1991. Sy ste matics s Agenda 2000: Integrating biologic al diversity and societal needs. Syst. Zool. 10: 520-52: Avise, J. . M. Ball. 1990. Principles of genealogic al cone — e in species conce pts and biulosic al taxon- Д у. Evol. Biol. 7: 45-67. hus P. 1999, i ia lip service ar action? Courier Forschunginst. Set enbere 215: —From Convention pm . pois: ethnobotany for the new ا‎ Ann. Missot i Bot. Gard. 83: 58—00. n Fles 1996. Species. The problem of units of man- -25 sness. races: agement in conservation. Pp. 247 4 in D. G. Klei- man, M. Alen, V. >. Lumpkin & H. Harris (e ditam. Wild Mammals in € ap tivity: Principles and Тес — s. Univ. Beresford, Cracraft. 1999, Speaistion in African orest — (Stindroruis) Species limits, phylogenetic Thompson, 5 Chicago Press, Chicz ago. re а and molecular biogeography. Amer. Mus. 22. Novit. 3270: 1— Bisby. F. А. 2000. The quiet т — in- formatics and the internet. Sci » 289: 23 ке S. р. Cutler. 199 o. dr henna The Challenge for Europe. Linn. Soc. Occas. London. wa No. 1, | Bock. W. J. 1987. рене ies concepts, speciation, mac- evolution Pp. 31- 1D. Iwatsuki, P. Raven №, J. Bo k fedi Moden "n о of Species. Ü niv. To- kvo Press. Tok » Bojorquez- "Tu a. P. Balvanera & A. D. Cuaron. 1991. iological inventories and computer databases: Their role in environmental assessments. Environm. Mana- vem. 18: 545-551. d p. a D cLennan, J. M. ‚ 1995. Systematics, ecology. Carpenter, 5. 6. a id De — "Bias ience 15 687-095. Cameron, C. B., J. R. Garey & B. J. Swalla. 2000. Evo- lution of the chordate body plan: New insights from phylogenetic analyses of deuterostome phyla. Proc. Natl. Acad. Sei. U.S.A. 97: 4409-4474 hua. K. B.. W. J. Bellini, P. A. Rota. B. H. Harcourt. A. lamin. 5. K. Lam, T. G. Ksiazek, P. E. Rollin, S. R. aki. W.-J. Shieh, C. S. — D. J. wes JT Roehrig. B. Eaton, A. R. Gould, J. Olson, Field, P. Daniels, A. E. Ling, C. J. T Leds our & B. 2000. Nipah А rec ently | emergent de dh param) xovirus. Science - 288; 432-14. Claridge. F. 1995. Introducing eat s 2000. ge Badio sity & Conservation 4: 451-454, HA. Dawah 8 Wilson (editors). 1997. Species: The Units of Diversity. Chapman & Hall, Lon- don — J. ^.. E. Griswold, aranda & S. F. I w d 199]. sampling air ie to estimate ег in tropical ) in E. C. d (editor), The Unity of E — Biology, J . Dioscorides Press, Portland. ( regon. а A. Orr & D. J. Futuyma. need a new species Cracraft. J. 1983. Species concepts and speciation anal- Ornithology 1: 159-187. 36. Origin and t. Men of continental biotas: D. S. Dávila, E. Pen- Designing г testing ecosystems. Pp. 41-6 1988. Do we 190—200. Coyne, J. oncept? Syst. Zool. 37: ysis. Curr. . 19% Speciation and historical congruence within the Austra- lian m E volution 40: 977 ). 1987. Species concepts Ме: the ontology of evo- ‘ition: Biol, : Philos. 2: 329-346. 89a. Speciation and its ontology: Т he empirical consequences of alternative species concepts for un- de — patterns Ke processes of differentiation. —59 in D. Otte & J. A. Endler (editors). Specia- tion dd its Consequences. Sinauer, Mas- Sunderland, sachusetts 198 0b. Species as entities of biological theory. Pp. 31-52 in M. Ruse (editor), What the Philosophy of Kluwer, Dordrecht Systematics, s Abstract, 5th I Meeting, Soci- 'onsin. 995. The urgency of building global capacity Biology Is. А 199] ecies concepts, and con- servation biology. ety of Conse rvation Biology, Madison. Wis | [ог — science. Biodiversity & Conservation 4: 163—475. 1997. Species concepts in systematics and con- servation ME An — 'al 325—339 in M. A. Claridge, Н. A. Dawah & Wilson (editors), Species: The т nits s of Diversity. Chap- man & Hall, viewpoint. Pp. M. R. London. 142 Annals of the Missouri Botanical Garden 000. Species concepts in theoretical and ap- plied — A systematic debate with consequences. Pp п Q. D. Wheeler & R. Meier (editors), Spe- cies C once m and Phylogene tic Theory: A Debate. Co- lumbia Univ. Press, New Yor Crandall, — 1999, — Hopkins L niv. Press, Balti 295. Systematics aid ibli health. Bio- 14. > — Te ны of HIV. . 1992. Archaea in costal marine environ- . Natl. Acad. Sci. U.S.A. 89: 5685-5689. & W. S. Alverson. 2000. A new age of -Momon Bot. — 87: 110-126. . А. Lane & E. S. Nielsen. 2000. Inter- ope — ol — data — mation on every desktop. Science 289: 2312-2314. Eisen, J. A. 1998. Page s: Improving functional predictions for unc — d analysis. 2 ıe Res. 3-167 З M. (editor). 1992. The Units of Evolution: Es- sa) Nature of Species. MIT Press, Boston. Eshbaugh, W. H. 1995. Systematics Agenda 2000: An 2g toric ‘al perspective. Biodive sity & Conservation 4: 455 m - X = = 5 =ч Biodiversity infor- genes by evolutionary Mis P. L., C. J. Humphries & R. I. Vane-Wright. 1994, Syste ipm s and Conservation Evaluation. Oxford Univ. Press, Oxford. Funk, V. A. 1997. Using collections data and GIS to ex- amine biodiversity information levels in Guyana. Pp. 17-128 in К. К. Hoaglund & A. Y. Rossman (editors), Global Genetic Resources: Access, Ownership, and In- tellectual Property — баке. "lation Systematics Col- lections, — D. Futuyma, D. J. 1998. Evolutionary Biology, auer, ж a Massachusetts. Geist, je np species and the law. Nature Third Ed. Sin- . "uos, V. V. Dung, E. D. W — ato E Arctander & J. R. MacKinnon. 1998. De- sc иса of Muntiacus id ee из, a new species of muntjac (Artiodactyla: Muntiacidae) from central Viet- ат, and implications pu conservation. servation 1: 61-68. Graham, L. E., M. E. Coo Animal Con- & J. S. Busse. 2000. The origin of plants: Body plan — contributing to а — evolutionary radiation. Proc. Natl. Acad. Sci. U.S ii 40. Groves, P. 2000. The genus Cheirogaleus: Unrecog- nized bc in dwarf lemurs. Int. J. Primatol. 21: 943—902. Harvey, P. H., „ Brown, J. M. Smith € S. Nee. 1996. Ne ^w Uses 8 Ne W — nies. Oxford Univ. Press, Ox- Holmes, E. C., P. L. Bollyky, S. Nee, A. Rambaut, € Garnett & P. — 1996. Using phylogenetic trees to rec т the history of infectious disease epidemics. Pp. 169-186 in P. Н. Harvey, A. J. L. Brown, J datus j^ S. Nee (editors), New Uses for New Phyloge- nies. Oxford Univ. Press, Oxfor Horikoshi, К. & K. Tsujii (editors). 1999, He ши in Deep-Sea sta nts. Springer-Verlag Howard, D. J. & . Berlocher p ms tm Endle 88 ا‎ Species — Speciation. Oxford Univ. Press, w York. ы Т. 1995. Down in the woods the y have no names— BioNET-INTERNATIONAL. р) — o = Strengthe ning systematics in developing countries. Biodiversity & Conservation 4: 501—509. Jousson, O., J. Pawlowski, F. Dini, G. Di Guiseppe, — 200 ). Invasive co Z. Zaninetti, F. W. Zechman, R. Woodfield, A. Millar & A alga reaches California. Nature — . B. Martin (editors м 1993. Ѕресіеѕ, Species С one vind and Primate Evolution. Plenum Press, * ork. Lacoste, Mauc wu — ul, " Lewis, M.-C. — & A. . 2000. KSHV-like her- pesviruses in chimps т “gorillas. Nature 407: 152. Lanciotti, R. S., J. T. Roehrig, V. Ds ubel, J. Smith, M. Parker, K. Steele, B. Crise, К. E 3. Crabtree, H. Se therret, R. Cropp, B. Panigrahy, E inson, C. 151- Ostland, B. Schmitt, м. Malk- J. Weissman, N. Komar, H vage, nias ا‎ for an — of «de T. a Vl ‚ B. Huey, R. K. Monson & R. J. Jensen. 05. ata s and the study of organismal form and — BioScience 45: 696—704. — J. G., M. Kottelat, G. R. Smith, M. L. J. Stiass- & A. C. Gill. 2000. So many fishes, so little time: (dip overview of recent ichthyological disc overy in con- tinental waters. Ann. Missouri Bot. Gard. 87: 26-65 Mackinnon, J. 2000. New mammals in the 21st century? Ann. Missouri Bot. Gard. 87: 63-66 Mayr. E. 1942 ystematics and the Origin of Species. Columbia — Vork. Press, New Yor ————. 1963. Animal Species and Evolution. Harvard Univ. n ress, Cambridge. Massachusetts. — ———. 1982. A local — and the biological species Amer. J. Bot. 79; 222-238. cone n — ———. 2000. The lone s spec ies concept. Pp .17-29 үз Wheeler & R. Meier (editors), Specie s Con- ea and Phylogenetic Theory: A Debate. Columbia Univ. Press, New York. McNeely, J. A. 1995. Keep all the pieces: Systematics 2000 and world conservation. Biodiversity & Conser- 9. e 1995. Systematics, bio- ч udis 680—680. K. . BioScience 45: € The u use of phylogenetic perspect in compara — puso and ‘velopment | bi- ology. Ann. Missouri Bot. Gard. 83: 3-16. Moritz, C. 1994a. aa of е 'hondrial DNA analysis i in conservation: A critical review. Molec. Ecol. : 401411. . 1994b. Defining rge — ant units” for conservation. Trends Ecol. Evol. 9: 373-375. — ———. 1995. Uses of molecular ph plis ا‎ for ¢ conser- vation. Philos. Trans. Roy. q^ 349B: 113-118 Nelson, G. & N. I. Platnick. 1 'stematics ind Bio- geography: C ladistics and Vic 'ariance. Columbia Univ. Press, New Yo 1999 . Systematics and conservation. Pp. 13 in W. Ponder € D. Lunney (editors), The Oth- er 99%, The em is Biodiversity of Inverte- brates. Trans. Roy. Zoo . New South Wales, Mos- man, New Pis Wales. & з. West. 1994. Biodiversity research ge bi- — ‘al ea Transfer of information. Pp. 121 in P. L. Forey 01— „С. J. Humphries & R. I. Mte Volume 89, Number 2 2002 Cracraft 143 Systematic Biology (editors), Systematics and Conservation Evaluation. Clarendon Pres 288, E ford. Nixon, K. € 2 poten species concept. Cladistics oon T. T Н.. Jr. 1996. Importance of systematics to public health: Tic : microbes, anc Bot. Gard. 83: 37—40. disease. Ann. Missouri Ou, C.-Y.. C. ri Ciesielski. G. Myers, C. I. Bandea. C.- C. Luo, B. T. M. Korber, J. 1 . Mullins, G. Schochetman. R. L. Berkelman L. A. N. Fconomou, + J. Witte, L. J. Furman, G. A. Satten, Mac s, J. W. E urran & І Jaffe, Laboratory ов iain. group, Epidemio- 2 logic Investigation group. 1992. Molecular epidemiolo- ey of HIV Ои in a die practice. 256: 1165-1 Pace, N. R. ard and the — re. Patterson, B. D. ery of new Noti ‘al mammals. Diversity and Distri- butions 6: Patterson, € Science n e scular view of microbial diversity ence 276: (34—740. ce rns and trends in the discov- 15-151. — 1987. Molecules and ub d in E лю Conflict or Compromise? Cambridge Uni Press, C ambridge. Peterson, A. Navarro-Sigiienza & H. Diaz. 1998. The — for continued scientifie collecting: \ geographic analysis of Mexican bird specimens. Ibis 140: 288—292. Peterson, K. J. & E. H. — 2000. Regulatory évo- lution and the origin of the bilaterians. Proc. Natl. Acad. E U.S.A. 97: — 33. Pimentel. D.. L. Lach. R. Zuniga & 2000. iim: ntal and economic costs p nonindigenous the United States. BioScience 50: 33-05 Pollock. "р, Э ЧЕ: . P. Cum- mings. Sol A case for evolutionary genomics and the со ало е of sequence diversity. Mo- iol. vol. : 1776-178 Ө. Prance, G. т E — M velopment. 490— Benitez- . Morrison. — Eisen, N. А. э. Syste matics, conservation and sus- Biodiversity & Conservation 4: $^ Beentje, J. Dransfield & R. Johns. 2000. The tropic al es remains undercollected. Ann. Missouri Bot. Gard. 87: rae Raff, R. А. 1006 tthe Shape of Life. Univ. Chicago Press. thicago. Ganzhorn. — * — . M. Goodman & J. U. 2000. Taxonomic revision of mouse lemurs ( os Me in the western portions of Madagascar. Int. J. Primatol. 21: 963-1019 Richardson. P. M. 1996. тары Introduction. Ann. Missouri Bot. l- Agenda 2000 Gard. 83 The Systematics 2, 2000. Our unknown planet: Recent discoveries and the future. Introduction. Ann. Missouri Bot. Gard. 87: 1-2. Rosen, D. E. 1978. Vicariant patterns and prb “al ex- planation in — Syst. Zool. 27: 159-188. Rossman, A. Y. & D. R. Miller. 1996. — matics solves — ms in ape ulture wal forestry. Ann. Missouri Bot. Gar d. х 17 Ruedas A.. T — zar-Bravo, J. W. Dragoo & T. Yales. od The importance of being earnest: W e anything, constitutes a de à "specimen examine ad?” 29-132. Ryder, O. A. 1986. Species conservation and systematics: . Wheeler. 1990. An —— = - The dilemma of subspecies. Trends Ecol. Evol. 10. Savage, J. M. )5. pou and the biodiversity cri- sis. BioScience 45: 673-679. Simpson, B. B. & J. i 1995. Systematics: The sci- ence of MICE BioScience 45: 670-072. Soberón, J.. J. B. Llorente & L. Onate. 2000. The use of specimen- — databases for conservation purposes: An example using Mexican M and pierid basher flies. Hives & 41 E. 1999, The D T global biodive rsilv 3-32 í Conservation 9: 1441—1466. and its ie E Pp. Cracraft & F. T. Grifo . The Living Blasi in € risis: Biodiversity Sci- . Press. New Yor Systematics Age nda 2000. 1994a. Systematics Agenda 2000: Charting ш Biosphere. Systematics Agenda 2000. Society of Systematic Biologists. American Soci- ety of Plant к a ciation of Systematics р. tions, New Yor 1994b. Systematics Agenda 2000: Charting the Biosphere. Technical Report Systematics Agenda 2000. Society of Systematic Plucis, American Soci- ety of Plant Taxonomists, Willi Hennig Society, Asso- ciation of Systematics C s sctions, New York. | Reprint- Willi Hennig Society, Asso- К. ed. with corrections, 5.] Systématique Agenda 2000. 1994. sphere. Société Frangaise de 8 А2000 — report, French клк Thalmann, U. T. Geissmann. 2000. КИ and variation in the western woolly lemur (Avahi species — uni- — levé de la Bio- sté са Paris geographic decida ntalis) with en of a new color). Int. J. ges 21: 915-941. Vane-Wright. R. I. 1996. Systematic s and the conse وا‎ of biological diversity. 47 Ann. Missouri Bot. Gard. § 51. Vecchione. M. & B. B. Collette. 1996. Fisheries agencies m marine biodiversity. Ann. Missouri Bot. Gard. 83: 1994. Conservation Biol. 8: 351— Diagnosing че of ›. — A. P. «€ R. DeSalle. consery ation management. W he e dl т, Q. D. 1905a. Systematics, the scientific basis for inventories of biodiversity. Biodiversity & Conservation 1: 176— 189. . 1995b. Systematics and biodiversity: BioScience Su uppl. 19 — of taxonomy in gene etic resource Pp. 59-70 in K. E. Hoagland & A. Y Rossman (editors), Global Genetic Resour Ownership, and E re Property Rights. Associa- Washington, Policies at higher leve ‘ls. Ө. 9‹ management. Acce ә. — — tion x Syste matics Collec & J. Cracraft. 1996. amó preparedness: Are we — to meet the biodiversity — Pp. 435- H6 in E. О. Wilson. M. Kudla-Reaka & (editors). Biodiversity H: Understanding and ba ting our a al Resources. National Academy of Scienc- ES Wilson Press. Washington. rar n E Meier. 2000. Species Concepts and Phylo- genetic Theory: A ا‎ Columbia Univ. Press. New Yor & N. I. Platnick. 2000. The phylogenetic spec les concept (sensu Wheeler and Platnic ae 55-09 in Q. D. Wheeler & R. Meier (editors). s Concepts and dii snetic Theory: A Debate. Сайыш Univ. Press, w York “у, R. A. 1999. Species: New Interdisciplinary Es- savs. MIT Press. Boston 144 Annals of the Missouri Botanical Garden Wolfenbarger, L. L. & P. R. Phifer. 2000. The ecological risks and benefits of + tically engineered plants. Sci- ence 290: 2088-206 Internet Resources [All Species Inventory |: http://www.all-species.org/ [DIVERSITAS]: http://www.icsu.org/diversitas/ |G BI К]: htty // rg/ [GenBank]: htt ыызы: ne ebi. nlm.nih.gov/ [С i Global Taxonomy Initiative: http://www. biodiv.org/ E ht ml »ec-tax/gti/ [ Dus BASE |: ip ihren harvard.edu/treebase [ pis of Life] NSF _ Workshop — esearch.amnh 1 nl O а TAXONOMY AND HERBARIA IN SERVICE OF PLANT CONSERVATION: LESSONS FROM MADAGASCAR’S ENDEMIC FAMILIES! George E. Schatz? ABSTRACT Our actions during the next two decades will largely determine how many of the world’s ca. fundamental data that define both the t vascular plants will survive for future generations. The within which species are cireumseribe d and delimited ч related 5 These herbaria, and the taxonomists who of those species, reside in the € s ca. 3000 herbaria. and must play a critical role i project that has focused on the species in М of reviewing, and then revising when necessary, the exist referencing of primary occurrence data га facilitated simple entifying as rapidly as possible those species 300.000 species of taxonomic frameworks the geographical — work 1 the m, can )ecies, as well as GIS analvses of Exter labs ence in ш. estimates of the number of * ‘subpopulations, species in the endemic families re tance: both those already incorporated in the protected areas system, Partnerships among the world’s herbaria can ef outside of the protection network. — and thei AUT presence assessment of extine tion risk. In s NA perse ne mapping of Pall assessment of the most threatened vascular plant species by focussing on taxa endemic at politic ‘al and regional (e.g., Hotspot) scales. The synthesis and analysis of the primary data house ai in the world’s record of plant life on planning. and thereby minimizing the loss of plant diversi ey words: Earth—constitutes the most effec herbaria—our only incontestab da tive and robust means of directly informing conservation conservation, natural history collections, primary occurrence data, taxonomy. As the term “biodiversity” has entered the in- ternational lexicon of the environment during the past decade, one of the most fundamental aspects of the concept has often been neglected: principal knowledge of biological diversity emanates from the study of natural history collections by taxonomists. Plant taxonomists utilize herbarium specimens to егесі taxonomic frameworks within which species are defined, circumscribed, and delimited from re- — ated species. In addition to portraying the limits of variability of individual taxa, those herbarium specimens representative of a given taxon also con- stitute primary occurrence data that collectively de- pict the verifiable (i.e... vouchered) geographical and the the distribution of the taxon. Thus, herbaria, taxonomists who examine, classify, and curate primary specimen data that reside in he rbaria. fur- "what," “where,” and "when." initially necessary to document plant life on Earth. As such. nish the * they have a unique and critical role to play in glob- al efforts to mitigate the loss of biodiversity. Recent studies of the families of vascular plants endemic 2000a) demonstrate that central to such a role is the revisiting, reattri- to Madagascar (Schatz et al., buling, and synthesis of the primary data itself. Tens of thousands of plant species have already been listed as threatened with extinction. The 7997 IUCN Red List of Threatened Plants (Walter & Gil- let, 1998). The World List of Threatened Trees (Old- field et al.. 1998), and their to-be-revised-vearly electronic Internet successor, 2000 [UCN Red List of Threatened Species (see [IUCN]). — serve as sober- с Birkinshsw, F. Magil 'T thank N Andrianjafy, ( vrl in the k ya mic F Ke Pojat R. IL and an anonymous reviewer > Missouri - Batani cal Gar t Piscic КОНЕ tesies extended by the Government of — ar (Direct tion Générale de Protégée mal Science пие grants DEB-9024749 and DEB- 062707 2 and from the John D. and Catherine z Claiborne and Art Ortenberg — LWO Inc . Louis. Missouri 63166-0299, U.S.A. george.se hats @шороп: Git the Association Nationale pour la Gestion des Aires 1 Nati Foundation, the 2 Missouri — “al Garden, P.O. Box 299, ANN. Missouni Bor. Randriantafika, | for generating е г for helpful comments on the — ript. F y vork was and the Pare Botanique et Zoologique de Tsimbazaza and the аот ану for 6x ir partic- ^ P. Lowry — and Y S statistics: and V. Hollowell. bo бг — eder Antananarivo, Madagascar. We gratefully acknowledge cour- a Gestion des Ressources Forestiéres) and by This research was conducted with — from U.S. . Mac Arthur . and the National Geographic Society GARD. 89: 145-152. 2002. 146 Annals of the Missouri Botanical Garden Table 1 Malagasy endemic plant families: Revised (since 1998) taxonomic frameworks. No. of Newly ewly Placed into species recognized described synonymy Asteropeiaceae 8 3 (38%) 2 (25%) — Didymelaceae 2 — Kaliphoraceae l = — Melanophyllaceae 6 | (17%) 1 (17%) 4 Physenaceae 2 — Sarcolaenaceae 60 25 (42%) 23 (38%) — Sphaerosepalaceae 18 2 (11%) — — Tota 97 31 (32%) 20 (27%) 4 ing baselines on the state of the Earth’s plant re- sources. As commendable as these efforts are, the fact remains that the majority of plant species have not yet been evaluated with regard to their risk of extinction. As would be expected, the lists are heavily biased in their representation to the com- paratively well-known floras of Europe and North America, and to a lesser degree Australia and South Africa, or to specific taxonomic groups (e.g.. conifers, palms) or life forms (e.g., succulents) pop- ular in cultivation. One suspects therefore that the number of threatened plant species identified thus far is only the tip of the iceberg. In the absence of an exhaustive survey of all ca. 300,000 (or more? species of vascular plants (Prance et al., 2000), a strategy to identify and safeguard those remaining species most at risk of extinction must be imple- mented as rapidly as possible. How can the daily work of herbaria and taxonomists contribute to such a strategy and thereby directly inform conservation — planning? CONSERVATION ASSESSMENTS BASED ON OLD TAXONOMY WILL BE INADEQUATE AND MISLEADING, BECAUSE THEY FAIL TO INCLUDE RECENT PRIMARY DATA: REVISITING THE “WHAT” The goal of the Endemic Families of Madagascar Project has been to provide current assessments of the conservation status of all species in the seven families of vascular plants endemic to Madagascar and the Comoro Islands (Schatz et al., 2000a). An essential first step in that process was the reeval- uation of existing taxonomic frameworks for each of the families, four of which were originally sub- sumed within non-endemic families: Asteropei- aceae within Theaceae (Perrier de la Báthie, 1951): Kaliphoraceae Melanophyllaceae within Cornaceae (Keraudren, 1958); within Capparaceae or Flacourtiaceae (Perrier de a Bâthie, 1946). The available taxonomy prior to our review dated on average to 1952 (ranging from and and Physenaceae — 1937 to 1963). Reevaluation of existing taxonomic frameworks entailed examination of all primary specimen data. In the case of three of the fami- lies—Didymelaceae (Leandri, 1937), Kaliphora- ceae, and Physenaceae—together comprising just five species, the most recent frameworks remain valid. But in the remaining four families—Astero- peiaceae, Melanophyllaceae, Sarcolaenaceae (Ca- vaco, 1952). and Sphaerosepalaceae (Capuron, 1963)—review of all primary data necessitated their taxonomic revision (Randrianasolo & Miller, 1999; Schatz et al., 1998, 1999a, b, 2000b, 2001: Lowry et al., 1999, 2000). In all but one instance, a synoptic format for the revisions was adopted for the sake o the taxonomic framework was illuminated through new identifi- cation keys emphasizing diagnostic features, and full synonymy for each taxon was given, but only the newly described species were provided with full expediency: revised descriptions. These revised taxonomic frameworks include 31 newly recognized species (or 34% of the 92 total species now enumerated in those four fam- ilies), 26 of which were newly described; 4 formerly recognized species and a number of infraspecies were placed into synonymy (Table 1). Primary data that had been collected since publication of the former frameworks accounted for 19 (73%) of the 26 newly described species. In the remaining 12 newly described or recognized species, specimens ascribed to other species in the previous taxonomy were judged to represent clearly discernible, dis- tinct taxa based upon morphological and ecogeo- graphic criteria. Conversely, in the cases of species and infraspecies now placed into synonymy, for- merly cited specimens, in conjunction with newly available material, revealed taxa that could no lon- ger be distinguished from earlier described species. Clearly, assessments of conservation status based on the taxonomies that existed prior to our reeval- uation would have been highly flawed and mislead- ing, and would have failed entirely to account for Volume 89, Number 2 2002 chatz 147 Plant Conservation one third of the species now recognized in the sev- en endemic families. CONSERVATION IS ALL ABOUT GEOGRAPHY, BUT Most OF PLANT DIVERSITY 15 Nor GEO- REFERENCED: REATTRIBUTING THE “WHERE” Mapping the primary occurrence data upon which a taxonomic framework stands has long been a preoccupation of taxonomists: most revisions in- clude distribution maps of the taxa. Distribution mapping is usually a tedious task insofar as the majority of primary data collected prior to the latter decades of the 20th century are only imprecisely located in relation to population centers or phys- iographie features of the landscape, but lack the precise geographical coordinates that can now be assigned so easily in the field with Global Position- ing System (GPS) technology. As a consequence, most of plant diversity must be post facto geo-ref- erenced in order to capitalize on the current revo- lution in computer mapping and spatial analysis brought about through Geographical Information Systems (GIS data representing the Malagasy endemic families tools. Geo-referencing of ~ primary has been greatly facilitated by the compilation of a Gazetteer to Malagasy Botanical Collecting Locali- ties [Schatz & Lescot, 2001]. Some historical spec- imens that lack coordinates may simply be impos- sible to geo-reference to a seale that has meaning for conservation. In the majority of cases, however, geo-referencing can be achieved to obtain a Mini- mum Mapping Unit of plus or minus one minute of latitude/longitude, more than sufficient for the coarse area estimates used in conservation assess- ments. Even in cases of uncertainty involving col- lections known to originate from within protected areas in Madagascar, but without any greater pre- cision, e.g., the Réserves Naturelles or RN collection series, mapping to the centroid of the protected area polygon still constitutes a valuable reattribu- tion that can be included in area and subpopulation estimates. Of even greater importance is that col- lections in the RN series document historical pres- ence within the protected areas networ With works in place that are based on the totality of new and reconfirmed taxonomie frame- = available primary data, the next step in assessing the conservation status of each species entails anal- vsis of its geography, including the geography of current land cover and threat. Conservation ulti- mately depends on sufficient space to maintain a viable population, the size of which will vary con- siderably from one species to another depending upon life histories and habitat requirements. Al- though ideally one would want to conduct a long- term Population Viability Analysis (PVA) to deter- mine exactly what constitutes “sufficient space" for each species, area measurements calculated di- rectly with GIS tools based solely on the primary occurrence data can provide a rapid first estimate of a species’ vulnerability to threat, and therefore a means of prioritizing concern. Among the as- sumptions inherent in such an initial evaluation are: (1) that the known collections and localities for a given species are a valid reflection of its abun- dance and distribution; (2) that widespread and common species will be at lower risk than restricted and rare species; and (3) that species occurring within protected areas will be at lower risk than species that occur only outside of protected areas. The IUCN Red List Categories (IUCN, 1994; IUCN/ SSC Criteria Review Working Group, 1999) serve as a guideline for incorporating the area measure- ments of Extent of Occurrence and Area of Occu- pancy into a hierarchical delineation of extinction risk. For example, among the species in the endem- ic Malagasy families, newly described Melanophyt- la modestet G. E Schatz, Lowry & . Wolf (Me- lanophyllaceae) (Schatz et al., 1999). is classified as Critically Endangered (CH) given an Extent of Occurrence less than 100 km’, an Area of Occu- pancy less than 10 km?, and a single known pop- ulation (Schatz et al., 2000a). On the other hand. \steropeia densiflora Baker is classified as Vulner- able (VU) due to an Extent of Occurrence less than 20.000 km? (Fig. 1), an Area of Occupancy less than 2000 km? (Fig. 2), and less than 10 known “subpopulations” (the number of non-contiguous occupied cells or cell clusters within the 10 km X 10 km grid utilized to estimate the Area of Occu- рапсу) (Fig. 2) (Schatz et al., 2000a). In addition to facilitating rapid area measure- ments that can be incorporated into evaluations of extinction risk, reattributing primary data with geo- graphical coordinates also allows the modeling of potential distribution (Skov, 2000; see also [BIODI] and [Species Analyst]. Utilizing the spatial analyt- ical functions of GIS, an envelope of independent physical and environmental variables associated with a set of primary occurrence data points can be described, and in so doing define an environmental niche, and hence, a potential distribution that cor- responds to the range of documented heterogeneity. When compared against current land cover as re- vealed from satellite imagery, remaining viable habitat coinciding with (preferably recently) docu- mented distribution can be identified. In the ab- sence of extant viable habitat that intersects the documented distribution. 1.e.. when all historical 148 Annals of the Missouri Botanical Garden Figure ent of Occurrence (= 17.830 km?) of Asteropeia pie чөн, Baker in Madagascar. = primary data originate from areas now converted to urban or agricultural uses, conservation hopes for a species may still lie within areas of potential dis- tribution and remaining viable habitat. Conservation in such places as Madagascar may ultimately depend upon encompassing a species within some type of protected area. An overlay of protected areas polygons in relation to primary oc- "gap" tection exists. Among the 97 species in the seven currence data instantly reveals if a in pro- families endemic to Madagascar, at least 28 species are not presently recorded from protected areas. The situation is potentially far bleaker for the spe- 75% of the 4011 endemic species are not yet documented in protected areas (Valencia et al., 2000). Similarly, the 1995 revised ROTAP (Rare or Threatened Aus- tralian Plants) listing (Briggs & 1996) re- veals that 47% of the 5031 listed taxa are not doc- cies endemic to Ecuador, where over Leigh, = umented from protected areas. If one conservation is to ensure that the maximum number of species is included within the protected areas network, dictate that centers of endemism, then. considerations. of. complementarity L.e., concentra- tions of co-occurring local endemic species, be identified (Williams, 1999). multaneous mapping of all species in the endemic For Madagascar, si- families revealed, for example, that the Ibity/Itremo goal o Figure 2. sis of * al y eer (7 individual cells + 2 cell « steropeia densiflora Baker in — ar as 10 km grid. Area of Occupancy (= 1800 km? ) € num- ‘lus- ers = 9) of / de fined by a 10 km X Massifs on the Central High Plateau, which are geologically characterized by a complex mosaic of granitic, marble, and quartzite substrates, consti- tute an important center of endemism that is not currently encompassed within the protected areas network. Of the ten endemic families’ species re- corded Ibity/Itremo, five (all Sarcolaena- ceae)—Leptolaena diospyroidea (Baill.) Pentachlaena latifolia H. Perrier, Perrierodendron — еи -F. Leroy, Lowry, Haevermans, Labat & . Schatz, Schizolaena microphylla H. Perrier, from Cavaco, anda an quede species of Xerochlamys/Sarco- are essentially restricted to the Massifs, laena and are therefore entirely lacking protection. The presence of numerous other local endemics in the Ibity/Itremo Massifs, gumes (Du Puy & Moat, including a number of le- 1998), forcefully argues for the immediate establishment of new protected areas in the region. Simultaneous mapping of all species in the endemic families also revealed the importance and management needs of existing pro- tected areas. In particular, the small Réserve Na- turelle Integrale of Betampona (2228 ha) shelters 20 species in the endemic families, including two newly described as a result of revised taxonomies within Sarcolaenaceae (Pentachlaena betamponen- sis Lowry, E. Schatz and Rhodolaena leroyana С. E. Schatz, Lowry & А.-К. Wolf) that are known only from this reserve. Map- Haevermans, Labat & G. Volume 89, Number 2 2002 Schatz Plant Conservation 149 ping reattributed primary occurrence data of re- stricted range species, i.e., country endemics. о species endemic to. Ecoregions (Olson & Dinner- stein, 2002 [this volume]) or Hotspots (Mvers et al.. 2000). robust. means of directly informing conservation may well constitute the most efficient and planning. Much or PLANT DIVERSITY IS “RARE.” LE. DOCUMENTED BY VERY FEW COLLECTIONS, AND HERBARIA AND THEIR ASSOCIATED DATA ARE IDIOSYNCRATIC: STRATEGIES AND NEW STRUCTURES FOR RAPIDLY ACHIEVING THE SYNTHESIS OF REATTRIBUTED PRIMARY DATA Faced with the very real prospect of losing a sig- 300.000 or more species of vascular plants during the coming nificant proportion of the estimated decades. it is incumbent upon the systematics com- munity lo synthesize the most relevant primary data. and disseminate that data to governments and the conservation community. By definition, such a synthesis must go beyond Red Lists per se as they are currently envisioned by IUCN (see [IUCN |). Al- though Red Lists attempt to draw attention to the species most at risk of extinction, because they lack the underlying geo-referenced primary data. thes have only very limited utility for conservation plan- ning. Nevertheless, Red Lists in their current form. along with country and. regional checklists. can help guide the prioritization of primary data syn- 25% of Ecu- adoran species known only from the type collection, thesis. By identifying the greater than and the greater than 50% known from only one or two populations, and conversely, the less than 10% that are “common” (known from ten or more pop- ulations), the Ecuador Red List of Endemic Plants Valencia et al.. 2000) serves to prioritize which species should be subject to comprehensive. pri- ROTAP lists nearly 200 Australian species known only from the type 1990). prehensive synthesis of all vascular plant species mary data synthesis. Similarly. collection (Briggs & Leigh. Surely, a com- known onlv from their type collection. and the geo- referencing of as many as possible, must be one of the very first priorities. Species known only from a single (tvpe) collec- tion are merely an extreme case of endemism. Within the context of the Convention on Biological Diversity (see [CBD]). governments are obligated to pay particular attention to all those species endem- іс within their borders. Among the nearly 54.000 species listed in the 1997 IUCN Red List of Threat- ened Plants (Walter & Gillet, 1998), 91% were sin- ele-country endemies. The recent Red Book of Iran (Jalili & Jamzad, 1999) evaluates all 1727 endemic species (22% of the total flora). and the Ecuador Red Book focuses solely on the endemic species. What is required next is the assimilation and geo- referencing of the primary data of all of these. as well as every other country’s endemic taxa. Priority should be accorded to species representative of en- demic higher taxa, i.e., families and genera. A sim- ilar approach should be adopted within the frame- work of the new Critical Ecosystem Partnership Fund initiative (see [CEPF]), dress conservation needs at a regional scale, as de- which seeks to ad- limited by conservation Hotspots. which are them- selves defined in part by a minimum number (1500) 2000). Knowing where the endemic species occur (or al of endemic plant species (Myers et al., least occurred at some point in the past, and might possibly still occur) within countries or regions 15 fundamental for the rational allocation of finite con- servation resources But as the review of all primary data represent- ing the Malagasy endemic families has demonstrat- ed. knowing where the endemic species occur must he predicated on knowing just what the endemic species are. Catalogues such as those for Ecuador 1999) (Brako & Zaruechi. 1993) involve some review of the pri- (Jorgensen & León-Yánez. and Peru mary data. but generally do not exhaustively inven- tory all existing specimens. Revised taxonomic frameworks, however, should in theory be based upon examination of all existing collections. and therefore represent the most appropriate and op- portune point at which to disseminate reattributed primary data. Indeed, deposition of reattributed pri- mary data into an Internet-accessible database should be a sine qua non for publication of a re- vised taxonomic framework, just as deposition of GenBank has become (in most cases) a nucleotide data into (see (NCBI) precondition for publishing sequence necessary phylogenetic frame- works. For revised taxonomies of the Malagasy en- demic. families. comprehensive, reattributed (i.e. geo-referenced) primary data have been deposited in the world’s largest botanical specimen database. to which Internet access is provided through W"TROPICOS (see [MBG]). Recent discussion of the state of bioinformatics for biodiversity has sounded the call for improved infrastructure, and highlighted various developments involving remote query and retrieval from multiple, so-called distrib- uted databases (= “interoperability”) (Bisby, 2000: 2000: Krishtalka & Humphrey. Nevertheless. in conjunction. specifically Edwards et al.. 2000). with the publication of revised taxonomic frame- works, it would seem appropriate and extremely 150 Annals of the Missouri Botanical Garden Table 2. Representativeness of TROPICOS specimen database on 10 October 2000. 1.5 + million specimen records e 922,068 geo-referenced e |. a million identified to spp. representing 117 SF .806 752 spp. (39%) represented by only р specimen ө ub 709 spp. (53%) represented by | or 2 specimens € 96,456 spp. (82%) represented by 10 or less specimens @ 4,742 spp. (4%) represented by 50 or more specimens useful lo establish a GenBank analog (“SpecimenBank”) for deposition of the underlying geo-referenced primary data. The utility of recent Species Plantarum—Flora of the World (see [ABRS]) treatments of Irvingiaceae (Harris, 1999), Stangeriaceae (Steyn et al., 1999), and Welwitschi- aceae (Steyn & Smith, 1999) would be significantly enhanced if the underlying primary data were ac- cessible from an Internet specimen database. Just as the nucleotide sequences in GenBank document the microgeography of biodiversity, natural history collections define its macrogeography. Obvious linkages between the two scales should be made, as well as to seedbanks and living germplasm col- lections (at [IPGRI] and [NPGS]. Although the TROPICOS database of the Mis- souri Botanical Garden does contain some exhaus- tive sets of primary data such as those assimilated in the course of the review and revision of the Mal- agasy endemic families, in general, it reflects the idiosyncratic, incomplete nature of herbaria and their associated specimen data. The sources of the representativeness of TROPICOS (mostly contem- porary collections from a limited number of re- gions), wherein 117,806 species are represented by 1.5+ million specimens, dictate that the majority of species are represented by only one or two spec- imens; conversely, only 4% are represented by 50 or more specimens (Table 2). Just as taxonomic re- visions require the pooling of primary data from numerous herbaria, the synthesis of such data for conservation must also involve “North/South” her- barium databasing partnerships. For given country or region of the world, usually a limited any number of internal and external herbaria hold the majority of unique primary data. Adopting a cies by species approach, and beginning with those spe- endemic to countries and regions, North/South her- barium partnerships should work to synthesize pri- mary data into a global plant conservation database. A model already exists to emulate for the devel- opment of such a database. Botanical Garden Con- servation International (see [BGCI]) serves as a co- ordinating body to galvanize ex situ plant conservation efforts at over 500 member botanical gardens, maintaining a database of the ca. 85,000 species currently in cultivation in those gardens. Similarly, the 3000 herbaria worldwide (Holmgren et al., 1990), and indeed, the ca. 215,000+ plant species not yet in cultivation, would benefit tre- mendously from an analogous coordinating body “Herbarium Conservation International”) to help facilitate databasing partnerships for the synthesis of primary data that could directly inform in situ plant conservation. Such a coordinating body would assist in the organization of regional herbarium net- works such as the highly successful Southern Af- rican Botanical Diversity Network (see [SABO- NET]. and ensure that they are partnered with the appropriate Northern herbaria. The ongoing Red ist Program within SABONET (Golding. 1999a, 1999b, 2000) would be greatly enhanced from for- malized partnerships with the Northern herbaria where the majority of primary data from the region are housed. The task of synthesizing primary data of the most threatened plant species is large, but by no means insurmountable. The analysis of over 4000 adoran endemic species was achieved within a little over a year of the publication of the Ecuador Cat- alogue. With ca. 300,000 species to track, and 3000 herbaria, each herbarium would need to take responsibility for only 100 species on average. /n- dex Herbariorum lists 8800 staff working at the 3000 herbaria: throw in an additional 1200 stu- dents and volunteers, and each person would be responsible for collating the primary data of just 30 species on average. In fact, the task is not even that great. There are numerous widespread species of little or no conservation concern (except when their invasive capacity leads to the displacement of indigenous species). The problem should thus be attacked from both ends, identifving both the most widespread and "weedy" species, as well as those represented only by one or several collections. Mo- mentum is building to synthesize information on invasive species globally (see [GISP] and [NBII]). but there is as yet no organized effort to tackle the atter, i.e., the rarest of species. Similarly, great pro- gress (sometimes even with redundant and — ping efforts) is being made to diffuse the “1 of biodiversity (see [ABI], ПРА, "ПОР, TITIS), [Species 2000)). servation organizations seek to prioritize and pro- ). However, as governments and con- tect remaining tracts of viable habitat, it is imper- ative that they have access to more than just the have “what” of threatened biodiversity. “Names” meaning only in relation to the primary data that Volume 89, Number 2 2002 Schatz Plant Conservation define them: originally by their types, and then sub- sequently all the other specimens assigned to them by taxonomists in the course of trying to order the grand diversity of life. Therefore, by definition, the meanings of plant “names” cannot remain static. Each new collection expands the meaning in space and/or time, and has the potential to significantly modify or alter the meaning, or even to define a new name. To maximize the number of plant spe- cies that will survive the current extinction wave, m 9 we must also furnish the “where” of rare and threat- ened plant species, continually revisiting and attributing the primary specimen data, our only in- contestable record of life on Earth. Literature Cited Bisby, К. А. 2000. The quiet pn Biodiversity in- formaties and the Internet. Science 289: 2309-2312 Brako, L. & J. Zarucchi cand 1993. Catalogue of the Flowering Plants and Gymnosperms of Peru. Mon- 3 ogr. Syst. | Missouri Bot. Gard. 45: i-xl, 1-1280. Briggs. J. D. & J. I eigh. 1996. Rare or Threatened — stralian Plants, 1995 rev. ed. CSIRO, Australia. Capuron, R. 1963. Rhopalocarpacées. Fl. Madagasc. 127: 1-42. Cavaco, A. en wes Fl. Du Puy, D. J. & J. Moat. 199€ classification. in Madagase. 126: | . Vege — mapping abd Madagsas ar ene GIS mplications and recomme nations for the conse — of biodiver- v. Pp. 97-117 in € uxley, J. M. Lock & Culler — Chorology, Taxonomy, and Ecology af the Floras of Africa and Madagascar. Royal Botanic , Кем uw . Lane & E. S. Nielsen. 2000. Inter- ope кү at biodiversio databases: Biodiversity infor- on every desktop. Science 289: 2312-2314. 9‹ А — Red Data List for southern \frica SORORE News 4: 111-118 Е. 999}. Не pus and Red. Data Lists. SABONET Mews |: 202-20 2000. De aa with DDs: i neertainties m E Б ficiencies in — Listing. SABONET News 5: 28— Harris, D. J. ‚ Irvingiaceae. Species LUE я tralian — Resources Study, Canberra Holmgren, P. K.. N. Holmgren & . Barnett (edi- tors). 1990. Index Herbariorum: Part t The Herbaria of the World, 8th ed. International Association for Plant mation Taxonomy and New York Botanical Garden. Bronx. New York. IUCN, 1994. IUCN Red List Categories. IUCN. Gland. Switzerland. IUCI N/SSC Criteria Review Working Group. 1999. IUCN Red List criteria review provisional report: Draft of the — changes and recommendations. Species 31— 32: g A ЧЛ Jamzad. 1999, Red Data Book of Iran. Re- rch perce of Forests and Rangelands, мя nsen, P. M. & S. León-Yánez (editors). ogue of the Vascular Plants of к Bot. Missouri m Gard. 75: i—vii, 1—11 ke — n, М. 1958. Cornacées. Fl. — Tehran. 1999, Cata- аа Syst. 158: 13- — . & PS. Humphrey. 2000. Can natural his- tory museums capture the future? Bioscience 50: Oll- 617. Leandri, J. 1937. Sur l'aire et la position systé e du 3 genre Баре s — Thouars. Ann. Sei. Nat. Bot., sér. 10, 19: —319. Lowry. P. P., Il. n J.-N. Labat, С. E. Schatz, E Leroy & A.-E. Wolf. Madagascar. V. A s Pic 2000. families of revision of Kremolaena. Pen- tachlaena and ʻ ian (Sarcolaenaceae). Ad- ansonia, sér. : 11-31. ‚ G. К. Se а J.-F. Leroy & A.E. Endesa families of Madagascar. Ill. sion of Schizolaena (Sarcolaenaceae). 21: 183-212. Mvers, N., R. var ler, ne it. 2000 da Fonseca & J. Ken — Wolf. 1999. А synoptic revi- Adansonia, sér. 3. м 2 . Mittermeier, ES sity Кыр bor : 853—858. . Mac — mn. 1998. е vis — 40: Oldfie lk м The World . Lusty List « рез ne a Tre es. World Conservation Pres ( ч жн U.K Olson, D. M. & E. Dinerstein. 2002. The Global 200: Priority ecoregions for global conservation. Ann. Mis- souri Bot. Gard. € “ E. 9-22 Perrier p : Báthie. Н. 1946. Flacourtiacées. Fl. Mada- gasc. : 1-13 T 23 Thea ‘асве Madagasc. 13. Prance, G. E, H. Beenie, i — i — 2000. The tropical е remains undercollected. Ann. Missou- T" ri Bot. Gard. 87: 67-7 Randrianasolo, А. 8 J. S. Miller. 1999. ‘Taxonomic revi- sion of the genus Sarcolaena (Sarcolaenaceae). Ann. ТИЕ Вог. Сага. 86: 7 22. — G. : ~ P. P. Lov 7. Wolf. 1998. Endemic Madagascar. 1. — revision of Me- amphi В r (Me Mor dic eae . Adansonia, sé — 999a. Endemie \ — revision of Sphaerosepala- Adansonia, sér. 3, 21: 10712 ————. 1999b. A — IV. A syn y od families of Madagascar. П. беае. Ende mic families of optic revision of . = (Аз- 3 y teropeiaceae). Adansonia, sér. 3, ed 208. . Birkinshaw, P. P. Lowry Il. ean i & F. | TUN 2000a. The endemic jdm families of Madagascar apes oq taxonomy and conser- vation. Pp. 11-24 i . Loure | S. Goodman (editors). Diversis T Rade mism in n Na ar. Mém. Soc. Вторбо Er. II & A.-E. Wolf. 2000b. байа "s of Madagascar. VI. A synoptic revision of 22. Endemic Rho- 930 Adansonia, sér. 3, 22: 239— . P. Lowry dolaena (Sarcolaenaceae). 2524 . P. P. Lowry H & A.-E. Wolf. 2001. ladagascar. УП. Endemic fam- ilies af 3 — revision of Lepto- laena sensu ids ricto (Sare — naceae). Adansonia, sér. З. 23: 171-1 Skov, E on climatic similarity. ' axon я 2 515. A. Smith. . Welwitschiaceae. з жа ub. al Resources 2000. Pote ntial plant distributia mapping based ji ) Steyn, E. M. é Plantarum. Study. Canberra. Spe cies F. Smith & K. D. Hill. 1999. Stangeriaceae. Spe cies жаа. Australian Biological Resources — Canberra. Valen ., N. Pitman, S. León-Yánez & P. M. Jorgensen (e — 2000. 1. ‘bro rojo de las plantas e 'micas del 152 Annals of the Missouri Botanical Garden Ecuador 2000 [Red Book of Ecuador's Endemic Plants]. PUCE editions, Quito. Walter, K. S. & Gillet (editors), 1998. 1997 IUCN Red List of Threate ned Plants. Compiled by the World Conservation Monitoring Centre. IUCN—The World Conservation Union, Gland, Switzerland, and Cam- bridge, U. Williams; P. H. Area- . Key sites for conservation: selection me — biodive sity. Pp. 211-249 in G. Mi Balmford & J. R. Ginsberg, Conservation байл Univ. P Сап асе, ina MERE World. bri ridg ress, Internet Resources [ABRS]. Plantarum. Flora of the WWW anbg gov | itm Assoc GALÓN for Biodiversity Information: ane Species a - Garden С Somésvadion International. I. org.uk [BIODI]. BIODI. Advanced computational approaches to environme Pia and biodiversity information. (biodi.sds« [С ed Каш о оп Biological Diversity. (www.biodiv. [С E te critical Ecosystem Partnership Fund. (www.cepf. ne s/he гаа htm) [GISP]. Global а (jasper. stanford.edu/G [IOPI]. The E xs Checklist Project. (iopi.csu. edu.au/iopi/iopigpe l.htr [IPGRI]. Intern: ШШ Plant Genetic Resources Institute. www.ipgri.cgiar.or [IPNI]. International Plant Names Index. (www.ipni.org/) [ITIS]. Taxonomic Information System (www.itis.usda.gov/plantproj/itis/index.html) [IUCN]. International Union for the Conservation of Na- ture. Species Survival Commission. IUCN Red List of Threatened Species. (www.redlist.org/) [MBG]. W3TROPICOS. (mobot.mobot.org/W3T/Search/ vast. html Species Program. | NBII]. Invasivespecies.gov: The Nation's Invasive Species Information System (invasivespecies.gov/index.shtml) [NCBI]. GenBank overview. (http://www.nebi.nlm.nih.gov/ Genbank/GenbankOverview.html) [NPGS]. National Plant Germplasm System. (www.ars-grin. gov/npgs [SABONET]. Southern — Botanical Diversity Net- work a кик — [Schatz, G. 2001 | Gazetteer to Malaga Botanical ( ‘alle ting Loc liis (www.mobot. org/MOBOT/ research/madagascar/gazetteet [Species 2000]. (http://www. sp2000. org/) [Species Analyst]. (www.speciesanalyst.ne — ar CHICAGO WILDERNESS: A Debra K. Moskovits,? Carol J. Fialkowski,2 ү E DIAM Ad CONSERVATION! 5 ABSTRACT ^ = = ре = In a of diverse and determined organizations launched a new initiative in the Chicago region. Our vision? Chicago Wilderness: a thriving mosaic of natural areas. connected by greenways and wildlife corridors; embed- ded in the nation’s third largest metropolis. In this vision, the region's human communities reclaim a cultural tradition of protecting and restoring these globally significant natural communities that enrich our lives. Today. more than 100 organizations join forces to transform this vision into reali ty. A regional Biodiver rsity Recovery Plan—the result of three years of assessment and planning by scientists, land managers, educators, and policy y strategists—sets priorities and " determines the lines of action for the coalition. This жайна! agenda stems from our vision and recovery goals for each ecological community; it encourages targeted research initiatives that focus on characterizing our native biological diversity and on analyzin g elements critical to its recovery, Ultimately, though, the long-term survival of our natural wealth rests on the support from "the public. " hile the challe ‘поез to conservation educators and Communicators are many, QC hicago Wilderness аши us lo work toge ‘ther in unde rstanding our — nces, Channeling our resources, and ff ey words: Biodiver sity Rec covery Plan, Chicago Wilderness, citizen science, coalition, conservation education, forest preserves, fragmentation, metropolis. mosaic, natural communities, oak savannas, oak woodlands, prairies, public outreach, remnants, restoration, urban. CHICAGO WILDERNESS: THE NATURAL CAPITAL west (Table 1). These remnants shelter a number of species that are rare worldwide and that are listed Chicago. the third largest metropolis in the Unit- as threatened or endangered in the states of Illinois. ed States, is anything but "untrammeled by man. Indiana. and Wisconsin. These species once lived where man is a visitor who does not remain" in the in the thriving, expansive grasslands and wood- words of the Wilderness Act (Section 2.e: [Wilder- lands that have now been turned into monocultures ness Асі). which has shaped our modern concept of corn or soybean. of wilderness. Nonetheless, "Chicago Wilderness At the turn of the century, Chicago was blessed — is the emerging force in our urban landscape. with far-sighted planners who created a system of Launched in 1996 by a coalition of 34 diverse and forest preserves in and around the city. In 1900. determined organizations, Chicago Wilderness joins architects Jens Jensen and Dwight H. Perkins led forces to conserve the globally important natural he Municipal Science Club in Chicago. which communities that survive in our metropolis. worked hard for many years to ensure that the best The surprise to many is that the boundaries of remaining natural lands were set aside for the pub- Chicago Wilderness capture a spectacular array of lie (see Sullivan. 1997: Chicago Region Biodiver- rare ecosystems: tallgrass prairies, oak woodlands. silty Council, 1999: 19-21). Famous architect and oak savannas, sedge meadows, marshes, bogs, fens. planner Daniel Burnham and fellow architect. Ed- and other prairie wetlands. In fact, within Chicago ward Bennet also contributed extensively to the Wilderness—which extends in a crescent around cause. The wilderness in Chicago now survives in Lake Michigan from southeastern. Wisconsin, parcels that range in size from half-acre remnants through six Illinois counties, into northwestern In- to. the 19.000-acre Midewin National Tallgrass diana—survive some of the best remnants of the Prairie. Some of the natural remnants survive in old prairie and oak-savanna communities of the Mid- settler cemeteries or along railway rights-of-way. ! We thank the members of Chicago Wilderness and of its teams, along with the citizen scientists in the region. Collectively, they are the power behind this re — effort and the focus of this paper 3E vironmental and e conservation Programs, The Field Museum. 1400 S. Lake Shote Drive, Chicago, Illinois 60605- 2496, U.S.A. dmoskovil Idmuseum.org. « mu Idmuseum.org '* Department of cari The Field Museum, 1400 S. Lake Shore Drive. С hicago, Hlinois 60605-2496, U.S.A. gmueller @fie — um.org Conservation Programs, Chicago Zoological Society, Brookfield Zoo, 3300 Golf Road, Brookfield. Hlinois 60513, l 4 \. кешу ТТИ СА ldzoo.org ANN. MISSOURI BOT. GARD. 89: 153—163. 2002. 154 Annals of the Missouri Botanical Garden Table 1 oak savannas, oak woodlands, prairie wetlands) with high floristic quality. Remaining natural areas in Illinois (prairies, Region Original acreage % Surviving Galena 371,000 0.00005 Chicago 1,998,000 0.04. West Central 3,217,000 0.00006 Springfield 2,041,000 0.003 Source: The Illinois Natural Area Inventory, Illinois De- partment of Conservation, 1976 Most survive in the preserves that total 200,000 acres (81,000 ha) of federal, state, county, and municipal protected lands and waters. All par- cels exist within the much larger matrix of public and private, built and natural lands that support the region's 8 million residents. Our natural communities suffer from a number of stresses, including habitat fragmentation, inter- ruption of natural fire regimes, air and water pol- lution, and native and exotic invasive species. The long-term survival of these endangered plant and animal assemblages depends on the proper ecolog- ical management and aggressive restoration of de- graded lands, even—or primarily—in the region's forest preserves and conservation districts. It is this need for coordinated and rapid action that now brings together more than 130 institutions includ- ing landowners, government agencies at local, state, and federal levels, education and research institu- tions, and conservation organizations to turn the vi- sion of Chicago Wilderness into reality. AND WHY CHICAGO WILDERNESS: WHAT, WHERE, Chicago Wilderness is both the name of the co- alition of organizations committed to protecting, re- storing, and managing the natural communities in the Chicago region, and the natural communities themselves, along with the plants and animals that depend on them for survival. Chicago Wilderness also is the people of the region, the human resi- dents whose quality of life depends on a thriving regional nature reserve. Urban living often results in an almost complete detachment of people from the land. An important goal of Chicago Wilderness is to reconnect a land- less urban population with the pulse of nature. The name Chicago Wilderness is meant in part to em- phasize the connections between wilderness in the inaccessible conventional sense—in remote and places—and wilderness in the places where people ive and work. | 1 work But why strive for a regionwide coalition instead over of working through more manageable, individual collaborations? The answer lies in the realities of the region: extreme fragmentation of the natural communities. The high-quality native patches in the region are tiny. Nothing short of a massive, re- gionwide, coordinated effort will ensure the long- term survival of these natural communities, through the management of much larger lands that surround and connect the remnants. In the vision of Chicago we reconnect the 200,000 acres of fragmented natural areas through greenways and Wilderness, wildlife corridors. And in this vision, the human communities reclaim a cultural tradition of restor- ing, protecting, and managing our natural commu- nities. CHICAGO WILDERNESS: THE ORGANIZATION Chicago Wilderness is a loosely structured or- ganization, with an unconventional organizational chart (Fig. 1). In Chicago Wilderness we are bound by our common goals and objectives and by our collaborative projects. We are in essence a volun- teer-driven effort. Four teams focus on the central lines of action: science, land management, policy and strategy, and education and communication. The teams attract participation of many non-mem- ber institutions, which adds to the scope and strength of the coalition. Current and past chairs of these four teams and staff from other member or- ganizations form the nucleus of a coordinating group that develops central strategies, maintains momentum, and resolves day-to-day problems. A steering committee of executives in the Chicago Region Biodiversity Council oversees the initia- tives, including approval of budgets and projects. The Council encompasses the chief executives of federal, state, and local all member institutions: government agencies; land-owning agencies; re- search and education institutions; and conservation organizations. These executive, or voting, members elect the steering committee. A proposals commit- tee solicits and reviews proposals for priority con- servation projects and recommends funding; the committee has one or two representatives from each And the Recovery Plan Task Force spearheaded the gigantic effort of developing, compiling, and writing the first draft of the Biodi- of the four teams. versity Recovery Plan for the region (see below). A few ingredients have been instrumental in al- lowing Chicago Wilderness to succeed as a massive coalition. We briefly highlight these below. (1) A critical mass of people thoroughly—pas- sionately—committed to seeing the work succeed, and ready to put an inordinate Volume 89, Number 2 2002 Moskovits et al. 155 Chicago Wilderness Figure 1. — — N w — Chicago Wilderness Structure Chicago Region Biodiversity Council Executive Council General Members Proposals Steering Committee | J— Committee / Science agemen Collaborative Projects amount of time and effort into making it hap- pen. For us, this core energy came from the coordinating group, with its 12-15 members. Early, visible, and popular successes. In Chi- cago Wilderness, these early successes Came from some of our public awareness materials. The Atlas of Biodiversity (Sullivan, 1997) the first publication of the Chicago Region the natural communities in the region, with their Was Biodiversity Council. It characterizes rich assemblages of plants and animals. and helps bring these natural areas to the atten- tion of a very broad audience. most of whom are unaware of the existence of such biolog- The Atlas was one of the first concrete examples of how ical riches in our own backyard. Chicago Wilderness could help individual members become more successful and effec- tive in their work. The whole region needed a powerful, visual publication about our nat- ural treasures. The Atlas is now in its fourth edition, with more than 40,000 copies al- ready distributed; see [Chicago Wilderness publ. |. Mission-related benefits to individual mem- bers. For the coalition to thrive, individual members must become more successful i their individual endeavors because of "e association. with the coalition. Conversely, each member must contribute in its own way and with its unique strengths to the coalition and to the pursuit of the collective mission. For example, one of the goals of The Field Museum is to direct its resources—scientific innovative expertise, worldwide collections, education programs—to the immediate Organizational chart for Chicago Wilderness. needs of conservation at the re pona and i in- ternational levels and to create a < be- tween academic institutions and i groups most closely tied to biological stewardship. Chicago Wilderness offers the Museum a tre- mendous advantage in tapping into a wide. with Chicago Wil- directed regional effort. In turn. the more of its efforts in research, inventory, and derness, Museum has much tools for conservation to the Chicago region. The potential for Chicago Wilderness to serve as a model for urban conservation attracted early at- tention of several federal agencies, including the S.D.A. Forest Service, the U.S. Fish and Wildlife Service, U.S. Ageney, who have provided significant operating and the Environmental Protection grants since 1996, bringing new conservation rev- enues to the region. Matching services from mem- bers. and funds attracted by members for projects catalyzed by Chicago Wilderness, exceed 70% of the funds dedicated to this regional effort. CHICAGO WILDERNESS: IMPACT ON SCIENTIFIC RESEARCH Research focusing on local biodiversity and land management issues in the Chicago region has been ongoing for over LOO years. But a recent resurgence of interest in local issues by the region's scientific community was coincidental with, and in large part a consequence of, the founding of Chicago Wilder- ness. This outcome can be attributed to several rea- sons, including that the consortium has been suc- cessful in bringing new resources to the area to than the new More funds, however, is that the partnerships and oppor- support research. important 156 Annals of the Missouri Botanical Garden tunities for collaboration fostered by Chicago Wil- derness have created new research initiatives, iden- tified important interest in the resulting data, and created an en- hanced atmosphere for the importance of biodiver- research questions, expanded sity and ecological research focused on the region. Several Federal and Illinois State agencies have provided grants to Chicago Wilderness to help en- able the initiative to meet its goals. Chicago Wil- derness makes requests for proposals each year to distribute these funds to outstanding projects. A proposal committee (see above) reviews submitted proposals and ranks them based first on merit— and then on the rel- scientific, educational, policy evance to the mission of Chicago Wilderness and on the degree to which the results will make unique contributions to the initiative; i.e., will serve as a model for other projects or will provide data that may serve as the underpinning for a series of other studies. The topics of funded research proposals fall into one of the following broad categories: Re- search/Inventory/Monitoring; Restoration; and Tools for Research—imaging projects, database projects, or other conservation tools. The following exemplars of science projects funded through Chicago Wilderness provide an in- dication of the breadth of the research undertaken by Chicago Wilderness member institutions. All projects funded through Chicago Wilderness are summarized in www.chicagowilderness.org. Al- though Chicago Wilderness funds only a small por- tion of the flourishing biodiversity research activity in the area, it has been greatly responsible for in- vigorating the interests of scientists in conducting research in the region. RESEARCH/INVENTORY/MONITORING PROJECTS Projects within this category focus on increasing our understanding of the native biological diversity in Chicago Wilderness. Research projects concen- trate on factors influencing that diversity and on the efficacy of current management practices to main- tain and restore the region’s native diversity. Fund- ed projects have ranged from studying the effects of restoration efforts in oak woodland communities on a broad group of taxa (fungi, beetles, millipedes, amphibians, mosses, birds, and flowering plants). to documenting the diversity and distribution of the region's bats, to landscape-level inventories of wa- tersheds, to studying the role of arbuscular mycor- rhizal fungi in prairie restoration. RESTORATION PROJECTS The focus of projects in this category has been (1) to develop best practices transferable to other sites or (2) to restore specific sites of special sig- nificance. Funded projects have ranged from de- veloping and testing methods for restoring specific threatened and endangered plants, to restoring a degraded tallgrass prairie at Indiana Dunes Nation- al Lakeshore. TOOLS The goals of these projects are to develop useful tools that facilitate research and land management activities in the region. These tools include imaging projects to develop multispectral vegetation maps using aerial flyovers, creating an online metadata catalogue of existing spatial datasets owned by Chi- cago Wilderness member institutions, and using spectral data to track changes in land use and land cover over the region. Funded database projects in- clude compiling data from past breeding bird sur- veys into a user-friendly database [Bird census] and the development of a database of the current status and location of the region's threatened and endan- gered plant species [Chicago Wilderness grant]. An example of another type of tool is the development of a user-friendly computer model to analyze strat- egies for deer removal in different land-manage- ment programs [Ecology Software Server]. The continued growth and development of the Chicago Wilderness coalition and the impact that the coalition is having throughout the area, has en- hanced communication among scientific research- ers, land managers, educators, and policy makers n the Such communication is critical for effective development of research programs that en- region. hance the conservation of biodiversity. Through un- derstanding the types of data most needed by the users of research results (i.e., land managers, con- servation biologists, policy makers, educators), and how to present these data to the users, we have identified new research projects while enhancing the value of our research and expanding the audi- ence for our results. In turn, the region has gained better recognition of the value of research and has increased opportunities for its continued funding. The positive cycle continues to foster further re- search activities in Chicago Wilderness. OUR Roab Mar: THE REGIONAL BIODIVERSITY RECOVERY PLAN From the earliest days of the Chicago Wilderness coalition, member organizations recognized that to be effective, we needed a strong regional conser- vation agenda with clear priorities for action, The concept was to produce a biodiversity recovery plan, in some ways analogous to a single species Volume 89, Number 2 2002 Moskovits et al. 157 Chicago Wilderness recovery plan, which would provide an assessment of the current conditions along with a road map for how to reach desired vision and goals. An effective recovery plan would help shape decisions at both the policy and the land-management levels. We completed the plan in late 1999, after almost three years of dedicated work by more than 200 members and nonmembers—including the science, land management, education, and policy experts in the region [Chicago Wilderness biodiversity |. The Recovery Plan identifies the ecological communi- ties of the greater Chicago region. assesses the overall condition of these communities and the threats facing each, lays out the vision and recovery goals for each species assemblage. and recom- mends actions to achieve our goals. The plan also introduces a process of conservation and informa- tion design that leads to coordinated research. in- ventory, and monitoring agendas to support these conservation goals and strategies and to anchor them on the best available science. And the plan calls for a massive communication effort to galva- nize public support behind our conservation agen- da. ANATOMY OF THE RECOVERY PLAN The plan provides both an overarching vision for what Chicago Wilderness—the land and biodiver- can become and a broad treatment of the de- sily tails of how to get there. In developing the plan. we used a consensus-based approach, relying on the region's experts and the available data. The plan also provides the platform for organizing action at a further level of detail, and it provides a strong foundation for communication about the challenges facing biodiversity conservation in the region. The recovery plan document itself is intended for reference. It is a comprehensive treatment of the central issues, with 11 chapters and 11 appendices Council. 1999 Chapters cover a summary of the biodiversity val- (Chicago Region Biodiversity ues in the region: a history of the changes in the landscape and of conservation efforts to date: key ecological processes that shape and sustain the re- gions natural communities: the status. needs, and goals of terrestrial and. aquatic communities: the status of threatened species: research and monitor- tools for conservation: ing needs: goals for com- munication and education: and the role for. key players in executing the plan. To present the find- ings of the plan to broader audiences, Chicago Wil- derness developed а smaller, summary document focusing on the important. issues and. recommen- dations. The recovery plan process brought together ex- perts from many backgrounds and disciplines to share their knowledge. The outcome is not only the region’s state-of-the-art insights, but also a blending of our science, policy, and communications exper- tise. The process resulted in heightened interest and commitment for collaboration in more detailed plans for conservation action within a regional con- ех! The plan contains 140 separate rec ‘ommended actions that fall into the following six categories: (1 — Land management. Most natural areas in the region are not being managed in a manner that will sustain plant and animal diversity over time: more acres need to be under eco- f and the intensity ( - logical management, management must Increase. Qe М 2) Land protection. Additional land. will need protection as the region’s human population continues to grow. The primary concern in setting aside more land for conservation is to create larger, unfragmented protected areas, either de novo or by enlarging existing ones. Since much of the land in natural conditions is either protected already or is in extremely small parcels, the recovery plan. envisions the need for restoring to ecological health disturbed and converted areas, such as ag- ricultural lands, as a strategy for creating arger units. мм Stream protection, We must protect remaining high-quality streams, where most of the re- maining aquatic biodiversity resides, from additional degradation. Research needs — Research and monitoring. relate primarily to understanding how best to address the threats to biodiversity posed by human development. For monitoring, the plan puts forward a system of adaptive man- agement based on conservation planning (briefly discussed below). (5) Education and communication. The plan rec- ognizes that public understanding and sup- port is a precondition to the success of long-term, broad-based public undertaking such as biodiversity conservation. The plan lays out steps to achieve increased under- standing of the importance of biodiversity takes short- and long-term scales (see further dis- and what il to conserve it on both cussion, below). 6) Local and. regional development policies. The — region is at a critical time for developing pol- = icies that will control anticipated growth iı 158 Annals of the Missouri Botanical Garden Table 2. Vegetation communities in Chicago Wilderness: a consensus list developed in 1999 by members of C hicago Wilderness Science and Land Management Teams and by staff of the regional Forest Preserve Districts. Community classes (8) Communities (24) Subcategories (55) Forested communities Upland forest Floodplain forest Flatwoods Woodland Savanna communities Sand savanna Shrubland communities Sand shrub Prairies Fine- айша. aoil prairie Sand prairie Gravel prairie Dolomite prairie Wetland communities Marsh Bog Fen Sedge meadow Panne Seeps and springs Cliff communities Cliff — Lakeshore communities Akeshore Cultural (human-dominated) Croplan vegetation communities Unassociated growth Tree plantation Developed land Fine-textured-soil savanna Fine-textured-soil shrubland Dry mesic, Mesic, Wet mesic Wet mesic, Wet Northern, Sand Dry m Mesic, Dry mesic, esic, Wet mesic ', Wet mesic Dry, Dry — Dry mesic, Wet mesic Dry mesic, Wet mesic Basin, Streamside Graminoid, Low shrub, Forested Calcareous floating mat, Graminoid, Forested Neutral, Calcareous, Aci Eroding bluff, Dolomite cliff Beach, Foredune, High dune Forb, Shrub, Tree Grass, population and in urban space in the coming decades. The recovery plan stresses the im- pact of the region’s management of growth on biodiversity and suggests steps to mitigate that impact. THE PROCESS OF CREATING THE RECOVERY PLAN The primary units of conservation in Chicago Wilderness are plant communities. Initial natural area inventories for the region, upon which most conservation strategies to date have been based, fo- high-quality communities (White, 1978). The recovery plan also adopts plant cused on remnant communities as its primary conservation targets, ut it considers animal communities or assemblag- es, aquatic communities, and threatened species as well. A critical first step in the plan became to develop and adopt a plant community classification system that all Chicago Wilderness members could accept and upon which the region would base sub- sequent conservation and land management deci- sions (Table 2). ost land management agencies in the region traditionally have classified their holdings accord- ing to the community classification system used by the Illinois Natural Areas Inventory (INAI) in the 1970s (White, 1978). Rather than adopt a com- pletely new system for the recovery plan, Chicago Wilderness elected to develop a modified version of the INAI system. The modifications reflect a deeper current understanding of community gradi- ents in fire-dependent forested lands in the region (see Bowles & McBride, 1996). In addition. the new classification places more focus on degraded lands, recognizing those lands as the primary target of res- toration efforts. While the Chicago Wilderness clas- sification system does not follow the emerging na- tional standard for community (Grossman et al.. 1998), the recovery plan includes a crosswalk to these national standards. The resul- tant community classification system includes 8 major community groups, 24 community types, and 55 Ар distinguished primarily by soil mois- ture (Table While i * lays out actions covering all of the biodiversity of the region, not only what classification is unique or rare, it also incorporates a method to identify what is most important from a global per- spective and what conservation targets need the most immediate attention. We undertook this eval- uation primarily at the natural community level, but also included information from key groups of ani- mal assemblages and of threatened species. Volume 89, Number 2 2002 Moskovits et al. 159 Chicago Wilderness Regional experts first evaluated the status of an- imal groups or subsets of groups: reptiles and am- phibians. breeding birds. mammals, and some ter- restrial invertebrates. In workshops, the scientists grouped the species found in the region into broad assemblages associated with habitat types: for ex- ample. savanna birds or forest and woodland rep- tiles and. amphibians. These assemblages are not mutually exclusive in that some species will fall into more than one grouping. The experts then as- sessed the significance of the Chicago Wilderness region to the global survival of each assemblage and its constituent species. The three ranks we used were: Globally Critical, meaning that the Chi- cago Wilderness region is very important to the overall survival and. well-being of the assemblage: Globally Important, meaning that the region con- tributes significantly to the overall well-being of the assemblage: and Locally Important, meaning that the assemblage is an important component of the region's biodiversity, but its status in the region will not have a substantial impact on the global well- being of the assemblage. Experts also assessed the condition of the assemblage in the region. ranking 3 them as “of concern/declining” or “stable.” Com- bining the rank in importance on the one hand, and the condition of the assemblage on the other. helped to identify the Chicago Wilderness habitats that need most immediate attention to stem de- clines or to improve conditions. The exercise also helped us to determine the habitats in the region that are most significant in terms of global conser- vation. The moist grassland bird assemblage came out as both. Globally Critical and in poor condition. Ten other assemblages in Chicago Wilderness are Globally Important and are declining. АП of these assemblages and conservation targets are listed in the Biodiversity Recovery Plan. Prior to evaluating communities, planners also reviewed the threatened plant species in the region and placed them into six categories of priorities, from species that are globally rare to species that may be adequately protected or stable. but are re- stricted to communities that are rare within Chicago Wilderness. Presence of threatened plant species was an element in evaluating priorities for conser- vation targets within biological communities. To evaluate the relative need of different biolog- ical communities for additional conservation action, Chicago Wilderness members built on existing methods for ranking endangered ecosystems (Noss et al.. 1905: Noss & Peters. 1995). We modified these systems slightly, condition (quality), which is one of the most sig- nificant elements in the sustainability of biodiver- Table 3. Conservation targets for recovery based on status, importance, and distribution. to include the key factor of (highest) tic — (all i moisture classes) Fine-textured-soil savanna (all moisture classes) vanna Sand prairie (all moisture gradients in dune and swale lesic sand sa topography) Dolomite prairie (all) Panne Graminoid fen Fine texture 'd-soil prairie (all moisture classes) Second t Dry sa ind savanna ravel prairie (all) Bania i marsh Calcareous floating mal Calc 'areous seep Sand prairie xm than those in dune and swale topogr raphy Northern flatwood Streamside marsh sity in this urban region. As with the animal as- semblages. expert groups convened for one-day workshops over a period of several months to assess available data and to rank natural community types against defined criteria. Four rankings—quantity, quality, biological importance, and contribution to global conservation—combined to make the final ranking. The ranking variables at focus were: For quantity—(i) total amount (acres) remain- ing of the community, (ii) amount lost. since (111) arge enough to sustain a relatively intact as- pre-settlement, and number of patches semblage of constituent species over time. e For quality—(i) remaining percent in good quality, (ii) degree of fragmentation and isola- tion. and (iii) extent and effectiveness of cur- rent management efforts. For biological importance—(i) levels of species richness, (ii) numbers of threatened species. and (iii) key ecological processes. For contribution to global conservation—rank- ing of communities as (i) endemic to the re- gion. (11) at the center of the range. or (iii) at the edge of the range. This assessment also considered the effectiveness of conservation ef- forts for the community type in other regions. We combined the individual rankings to develop a tiered set of conservation targets, recognizing that all community types will require some degree of conservation effort to survive in the urban land- Table 3 shows the first two tiers of commu- nity targets. In all, 13 daylong workshops (in 1997—1999) for experts scape. Chicago Wilderness convened 160 Annals of the Missouri Botanical Garden to rank community types, and to discuss threats and conservation needs. Through a similar process, we evaluated the conservation needs of rivers, streams, and natural lakes in the region. The ranking system to produce the conservation targets was not quantitative, but it represents a con- sensus of the experts in the region. Key effort in the planning process went toward developing and adopting a process for further refining our conser- vation goals, developing strategies to obtain those goals, and setting up monitoring systems that en- able information to feed back into refining manage- ment strategies toward these goals. The Chicago Wilderness region adopted a pro- cess of conservation design, where planners (1) de- velop a vision and goals for each conservation tar- get, (2) identify key threats to achieving the vision/ goals, (3) set conservation strategies to abate the threats, (4) establish indicators for progress toward achieving the goals, and (5) develop a feedback mechanism to adjust management strategies if mon- itoring shows progress is insufficient. Chicago Wil- derness members are using additional consensus- based expert workshops to develop this iterative process for conservation targets at individual sites and across the entire region. Linking conservation design to efforts that involve citizens in collection of scientific data will help combine the scientific aspects of the recovery plan with education, out- reach, and public involvement programs. Once assembled, the draft of the recovery. plan went through a formal, legal hearing process under the auspices of the Northeastern Illinois Planning Commission and the Northwestern Indiana Region- al Planning Commission. Both Commissions and several municipal agencies adopted the plan within months of its publication. The value of the formal process was not only to gain feedback from a larger public, but also to seize on the opportunity to raise the issue of biodiversity in fora traditionally con- cerned primarily or exclusively with development and transportation issues. The final recovery plan is now on a level with other major plans that will guide the development of the region in the years to come. As a result, the issue of biodiversity will be discussed with business and civic leaders along with economic competitiveness, quality of life, and education. OUR EDUCATION GOALS Chicago Wilderness realizes that the future of our native landscapes depends on the support and involvement of our citizenry. On the encouraging side, Gallup surveys (1999) and research by Belden and Russonello (1996) reported that most Ameri- cans consider environmental protection to be a pri- ority. Concern about air and water pollution is widespread; citizens worry about destruction of tropical rain forests, toxic waste contamination, and habitat loss. But while concern about these issues is wide, it is not deep. Environmental issues often fade when economic matters compete. And biodi- versity recovery faces additional challenges: public surveys report that only 20% of the population rec- ognizes the term “biological diversity” (The Biodi- versity Project, 1999) Faced with these challenges, the Education and Communication Team of Chicago Wilderness has adopted (1) short-term communication strategies coupled with active public involvement projects, while (2) building and implementing longer-term educational strategies. To guide our actions, the more than 200 educators in Chicago Wilderness have embraced the internationally held definition of Environmental Education: “Environmental Education is a learning process that in- creases people's knowledge and awareness about the environment and associated challenges, develops the s and expertise to address these chal- attitudes, motivations, i com- mitments to make informed decisions and take respon- sible action” (UNESCO 1978: 5: highlights added). necessary lenges, and fosters The boldfaced words (our emphasis) spell out the five components essential for making a program in biodiversity education result in changed behavior and in informed decision-making. Chicago Wilder- ness Educators have used these five components— awareness, knowledge, attitudes, skills, and partic- ipation—to develop a matrix that assesses our ex- isting programs, from birth to death, against our targeted objectives. Table 4 presents an abbreviated sample of this education matrix. An “X” indicates presence of the component in that program. For example, the BETA project incorporates all five essential components. Our compilation revealed that while our programs are strong in awareness and knowledge, they are weak or altogether lacking in attitude clarification, skill development, and active participation, To ad- dress these gaps and to reach the 80% who have never heard of biological diversity, while also en- suring that our education systems produce biolog- ically literate graduates, we adopted program guidelines that (1) extend the scope of our existing, successful programs to include attitude, skills, and action components; (2) develop new programs that specifically address the lacking components; (3) reach beyond our current circles by working with community-based organizations; (4) expand the dis- Volume 89, Number 2 2002 Moskovits et al. 161 Chicago Wilderness Table 4 in that program. Sample matrix of existing programs in conservation education. An “X” indicates presence of the component Target audiences Existing resources Essential components of biodiversity education Awareness Knowledge Attitude Skills Participation Children (Pre-School-8th Ап Atlas of Biodiversity X X v Chicago Tribune Biodiversity Supple- X X X X ment Biodiversity Education Through Ac- X X X X X tion (BETA) Project Chicago Wilderness Communication X X Tools Mighty Acorns Expansion Program X X X X X IDNR Biodiversity of Hlinois CD- X X X X ROM Series EcoWateh X X X X X Adults An Atlas of Biodiversity X X Chicago Wilderness Magazine X X Chicago Wilderness Communication X X Tools EcoWatch X X X X X Its Wild in Chicago N X X tribution of our successful educational tools: and (5) ensure that students in the Chicago Wilderness region—from elementary through university—are literate in. biodiversity. Promoting programs jointly, creating complemen- {агу (rather than competitive) projects, sharing teacher-education expertise, and submitting collab- oratively developed proposals to schools are some of the new ways in which Chicago Wilderness reaches adults. children, and educators for greater impact. For example, Mighty Acorns is a steward- ship program for eight- to ten-year-olds created by The Nature Conservancy. This highly successful program used to be limited to a few schools in one county. Today, through the leadership of Chicago Wilderness and two of its members (The Nature Conservaney and The Field Museum). the program has expanded to 19 Chicago Wilderness organiza- tions in six counties in Illinois and one in Indiana, and it involves 90 schools, 210 teachers, and 6500 students. The majority of these 6500 students come from underserved communities. These elementary students are engaged actively in learning about the region's plants and animals, experiencing nature, and practicing ecological stewardship—removing buckthorn and garlic mustard, collecting and plant- ing seeds—to improve the health of our natural communities in the Chicago region. Chicago Wilderness also has partnered with large distribution networks, like the Chicago Tri- bune’s Newspapers in the Classroom program, to develop and promote a curriculum and activity guide on the biodiversity of the region, using the Chicago Wilderness Atlas as the central teaching tool. Through its website and a printed brochure, Chicago Wilderness provides to all in the region a full list of the local educational resources and how to obtain them [Chicago Wilderness publ.]. PUBLIC. INVOLVEMENT AND OUTREACH To inform the public about pressing issues relat- ed to the protection of our natural communities and to engage a large constituency in the support and work of restoration, the Biodiversity Recovery Plan outlines the following priorities: * Increase the public's understanding of the role of ecological management * Foster grassroots involvement in restoration and conservation activities * Expand opportunities for “citizen science * Engage a wide public as volunteers One opportunity that fosters wide public involve- ment is an annual spring event, “Ws Wild in Chi- cago." ' visitors to interact with Chicago Wilderness mem- Ihe four-day festival attracts thousands of bers at their displays through featured performanc- es, hands-on activities, and targeted programs. The festival provides a forum for the public to learn about Chicago's wilderness and for Chicago Wil- derness to reach new audiences with our messages. 162 Annals of the Missouri Botanical Garden Our National Public effort fostered by Chicago Wilderness, enhances Lands Day event, family involvement as volunteers and the engage- ment of a wide public in exploring the role of eco- logical management in the survival of our native communities. In 2000, more than 6000 people par- ticipated in restoration activities at 21 Chicago Wil- derness sites. Through a program of the Illinois Department of 8 prog | Natural Resources (IDNR; [EcoWatch of citizens ). hundreds adults and students— high-school gather data on the ecological condition of rivers, forests, wetlands, and prairies in Chicago Wilder- ness. Several more hundred “citizen scientists” gather targeted data on birds and butterflies in the instances, the volunteers. them- region. In. many selves enter the data on-line, to be analyzed by regional scientists and land managers, and by the state. Through our process of conservation design, we are now tying the data collection by citizens to the needs of land managers in monitoring and re- vising their management efforts. As a result of Chicago Wilderness, Field Muse- um scientists and educators partnered with IDNR to develop [UrbanWatch] with the following goals: * Record the presence and distribution of groups of organisms—plants, animals, fungi—in an urban environment (we selected organisms that can be used to assess the condition of natural areas in urban centers) * Examine environmental factors that influence the biodiversity in urban ecosystems Analyze and compile the information in a way that is readily usable by urban planners and park managers. Focused on city parks, cemeteries, golf courses, corporate Campuses, backyards, and tree-lined streets, UrbanWatch collects data to ex- schoolyards, amine environmental factors that influence the na- tive diversity in urban green space. Urban plan- ners, park districts, site managers, and landowners will have access to these data and will apply them i 1 the management of sites to improve or create urban spaces that are friendly to native communi- ties and that address the needs of migrating birds and butterflies in our region. OUR COMMUNICATION OPPORTUNITIES Significant current challenges in ecological res- toration require immediate strategies to increase public awareness and understanding. Chicago Wil- derness has complemented its long-term approach- es in public involvement, with actions aimed another achieving the following short-term goals for com- municating about biodiversity to the general public: * Recognize biodiversity in everyday experienc- es Understand human impacts on biodiversity * Recognize the connection between healthy na- ture and quality of life The Chicago Wilderness Communications Team has adopted the successful approach to crafting messages developed by the Biodiversity Project (1999) in Madison, Wisconsin. vocales having a clear goal and target audience, The approach ad- identifying core values of the target audience that the message embraces, and writing the message in short paragraphs. These paragraphs address (i) why the issue is important to the audience (leading with values), (ii) what the threats are and who is re- sponsible, and (iii) what people can do to address the threats. Through this method, the message di- rectly incorporates the values of the audience. Equally important, each message ends with an ac- tion that the target audience can easily undertake. We developed a number of attractive and highly effective communication tools using this method, including a video, tabletop display, and slide show [Chicago Wilderness]. These tools are available to all Chicago Wilderness members, to deliver a con- sistent, compelling message. Protecting Nature in Your Community, a companion piece to the Biodi- versity Recovery Plan, is one of our tools that fo- cuses on reaching public officials, policy makers, municipalities. zoning boards. and the over 300 governing bodies that review and guide the design of hundreds of proposals for land development and redevelopment every week. For the general public, the Chicago Wilderness Magazine, currently with over 9000 subscribers, is sold at local bookstores, including Barnes € Noble and Borders, at gift shops of our member institutions, and at other out- els. While the challenges to conservation education and communication are many, the Chicago Wilder- ness coalition enables us to work together in un- derstanding our audiences, challenging our re- sources, and creating novel approaches to engage the widest public in maintaining and restoring our local biodiversity. Chicago Wilderness began with a larger-than-life goal: nothing less than rescuing vibrant natural ar- eas by transforming the environmental culture of the people responsible for them. Through the part- nerships formed and the collaborative work, Chi- cago Wilderness is proving itself equal to the task. From scientifically rigorous approaches in ecologi- Volume 89, Number 2 2002 Moskovits et al. 163 Chicago Wilderness cal inventory, restoration, and monitoring to crea- tive advances in planning, policy, education, and outreach, the accomplishments of Chicago Wilder- ness are building a living legacy for the future. Literature Cited Be * ain and Russonello Research and Communic ations. Human Values Mute on Biological Diversity. vello p search and Communications. ся D.C. & J AcBride. 1996. Evaluation and ina, Woodland. жк Barrens Nat- The Morton Arboretum. Bowles, шй C iis ‘ation of Sani Ju is in Northern Illinois. asle, ili i Biodiven TRIN — Making the Connection: tions J The Biodiversity The Public. Communica- Madison. 1999. Life. Nature. A Biodiversity Project. { ‘onsi n. C аз d Biodiversity Council. 1999, Biodiversity Hec v Plan. Chicago Region. Biodiversity Counc il. C hicago. Illinois. Gallup News Service, the Gallup Organization. 1999. U.S. public worries about toxic waste, air and Mer er pollution as key environmental — CNN/USA Today/Gallup Poll conducted March x D. ж D. — tpe doen, А. S. Weakly. M Bourgeron, R. Crawford. K. Goodin. S. Төк K. D. Patters yr . Reid 1998. International Classification of Ecological Communities: Terrestrial Vegetation of the United States. Vol. I. The National Vegetation Classifi- Grossman. cation System: Development, Status, and Applications. The Nature Conservancy. Arlington. Virginia. Noss. et & R. L. Peters. 1995. — Status Report on America’s Vanishing Habitat m Defe nders of Wildlife, ш C. LaRoe IH & J. M. Scot . Endangered Ecosystems of the United States: | go — nary As- sessment of Loss and Degri — Biological Report 28. National Biological Sen vice, U.S. Department of the Interior, Washington Sullivan, J. 1997. An Atlas of eas rsity. Chicago Re- gion xr pis Council, Chicago, Illinois. UNESCO. 1978. Final Report, ELE mtal Confer- Environmental Education. ned y UNESCO inc voperation t UNEP. Tbilisi. USSR. 26 October 1977. UNESCO ED/MD/19. White, J. 1978. Illinois N vatural Areas Inventory Technical Report. Illinois Natural Areas Inventory. Urbana. Hli- nois ecosystems W шы Internet Resour [Bird census]: — //www. heldmuse 'um.org/birdeensus/) [Chicago Wilderness|: (www.chic: [Chicago W — ы ey (http//wwwe hic албаа. r hin [Chicago Wilderness grant]: (http://www.chicagowilderness. — bom tml [Chicago Wilderness publ. (http/www.chicagowilderness. bi ations.htm : (http: dl i state.il.us/or : (http://nhsbig. im uiuc.edu/ асом ilde mess.org ixl p/inrin/ecc h) software ww/deer. — hin I). [Urban We uch]: — fieldmuseum.org/urbanwatch) № ла ness Act]: (www.fs.fed.us/outernet/htnf/wildact. htm SAFEGUARDING SPECIES, Gary Paul Nabhan, Patrick Pynes,? and LANGUAGES, AND кш CULTURES IN THE TIME OF DIVERSITY LOSS: FROM THE COLORADO PLATEAU TO GLOBAL HOTSPOTS! ABSTRACT Hotspots of — — have become priority areas for land conservation initiatives, oftentimes without recognition that these areas are hotspots of cultural diversity as well. Using the Colorado Plateau ecoregion as a case study, this inquiry (1) — the broad geographic patterns of biological ds and ethnolinguistic diversity within this ecore- gion; (2) discusses why these two kinds of diversity are often influenced by the same geographic and historic factors; and (3) suggests what can be done to integrate traditional ecological knowle dge of indigenous peoples into multicultural conservation collaborations. Key words: biodiversity, conservation planning, linguistic diversity, traditional ecological knowledge. “Along come the sc ientists and make the words of our tablishment of Mesa Verde National Park in 1906 sto | ; EURO пай and the Grand Canyon National Park in 191« —Ag 1966 А — (Burnham, 2000), more than 11.1 million acres of "hs deir or de da — n — kanaat sa i ihe the Colorado Plateau's 130 million acres have been diversity of human cultures. ... The intriguing ques- : . | tion is this: apart from о md over resourc- federally protected for their natural and cultural re- es, will the local communities bring back some of their sources. A diverse collection of national parks and earlier cultural traditions of conservation of biological monuments, wildlife refuges, recreation areas, con- diversity? NIE _ servation areas, preserves, wilderness areas, and —Gadgil (1987) : 1 . й national historic parks апа sites, these protected The Colorado Plateau of North America (Fig. 1)? lands are managed by the National Park Service, has received international recognition for nearly a the U.S. Fish and Wildlife Service, the U.S. Forest - x . А Servire : а "Or i € € 3009 Ta. century because of the pioneering efforts there to Service, and the Bureau of Land Management (Ta formally protect its spectacular natural and cultural bles 1 and 2). Conservation efforts on the Colorado landscapes (Sellars, 1997). Despite that recogni- aleau were initiated long before our belated rec- t $ 1- ` ars, . $ c : : tion, the region's resource managers and conser ognition that the ecoregion harbors a remarkably LI ^ р ` € D € D > Е : е ы vationists have yet to work with much understand- high diversity of plants, butterflies, tiger beetles, ing of how biological and cultural diversity have and mammals compared to 109 other ecoregions of € € € ` € interacted within this four-state area. Since the es- similar size in North America (Ricketts et al., are grateful for the support of the Arizona Community Foundation, the Town Creek Foundation, the Ottens Foundation, and Agnese Haury, all of whom enabled this synthe sis. Our manuscript was greatly improved as a result of comments from Vic toria Hollowell, our Fl. agstaff colleagues Larry Stevens and Tom Sisk, David Armstrong, and two anonymous reviewers. The numerous participants in the NAU/Te Se | кана sponsored retreat, “Bridging Ecological Res- toration des Linguistic Revitalization on Indian Reservations on the ( ps rado Plateau," have greatly influenced our thinking. In particular, we thank Vernon Masayesva, Ferrel Sekakaku, Luisa Maffi, David Harmon. and — Albert. Lave * asayesva Jeanne and Donna House have also offered us guidance and insight into issues particular to the Hopi and Navajo, respectively. Aimee Goodwin was оч м for GIS analyses in ARC-VIEW, and toge et with Dan Boone and Ron Redsteer, kindly assisted us with ma ? Cente de or Sustainable Environments, Northern — University, Flagstaff, Arizona 86011, " definition of the Colorado Plateau does not exist. ШЕ researchers draw the — boundaries differently, but most agree that the Plateau's geographic heart is located in the Four Corners region, where the states of Arizona, Utah, Colorado, and New Mexico intersect. The three — ine ін d in this essay show two different boundaries of the Colorado Plateau, espec Чай, along the northeastern margin. For the purposes of statistical analysis (of the total area of Indian lands versus federally protec ‘ted lands, ete.) in this dd we used the ( Colorado Plateau. boundary outlined in Figure 3. The boundary shown in both Figures 1 and 2 is an “alter- native? delineation. .. s far as we are concerned. an objective ANN. Missouni Bor. GARD. 89: 164-175. 2002. Volume 89, Number 2 abhan et al. 165 2002 Safeguarding Species dux AA - — · - ==" m.s Un М. UINTA MOUNTAINS I. Colorado © | Salt Lake p —— m —— Yampa__ARiver Ф ° / D 2 | { | $ L N | \ A 40 4 i N / $ \ т à i | | 483 | Na * a Kilometers 134 ` 2 [. p C dM ; | © a 0 25 100 Y à | ue ] Y ; / $ WwW 2 A TS / y | ё 2 / А o | ak 2 y < С | SCR І K N | | ^ Ё l | 7 38 + | ABAJO MINS * Say, | i ' V Lapara ^ MTS. | ' v TS. —X — 4 San Juan Rivet | | | / ' / — * Y T 20 E SAN JUAN BASN f ig { we: Colorado | 8 l = : ' | d « Arizona | New Mexico $ І b sf ) Бч с СА : Mt. o, Flagstaff Ге) | dan I 6 < ' Ayr uerque Pio, | l 1 | 120 110 100 = | ip Y T Li ' 45 E — SONORAN | Ө WHITE DESERT 114 110 uns | 40 | 1 & = as | Z I2 NC 7 Nilo 435 30 | 430 25 L L 1 DE 115 110 105 =н Figure Boundary of the Colorado Plateau. For additional maps, articles, and photographs, see (http:// www.cpluhna.nau.edu/i ndex.htm). 1999a, 1999p). Although recent National Monu- ment designations such as the Grand Staircase/Es- calante have, in fact, taken into account the area's biodiversity and rich cultural heritage, these two factors have rarely been conceptually linked. More typically, they have been offered as "twin" attrac- tions for ecotourists intrigued by redrock land- scapes. Indeed, the 8 and 10 million tourists who annually visit national parks and monuments on the Plateau may receive some unanticipated exposure to this biodiversity and the ancient cultural influ- ences upon it, but that is seldom what attracted 166 Annals of the Missouri Botanical Garden Table 1. National parks of the Colorado Plateau. Sourc e: National Park Service website (http://www.nps.gov) Park State Acreage Arches Utah 76,518 Black Canyon of the Gunnison Colorado 27,705 Bryce Utah 35,845 Canyonlands Utah 337,597 Capitol Reef Utah 241,904 Grand Canyon Arizona 1,217,403 Mesa Verde Colorado 52.121 Petrified Forest Arizona 3.53: Zi ah 146,592 Zion 9 National Parks 2 in AZ; 2 in CO: 5 in UT 2,229.218 total acres ( 2.3 million acres) them to the parks and monuments of the Painted Desert or Canyonlands in the first place. Because about 13.5 percent of the Colorado Pla- teau’s landmass is already protected by federal agencies, the ecoregion’s extant biodiversity has not been considered as gravely imperiled as the bio- diversity of other regions of North America. Nev- ertheless, our recent (unpublished) survey of more than 70 environmental professionals (including Na- tive Americans) working on the Plateau indicates that this ecoregion remains unusually vulnerable to Table 2. National monuments of the Colorado Plateau. agement websites (http://www.nps.gov) and (http://www.blm. threats such as the damming of rivers; oil, gas, coal, uranium, and aquifer mining; competition from in- and the fragmentation of wildland habitats by rap- idly increasing urbanization. Such threats continue vasive species; mismanagement of wildfire regimes to diminish native biodiversity, both within and be- yond national parks. The Nature Conservancy (TNC) has therefore ranked the Colorado Plateau within the third tier of hotspots of imperiled bio- diversity. At the same time, TNC recognized that ۰ this ecoregions “rarity-weighted species richness’ Source: Nalional Park Service and Bureau of Land Man- gov.) Le) Name State Managing agency Acreage Aztec Ruins New Mexico NPS 31 Canyon de Chelly Arizona NPS (leased from Navajo Nation) 83,840 Canyons of the Ancients Colorado BLM 164,000 Cedar Breaks Itah NPS 6,154 Colorado Colorado NPS 20,453 Dinosaur J NPS 210,277 El Malpais New Mexico NPS 114,277 El Morro New Mexico NPS 1,278 Grand Canyon-Parashant Arizona BLM 1,014,000 Grand Staircase-Escalante Utah BLM 1,900,000 Hovenweep Utah NPS 784 Montezuma Castle Arizona NPS 857 Natural Bridges Utah NPS 7.036 Navajo Arizona NPS (leased from Navajo Nation) 360 Pipe Spring Arizona NPS 40 Rainbow Bridge Utah NPS 160 Sunset Crater Volcano Arizona NPS 3,040 Tuzigool Arizona NPS 10 Vermilion Cliffs Arizona BLM 293,000 Walnut Canyon Arizona NPS : Wupatki Arizona NPS 35,422 Yucca House Colorado 33 22 National Monuments 10 in AZ: З in CO; 3in NM; 6 in UT NPS BLM and NPS lands 3,859,549 total acres Volume 89, Number 2 2002 Nabhan et al. Safeguarding Species was considered more significant and less well- known than it deserved to be (Stein et al.. 2000): in other words. the Colorado Plateau harbors mans biological rarities whose vulnerability to threats has not yet been adequately assessed. Consider the fact that the relatively well endowed Grand Canyon Na- tional Park has continued to lose an average of one species per year during the last two decades: the park’s current research budget and resource man- agement strategies have somehow not been. suffi- cient to prevent the local extirpation of rare spe- cies. Putting aside for the moment the degree to which current threats imperil the ecoregion's biota. it is clear that the Colorado Plateau is indeed rich i such rarities. including endemic species (that is. species with narrow distributions that occur in this ecoregion and nowhere else). Continent-wide floris- Kartesz and Farstad (1999) affirmed that the Plateau is the ecoregion of con- пе analyses by have tinental North America with the highest rate of vas- cular plant endemism, reporting 290 species re- For the David Armstrong (in stricted to this ecoregion. the fauna of Colorado Plateau ecoregion, prep.) has recently determined that 23.6 percent of the mammals and 36 percent of the rodents exhibit endemism at the levels of species or subspecies. Vhile appreciation of the Plateaus biological uniqueness has grown, recognition of its cultural and linguistic uniqueness still lags far behind. The Colorado Plateau is home to more speakers of Na- tive American languages than all other regions in the United States combined. The ecoregion’s indig- enous peoples belong to 24 different tribes. bands. or dialect communities and represent six different language families (Table 3).* (English. of course, as well as Spanish and Basque are also spoken here.) Among the Plateau’s indigenous languages. Zuni is a language isolate, or what biologists might call an Ас- (a:shiwi) has no language of the Colorado Plateau. Zuni close relative in any other language family (Camp- bell, 1997). In addition to Zuni, the other language families indigenous to the Colorado Plateau include “endemic” cording to many linguists. Keres. a family represented by Acoma Pueblo, La- the language family represented at Jemez Pueblo: Uto- guna Pueblo, and Zia Pueblo; Kiowa-Tanoan. Aztecan. the language family to which the Hopi. Ute, Athabaskan. represented by the Apache and Navajo languages: and Paiute languages belong: and Yuman. represented by the Yavapai, Havasu- 'The numerous sources for Table 3 are found in the Literature Cited. and each source is preceded by an as- terisk (*). pai. and Hualapai tribes on the Colorado Plateau. Figure 2 shows the approximate geographic bound- aries of these six indigenous language families in 1850. when the Colorado Plateau officially became part of U.S. territory. Tribes and communities who spoke (speak) a language that belongs to one of the six families were living in these areas at that point in time. For example, Navajo and Apache speakers occ upie d the lands shown as "Southern- Mhabas- can" on this historical map. The Colorado Plateau undoubtedly ranks among the top five American regions north of the Tropic of Cancer in terms of linguistic. cultural. and bio- logical diversity. as well as in biological and lin- guistic/cultural endemism. Nevertheless. there is not a single conservation plan that takes into ac- count both the cultural diversity and the biological diversity of the region. It is as if the historic and geographic relationships between “nature” and “culture” on the Plateau are somehow irrelevant. o too hotly debated to be of value in conservation planning. While it may be reasonable for conser- vation planners to be skeptical of painting all Na- tive American land and water management practic- es as “ecologically noble,” it is also tragic that so few Native American communities have been in- volved in planning national parks and monuments adjacent to their current reservations. These federal ands were clearly parts of their historic homelands 2000). resources and the management of natural (Burnham, As a result. the management of cultural resources have typically been done by different sets of specialists, sometimes involving Native Ameri- cans in the former but nearly always ignoring their traditional ecological knowledge of the latter. This historic failure of the vast majority of con- servation biologists and environmentalists to sub- stantively engage Native American communities in collaborative work based on shared goals is both the residents of the Colorado Plateau have disappointing and ironic. It is ironic because long-term 1 substantial knowledge about the history of the local flora and fauna that is not available from other sources. Even if all their current hunting. foraging. or farming practices are not considered to be eco- logically benign by conservation. biologists (Dia- mond, 1993). this does not negate the value of their traditional ecological knowledge (as defined bs Berkes. animal distributions, densities, and vulnerabilities. 2000) about factors influencing plant and [Incidentally. Diamonds (1993) widely cited con- demnation of prehistoric peoples of the Colorado P obtain timber ateau for deforesting the Chaco Canyon area to build multi-storied pueblos has been refuted by recent strontium isotope evidence Annals of the 168 | Garden issouri Botanica M 591615 ]u313]]rp 1n0j ut c9reee sor[rurej ozensue[ O sosensue[ QI ZESTZZO OZ вәшипшшо 10 *spueq “SIGLO JOLOSIP pz 209 2,99 0696 (s93Ane]9.1 UMOUY ou) типу iun Олс EOF o[q9ng tunz E> AOL LL UESSI9N 8949M use 000 ZI I o[qonq BIZ AOC 25€ ! €90'€ UPOUP? [-емогу PMOL 619°68 ojqənd zəwəf Me ACO 009 UBOUB | -PAOIN ema] (ema, ido) ouey 28> 081 обот BEY O] Чоң 90€ СРС o[qong idoy We WOE 969"! ивѕәләҹ̧ вәләў UASA LE L6P o[q9ng випЗет HT 250€ ТРО ивзәләӘ 5919 y ULISIA GOC BLE o[qang BWOOY LE 0с 000€ Неее] эп 812909 онно зд YEC> LEE 00€€ UE o] 91 000'00€ t 219 Хет) цеш %G %SI 1 usen yon An B8C ELE ап Эрн) 2501 жс} 00€ U£29]ZV-017] әр шәціпос$ 000'€ 3Inieq urouinog uenf ueg %0 % с> ESL mor AU шәшпос 000 0t Чел jo эди], этед 20 E> OVE AA Bie amed ureqinog LEPOTI вивтрц әйе Jo pueg чесе %с> E> 8S I Heina Tampon) ed coe Hoosa14-redeaeq Mg 2562 C/9'[ uexseqeuiy es x чешпд -ruyo эцоваұ Y ed 9€9 ayoed y -redeaex opio. dure?) ec 750€ DOPZ AA TIOS ed 000'000' I redepeng 7506 7586 6t9 Еши IO ed 220'881 rednseae yy 25€ CL coe'c uexseqeui y -yey OÍBABN 000'08 (sofeAeN Jo pueg oj20ue7)) gyoopufeyoy, 20€ 7509 єс UP ASE qeu y -Jei ofPary ESOP I ofeAeN, yewey 200€ 251€ 9cc'occ ueyseqeury YEA ofeary 000'000' 2 1 uoney ofeaev 2506 25€6 000'c uByseqeyly-xery aqoedy 000'£9 ofesey oure[y LEI 250€ 000'CT uexseqeqy -e43 ayoed y 000'009" 1 əyəedy urejunopy AYM %0 250€ OL ueseqeuy YEA aqoedy cg auoedy ошо], 256 HET 00€'€ ueyseqeu y -xe4 ayoedy [09°68 auoedy ejpueorf 81—@ so3e siaxeads зләчшәш uonerpgje A ruej dnou3qns a3en3ue] (sə1ə2) Хиипшшоә 10 *pueq *aqup[ ‘sraypads əaneu uan [equa adenzue] Jo 93en3ue"] 3seqpue[ [equ едо, әлцви Juang jo a3ejuadiad po[[o1ua [eio] jo a3ejuadiad ojeuixo1dd y 3jeuirxoadd y (м) Х51ЛӘ15Р ue Aq popaoaud S1 991nos qoe] “UOLIS вәәпәлә}ә 34] ut UMOUS әле eyo sių} 10} S3JINOG :310N ‘sağen due] snouaZiput S,ne3]e[d оре1о[о? эч jo snes СС 3IqeL Volume 89, Number 2 2002 Nabhan et al. Safeguarding Species 169 е € 5 Via ou ENS e foes TEGAN SOUTHERN-ATHABASCAN Locations of Indigenous Language Families on the Colorado Plateau in 1850 GIS Support From Figure 2. y 190 Kilometers E Locations of indigenous language families on Center for Sustainable Environments that the Chaco Anasazi obtained their ponderosa and spruce/fir beams by cutting a single age class of trees selected from two large mountain ranges 50-75 miles away (English et al . 2001).| Indeed, CANYON CONSERVATON TECHNOLOGIES the Colorado Plateau in 1850. (From data developed by the the more pertinent question is whether the tradi- tional ecological knowledge of Plateau tribes is be- ing lost at such a rapid rate that it will no longer be of service in biological conservation (Nabhan 172 Annals of the Missouri Botanical Garden ligations (L. Masayesva Jeanne, in Nabhan $ Reichhardt, 1983). Traditional ecological knowledge can also be useful in locating and staging rare species reintro- ductions and habitat restoration. At a time when the ( were on the verge of failure, Rea (1981) proposed that the recovery team broaden its view and con- sider reintroductions in the historic refugium near the Colorado River's Grand Canyon, where oral his- tories from the Hopi recall sightings of the giant call Awaatoko (Whiting unpublished, in 1993). This has, in fact, turned out to be one of the more successful areas for Condor rein- ;alifornia Condor recovery efforts in California bird they Seaman, troduction, even though it is outside of California where many biologists erroneously presumed that the California Condor belonged (Rea, 1981). INVESTMENT IN CONSERVATION CAPACITY-BUILDING FOR NATIVE REMAINS CRITICAL AMERICANS For decades, the Bureau of Indian Affairs hardly involved Native Americans in so-called technical decisions regarding biological resource and land management options on tribal lands. While the number of Native Americans trained in biology and nature resource management increased fivefold be- tween 1970 and 1999, there remains a chronic un- derinvestment in training Native American profes- sionals in biodiversity conservation on tribal lands relative to the training available for professionals managing federal lands. The Navajo Natural Heri- tage Program, first sponsored by The Nature Con- servancy, has become an outstanding example of the "conservation -off" of such investment (Stein et al., 2000). The Hopi, Zuni, and White Mountain Apache tribes have also developed pro- grams that include wildlife management, endan- gered species recovery, and ecological restoration. In the case of the Zuni, their program to reintro- duce beavers as part of the tribe's riparian resto- ration efforts has involved elders teaching youth about the cultural significance of beavers and other wetland dwellers (Albert & Trimble, 2000). When both Western academic and traditional ecological knowledge are valued by such programs, it ensures that Indian youth interested in natural history are not forced to choose between their own cultural way of looking at the natural world, and the way of mod- ern science. One superlative example of such ca- pacity-building efforts among Native Americans is the EPA-funded Institute for Tribal Environmental Professionals based at Northern Arizona University, which has already trained more than 500 tribal members from over 200 communities in inventory, monitoring, and restoration techniques for environ- mental protection. MORE LINKAGES ARE NEEDED BETWEEN ECOLOGICAL RESTORATION AND LINGUISTIC REVITALIZATION In addition to the many ecological restoration projects recently initiated on tribal lands, most Na- tive American communities on the Colorado Pla- teau are actively engaged in language stabilization 1996). Т include hese language the elaboration of the entire extant lexicon of a tribe, and revitalization efforts (Cantoni, restoration projects often including its names for plants, animals, and their habitats. In eliciting these names, linguists often do not know the particular scientific taxa the words are describing, but nevertheless record. ecological in- formation that may be of use to ecological restora- tion. and recovery projects (Nabhan, 2000a). For instance, the Diné name for the Indian paintbrush, Castilleja lanata A. Gray, is dah yütih- "big hummingbird's food," to spec ies idaa tsoh, meaning distinguish it from the beardtongue, Penstemon bar- batus (Cav.) Roth, called dah yiitihidaa'ts‘ooz, or “food of the slender hummingbird.” It may be that Diné elders recognize that these two flowers are frequented by different sizes and species of hum- mingbirds, and that such information can be inte- grated into ecological restoration efforts to increase = forest understory biodiversity. CONCLUSIONS: IMPLICATIONS FOR GLOBAL CONSERVATION STRATEGIES Over the last decade, tremendous effort has gone into defining, locating, and rapidly assessing the hotspots of biodiversity” (Mit- 1998; these species richness of * termeier et al., 1998; Olson & Dinerstein, Ricketts et al., 1999a). hotspots not only capture a large proportion of the planet's biodiversity, but they also define an agenda By some accounts, for conservation investment, which is largely ex- pended on land purchases and infrastructure de- ve opm nt for protected areas (Mittermeier et al., 38). By investing an average of $40 million/year on land purchases in hotspots, rather than on the — current “scattergun” approach of current conser- vation expenditures, these conservation strategists have proposed a new “silver bullet” to biodiversity loss that could protect areas containing a high pro- portion of the world’s species, while business out- side the hotspots could be allowed to go on as usu- ^ mE — — Ehrenfeld (in press) has pointed out the serious 170 Annals of the Missouri Botanical Garden 2000a). To date, this rapidly disappearing orally Americans on the Colorado Plateau have vet to re- transmitted knowledge has rarely been systemati- cally taken into account and valued by land man- agers, who often consider such knowledge to be the quaint but anecdotal or superstitious recollections of scientifically untrained old-timers. Even contemporary Native Americans’ roles in managing the lands of the Colorado Plateau have been chronically underestimated. While the Grand Canyon Trust considers “Indian country” to com- prise just a quarter of the Colorado Plateau’s 130 1996: Wilkinson, 1999), our GIS-based analyses more accurately es- tablish that 29-32% of the Colorado Plateau is be- ing managed by Native American communities (de- million acres (Hecox € Ack, ending upon which definition of the Plateaus | 8 Ul t would be presumptuous if not impossible to develop boundaries is used). By virtue of this fact alone, a systematic conservation plan for this ecoregion that did not take into account Native land stewardship and traditional ecological knowl- American edge of endangered species on the Colorado Pla- teau (Nabhan, 2000a, b). In a modest effort to begin to bridge the historic gap between studies of biological and cultural di- versity, Northern Sustainable Environments (CSE) has initiated a pi- Arizona University’s Center for lot study with the Grand Canyons Wildlands Coun- cil to assess ways to better safeguard and restore both the biological and the cultural uniqueness of the Colorado Plateau. In designing this pilot study, we have investigated the potential linkages between biological and linguistic diversity elucidated by the scholars involved in “Terralingua: Partnerships for Linguistic and Biological Diversity,” an internation- al non-profit organization hosted by CSE, which has recently published a global analysis of biocultural (Maffi, 2001). emphasizes why efforts to conserve biological and diversity The following discussion linguistic diversity should be linked, whenever pos- sible, using examples from the Colorado Plateau. BIODIVERSITY ON NATIVE AMERICAN LANDS CONSERVATION INVESTMENT THE DESERVES MORE Indian reservations in the United States cover tens of millions of acres of the North American con- tinent, comparable in extent to the acreage that the National Park Service manages for North America’s biodiversity. In particular, reservation lands on the Colorado Plateau (Fig. 3) cover nearly twice the area that national parks, monuments, preserves. conservation areas, wilderness areas, and wildlife refuges cover in the same ecoregion. Nevertheless, the 29.6 million acres of lands managed by Native ceive much investment from federal or private sources for the inventory of their biodiversity, mon- itoring, and recovery of their rare species, or local capacity-building in environmental protection— relative to the considerable support given to those working on adjacent federally protected lands. If all the species found on Indian lands were also found on lands rigorously protected by the National 'ark Service or The Nature Conservancy, perhaps this issue would be easier for conservation biolo- gists to ignore. However, the narrowly distributed endemics of the region are often restricted to hab- itats found only on reservation lands, and not on park lands. The Navajo sedge, Carex specuicola J. T. Howell, is an endangered species found around only three springs and seeps used by Diné (Navajo) livestock herders, and one spring used by Hopi farmers. In other words, its range is restricted to the Navajo and Hopi Reservations (D. House, pers. comm.; Nabhan et al., 1991). Similar situations oc- cur for the endemic Hopi chipmunk (Tamias rufus). a subspecies of the Spotted Ground Squirrel (Sper- mophilus spilosoma cryptospilotus). the Chuska and unitcha Mountain subspecies of Abert's squirrel (Sciurus aberti chuskensis), and a subspecies of Ste- phen’s woodrat (Neotoma stephensi relicta) found only on Navajo lands. Even if conservation biolo- gists continue to feel ill-equipped to deal with the cultural and legal (sovereignty) issues regarding plants and animals restricted to tribal lands, they can no longer ignore the fact that the only means to sustain these species is by providing Native American land managers with the resources needed lo protect or recover these rare populations and their habitats. CONSERVATION COMMUNITY NEEDS TO TRADITIONAL ECOLOGICAL KNOWLEDGE THE EMBRACE Until recently, endangered species recovery teams and ecological restorationists disregarded the Native Perhaps the first formal traditional ecological knowledge found American communities. breakthrough occurred when Diné biologist Donna House incorporated traditional ecological and eth- nobiological knowledge about the Navajo sedge into its federal species recovery plan (House in Nabhan et al., 1991), acknowledging that Diné herders had been stewards of this plants habitat for decades further efforts to and deserved to participate ii safeguard the sedge. In addition to their role as habitat managers, some cultural communities on the Colorado Plateau intentionally protect rare species, as is the case Volume 89, Number 2 2002 Nabhan et al. Safeguarding Species t % = — z 2 XK © Sr A = / \ % e Northern Ute / s N * è oA р p, 1 \ $ A d ~ k OR сусту" \ T. Oph ute Р \ ' ۱ | * | [фр ) ee / / / | 4 е Grand Junctíon Paiute / е FI X | L p A. E 24 N ) f А / E / Ё * Ж E Southern’ A Paiute, Ute Mountain | Southern д ; \ ( Southern Ute Paiute outhern Durango g^ | — m ha 3 Jicarilla Apache 5: Cur | — y SER \ А — Г Havasupai / ° \ Flagstaff Prescott Camp Verde $ * \ | \ Native American Indian Reservations ~~ on or near the Colorado Plateau Native American Reservation p Colorado Plateau 5 Bound Figure 3. Native American Indian Reservations on or near the Colorado Plateau. with the anomalous sunflower (Helianthus anomalus tices (Nabhan & Reichhardt, 1983). Its flowers may S. F. Blake). Of the twenty-five known localities for be the sole source of a ceremonial facepaint pre- this species, at least seven are in or adjacent to pared for the Lakon mana (maiden) ceremonies of Hopi fields and kivas, where farmers and priests early fall, so that these sunflower populations are spare it from their otherwise rigorous weeding prac- protected as a necessity for fulfilling spiritual ob- Volume 89, Number 2 Nabhan et al. 173 2002 Safeguarding Species flaws associated with such an agenda, the most se- lands held in trust for future generations of in- rious being that most hotspots are not only inhab- digenous peoples (Burnham, 2000). ited by diverse cultures, but they are also not for œ It is — that for such high levels of biodi- sale. By the mid 1990s prominent ethnobiologists, versity to persist in any hotspot until this day, anthropologists, linguists, and biogeographers had —— residents consciously or unconscious- brought attention to the fact that the 10-20 richest ly developed active and passive means of man- biodiversity hotspots were also extremely rich in aging particular landscapes, microhabitats, or cultural diversity, which was typically defined in species guilds to maintain them. The potential terms of the richness of extant indigenous languag- utility of such knowledge, skills. and practices is es (Harmon. 1995). For instance, Toledo (1994) ob- great (Nabhan, 2000b), so much so that species served that of the nine countries considered to have recovery teams and formal land managers should the highest species richness of mammals, six of be encouraged to integrate such knowledge into these were also among the richest in indigenous their conservation management plans (Tuxill & language groups: Indonesia, Brazil, Mexico, Zaire, Nabhan, 1998). India, and Australia. e It is likely that proponents of biodiversity con- Three commentaries are typically offered to ex- servation will forge stronger, more effective col- plain this correlation and to hypothesize its causes laborations with indigenous residents in biodi- (as cited by various authors in Maffi, 2001): versity hotspots if they listen to and respect indigenous concerns about sovereignty, cultural * Geographic determinism: Both species and lan- property rights, and secrets associated with eso- guages diversify in heterogeneous. landscapes LUPIS aul — — c У. » f H sts with formidable geographic barriers. : . ‘ a af A EOM "EET should attempt to collaborate on community- * Ecological determinism: Linguistic diversification DUM ueste Qai Ner ihe асана bd . . . H . “гг c 5 >C S c 5 > > M ? t "n e “AR : o rere qe эү. occurs in response to high biodive хаку, as differ transmission of traditional ecological knowledge, е — фо — — T DM sss M indicis languages, eds niches for themselves and encode their knowl- | Frich bi НТ, tional subsistence or ceremonial practices, and edge of rich biotas in different ways. ; : — "ir ea P SS the generation of health benefits or income for a * Historic determinism: Areas of high linguistic di- eh АВ i Жолини ашин” , ; с e Oi ross section of co / members. versity today are "residual," persisting because We i m i "ede. e ee — uc uw While such community-based conservation efforts of their geographical isolation from extensive ag- BE e lta asia fs | | жент! devel а, 1 are already given lip service from governments ricultural development (implying that now-impov- ~ І i І . ished edi and NGOs alike. there remains a disproportionate erished areas were once more diverse). ; investment in “top-down” conservation strategies Clearly, our Colorado Plateau pilot study reiterates and only scattered. investment in community- a pattern seen on other continents as well: where based “bottom-up” strategies for maintaining bio- hotspots of biodiversity or endemism are inhabited diversity. by a diversity of cultures, the ways in which these * Given the fact that traditional ecological knowl- cultures encode traditional ecological knowledge edge about landscape-level biodiversity itself i about species in their native languages has tremen- being diminished (Berkes, 2000), the fragile re- dous potential for helping to conserve this biodi- lationship between the two needs to be more tan- versity (Nabhan, 2000b). From this frequently ob- gibly explored. While many biogeographers and served correlation, several management linguists are already involved in efforts to under- implications must be considered: stand global patterns of biodiversity and cultural diversity, it is hoped that our pilot project to safe- e It may not be feasible or even advisable for gov- guard the uniqueness of the Colorado Plateau will ernment or non-government organizations (NGOs) encourage other community-based. practitioners to purchase the lands within these hotspots for to experiment with more practical MAN ION biodiversity protection, because much of the bi- tegrating indigenous knowledge into collaborative ologically diverse area may be “common lands” efforts to conserve the world’s hotspots of diver- held in trust by these cultural communities, such sity and endemism. that they cannot be purchased, traded, or con- demned. New efforts to conserve. biodiversity elsewhere in the world need not make the same Literature Cited Latera + Naton ork arias Ha do TEN Е mistake the National Park а historically Agnon, S. Y. 1966. Editorial commentary. The Observer made on the Colorado Plateau, by stealing, con- 5(1): 6. demning. or finagling the rights to species-rich Albert, S. & T. Trimble. 2000. Beavers are partners in 174 Annals of the Missouri Botanical Garden riparian restoration. on ed — Indian reservation. Ecol. Restoration 18(2): € Armstrong, . in prep. е. in mammals of the Colorado Plateau. Draft analysis in preparation by the author, Department of Environmental, Population and Organism Biology, University of Colorado, Boulder, Col- orado. *Benson, — dad o of Utah. view. 12 Decembe — 2000. el Ecology Telephone inter- Yale Univ. Press, New — P. 2000. Indian Country, God's Country: Native Americans and the National Parks. Island Press, Wash- ingtor *Campbe ll, т, 1997. American Indian Languages: The Historical ee of Native America. Oxford Univ. Press, New Cantoni, G. (editor). 1996. Stabilizing Indigenous Lan- guages. Northern Arizona University Center for Excel- lence in Education, Monograph Series, Flagstaff, Ari- zona. *Cuch, Mariah. Ute Tribe Public Relations. P.O. Box 400, Ft. Duchesne, Utah 84026, Electronic mail correspon- dence, 27, 29 December 2000. оо J. 1993. The Third Chimpanzee. W. W. Norton, or Ehrenfeld. In press. Hot Spots and the Globalization of coran Oxford Univ. Press, te ford tu . Betancourt, J. 5. Dean & J. Quade. 001. Ае isotopes reveal d sources of ar- chitectural timbe sr in Chaco Canyon, New а! . Acad. Sei. 98: 11891-11896 . Wilfred. — red Zuni Public . Personal intery November 2000. Zuni ; Mexico. *Estes, James. October 1999. “How Many Indigenous Languages Are Spoken in the United States? By How Many Speakers?" National Clearinghouse for Bilingual Education. (htt tp://www.ncbe.gwu.edu/asknebe/faqs/ 20natlang. epe Gadgil, M. 1987. шо, Trends EU Evol. 2(12): : *Gallegos, Estrella. Sia nd 2 anguage De — Ute Mountain Ute Tribe. Telephone interview. 1 Noven ber 2000 = = Mexico. Proc. ч and biological. 13. *Glassco, Greg. Cultural Resources Office, Hualapai 2 Tribe. Peach Springs, Arizona. Telephone interview. x anguages of the World. dated February 1999 at TT ل‎ org/e e le йш, D. 1995. The status of the world’s languages n ntes in ioe Southwest J. Linguistics 14: 2" — W. К. Ack. 1996, Charting the Colorado Plateau: An — ne De ‘mographic Exploration. Grand —— Trust, Flagstaff, Arizona. *Hussey, Jenni en es Direc tor, Paiute Tribe of Utah — чанган, Cedar City, Utah. Telephone in- terview. 20 De М. — J ^cember 2 A. Farstad. 1999, Multi-scale analysis of in- — all web sites shown) that were used to construct Table — *All references denoted with an asterisk are sources 1 endemism of vascular plant ege ies. Pp. 51-55 in T. H. Ricketts, E. Dinerstein, D. M. Olson & C. Loucks, Ter- restrial Ecoregions of North America: A Conservation Assessment. Island Press, Washington, D.C. *Krauss, Michael. 1996. Status of Native American Lan- suage Endangerment. Pp. 16-21 in G. Cantoni — Stabilizing Indigenous Languages. Center for el- lene e in Education, Northern Arizona pop Flag. aff. Maffi, L. 2001. On Biocultural Diversity. Smithsonian Uni- ity Press, Fie ass ee py a Mittermeier, R. . Myers, . Thomsen, G. A. b. Fonseca & S. ‘Olivieri 1998. Biodive 'rsity hotspots and major tropical wilderness areas: Approac 3 ч conservation priorities. Conservation Biol. 1 520. Nabhan, G. Р, 2000a. Interspecific M. jene endangered species recognized by sire and Com- caac о ures. Ecol. Appl. 10: 1288—1295 2( The — Plateau as a center of « versity. idend Newsletter, The ( Consultative Group on bu dn Diversity, San Fran ncisco. . L. Reichhardt. 1983. Hopi protection of He- — avt x a rare sunflower. South W. Naturalist 28: 231-2306. ‚р, — ue Suzan, W. Hodgson, L. Hernandez & G. Malda. . Conservation and use of rare plants by — culo of the U.S./Mexico borderlands. 1: ‚ Oldfield & J. Alcorn (editors), Bio- aren Conservation and Ecodevelopment. Westview Press, Boulder, Colorado. ro Dorothy. Cultural Preservation Office, Southern Iribe. Ignaci io, Colorado. Telephone interview. 1: : E . Dinerstein. 1998. The Global 2000: A representation approach to conserving the world's most biologically valuable ecoregions. Conservation Biol. 12: 516—520. *Pino, Jackson. Middle School and High Sc ue teacher. Alamo Navajo ош Ala p o, New Mexico. Tele- phone interview. 15 November = *Polacca, Mary. Offic of Hopi Tribal Enrollment. Tele- phone interview. 10 December 200( *Poleahla, Anita. Office of Hopi — Preservation. Telephone interview. 10 December 2000. Rea, A. M. 1981. California Condor — E A recovery proposal: Environm. ree 192: € Ricketts, T. H., E. Dinerstein, Olson & A bs icks. 1999a. Who's where in North fue rica? BioScience 49: 369—381. Я 3 & 1999}. Terrestrial Ecoregions of North America: A Conservation Assess- ment. Island Press, Washington, D.( *Rollo, Gail. Tribal Administrator. Paiute Tribe of Utah. Telephone interview. 11 September 2000. *Salari, Nasim. Program — и Juan Miro Paiute Tribe. Telephone interview. 11 September 20( , Merton. C ultura Tribal Secretary, Tonto Apache Tribe. ayson, Arizona. Telephone interview. 11 September JO. Seaman, P. D. 1993. Users Guide and Index for the A. ting Collection. Northern Arizona Anthropologic ај ee 1, F — aff. *Sec — ro, Tony. President of Tohajilee (С e ito Nava- j. Telephone interview. 23 October 200( Volume 89, Number 2 2002 Nabhan et al. Safeguarding Species 175 Sellars. R. W. 1997. Presen туш Nature in the National Parks: A Historv. Yale Un y Se s. New Haven. “Smith, Joni. Acestes ve qoid Camp Verde Y — Apache Tribe. Telephone interview. 11 Septe “э г a L.S. Kutner & J. S. Adams. 2000. gre ‘TOUS Heritage: The Status of Biodiversity in the ed . Press, Oxford. *Tohtsoni, Nathan. ° W ho is a Navajo?” Navajo Times. 31 August 2000: 1. Toledo, У. М. — Biodiversity and к diversity in Mexico. Different Drummers: Ince North American Biodiversity 1: 16- 9. Tuxill. J. & G. P. Nabhan. 1998. Plants and Protected Areas. World Wildlife Fund чен and People Manual. Stanle y Thornes Publishing. Cheltenham. — Office. White нанні Ара he Arizona. Telephone interview. 13 Also (http://primenet.com) уж ate informa- уез for Protecting J q Nation. November tion). "Ме mvylewa. l'elephone interview. 11 S Wilkinson, €. 1999. Fire on the Plateau: Conflict and En- durance in the Southwest. Island Press. Edward. Zuni — nt of Game and Fish. eptember 2000. American Director of Cultural Center. *Wil Jicarilla Apache Nation. Dulce, y Mexico. Telephone inter- view. 13 November 2 o. E Prescott Indian Tribe (brochure). The Yavapai Prescott Tw Tribe. 530 E. Merritt; Prescott. 80301. Arizona *Web Sites Consulted (Unless otherwise noted, all web sites listed below were 1) accessed during December 2000 ог January 200 Ap //www.itcaonline.com) ii c — Hopi. Kaibab 1 Band of Paiute ен Yava ДЇ na.com) [Camp Уе PU Pid \pac i | — yavapai-apache-nation.com) [Camp Verde Ya- vapai- Apache] http:// Iced.state. » — [Paiute Tribe of Utah] 1 up н rnment/ е. мт) ‘Southern Ute Tribe] http: //www.ubtanet.com/~northernute) tribe | (http://www.dced.state.ut.us/indian/Today/ute.html) [Northern Ute tribe] (http://epa.gov) [Pueblo of Acoma] (http://www.zuni.k12.nm.us) — Pueblo] (htup://www.swr-hq.nm.bia.gov/ACRES-19.99.htm) [Bu- reau of жк Affairs: ا‎ st Region Indian Acre- As of 31 December 1999” | http:/ hq.nm.bia.gov/99ABQIf.htm) [Bureau of dion Affairs: (http://www.indigenous- ias org) [Indige nous Lan- titute. Santa Ке, New Mexico: “Endangered Accessed 5 December 2001 | i Prescott Tribe| Li "n :// ~ [Northern Ute “Баенов Region Combined Totals” | guages Ins Languages Database.” Richard N. Mack? and Marianne Erneberg? THE UNITED STATES NATURALIZED FLORA: LARGELY THE PRODUCT OF DELIBERATE INTRODUCTIONS! ABSTRACT Iver the past 400 years plant immigrants have arrived in the United States in huge numbers, the results of accidental and deliberate introduction by humans. Very few immigrations, however, result in naturalizations: the environmental hazards for immigrants in a new range are usually acute and chronic. We traced the history of introduction among the naturalized taxa within a group of U.S. regions and states that span much of the environmental amplitude of the United States. Despite differences among these regions in physical environments and the length of their histories of plant ае the — of their naturalized floras that likely arrived through delibe rate introductions is greater than 50 pe t. Many of the first species to become naturalized in the northeastern United States were introduced as food seasonings or putative sources of medicine. Erstwhile ornamentals are prominent components of all the naturalized floras, especially in Florida. Some species introduc ‘ed as sources of forage or fiber also became naturalized. Before 1900 many now naturalized species were introduced re peatedly and widely into the United States as seed contaminants through an extensive international commerce in crop seeds. The likelihood that the majority of species now naturalized in the United States has a history cultivation provides a plausible explanation for their biotic invasion, environmental stochasticity, naturalization, ornamentals, seed catalogs, seed contami- and post-immigration ¢ Key words: nants The mode of entry is unknown for approximately 30 percent. of deliberate introduction r persistence in a new range. Comparisons of the sizes of regional naturalized floras with the total number of species that have likely arrived in these new ranges produce a com- monly shared conclusion: few immigrant species ever become naturalized (sensu Mack et al., 2000). Williamson (1993) argued that this difference in the number of immigrant species, the number that re- side temporarily (adventive or casual species), and the number that subsequently become naturalized differ in turn from each other by an order of mag- nitude. Accurate determination of these proportions is rarely possible (but see Holdgate, 1964). Nev- A species that arrive in a new range and the number that eventually persist appears huge (Williamson € Fitter, 1996). Today more than 30,000 plant taxa ~ ertheless, the discrepancy between the number are commercially available in Australia (Hibbert, 1999), even more in the United States (Isaacson, 1996); most are routinely cultivated outside glass- houses or other shelter. Yet even liberal estimates put the total size of the naturalized flora for each country at less than 3000 species (Hnatiuk, 1990; Kartesz, 1994). Pursuit of explanations for the long odds against naturalization that confront immigrant species have long attracted the attention of biologists (de Can- dolle, 1855 planation would aid substantially the prediction of ; Gray, 1879) because any general ex- which non-indigenous species will persist. This fas- cinating topic in ecology and biogeography has re- cently taken on added importance as nations grap- ple with erecting legally defensible scientific ! We thank D. R. Gordon, С. Kinter, К. . R. Wunderlin for valuable — е and олш, and K. manuscript. Lonsdale, H. 2. roves r Agricultural Library. Support for M. E. was provided by Effects and. Risks of Biotechnology in — ulture, Theme I and the Royal Veterinary and Agricultural Univer ? School of Biologic al Sciences wsu.edu R ed E Ecology, Royal Veterinary and Agricultural University, 1871 Frederiksberg address: Depart — — dk ANN. Missourt Bor. SIRO- E (Ca кыы, provided invaluable advice during the comple tion « б this manusc e "We thank the staf Academy of Natural Sciences oe the New York Botanic al бш. Old, K. Schie те — k. C. Burks and V Weidema, L. Wester, a — Il for helpful comments on the sistance and the libraries of Washington State | Iniversity, the the — of Congress, and the National e Danish Environmental Research Programme (Centre for II), the Danish National — Research Institute, ‚ Washington wl University, Pullman, Washington 99164, U.S.A. rmackG mail. C. Denmark. Present rent of. Terrestrial. Ecology, National Environmental Research Institute, 8600 Silkeborg, Denmark. GARD. 89: 176-189. 2002. Volume 89, Number 2 2002 Mack & Erneberg US Naturalized Flora 177 protocols by which they may exclude harmful non- 1999). Despite numerous attempts to identify the few that indigenous species (Glowka € de Klemm, will become naturalized among a host of immigrant . Rejmanek € Richardson, 1996; Rei- 1997), explanation for this numerical discrepancy, much ) spectes (e.g. chard & Наша, we still lack a general less a predictive protocol (Williamson, 1999). Explanation lies in a better understanding of the interaction between immigrants and the environ- ment in a new range. All populations are subjected to multiple, random environmental events and cir- cumstances, whether in their native or new range (Simberloff, 1988). detrimental events, e.g.. prolonged or acute drought The consequences of. random or frost, sudden and massive attack by predators or virulent parasites, are particularly devastating for size of a small populations. Although the sheer arge population provides some numerical buffering from repeated losses of a percentage of its mem- bers. the same percent losses for a small population can rapidly bring it to extinction, unless members are replaced through reproduction or immigration. or both (Menges, 1992, 1998). Many, and perhaps most, immigrant populations are small founder pop- ulations for which there is no likelihood that indi- viduals will be added through recurring immigra- tion. The randomly occurring losses caused by ubiquitous environmental stochasticity can readily prove dire (Mack, 1995). erant population with attributes that allow it to tol- Thus, even for an immi- erate the average expression of conditions in a new still chance, repeated occurrence. of extreme environ- 1989: Crawley. 1989). The raison d'être for cultivation is the deliberate environment, it. could be destroyed by the mental events (Crawford. protection of plant populations from environmental hazards, including those with stochastic expression. Such protection can allow the population to reach a numerical threshold, such that it can sustain loss- es arising from subsequent stochastic events. M that threshold size the population may become nat- uralized. even if cultivation is withdrawn. Thus. cultivation emerges as a potential counter-force to environmental stochasticity and may well facilitate naturalization (Mack. 2000) If there is a causal link between cultivation and plant naturalization, we hypothesized that there should be a corresponding correlation between cul- tivation and the history of those species that have become naturalized. For example. in any natural- ized flora. the majority of species should have a history that includes cultivation upon entry into the new range. Clearly, some species have no such link: immigrants could arrive in a new range and become naturalized without human dispersal or cultivation Ridley, 1930; Mack & Lonsdale, 2001 hypothesis is correct, these examples would. how- (e.g.. ). If our ever, form a minority of naturalizations. To test our hypothesis we investigated the mode of pre-1900 introduction among the naturalized species within a group of continental U.S. floras and compared this evidence with similarly con- structed tallies for naturalized floras from South Australia (Kloot, 1987), Hawaii (Wester, 1992), and 2000). substantially Time to natu- (Mack et al. 2000). We attempted to circumvent most uncertain northern Europe (Weidema, ralization can vary cases of naturalization by restricting our investi- galion to species that had arrived in the United States before 1900. uralized is based on at least 100 years of residence in the United States. As a result, their status as nal- METHODS We recognized that our investigation could be complicated and handicapped by the sometimes conflicting synonymy of binomial names across the last two centuries and differences among floras in the de finition of naturalized species. We selected published U.S. floras or checklists that identify nat- uralized species as those that persist without aid of recurring human cultivation: furthermore, these flo- ras clearly make a distinction between adventive (temporary, waif, transient) species and naturalized species. Our selection criteria are met by recent floras or checklists for New York (Mitchell, 1986). Rhode Island (Gould et al., 1998), Florida (Wun- Чегип. 1998). and North-Central Texas (Diggs et al.. 1999). We did, however, delete 46 species that Wunderlin (1998) considered. naturalized because he reported that each of these species has been collected only once or has nol been collected poe- cently. We employed Fernald’s (1950) Gray s Man- ual of Botany, 8th ed., the northeastern quarter of the United States. a re- because of its coverage of gion with diverse habitats, and its frequent descrip- lions on the mode of introduction for non-indige- nous species. These floras collectively represent much of the diversity within the continental U.S. naturalized flora. We attempted to determine the history of intro- duction for each naturalized angiosperm in these floras. We deliberately excluded other plant groups from our study as the records for some of these groups (e.g.. eryplogams) are much less complete and reliable. Information is rarely collected on a species’ introduction event(s), although some intro- ductions can be traced detail (e.g... Haughton. Annals of the Missouri Botanical Garden 178 Table 1. Summary of the likely mode of entry (deliberate, seed contaminants, as both deliberate & seed contami- nants, and unknown) for angiosperm species now naturalized in Florida, north-central Texas, the central northeastern United States, Rhode Island, and New York as derived from published floras (see footnotes 1—5). Numbers in all but the last column are the percentages of the total naturalized flora that arrived by a mode of introduction. Delibrate : Deliberate & seed introductions contaminants contaminants Unknown Number of taxa Florida! 67 1 l 31 1161 North-central Texas? 62 2 3 34 417 Central northeastern United States? 59 3 3 5 559 Rhode Island 59 3 3 35 408 New York? 57 3 4 7 391 ! Wunderlin (1998). ? Diggs et al. (199€ * Fernald (1950). ' Gould et al. (1998). 5 Mitchell (1986). Sources for the modes of introduction: Anon. (1882), Austin (1978), Bailey (1906), Barton (1818), Blake (1922), Coon (1974), Darlington & Wylie (1955). пе (1826). I (1832), Diggs et al. (1999), Duke (1985), (1997), Gordon & Thomas (1997), Grieve — Josselyn (1672), Lamson Scribner ey ls Stre Mack (1991), Mahler (1980), Mohr (1878). Morton (1976). Morton (1989), Muenscher (1955 Old (1981), үре г (1915), Pursh (1814), Rhoads & Klein (1993 Spencer — Sullé & Maisch (1880), "ке (1919), ' 8). Smith (190 (1969), a rlin (1 1978; Mack, 1991; Schmitz et al., 1991, and ref- erences therein). In the absence of contemporane- ous records of a species! introduction, we estab- lished criteria for identifying the most. plausible introduction scenario. Species considered here to have arrived through deliberate introduction have a history of pre-1900 human use within the species native range or in the United States. We reasoned that if a species had been used for centuries as, for example, a seasoning or herbal in western Europe (Sturtevant, 1919; Grieve, 1959), troduced deliberately by human immigrants to the Dipsacus ful- it was likely in- United States (e.g., Nepeta cataria L., lonum L., Taraxacum officinale Weber Crop seeds have been imported to North America by Europeans for more than 400 years (Viola & Margolis, 1991). Throughout most of this time seed cleaning has been either not practiced, ineffective. or even subverted (see Discussion); consequently, arriving as a seed contaminant became a likely ac- cidental mode of entry for some species. Species were considered to have arrived by this mode if they had a pre-1850 history as seed contaminants in Europe or the United States (e.g.. de Schweinitz, 832 For many species a pre-1900 record of use is in herbals and horticultural compendia e.g.. Bailey, 1906; Fernald & Kinsey, 1943; Grieve, 1959; Uphof, 1968), plant materia medicas recorded Jarlington (1859), Darwent & Couplanc — (1989), Fernald (1950), Fernald & Kinsey (1943) ‚ Haughton (1978), Hitchcock (1950), Howell (1959), ‘Hume et al inge (1977), Long (1922), McCarthy (1888), McCartney (1984), s Schweinitz eira et a (1960), э). — & Westover Я , Ridley ei Schery (19€ Schmitz et al. (1991), Tanaka (1976). U phof (1968), — et al. — == (Sullé & Maisch, 1880), pre-1860 U.S. floras (e.g.. Pursh, 1814; Darlington, 1859), and 19th century state and federal agricultural publications (see foot- notes in Table 1). In addition to the primary liter- ature, we searched 19th century seed catalogs that had been distributed in the eastern United States (Mack, 1991, logs not only report a large array of non-indigenous and references therein). These cata- species for sale, but most also state each species’ binomial name and uses (Appendix 1). Fernald (1950), Wunderlin (1998), and Diggs et al. (1999) indicated a mode of introduction for some species, and their determinations supplemented our other sources. We initially searched in the pre-1900 literature for species under their currently accepted names. Fortunately, many plant species names have been retained since their initial description [e.g.. Rumex Vill.]. although we crispus L., Stellaria media (L.) usually cannot verify the species’ identification in a historic account. Some naturalized species in the modern floras were left without an assigned mode of introduction because we could not reliably iden- tify them in the 19th century literature. These were placed in the Unknown Category. The state or re- gional floras from which we prepared our tallies all include some naturalized subspecies and varieties. We usually could not identify these taxa within the pre-1900 botanical literature, and unidentified taxa Volume 89, Number 2 2002 Mack & Erneberg US Naturalized Flora 179 were also assigned to the Unknown Category. The Plant (http:// www.ipni.org) was employed as the nomenclatural International Names Index standard for plant names and authorities. RESULTS Results of our investigation of the history of in- troduction among these naturalized floras are sum- Table | Information on the mode of introduction for marized in ‚ (Only this summary is provided here. each investigated species is available from the au- thors upon request.) The majority of angiosperm species within each naturalized flora has a pre- 1900 history of use in their native range or the United States, or both. The percentage of such de- liberate use is similar among the floras: lowest with- in the New York flora (57%), highest in the Florida flora (07%). Some species have both a history of use and were 19th century seed contaminants (e.g.. Agrostemma githago L., Bromus secalinus |... Cen- Convolvulus arvensis L.. Ranun- culus acris L.) (de Schweinitz, 1832; Fernald. 1950: 1959; Haughton, 1978; Mack, 1991). species, which form 1 to 4% of the naturalized an- taurea cyanus |... Grieve, These as both The total percentage of these naturalized floras with a post- giosperm floras, are recorded in Table 1 intentional and accidental introductions. immigration link to cultivation, introduced either deliberately or as contaminants in crop seeds, is between 64 and 69%. under which the remainder (31-37%) arrived their new U.S. Similarity The modes or circumstances range are unknown. n the percentages among the floras surveyed for the central northeast. United States (Fernald, 1950). Rhode Island (Gould et al., 1998). and New York (Mitchell, 1986) is due in part to the similarity of these three species lists. Differences do occur, however, and may reflect the 40- to 50- year differences in the collection spans between Fernald (1950) and the other two accounts. as well as differences in the intensity of collection. Flori- da’s climate and less intense pattern of human set- tlement until the 20th century (Gannon, 1996) has vielded a naturalized flora that is largely not rep- the United States. Most of its naturalized species have sub- resented elsewhere in conterminous tropical or tropical native ranges. (Wunderlin. 1998), and many of these species were imported for potential use as ornamentals and maintained in government test gardens (Gordon & Thomas, 1997). This intensity of plant introduction for ornamental horticulture perhaps explains the high percentage (69%) of its naturalized flora that has a post-intro- duction link to cultivation. DISCUSSION Our results reveal a strong correspondence be- tween naturalized species and these species’ delib- 1). Other species that arrived as seed contaminants in crop erate introduction and cultivation (Table seeds would have been the collateral beneficiaries of cultivation. Together, the proportions in these two categories support the contention that cultivation, whether deliberately or inadvertently supplied for an immigrant species, could have contributed to the persistence of at least 60 percent of the naturalized angiosperm species in the regions we assessed. THE FATE OF DELIBERATELY INTRODUCED SPECIES The major agent for spreading plants into new д oD o anges around the globe for at least the past 400 vears has been human immigration. In embarking on an oceanic voyage of colonization, all peoples. whether Polynesians and Melanesians across the Pacific (Whistler, 1991) (Mack. 1999, 2001). they carry their domesticated plants with them. Europeans worldwide have carefully ensured that This motivation springs from a deep-seated need to anticipate and resolve the dilemma caused by un- known. uncertain, or at least not assured sources of food, fiber, forage, and other essential plant prod- ucts in a new range. Even if would-be colonists had advance knowledge that indigenous species in their new homeland could sustain them, plans for trans- oceanic. colonization have carefully included the transport of the germplasm of essential crops in the initial voyage. Invariably, this transfer of species that were deemed desirable, if not essential, has been maintained long after the colony’s survival 1999, 2001). e rapidity with which the was assured (Mack, early colonists 1 eastern North America established European crops is remarkable: dire necessity is indeed a powerful 1672) cited an account by Higginson in 1629 in which he raved stimulus. Tuckerman (see Josselyn, about the vigor and diversity of the European crops already available in New England, including beets. carrots, cabbage. asparagus, radishes, and lettuce. From the standpoint of species that were to become members of the naturalized flora, Higginson’s list of introduced herbs is revealing: sorrel, parsley, cher- vil, and marigold for pot-herbs, along with sage. thyme, clary. anise, fennel, coriander, spearmint, and pennyroyal as “sweet herbs.” Cultivated fields and gardens with these Euro- pean species produced some of the earliest natu- ralized species in North America. By 1672 Josselyn reported seeing “Dandelion” (Taraxacum officinale (Artemisia absinthium L.). Weber), “Wormwood” 180 Annals of the Missouri Botanical Garden and “Black henbane” (Hyoscyamus niger L.) grow- ing outside cultivation in New England. A century later, Kalm (1770) found other species, including Tanacetum vulgare L. and “Datura” (Datura stra- monium L.), that had also escaped cultivation. The diversity of species used for medicinal pur- poses and as seasonings (Grieve, 1959) likely pro- vided the largest single array of species naturalized in the United States by ca. 1800. Of the 559 spe- cies listed by Fernald (1950) as naturalized, at least 65 were in use before 1900 as herbal remedies or seasonings. Early 19th century seed catalogs often contained sections devoted to “pot-herbs” and me- dicinal species: many of the species in these lists are now naturalized throughout the United States (Mack, 1991). list, it is likely that many of these species had been As suggested in Josselyn's (1672) introduced much earlier. By the early 19th century some were already being listed as naturalized in regional floras, such as Anthemis cotula L., Cynog- lossum officinale L., Inula helenium L., Nepeta ca- taria L., Solanum nigrum L., Solidago odora Ait., (Barton, 1818). Pursh (1814) noted the persistence of Cannabis sa- and Urtica dioica L. In addition, (“Cannabis sativa”), Cichorium intybus L. (“Cichorium intybus”), Conium maculatum L. (“Co- nium maculatum"), Hypericum perforatum L. (“Hy- pericum perforatum"), Linum usitatissimum L. tiva L. (“Linum usitatissimum”), Marrubium vulgare L. (“Marrubium vulgare”), and Ricinus communis L. (“Ricinus communis”). Forage grasses, an early per- ceived deficiency within the native flora, were also actively imported (Cronon, 1983). As a result, west- ern European pasture species were members of the pre-1800 naturalized flora: Aira praecox L., Holcus lanatus L., Lolium perenne L., Phleum pratense L., Poa compressa L., Poa annua L., Poa pratensis L. (Barton, 1818). European fruit trees were planted very early in the settlement of New England (Young, 1846). By 1671, quince, apple, pear, cherry, damson |Prunus domestica var. insititia (L.) Fiori & Paoletti], plum, and common barberry were all commonly grown (Josselyn, 1672). avium (L.) L., Malus pumila Mill.] have become Several of these species [Prunus naturalized but appear innocuous. However, com- mon barberry’s introduction was soon to plague the colonists. Berberis vulgaris L. is a host for Puccinia graminis f. sp. tritici, the stem rust of wheat, and it would be almost three centuries before common barberry was controlled effectively in the United States (Anon., 1937). Ornamentals, i.e., species introduced purely for aesthetic reasons, were introduced surprisingly ear- ly, given the colonists’ need to first establish reli- able sources of food, fiber, forage, and medicine. By 1672 Josselyn was commenting on the imported ornamentals (lavender cotton, hollyhocks, satin, gil- lyflowers, pinks, English roses, and eglantine) that he encountered in New England. Eglantine (Rosa eglanteria L.) is now naturalized in the United States (Fernald, 1950). Given its early introduction, it may be one of the first European ornamental spe- cies to become naturalized in the United States. In introducing ornamentals, the colonists were greatly expanding the taxonomic breadth and geographic range from which naturalized species would be drawn. We lack adequate records of the market in ornamental species that emerged in the 18th cen- tury. Cutler factly that seeds of an Antirrhinum species were However, (1785) reported matter-of- imported b seed-sellers in New England, and a f "Garden and Grass Seeds, with a Roots, & Seeds, Just Imported" was produced as early as 1793 in Rich- broadside « choice — tion of Flower mond, Virginia (National Agricultural Library ar- chives, as cited in Pennsylvania Horticultural So- ciety, 1976). With the apparent proliferation of seed catalogs by 1800, scores of species were arriving in the United States from a worldwide list of native rang- es. Ornamental species (and earliest date of their appearance in а seed catalog published in the Unit- ed States) include Lonicera japonica Thunb. (1823), — camara L. (1804), Melia azedarach L. (1807), Mesembryanthemum crystallinum L. Mimosa pudica L. (1804), Myrica faya Dryand 1823). Rhamnus cathartica L. (1807), Rosa mul- tiflora Thunb. (1826), Schinus terebinthifolius Rad- „ (1807) (Mack, 1804 a Philadelphia seed merchant (1807), — di (1832), and Ulex europaeus 1991). By could describe himself as someone who had for sale, “... South- Sea Islands, African, and European Seeds, of the an extensive variety of Asiatic, most curious and rare kinds: and is daily adding to his collection, as he avails himself of every oppor- tunity to procure seeds from all parts of America, as well as from every part of the world, to whic h the enterprise of American commerce extends . . . (B. M’Mahon [1804: 1] “A catalogue of American Seeds ...”; archives of the National Agricultural Library, Beltsville, Maryland). The speed with which regions only recently colonized by Europe- ans were contributing species to this global traffic in ornamental plants is impressive. D. & C. Lan- dreth, seed merchants in Philadelphia, offered nine Melaleuca species from “New Holland" (Australia) in their 1832 catalog (archives of the Pennsylvania Horticultural Society, Philadelphia). In all likeli- hood, these species had been imported even earlier Volume 89, Number 2 2002 Mack & Erneberg US Naturalized Flora Table 2 Naturalized species in the United States considered by de Schweinitz (1832) to have been deliberately introduced. Some of the names employed by de Schweinitz are not in current nomenclatural usage, and he did not include scientific authorities for the species he lis ted. His names for these species appear parenthetically. The Inter- national Plant Names Index (http://www.ipni.org) was employed as the nomenclatural standard for plant names and authorities / —— a Roth. (Agrostis p "stis tenuis Sibth. (Agrostis vulgaris) Vnthoxanthom мочит L, e 208 odoratum) чое vulgaris) ‚ Jakoch (Sinapis nigra) (€ я saliva) Chelidonium majus L. (Chelidonium majus Cynoglossum officinale L. (Cynoglossum officinale) (Daucus carota) ra stramonium) Holcus lanatus) Daucus carota L.. Datura stramonium L. (Datu Holcus lanatus L. ( Leonurus cardiaca L. (Leonurus cardiaca .( Marrubium vulgare а sativa Phleum рону nse L. (Phleum pratense) (Plantago major) od pratensis (Rosa — umex crispu — А РИ (sic)] Plantago major L. Poa pratensis L. (P^ Rosa eglante ria L. vumex crispus l. ( Rumex pini Ls Salix alba L. (Salix — i ua annuus L. (Scleranthus annuus) S Stellaria media L. (Stellaria media) Taraxacum officinale Weber (Leontodon taraxacum) Trifolium pratense L. (Trifolium pratense) Trifolium repens L. (Trif lium repens Verbascum blattaria lL. (Verbascum blattaria) . (Verbascum thapsus) (Veronica officinalis) Verbascum thapsus L Veronica officinalis L. to Britain before being introduced in the United States. The flurry of published floras that appeared after 1800 provides some of the best evidence we have of which deliberately introduced species were be- (“Salix vi- minalis”) (Pursh, 1814), Crataegus monogyna Jacq. ("C. oxyacantha 1.7) (Barton, 1818), L.. Salix babylonica 1... and Salix alba L. coming naturalized. Salix viminalis L. Acer negundo (Darling- ton, 1820) were all recognized as new members of the eastern U.S. flora. Sometimes the flora’s author even knew the circumstances of a species’ intro- duction. Darlington (1826) wryly attributed the es- tablishment around West Chester, Pennsylvania. of — = Leonurus marrubiastrum L. Humphrey Marshall, a local horticultural enthusiast (Wilbert. 1908). Je Schweinitz (1832) assembled comprehensive the gardening of information about deliberately introduced species that in the United States. His observations (de Sehweinitz. 1832: 148) par- were becoming naturalized based on major categories of introduction are ticularly informative: species “purposely brought hither to be cultivated, for the purposes of agricul- ture, or for some real or fancied value they possess? and others that had “been evidently involuntarily introduced with the imported seeds of agricultural plants...” АП 31 species that he listed as delib- erate introductions (Table 2) remain naturalized in the United States, and a few have become invasive (e.g.. Poa pratensis l.. Rumex crispus L.. Verbascum thapsus 1..). Apparently, most of the current worst invaders in the United States had yet to arrive or were still undetected [e.g.. Bromus tectorum |... Lonicera japonica Thunb.. Polygonum cuspidatum Jal T / Sieb. & Zucc.. Sorghum halepense (1...) Pers.|. SEED CONTAMINANTS: EARLY IMMIGRANTS In an era before herbicides and diligent. seed sieving and inspection, seed lots of crop species varied radically in the extent to which they were contaminated with the seeds of extraneous and un- wanted species. Some of the earliest records of non- indigenous plants in North America include spe- that likely contaminants, e.g., Rumex acetosella L. and Rumex 1983. and references therein). Many of these species have been intimately asso- cies arrived from Europe as seed acetosa l. (Cronon, ciated with crops through the strong selection pro- vided by cultivation and post-harvest storage. Prob- ably all seed-sown crops have their own array of seed mimics (Barrett, 1983); each mimic’s phenol- ogy from germination to seed maturation is under selection by the cultivation cycle for its associated crop. Through this close synchrony between the life eyele of the crop and its mimics, cultivation applied to the crop could simultaneously benefit the mim- . leading to their naturalization (Mack. 2000). We have only a sketchy list of species reputedly introduced as seed contaminants in the early set- tlement of the United States (Table 3). general neglect of seed cleaning, there were. how- Given the ever, many possible immigrants among the ruderals and crop weeds of Europe. The native forage grass- Ss in New England were deemed so unsuitable as dnd that by the 1640s a market in European grass seed had already emerged around Narragan- Annals 182 of the Missouri Botanical Garden Table 3 contaminants among agricultural seeds. Some of the names employed by de Schweinitz are not in current nomenclatural Non-indigenous species in the United States considered by de Schweinitz (1832) to have arrived as seed usage, and he did not include scientific authorities for the species he listed. His names for these species appear parenthetically. The International Plant Names Index (http://www. ipni.org) was employed as the nomenclatural standard for plant names and authorities. Achillea millefolium L. (Achillea millefolium) Agrostemma githago L. ( grostemma githago) ا‎ ; le Capsella и (L.) Medik. (Thlaspi bursa pas storis Cerastium fontanum Baumg. (Cerastium vulgatum) Cerastium glomeratum Thuill. (Cerastium viscosum) Cerastium semidecandrum L. (Cerastium semidecandrum) Chenopodiura album L. (Chenopodium album) Cirsium arevnse (L.) Scop. (Carduus arvensis Cirsium vulgare may Ten. (Cnicus lanceolatus) Commelina sativa* Elytrigia repens var. repens (L.) Desv. (Triticum repens Hypericum perforatum L. (Hy peric um perforatum) Lamium amplexicaule L. (Lamium amplexicaule) Leucanthemum vulgare Lam. (Chrysanthemum secu m) Linaria vulgaris L. Lithospermum « arvense L. (Litho olium pere „ (Plar dies S NONE innua L. (Poa a a Polygonum aviculare L. (Polygonum aviculare) Raphant us — L. (Raphanus raphanistrum) „) Beauv. (Setaria glauca) Scop. (Erysimum officinale) „. (Sonchus oleraceus) гиса dioica) „ (Urtica urens) Veronica — L. (Veronica agrestris) (Veronica arvensis) (Antirrhinum linaria р ermum arvense) — > I = = = ош = У > ® a y = = = — SL. = D^ = М 3 = > = ~ LIÉ = з > = (А د چ‎ B > A ~ ~ = = 2 dm — — — J eronica arvensis L. * Probably a corruption of Camelina sativa L. sett Bay (Cronon, 1983, and references therein), providing ample opportunity for the importation of (1672: 216) list, “Of such Plants as have sprung up since seed contaminants. Furthermore, Josselyn’s the English planted and kept Cattle in New Eng- and,” includes non-indigenous species that are un- likely to have been introduced deliberately as pas- ture often found as seed species but are contaminants: “Shepard's purse” [Capsella bursa- pastoris (L.) Medik.], “Groundsel” L.], “Sow-thistle” [Sonchus sp.], “Cheek-weed” [Stellaria media (L.) Vill.]. Cronon (1983: 143) pro- vided an account from 1652 in which settlers in [Senecio vulgaris the New Haven (Connecticut) colony were already debating without avail as to whether the “ spreading of sorrill [probably Rumex crispus L. or Rumex acetosella L.| in the corne fields .. .” could be stemmed. Kalm (1770) commented on European introductions that he saw along the North American 118) re- ported the informed opinion of John Bartram and eastern seaboard in 1748. Kalm (1770: other American botanists that “Chenopodium album [Chenopodium album L.| ... is not a native of America, but has been bruit over amongst other 770: 119) also claimed that Tanacetum vulgare L., “which grows here and seeds from Europe." Kalm (1 there in the hedges, on the roads, and near houses, It likely ar- rived both through deliberate introduction (Mack, was produced from European seeds.” 1991) and as a seed contaminant. Some early 19th century local and regional Unit- ed States floras also cited species that reputedly (Crantz.) (false flax). arrived as seed contaminants. These statements re- flect informed opinion, rather than documented cas- es, Nevertheless, they are among the very few ac- counts of these species in the United States that are nearly contemporaneous with their arrival. Bar- ton (1818) described the flora in and around Phil- adelphia. He listed both Lithospermum arvense L. and Lithospermum latifolium Michx. as "introduced among grass seeds from Europe, but now natural- Pursh (1814) attempted to assemble a flora of the United States, although most of his own col- ised.” lections and exchanges originated in the eastern states from Virginia northward. He reported that Anthoxanthum odoratum L. ("Anthoxanthum odor- atum"), Festuca pratensis Huds. (“Festuca elatior”), and Centaurea cyanus (“Centaurea cyanus”) were either “... introduced with grass seeds from (Pursh, 1814: 65, 83) or “brought from Europe with the grain” (Pursh, 1814: 576) De Schweinitz (1832: 151) provided the first spe- cific attention to non-indigenous species arriving in Europe” the United States as seed contaminants. Although he did not provide explicit information as to how he determined which species were “introduced for- tuitously with agricultural seeds,” his list is none- theless illuminating (Table 3). These species in- clude many that remain today as crop seed contaminants or are ruderals, or both. All are now naturalized in the United States (Fernald, 1950). Some in his list also arrived through deliberate in- troduction, e.g., Hypericum perforatum L. (Haugh- ton, 1978) and Urtica dioica L. (Uphof, 1968). Volume 89, Number 2 2002 Mack & Erneberg US Naturalized Flora 183 SEED CONTAMINANTS: A LONG-TERM MODE OI IMMIGRATION The opportunity for non-indigenous species to arrive in the United States as seed contaminants grew throughout the 19th century, in part. because the United States remained surprisingly dependent on the routine importation of many crop seeds. Hicks (1895: dinary list of crop and forage species. including. "alfalfa. beet, cauliflower, chicory, 391) maintained that for an extraor- Brussels sprouts. kohl-rabi. ish, salsify. spinach, turnip, the seeds are grown borage, broccoli. cress, endive, rad- abroad, as are also the seeds of many of our grasses, such as crested dog's tail, sheep fescue. meadow foxtail, perennial rye grass, and sweet vernal grass." In addition, “Of the following vegetables about one-half. of the seeds are imported: Carrot, eggplant. leek, onion. parsley. parsnip, and pep- рег.” Large fractions of the seeds needed for do- mestie production of cabbage. celery, chervil. kale. and lettuce were also imported (Hicks. 1895: 391). des (1895) and others (e.g.. Ledoux. 1880: Ball. tation had created enormo'is opportunities for the 1898) realized that such massive seed i impor- inadvertent introduction of unwanted non-indige- nous species. Furthermore, they recognized that a cultivated field, carefully tilled to enhance the erop. was equally advantageous for the emergence of ex- 1880). had been heightened by recent events. traneous species (Ledoux. Their concern An ageres- sive invader, Salsola kali L.. had arrived in the 1880s as a contaminant in flax seed from Russia. By 1894 it had already invaded more than 90.000 km? in the wheat-producing regions in the Dakotas 1894). McCarthy (1888) contended that most of the weed flora in the United States was (Dewey. originally introduced and disseminated in the pack- ages of imported seeds, an unsubstantiated claim but with some element of justification. The problem was not, unfortunately, limited to inadequate seed-cleaning. Foreign seed merchants deliberately adulterated crop seeds with commer- : “charlock” cially worthless species (Sinapis arven- sis L.) mixed with turnip and rutabaga seeds. ~ [^ “black medic" (Medicago lupulina V.) mixed with red clover, valu- able tall fescue and Italian rye (Hicks, 1895: 391). Ledoux (1880) reported that the seeds of ruderal English rye mixed with the more species were routinely gathered in Austria and Ba- use as seed adulterants. In one 24.5 g Nobbe (1871) found 3329 extraneous seeds representing 31 taxa. in- varia for sample of Phleum pratense 1... cluding Rumex acetosella L.. Prunella vulgaris L.. Cirsium arvense (L.) Scop.. Sonchus asper (L.) Hill (“Sonchus asper Villars”), and Spergula arvense L.. In one extreme case, 90 percent of a Canadian seed lot sold in Michigan as clover consisted of extra- neous and non-indigenous seeds. The unwanted seeds averaged 132,000 per kilo in this contami- nated lot (Hicks, 1895: 393) ulation that could readily benefit. from any culti- a large founder pop- vation upon sowing. Some seed merchants in the United States were aware of this imported hazard. The Philadelphia seed firm LV. Faust assured cus- tomers in its 1868 catalog that “We are most par- ticular in the purchase of our grass seeds to procure them from a source where there is no danger of foreign seeds having become mixed with them, as we fully appreciate the great damage which some of these will create if once introduced upon the soil.” The response to foreign and domestic contami- nation was a flurry of state and federal legislation to examine commercial seeds, including seeds that had been directly imported from Europe (McCarthy, 1888: Hicks. 1895). lots from domestic and foreign sources. Chester (1889) examined seed Although it is difficult to distinguish between results for do- mestic and imported seed lots in his data. the litany of non-indigenous species he detected is consistent among all samples: naturalized species typically found in arable land and roadsides were being re- (Table 4). state-appointed seed analysts at the turn of the cen- peatedly introduced The diligence of the tury led undoubtedly to curbing the introduction of United Unfortunately, these regula- unwanted non-indigenous species in the 1941). tory practices were enacted long after many non- States (Brown. indigenous species had repeatedly entered the United States and become naturalized. Despite the ample opportunity for non-indige- nous species to arrive as seed contaminants before 1900. we detected few species for which there is a historic reference to their arrival in that mode: e.g., only 14 species within the central northeastern United States flora. nevertheless been significant. Some species not re- This mode of introduction has ported by any pre-1900 observer likely arrived in this manner and were simply overlooked. The Unit- ed States probably derived many of the weeds of European arable fields simply through the frequen- cy of their importation as seed contaminants. Fur- thermore, many species, such as Amaranthus hy- bridus L., Anthemis cotula V... Capsella bursa- pastoris (L.) Medik.. cataria, Plantago major L., and Rumex acetosella Chenopodium album, Nepeta were continually being introduced and dispersed through the eastern United States by seed mer- chants (Table 4). Although these species had ar- 184 Annals of the Missouri Botanical Garden Table 4. in the late 19th ce Non- sane species detected repeatedly as seed contaminants in domestic and imported crop seeds гу (Chester, 1889). Some of the names employed by Chester are not in current nonmenc ‘latural usage, and he did not include scientific authorities for the species he listed. His names for these species appear parenthetically. The International Plant for plant names and authorities. I Names Index (http://www.ipni.org) was employed as the nomenclatural standard (Agrostemma githago) (Amaranthus hybridus) ithe pr arvensis Agrostemma githago L. Amaranthus hybridus L. Anthemis arvensis Capsella bursa- -pastoris s(L .) N Carduus arvensis (L.) Scop. (Cnicus arvensis n (C the — album) ulgar Hypericum perforatu „ (Hypericum perforatu m) Leucanthemum ln tu (Chrysanthemum oun e) Lithospermum arvense L. (Lithospermum arvens edik. (C 'apsella bursa-pastoris) s) Nepeta cataria L. (Nepeta cataria) Plantago lanceolata L. (Plantago lanceolata) Plantago major L. (Plantago major Polygon › Ranunculus sp. umex acetosella — ad L. (Rumex crispus) Setaria Stellaria media L. Verbascum thapsus L. (Stellaria media) (Verbascum thapsus) м rived before 1800, the potential for an increase in their genetic variation in the United States would have continued long after these species’ initial in- troduction, a function of the different European lo- cales from which later-arriving populations were drawn (Novak & Mack, 2001). SPECIES WITH UNKNOWN MODE ( THE UNITED STATES JF IMMIGRATION TO We were unable to identify a pre-1900 use or other mode of introduction for approximately 30 percent of the species now naturalized in the re- gions we examined. Any assessment of the modes of introduction in naturalized floras is handicapped by the paucity and reliability of historic records. We avoided relying on common names to trace unless a common mode of introduction, species’ name has been used consistently for several hun- = dred years: e.g., henbane (Hyoscyamus niger L. foxglove (Digitalis purpurea L.), shepherd's purse (Capsella bursa-pastoris). This criterion limited our ability to trace introductions before ca. 1780. Nev- ertheless, keen observers such as Josselyn (1672) and Kalm (1770) made invaluable observations. Additional underestimate of deliberate introduc- tions was likely because some species were intro- duced for unrecorded purposes. Their naturalized descendants are, however, a living link to a pre- 1900 agrarian-based economy in the United States that relied on few imported commodities. Solutions to almost all material needs and desires were lit- erally “home grown." Thus, Hypericum perforatum was used for medicinal purposes (Darlington, 1859) and as an object in religious services in 18th cen- 1978). were used to comb tury Pennsylvania (Haughton, The dried heads of Dipsacus fullonum L. wool (Fernald, 1950). If a plant was deemed valu- able, its germplasm was imported, even if the like- lihood of successful cultivation anywhere in the United States was low 1959; Sullé & Maisch, 1880). We may never discover all the pur- poses that our resourceful ancestors had for the (Grieve, range of species they so methodically imported. Species introduced by all accidental (but unde- tected) modes occur in the Unknown Category, in- cluding those that can survive attached or within a vast array of cargo: hay, ballast, packing material, attached to livestock and clothing (Ridley, 1930; Mack, in press). But collectively they appear far less important than deliberate introduction as the mode by which plants have arrived in new ranges in the United States in the last 400 years (Mack, in press). post- 1900 IMMIGRATION AND NATURALIZATION Our emphasis has been on the link between pre- 1900 plant introduction and subsequent cultivation and naturalization. Plant importation has contin- ued, however, and new species continually become naturalized. Rejmanek and Randall (1994) reported that nine species had become naturalized in Cali- fornia between 1968 and 1993; deliberately intro- duced species are prominent in this list (Catalpa bignonioides Walt., Nerium oleander L., Pinus pinea L., Pinus halepensis Mill.). The post-1900 growth of the naturalized flora in the United States has likely been substantial. Henry and Scott (1981: 318) tallied the dates of introduction for the “alien component of the spontaneous Illinois vascular flo- » ra," species that apparently include all naturalized but also adventive species as well. They concluded that the woody and herbaceous component of this non-indigenous flora before 1922 was composed of 440 species; 163 species were added between 1922 and 1955, and another 208 non-indigenous species were added from 1956 to 1978. Many of these post- Volume 89, Number 2 2002 Mack & Erneberg 185 US Naturalized Flora Table 5. naturalized angiosperm floras of S outh Australia, Hawaii. and Northern Europe. Numbers in Tallies for the likely mode of entry (deliberate, accidental, deliberate & accidental, and sê for the each column except the last are percentages of the total naturalized flora surve yed. Sources for the modes of introduction are listed in the footnotes. Deliberate & | Unknown mode Number Deliberate Accidental accidental of introduction of taxa South Australia! STI 24 0 19 901 Hawaii? 51 39 0 1 813 Nordic continental countries” \ 36 14 13 <8 559 North Atlantic Islands” " 16 ol 9 14 10 ! Kloot ( 1981). ? Wester (1992). ' Weidema (2000). VN (2000) from the data of Weidema (2000). 1900 plant immigrants to Ilinois would have been if not all. Hamilton, including most, (Reichard & deliberately introduced, the woody 1997) The proportions among deliberately and acciden- immigrants tally introduced 20th century immigrants that have become naturalized are better documented in Aus- Of the 290 weed, cursions into Australia from 1971 to 1995 that have tralia. i.e.. deleterious plant. in- led to naturalizations, 65 percent of these species were introduced as ornamentals, and an additional 7 percent arrived as intended additions to agricul- 1998). these species have had some degree of post-immi- Groves, — ture gration cultivation. The proportion of species arriv- ing in Australia through deliberate action continues pattern set into motion centuries earlier among European colonies and their trading partners. Detailed examination of the fate of introduced woody ornamental species across much of the 20th 2001), suggests an additional aspect of the importance of century in. Canberra, Australia (Mulvaney, post-introduction cultivation. Mulvaney (2001) con- tended that the probability of a species becoming naturalized is a direct function of the number of 1909 A similar explanation has — recorded plantings of from through the mid 1980s. been proposed to account for the naturalization of the spec les non-indigenous birds in New Zealand: persistence correlates with the intensity of the introduction ef- Veltman et al., 1996). The more separate op- — forts portunities for non-indigenous species to be culti- vated, the greater the probability some of its immigrants will be initially spared the full force of environmental stochasticity in the new range. ! Numbers are means of percentages for Greenland. Iceland. the Thus, more than two-thirds of umbers are means of percentages for Norway, Sweden. Finland. and Denmark as derived from the data of Weidema Aland Islands, and the Faeroe Islands as derived NATURALIZED SPECIES ARISE FROM DELIBERATE INTRODUCTIONS: A RECURRING PHENOMENON WORLDWIDE Our chief goal was to evaluate in an objective manner the hypothesis that much of the U.S. nat- uralized flora has a historic link in its mode of entry to deliberate: introduction. and post-immigration cultivation. At the benchmark for rejecting this hypothesis would be outset, we decided that our the failure to detect that even half of the naturalized species had a history of pre- 1900 use. Our hypoth- esis appears supported by our tallies (Table 1). The plausibility of this link is further supported by evidence gathered among naturalized angio- Wester (1992) examined the modes of introduction within the Hawaiian nat- кф, sperm floras worldwide. uralized flora (Table 5). He concluded that the ma- jor mode of introduction had been deliberate and that a large fraction of these species had been introduced in ornamental horticulture. By coincidence, Kloot (1987) also found that at least 57 percent of the South Australia naturalized flora owed ils arrival to deliberate transport. (Table 5) Esler (1987) determined the modes of introduction (imported for use in either horticulture [including i shelter tree timber and specie r agriculture | crop. pasture, and land —— or accidental) for the 303 angiosperm species now naturalized in urban Auckland, New Zealand. Almost 93 percent of this naturalized flora was introduced deliberately, although the degree of post-immigration cultivation probably varies. the mode of introduction has been as- Denmark. Norway. and Sweden along with Recently, sessed for non-indigenous species in Finland. Iceland. 186 Annals of the Missouri Botanical Garden the Aland Islands, the Faeroe Islands, and Green- land (Weidema, 2000). Here again, many natural- ized species appear to have a history of deliberate introduction (e.g.. /nula helenium L., Lychnis chal- cedonica L., Ornithogalum nutans L., Syringa vul- garis L., Spiraea salicifolia L.), but the role of ac- cidental introduction appears to be much larger than in the United States South Australia. Among Nordic countries in continental Europe. about one-third of the naturalized species are con- sidered to have been deliberately introduced, but more (44%) on average are believed to have been introduced accidentally as seed contaminants of cargo or carried by domesticated animals (Table 5). These values from northern Europe require fur- Among Nordic investigators or ther interpretation. there is apparently neither a consensus on the def- inition of “naturalized” (the values reported may also include adventive species for the floras of some ~ countries) nor on an arrival date before which an immigrant species is deemed native (Weidema, 2000). Potentially more important is the much lon- ger history of agriculture in northern Europe than farming by European colonists in the United States or South Australia. Several millennia of raising crops in northern Europe has given ample oppor- tunity for species to have been introduced, both deliberately and inadvertently (Iversen, 1973). For some of these species, their mode of introduction is unknown [e.g.. Helleborus foetidus L., Potentilla micrantha Ramond ex DC.. Digitalis lutea L., Sil- ene tartarica Pers.| (Weidema, 2000). Furthermore, some fraction of those species now considered na- tive arrived so long ago with human settlement that any identification of erstwhile deliberate use is problematical. The much shorter histories of plant introductions into the “New Europes” in North America, Australia, and South America provide us with a sharper picture of the causes of plant natu- ralization than can be reconstructed from the re- cords of plant dispersal by humans in Europe. Needed now is experimentation that bridges the gap between two growing bodies of information: knowledge of the modes of plant introductions and naturalizations since A.D. 1500 (Kloot, 1987; Rei- chard & Hamilton, 1997; 1998; Mulvaney, 2001) and understanding of the stochastic forces to which immigrant populations are usually vulnera- ble (Menges, 1992, 1998; Mack, 1995, 2000). The design of experiments on the fate of immigrant pop- ulations (Panetta & Randall, 1994) could benefit from clues derived from the history of plant intro- ductions. Experimental variables, including the siz- es of the immigrant populations, the initial entry locales in the new range, and the character and Groves, extent of cultivation, could duplicate what is known about a species’ early history in its new range. Once such investigations become widespread, our under- standing of the causes of plant naturalization and subsequent invasion will likely become much clearer (Williamson, 1999; Williamson & Fitter, 1996). Literature Cited 1882. Sample No. 1,398. North Carolina Agric. . Sta. Annual Rep. p. 147. 1937. A Report on the I — ~ in Bar- — Ersdia ation. During the Ca 1936 in- cluding Summarized — or the Period 1918-19: 30. U.S.D.A. dE Pl. Dis. Ctrl Austin, D. 1978. Exotic plants and their effects in ssi does Environm. Conservation 5: 25-34, Bailey, L. Юю. Cyclopedia of American Horticulture, Comprising (ipee n for Cultivation of Horticultural Plants, Descriptions of the 'ruits, Vegeta- bles, Flowers and Ornamental Plants Sold in the United States and Canada, Together with Geographical and Biographical Sketches. 4th ed., 6 vols. Macmillan, New — Ex Species of ОГК. Ball, С. К. 1898. Seed testing: Its importance, — and some results, with a partial bibliography. lowa State Coll. Agric. Mech. Arts, Contr. from the Botanic al De- partment, Barrett, S. С. н. 1983. Crop mimicry in weeds. Econ. Bot. 37: — 32 Barton, W. P. C. 1818. Compendium Florae Philadelphi- cae: — a Description of the Indigenous and Naturalized Plants found within a Cire 2 Е Ten Miles around Philadelphia, Vols. 1 and 2. M. Carey & Sons, ^ M 1922. Two new western weeds. Science 55: — A brief record of seed testing in the Unit- es Department of Agriculture. Pp. 32-39 in M T Munn (editor), History of the Association of Official Analysts. Assoc iation of Official Seed Analysts, iar. A. de. 1855. — ا‎ е вЫ e. des Faits Principa des Lois Ca ant la ирне ean ane des Plantes ds Г Epoque x uelle. 2 vols. V. Par — F. 1889. s ГЕ De bane Coll. Exp. Sta. Ball 5: 5- zT Masson, Agric. Coon, N 1974. The o of Useful Plants. Rodale Press/Book — Emmaus, Pennsylvania. Crawforc . M. 1986 Sudie 's in Plant Survival: Eco- logic al Са азе "Histories of Plant Adaptation to Adversity. Blackwell Scientific, Oxford. Crawley, M. J. 1989 Chane e and timing in biological in- vasions. Pp. 407-423 ¡ Jrake et al. (editors), id logic ‘al Invasions: A Global Perspective. Wiley, ork. — W. 1983. Changes in the Land: Indians, Colo- nists, and the Ecology of New England. Hill and Wang, New Y Cutler, М. 1785. An account of some of the vegetable Dia шен, мна growing in this part of America. Amer. s Sci. Mem. 1: 396—493. Darlington, C. p. & A. P. Wylie. 1955. Chromosome Atlas of Flowering Plants, 2nd ed. Allen & Unwin, London. ОГК. Асаа. Volume 89, Number 2 Mack & Erneberg 187 US Naturalized Flora Darlington, W. 1826. Florula Cestrica: an Essay towards a Catalogue of the Pheenogamous Plants, Native and — м Growing in the Vicinity of the Borough of »st-Chester, in C * 'sler County, Pennsylvania. 5. Sieg- friec * 1859, American Weeds and Useful Plants, 2nd ed. (revised by G. Thurber Darwent, A. T. bu uf: Оор panic wate Weeds 14: : 31: 31 Dewey, L. H. 189 e Russian thistle: [ts (m as a weed in the U р States, with an ace — of the means — for its eradication. U.S.D.A. Div. Bot. Bull. ES Diggs. G. M.. Lipscomb & R. J. O'Kennon. 1999, Shinners & г. s Illustrated Flora of North Central x xas. Botanical Research Institute of Texas. Fort d | 1985. CRC Handbook of Medicinal Herbs. CRC AUN жа, ‘a Raton. Esler, A. К. 1987. The м: of plants in urban Auckland Zealand. Catalogue of naturalised s. New Zealand J. ES 25: 53 8. 989. Commelina caroliniana E ommelina- eae): À вине сб species in the United States is an old — ‘tion from Asia. Taxon 38: "tan Femald. L. 1950. Gray’s Manual of Botany. Sth ed. sana Press, Portland. Ore & A.C. Kinsey. 1943. Edible Wild Plants of East- Idlewild Press, Cornwall-on-Hud- —55€ ern. North — “a. son, New Ferreira, J. F. s. 1. Simon & J. Janick. 1997. Artemisia annua: Botany, Шейше, pharmacology. Hort. 19: 319-371. Gannon, M. (editor). 1996. The New History of Florida. Univ. Press — Gainesville. Glowka, I e Klemm. 1999, International instru- ments, proc organizations and non-indigenous species introductions: Is a — to the Convention on Biological Diversity necessary? Pp. 389-405 in O T. Sandlund, P. J. Schei & A. Vike n adiu Invasive p and Biodiversity Management. Kluwer, Dor- drec ddon. 5. R. & K. 1997. y de ig п ry, screening and ге ps Pp. 21-37 in D. Simberloff, D. C. Schmitz & T . Brown (editors), Strangers in Paradise. Island Press, C esses, Thomas. Florida's invasion E sy L. L., R. W. Enser. R. E. Champlin & I. H. S 998. Vascular Flora of Rhode Island. Rhode Island Natural History Survey, Kingston, Rhode Island. Grey, А. 1879. The pertinacity — tque of — a Sci. Arts 18: 161— Grieve, M. 1959. A го 1 Herbal, D и C. F. Leyel (editor). ane. 3 Yor! ; . Recent incursions of weeds to pos Tech. S Groves, R. tralia. nC à — л Systems, No. 3. University of Adelaide, Ade laide. Haughton, C. S. 1978. Green Immigrants: Plants that Transformed America. Harcourt, Brace and Jovanovich. New York. Henry, R. . R. Scott. 1981. Time of introduction of the кж шол of the — ous Illinois vas- cular QE — . Midl. Naturalist 106: 318—324. Hibbert, M. The Aussie Plant Finder 1999/2000. Horum po New South Wales. Hicks, G. H. 1895. Pure seed investigations. Pp. 389-408 g7 in Yearbook of the United States Department of Agri- Е D.C. 50. Manual of the — the Unit- | — 2nd a (rev. by A. Chase). U.S.D.A. Misc. 00. Washington. D iC. Hnatiuk, as J. (editor). 1990. Census of Australian кс cular Plants. "i Flora and Fauna series: no. Australian Pt Pub. Service, Canberra. Holdgate. M. W. n experimental introduction of 94. A p to the pie ы Bull. Brit. Antarc. Surv. Bull. 13-1 — J. T. 1959. V وف‎ ar data on weedy thistles in western өп America. Le . Bot. : 17-32. Hume, L.. J. The biology of Canadian weeds. 60. ppm — ulus L. J. Pl. Sci. 63: 959— Isaacson, R. T. 1996. S ca List of Plants and Seeds, 4th ed. Andersen Horticultural Library. University of Min- nesota — Minneapolis. Iversen, J. 1973. The ica of Denmark’s nature since the last е ial. (Translation from Danmarks natur. 1: 345-445, 1967, by M. Robson.) C. A. Reitzels For- lag. a ques n. Joselyn, 2. New-England’s Rarities, Discovered in Birds, Beasts, Fishes, Serpents. and Plants of that Country. [Reprint. with introduction by E. ' 105-238 in Transactions and ( ‘ollections of the American Antiquarian Society, Cambridge, 1860. | Kalm, P. 1770. The America of 17 50; Peter каз travels i ion of 1770. Martinez & K. E бырай, 94. A =r Mie Checklist of the Vas- dila Flaca Bi the United States, Canada. and Green- land, 2 08: a ed. Timber Press, Portland, Oregon. api Р. . The naturalised flora of South Australia ts origin, S d — — forms i signifi ance. J. Adelaide Bot. Gard. 99-111 „amson Scribner, ЕЁ 1809. Weeds he Mas ron Office, Augusta, Maine. pus A. R. 1880. Seed examinations. ric. Exp. Sta. Annual Rep., pp. 22— TE B. 1922. Sonchus uliginosus occurring in the Phil- adelphia area. Torreya 22: 91-98. McCarthy. G. 1888. The purity and vitality of seeds, with recent tests of grass and clover seed sold in North pas ar- olina. Коп т — Exp. Sta. Bull. 59: 4—32 — ous Carolina л McCartney, С. rehids by ba side of the ded Amer. = bae Ed. 53: 18 Mack, R. The aa ы trade: An early tem " Me Econ. Bot. 45: 257-273 — ——-. 1995. Understanding the processes of weed in- vasions: The influence of environmental stoc bass Pp. 65—74 in C. Stirton (editor). Weeds in a Changing World, Symposium British Crop Protection Council, I pira No. 64. Brighton. — ———. 1999, The motivation for poring invasive plant species: A primal urge? Pp. —502 in ите of "n VI ао Rangeland Con- gress, Townsville, Australia. acie Al, 00. Lim à fosters plant naturalization by reducing environmental stochasticity. Biological Inva- sions 2(2): 111—122. — ———. 2001. Motivations and consequences ^e the hu- man a renal of plants. Pp. 23-34 in J. McNeely (editor), The Great Reshuffling: Human — nsions in Invasive Alien Species. International Union for — 188 Annals of the Missouri Botanical Garden Conservation of Nature, Gland, Switzerland, and Cam- bridge, U.K —. In press. Global plant — sal: Pathway modes and circumstances. /n G. Ruiz & J. Carlton (ed itors), Global Pathways of Biotie pcre Island Press, Washington, D.C. — ——— & W. M. Lonsdale. 2001. Humans as global е dispersers: Getting more than we bargained for. Bio- Science 51: 95-102. ——, D. Pan W. M. Lonsdale, H. Evans, M. Clout € F. A. Bazzaz. 2000. Biotic invasions: Caus epide mol Vou consequences and control. Ecol. Appl. 10: 689—710. Mahler, M. 3 1090, Limnophila—A new exotic pest. в 2: 4- Menges, E. S. 1992. Stochastic modeling of extinction in plant populations. Pp. 255-275 in P. L. Fiedler & 5 Jain (editors), Conservation Biology. Chapman and Hall, New York. 1998. Population viability analyses in — че and opportunities. Trends Ecol. Evol. Mite he ^a R. S. (editor). 1986. A Checklist of New York State Plants. Unive a of the State of New York, State Education "i pt., Alb: Mohr, €. 187 —— piant introduced into the Gulf States. Bot. p 3: 42-10. Morton, J. F. 1976. Pestiferous spread of many ornamental and fruit Ае! ies in South Florida. Proc. Fla. State Hort. Soc. 89: : 989. C ree ping Indigo (Indigofera spicata Forsk.) Flo — rd to herbivores in Florida. Econ. Bot. 43: Muensc her T, E P „ 1955. Weeds, 2nd ed. Macmillan, New Yor Mulvaney, M. 2001. The effect of introduction pressure on the naturalization of ornamental woody plants in south- eastern Australia. Pp. 186-193 in R. H. Grov Panetta, J. G. Virtue (editors), Weed Risk ppa * CSIRO, Melbourne. Nobbe, F. 1871. Mittheilungen aus der physiologischen versuchs-station Tharand. IV. Botanische untersuchun- gen kauflicher landwirthsc haftlic her Samereien. Land- wirtschaftlichen Ve poe ‘hs-Stationen bd. 13: 233-235. Novak, S. J. € R. N. Mack. 2001. Tracing plant introduc- tion and spread into naturalized ranges: Genetic evi- депо e from ен tectorum (cheatgrass). BioScience Oakley, R. A. 4 H. L. Ау slover. 1910. — rcial vari- eties of alfalfa. U.S.D.A. Farmers Bull. 1-23. Old. R. R. 1981. iin skeletonweed (С нін d ea L.): Its Biology, Ecology and Agronomic Hist Thesis, ЕИ State University, Pullman Panetta, E D. € R. P. Randall. 1994. An assessment of — ability of Emex — Austral. J. Ecol. 19: 76—82. Pennsylvania Horticultural Society. 1976. From Seed to Flower: Philadelphia, 1681-1876: A Pun ultural Point of View. Pennsylvania Horticultural Society. Philadel- phia. Piper, C. V. 1915. Forage Plants and their Culture. Mac- — New York. Pursh, F. 1814. Flora Americae Septentrionalis. White. C e London. Reic Бай, S. H. & С. W. Hamilton, 1997. Predicting in- vasions of woody xt introduced into North America. ) Conservation Biol. 11: 193-2 Rejmanek, М. & J. M. Randall. 1994. Invasive alien plants in California: 1993 summary and aeter with other areas in North America. Madrofio 41: 161— 177. ). M. Richardson. 1996. What attributes — some * species more invasive? Ecology 77: 165 61. — A. F. ‚ M. Klein. 1993. The Vascular Flora of dud шша] Checklist and Atlas. Amer- ican P — q Society, Philadelphia Ridley, H. The Dispe rsal of Plants Throughout the Wor a | ere Ashford, Kent. Sche ery, R. ы, . The migration of a plant. Nat. Hist. : 41— Se tu D: ‚В. V. Nelson, L. E. Nall & J. D. Schardt. 99]. Exotic aquatic plants in Florida: A historical per- spective and review of the present aquatic plant regu- ation program. In T. p. Center et al. (edit ors), Proceed- ings of the Symposium on Aquatic Pest Plants. Technical report NPS/NRE VER/NRTR-91/06. U.S. De- partment of Pu eres National Park Service, Denver. Schweinitz, L. D. de. ‚ Remarks on the plants of Eu- rope which en — naturalized in a more or жа degree, in the Unit E States. Ann. Lyceum Nat. His New York 3 8-15 Simberloff, D. 1988. The ат of population and ое — to conservation science. Annual col. S 19: 473-5 " E J. G. 1900. diga and Forage Plants. U.S.D.A. Div. Agrostol., (p rev. ed.). Spencer, N. R. 1984. Ve ‘lvetleaf, Abutilon theophrasti cla eae), histon — economic impact in the United States. Econ. Bot. 38: * * Stillé, A. & J.M. М b The National Dispensa- tory. Containing the Nara s y, Chemistry, Phar- macy, Actions and Uses of Medicines, inc luding those Recognized in the — opoeias of the United States, Great Britain, and Germany, with Numerous References to the French E _ x, H. C. Lea, Philadelphia. Strange, Ң. le. A History of Herbal Plants. Arco, New York. Sturtevant, E. L. 1919. Sturtevant’s Edible Plants of the World. /n U. P. Hedrick (editor), New York Department of Agriculture and Markets. Annual Report; 27th, 1918-19, v. 2, pt. Tanaka, T. 1976. 1 Tanaka's Cyclopedia of Edible Plants of the World, S. Nakao и Yugaku-sha; distributed by Ke igaku Pub., То Uphof, J. €. T. 1968. Dictionary of Economic Plants. Wel- don & We се js w Yor Veltman, C. J.. ee & M. J. Crawley. 1996. Correlates of — success in exotic New Zealand birds. Amer. Naturalist 147: 542—557. Viola, H. J. € С. М: irgolis (editors). 1991. Seeds of Change: A Quintcentennial Comme moration. Smithson- ian Institution Press, Washington, D. Weidema, I. (editor). ped — ‘ed Snes ies in the Nor- dic е — 213, Axe Weldon, L. W.. uc ios & D. S. Harrison. 1969. Common Minim weeds. Pp. 1-43 in USDA. Agricul- tural Жү No. E Wester, L. 1992. Origin and distribution of adve BUE alien quas СЕ іп Hawai’ ii. Pp. 99-154 in С. P. Stone nith & J. Т. Tunison (editors), Alien is In- vasions in Native ben ms of Hawaiʻi: Management and Research. University of Hawaii Cooperative Na- tional Park Resources Studies Unit, Honolulu. “= = = T Volume 89, Number 2 2002 Mack & Erneberg 189 US Naturalized Flora Whistler, W. A. 1991. Polynesian plant introductions. Pp. 11-66 in A. С. Cox & S. A. Banack (editors). Islands, Plants, and Polynesians: An Introduction to Polynesian ‘age i Dioscorides, Portland, Oregon. Wilbert, M. I. 1908. Some early botanical and herb gar- r J. — 80: 412-427. Invaders, weeds FE dens re Williamson м. ; and the risk from etically modified organisms. Experientia 49: 219— . 1999, ои ا‎ 22: 5-12, — А. — 199 s le varying success of invad- ok ор) : 166 * 660. — RI 1998, S to the Vascular Plants of "lorida. Univ. Press Fk rida, — le. 1846. Chronic les of the Pilgrim Fathers of wa Colony of Plymouth, from a to 1625: Мө ow First Ce lected from Original Records and С Printed Documents. and имне d with Notes "s. L ittle & Bostor Young, A. Brown, APPENDIX | Pre-1900 use of some of the perae angiosperm — that — the tallies in Table | was indicated by their sale i > 19th e (Archives of the Rational rae cultural | ibrary. Beltsville, Maryland). entury seed — Allen’s annual catalogue of seeds. 1870. New York. New \ze 1 Вані h. 1854. seeds for sale. Boston, Barr. G. H. and Co. 1853. Catalogue of agricultural and \ descriptive catalogue of flower Massachusetts. — ultural implements field and garden seeds and fertilizers. J. clone and Son. 1807. and herbaceous plants, indigenous to the United States p America. Philadelphia, Pe nnsylvania 1870. A catalogue of trees. shrubs. Bliss and Sons. Catalogue al seeds. Spring- | Pe | i eng husetts Charles H. B. Breck. tree, ee and grass seeds, for sale store. Boston, Massachusetts. Robert Carr. 1828 Bartram’s catalogue of trees. shrubs. and herbaceous plants, indigenous to the United States of America. Philadel Ма, Pe nnsylvania. Bernard M'Mahon. 1804. catalogue of American seeds, & c. Philadelphia, воа Bernard M'Mahon. s. d. A catalogue of flowers, tree, shrub and Pennsylvania Pomeroy and Marshall. 1857. 1842. Catalogue of vegetable, herb, al the New Seed garden, herbs. grass seeds. Philadelphia. Sar title), Mobile. Alabama. Warren's Garden and Nurseries. 1845. Annual descriptive catalogue of fruit and ornamental trees. Boston. Mas- sachusetts Me essrs. Winship. 1833. Annual catalog of the ornamental trees and plants, cultivated for sale at the Brighton Nurseries by Messrs. Winship. Boston, Mas- sachusells. Winter * Co. Catalogue of fruit and ornamental — innéan QU y Dude n and nursery (late of the messrs. Princes).| Flushing. New Yor Winter and Co. 18 14-1845. Catalogue at fruit and orna- mental trees. [Linnean Коше (late of the messrs. Princes).] Flushing, New Yor fruit and Garden and nursery THE DODO WENT EXTINCT (AND OTHER ECOLOGICAL MYTHS) Stuart L. Pimm! ABSTRACT The scientific consensus is that human impacts are driving species to extinction hundreds to thousands of times — they deserve, | examine four concerns. aster than ed from the natural background rate. Crities challenge this. Perhaps giving them more credit than First, that the extinction crisis is not real. It is and high rates of extinction are the rule, not the exception, within well-known taxa. The second criticism — the problem as one restricted just to islands. It is not. Island species have special аа н but they ranges than are continental species with the same range species with small a ranges that are threatened by "nn impacts. the clearing of forests from eastern that became extinct followin e far more locally abundant within their There are de numbers of loca lly rare, continental third criticism notes the few species North America in the 19th century, casting doubt upon the re PERRA be — habitat loss and spec ies loss. — of this case history shows that exactly as many spec ies 0 MI such as =a eae Asia and the This leads to the final criticism: that there have been too deforestation is more recent and species do not go extinct — Some doomed species can linger for decades as did the now-extinct species in eastern North Ame ric Key words: were lost as expected, for the region had v ry few species to lose. Extensions to species-rich areas Atlantic. coast. of Brazil conim the expected calibrations with an interesting caveat. Forest losses predict the number of threatened species—tho n the verge of extinction—not the number of extinctions. ew recent extinctions. The re ply is that in these regions the deforestation, extinction, species-area curves. Among scientists there is a broad consensus that species are going extinct in unusual numbers. I will not assemble the evidence for this directly because there are recent reviews (May et al., 1995; Pimm, 2001; 1995). Rather, I wish to tackle the critics who dispute this consensus. Whatever Pimm et al., one thinks of them and those who finance some of them, however one scorns their willingness to ig- nore volumes of inconvenient facts, the critics per- sist. They will likely continue to do so while indi- viduals gain financially environmental destruction. from short-term Over the last decade, | have listened to these critics and, perhaps giving them more credit than they deserve, assembled the science to rebut them directly. The synthesis I pre- sent here is one based largely on my own work on birds. This is not because it is unique—far from it; there is an abundance of evidence to counter these critics. Rather, it is an attempt to lay out cohesive, linked arguments into a recipe that readers can readily apply to other taxa. There are four criticisms. The extinction crisis is not real. Rather, it is (Budiansky, 1993). It is the "facts, not the species" that are endangered (Simon & Wildavsky, 1993: A23), the estimates of extinc- “strident, inconsistent, and data-free" a “doomsday myth” tion rates are (Mann & Plummer, 1995). Since the humid tropical forests of the Amazon, Congo and New Guinea, and elsewhere hold the majority of species, their fate is closely tied to the fate of species. Stott (1999) had this to say about them: “Tropical rain forest’ does not exist and never has existed.” Those who accept that unusual numbers of extinctions have occurred can still proceed to dis- miss their significance. “The dodo went extinct” proclaims the Oxford English Dictionary. Most re- cent like the dodo, have been on is- The implication is that island species are wimpy, naive, and unsophisticated. Perhaps island species had it coming to them and the urbane, so- phisticated species that populate continents may not share their fate. extinctions, lands. 3. Habitat destruction does not cause extinc- tions—look at eastern North America, Budiansky (1993) urged. Projections of future high extinction rates such as those by Wilson (1988) and Raven (1988) combine well-documented rates of tropical forest destruction and a model to predict species loss from habitat loss. How good are those predic- tions? Eastern North America was cleared of its deciduous forests from 1750 to 1900, vet suffered few known extinctions. Critics argue that we sim- ply do not know how to predict the numbers of ! Center for Environmental Research and Conservation, York, New York 100027, U.S.A. StuartPimm@aol.com ANN. Missourt Bor. MC 5556, Columbia University, 1200 Amsterdam Ave., New GARD. 89: 190-198. 2002. Volume 89, Number 2 2002 Pimm Species Extinction species that will be lost as tropical forests disap- pear. 4. In what is mostly a rehashing of earlier myths, Lomborg (2001) seemed to be asking where are the bodies to prove an extinction crisis? Some early efforts did indeed suggest that there should be lots of extinct species by now. For example, “one sev- enth to one fifth of all species” extinet within what would now have been the last two decades (Barney, 1980: 328). There are not nearly enough, though to continue the metaphor, there are the requisite num- ber of seriously wounded ones. So are these species really dying off at the expected rate—or are our concerns about them misplaced? l consider each of these myths in turn. 1. The EXTINCTION CRISIS Is Nor REAL. Has humanity increased extinction rates beyond the background rates expected without our im- pacts? Those who argue that we have not are claim- ing that far too few species have gone extinct in the recent past. Where should we look for the extinct species that would reject this assertion? Pacific islands are the obvious place to start, for they were the planets last habitable areas to be colonized. Polynesians reached them only within the last 1000 to 4000 years. The evidence of human impact is freshest here. (The evidence of human- caused mass extinctions in Australia, Madagascar, and the Americas grows more compelling each 1999). Pacific birds provide unambiguous evidence of massive ex- tinction (Pimm et al., 1994; Steadman, 1995). The bones of many bird species persist into, but not year, however (Flannery, island through, archaeological zones showing human pres- ence. | will consider the Hawaiian. islands in detail. We know 43 bird species only from their bones. Yet bird bones are fragile and easily destroyed. We may never find bones of all the now-extinct species, s how many are missing? The bone record would be complete only if all the recent species—those col- lected or seen in the last two centuries—were also found as bones. The proportion of recent species also found as bones estimates how complete the sample of species found only as bones is. The pro- portion is about a half: across the Hawaiian islands. we estimate there are about 40 species missing from the record (Pimm et al., 1994). Add this num- ber to the 43 known species extinctions and the body count rises to 83. James Cook found the Hawaiian islands in 1778. International trade and colonization followed within a generation. These new people cleared forests and introduced cattle and goats. These destroyed native plants as unprepared for large mammalian herbi- vores as the birds were for the rats and. pigs the Polynesians brought with them on their earlier col- onization. Today, our only records of 18 species of birds are the specimens collected by 19th century naturalists. The body count rises to LOL. What remains in the Hawaiian islands today? Pimm et al. (1994) recorded that a dozen species are so rare that there is little hope of saving them. If we cannot find these species, then they probably cannot find each other. A further dozen we can find, but in numbers so small that their future survival is uncertain. Of an estimated 136 species, only 11 survive in numbers that suggest a confident future. Similar extinctions followed across the Pacific. Over the roughly 1000 years, as the Polynesians colonized the Pacific from New Zealand, north to Hawai'i, and east to Easter Island, they extermi- nated ~1000 species of birds or ~10% of the world total (Pimm et al., 1994: Curnutt & Pimm. 2001). On some islands, they exterminated all the bird species they encountered. They extensively cleared lowland forests, espe- cially the drier ones (and used only Stone-Age technology to do so). Birds were not the only vic- tims of these colonizations, incidentally. Of 980 na- live Hawaiian plants, 84 are extinct and 133 have wild populations of fewer than 100 individuals Sohmer, 1994). These plant extinctions were the consequence of recent human colonizations. Quite how many plant species the Polynesians extermi- nated we may never know. Few species groups leave traces; land snails are one and their losses illustrate the bizarre but ex- tensive devastation that human actions can effect. A predatory snail, Euglandina rosea, introduced to many Pacific islands to control another introduced snail, Achatina fulica, ate to extinction hundreds of laxa of native Achatinella and Partula land snails (Hadfield. 1986: WCMC., 1992). (1 use the term axa to include recognized geographically distinct populations. Taxonomic uncertainties often raise and sink their specific status. For those that are now exlinct we may never resolve the issue.) Nor are Pacific islands unusual in their species losses. As European explorers moved from their coastal waters from the early 1500s, Mauritius, Ro- drigues, and Réunion in the Indian Ocean lost 33 species of birds, including the dodo, 30 species of land snails, and 11 reptiles. St. Helena and Ma- deira in the Atlantic Ocean have lost 36 species of land snails (WEMC, 1992). These examples raises two obvious questions. The first is whether their numbers are unusual or, 192 Annals of the Missouri Botanical Garden alternatively, how many species should we expect to go extinct each year? The background, that is, pre-human rates of extinctions fluctuate consider- ably over time and surely vary from one species group to the next. However, a convenient (and like- ly conservative) background rate of extinction is about (May et al., 1995; Pimm 1995). Only about one in a million species “one in a million” et al., should terminate their existence naturally within a year. The bird extinction rate is closer to one bird species per year from a sample pool of only 10,000 bird species. This means that bird extinctions are running 100 times the expected rate. The second obvious question: Do we find evi- dence of massive extinctions only on islands? THE Dopo WENT ExriNcr (Too Bab!) If island species were the only ones at risk, then we consider their loss to be unfortunate, but relax in the confidence that they were especially vulner- able. This argument fails at two levels. The first failure is that high rates of extinctions occur in places other than islands. Here are three examples: (1) A distinct and unusual flora defines the Cape Floristic Region, which occupies a small area of the southern tip of Africa. It comprises several vege- tational types of which the fynbos is dominant in area and contributes the most species. Of the Re- gion’s 9030 species (Goldblatt & Manning, 2002 this issue), 36 species have become extinct in the last century, and some 618 species are threatened (Cowling, 1992). (I will always use "threatened" a specific, technical sense to mean those species thought likely to become extinct within at most a few decades. Quite how long threatened species are likely to last is a topic I discuss later. (2) In. North America's rivers, Williams et al (1992) described the mussels and clams in the Mis- sissippi and St. Lawrence river basins. Of the 297 North American taxa of the two families Unionidae and Margeritifidae, an estimated 21 have likely gone extinct since the end of the last century. An- other 120 taxa are threatened. Miller et al. (1989) found that of ~950 taxa of freshwater fish in the United States, come extinct in the last 100 years. Northern lakes, Canada, and Mexico, 40 have be- southern streams, wetlands, and desert springs are very different habitats, vet all have lost species. (3) Of the 60 species of recent mammal extinc- tions. worldwide, 19 are from Caribbean islands (WCMC, 1992). extinction rates of islands, and I will not consider them further. Interestingly, 18 more were їп Aus- tralia (WCMC, 1992), representing ~6% of its non- This repeats the pattern of high marine mammal species. The extinctions have been equally divided between the southern arid zone a sparsely inhabited area of mostly spinifex desert and extensive pastoralism—and the wheat belt of the southern tip of Western Australia—where 95% of the natural woodland has been cleared (Short & Smith, 1994). Another 43 Australian mammal spe- cies have been lost from more than half of their former ranges or survive on protected offshore is- lands (Burbidge & McKenzie, 1989). These examples refute the criticism that extinc- tions are restricted to islands. In reviewing these and other examples | am struck by the diversity of taxa and ecosystems they encompass. Across these examples, indeed for all well-known taxa, recent extinction rates are 100 to 1000 times the expected rate (Pimm et al., 1995; Lawton & May, 1995). The second failure of the “it’s just islands” crit- icism is more surprising. Certainly, greater numbers and greater fractions of recent species extinctions have been on islands than on continents. For in- stance, since 1600, 97 of the total 108 bird extinc- 1994). sland biotas are uniquely vulnerable to the human tions have been on islands (Collar et al., introduction of previously absent herbivores, pred- alors, diseases, and other natural enemies (Pimm, 1991). Species on continents are not so ecologically naive. There is another major factor that determines threat. Most threatened species have small geo- 1998) and island species’ ranges are inevitably smaller than conti- graphic ranges (Stattersfield et al., nental ones. For a given range size, how do the island and continental fractions of threatened spe- cies compare? Manne et al. (1999) calculated the ranges of all the passerine birds in the Americas and their as- sociated islands. (They comprise roughly a quarter of all bird species.) To separate the effects of range size, and island versus continental distribution, we calculated the breeding range—henceforth, just “range.” Some of the continental species inhabit montane habitat “islands” isolated by a “sea” of lowland habitats. We ask whether these montane species suffer different levels of threat and so sep- arate them from lowland species. showed that for the 14 lowland, 8 montane, and 27 island species with ranges smaller Manne et al. than 1000 km? the proportions of threatened spe- cies are high, but uncertain because of the small sample sizes. Despite these uncertainties, for these small ranges there is no evidence that island spe- cies are more vulnerable than those on continents. Perhaps one should expect this. Tiny ranges should Volume 89, Number 2 2002 Pimm 193 Species Extinction make species vulnerable to habitat loss, hunting, and other threats wherever the species live. Their most surprising conclusion emerges for range sizes between 1000 km? and 100,000 km’. Much smaller fractions of montane and island spe- cies are threatened than of lowland species. At ranges larger than 100,000 km? the proportions are uniformly small in all three groups. This unexpected result at intermediate ranges has several possible explanations, but we consider that local abundance is the most likely (Manne & Pimm, 2001). We find that island species with a range of (say) 10,000 km? are often locally abun- dant on their island. Montane species with small ranges are also locally numerous within their rang- . These examples of numerous species with small ranges have no match in continental lowlands. There, species with such small ranges are almost always very rare within those ranges (Brown, 1984; Gaston et al., 1997). A reasonable explanation for the abundance of island species is competitive re- lease (MacArthur et al., 1972). petitors, island species are able to attain. higher With fewer com- densities and are thus less likely to be threatened. In sum, corrected for range size, continental spe- cies are more—not less—likely to be threatened. This unexpected vulnerability of continental. spe- cies offsets their putative ecological sophistication and experience of predators and other threats. There seems to be no reason why continental spe- cies will be spared the high rates of extinction hu- manity first vested on insular species. AS AMERICANS HAPPEN 3. TRASH THE RAINFORESTS JUST Dip THEIR FORESTS: NOTHING WILL Extensive reductions in the forests of eastern North Surprisingly, only four bird species went extinct: America occurred during the 19th century. passenger pigeon, Carolina parakeet, ivory-billed woodpecker, and Bachman’s warbler from reasons wholly or mostly from habitat loss. Birds are well- known, so we cannot plead ignorance of their ex- tinctions. Critics use this apparent discrepancy to claim that fears about massive global extinctions (Budi- 1994). Just how many species should have based on habitat losses are “simply wrong” ansky, gone extinct as a consequence of the loss of forests? The answer comes from an extension of one of the most well calibrated ecological relationships 1995). quently provides a good description of the relation- (Rosenzweig, The function, 5 = fre- ship between the size of an area, A, and the number of species, S, that it contains; ¢ and z are constants. For real islands surrounded by sea, z is usually about !4 (Rosenzweig, 1995). “islands” We can then predict the reduction in numbers of species Srinat OES from A Now, suppose we extend this to forest * that remain amid a “sea” of deforestation. „и аз the habitat's area is reduced from —the original extent of forest—to A original пой the area that now remains as forest “islands.” The proportion of species lost = should ۴ (A JA usu). Thus S, equals Sigina (Ano A and the number of extinctions. S. и ongin euina equals Signa — Su. Notice that we need an satiate of the — of z, bil not of c. Does this recipe work or are one or more of its assumptions flawed? In North America, some 48% of the area covered by the eastern forest at the time of European set- tlement (1620) was still wooded at the point of its lowest forest cover (roughly 1872: Pimm & skins. 1995). With A,4:/A,,, = 0.48 and z = 0.25, we predict that — 179€ of the region's 160 forest birds (27 species) should have become extinct. It is this than the four prediction, some six times greater well-documented extinctions, that causes contro- versy. Does this discrepancy cast doubt on the predic- tions of species losses from habitat reduction? It does not. Those who point to the small number of observed extinctions in the eastern forests mean global extinctions—species that are lost every- where. The prediction of 27 extinctions is based on the number of species lost only within the region. Some of these 160 bird species would survive even if all the eastern forests were cleared. Their distri- butions across the boreal forests of Canada or into Central America would afford them a refuge while U.S. forests were cleared. How many species could become globally extinct if all the eastern forests were felled? Which species are found only in these forests, that is, how many species are endemic to them? The answer is only 28. Now 17% of 2 : 4.76 roughly еи of a species higher than the This prediction is number of extinctions observed. | will not push my luck woodpecker, Picoides borealis, is three-quarters of o argue that the endangered. red-cockaded its way to extinction. The observed and predicted numbers are remarkably close. This case history is not the counterexample critics claim it to be. North America lost few species because it had relatively few species to lose What happens in areas of the world that stand lose many species? My colleagues and | have applied this recipe to two such areas. The first is 1997, 1999a). The region comprises four archipelagos: the Phil- insular Southeast Asia (Brooks et al.. 194 Annals of the Missouri Botanical Garden ippines, the Greater Sundas (Java, Sumatra, and Borneo), northern Wallacea (Sulawesi and the Mo- luccas), and the Lesser Sundas. Their forests hold 585 endemic species of bird—roughly 20 times in half the area. About 10% of the original area is cleared per de- that of America’s eastern forest, cade. Most of this deforestation has occurred re- cently and ~60% of the original area is still for- ested. Unlike the previous example, deforestation has not yet caused any confirmed bird extinctions in insular Southeast Asia. Extinctions take time fol- lowing habitat loss, a point to which I must return. hat does the species-area recipe predict about the details of where extinctions will eventually oc- cur? Across the region, some areas still have most of their forests: Borneo had ~67%, for example, when we assembled the forest cover data. (Forests are shrinking rapidly, however.) Other areas have almost none: Cebu, in the Philippines has < 1%. And some areas have more endemic species than others. Using the recipe, we predicted the numbers of threatened bird and mammal species in each of these four archipelagos, island by island. With a few, interesting exceptions, there is a statistically striking correspondence between the numbers of species we predict should go extinct and those that are currently threatened. Borneo, for example, has 38 endemics of which only 3 are considered to be threatened: the recipe predicts 4. Sulawesi and as- sociated islands also have about two-thirds of their forest remaining, but there are 146 endemic spe- cies strewn across these islands. The recipe pre- dicts 14 should be threatened with extinction and 16 actually are. In contrast, in the Philippines, the islands of Mindoro and the western and central Vi- sayas have 19 endemic species; all are threatened, while the recipe only predicts that 10 should be. Where the recipe fails it usually does so by under- estimating the number of threatened species: when little forest remains, other factors—including hunt- ing and invasive species—add to the threats. The second area is the Atlantic coast forest of Brazil (Brooks & Balmford, 1996). It has 214 en- demic bird species. The area has four major sub- divisions and for each there is a close match be- tween the numbers of threatened species and those we predict should become extinct solely on the ba- sis of habitat loss. (The lowland forests are partic- ularly hard hit, with only 296 of the forest remain- 11 endemic species should be threatened; 9 are threatened.) ing; the recipe predicts that 7 of In sum, we have a well-calibrated ecological re- lationship that predicts how many species should become extinct following the loss of habitat across three continents. Given enough time for the species to die, as in North America, the predictions are supported. Worldwide, for every extinct species of bird there are 10 that are threatened. We predict these much larger numbers, too, from the loss of habitat in endemic-rich parts of South America and tropical Asia. But we are still left with the criticism that the species have not yet expired. There is a lingering uncertainty that perhaps our worst fears will not be realized. That leads to the final criti- cism. 4. WHERE ARE THE BODIES? If we are in the midst of an extinction crisis, why are more species not going extinct? The reply is that it takes time for (metaphorically) fatally wounded species to expire. The point is made by the extinctions of birds in eastern North America. he low point of forest cover for these forests was about 1870; the four fatally wounded birds lingered for several decades, perhaps even a century, before finally expiring. This “many decades” matches many other sourc- es of information. It fits with the IUCN definition of “threatened”—a widely held expert opinion that threatened species will likely go extinct within a few decades. And it fits exactly with the few studies that have explicitly examined forest fragments and watched how fast species disappear from them (Brooks et al., 1999b; Pimm & Brooks, 2000) These studies suggest a species survivorship curve with a half-life of roughly 50 years. That is, half the species that will eventually expire do so within the first 50 years, half of what remain expire in the next 50 years, and so on. Given these results, over what time period might the pending massive loss of species from human actions unfold? Pimm and Raven (2000) provided several an- swers. The first comes from considering the large fraction of species living within tropical forests and how fast those forests are shrinking. A second an- swer comes from looking at the hotspots—such places as the Atlantic coast forests of Brazil and Southeast Asia where endemic species are partic- ularly concentrated. About two-thirds of all species occur in the trop- ics, most of them in tropical humid forests (Raven, 1980). Such forests include both evergreen rainfo- rests and more seasonal ones. They originally cov- ered from 14 to 18 million km?, depending on the exact definition, and about half the original area remains (Skole & Tucker, 1993). Much of the forest reduction is recent, and clearing now eliminates about 1 million km? of tropical forest in 5-10 years. Burning and selective logging severely damages Volume 89, Number 2 2002 Pimm Species Extinction several times the area that is wae چ‎ (Nepstead et al., 1999; Cochrane et al., 199€ To convert habitat loss to species loss, one ex- tends the species-area relationship derived for is- lands to predict how many species will not survive ” that remain amid a in habitat fragment “islands as described above. Then “sea” of converted land one updates the numbers each year as the total forested area shrinks. Species that are classified as threatened. will expire in decades to come and they will be joined by other species for which we are only now destroying their habitats. The doomed species do not all die at once, but are spread over time as determined by the species survivorship curve. Combining these results gives an extinction curve that I view as no more than a first sketch that captures a few salient features. Because the species-area curve is non-linear, the clearing to date of half the humid forests should have fatally wounded 15% of their species. This is the case. Some 12% of all plants are threatened (Walter & Gillett, 1998). be an underestimate since many rare species hav e This estimate is likely to vet to be described. Of course, clearing the re- maining half of these forests would eliminate the other 85% of their species. Thus, the numbers of fatally wounded species should accelerate. rapidly to a peak by mid century. They will be joined by ever-larger fractions of species jeopardized by the interaction between the assumed constant rate of forest clearing and the non-linear species-area curve The relative height of the peak depends critically on the fraction of habitat that remains. 5% would protect about 50% of all the forests’ spe- cies. Smaller percentages of remaining forest would lead to very much smaller estimates of surviving species. (About 5% of the world’s land surface is protected at present, but that percentage includes disproportionately large areas of desert and tundra ecosystems. Protecting 5% of tropical forests will require a considerable effort.) that The time delays before extinction mean there will be far fewer species going extinct at pre- sent than are being fatally wounded. The mode predicts that current extinction rates should be modest—on the order of a hundred species per year, per million species. This matches current es- timates (Pimm et al., There are as many bodies as we expect, not far fewer. Extinction num- bers will also peak in mid century, but will be spread out over a century or more thereafter. Modest tinkering with parameters does not alter the “fewer extinctions now, many more later” fea- ture of this curve, but the contribution of Myers el A value of does. They show that roughly 30 to 50% al. (2000 of plant, amphibian, reptile, mammal, and bird spe- — cies occur in 25 hotspots that individually are no more than 2% of the ice-free land surface. These diverse taxa demonstrate that species with small ranges are numerous and they are extraordinarily concentrated. Nature has put her terrestrial species in relatively few baskets. The sample applies to the oceans: fishes and other organisms dependent on coral reefs are similarly concentrated (Mc Allister et al., 1994). Myers el a (2000) showed that human impacts are malevolent, not random. Across the 25 hotspots, an average of 12% of the original primary vegeta- tion remains. This percentage should be compared to the roughly 50% for tropical forests as a whole. Even within the hotspots, Myers et al. found that the areas richest in endemic plant species have proportionately the least remaining vegetation and the smallest areas currently protected (Fig. 1). A second way to sketch the unfolding extinction assumes that conservation actions immediately pro- tect all the remaining habitat areas within the hot- spots. Applying the species-area curve to the in- dividual hotspots predicts that 18% of all their species would eventually go extinct. [Since Myers et al. (2000) showed that hotspots hold 30-50% of all species, see above, this percentage is also con- sistent with the fraction of currently threatened spe- cies.| Yet another sketch assumes that the hotspots higher than global average rate of habitat loss con- tinues for another decade until only the areas cur- rently protected remain. The hotspots would even- tually lose 44% of all their species (Pimm & Raven. 2000). None of these three sketches captures the inad- equacy of some of the protected areas. the so-called "paper parks." Nor do these ideas consider the added threat of global warming that will doubtless limit the effectiveness of sharply delimited, small reserves. Also excluded are the major threats that and weedy species invasive s pose to the remaining species. Often listed as the most important factor in causing threat and extinc- tion. the impacts of invasive species on islands are Plant introductions are a major threat to the Cape Region of South Africa, 1992). The distinction between these three sketches is well known. Continents. are. vulnerable. too. for instance (Cowling. artificial. Many species live in tropical forests that are also hotspots. Yet others live in tropical forests that are not and some live in hotspots that are not tropical forests. Nonetheless. the sketches capture views of the size and time-scale of the processes driving the unfolding extinction. 196 Annals of the Missouri Botanical Garden 50 yo] 2 $ 404 О o > a Е 304 ш о © ga 3 1 = m Ф E ш НЕ 201 > шш. a B £ © s 74 m^" E g : E и S БЕ a = > © as © HB о 59 o m o o bd T Y LI 1 0 5000 10000 15000 20000 25000 Number of endemic plant species Figure 1. The percentage of the original habitat remaining and the percentage of the original habitat protected are сона smaller іп those hotspots that contain the greatest number of endemic plant species. From data in Myers et al. (200 The first process is the rapidly accelerating loss of presently extensive, but rapidly shrinking, trop- ical forests. Protecting substantial and representa- tive areas requires prompt action. This is unlikely happen unless industrialized nations become more deeply involved with funding conservation in developing ones. Without such action, the loss of species from these areas will overtake the loss of species from hotspots within a few decades. The second process is the rate of loss of species from hotspots. Losses here should dominate for the next few decades, since hotspots are already se- verely fragmented. [By definition: Myers et al. (2000) defined hotspots to have unusual numbers of endemic species and to have suffered dispropor- tionate habitat losses.] Only immediate conserva- tion actions, including restoration of damaged hab- itats, can prevent further species loss. And unless there is immediate action to salvage the remaining unprotected areas, the species losses will more than double. As Myers et al. pointed out, the current unprotected areas constitute only a little more than l million km?. High concentrations of small-ranged species make many species vulnerable, but equally they permit a concerted effort to prevent future ex- tinctions. CONCLUSIONS The dodo did not go extinct. Humanity blud- geoned it into oblivion. With it went 10% of the planet’s birds and, in all probability, similar frac- tions of other poorly known species of plants and animals. That we did not identify and name all the species that disappeared is not a credible argument or their continued survival. The Vietnam memorial on the Mall in Washington, D.C., is a poignant list of all the Americans who died in the U.S.A.’s war in that country. A far smaller list of names appears on a memorial in the village in England where I was born to men who died in France between 1914 to 1918. I recognize those names as just a sample and, relative to the village’s small population, read- ily extrapolate to the massive slaughter of men across the entire country. While a complete list of extinct species would be useful, it is not essential to perceive or to estimate the size of the current crisis най the threats of future extinctions from the few that have occurred in North America is likewise the consequence of misinterpretation. Most of the recent known bird extinctions on continents happened in North America following European Volume 89, Number 2 2002 Pimm 197 Species Extinction colonization. Quite what happened in Europe when its forests were cleared centuries earlier we may never know. Consequently, North America is the crucial case history of forest and species loss. It teaches that we lost 4 of 28 of its endemic forest bird species, almost exactly what the species-to- area calibrations predict on the basis of a 50% re- duction of forests. (Three more species were hunted to extinction: the great auk, the Labrador duck, and the Eskimo curlew.) The major tropical forests in the Amazon, the Congo, and New Guinea have al- ready lost half their area, are shrinking by the day. and yet they hold more than 10 times the number of bird species that were found in eastern North America. The hotspots are already depleted even further. The North American case history is most telling when scaled appropriately. Some scientists have overestimated the numbers of species that should be going extinct per year at present. The fault lies solely with the assumption that species would die out immediately. Some do, but most manage to linger. We have yet to realize the 10% loss of species—roughly the fraction of because well-known species that are threatened the destruction of the most species-rich ecosystems has only unfolded in the last half century. Yet this overestimation is simply fixed by changing the text from predictions of “actual extinctions” to predic- tions of species “being on an inexorable path to extinction.” Unless we protect more of the planet’s remaining natural areas, by the end of this century that distinction will seem absurdly trivial. Literature Cited Barney, G. О. x The Global 2000 Report to the Pres- ident, Vol. 2 . Government Printing Office; Wash- — D.C. Brooks, T. A. cape tinctions. Nature 380: S, Pimm & N. E: Collar. 1997. f threatene n Conse rvation Biol. 382-38 — — apos & C. FK: — 1999a. Threat from dota to montane and lowland birds and mammals in insular Southeast Asia. J. Animal Ecol. 6 1996. Atlantic Forest ex- Deforestation edie ts үз insular — in Southéast Asia. 1061-1078 ———— & J. О. Ovugi. 1999b. Time lag between deforestation and bird Pa * in tropical forest frag- its. Conse — Biol. 13: 1140-115 Brown, J. Н. 1984. On the re — Бен dane 'e and — of species veen abun- ‚ Amer. Naturalist 124: 55 апей, 8, 1993. The d myths. U.S. News and World Report, December 13, 1993, 81—83 —— ———. 1994. Extinction or miscalculation? Nature 370: 105, Burbidge, „ McKenzie. 1989. Patterns in the modern ox ci. of — Australia’s vertebrate fauna: C ‘auses and conservation implications. Biol. Conserva- tion 50: 143 pu Сос ev. MA A. Alencar, М. D. Schulze, С. М. Souza. Jr., a nem P. Lefebvre & E. A. ee 1999, a feo ks in the fire — of closed canopy — al forests. Science 284: 1832-183: Collar. N., M. Crosby A. Statte si ld. Watch 2. — Institution Press. D.C. Cowling, R. M. (editor). 1992. The 0: iv. Press, Cape Town. Curnutt, J. & S. L. Pimm. 2001. in Hawaiʻi and the E Pacific before first contact? Pp. 15-30 in J. M. Scott, S. Conant & C. van Hiper IH (editors), Eula P Conservation, and Man- agement of Hawaiian non Studies in — Ese Н Байа extinction. 199 1. Birds to Washington. — Ecology of Fynbos. How many bird species Vanishing Avifauna. Flannery, i E Science 283: 182- Gaston, a T Blackburn & J. abundance-range Lawton. 1997, Interspecific size relations —— An appraisal of mechanisms. J. eis Ecol. 66: 579-601. Goldblatt, P. & J Manning. 2002. Plant diversity о the Cz T = е southern Africa. Ann. Missouri Bot. 813 02. . 1986. Extinction in Hawaiian achatinel- Malas ologia 27: ool. Lawton, J. H. & . Extinction Rates. Ox- ford Univ. Press. Oxfor Lomborg. B. 2001. The Skeptical Environmentalist: Mea- suring the Real State of the World. Cambridge Univ. Press, Mey 7 MacArthur, I . J. M. Diamond & J. Karr. compe nsation in island faur » С. & M. L. Plummer. fred A. Knopf, New York. & S. L. Pimm. 2001. rarity: Which species are threatene d and which will next? — C — 4: 221-230. 1999, Relative risk of extinction of passe rine p on continents and is ;. Nature 399: 258-201. May. R. M., J. H. Lawton & N. E. Stork. ‚ Assessing extine tion rates. Pp. 1-24 in J. H. — R. M. May С Extinc а Rates. КИ Univ. Schueler, C. A apping and GIS — of the global dais hr nsity nas. Ecology 53: 342 1995. Noah’s Eget м. Beyond eight forms of . Brooks € S. L. Pimm. distribution of coral reef fishes on an equal-area grid. Pp. 155-17 Miller (editor). Advances in Map- ping the — of Nature. Chapman & Hall, London. Miller, R. R., J. D. Williams & E. Williams. 1989, Extinctions of North American fishes during the past Fisheries 14: 22-38. yers, N., iae p C. G. Mittermeier, G. A. B. da onsena & J. Kent. 2000. Biodive 2: — for conservation Priorities. Nature 403: 853- epstead, i. A. Verssimo, A. Aler Lima, P. Lafe hië, P. Schlesinger, C. Pete P. Moutin- Mendoza, M. Cochrane € V. Brooks. 1999, Large-scale a Ren us as forests by logging and z ^s. Nature 398: ) The lance of — ? Univ. Chicago century. di TM E. )5—508. . The World ч to Pimm: A Scientist Audits he Кан, McGraw Hill, York Askins. 1995. Forest um s predict bird ex- 198 Annals of the Missouri Botanical Garden tinctions in eastern North America. Proc. Natl. Acad. Sci. U.S.A. 92: 9343-9347. & Т. М. Brooks. 2000. The — Extine tion: How large, how soon, and where? Pp. in P. Raven (editor), Nature and Human Society: me Quest for a Sustainable World. National Academy Press, Washing- ton, & P H. €— 2000. Extinction by numbers. Na- ture 403: 843— ———, М. P. viral & L. J Justice. 1994. Bird extinc- tions in the pop" Pacific. Philos. Trans. Roy. Soc London B344: 27— — ———, J. Russell, n (а Gittleman & T. М Brooks. 1995. The future of biodiversity. Science 269: 347-350. Raven, P. H. (editor). 1980. Research Priorities in Tropical Biology. National Academy of Sciences Press, Washing- ton, ‚ 1988. Our араа — al forests. Pp. 119— 122 in К. О. Wilsor 'r (editors), Biodiver- sity, National Ac a my Press, Washington, D.C, Rosenzweig, М. L. 1995. Species Diversity in Space and Time. Cambridge Univ. Press, Cambridge, U Short, J. & A. Smith. 1994. Mammal — and recovery in car J. Mammalogy 7: Simon, J. & А. Wildavsky 1993. are — = York Times 142 May 13, : A23. Skole, D. & C. J. Tucker. 1993. Tropical — and one b — ntation in the — Satellite data from 1978 to 1988. Science 260: 1905-1910. Sohmer, n 1994. Pp. 43-51 de. .-1. & rial Eco (editors), Biodiversity and Terre stems, Monograph Series Vol. 14. E of Шы e ade- mia Sinica, Taipei. piy A. . Crosby, A. J. Long & D. C. We 8. Endemic Bird Areas of the World: Priority ы, or — Conservation. BirdLife Conservation Se- 7. Cam ric Sie aan: D. . 1995. ride extinctions of Pacific — d „Biodiversity meets zooarchaeology. Sci- e 261: Si. P i ‘Dopieal Rain Forest: A Political Ecology ji еш пш Studies in the Environment 15, Institute of Economic Affairs, London. Wa * K. S. & н. J. Gillett. 1998. IUCN Red List of Threatened deis E. oe World Conservation Jnion, Gland, Switz WCMC. 1992. Global а Status of the Earth’s Living Resources. World Conservation Monitoring Cen- tre, Chapman & Hall, Londor Williams, J. D., M Warren, J ‘ummings, J. L. Harris & R. J. Neves. 19€ servation status of freshwater mussels of the da oy ач and Canada. isheries 18: 6-22 Wilson, E. О. 1988. Pp. 3-18 in E. O. Wilson € E Pet ا‎ F: тт National Academy Press, Washington. D. THE GLOBAL 200: PRIORITY David M. Olson?’ and Eric Dinerstein? ECOREGIONS FOR GLOBAL CONSERVATION! ABSTRACT A global strategy to conserve biodiversity must aim to protect represe ntative examples of all of the world’s ecosystems, as e as those areas that zc exc eptional concentrations of species and endemics. Although lacking the richness of tropical forests, deserts, tropical lakes, and subpolar seas all contain distinct species, communities, and ecological phenomena. We analyzed global — of biodiversity to identify a set of the Earth’s terrestrial, freshwater, and marine ecoregions that harbor exceptional biodiversity and are representative of its ecosystems. We placed each of the Earth’s ecoregions within a system of 30 biomes and biogeographic realms to facilitate a re — — Biodiversity features were compare E — ecoregions to assess their irreplaceability or pere tive These features included species richness, endemic species, пеем higher taxa, unusual ecologic al or a en phe nomena, * = global rarity of habitats. This | process yielded 238 ecoregions—the Global 200—compis] of 142 terrestrial, 53 freshwater, and 43 marine priority ecoregions. Effective conservation in this set of ecoregions uña help conserve ia most outstanding and representative habitats for biodiversity on this planet. Key words: biodiversity, conservation, ecoregions, endemism. global, phenomena, priority-setting, representation. Tropical rain forests rightfully receive much con- ~ yr those with unusual ecological or evolutionary servation attention as they may contain half of the phenomena. We must also target representative ex- world’s species. A comprehensive strategy for con- amples of all of the world’s biomes within each bio- serving global biodiversity, however, must strive to geographic realm where they occur (Fig. 1). Be- save the other 50 percent of species and the dis- cause of distinct biogeographic histories, similar tinctive ecosystems that support them. For example. kinds of ecosystems found on different continents while they may not support the rich communities or in different ocean basins support unique assem- seen in tropical rain forests or coral reefs, tropical — blages of species and higher taxa. For this reason, dry forests, tundra, polar seas, and mangroves all global strategies should strive to conserve examples harbor unique species. communities, adaptations, of every biome in each realm where it occurs for and phenomena. Some of these biomes, such as terrestrial, freshwater, and marine biodiversity (Ol- tropical dry forests and Mediterranean-climate son & Dinerstein, 1998; Udvardy, 1975: Dasmann. shrublands, are more threatened than are tropical 1974). Here we present the Global 200—an al- rain forests and require immediate conservation ac- tempt to identify a set of ecoregions whose conser- tion. To lose examples of these assemblages would vation would achieve this goal of saving a broad represent an enormous loss of global biodiversity. diversity of the Earth's ecosystems (Figs. 2. 3). This Limited funding compels the conservation com- paper expands and updates an earlier analysis by munity to be strategic and earmark the greatest Olson. and Dinerstein (1998). Several additional amount of resources to protect the most outstanding ecoregions have been identified through ongoing re- and representative areas for biodiversity. On a glob- gional analyses (e.g.. Wikramanayake et al.. 2001) al scale. this requires identifying large regions with and the marine Global 200 have been reduced. exceptional levels of species richness or endemism, largely due to combining several adjacent areas ! We thank the regional experts, biologists, — conservationists who contributed their time and knowledge to the conservation analyses that went into the Global 200. J. Leape and С. Hails have provided critical — for this effort. The staff of WWE contributed greatly to A — assessments from which the map is derived. We thank World Wildlife Fund’s Conservation - ience P rogram for their contribution 2 the analysis and preparation of is Global 200, specifically R. Abell, T. Allnutt, C. Carpe F p — o, P. Hu rley ‚ Kassem, Н. Strand, M. Taye, M. Thieme, W. Wettengel, E. Unde — E. W TS . Itoua, C. Loucks, Y. Kura, J. Morrison, and G. Powell. J. Martin- Po anc Ü . Lagler he [pedi in many ways to facilitate the completion of this project. We thank the staff of pad WWF Network, including all of the ont! organizations, various field offices and programs, and associates, for their review and comments on си "dra ? Conservation Science Program, World Wildlife Fund-US, 1250 24th — NW, Washington, dolson@wes.org, eric.dinerstein@ww 3 WCS South Pacific Program, У Box. 3080. Lami, Fiji. = = Ф 7 Es = تع‎ = 3 T DI x E - =! — S a ).C. 20037, U.S.A. ANN. Missouni Bor. GARD. 89: 199-224. 2002. 200 Annals of the Missouri Botanical Garden Oceania — p Tropical — зазора Moist Broadleaf Forests cal Dry B 4 Pis ك‎ d qu Шере vi nega Grasslands, Savannas, and Shr Figure 1. Terre (sensu Dasmann, 1974; Udvardy, — into larger units. In addition, the conservation sta- tus of all ecoregions has been estimatec Ecoregions are regional-scale (continental-scale) units of biodiversity. We define ecoregions as a rel- atively large area of land or water containing a characteristic set of natural communities that share a large majority of their species, ecological dynam- ics, and environmental conditions (Dinerstein et al., 1995; Groves et al., 2000). They function effective- ly as coarse-scale conservation units because they and their extent roughly coincides with the area over which key ecological processes interact most strongly (Orians, 1993). For each of the Earth's 30 terrestrial, freshwater, encompass similar biological communities, and marine biomes (formerly referred to as major habitat types in our previous analysis), we com- pared the biodiversity of each constituent. ecore- gion. Those ecoregions whose levels of biodiversity were considered exceptional (that is, highly dis- tinctive or irreplaceable; 1995; Pressey et al., see Dinerstein et al., 1994) for their biome, or which were considered the best example of a biome within a realm (even if none of the candidates har- bored exceptional biodiversity), were identified as areas of particular importance for achieving global conservation goals, This prioritization yielded 238 ecoregions—the Global 200—comprised of 142 С Temperate Grasslands, Savannas, and Shrublands ШЕШ Flooded Grasslands and Savannas WÎ Montane Grasslands and Shrublands И Tundra IN Mediterranean Forests, Woodlands, and Scrub ШЕШ Deserts = Xeric Shrublands ШШ Mangrov sstrial biomes, terrestrial and freshwater biogeographic realms, and marine biogeographic realms 975 terrestrial, 53 freshwater, and 43 marine ecoregions nested within 30 biomes and 8 terrestrial and fresh- water biogeographic realms and 13 marine biogeo- graphic subdivisions (Table 1) DELINEATION OF ECOREGIONS TERRESTRIAL ECOREGIONS Dasmann (1974) and Udvardy (1975) were the first to conduct a global representation analysis for terrestrial conservation. Dasmann's system of 198 biotic provinces and Udvardy's 193 units are nested within 7 biogeographic realms and 13 terrestrial biomes and 1 freshwater biome. Both these geo- graphic models serve as logical frameworks for analyses of global representation. and Udvar- dy's biotic provinces, however, limits their utility as The relative coarseness of Dasmann's regional conservation planning tools as many dis- tinctive biotas may remain unrecognized. We be- lieved a more finely resolved map of biodiversity patterns was required, one that mapped distinctive biotas within single, continuous biomes. This called for intensive regional analyses of biodiversity pat- terns across five continents by synthesizing existing 201 Olson & Dinerstein The Global 200 Volume 89, Number 2 2002 әче, ur paist] SUOIBAIOI9 ou 0] puodsoiio) ләш up "Jn[e^ X022 QOS 1694019) [8459.19] əy С 21n31 y aAejuoasoadoa pue soinjeoj Nprsaoarporq aurpue]sjno цим SUDIDAIOIA ӘЗІР] suor. spuejqniys pue se^ojbuew EN 'seuue^es "spue|sseJc једіодапѕ pue үедіо К spue|qnaus эпәх pue sueseg Ш qnaos pue 'spue|pooM 's153104 ueaue1123!PaN ни spuejqnius pue spue|sseJc әйезиору аш SeuueAePs рие spue|sseJ9 рәрооу spuejqniys pue 'seuueAes 'spue|sseJ) ajejeduijg| — — 202 Annals of the Missouri Botanical Garden Table 1. and biogeographic realms Global 200 ecoregions organized by biomes . the estimated conservation sta- low ritical and RS dor relatively arked by asterisks (*) rep- tus for each ecoregion is noted as fol ¿ for сг or endangered, V for vulnerable, e or intact. Ecoregions m areas presently under review for elevation to С Siobal 200 status based on their biodiversity features and representation value. TERRESTRIAL REALM TROPICAL AND SUBTROPICAL MOIST BROADLEAF FORESTS Afrotropical Guinean Moist Fore CE 2. Congolian Coastal For 's CE 3. Cameroon Highland — CE 4. Northeastern Congo Basin Moist em V 5. — Congo Basin Moist Fores RS 6. Western Congo Basin Moist Sat V Albertine Rift Montane Forests CE ‚ East African Coastal Forests CE A Eastern Arc Montane Forests CE 10. Madagascar Forests and Shrublands CE 11. Seychelle and Mascarene Moist Forests CE Australasian 12. Sulawesi Moist — CE 13. Moluccas Moist For ү 14. Southern New ме — Forests V New Guinea Montane Forests RS v Solomons- Vanuatu- н н Moist Forests V . Queensland Tropical Fore V tà New Caledonia Moist Ё CE 19. Lord а Norfolk энге Torati CE ме — ). lia Ghats — Forests CE * Sri Lankan Moist Forests CE 22. сое Indochina — ‘al Moist For- ү 23. — China-Hainan Moist Forests CE 24. Taiwan Montane Forests ү 25. Annamite Range Moist Fores 26. Sumatran Islands Lowland and ‘Montane For- ests CE 27. Philippine Moist m CE 28. Palawan Moist F CE S Kayah- Karen/Tenasserim Moist Forest [ ‚ Peninsular Malaysian Lowland and — Forests V 3l. Borneo Lowland and Montane Forests CE 32. Nansei Shoto Archipelago Forests CE " Eastern Deccan Plateau Moist Fore CE 34. Naga-Manupuri-Chin Hills — forests ү 35. = Mountains Moist Forest RS 36. Western Java Montane Forests CE Neotropical 37. Greater Antillean Moist Forests CE * (Lesser Antillean Moist Forests) CE Talamanc an⸗ d Pacific Forests RS E Chocó-Darién Moist Forests RS 40. Мору panne Montane Forests CE 41. Coastal Venezuela Montane Forests ү 42. uel Moist Forests RS 43. Napo Moist Fores V 44. Rio Negro-Juruá Moist Forests CE 45. Guayanan Highland Moist Forests RS Table 1. Continued. 46. Central Andean Yun CE 47. Southwestern — Moist Forests RS 48. Atlantic Forests CE Oceania 49. South Pacific Island Forests CE 50. Hawaii Moist Forests TROPICAL AND SUBTROPICAL DRY BROADLEAF FORESTS Afrotropica 51. Madagascar Dry Forests CE ок . Nusa Tenggara Dry Forests CE . New Caledonia Dry Forests CE i^ — . Indoc hina Dry Forests CE 55. Chhota-Nagpur Dry Forests CE Neot ropical 56. Mexican dry Fore CE di Tumbesian- Andean ‘Valleys Dry Forests CE C hiquitano Dry Forests CE M Atlantic Dry Forests CE Oceania 60. Hawaii Dry Forests CE TROPICAL AND SUBTROPICAL CONIFEROUS FORESTS Nearctic 61. Sierra Madre Oriental and Occidental Pine- Oak Forests Neotropical 62. Greater Antillean Pine Forests CE 63. Mesoamerican Pine-Oak Forests CE TEMPERATE BROADLEAF AND MIXED FORESTS Australasia 64. Eastern Australia Temperate Forests CE 65. Tasmanian Temperate Rain Forests ү 66. New nd Temperate Forests ү m d р Himalayan Broadleaf and Conifer 68. West stern n Himalayan Temperate Forests CE Nearctic 69. Appalac hian and Mixed Mesophytic Forests V Palearcti di South China Temperate Forests V . Russian Far East Temperate Forests V TEMPERATE CONIFEROUS FORESTS Nearctic 72. Pacific Temperate Rainforests CE 13. Klamath-Siskiyou Coniferous Forests CE 74. Sierra Nevada Coniferous Forests CE ТӘ; — Coniferous and Broadleaf For- ( КО. al . Valdivian Temperate Rainforests/Juan Fer- nández Islands CE E — Islands and Desventuradas slands CE Paleare ис 77. European-Mediterranean Montane Mixed Forests Volume 89, Number 2 2002 Olson & Dinerstein 203 The Global 200 Table 1. Continued. Table 1. Continued. 78. Caucasus-Anatolian-Hyrcanian Temperate Orests 79. Altai-Sayan Montane Forest V 80. He ngduan Shan Coniferous Todd RS BOREAL FORESTS/TAIGA Nearctic 81. Muskwa/Slave Lake Boreal Forests RS 82. — Boreal Forests S PEE Ural Mountains Taiga V ^ Eastern Siberian Taiga RS 85. Kamchatka Taiga * Grasslands RS TROPICAL AND SUBTROPICAL GRASSLANDS, SAVANNAS, AND UBLANDS Horn of — Acacia Savannas ү 87. East African Acacia Savannas \ 88. тте al d Eastern Miombo Woodlands V 89. Sudanian Savannas CE Australasia 90. Northern Australia and Trans-Fly Savannas RS a“ Malayan Ter rai-Duar Savannas and Grasslands CE ub us a 92. Llanos Savannas M 93. Cerrado Woodlands and Savannas \ TEMPERATE GRASSLANDS, SAVANNAS, AND SHRUBLANDS Nearctic 94. Northern Prairie CE * Ta * prairies ls pow 95. peo Steppe CE PED tic . Daurian Steppe \ FLOODED GRASSLANDS AND SAVANNAS Afrotropica DE. S — Sahelian Flooded Grasslands and Sa 98. Zambe ian Flooded Savannas \ Indo-Malayan ‹ ann of Kutch Flooded Grasslands CE Veotropica 100. Everglades Flooded Grassland \ 101. Pantanal Flooded Savannas CE MONTANE GRASSLANDS AND SHRUBLANDS ji e al 2. Ethiopian Highlands CE 14 Southern Rift Montane Woodlands CF 104. East African Moorlands RS 105. Drakensberg Montane Shrublands and Wood- СЕ Australas 106. Central Range Subalpine Grasslands RS ne Malay Kinabalu Montane Shrublands RS a ee 108. Rehet Andean Paramo RS 109. Central Andean Dry Puna ۷ Paleartic 110. Tibetan Plateau Sleppe ү 111. pires Asian Montane Steppe and Wood- lar 112. * rn Himalayan Alpine Meadows RS TUNDRA Vearcti 113. AL askan North Slope Coastal Tundra RS 114. Canadian Low Arctic Tundra RS Palearctic 115. Fenno-Scandia Alpine Tundra and Taiga V 116. Taimyr and Siberian Coastal Tundra RS 117. Chukote Coastal Tundra RS MEDITERRANEAN FORESTS, WOODLANDS, AND SCRUB l Afr реа а 118. Fynbos СЕ — —119 ee stern Australia Forests and Scrub CE 120. Southern Australia Mallee and Woodlands C Nearctic 121. California Chaparral and Woodlands CE t Veotropical 122. Chilean Matorral CE Palearctic 123. Mediterranean | Forests, Scrub Woodlands. and DESERTS AND XERIC SHRUBLANDS Afrotropical 124. Namib-Karoo-Kaokoveld Deserts 125. Madagascar Spiny p” СЕ 126. Socotra Island Deser 127. Arabian ШМ Woodlands and Shrub- ls \ lands lustralasia 128. Carnavon Xeric Serub CE 129. Great Sandy-Tanami Deserts RS Vearcti 130. Son ioran-Baja Des RS 131. Chihuahuan-Te өр icán Deserts \ Veotropical 132. Galapagos Islands Serub \ 133. Atacama-Sechura Deserts V Palearctic 134. Central Asian Deserts CE MANGROVES \frotropical Atlanti 135. Gulf of Gaines Mangroves CE ыст E пар African Mangroves СЕ : 37. en ar Mangroves CE Australasia 138. New Guinea Mangroves RS Indo- Malayan Indo-Pacific 139. Sundarbans Mangroves CE 140. Greater Sundas Mangroves CE Neotropical Atlantic 141. Guianan—Amazon Mangroves RS e ta Pacific 142. Panama Bight Mangroves RS 204 Annals of the Missouri Botanical Garden Continued. Table 1. Continued. Table 1. FRESHWATER REALM LARGE RIVERS Afrotropical 143. Congo River and Flooded Forests Indo-Malayan 144 Lm ‘kong River Nea 45. pr Hiver 146. Lower Mississippi River Neotropical 147. Amazon River and Flooded Forests 148. — o River and Flooded Forests Palearct 149. Yangtze River and Lakes LARGE RIVER HEADWATERS — a 150. Congo Basin Piedmont Rivers and Streams ата ic Mississippi Piedmont Rivers and Streams Veni al 2. Upper Amazon Rivers and — СЯ Upper Paraná Rivers and Streams 154. Brazilian Shield {тэн иде Rivers and Streams ARGE RIVER DELTAS ye 155. Mas — Delta e Malay Indus River Delta — ic 157. Volga River Delta 158. Mesopotamian Delta and Marshes 159. Danube River Delta 160. Lena River Delta SMALL RIVERS — 'al 61. Upper Guinea icd and Streams y Madagascar Freshw 163. Gulf of Guinea Rivers and Streams 64. Cape Rivers and Streams nd Ce . New Guinea Rivers and Streams Ti New Caledonia Rivers and Streams 167. Kimberley Rivers and Stream: 168. Southwest Australia Rivers and Streams 169. Eastern Australia Rivers and Streams New Zealand Rivers and Streams) ens Malayan . Xi Jiang Rivers and Streams | Western Ghats Rivers and Strea 172. Southwestern Sri Lanka Rivers and Streams 173. Salween River 174. Sundaland Rivers and Swamps Nearctic 175. Southeastern Rivers and Streams 176. Pacific Coastal Rivers and Streams 177. Gulf of Alaska Coastal Rivers and Streams Neotropical 178. Guianan Freshwater 179. Greater Antillean Freshwater * (Southern Cone Freshwater, especially Val- divian region) * (Atlantic Coast rivers of SE Brazil, Uruguay) Palearctic 180. Balkan Rivers and Streams * — expansion to Mediterranean region in gen- eral including western North Africa 181. Russian Far East Rivers and Wetlands * (Aral Sea Basin, particularly Syr- and Amu- Dar'ya Rivers) ARGE LAK пто uu 82. Rift Valley Lakes Netopia l 83. High Andean Lakes Pia Lake Baikal 85. Lake Biwa Afroropi al 86. — Crater Lakes — 187. Lakes Kutubu and Sentani 188. Central Sulawesi Lakes — Malayan 9. Philippines Freshwater Ts Lake Inle 191. Yunnan Lakes and Streams rei 'al . Mexican Highland Lakes XERIC BASINS Australasia adm — Australian Freshwater Nearc veg С hihuahuan Freshwater — tic 95. Anatolian Freshwater MARINE REALM POLAR Antarc 196. — ‘tic Peninsula and Weddell Sea Arctic 197. Bering Sea 198. Barents-Kara Seas TEMPERATE SHELF ee 19¢ editerranean * Poda Temperate Atlan 200. Northeast Мя Shelf Marine 201. — nks 202. Ches i p. Temperate — Pac iſic 203. Yellow AA meh Sea AND SEAS Volume 89, Number 2 2002 Olson & Dinerstein 205 The Global 200 Table 1. Continued. Southern Ocean 205 200. Patagonian Southwest Atlantic Saam Australian Marine New Zealand Marine TEMPERATE UPWELLING North 208 2] = C Te mperate Indo- Pacific Agulhas Current TROPICAL UPWELLING Central indo- Pacific = stern — Marine Eastern lids: Pac 213. 214 Panama B i . Gulf of California э. Galápagos Marine Eastern Tropical Atlantic 216. Canary Current TROPICAL COR! pn — Pac ific 222. 223. el : Nansei Shoto New Caledonia Barrier Reef Great Barrier Reef Lord Howe-Norfolk Islands Marine Palau Marine Andaman Sea fal astern Indo-Pacific 220. 228. 228. = 233. 234. w = Societies/Marquesas/Tuamotus Marine Hawatian — Nui Marit Fiji Barrier Reel 8 Marine We ste rn Indo-Pacific Maldives. Chagos. Lakshadweep Atolls East Afric ‘ап Marine West Madagascar Marine * (The Mascarene Islands are under consid- eration due to high numbers of endemic reel ish) * (The Maldives are under consideration for extension to include Sri Lanka and southern Indian coast) We stern — Atlantic 23: 236. 7. Southern Caribbean Sea oamerican Ree ost Antillean Marine Northeast Brazil Shelf Marine RS ү ENE classifications from finer scales. Furthermore, de- lineations were conducted in collaboration with hundreds of regional experts and included exten- sive literature reviews. The result is a digital map of 867 terrestrial ecoregions, classified within bi- omes and realms, to be used for prioritv-setting analyses (Olson et al., 2001). This map provides a much more detailed picture of how species assem- blages are distributed across the world. The in- creased resolution is most apparent in the tropics where Dasmann (1974) and Udvardy (1975) iden- tified 115 and 117 provinces respectively, com- pared to 463 terrestrial ecoregions. Dasmann and Udvardy both went on to assess how well existing protected areas represented the Earth's terrestrial biomes and realms. Biotic prov- inces with little or no protection were identified as priorities. The Global 200 analysis frames the goal of prioritization differently: We ask which regions should be a priority for conservation action (e.g.. designating and strengthening protected areas) be- cause of their outstanding biodiversity features or their representation value. We also apply this ques- tion to the terrestrial, freshwater, and marine realms. FRESHWATER ECOREGIONS Separate analyses of freshwater and terrestrial ecoregions were conducted because the distribution of freshwater biodiversity diverges from terrestrial patterns. Freshwater ecoregions were based on sev- eral regional analyses and consultations with re- gional experts.” Currently, the Global 200 analysis effectively targets the majority of freshwater prior- ities. Some targets. however, may change as we near completion of a global map of freshwater ecoregions that is based on a standard level of biogeographic resolution and relevant biomes. ' Victor (1955), Fre x (1971). Zohary (1973), Miva- — 1975), Yim da a Chinese bn. 78 Map Com- Department of Conse ation (1087). Noirfalise (1987), Changchun и of Geography and Chinese Academy of Sciences (1990), Kurnae 3 Bohn (1994), Krever et al. (1991), et al. (1995). Ecological Stratifi- cation Working Group (1995), ed ant et al. (1995). — (1995). Omernik (1995), Thackway € Cresswell (1995 Mongolian Ministry for — hi the Environment et К) (1996). European Topic € > on. Nature Conse "n (2000), Ricketts et al. — К WF/IUC \ (1994, 1997), Bohn & Katenina p Vra Mentes el a I ^ Hocutt & Wiley (1986), Frest & Johannes (1991), WCMC (1992), Maxwell et al. (1995). — 1 al. (1995). Kottelat & Whitten (1996), — et al. (1999). Abell » al. (2000), Thieme et al. (in pr Annals of the 206 Missouri Botanical Garden °1 AQEL Ul рә1ї5 SUOI391000 әш 01 puodsauı02 si9qunu 3y |, *SUOI391099 ООС [FqO[9 эчоеш pue I9Jemysa31] эц, "€ әлпйї 4 uiseq энәх КО uiseq зәли jews йй Jejod В aye] eus Шш ¡e109 jeoido.n ¡e3seo) Ш Jayempeay зәли 26187 бицәмап үрә!йод jeyseo> 8 еўәр әли 96187 HEN Buijjamdn ayesadwe] ¡eyseo) 8 эли 96187 BN eas pue Jays әдезәйшәј NNI aye] 96107 NI SUL зәдемцѕә:ч Да 7 ABR, [ [тэп Th. Volume 89, Number 2 2002 Olson & Dinerstein The Global 200 207 MARINE ECOREGIONS Relative to most terrestrial and freshwater ecore- gions, marine ecological and biogeographic units are more spatially and temporally dynamic (Sher- man et al., 1990) and therefore more challenging to delineate. Marine ecoregions delineated by the Global 200 are derived from a synthesis of global and regional spatial schemas, review of the avail- able literature” and consultations with experts. Kel- leher et al. (1995), Sherman et al. (1990), Lon- ghurst (1998), and Bailey (1998) served as the primary sources for the Global 200. Our base map does not cover deep water ecosystems (i.e.. pelagic, abyssal, or hadal) nor are its biogeographic units as finely resolved as the maps used in the fresh- water or terrestrial analyses. We believe that sev- eral forthcoming and detailed analyses of marine з н. around the world (Callum Roberts. 2001. Allen. 2002) will be useful in testing and improving the ers. comm. Gerald pers. comm. accuracy of our results. As in the land-based anal- yses, the delineation of marine ecoregions is in- tended to highlight general regions within which characteristic. animals, plants, ecological interac- tions, and biophysical processes occur. BIOGEOGRAPHIC RESOLUTION The majority of Global 200 regions are composed of an aggregation of continental-scale ecoregions. This reflects the coarser level of biogeographic res- olution applied on a global scale. For example. whereas 12 terrestrial units were delineated for the island of New Guinea in the regional analysis (Wik- 2001), only 5 Global 200 units The ecoregions that were combined ramanayake et al., are recognized. are adjacent, related by habitat type. and are bio- geographically similar at a global scale. Appoxi- mately a third of the ecoregions used in the regional analyses correspond directly to Global 200 ecore- The boundaries of Global 200 ecoregions do not present gions. ecific location and configuration of exact target areas for detailed planning. Rather, Global 200 ecoregions are primarily intended spotlight regions of exceptional importance for stra- tegic decision-making. 200 SELECTION CRITERIA FOR THE GLOBAL Selection of the Global 200 draws heavily from the results of intensive regional analyses of biodi- ^ Hayden f id SU db IUCN (1988), — (1990). J et al. 32), Ray & Hayden (1 Kel al. 5). & Jer up : Sullivan & Bustamante ( leher et - V! 1990). ек (1998 versity conducted over the last several years (Krev- er et al., 1994; Dinerstein et al., 1995; Olson et al., 1999; Ricketts et al., 1999; Abell et al.. 2000; Wik- ramanayake et al., 2001: in press: Thieme et al Burgess et al.. in press). Within each biome and biogeographic realm, the relative importance of ecoregions was Classified at one of four levels: glob- ally outstanding, regionally outstanding (e.g.. Neo- tropics, Atlantic Ocean), bioregionally outstanding (e.g.. Caribbean), or locally important. The criteria used to prioritize ecoregions for the Global 200 are the same as those used for the regional assess- ments. We chose the set of ecoregions within each biome that were considered to harbor biodiversity that was globally outstanding or regionally out- standing based on the parameters described below. These parameters were weighted and measured in the regional analyses as illustrated in Appendix The weight assigned to the different parameters varied by biome to better address specific patterns of biodiversity and ecological dynamics. SPECIES RICHNESS AND ENDEMISM Richness values were first corrected for area. We then divided the range of values for the set of ecore- gions sharing the same biome and realm into four categories based on natural breaks. Globally out- standing ecoregions were compared with those identified for other realms to ensure consistency. In general, widely recognized global and regional cen- ters of richness and endemism were selected for Global 200 status. The precision of the data varied considerably as illustrated by richness and ende- vascular mism values for ` plants in temperate co- nifer and tundra biomes (Tables 2. 3). HIGHER TAXONOMIC UNIQUENESS The presence of an endemic higher taxon (genus or family) would contribute more to an ecoregion’s biotic distinctiveness than would an endemic spe- cies. Some ecoregions are notable for biotas that contain unique taxa at higher — levels than species (Vane-Wright et al., ; 1991; Gaston & Williams, Williams & Humphries, moist Villiams et al., =ч Forey et al., 1994; 1994). For example, the forests of northeastern Australia. northern New Zealand. and New Caledonia are recognized as having a number of the most primitive lineages of conifers and flowering plants in the world (МАЕ IUCN, 1994-1997) UNIQUE ECOLOGICAL PHENOMENA OR EVOLUTIONARY 200 status because of their extraordinary ecological Some ecoregions were elevated to Global 208 Annals of the Missouri Botanical Garden Table 2. Estimated richness and endemism (expressed as number of species) of native vascular plant species for temperate coniferous forest pane around the world. Data for ecoregions of the United States and Canada are derived from the Biota of North America Program databases developed by Kartesz and Meacham (1999). The estimates for Eurasian ecoregions may be lied: higher than values for the Americas because the former typically en- compass biogeographic areas that are broader in scope (i.e., they include mixed-conifer and broadleaf forest habitats) than ecoregions delineated for the Americas (WWE/IUCN, 1994-1997; Mittermeier et al., 1999). Ecoregion Species richness Endemism Nearctic Southeastern Conifer Forests 3095 >201 Sierra Nevada Forests 2373 51-75 Arizona Mountains Forests 2204 76-110 South Central Rockies Forests 1933 51-75 — ae n Forests 1859 111-151 Piney Woods Forests 1729 4-10 North Central * kies Forests 1695 21-50 Colorado Rockies Forests 1626 76-110 Middle Atlantic Coastal Forests 1488 11-20 Okanogan Forests 1355 1-3 Cascade Mountain Leeward Forests 1328 11-20 North Cascades Forests 1325 4-10 Central and Southern Cascades Forests 1296 21-50 Eastern Cascade Forests 1224 21-50 Northern California Coastal Forests 1212 11-20 Blue Mountain Forests 1134 21-50 Wasatch and Uinta Montane Forests 1109 51-75 Central Pacific Coastal Forests 1109 11-20 Puget Lowlands Forests 1100 1-3 Great Basin Montane Forests 1043 21-50 Fraser Plateau and Basin Complex 1012 0 Florida Sand Pine Scrub 951 21-50 Northern British Columbia Mountain Forests 909 0 Northern Transitional Alpine Forests 876 0 Alberta/British Columbia Foothill Forests 740 1-3 Alberta Mountain Forests 660 1-3 Northern Pacific Coastal Forests 615 1-3 Queen Charlotte Islands 459 1-3 Atlantic Coastal Pine Barrens 632 1-3 Neotropics Valdivian Temperate Rainforests 463 >33 Palearctic Caucasus Mountains ~6300 ~ 1600 Middle Asia Mountains* ~5500 ~ 1500 Pyrenees ~3500 ~200 Balkan—Rhodope Mountains ~3000 ~900 Alps ~3000 ~350 Carpathians ~2000 ~100 Central China Mixed-Conifer Forests ~ 1900 ? Eastern Himalayan Temperate Conifer Forests ~1500 ? * Kopetdag, Tienshan, Pamiro-Alai, Pamir, Dzhungarian Alatau. phenomena. Ecoregions that contain extensive in- ests of the Amazon (varzea forests) (Goulding, tact habitats and large vertebrate assemblages were 1980; Goulding et al., 1996). Such phenomena recognized. Also considered were the long-distance were once widespread but are now rare due to the migrations of larger terrestrial vertebrates such as prevalence of human disturbance around the world. caribou or wildebeest, and the tremendous seasonal This is the only situation where we consider global fish migrations and fish frugivory in the flooded for- — patterns within the context of threats. Otherwise, Volume 89, Number 2 2002 Olson & Dinerstein The Global 200 209 Table 3. comm.). Actual or estimated vascular plant species richness and endemism (expressed as species — of some tundra ecoregions or regions based on data from WWF/IUCN (1994). Ricketts et al. (1999), and J. ' Kartesz (pers. Ecoregion or region Species richness Endemism Nearctic Pacific Coastal Mountain Icefields 792 0 Alaska/St. Elias Range Tundre 747 1-10 Interior Yukon/Alaska Alpine ‘Tundra 017 1-10 Brooks/British Range Tundra 193 1-3 Ogilvie/MacKenzie Alpine Tundra 589 1-10 Arctic Foothills Tundra 580 0 Beringia Lowland Tundra 953 0 relic Coastal Tundra 539 1-3 Beringia Upland Tundra 38 1-3 Low Arctic Tundra 197 0 Aleutian Islands ‘Tundra 388 1-10 Middle Arctic Tundra 371 1—3 Torngat Mountain Tundra 286 0 High Arctic Tundra 245 0 David. Highlands Tundra 216 0 Baffin Coastal Tundra 135 0 Palearctic Chukotsky Peninsula 939 —50 240 5 Taimyr Peninsula the Global 200 emphasizes biodiversity features that were in place prior to major human impacts of natural habitats and species populations. Both ecological and evolutionary phenomena are a critical. but widely overlooked. aspect of biodi- versity conservation. Unusual evolutionary phe- nomena such as the extraordinary adaptive radia- tions seen in Hawaiian plants, birds. and insects, the radiation of Galápagos finches. the radiation of cichlids in Rift Valley lakes of Africa, also elevated While evolu- every some ecoregions to the Global 200. tionary or ecological phenomena occur in ecoregion, we highlight those that are recognized as exceptional in global comparisons. GLOBAL RARITY All ecoregions in globally rare biomes were con- sidered priorities. We elevated ecoregions to Global 200 status if their biome or major habitat type was represented in fewer than eight distinct regions around the world. Examples of rare biomes include the six Mediterranean woodlands, forests. and scrub, all of limited area. Temperate rain forest eco- systems (a major habitat) occur in seven relatively localized areas around the world. Paramos, or wet tropical alpine shrublands, occur in only a few ar- eas of the northern Andes and Central America, a African and in New For this criterion, we counted only natu- few Каз! mountain ranges, Guinea. rally occurring rarity, although human-induced rar- ity is an important condition to assess when devel- oping conservation strate gies, INTACTNESS For ecoregions in the same biome that were as- sessed at a similar level of biological importance, we selected the ecoregions that had relatively more intact habitats and biotas (see conservation. status below). REPRESENTATION Ecoregions were also elevated to Global 200 sta- tus if they were the best example of their biome within a realm in situations where no other ecore- gion had been selected due to its outstanding bio- diversity. In this selection we emphasized those r most en- = ecoregions that harbored the richest « demic biotas. or had the most intact natural eco- systems if biological importance was similar among candidates, The Global 200 focuses on biological values as the critical first step in setting global conservation priorities. There are many other factors that may be used in the prioritization process. We purposefully did not use ecological function, conservation fea- social, economic, cultural sibility (i.e. political, factors), or human utility as discriminators to iden- Annals of the Missouri Botanical Garden tify the Global 200 as these features are either dif- ficult to measure or are highly fluid. The develop- ment and implementation of ecoregion strategies ry however, require careful attention to ecological function and non-biological factors. The recognition of remaining wild animal migra- tions and other contemporary ecological phenome- na is the only criterion where human impacts to the environment are recognized, because areas of ex- tinguished phenomena are ignored. Otherwise, the Global 200 emphasizes biodiversity features that were in place prior to major human impacts on nat- ural habitats and species populations. CONSERVATION STATUS OF THE GLOBAL 200 ECOREGIONS Ecoregions vary greatly not only in their biolog- ical distinctiveness, but also in their conservation status. Conservation status represents an estimate of the ability of an ecoregion to maintain viable species populations, to sustain ecological process- es, and to be responsive to short- and long-term environmental changes. Conservation status assess- ments of the Global 200 ecoregions were based on landscape or aquascape-level criteria, such as total water quality, and estimates of future threat. From a prac- habitat loss, the degree of fragmentation, tical perspective, a measure of conservation status can dictate the urgency, kinds of conservation ac- tivities, and level of effort needed among ecoregions or biomes. Conservation status can also indicate ar- eas with relatively high opportunity for far-reaching conservation measures. e estimated the conservation status of ecore- gions specifically to enable us to make decisions about elevating ecoregions when the similarity of their biodiversity features made discrimination challenging. Conservation status was also used to assess broad trends in threats among different re- gions and biomes. Again, we drew heavily from re- gional conservation assessments to estimate con- For the Global 200, we classified ecoregions into one of three broad categories: erit- servation status.’ ical/endangered, vulnerable, or relatively stable/ relatively intact over the next 40 years. For terres- trial ecoregions, the most prominent contributor to conservation status is habitat loss, followed by the * IUCN (1991, 1992), Krever et al. (1994), BSP et al. (1995), Dinerstein et al. (1995), — е! а, MacKinnon & oe (1996), Bi i bs E nerstein et al. (19‹ p et al. )7). E. ketts el al. (1999), Abell et 2 (2000). бош. et en (2000). Con- servation — (2000). Wikramanayake et (2001). Burgess et al. ( n press) size of remaining habitat blocks, degree of frag- mentation, degree of degradation, and degree of protection (see Appendix 2). Weightings for factors varied by biome for freshwater and marine ecore- gions, THE GLOBAL 200 ECOREGIONS We identified 238 ecoregions whose biodiversity and representation values are outstanding or sig- nificant on a global scale (Table 1). They represent the terrestrial, freshwater, and marine realms, and the 30 biomes nested within these realms. Among 142 (60%) are terrestrial, 53 (22%) are freshwater ecoregions, and 43 (18%) are the three realms, marine. Terrestrial ecoregions outnumber those of the other realms largely because there is more lo- calized endemism in terrestrial than in marine bi- otas. Gaps in biogeographic information for fresh- water and marine biodiversity also account for some of the variation. TERRESTRIAL REALM Tropical and subtropical moist forests Among the 14 terrestrial biomes, the largest number of Global 200 ecoregions falls within the tropical and subtropical moist forests biome (50 ecoregions or 35% of all terrestrial ecoregions) (Ta- ble 1). The high number of ecoregions reflects the biological richness and complexity of tropical moist forests. Although there are more tropical moist for- est ecoregions in the Indo-Malayan Biogeographic realm (17) than in the Neotropics (12), this is partly due to the archipelagic distributions of Asian trop- ical moist forests and their characteristic biotas (Whitmore, 1986, 1990; Whitten et al. 1987a. 1987b, 1996; Wikramanayake et al., 2001). Four of the Asian tropical moist forests are small island systems, and the original extent of all of the Asian ecoregions fits easily within the area covered by western Amazonian moist forests. The most diverse terrestrial ecoregions occur in the Western Arc forests of the Amazon Basin, with close rivals in the Atlantic Forest ecoregion of Bra- zil, the Chocó-Darién ecoregion of northwestern South America, Sumatra, and Peninsular Malaysia and northern Borneo forest ecoregions. The mon- tane forest biotas of the Northern Andes are re- markable for their globally high rates of beta-di- versity and extraordinary local endemism (Terborgh & Winter, 1983; ICBP, 1992; Hamilton et al., 1995; Wege & Long, 1995; WWF/IUCN, 1994—1997). The forests of the Guayanan region and Cuba are known for their pronounced endemism and unusual Volume 89, Number 2 2002 Olson & Dinerstein The Global 200 biogeographic relationships (Whitmore & Prance, 1987: Borhidi. 1991; Dinerstein et al., 1995; Stey- ermark et al.. 1995; Hedges, 1996). The forests of the Greater Antilles also are notable for a number of relict mammals, such as solenodons and hutias. The Congolian coastal forests are likely the most diverse in the Afrotropics, although diversity infor- mation is scarce for several ecoregions in the cen- tral Congo Basin (Oates, 1996; Kingdon, 1997: Burgess et al., in press). The Guinean moist forests support many species not found in the Central Af- rican region (IUCN/UNEP. 1986a: IUCN. 1990: Martin. 1991: IUCN., 1992: Mittermeier et al.. 1999). The Albertine Rift montane forests are ex- tremely rich for some taxa, such as birds. and have a high degree of endemism (Collar & Stuart. 1988: Kingdon, 1989: WWE/IUCN, 1994). The distinc- tiveness of the Eastern Are montane and East Af- rican coastal forests is attributable to their great age and isolation (Hamilton € Bensted-Smith, 1989: Lovett & Wasser, 1993: Hamilton et al., 1995; Bur- gess et al., in press). Madagascar forests and shrub- lands are also highly distinctive on global scales. especially at higher taxonomic levels (Nicoll & Langrand, 1989; Preston-Mafham, 1991: WWE/ IUCN, 1994). Tropical moist forests of New Guinea are highly distinctive (Brooks. 1987: Flannery, 1990. 1994: WWF/IUCN, 1994; Mittermeier et al.. 1996: Wikramanayake et al., 2001). although Aus- tralian moist forests do share many affinities with New Guinea. The long-isolated forests of New Cal- edonia are exceptionally unusual, with so many en- demic and relict higher taxa and species that the island is considered the “Madagascar of the Pacific.” The forests of Sulawesi are noted for their region- ally high degree of endemism in a range of taxa, a phenomenon also seen in the Philippine moist for- Lesser Sundas semi-evergreen for- (IUCN/UNEP, 1986b; IUCN, 1991: ICBP. 2: Stattersfield et al.. 1998; Wikramanayake et al.. 2001). The Western Ghats and southwestern Sri Lankan moist forests are distinctive due to their ests and in the isolation and stability of conditions over millions of years. Tropical moist forests on oceanic islands are often highly distinctive due to high rates of ende- mism, extraordinary radiations of taxa and adaptive radiation, and relictual or unique higher taxa (Dahl. 1986: Mitchell, 1989; Johnson & Stattersfield. 1990; Flannery, 1994: WW F/IUCN, 1994; Wagner & Funk. 1995). Tropical and. subtropical dry forests The most diverse dry forests in the world occur in southern. Mexico and in the Bolivian lowlands (Gentry, 1993; Parker et al.. 1993: Bullock et al., 1996). The dry forests of the Pacific Coast of north- western South America support a wealth of unique species due to their isolation (Parker & Carr. 1992: WWF/IUCN, 1994: Bullock et al.. 1996). tropical forests of Maputaland-Pondoland in south- eastern Africa support many endemies (Cowling € Hilton-Taylor, 1994: WM F/ IUCN, 1994). The dry forests of central India and Indochina are notable tebrate faunas (Corbett & Hill. 1995). Dry forests of Madagascar and New donia are globally distinctive because of their high The sub- are diverse and for their diverse large ver- 1992: Stewart-Cox, Cale- taxa and extreme endemism 1987: Preston-Mafham. 1991: Wikramanavake et al.. 2001). number of relictual (IL CN/UNEP/WW EF. WWE/IL CN, 1994; Tropical and subtropical coniferous forests the world’s richest and most complex subtropical coniferous forests (Perry. 1991: Peterson et al., 1993: Ramamoorthy al.. 1993; WWE/IUCN, 1994). the Antilles contain many endemics am relictual taxa (Borhidi, 1991). Subtropical conifer forests of Indochina are incorporated into the dry Mexico harbors The conifer forests of Greater and moist forests of the region. Temperate broadleaf and mixed forests Temperate broadleaf and mixed forests are rich- est in central China and eastern North America, with other globally distinctive ecoregions occurring in the Caucasus, the Himalayas, southern Europe. and the Russian Far East (Table 2) (Zhao et al., 1990: Martin et al., 1993: Oosterbroek. 1901: WWF/IUCN, 1994; Mackinnon € Hicks. 1996: Ricketts et al.. 1909), Temperate coniferous forests Temperate rain forests only occur in seven re- gions around the world—the Pacific Northwest. the Validivian forests of southwestern South America, rain forests of New Zealand and Tasmania. the in the 1 Northeastern. Atlantic (small. Ireland. Scotland. and Iceland). pan. and those of the eastern Black Sea (Kellogg t al.. 1992: WWE/UCA. 1994). Forest commu- nities dominated by huge trees (e.g.. giant sequoia. Sequoiadendron gigantea (Lindl.) J. Buchholz: red- wood, Sequoia sempervirens (D. Don) Endl.: moun- tain ash. Eucalyptus regnans F. Muell.) are unusual isolated. pockets southwestern Ja- ecological phenomena that are found only in west- ern North America, America, in the Australasian region in such areas as southwestern South and Annals of the Missouri Botanical Garden southeastern Australia and northern New Zealand. The Klamath-Siskiyou ecoregion of western North America harbors diverse and unusual assemblages and displays notable endemism for a number of plant and animal taxa. Chile are notable for their diversity of tree genera, many of which are monotypic and have Gondwan- These long-isolated forests have aland origins. many other unusual taxa and unique communities. Boreal forests and taiga Low species richness and endemism are char- acteristic of circumboreal and cireumpolar ecore- gions (USSR Academy of Sciences, 1988), thus the presence of intact ecological phenomena denoted outstanding ecoregions. Large-scale migrations of caribou, or reindeer (Rangifer tarandus), and intact predator assemblages can still be found in some regions. For example, the Northern Cordillera bo- real forests of Canada have been called the Ser- engeti of the Far North due to their abundance and ‹ -— м diversity of large vertebrates (Ricketts et al., Extensive tracts of boreal forest and taiga still exist in the northern Nearctic and Palearctic, the largest expanses being in central and eastern Russia (Stewart, 1992; Krever et al., 1994). also enjoys relatively unaltered natural disturbance This biome regimes, an increasingly rare situation in other bi- omes. Tropical and subtropical grasslands, savannas, and shrublands In many parts of the tropics large mammal fau- nas have evolved to take advantage of the produc- tive grasses and browse typical of this biome. These large mammal faunas are richest in African savan- nas and grasslands. Presently the most intact as- semblages occur in East African acacia savannas and Zambezian savannas comprised of mosaics of miombo, mopane, and other habitats (McClanahan & Young, 1996). Large-scale migration of tropical savanna herbivores, such as wildebeest (Conno- chaetes taurinus) and zebra (Equus zebra), are con- tinuing to decline through habitat alteration and hunting. Only in East Africa, the central Zambezian region, and in the Sudd region (Uganda kob or Ko- bus kob) do sizable migrations still persist. Many of the extraordinary migrations of the Guinean and Sa- helian savannas have disappeared. Sahelian ecore- gions support a large number of endemic rodent species, while the Somalian bushland and thickets harbor a concentration of endemic mammals, from rodents to antelopes. Both the Cerrado and the Lla- nos are noted for complexity of habitats and the The Valdivian forests of unusually high levels of endemism and beta diver- The tropical savannas of northern Australia and southern New sity in plants for tropical savannas. Guinea support distinctive communities with sev- eral pockets of endemism for a range of taxa (Stat- tersfield et al., 1998). Temperate grasslands, savannas, and shrublands The vast expanses of grass in North America and Eurasia once sustained vast migrations of large ver- lebrates such as buffalo (Bison bison) and saiga (Saiga tatarica). Such extraordinary phenomena now occur only in isolated pockets, such as on the 1994; Hilbig, 1995; The extraordinary floral communities Daurian Steppe (Krever et al., 1996). of the Eurasian steppes and the North American Finch, Great Plains have been largely extirpated through ЗОО different plant species can occur on a few hectares of North conversion to agriculture. Nearly American tallgrass prairie. The Patagonian steppe and grasslands are notable for endemic higher taxa for mammals. Flooded grasslands and savannas Some globally outstanding flooded savannas and grasslands occur in the Everglades, Pantanal, Sa- helian flooded savannas, Zambezian flooded savan- nas (including the Okavango Delta), and the Sudd. The Everglades are the world's largest rain-fed flooded grassland on a limestone substrate. The looded savannas and erre selected are gen- erally the largest complexes in each region. Anoth- er extraordinary inland delta, the Mamberamo Riv- er inland delta, is captured. within the montane forests of the New Guinea ecoregion, Montane grasslands and shrublands The paramos of the northern Andes are the most extensive examples of this biome. Paramo ecosys- te ms occur in only a few other localities in the trop- . The heathlands and moorlands of East Africa (e. Mt. Kilimanjaro, Mt. Kenya, Rwenzori Mts., Ethiopian Highlands), Mt. Kinabalu of Borneo, and the Central Range of New Guinea are all limited in extent, extremely isolated, and support highly en- demic plants and animals. A characteristic feature of many tropical paramos is the presence of large rosette plants from a variety of plant genera, such as Lobelia (Africa), Puya (South America), Cyathea (New Guinea), and Argyroxiphium (Hawaii)—these plant forms can reach elevations of 4500-4600 m above sea level. Drier, yet distinctive, subtropical montane grasslands, savannas, and woodlands in- Volume 89, Number 2 2002 Olson & Dinerstein 213 The Global 200 clude the Ethiopian Highlands, the Zambezian montane grasslands and woodlands, and the mon- tane habitats of southeastern Africa (Werger, 1978: White, 1983; Huntley, 1989, 1994; Timberlake & Müller, 1994; WWF/IUCN, 1994). The montane erasslands of the Tibetan Plateau still support rel- Pan- — atively intact migrations of Tibetan antelope tholops hodgsoni) and kiang, the Tibetan wild ass (Equus hemionus). The puna grasslands of the high Andes support over 30 species of endemic rodents (45 total species). Tundra Tundra ecoregions were selected. primarily. be- cause of extraordinary seasonal concentrations of breeding waterfowl and shorebirds. and caribou (Stewart, 1992: Krever et al., 1994; Ricketts et al.. 1999). Relatively intact. tundra ecoregions were chosen, wherever possible. The Chukotsky tundra ecoregion is unusual with nearly 50 endemic plant Knystautas, 1987; USSR Academy of Sci- 1988: WWE/IUCN, 1994). species — ences, Deserts and xeric shrublands The support the world’s richest desert floras (Cowling & 1994: Maggs et al., 1994: WWF/ while the Chihuahuan Desert and Namib—Karoo deserts of southwestern Africa Hilton-Taylor, IUCN, 1994), central Mexican deserts are a close second and are the richest Neotropical deserts (Cowling et al.. 1989: Hernandez & 1995; Ricketts et al.. 1999). Australian deserts support the richest reptile faunas. The Carnavon Xeric Scrub of west- Barcenas. ern Australia is a regional center of endemism for a range of taxa. Unusual desert communities dom- inated by giant columnar cacti occur in the Sonoran and Baja Deserts of North America (Brown, 1991). while the spiny thickets of southwestern Madagas- car are globally unique in terms of structure and taxa. Some Baja California communities are par- tially convergent in structure with the Madagascar thickets. South America (including the adjacent transition The Atacama Desert ecoregion of western area of the Monte/Puna/Yungas) and the Horn of Africa deserts were recognized as some of the more outstanding regional centers of richness and en- demism. The Central Asian deserts. while not as rich as Afrotropical or Neotropical deserts, are rep- resentative of the region’s deserts with diverse rep- tile and mammal faunas. Mediterranean forests. woodlands, and scrub АП five Mediterranean-climate ecoregions are highly distinctive, collectively harboring 20 percent of the Earth's plant species (Cody, 1980; Kalin Ar- 1995; Picker & Samways, 1995). Phy- togeographers consider the Fynbos as а separate floral kingdom because 68% of the 8600 vascular plant species crowded into its 90.000 km? are en- royo el al., demie and highly distinctive at several taxonomic 1989, 1996; Cowling & Hil- ton-Taylor, 1994). In terms of species densities, this levels (Cowling et al., is equivalent to about 40 percent of the plant spe- cies of the United States and Canada combined, found within an area the size of the state of Indiana (N. Myers, west Australia shrublands have floras that are sig- pers. comm.). The Fynbos and South- nificantly more diverse than the other ecoregions, although any Mediterranean shrubland is still rieh in species and endemics relative to other non-forest ecoregions (Cowling et al, 1996: Oosterbroek, 199 Mangroves The diversity of mangroves in the Indo-West Pa- cific (WP) region is much greater than those of the Atlantic—Caribbean—East Pacific (ACEP) region— the former supporting 17 genera and 40-42 species of true mangroves and the latter having only 4 gen- era and 7 (MaeNae, 1968: Lacerda, 1993; Olson et al... 1996; Spalding et al., 1997; Rickleffs & Latham, 1993). A single site in the ACEP typi- cally contains 3 or 4 true mangrove species, while species 30 species have been recorded from one locality in the IWP region (Ricklefs & 1993 grove forests on the western coast of НЫЕ Latham, lan- support a number of endemic bird species that are endangered. The mangrove swamps and forests of the Indo-Malayan and Australasian realms are the world's most extensive. South and Southeast Asia 12% of the total area of the world’s 1997). are the largest contiguous mangrove forest in the alone contain mangroves (Spalding et al.. The Sundarbans world. The vast floodplains of New Guinea also sup- port extensive mangrove swamps unrivaled. else- where in the world. and terrestrial If all of the marine, freshwater, species that occur in mangroves are considered, these seemingly simple forests can be considered as one of the more diverse ecosystems in the trop- ics. Mangroves are keystone habitats in the sense that they have an inordinately strong influence on species populations and ecosystems well beyond their limited area. In addition to providing habitat and resources to a wide range of species, mangrove forests and swamps also protect inland habitats and shorelines from damage by damping storm waves and tidal action. Mangroves filter silt and pollutants 214 Annals of the ende Botanical Garden from terrestrial runoff that would otherwise damage seagrass beds and coral reefs FRESHWATER REALM Large rivers Faunas adapted to high-flow regimes of large riv- ers are uncommon and best developed in the Yang- tze, Colorado, and lower Congo Hivers. A relatively small area of rapids in the latter region supports 22 endemic species of fish that are rapid specialists (Lowe-McConnell, 1987). raná, and Amazon—Orinoco Rivers harbor the four The Mekong, Congo, Pa- Р, B great large tropical river fish faunas (Mori, 1936; Roberts, 1975; Hocutt & Wiley, 1986; Lowe- McConnell, 1987; Kottelat & Whitten, 1996). The waters of the Lower Yangtze and Mississippi Rivers contain outstanding examples of large-river fishes, amphibians, reptiles, and invertebrates, including relicts and many endemics (Abell et al.. 2000). Large river headwaters Species, assemblages, and processes in head- water areas are distinct from those of thei мг larger mainstems. The Mississippi Piedmont, Guayanan highlands, Upper Amazon, Upper Parana, Brazilian Shield, and Congo Basin Piedmont harbor a tre- mendous array of species, including numerous en- demics adapted to life in these waters. In turn, these river systems ultimately feed a number of the world’s largest and richest rivers (Hocutt & Wiley, 1986; Kottelat & Whitten, 1996; press). The most diverse vertebrate assemblages on Thieme et al., in Earth occur in freshwater communities of the Am- azon and the Orinoco River basins. Over 3000 spe- cies of fish are estimated to occur in the Amazon Basin alone (Goulding, 1980). Large river deltas Delta complexes of several large temperate and polar rivers are identified, including the Mesopo- tamian, Volga, and Lena River deltas. The Niger River delta, the most extensive river delta in Africa, is characterized by high species richness (Wetlands International and The World Bank, 1996; et al., in press). The extensive deltas of the Orinoco Thieme and Amazon Rivers are encompassed in their re- spective large-river ecoregions (see above). Small river basins The Mississippi River embayment, the Mobile River basin, and numerous coastal streams and riv- ers of southeastern North America together support one of the Earth’s richest temperate freshwater bi- otas (Hocutt & Wiley, 1986; Hackney et al., 1992; Abell et al., 2000). The headwater streams and riv- ers of the Yangtze River in central China are also extremely diverse (recognized as a large river bi- ome in this analysis) (Mori, 1936; Nichols, 1943; Taki, 1975). Secondary centers of temperate diver- sity occur in the rivers and streams of southeastern North America, the western coast of North America, and the Russian Far East (Zhadin & Gerd, 1961; ee et al., 1980; Hocutt & Wiley, 1986; Groom- bridge & Jenkins, 1998; Abell et al., 2000). Several freshwater biotas on islands are highly distinctive, including those of Madagascar, New Guinea, the the Greater Antilles, Sri Lanka, New Caledonia (IUCN/UNEP/WWE, 1987; Zakaria-Ismail, 1987, 1994; Allen, 1991; Preston- Mafham, 1991; Oberdorff et al.. 1995). The South- west Australian Rivers and streams ecoregion is a Greater Sundas. and center of endemism, while also harboring a number of primitive higher taxa and several species with highly unusual freshwater life histories (McDowall, 1996; State of the Environment Advisory Council, 1996). Rivers and streams along the Gulf of Guinea harbor some of the richest and most endemic riv- erine freshwater biotas in Africa (Kingdon, 1989; 1992; Lévéque, 1997; al.. in press). The Salween River of Southeast Asia Lévêque et al., Thieme et is recognized for its rich and endemic freshwater fish fauna (WCMC, 1992). The rivers and streams ` New Guinea, including the inland delta of the Mamberamo River of New Guinea, support a large number of unusual and endemic species and higher taxa. Large lakes The Global 200 also identifies the most outstand- ing examples of diverse and endemic freshwater faunas in large lakes found in temperate and trop- ical regions, many displaying extraordinary species flocks and adaptive radiations in fish taxa. Some particularly notable lake biotas include those of the African Rift Lake Baikal, high-altitude lakes of the Andes, and the highland akes of Mexico (Myers, 1960; Roberts, 1975; Ho- cutt & Wiley, 1986; Allen, 1991; Stiassny et al., 1992; WCMC, 1992; Nagelkerke et al., 1995; Kot- telat & Whitten, 1996; Olson et al., 1999; Thieme et al., Lakes and Lake Tana in Ethiopia, Lake Biwa of southern Japan, the in press). Small lakes Similarly, a number of smaller lakes around the world host extraordinary expressions of freshwater Volume 89, Number 2 2002 Olson & Dinerstein 215 The Global 200 Lake Kutubu and Lake Sentani of New Lakes Highland Lakes, the Cameroon Crater Lakes, Lake biodiversity. Guinea, Yunnan and Streams, Mexican Lanao of the Philippines, Lake Inle Myanmar (Burma). and the Central Sulawesi Lakes have been selected for their globally outstanding biodiversity features. Yeric basins Ephemeral streams, rivers, and lakes. and per- manent springs characterize ecoregions in this bi- ome. Low richness and high endemism in fish and invertebrates (e.g.. molluses) is typical of the Chi- Anatolian, and Central Australian fresh- 1980: Balik. The Cuatro Ciénegas huahuan, water ecoregions (Hocutt & Wiley, 1995; Abell et al., 2000). spring and pool complex in the Chihuahuan Desert is globally unique in its high richness, extreme en- demism. and unusual evolutionary adaptations (Contreras-Balderas, 1978; Hocutt & Wiley. 1986). Freshwater habitats in the Anatolian region of Tur- key support many endemic species (Balik. 1995). MARINE REALM The distribution of marine biodiversity varies widely throughout ocean basins (Briggs. 1974: El- der & Pernetta. 1991: Angel, 1992, 1993: Clarke. 1992: Kendall & Aschan. 1993; Kelleher et al.. 1995; Groombridge & Jenkins, 1996; al.. 1997). The abundance and diversity of most taxa tend to be highest near continental and island Ormond et margins that are less than 2000 m deep (Кау. 1991: Johannes & Hatcher, 1986; Gray, 1997). These ar- eas experience nutrient enrichment from upwelling processes and terrestrial runoff (Ray. 1988: Norse. 1995). Areas where significant upwelling occurs are often. extraordinarily productive in tropical. tem- perate, and polar regions. Within biomes, species richness and endemism also vary enormously around the globe. Current biogeographic data suggest that species endemism tends to be less pronounced in marine ecosystems than in terrestrial or freshwater ecore- gions, but several regional centers of endemism are recognized, including the southern coast of Austra- lia, New lands. the northern coast of South America, the Yel- Каз! Hed Sea. the Mediterranean Sea, the Sea of Cortez, the Caledonia. Lord Howe and Norfolk Is- low and China Seas, the Greal Barrier Reef, and tropical Pacific Islands such as Hawaii. the Marquesas, the Tuamotus and Socie- ties. and Easter Island (Robbins. 1991: Lieske & Myers, 1996; Vernon, 1995; Groombridge & Jen- kins, 1996). In general, marine ecoregions associ- ated with isolated islands and enclosed seas tend to display pronounced endemism (Kelleher et al.. 1995; Groombridge € Jenkins, 1996). We categorized the marine realm into 10 biomes. Pelagic (trades and westerlies). abyssal. and hadal biomes, however, were not assessed for the Global 200 marine analysis because of the large scale of these units compared to other Global 200 ecore- gions. the lack of consensus on their classification. and the limited biodiversity information for these 1991: Grassle. 1991: Large biogeographic ecosystems (see Gage « Tyler, Grassle & Maciolek, 1992). units have been identified for pelagic and abyssal biotas (e.g.. Brinton, 1962: Angel. 1993: Longhurst. 1998: Pierrot-Bults, 1997: Vinogradova. 1997). but their scale is several orders of magnitude greater Global 200 units may be biogeographically and dynamically than most ecoregions. These larger appropriate for open ocean environments. The vast size and dynamic nature of these biomes precluded delineating biogeographic subunits at an appropri- ate level of resolution for the Global 200. Pelagic species are noted for widespread distributions. while the few ocean trench surveys that are avail- able suggest many species are endemic to single trenches. The paucity of species data for these eco- systems also reduces our confidence to undertake comparative analyses. Polar The Weddell Sea and Peninsular Antarctica were identified as the most productive and diverse ecore- gions of the Antarctic large marine ecosystem. The Bering. Beaufort, and Chukchi Seas and Barents— Kara Seas ecoregions are arguably the two most diverse and productive Arctic marine ecosystems (USSR Academy of Sciences. 1988: Reeves « 1994). Marine ecosystems near southern Greenland require further evaluation, Leatherwood, Temperate shelf and seas Some of the most productive marine ecosystems occur in the Grand Banks and New Zealand plus The South coastal waters are remarkable for unusually high the Patagonia ecoregions. Australian invertebrates and = levels of endemism in some groups of fish, in addition to the diverse marine mammal assemblage found there. Two of the world’s argest temperate estuaries, the Chesapeake and Atlantic Shelf are elevated to the Global 200 due to their size. Delaware Bays. and the Northeast productivity, and habitat diversity. Some of the most distinctive enclosed temperate seas. the Mediter- 216 Annals of the Missouri Botanical Garden ranean Sea and the Yellow—East China Seas, are recognized in the Global 200. Temperate upwelling Highly productive and diverse coastal upwelling areas occur along the West Coast of North America where the California Current moves southward. Along the southwest coast of Africa the Benguela Current exhibits similar dynamics. Tropical upwelling The Humboldt Current along the West Coast of South America and the Canary Current along the West Coast of Africa bring rich nutrients to the sea surface where they support highly productive ma- rine systems. Important tropical upwelling and cur- rent areas also occur in the Panama Bight ecore- gions. Tropical coral Southeast Asian seas support more than 450 spe- cies of hard (scleractinian) corals, the western In- dian Ocean around 200, and the Caribbean only 50 species (Vernon, 1995). Variation in reef fish and non-coral invertebrate diversity follows a similar biogeographic pattern (McAllister et al.. 1994; Lieske & Myers, 1996). Overall, the coral reef com- munities of the central Indo-Pacific seas are the most diverse in the world, with the Sulu, Sulawesi, Banda, and Coral Sea ecoregions being the most diverse on Earth (Vernon, 1995; Lieske & Myers, 1996). The largest barrier reef in the world is the Great Barrier Reef. Other world-class barrier reefs include the barrier reefs of New Caledonia, the Me- soamerican and Bahamian barrier reefs, and the large barrier reefs of Fiji. The largest coral atoll complexes occur in the Maldive-Lakshadweep ecoregion of the central Indian Ocean and in the Tuamotus of the central Pacific. CONSERVATION STATUS OF ECOREGIONS Among all terrestrial Global 200 ecoregions (142 in total), 75 ecoregions (53%) are considered crit- ical or endangered, 39 ecoregions (27%) vulnera- ble, and 28 ecoregions (20%) relatively stable or intact (Table 1). Terrestrial ecoregion boundaries do not reflect the extensive habitat loss, fragmentation, and degradation that have occurred in many of the terrestrial ecoregions. In ecoregions that have been dramatically altered, characteristic species and communities survive only in the few remaining small blocks of habitat (e.g.. Collar & Stuart, 1988; Dinerstein et al., 1995). Among the terrestrial bi- omes, ecoregions falling within the tropical and subtropical dry broadleaf forests, temperate grass- lands, Mediterranean shrublands, and temperate broadleaf and mixed forests are the most threat- ened. Virtually all biotas on small islands are vul- nerable or critical/endangered due, in large part, to their limited habitat area and extreme sensitivity to anthropogenic disturbance — alien species (Ra- 1988; Wilson, 1988, 1992; WCMC, 1992; Su- jatnika et al., 1995; Brooks et al., 1997; Reaka- Kudla et al., 1997). Island ecoregions are projected to experience a wave of extinctions over the next ven, two decades given the fragility of island ecosys- tems, the sensitivity and endemicity of island spe- cies, and the severe threats native island biotas face worldwide. Mangrove habitats are threatened worldwide from a range of threats including clear- ing and channelization for shrimp ponds, aquacul- ture, and agriculture, the extraction of timber and fuelwood, pollution, and habitat loss due to urban and industrial expansion. Assessment of conservation status for freshwater North America and South America was based on existing regional analyses (Abell et al., 2000; Olson et al., 1999). In Africa and Europe, analyses currently under n press) provided the basis for rankings presented here. In areas where no regional assessment has ecoregions way (Thieme et al., i been undertaken, review of relevant literature fa- cilitated decisions on the levels of threat faced by native biotas. Worldwide, freshwater organisms rep- resent a disproportionate number of endangered species; thus, it is not surprising that so many freshwater ecoregions received a critical rating in the assessment. In particular, seasonally flooded forests, cataracts, and freshwater communities in xeric areas, are endangered worldwide (Goulding et al., 1996; Abell et al., 2000; Olson et al.. 1999). Moreover, most temperate freshwater biotas are threatened by invasion of exotics, pollution, dams, and habitat degradation. Among the 53 freshwater ecoregions 31 (58%) were deemed to be critical or endangered, 10 (19%) were assessed as vulnerable, and only 12 (23%) were assessed as relatively sta- = = The individual status of marine ecoregions was estimated through review of the literature and con- sultations with regional specialists. Twelve marine ecoregions (29%) were considered relatively stable or intact, while another 12 (29%) were considered critical or endangered. In marine biomes, upwelling areas are heavily overfished, enclosed seas are de- graded, and coral reefs and mangroves are severely affected by habitat destruction, degradation, and overfishing around the world (Sherman et al., 1990; Volume 89, Number 2 2002 Olson & Dinerstein 217 The Global 200 Suchanek, 1994; Kelleher et al.. 1995; Bryant et al.. 1995: Olson et al.. 1996; Ormond et al.. 1997). Increasingly rising sea surface temperatures from global warming may endanger all coral reef ecore- gions within several decades. DEGREE OF OVERLAP OF TERRESTRIAL, FRESHWATER, AND MARINE GLOBAL 200 ECOREGIONS The linkages among terrestrial, freshwater, and marine conservation are often overlooked. Among the Global 200, 33 (23%) of the ТАЗ terrestrial ecoregions overlap extensively with freshwater ecoregions (i.e., more than 50% of the original ex- tent of the terrestrial ecoregion is covered by a freshwater unit). Thirty-four (23%) of the terrestrial ecoregions share at least 50% of their coastline with a marine ecoregion. Ten (6%) of the terrestrial ecoregions do both, overlapping extensively with a freshwater ecoregion and sharing at least 50% of their coastline with a marine ecoregion. The terres- trial ecoregions of this third group are the Mada- gascar dry forests, Congolian coastal forests, Great- er Antilles moist forests, Pacific temperate. rain forests of North America, Queensland tropical moist forests, southeastern. Australia. Aucalyptus— Acacia forests, New Caledonia moist forests, New Caledonia dry forests, New Guinea lowland forests, Sulawesi moist forests, Philippine moist forests, Northeast Borneo/Palawan moist forests, and Rus- sian Far East temperate forests. Carefully designed conservation activities in these 13 units could ul- timately affect 39 ecoregions. THE GLOBAL 200 as A CONSERVATION TOOL The Global 200 is based on the best available information and biological insights. As new inter- pretations of biogeography and better information on the distribution of species and phenomena be- come available, we expect to periodically revise the Global 200. The present list and map incorporate a number of changes from an earlier version (Olson & Dinerstein, 1998). For example, the highly un- usual freshwater biota of southwestern Australia is now recognized, and the terrestrial ecoregions of the Amazon Basin have undergone major revisions based on a recent biogeographic analysis by Silva (1998). EXPANDING CONSERVATION GOALS The Global 200 goes beyond the conservation targets of other prominent global priority-setting ef- forts by explicitly incorporating representation guidelines for biomes within realms. Biological phenomena are also important criteria used in ils selection protocol. The Global 200 also emphasizes freshwater and marine. biodiversity. The Hotspots analysis (Mittermeier et al.. 1999; Myers et al., 2000), for example, mostly targets very large and threatened terrestrial regions with concentrations of range-restricted (locally endemic) species. The Hotspots are largely nested within the Global 200 — > 90% congruence) because both analyses em- phasize exceptional levels of endemism for species and higher taxa. The Global 200 can complement hotspot analyses by corroborating the vast majority of their priority areas and, in some cases, by pro- viding a finer resolution of the variation of biodi- versity features within important regions. For ex- ample, the Madagascar Hotspot identified by Myers et al. (2000) corresponds to five separate Global 200 ecoregions and the Indo-Burma Hotspot over- laps with 14 Global 200 terrestrial and freshwater ecoregions. The Global 200 also encompasses dis- tinct freshwater and marine hotspots and warm- spots, as well as ecoregions important for their ex- traordinary ecological or evolutionary phenomena and their representation value. Endemic Bird Areas of the World highlights concentrations of bird spe- cies with restricted ranges (Stattersfield et al., 1998). Like hotspots, the majority of the Endemic Bird Areas are nested within the Global 200. Both Tropical Forest. Wilderness Areas (Mittermeier el al.. 1999) and Frontier Forests (Bryant et al., 1997) map larger landscapes of relatively undisturbed natural forests around the world. Although the Global 200 does not specifically employ forest wil- derness as a discriminator, again there is extensive overlap with these wilderness areas because such areas often harbor rich assemblages of species and endemics, and unusual phenomena such as intact predator-prey systems. OTHER CONSERVATION TARGETS Other conservation targets, such as species of special concern, keystone species, habitats, and phenomena, large-scale ecological phenomena (e.g. bird, butterfly, caribou, cetacean, sea turtle migrations), wilderness areas, ameliorating climate change impacts, reducing toxins, and maintaining ecosystems with low impacts from alien species are the Global 200. Again, effective conservation within priority ecore- also not directly addressed by gions and coordinated efforts among ecoregions will help achieve conservation goals for these targets. 218 Annals of the Missouri Botanical Garden AN AMBITIOUS BLUEPRINT FOR GLOBAL CONSERVATION One tactical concern of the Global 200 is that it is ambitious, and that by focusing on 238 ecore- gions rather than on a handful of conservation units, we run the risk of placing less emphasis on the most diverse and distinct ecoregions. In re- sponse, we maintain that the broad geographic reach of the Global 200 makes almost every nation on Earth a stakeholder in a global conservation strategy. From the global scale to regional and na- tional-level conservation strategies, the Global 200 lends weight to shared priorities and provides a global perspective for lobbying efforts by local con- servation groups. The Global 200 also can help ma- jor development agencies better recognize and mit- igate the effects of projects that result in land use change, or forego development activities in partic- ularly important and sensitive ecoregions. The targets of the Global 200—representation, outstanding ecoregions, and ecological phenome- na are all essential elements of a global conser- vation strategy. The conservation community should not shrink from this ambitious but necessary agen- da. The widespread destruction of the Earth's bio- diversity occurring today must be matched by a re- sponse at least an order of magnitude greater than currently exists. The Global 200 provides a nec- essarily ambitious template for a global conserva- tion strategy. Literature Cited Abell, R., D. M. Olson, E. Dinerstein, P. Hurley ‚1. Т. Diggs, W. Eichbaum, S. Walters Wettengel, T. All- nutt, C. Loucks & P. Hedao. 2000. Freshwater — A C conservation Assessment, Island Puer АЕ . D. Alle "n, G. 1991. Field Guide 1 to He pede Fishes of New ae Publication No. 9. Christensen. Re- searc А pros Madang, Papua Ne W Guin Angel, 1992. Managing biodiversity in he oceans. Pp. У, — їп М. М. A. Peterson (editor), Diversity of Ocean Life. Center for Strategic and e Stud- ies, ees DA ‚ 19 lives вну of the pelagic ocean. Conser- vation B a T: — Bailey, R. G. The Ecosystem Geogra- phy of the D eans — Continents. Springer, New York grs: S. 1995. — in Anatolia, Turkey. Biol. onse "rvalion 7 72 21: Bohn, | . 1994. E. ы project for the construction of a map of the nat e vegetation of Europe at a scale of 1:2.5 million—lts concept, problems of harmoniza- tion — ЛУ ation for nature protection, Colloq. Phy- to 23: . Ecoregions: 2. с an mina. 1996. General Map of Natural Vege tation of Europe. Map (1: 10,000,000). Federal Agency of Nature Conservation, Bonn, Germany. Borhidi, A. 1991, Phytogeography and Vegetation Ecology of Cuba. Akadémiai Kiad6, Budapest, Hungary Briggs, J. С. 1974, Marine Zoogeography. MeGraw- Hill, New ork. Brinton, E. 1962. ne distribution E Pac — кашан Bull. — Inst. Oceanogr. 8: 51-27( Brooks, R. R. 1987. Se rpentine and a Veg >getation: A Mul- У Approach. Ecology, Phytogeography and Physiology Series, Vol. 1. Dioscorides Press, Portland, gon. Brooks, T. M., S. L. Pimm & N. J. Collar. 1997. Defor- estation predicts the number of threatened birds in in- sular S . i : 383-394, Brown, 1994. Biotic atari South- western United States and Northwestern Mexico. Univ. Utah i Salt Lake ep Bryant, D., E. Rodenberg, ' : D. Nielsen. 1995. ( E. al Risk: 2 m "x ‘af ien Development- related Threats to Coastal. Ecosystems. Indicator Brief, World Resource Instit >, о D.C. . D. Nielsen & L. Tangley. . The Last Frontier Forests: Ecosystems and Economies on the Edge. World Resources Institute, Washington, p — 1 — ү McManus & Spalding. 2000. Ree ^ at Risk: A Map-based Ei сы of Threats to the — um s fs. World Resources Institute, Wash- ing BSPIC Vi Ne ЛУ RI/WWF. 1995. A Regional Analysis of Geographic Priorities for Biodiversity Conservation in atin America and the Caribbean. A Re port for US; AID. о sea) Program, Washington, Bullock, Medina (editor). 1996. л, Dry Tropic al Forests. С cambridge Univ. ress, Cambridge. . J. D'Amico, is — = М A H. A. Mooney & E. t HOUR & D. M. : A Con- servation Se Island Bue Washington , D.C. C hange ‘hun Institute of Geography and Chinese Ас ademy of Sciences. 1990, Hi C Atlas of China. чепсе Press, Beijin C — Vegetation Map ЕРЕ Committee. 1979. Veget tation Map of China. Map (1:10,000,000). Science — го, ;onservation . 1992 ». Is ай" а — oll sity cline in the — a Ecol. Evol. Cody, M. L. Dive he ra ян pe conservation in Me ана — regions. Pp. 122-152 in M. E. Soulé (editor), Conservation Bio logy: The Sci ience of Scarcity and Diversity. Sinauer, Sunderland, Massachu- setts. Collar, N. J. & S. N. Stuart. 1988. Key Forests for Threat- ened Birds in Africa. International Council for Bird Preservation Monograph No. 3. S-Print, Cambridge, U.K. Conservation International. 2000: Tropic ‘al forest. wilder- tt nservation.org). эш D.C. e ntal impacts in . Mexico: A commentary. J. 9: ness areas. Web Site: Conservation Inte emational, ( ORR en s. 1 97 gm, 5 — pls ); 85—88. 1992. The Mammals of the ( l. “aly — tud Univ. Press, Oxfor .M.& ( Cowling, |. Hilton-Taylor. 1994. Patterns of plant — and — in — Africa: An over- view. Pp. 31—52 in B. luntley (editor), Botanical Dive a in Southern is ‘a. NE Botanical Insti- lute, Pretoria, South. Africa. Volume 89, Number 2 2002 Olson & Dinerstein 219 The Global 200 Г. Hoffman & C. Hilton- ‹ diversity in southern Нал (editor), Biotic Di- Concepts and Conservation. South Afric: M. Simmons. Gibbs Russell. ~ a Tavlor. s Africa. Pp. 19-50 in B. J. versity in Southern Mrica: Oxford. Univ. Press, Cape Town. . 1. A. W. MacDonald & 1 1996. The Cape Peninsula. South Africa: Різне aphical, bi- ological and historical background to an — hot-spot of biodiversity. Biodive rsily and Conservation 5: 521-5 Croom, M. . Wolotira & W. Henwood. 1992. Pro- Subdivisions of the North East Pa- m. National Ocean Service, Washing- ada—West Coast British Colum- Environment Ca on. Coordination Office. North Vancouver. ‚ 1986. Review of the Protected Areas System in 0c eania. IUCN. Gland. Switzerland. and UK/UNEP, airobi, Kenya. а R. E 1974. Biotie Provinces of the World: Fur- er Development of a System for Defining and Classi- Hen Natural Regions for shes es of Conservation, ICN lo. 9. International Union for Conse rvalion of Nature and Natural Resources. Morge Occasional Paper No zerlan „ Webster, 1995. estrial Ecoregions World Bank. Dine йен, E. D. M. Olson. D. J. Graham. A. | S. A. Primm, M. Р. O & С. Conservation Ledec. of Latin America a aind the Caribbean. The Washington, D Dobson. A. P. J. к Жой; W. M. Roberts & D. Vilcove. 1997. E distribution of endangere : species in the United States. Science 275: 550—553. Ecological Stratification Working Group. 1995. A National Касе Framework for Canada. Agriculture and Agri-food — Research Branch, Centre. for i i urces Research: and Environment — nt Directorate; Ecozone Analysis quid наалы Canada. Land Elder. D. € J. Pernetta (editors). 1991. Oceans. IUCN. Mitchell Terba London. European. Topic Centre on Nature Conservation. 2000. The Digital Map of European Ecological Regions (DME ER). Muséum — d'Histoire Naturelle, Par- is. Franc Finch. С. (editor). 1996. Mongolia’s Wild Heritage. Mon- golia Ministry for Nature and the Environment, UNDI GEF. WWE. Avery Press, Boulder. ата 0. uinea. The Austra- Queensland. Braziller. 1900. Mammals of New Flannery, T. ian mo Em Robert Brown. . 19 New . The Future Eaters. George ak: Forey, P. L.. . Humphries & R. L Vane-Wright (edi- tors). — [m malies and Conservation. К valuation. The Systematics Association Special Volume No. 50. Clare айап Press. — Freitag. H. 1971. Studies in the ni oan vege tation of Af- vie кН Pp. 89-106 in m H. is, Р. Harper & 1. C. Hedge (editors). Plant Life : P Asia. The Botanical Society of Edinburgh. — Frest. T. J. & E. J. Johannes. 1991. Present and Potential inse Mi illu iscs Occurring ithin the Range of the orthern Spotted Owl. Report prepared for Northern — i Pus rv Team. USES, Portland. Oregon. Gage. J. D. A. Tyler. 1991. Deep-sea Biology: A Natural B of Organisms at the Deep-sea Floor, Univ. Cambridge Press. Cambridge. Gallant, A. L.. E. E Binnian, J. M. e & M. B. Shasby. 1995. Alaska. U.5 Survey Professional Paper — U.S. Government Printing LR ie m Gaston, K. J. 4 Williams. Ten Mapping the world's spe — d taxon approach. Biodiversity Lett 1: 2-8. Gentry, А. 1993. Ecoregions of Geological Diversity d floristic ru of Neotropical dry forests. Pp. 146—194 in . Bullock. H. A. Mooney & E. VAN (editors). Se pali Dn Tropical А Cambridge Univ. Press. Cambridge. Goulding, M. The Fishe s and the Forest: Explora- tions in — Natural History. Univ. California Press, im kelev N. J. H. fuilh & D. J. Mahar. 1996. Floods of Along the Amazon. Co- Ecology and Economy P ress; New Yor Deep-sea b nthic biodiversity. Bio- Portum: lumbia Univ. Grassle, J. F. 199 р. Maciolek. 1992. Deep-sea species rich- ness: — and local diversity estimates from quan- Amer. Naturalist 139: 313- Science titative bottom 341. Gray, J. S. conservation needs. samples. Patterns, threats anc and Conservation 6: ‚ Marine biodiversity: Biodive тэу 53-175. Groombridge, В. & M. Seas: Jenkins (editors). 1996. = Diversity of the Я Regional Approach. WEN Biodive silty Series No. 4. World Conservation Ment ing Ce nire, — К gu c Press; Cambridge. . . Freshwater Biodiversity: A e re- mir Glob: ib Assessment. WCMC Biodiversity Series 8. World Conservation Monitoring Centre. Cambridge. ЕРИНИ ‚С. L. Valutis; D. Vosiek, В. Neely. К. Wheaton. Je Тош: àl & B. Dich Is. 2000. Designing a Geography of Hope: A — for Conservation Planning. Special ed.. 2 vols. The — Virginia. M. Adams & W. H. Martin (editors). 992. B А of the Southeastern United States: * uatic Communities. John Wiley and Sons; New York. Hamition, А. C. & R. Bensted-Smith. 1989. Fores i Fasl Usambara Mountains. Tanzania. Practitioners Ecoregional he Nature Conservancy, Ыт Кпеу. Hamilton. L. S.. d. 0. atena (editors). 1995. — a) Montane С loud Fo oresls. d Teologi al Stud- ies 110. Springer-Ve rlag. New York. . Billington (editors). 1990. Г Tropical и The Amer- IUCN, WCMC, Cam- Harcourt. €. S.. J. — The Conservation World буена Union. — idge Hayden. B. . Ray & R. Dolan. 1984. Classification of coastal i marine environments. Environm. Conser- — 11: 199—207. . B. 1996. Historical mogeography o of * In- dian verte brates. e Rev. Ecol. Syst. 27 ‹ Hernandez. Н. R. T. Вагсепаз. 1995. cacti in the € ae ihuan De sert: Ll Distribution patterns. 1 Conservation Biol. 9: 1176-1188. . 9 Hedges, Hilbig. W. 1995. The Vegetation of Mongolia. SPB Aca- demic Press. Amsterdam. The Netherlands Hocutt, С. H. & E. ( ¡ley (editors). 1986. The Zooge- ography of orth — ап Freshwater n ishes. John Huntley. B . J. (editor) "1999. Biste Diversity in Southern 220 Annals of the Missouri Botanical Garden Africa: Concepts and Conservation. Oxford Univ. Press, Cape Town, South Africa. — — (e ditor). 1994. Botanical Diversity in Southern Af- rica. National Botanical Institute, Pretoria, South Afri- ICBP 1992. Putting Biodiversity on the Map: Priority Ar- eas for Global Conservation. Inte rnational Council for Bird Preservation, Cambridge, U. IUCN. 1988. Coral Ree da of ihe "World. WCMC, € Sunbridge. | . 1990. Biodive аа in Sub-Saharan Africa and its СБа. Occasional Papers of the IUCN S vival Commission No. 6. IUCN Gland, Switzerland. ‚ 1991. The Conservation Atlas of Tropical Forests: Asia sud the Pacific. MacMillan, London. — ——. 1992. The Conservation Atlas of Tropical Forests: ux Millan, London. IUC NIUNEP_ 1986a. Review of the Protected Areas Sys- tem in the Afrotropical Realm. IUCN, Gland, Switzer- and. 3 vols. IUCN and Sur- Unit, )ecies Publications — 30b. Review of the P — Areas System in the — Malayer Realm. IU( Gland, Switzerland. IUCN/UNEP/W WF. 1987. — ar: An Environmental Profile. IUC М, G — Switzerland. Johannes, R. E. & B. G. Hatcher. 1986. vid tropical marine environme * Pp. 371—392 in М. yulé (ed- itor), Conservation Biology: The Science of m arcity and Diversity. Sinauer, Sunderland, Massachusetts. Johnson, T. H. & A. J. Stattersfie = 1990. A global review of island endemic birds. Ibis : 167-180. Kalin rm M. T., P. H. 4 эж = M. D. Fox (editors). 1995. Ec ology and Biogeography of Mediterranean Eco- Sus ms in Chile, California, and Australia. Springer- erlag, js w York. — J. T. & C. Meacham. 1999, Synthesis of the North ral rican Flora, Database. Version 1.0. North — Botanical e C — Hill. Kelleher, G., C. Bleakley Wells. 1995. Represe — System a — Protected |. vols. Great Barrier Marine Park Authority, hi — Fa World Bank, IUCN, Washington, D.C Kellogg, E., J. Weigand, A. Mitch ell & D. Morgan. 1992. Coastal Temperate Rain Forests: Ecological Character- istics, маи and Distribution Worldwide. Occasional Paper Series No. ;cotrust — Put Orekan: Kendall, & M. Aschan. 1993. Latitudinal gradients in the struc — of macrobenthic communities: A com- — of Arctic, temperate and tropical sites. J. Exp. r. . Ecol. 172: 157-169. Kingdon, T 1989. Island Africa: The Evolution of Africa’s Animals and Prine ier 997. African Mammals. and Conservation Inter- Plants. Princeton Univ. Press, Academic Press, San Disgo gy A. 1987. Pun Natural History of the USSR. w-Hill, New Rotts on м. & T. mos п. 1996. Asia-wide Assessment of Freshwater Biodiversity. World Bank Technical Paper Number 281. The World Bank, e d C. Krever, V, E. Dinerstein, Olson & L. Williams. 1994. Conserving Russia's — a Dive rsity: An An- лш al Framework and Initial Portfolio. е D. Kuma S. F. 1990. — regionalization of the USSR. Map (1: DA. Sheet 15 of Forests of the USSR. M Investment Main Division for — and Cartography of the So- viet of Ministers, Lacerda, L. D. (с a a 1993. Conservation and Sus- amable Utilization of Mangrove Forests in Latin Amer- ica and Africa Regions. Parts | and 2. Society for Mangrove Ecosystems and Tropical Timber са Okinawa. ee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister & J. R. a Jr. 1980. Atlas of North American Freshwater Fishes. North Carolina Biological Survey L — — — International International ‚ 1997, jug беш Dynamics and Conserva- > Freshwater Fish of Africa. Cambridge Univ. Press, pon ae tta LK. —— D. Paugy & G. G. Teugels (editors). 1992. Faune des Passani d'Eaux Douces et Saumátres de l'Afrique IC tion: de Ouest. Tome 2. — TOM Lieske, E. & R. — р Reef Fishes. Prince- ton Univ. Press, Pr rinc Longhurst, A. 1998. Ec — al Geography of the Sea. Ac- ademic P ress, London. Lovett, J. C. & S. К. Wasser (editors). 1993. Biogeography and E — of the Rain Forests of E astern Africa. Cam- bridge Univ. Press, Cambridge, Lowe-McConnell, R. H. 1987. Ec logic ‘al Studies in Trop- ical F a Саш Cambridge Univ. Press, Cam- bridge, l Mac B. T & N ress, C ambridge . Hicks. „ Massa ‘tts. — ——— & G. Bunting. 1996. ууш ТЭЙ Natural Habitats of the Indomalayan — ıl Database. World Conse p „Moniióring Centre, Cambridge, MacNae, W. 1968. A general account of i — and flora of mangrove swamps and forests of the Indo-West- Pacific re — Adv. Mar. Biol. 6: 73—270. Maggs, С. . H. Kolberg & C. J. H. Hines. 1994. НИ dise ity in Namibia. Pp. 93-104 in B. J. Huntley (editor), Botanical Diversity in Southern Africa. National Botanical Institute, Pretoria artin, С. 1991. The Rainforests of West Afric y Ecolo- 19 Duca C ees — Martin, W. H., . Boyce & A. Echternac bh (editors). 1993. — of Mie Sout us astern United States: Lowland Terrestrial Communities. John Wiley and Sons, б York Maxwell, ia C.J, — M. E. Jensen, S. J. Pau tian. H. ae & D. M. Hill. 1995. A Hie — ا‎ of Aquatic ee ‘al Units in North Amer- 1 (Nearctic ты, USDA Forest Service General Тес мк — NC- 176. St. — — McAllister, ‚ F. W. Schueler, C. M. Roberts & J. P. Hawkins. А Mapping and G É analysis of the global distribution of coral reef fishes on an equal-area grid. Pp. 1 К.Т. “Miller (editor), Mapping the Di- ‘sity of vv Chapman and Hall, London. McClanahan. T. R. & T. P. Young (editors). 1996. East African E and Their Conservation. Oxford Univ. Press, New Yor McDowall, R. (editor). 1996. Freshwater Fishes of South- eastern Antal Reed Books, Chatswood, Australia. Mitchell, Fragile Paradise: Nature and Man in the ке C ollins, London. Mittermeier, R. ` B. Werner & 1996. New Caledonia—A orion imperative for an ancient land. Oryx 30: 104-112. >. Mittermeier & P. R. Gill. 1999, Earth’ s Biologically Richest and Most Endan- 1996. Wild China. The MIT > Basel. А. Lees. . N. Hotspots: Volume 89, Number 2 2002 Olson & Dinerstein The Global 200 gered Terrestrial Ecoregions. CEMEX and Conservation — Washington JC. — A. 1975. Outline of Japanese vegetation. Pp. 1 K. Nana К. Yoshida & М.К a as in Conservation of Natural Terrestrial Ecosys- " ms in ZA JIBP Synthesis Vol. 8. Univ. Tokyo Press, . Jap — Ministry for Nature and the Environment, Unit * ations De -velopme nt Programme (UNDP)/G "m Environme ent dy ility (GEF). and World Wide Fund for 1996. Mongolia’s Wild Heritage. Avery S. Poe Colorado. Mori, T. 1936. Studies on the Geographical Distribution of Freshwater. Fishes Univers sity, v Vor n. poe Myers, G.S. and the e Do T higher taxonomic lution 14: 323-33: ato (editors). zu | Eastern. Asia. Keijo Imperial . The jene mic fish fauna of Lake Lanao. categories. Evo- — чег, C. G. Mittermeier, G. A. B. Kent. 2000. Biodiversity hotspots for 858 vers, N., Ag A. da Fonseca & J. conservation priorities. Nature 408: 853— Nagelker Н . M. V. Mina, T. Wudneh. F. A. Sib- bing & qo Dass. 1995. In Lake Tana, a unique fish fauna — protection. BioScience 45: 772-775 New Zealand Department of Conservation. 1987. Ecolog- ical Regions and Districts of New Zealand. Map (1 500.000). Department of Conservation, Wellington, New Zealand. Nichols, J. Л. 1943. The Fresh-water Americ due i Muss um of Natural History, l. (12—44 id. Fishes of China. Ne "Ww You К Nicol & О. — 1989, Madagascar: Revue de la Es onservation et des Aires Protégées. WWE Inter- National, : — Switzerlan d. Noirfalise ap of the Natural Vegetation of the Me — ountries of the European Community and the Council of Europe. Map (1:3,000,000). Council of Eu- rope, жае France. Norse, К. A. 199 ‚ Маййайв the world’s marine biolog- ical 7 a Bull Mar. Sci. 57: 10—13. Oates, К. Apis Primates: Status Survey. and Conservation p tion Plan, revised ed. IUCN, Gland, Switzet ia Oberdorff, T.. J.-F. Guegan € B. Hugueny. 1995. Global scale us rns of fish species richness in rivers. Ecog- raphy 18: 345-352. Olson, D. M. & E. Dinerstein. 1998. The Global 200: A represe ntation approach lo conse туш the Кап» vu biologically valuable ecoregions. Conservation Biol. 502—515. A Con- America . Cintron & P. lolster. 1996. servation — sssment of Mangroves of Latin and the Caribbean. Conservation Science WWF-US. oe D.C. Program, — ———, В. Chernoff, C. Burgess, I. Davidson, P. Canevari, Dinerstein, "Gt Castre Morisset. R. Abell & E. e 1999, Fre ut T dodici of Latin America and the Caribbean: A Conservation Assessment. WWF- US, Wetlands International, Biodiversity Support. Pro- gram, е USAID, W: — D.C. ——— F. Dinerstein, E. ici E И . D. Bur- gess, e V. N. Y vell. E. C. U nde rwo * a . D'Amico, * Morrison, € . J. Louc T. F. АП. | H. Rie kets, I. Tue e W. Wetten- Pp. Hedan & K. Kassem. 2001. мамы мину ч ths world: A new map of life on Earth. BioScience ql : Omernik, J. M Terrestrial 933-938. 1995. bud Ш Ecoregions of the Conti- National Academy, Washington, — — 199 nental US. National — and Environment Effects Research aborato S Environmental Protection Agency, Washington, n. C. Oosterbroek, 1994. Biodiversity of ^x Mediterranean region. Pp. ege in P. L. Forey, C. J. Humphries & { Vane-Wright (editors), Syste matics and Conserva- tion EvalusBon. Vol. No. 50. Clarendon Press, — Orians, G pim Endangered. Applic 200-20 башы. E The Systematics * iation Spec ial ~ den level? Ecol. J Gage & M. V. Angel — 1997. Marine p е rsity: Patterns and Proc и Univ. Press, Cambridge, U.K Parker, ine iu J. L. Carr (editors). 1992. Status of — Ren — in the Cordillera de | Southwestern Ecuador. RAP Working Papers 2. Conservation Inte rational, jacent Areas of ' Washington, D.C. . H. Ge * R. 'oster, L. H. Emmons & J. V. M msen, Jr. 1993. The land Dry Forests of Santa Cruz, Bolivia: A d Tobal Conservation Priority. RAP Working Papers 4. Conservation International, Wash- ington, D.C. А ОР. Jr. 1991. The I Timber Press, Portland, Peterson, ^. T. О. es- Villela, Llorente- hocque Is, M. A. evs rro-Sigiienza, Pines of Mexico and Central id us LS. — Paniagua, p Martinez, A. G. Torres-Chávez — | & l. — Fernández. 1993. Conservation. priorities in. Mexico: — ap in the world. Biodiversity Lett 33-38. Picker, M. D. & M. J. Samways. 1995. Final diversity and endemicity of the Cape Peninsula, South Africa: A irst assessment. Biodiversity and Conservation 5: 591— 6006. Pierrot-Bults, A. С. 1997, Biological diversity in oceanic mac rezoaplankton More than counting species А 93 in R. & М. ^p. . Ormond, J. D. Gage ¢ V. Angel, Marine duced sity: — ‘rns s and Processes. Cambridge iniv. Press, Cambridge, І gral & D. A. Keith. 1994. New procedures for reserve selection in New South Wales: Pressey, R. L., Maximizing the chances of achieving a representative network. Pp. 351-373 in P. L. Forey, < R. I. Vane-Wright (editors), Systematics and Conser- vation Evi — The J. Humphries > Systematics For lation Speck i ol. No. 50. Clarendon Press, Oxford. о Mos K. 1991. | File, Oxford, U. был T. P, К. Bye, Madagascar: A — History. — ^. Lot & J. Fa (editors). 1993. Biological Diversity of Mex tion. Oxford Univ. Raven, P. H. 1988. Our diminishing topi cal forests. Pp. п 9-122 in E. O. Wilson (editor), Biodiversity. National ademy, eec d nite С. ( oceans. » 36— ico: Origins and Distribu- Press, New Yor ~ 88. Ecologic ‘al dive ersity in coastal zones and 50 in E. О. Wilson (editor), Biodiversity. C. Coastal-zone biodiversity patterns. Bio- — 11: 490—498. & B. P. Hayden. 1993. Marine biogeographic коө of the Bering, С -— ‘hi, and Beaufort Seas. Pp. 175-184 in K. Sherman, L. M. Alexander & B Gold байо, Large Marine — Patterns, Pro- Yields. vancement of W ience, Washington, D.C. Reaka-Kudla, M. L., D. E. Wilson € E. O. Wi tors). 1997. M Rs Il: Understanding and Protect- — cesses and American Association for the Ad- Ison (edi- 222 Annals of the Missouri Botanical Garden ing our Biological Resources. Joseph Henry. Press, Washington, D.C. Reeves, R. R. & — Dolphins, Por- Action Plan for the Gland, S See awn 5. —€— poises, and Whales. 1994—19 Conse dos of Cetaceans. IUCN, Ric ke tts, . Dinerstein, D. > Olson, C. J. Lou . Eic РА p. x llaSala, K. Kavanagh, P. — Р T. Hurlev, Carney, R. Abell € S. Walters. 1999 ‘ial Ecoregions of North ; : А Сопзегуа- 1 Press, Washington, D.C. Terrestr tion Assessment. Islan E. & ROE mangrove floras. Рр. 215- Ricklefs & D. Schluter (editors), Species pote in Ecological Communities. Univ. Chicago Press, Chicag Robbins. C. R. 1991. Regional dive ape among Caribbean fish species. Bios ence 47: 458— Roberts, T. R. 1975. Е val — vin 'an freshwater fishes. Zool. Ann. Soc. 57: 249-: Sherman, K., 1 Gold — 1990. Large Manne Ecosystems: Patterns, and Yields. American Association for the Advancement of Science, е С. Silva, J. М. С. 1998. Um Método рага о ipei 'imento de Áreas Prioritárias para a Conservação na Amazonia um Report pre pared for World Wildlife Fund. Bra- P rocesses Spalding, M., F. Blasco & C. Mangrove Atlas. Ecosystems, Okin Field (editors). 1997. World dbi International Society for Mangrove wa, Japan State of the — nt Advisory Council. 1996. Austra- : State of the Environment. De ‘partment of the Envi- ronment, Sport and Territories. CSIRO Publishing, Col- Australia. n Stattersfield, A. J., M. J. Cros sby, A. J. Long & ae 1. Wege. 1998. Endemic s Bird Areas of the Wo — irdlife Con- servation Series No. 7. Birdlife но riae zam- bridge, U. Stewart, J. M. 1992. The Nature of Russia. Boxtree, Lon- don. Stewart-Cox, В. 1995. Wild Thailand. MIT Press, Cam- bridge, Massachuse tts — J. A.. P. Bens & B. К. Holst — )5. Flora of the Ve 'nezuelan Guayana. Vol. 1: Intro- rii Missouri Botanical Garden, St. Louis Stiassny, M. L. . K. Sehliwen € W. J. — 1992. new species Hoe k of cichlid fishes from Lake Bermin. Cameroon, with a description of eight new species of Tilapia (1 ^ bruidei: Cichlidae). Icthyol. Explor. Fresh- waters 3: 311—346 Suchanek, T. H. 16 9 I. Temperate coastal marine com- munities: Biodiversity and threats. Amer. Zool. 34: 100-114. Sujatnika, J. P., T. R. Soehartono, M. J. Crosby - diastuti. 1995. p onserving Indone — Bio liv A. ` Endemic Bird Area Approach. BirdLif Indonesia : "rogramme, Bogor, Sullivan, K. . Bust vation ns 'ssment of the Freshwater Marine Ecore- nd ‘aribbe ап. BSP/WM ife we ed sa sia. tamante — 1996. A Conser- gions of Latin America папі the ( TNC, Washington, Taki, Y. 1975. primary freshwater fishes in кш — ipal areas of Southeast 0—214 м — — distribution. of Asia. $ Asian Stud. (Kvoto Univ.) 13: 20 Terborgh, 1. & В. Winter. ion \ me Mar for siting patio and reserves with special reference to Colombia and Ecuador. Biol. Conservation 27: 45-58. Thackway, R. & I. D. Cresswell (editors). 1995. An Inter- Latham. 1993. G — IE rns of R. E. Ec alos of Java and Bali. im р Regionalisation for Australia: Framework for Setting Priorities in the National Re- serves System на Program. Version 4.0. Aus- tralian Nature Conservation Agency, Canberra. Thieme. M. L., R. A. Abell, D. M. Olson, E. Dinerstein, M. L. J. Stiassny, J. D'Amico & E. M. Underwood. In press. Freshwater regions of Africa: A Conservation Assessment. pes Press, Washington, D.C. Timberlake, J. R „ Müller. 1994. Identifying and de- scribing areas ior vegetation conservation in Zimbabwe. Pp. 125-140 in B. J. Huntley (editor), Botanical Diver- sity in S Africa. National Botanical Institute, Pretoria, wy Africa. Udvardy, M | 1975. World ре Provinces. CN B al Paper No. 18. Gland, Switzerland. lisse Academy of Sciences. 1988. Red Book—Russian Soviet Federal Socialist Re public. 2 vols. USSR Acad- у of Science, is cow Vane- Wright R. L, C. J. a & P. H. Wil 1991. What to potect?—Syst ematics and b agony of qu Conse — 55: 235-254. быш, 1. N. 1995. Corals in Space and Time: The goin and Evolution of the Scleractinia. Com- 5 те — — New Yor . Ice-Geography ‘of Greenland. Map (1: 5. 000,000). Genetisk Institut, Copenhagen, De nmark. Vinogradova, N. G. 1 liams. 77 LI me o 7. Zooge dw of the abyssal and 87 in ` . Blaxter, A. J. A. V. Gebruk & P. A r. r (editors), Ad- vances in manne parma bur 32: ‘The Biogeography of San Diego. A Funk (editors): 1995. Hawaiian — on a He и Spot Archipelago Wagne Bioge gr — Earth's Living Resources. Chapman and Hall, London. C. & A. J. Lc 5. Kev Areas for Threatened — in the Neoiropitis Birdlife gc Series No. 5. BirdLife International, Cambridge. — М. J. A. (editor). 1978. Hinge pe Ecology of Southern Africa. Junk, The Hague. Wetlands International and The I — lands of International Importance. i World Bank, Washington, D.C. i | 33. The Vegetation of Africa: A Descriptive Memoir to Accompany 1 the UNES SCO/AE! l'FAT/UNSO Afric a. Natural Resources Research 1996. Wet- Asia. Map. The Whitmore, T. C. Hi n Rainforests of the Far East. Oxford Univ. s ss, Oxford, U.K. 1990. An ntti to Tropical Rainforests. Clare ndon tress, Oxford, U.K. & ( ance (editors). 1987. Bioge — — umen tina in Tropical America. Oxford Mor ogr. de gee No. : و‎ Press, Oxford, UK. Whiten: A. J.. M. Mustafa & G. 1987a. The Ecology : те AR U niv. Press, Jakarta, Indonesia. S. J. Damanik, J. . Henderson. Anwar & N. Hisyam. 1987b. The Е — of Sumatra. Gadjah Mada Univ. Press, Yo- gvakarta, Indo mesia. R. E. Soeriaatmadja & S. Afiff. 1996. The The Ec MS of — Se- ries, — ne Il. Periplus E — s, Singapo Wiker . (compiler). 19 Te earal Fco 'oregions of ina и al Land С Pa ‘ation Series, No. 19 Environment Салай, Hull, Quebec. Volume 89, Number 2 Olson & Dinerstein 223 The Global 200 Wikramanayake, E.. К. — C. Loucks, Olson, J. Morrison, J. — M. Knighi & Е — 2001. Terrestrial Ecoregions of j Indo- Pac ifie: A Con- servation enn, C Williams. киюе» ге aede sss, and endemism in conservation. E 269-287 in Р. L. J. Humphries Vane-Wright E aul tiation 50. Clarendon Press, Oxford. & R 'orey, С. (editors). Systematics and Conservation The Systematics Association ай ial Vol- ume No. | ane-Wright. 1991. Measuring Бейге: € re d 388 for conservation лое Is 5—07 out Biodiversity — Acad- 2. The Diversity a Life. Belknap Press, Cam- bridge, е husetts. WWF/IUCN. 1994—1997. Guide and Strategy for their Conservation. LK. Centres of Plant Diversity: A 3 Vols. IU Publications Unit, Cambridge, Û Yim. Y.-J. 1977. Distribution of forest vegetation and cli- mate in the Korean Peninsula: IV Zonal distribution of orest vegetation in relation to thermal climate. Jap. J à — ү 27: “Ж” Li 87 A Zoogeography of some primary Asia. Walla- CO Zakaria- Ismail. freshwater EN s in South and Southeast ceana 17: 6-9. . 1994. Zoogeography and biodiversity of the fresh- Wate r fishes of Southeast Asia. Hydrobiologia 285: 11 — V. J. & S. V. Gerd. 1961. Fauna and Flora of the ivers. Lakes. and Reservoirs of the USSR. Smithson- ian Institution and National Science Foundation, Wash- ДН inglor Zhao. he ng Guangmei, Wang Huadong & Xu Jialin irs 1990. The Natural History of China. McGraw- Hill. New York. Zoharv. M. 197 » Geobotanic al Foundations of the Middle East. x Is. ustav Fischer Verlag. Stuttgart. Ger- many. APPENDIX l. WEIGHTING AND MEASURING BIOLOGICAL DISTINCTIVENESS CRITERIA The we ighting and measurement of the parameters used 1 s of terrestrial ecore- to assess iological distinctis gions of Nort i merica are d here to detta how different biodiversity features were evaluated as con- servation targets and how analyses were tailored to differ- ent biomes. Comparisons among biodiversity. parameters were only conducted within the set of ecoregions sharing the same biome. SPECIES RICHNESS ® G lobally outstanding 100 ligh lo Medium 10 A *Only nalive species were used in species counts. ENDEMISM Globally outstanding 100 High 25 Medium 15 W 5 For species richness and endemism, the total number he of species that occurs within each ecoregion, and the total al ‘Species — and native vascular plants, regional experts — Балу Roth analyzed land snail Histon for western and ed rich- i John ness — endemism —— for native v ort th A meric a, cies was consider ndemic^ to an ecoregion if its es- imated range fell « entire сн within a single ecoregion (strict vithin a single endemic), 75% or more of its range fell y ecoregion (near-endemic), or its range was less than 50,000 km” (range-restricted). If a species had a signifi- cant distribution outside of the United States and — а, il was nol conside red as an ende mic. Higher taxonomic uniqueness—e.g. cies or communities, primitive lineages—was : ered for identifving globally outstanding ecoregions from an endemism perspective. The actual number of species and endemies for each laxon found within an ecoregion were log transformed to uce the influence of very species-rich groups. The logs > then summed to derive a single richness and ende- ‘hese scores were plotte d for the ec oreglons < mism score. within each biome and the curves broken subjectively into high, medium, and low scores. Globally outstanding scores were determined through ecoregions within the same biome found throughout the comparisons with values for work UNUSUAL ECOLOGICAL OR EVOLUTIONARY PHENOMENA Globally outstanding 100 Regionally outstanding 2 No globally or regionally unusual phenomena 0 e P, nm J amples of unusual оне ев volutionary phe- | Ex nomena al global or гери за SC ‘ales i me lie re alive M in- tact, caribou. intact predator a centrations of breeding waterfowl and shorebirds. extraor- arge-scale migrations of large vertebrates such em ibl: age 8: supe abundant СОП" dinary levels of adaptive radiations, rain-fed flooded grass- lands on limestone, and conifer forests dominated. by gigantic trees. GLOBAL RARITY OF BIOME Global rarity 100 Regional rarity 5 Not rare al global scale 0 ts that were considered globally rare woodlands, an Biomes or habitat editerranean-climate forests. include тар. te mperale rainforests. and paramo. TOTAL SCORES FOR DETERMINING BIOLOGICAL DISTINCTIVENESS INDEX The points from each criterion were summed to arrive at a final score. This score was then translated into a bi- ological distinctiveness category as follows Globally outstanding 15, А or 55+ points Regionally outstanding 30, 35. 40 Bioregionally outstanding — 20. T Locally important 224 Annals of the Missouri Botanical Garden Ecoregions identified as globally outstanding were sub- sequently compared with similar ecoregions around the world to validate their hit status. APPENDIX 2. ASSESSING CONSERVATION STATUS OF ECOREGIONS Conservation status measures landscape and ecosys- tem-level features and relates inet to the ecological in- tegrity of with increasing habitat loss, degradation, and bun ia ecological processes cease to function naturally, or at all, resiliency to distur- bance declines, and major components of biodiversity are steadily eroded. We a conservalion status « ecoregions in the tradition of IUCN Red Data Book cate- gories for threatened and endangered species: critical, en- dangered, and vulne га For ес — we — Hie fol- lowing | conservatior мац categories: ritical, endangered, vulne e re dis sly zie and re Чай ly intact. Throughout all of the regional analyses, the specific parameters and thresholds used for assessing conservation status were tailored to the characteristic patterns of a у, Kin ‘al dynamics, and responses to distu bance of different biomes. ecoregions, namely, how assess the diverait TERRESTRIAL ECOREGIONS We present the method used to assess conservation sta- tus for the terrestrial ecoregions of North America to il- lustrate the approach (Ricketts et al., . The re ае contributions of different 4096—habi ze of remaining blocks of — o 20%—de gree of habita fragmen- tation, and 1 servation status apshot con- vel of en- dangerment. The point thresholds for different categories of conservation status were as follows: critical 89-100 5—88, vulne rable 37—64, РА ج‎ = b | . were determined by summing points — for each ра- rameter. Individual parameter point values were associ- © ated with different landscape scenarios. For example, — habitat loss scenarios were related to points as follow % Original habitat Heavily altered Ме 90-100% 40 20 75-89% 30 15 50-74% 20 10 10-49% 10 5 9% 0 0 3 3 w A © E: S es £z ص‎ — = — habitat (10 pointa), and therefore 10% intact hab- . By combining the two scores, the ecoregion would receive a total score of 20 polola: "рите rent quantitative and qualitative biodiversity and landscape ecology char- acteristics are used to define intact, altered, and heavily altered states tailored to the specific patterns and dynam- ics of different biomes. Total scores for each of the param- eters are summed to give a total conservation status index SCC ОР: = apshot scores were subsequently modified by a 20- s assessment as facing high threat were elevated to a servation status. The threat analysis estimated the cumu- lative impacts of all current and projec = threats on hab- habitat degradation, and wildlife exploitation using a point system assoc — with different qualitative and quantitative impacts. Using an n — 0— 100 points, pending threats within an ec oregion sessed and point totals assigned for each of f the > ae cate gories. c onversion threats were considered to be the most serious, and thus habitat loss c omprised half (50) of all possible points in the weighting of threats. For exam- ple, 50 points were assigne or more of remaining habitat would be categorize heavily altered within 20 years. For conversion of be sit 10% and 24% of remaining enn a score of 20 points was assigned. The remaining two threats, е —— dation and wildlife e — ation, were assessed u imum point totals of 30 and 20 respectively ен asc * based on high, а edi or no threat. itat conversion, > EL GENERO JUSTICIA Cecilia Ezcurra? (ACANTHACEAE) EN SUDAMÉRICA AUSTRAL! RESUMEN El género Justicia es el más grande y — de la familia Acanthaceae, y está muy diversificado en las regiones iste trabajo с ompre nde la revisión de las espec ies de Sudamérica austral Las 38 especies de Justicia de esta región se agrupan en 8 secciones cápsula y semilla. Todas las especies e da una clave para tropic Шш y subtropicales de América del S sobre la base de material de Argentina у — ay. en relación a caracteres de su morfología. espec ialmente de inflorescencia, flor, se describen, se dan referencias de ilustraciones ya publicadas o se ilustran por primera vez, y s para la region, y la nomene — de las mismas incluye varias sinonimias identificarlas. Algunas no habían sido citadas rales. También se dan notas sobre su fe nologí . hábitat, distri- nuevas basadas en el estudio de los tipos nomenclatu bución geográfica, caracteres distintivos y afinidades taxonómicas. ABSTRACT The genus Justicia is the largest and most complex of the Ac — ‘eae family, and is highly diversified in tropical and bolo ‘al regions of South America. This work revises the species of southern South America based on material from Argentina and Paraguay. The 38 spes ies of Justicia that are found in this region are grouped in 8 sections in relation to their — ‘al characters especially inflorescence, flower, capsule, ed seed morphology. All the spe cies are described, references to published illustrations are given or they are illustrated for the first time. and i is provided for their identification. Some had not bee n reported pre ‘viously f for the region, and their nomenc Шз, inc — several new synonymies based on the study of type material. Notes on phenology, habitat, geographic distribution, distinguishing c characters, and taxonomic affinities are also give Key words: Acanthaceae, Justicia, morphology, taxonomy, southe rn South America. El género Justicia L. es el más grande y com- (Burkart, 1943) u ornamentales (Bailey, 1949; Pa- rodi & Dimitri. 1980). Muchas de sus especies tam- plejo de las Acantáceas a nivel mundial, y el de bién tienen importancia ecológica por ser elemen- mayor número de especies en los regiones tropi- cales y subtropicales del continente americano. tos abundantes en el sotobosque de selvas у Este género comprende aproximadamente 600 es- bosques húmedos, o por ser frecuentes o dominan- y arbustos perennifolios (Graham, tes en ambientes semiáridos. En México está pecies de hierbas 1988). varias de ellas importantes como forrajeras presentado por alrededor de 75 especies (Daniel. ! Agradezco a Tom — | por е srme cedido generosamente el estudio de las Acantáceas de Paraguay que origi- defines nte. pensaba realizar él. Mi reconocimiento a Angel Cabrera por sugerirme este tema hace varios años, Vladimiro Dudas por las ua iones, a Ana Ladio y — Amos por ejemplares e índices, y a los curadores y personal de los siguie ntes he rbarios por facilitar material para este estudio: BM, CORD, CTES, E, ЕСО. G, K, LIL, LP. М, P NY. MO, SI, US, W. También agr adezco a Fátima Mereles, HM Soria su valiosa ayuda en Parto tanto en el herbario (FCQ) como en el campo. colección y sus interesantes datos sobre distribuci ion a asistencia en la edición de las listas de Basualdo. Rosa Diesen у Nélida S a Carlos Saravia Toledo su invalorable colaboración en viajes de y ecología de las Acantáceas del Chac "te trabajo se realizó en parte con una beca « Tinker para realizar — en el Jardín Botánico de Missouri, St. Louis, realizar estudios en el National Museum of Natural History. Washington, y una beca del Jardín РН о de Kew (cor financiación de Mellon Foundation) para trabajar en herbarios europeos. También el proyecto PIP CONICET 1 20/96. financió parte de mis visitas a herbarios de Argentina. Agradezco a todos los que me ayudaron en estas instituciones, Dick Brummitt, Enrique Forero, Kaj Vollesen y Dieter Wasshausen. Mi reconoci- Missouri Botanical Garden (con financiación de la Fundación una beca de la Smithsonian Institution à 1 y muy especialmente a Luis Ariza, miento también a los que de m y otra manera me estimularon para que ea, este trabajo. en especial a Vicloria Graham y Lucinda Me Dade. Y mi especial consideración a los tres revisores anónimos y a los editores que gracias a su detallada y rigurosa larea me — a — Mi trabajo es financiado por el Consej jo Nacional de Investi- gaciones Científicas y Técnicas (C ONICE T) de Argentina a través de la Carrera del Investigador a la que pertenezco. ? Departamento de Botánica, Centro Ee ded Universitario Bariloche, Universidad Nacional del Comahue, Quintral 1250, 8400 Bariloche, Argentina ANN. Missourt Bor. GARD. 89: 225-280. 2002 226 Annals of the Missouri Botanical Garden 1993), en Colombia por 85 (Leonard, 1951—1958), en Ecuador por 27 (Jörgensen € León-Yánez, 1999), en Perá por 50 (Brako & Zarucchi, 1993), y en Argentina por 28 (Ezcurra, 1993a, 1999). Hasta ahora no existen trabajos sobre la siste- mática de este género en la porción austral de América del Sur, excepto algunos tratamientos re- gionales (Lindau, 1894, para Argentina; Rambo, 1964. para Río Grande do Sul. Brasil: Dawson, 1965 y 1979, para Buenos Aires у Entre Ríos, Ar- 1969, para Santa Catarina, Brasil; Ariza-Espinar, 1971, para el cen- gentina; Wasshausen & Smith, tro de Argentina; Ezcurra, 1993a, para Jujuy. Ar- gentina). Debido a esto se propuso realizar un es- tudio de las especies de Argentina y Paraguay, países donde el género está especialmente repre- sentado y presenta una diversidad importante. El objetivo principal de este trabajo es describir la morfología y resolver la nomenclatura de las es- pecies de Sudamérica austral, lo que facilitará la identificación de la mayoría de las especies de la región subtropical y templada de América del Sur. Además este estudio comprende un análisis mor- fológico teniendo en cuenta especialmente la ar- quitectura de la inflorescencia, la morfología de la corola, y la forma y superficie de la semilla, carac- teres que ya han sido sefialados como importantes en la delimitación de grupos de especies afines dentro de Justicia (Ariza-Espinar, 1971: 1988: Lester & Ezcurra, 1991). Graham, — HISTORIA TAXONÓMICA Y DELIMITACIÓN DEL GÉNERO El género Justicia fue propuesto por Linné en 753. Un siglo después Nees, en su tratamiento de las Acantáceas a nivel mundial (1847b). redelimitó el género excluyendo más de 600 especies que se habían descripto bajo el nombre de Justicia, re- duciéndolo a solo 12 especies de Asia y África. Por otro lado creó varios géneros nuevos afines a Jus- ticia para ubicar un gran número de especies del Nuevo Mundo recientemente descubiertas. Ben- tham (1876) y Lindau (1895) ampliaron el concepto de Justicia de Nees y redujeron a la sinonimia mu- chos de los géneros creados por este último autor. Durante este siglo, los estudios taxonómicos de Jus- ticia han seguido estas dos tendencias opuestas a través del tiempo, ya sea reconociendo un número grande de géneros afines a Justicia s. str. (por ej., Rizzini, 1949, 1951; tando una definición amplia que incluye a la ma- Bremekamp, 1969), o adop- yoría de éstos como sinónimos (por ej.. Leonard, 1951-1958). El problema de los límites del género ha sido discutido ampliamente en épocas recientes (por ej., Gibson, 1972; McDade, 1982; Ezcurra, 1988; Graham, 1988). Actualmente la mayorfa de los autores tratan a Justicia en su sentido más amplio siguiendo la pro- puesta de Graham (1988) (por ej., Wasshausen, 1992: Profice, 1993; Daniel, 1995). Este concepto amplio ha resultado en esti- 1995; Kameyama, maciones de 600—700 especies para todo el género 1988; McDade & Daniel, 1998). Sin embargo, no se sabe si la gran en los últimos años (Graham. heterogeneidad de números cromosómicos que han sido citados para el género (Daniel, 2000) estarían reflejando evolución diversificadora a nivel cromo- sómico, o características polifiléticas del mismo (Daniel, 1993; Daniel et al., 1984; Piovano & Ber- nardello, 1991). Análisis moleculares preliminares (McDade & Daniel, 1998) indicarían que Justicia s. lat. conforma un linaje separado del resto de la tribu Justicieae, y que la gran diversificación mor- fológica y la alta tasa de especiación en este género se ha dado concomitantemente con muy poco cam- bio a nivel molecular en los loci analizados hasta el momento, lo que apoyaría la primera hipótesis. Actualmente se está trabajando para dilucidar las relaciones filogenéticas entre los taxa afines y per- tenecientes a Justicia utilizando caracteres mole- culares, morfológicos y cromosómicos, y un número mayor de representantes especialmente del Viejo Mundo (McDade et al.. 0) Siguiendo el concepto de Graham (1988) el gé- nero Justicia comprende especies caracterizadas por corolas bilabiadas con surco estilar en la parte posterior interna, presencia de 2 estambres exsertos bajo el labio posterior, con 2 (raro 1) tecas, ausen- cia de estaminodios, polen subprolado a perprolado 2- o 3(4)-porado o colporado, y cápsulas 4-semi- nadas. Este criterio de delimitación del género es el que se sigue en este trabajo. MATERIALES Y MÉTODOS El estudio se realizó sobre la base de material de los principales herbarios de Paraguay y Argen- tina, y de varios herbarios extranjeros (BM, CORD, CTES, E, FCQ, G, K, LIL, LP, M, MO, NY, P, SI, US, W (abreviaturas según Holmgren et al., 1990); Apéndices 1, 2). Los ejemplares citados se selec- cionaron como para ser representativos de la dis- tribución y diversidad de cada especie en la región. La identificación de las especies se hizo por com- paración con material tipo o fotografías del mismo, lo mismo que la resolución de los sinónimos no- menclaturales. Las fotografías de ejemplares tipo de la colección del Field Museum de Chicago, se designan con la abreviatura "fot. Е y el número Volume 89, Number 2 2002 Ezcurra 227 Justicia en Sudamérica Austral de la colección fotográfica. Este estudio incluvó la revisión de más de 220 ejemplares tipo. En la mayorfa de los casos de nombres creados por Nees (1847a. 1847b). este autor no designó ho- lotipo sino que citó varios ejemplares representa- tivos que deben considerarse sintipos. Debido a que no pude consultar todos los sintipos de cada nombre por estar estos ejemplares distribuidos en varios herbarios europeos, en este trabajo general- mente no designé lectotipos de entre estos sintipos. Nees (1847a. 1847b) tampoco designó claramente una variedad tipo cuando estableció varios taxa in- fraespecíficos para una especie, sino que las desig- nó con letras del alfabeto griego. Por las razones ya expuestas previamente (Ezcurra. 1993€). consi- dero en esos casos a la variedad alfa como la va- riedad tipo de cada especie. En los casos en que esle autor no designara una variedad alfa pero des- eribiera variedades con letras subsecuentes (beta. gama, etc.). considero a la descripta en la descrip- ción general de la especie como la variedad tipo. Otro problema que surgió en cuanto a la tipifi- cación se relaciona con los ejemplares de Sellow. Muchas de las especies descriptas por Nees (1847a. 1847b) que se tratan en este trabajo fueron descriptas sobre la base de material de Sellow del sur de Brasil que estaba depositado en el herbario de Berlín que fue destruido durante la segunda guerra mundial. Existen. duplicados de varios de estos ejemplares en Kew, los que se reconocen por los datos “Herb. Reg. Berolinense. Brasilia. Sellow legit." que figuran en las etiquetas. Como en gen- eral estos son los tinicos datos que figuran en sus etiquetas (falta información sobre la localidad y nú- mero de colección). no se puede saber con segu- ridad si realmente son duplicados de los holotipos que estaban en Berlín que utilizó Nees para sus descripciones. Pero cuando en estos ejemplares fi- gura el nombre de la especie manuserito por Nees. considero que muy posiblemente sean duplicados del este trabajo los cito como "probables? mismo material que Nees vio en Berlín. y en isotipos о isosintipos. El estudio de la distribución geográfica у de los requerimientos ecológicos de las especies se hizo sobre la base de datos del material de herbario complementados con observaciones a campo de la тауогѓа de las especies tratadas. La cita de una especie para otros países siempre implica que he visto material de esos países. Los caracteres morfológicos se analizaron en for- ma comparada en todas las especies presentes en el área. agrupándose las mismas mediante búsque- da de correlaciones que considero señalan posibles afinidades filogenéticas. Muchas de las especies de área en estudio eran muy poco conocidas y no ha- bían sido tenidas en cuenta al elaborar el sistema de clasificación vigente (Graham. 1988). MORFOLOGÍA La diversidad morfológica del género ha sido tra- tada por Graham (1988). Aquí se discuten princi- palmente los caracteres con valor taxonómico para diferenciar las especies presentes en Argentina y Paraguay. y se realizan observaciones en espectal sobre las especies que no fueron incluidas en el trabajo de Graham (1988) para ubicarlas dentro del esquema de clasificación de esta autora (Tabla 1). Irquitectura de la inflorescencia. Los tipos de inflorescencia de las Acantáceas en general, inclu- vendo varias especies de Justicia. han sido tratados en detalle por Sell (1969, 1976). Graham (1988) utiliza un sistema más simple de clasificación de inflorescencias. que es el que se sigue en este tra- — bajo. Las especies de Sudamérica austral, coinci- diendo en general con las del Nuevo Mundo (Gra- ham. 1988) poseen (1) inflorescencias compuestas (espigas de espigas, o panojas de espigas) (dibo- trioides o tribotrioides según Sell. 1969). como en J. corumbensis, J. jujuyensts, J. oblonga. J. oranen- sis y J. saltensis, o (2) inflorescencias simples (es- pigas simples) (monobotrioides segtin Sell. 1969). como en J. axillaris, J. chacoënsis, J. cuspidulata. J. gilliesii, J. hunzikeri, J. lilloana. J. phyllocalyx y algunas formas de J. pectoralis y J. saltensis. o (3) flores solitarias, como en J. riojana y J. twee- diana y en algunas formas de J. axillaris. Las es- pigas de Justicia muchas veces son secundifloras. como en J. glaziovii, J. goudotii, J. laevilinguis y J. saltensis. A veces en una especie se combinan inflorescencia simples y compuestas, como las es- pigas y racimos de espigas que se encuentran en J. saltensis y J. xylosteides. Las flores solitarias y las inflorescencias simples probablemente deriven de inflorescencias compuestas (Sell, 1976). Además de este tipo de condensación, en la evolución del género también pudo haber habido elaboración de inflorescencias simples a compuestas por combi- nación de espigas axilares con espigas terminales (Graham. 1988). Morfología de la corola. Las corolas de Justicia son marcadamente cigomorfas. Típicamente están compuestas por un tubo basal cilíndrico, una gar- каша más o menos tubulosa o campanulada api- calmente dividida en dos labios, un labio superior angosto, erecto. más o menos convexo y bidentado, v un labio inferior ancho. más o menos patente y trilobado. En las especies de Sudamérica austral Annals of the Missouri Botanical Garden Table 1. Especies de Justicia presentes en Sudamérica austral, sección a la que pertenecen, probable tipo de polinización, clase de ornamentación de la superficie seminal Graham, 1988: A, tuberculada, verrugosa o pusticulada; — В, pilosa o papiloso-pilosa; С, lisa a muy levemente rugosa), y provincia fitogeográfica en la que se encuentran. (Con asterisco* las conoce su semilla madura). especies que no habian sido classificadas por Graham, 1988, con signo de pregunta ? las que no se Especie Sección Polinización Semilla Fitogeogr. l. Justicia aequilabris Orthota Ornitofilia A Cerrado 2. Justicia axillaris* Dianthera. Ойна ct. Saglorithys Melitofilia A Paranaense 3. Justicia baenitzii Incierta Ornitofilia ? Yungas 4. Justicia brasiliana Plagiacanthus Ornitofilia C Paranaense 5. Justicia carnea Cyrthanthera Ornitofilia ? Paranaense 6. Justicia chacoénsis* Orthotactus Ornitofilia A Chaquefia 7. Justicia comata Incierta Melitofilia A Chaquefia 8. Justicia corumbensis* — Melitofilia C Chaquefia 9, Justicia cuspidulata* anther ra Saglorithys Melitofilia Cerrado 10. Justicia dumetorum сан Ornitofilia C Chaquefia 11. Justicia floribunda Plagiac — Ornitofilia C Paranaense 12. Justicia gilliesii* Sarotheca Melitofilia A Chaqueña 13. Justicia glaziovii Sarotheca Melitofilia A Cerrado 14. Justicia glutinosa Sarotheca Melitofilia A Chaquefia 15. Justicia goudotii Chaetothylax Lepidopterofilia B Chaqueña 16. Justicia hassleri Leucolomc Lepidopterofilia B Paranaense 17. Justicia hunzikeri* Orthotatus Melitofilia A Chaquefia 18. Justicia jujuyensis* Plagiacanthus Melitofilia C Yungas 19. Justicia kuntzei Sarotheca Melitofilia / Yungas 20. Justicia laevilinguis Dianthera subsect. Dianthera Melitofilia C Paranaense 21. Justicia lilloana* Incierta Melitofilia С Chaquefia 22. Justicia lilloi Simonisia Melitofilia C Chaquefia 23. Justicia lythroides Chae — Melitofilia B Paranaense 24. Justicia mandoni Simo Ornitofilia A Yung 25. Justicia oblonga* — Melitofilia B Paranaense 26. Justicia oranensis* Sarotheca Ornitofilia A Yungas 27. Justicia pectoralis Sarotheca Melitofilia A Chaquefia 28. Justicia phyllocalyx Simonisia Melitofilia ? Cerrado 29, Justicia polygaloides Dianthera subsect. Saglorith ys Melitofilia A Paranaense 30. Justicia ramulosa Chaetothylax rnitofilia B Paranaense 31. Justicia deg a* Plagiacanthus Melitofilia C Chaquefia 32. Justicia rusbyi Simonisia Melitofilia C Cerr. y Par. 33. Justicia saltensis* Sarotheca Melitofilia B Chaquefia ЗА. Justicia squarrosa Simonisia Lepidopterofilia C Chaquefia 35. Justicia tocantina Chaetothylax Ornitofilia B Chaqueña 36. Justicia tweediana е, Melitofilia C Chaquefia 37. Justicia xylosteoides* Orthot Ornitofilia A Chaquefia 38. Justicia yhuensis a — Strobiloglossa Melitofilia B Cerrado esta estructura básica está ampliamente diversifi- cada, probablemente en respuesta a la presión se- lectiva de diferentes tipos de polinizadores durante la evolución del género. Esta gran diversidad mor- fológica de las corolas, similar a la descripta para las especies de Ruellia de Sudamérica austral (Ez- curra, 1993c), comprende principalmente tres tipos (Fig. 1, El primer tipo se caracteriza por las corolas de Tabla color blanco o lila con el tubo, garganta y labios proporcionalmente cortos y anchos, generalmente con manchas y marcas venosas en la fauce, pro- bablemente polinizadas por insectos del grupo de las abejas (flores melitófilas, Fig. 1A, B). Este es el grupo más numeroso en Sudamérica austral y es muy variable en tamaño, ya que existen especies con flores de este tipo de unos pocos mm de lon- gitud (por ej., Justicia comata, J. glaziovii, J. poly- galoides) hasta de varios centímetros de longitud (por ej., Justicia lilloi, J. rusbyi). El segundo tipo se caracteriza por las corolas de color rosa, rojo o morado, grandes, con el tubo, gar- Volume 89, Number 2 229 Ezcurra 2002 Justicia en Sudamérica Austral Figura 1 lo role as ornit C. D, J. squarrosa. — tweediana. jujuyensts. J. le ма rófilas: —E. J. жиш ой. —F ganta y labios proporcionalmente angostos y largos, aparentemente adaptadas a la polinización por pi- caflores (flores ornitófilas, IC. D). En general este tipo de corola mide de 2.5-5 ст de longitud. Fig. y también está muy representado en la región (por | J. brasiliana, J. dumetorum., j-. Justicia baenitzii, xylos- J. floribunda. J. teoides), lo que se relaciona con la importancia de mandoni, J. ramulosa. J. la polinización por picaflores en América del Sur tropical y subtropical (Gentry. 1982) Un tercer tipo se caracteriza por las corolas de color rosado, lilacino o blanco, con el tubo y gar- ganta Muy estrechos у el labio anterior extendido. aparentemente adaptadas a polinización por mari- posas (flores lepidopterófilas, Fig. IE. К) (por ej., Justicia goudotii y J. squarrosa de flores lilacinas, y J. hassleri de flores blancas). Muchas veces los distintos tipos de flores están representados dentro de una misma sección de Jus- ticia. Esto sugiere que, al igual que en Ruellia (Ez- curra, 19936), la morfología de la corola de Justicia pr "A Tipos de corolas de las especies Na Justicia presentes en Sudamérica austral. J. | C. B E y Bn F ( orolas me а Шах. . B, іаѕ. — . mandoni. — 1). as oranensis. E. "ui ds (Esque Maro RA à partir de ilustraciones de Ezcurra. 19932. ha resultado un carácter evolutivamente muy plás- tico ante la presión selectiva de diferentes tipos de flor grupo. polinizadores. De esta manera la forma de la parece haberse diversificado mucho en el incluso dentro de diferentes linajes, diferencián- dose entre especies muy relacionadas en otros as- pectos (como por ejemplo la ornitófila J. ramulosa y la psicofila J. goudotii de Chaetothylax), | general sea un carácter poco a sect. lo que hace que et importante en la delimitación de grupos de espe- cies relacionadas. Morfología del polen. La llamativa diversidad del polen de las Acantáceas en general y de Jus- licia en particular ha sido descripto por Lindau (1893, 1895), Raj (1961, 1973). Petriella (1968 Graham (1988), у Daniel (1998). Justicia se teriza típicamente por el polen 2(—1) porado o 2(4) Carac- colporado, prolado, dorsiventralmente aplanado cuando biaperturado, con el mesocolpio reticula- damente esculturado, y la apertura con 1-3 hileras 230 Annals of the Missouri Botanical Garden de ínsulas o esculturas sexinosas a cada lado del poro. Sin embargo, existen muchas variaciones de 1988). Parte de la diversidad morfológica del polen de las este esquema dentro del género (Graham, especies de Justicia del sur de Sudamérica se de- scribe en Petriella (1968) у Wasshausen y Ezcurra (1997). damente del resto del género por presentar toda la Justicia phyllocalyx se diferencia marca- superficie cubierta de ínsulas ordenadas en líneas sub-paralelas al eje longitudinal (Wasshausen « Ezcurra, 1997). se asemeje a especies del género Potkilacanthus, La cobertura de fnsulas hace que un género muy afín a Justicia cuyas estrechas re- laciones deberían estudiarse (Daniel, 1998). Forma y superficie de la semilla. Como fue su- gerido por Ariza-Espinar (1971), ciertos caracteres de la morfología de las semillas han resultado muy útiles para el reconocimiento de taxones infrage- nérica en Justicia, en especial el grado de compre- sión, el margen, y la superficie (Graham, 1988; Lester & Ezcurra, 1991). Recientemente se ha pu- blicado un estudio con microscopio electrónico de barrido de semillas de doce especies de Argentina (Peixoto, 1998). En el presente trabajo se presentan microfotografías con MEB de otras nueve especies del sur de Sudamérica (Figs. 2-4) que no fueron tratadas por Peixoto (1998), y se dan las caracte- rísticas macromorfológicas de las semillas de todas as especies en las descripciones generales. Estas características han sido utilizadas para clasificar o confirmar la clasificación infragenérica previa de 2 as especies aquf tratadas (Tabla 1). La morfologia de las semillas coincide con la descripta por Gra- ham (1988) para las secciones en las que se ubican las especies. Las especies de Sudamérica austral poseen en su mayoría semillas comprimidas, más o menos bi- convexas y lenticulares, como J. cuspidulata (Fig. 2C), J. hassleri (Vig. 2E). J. lythroides (Fig. ЗС). J. pectoralis (Vig. 3E). J. saltensis (Fig. АС) v J. xyloes- toides (Fig. 4E). Algunas semillas comprimidas son plano-convexas y tienen un reborde grueso en el Fig. A). J. brasiliana, y J. jujuyensis de la sección Pla- — margen del lado plano, como J. corumbensis giacanthus. Una minoría de las especies de esta región presentan semillas esféricas, como J. lilloi (Fig. ЗА), J. rusbyi (Fig. ФА), J. phyllocalyx y J. squarrosa, pertenecientes a la sección Simonisia. La superficie en general varía entre las tres cla- ses principales descriptas por Graham (1988) para el género: (A) rugulosa, pusticulada, verrugosa о tuberculada, (B) pubescente (pilosa o papiloso-pi- losa), y (C) más o menos lisa (Tabla 1). Ejemplos de superficie verrugosa a tuberculada (A) se en- cuentran en J. cuspidulata (Fig. 2D) y J. axillaris de la sect. Dianthera subsect. Saglorithis, J. ob pa (Fig. 4F) de la sect. Orthotactus, y J. comata de posición incierta. Ejemplos de superficie pubescente (B) con pelos cortos aparecen en J. hassleri (Fig. 2F) de la sect. Leucoloma, con pelos más o menos gloquidiados en J. lythroides (Fig. 3D). J. oblonga y demás especies de la sect. Chae- tothylax, y con pelos levemente capitados en J. pec- toralis (Fig. 3F). En J. saltensis (Fig. 4D) de la sect. Sarotheca los pelos son capitados y gloquidiados. Superficies lisas (C) aparecen en J. laevilinguis de la sect. Dianthera subsect. Dianthera y J. lilloana de posición incierta, más o menos lisa en las es- pecies de la sect. Plagiacanthus como J. corum- bensis (Fig. 2B). v lisa y lustrosa en las especies de la sect. Simonisia, como J. lilloi (Fig. 3B) v J. rusb- 1 (Fig. 4B). Cromosomas. Se han estudiado los cromosomas de varias especies de Justicia (Grant, 1955; Pio- vano & Bernardello, 1991; Daniel et al., 1984, 1990; Daniel & Chuang, 1993; Daniel, 2000). El número básico del género parece ser x = 7 y el número más común n = 14. Sin embargo, no está claro si la gran diversidad de los números (л = 7, 9-18, 20, 22-25, 27-29, 31 y 34) encontrados en Justicia s.l. sugiere una evolución diversificadora o un origen polifilético del género (Daniel et al., 1984). Aunque se han encontrado algunas coinci- dencias preliminares entre la clasificación infra- genérica en secciones de Graham (1988) y los nú- meros cromosómicos de las especies, está claro que Justicia necesita todavía más estudio a nivel cro- mosómico (Daniel, 2000). Los pocos recuentos pu- blicados de especies de Argentina y Paraguay han dado n = 14 o 2n = 28 (J. brasiliana, J. gilliesii, J. oranensis, J. squarrosa y J. tweediana) excepto en J. xylosteoides, 2n = 32 (Piovano & Bernardello, 199] Distribución geográfica. Justicia es un género muy diversificado en sus requerimientos ecológi- cos. En Paraguay está representado en todos los ambientes de las provincias fitogeográficas en que se divide el país: del Cerrado, Paranaense y Cha- quefia, у en Argentina también se encuentra en las provincias de las Yungas, de la Prepuna, del Monte, del Espinal y Pampeana (Cabrera & Willink, 1980) (Tabla 1) Especies xerófilas típicas de la provincia Cha- queña son, por ej., Justicia squarrosa y J. xylos- teoides, que aparecen exclusivamente en ambientes semiáridos de la porción occidental de la región, mientras que especies hidrófilas típicas de la pro- J. brasiliana y J. vincia Paranaense son, por ej.. Volume 89, Number 2 Ezcurra 231 2002 Justicia en Sudamérica Austral Figura 2. Tint de semillas de las especies de Justicia presentes en Sudamérica austral. A, В. J. corumbensis (Bernardi 20142). — specto general de la cara cóncava con reborde. —B. Detalle de la superficie iu C, D; cuspidulata (Se hinini ) Bordas 20534). —C. Aspecto ge * өр —D. Detalle de la superficie verrugosa. E. F, J. ice ri (Zardini y Velázquez 9943). —E. Aspecto general. —F. Detalle de la superficie cortamente verrugoso- — carnea, que solamente se encuentran en bosques arenosos sometidos a incendios periódicos (por ej.. húmedos del noreste de Argentina y Paraguay ori- J. axillaris), o de bosques en galería de la región ental. Existe un pequeño grupo de especies de Jus- chaqueña (J. corumbensis). ticia endémicas de Sudamérica austral (J. gilliesii, No existe relación entre la taxonomía infrage- J. hunzikert, J. riojana. J. lilloana). aunque la ma- nérica y la distribución geográfica. ya que dentro voría en general tienen una distribución que excede de una misma sección puede haber especies de to- los límites de la región. Muchas además son carac- dos los ambientes. Aparentemente el área extensa lerísticas de condiciones locales y edáficas parti- y la diversidad de ambientes de la región subtro- culares, como las especies de ambientes anegados — pical y tropical de América del Sur habrían per- (por ej. J. laevilinguis), o de campos abiertos y mitido la evolución y diversificación de varios li- 232 Annals of the Missouri Botanical Garden Figura 3. Tipos de semillas de las especies de Justicia presentes en Sudamérica austral. A, B. J. lilloi (Mereles 2718). —A. Aspecto general de la forma esférica. —B. Detalle de la superficie lisa. C, D, J. lythroides (Degen 1522). —C. Aspecto general. —D. Detalle de la superficie pilosa, con pelos gloquidiados. E, F, J. pectoralis (Schinini y Palacios 25788). —E. Aspecto general. —F. Detalle de la superficie pilosa, con pelos levemente capitados. najes distintos de Justicia, muchos de ellos representados por las numerosas especies presentes en la gran heterogeneidad del territorio. TAXONOMÍA El análisis morfológico permitió confirmar la po- sición taxonómica de varias de las especies pre- sentes en la región, y proponer la ubicación de otras 14 (marcadas con asterisco en la tabla) que no habían sido tenidas en cuenta dentro del esque- ma clasificatorio propuesto por Graham (1988) (Ta- bla 1 rece ser robusto, ya que permite acomodar estas — . El esquema clasificatorio de esta autora pa- especies. En Sudamérica austral están representadas ocho de las nueve secciones del género que existen en el Nuevo Mundo: Justicia L. sect. Chaetothylax (Nees) V. A. W. Graham, sect. Cyrtanthera (Nees) V. A. W. Graham, sect. Dianthera (L.) V. A. W. Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral дм 4 Wes ДТ ~ Figura 4. Tipos de semillas de las es 2 | | - Aspecto general de la forma esférica. —D. Detalle a superficie | F. Aspecto general, === de | Graham subsect. Dianthera. sect. Dianthera (L.) V. \. W. Graham subsect. Saglorithys (Rizzini) V. A. W. Graham, sect. Dianthera (L.) V. A. W. Graham subsect. Strobiloglossa (Oersted) V. A. W. Graham. sect. Leucoloma V. A. W. Graham, sect. Orthotactus W. Graham, sect. Plagiacanthus (Nees) . Graham, sect. Sarotheca (Nees) Bentham. sect. Simonisia (Nees) V. А. W. Graham. El hecho de que la región subtropical de América del Sur sea un centro de diversidad tan importante para el есте de Justicia presentes en Sudan B.D osa Detalle de li SA ee: 0 > A , 4 y 2 Ve Wey „ЕТ э? T 29, 11 atl AM al M. NJ єс. а 2] чў — 51 ( vee rusbyi (Arbo rica austral. A, В. J. isa. С, D, J. saltensis (Duré 428). etalle de la superficie li ч vylosteoides (Krapovickas э sd . соп pelos capitados. E. F, (superficie verrugosa. género sugiere una larga historia evolutiva en la región y un posible origen sudamericano para el mismo. a siguiente. sinonimia incluye solamente los nombres genéricos que han sido utilizados para es- pecies presentes en Sudamérica austral. Para una sinonimia más completa, véase Graham (1988) у Daniel (1995). En esta lista se da el nombre de la especie tipo de cada nombre genérico y su sinó- nimo en Justicia. Los sinónimos homotípicos se Annals of the Missouri Botanical Garden marcan con el símbolo =, y los basados en tipos diferentes con el símbolo =. En Apéndice 3 se presenta un índice de nombres científicos y de nombres vulgares. Justicia L., Sp. Pl. I: 1753, nom. cons. prop. TIPO: (lectotipo, designado por Hitchcock en Hitchcock & Green, 1930) Justicia hyssopifo- ia L. ‚ 1753. oe L., Sp. РІ. TIPO: Dianthera ame- 1 L. " Justicia americana (1. a e — P Miller, Gard. Dict. Abr. Ed. + 754. TIPO: Adhatoda zeylanesium Miller |= A adhatoda Asiat. Rar. 3: 77, 112. ~ бии — Nees, in Wall., РІ. 1832. TIPO: mohos opia beyrichii Nees. Beloperone Nees, in Wall., Pl. Asiat. Rar. 3: 76. 1832. ТРО: (le ectotipo, designado por Graham, 1988) Be- loperone amherstiae Nees |= Justicia brasiliana Roth]. Leptostachya Nees, in Wall., Pl. 1832. TIPO: (lee totipo, 1988) — virgate Asiat. Rar. 3: 76, 105. designado por Graham, ı [= Justicia virgata Nees) T. s.]. dis a s i i indi. Nat. Syst. Ed. 2, . 1836. TIPO: (lec — de signado por л 196€ ) Акуніна origanoic p ^s Nees [= Justicia a (Nees) Griseb., non Vahl]. Ethesia Raf., Fl. Tellur. 4: 63. 1838, non Raf. 1837. TIPO: thesia carnea (Lindl.) Raf. [= Justicia carnea b ana Raf., Fl. Tellur. 4: 61. 1838. TIPO: i ce pectoralis (Jacq.) Raf. [= Justicia pec toralis Ra MEME Nees, in Martius ras. 9: 847. TIPO: (lectotipo, aquí Жашай) Orthotactus veno- SUS Nees [= Justicia aequilabris (Nees) — Heinzelia Nees, in Martius, Bras. 9: 153. . TIPO (lectotipo, designado por Graham, ped Te lia lythroides Nees |= Justicia lythroides (Nees) Gra- ham]. rs Nees, in Martius, Fl. Bras. 9: 153. "i ): Chaetoth ylax toc antinus Nees [= Justicia to- A.W ~ - Ww Ww -— a (Nees) V um C ae Ne es, in Martius, FL js 9: 99, 1847. TIPO: (lectotipo, designado por Leonard, 1951-1958) — magnifica Nees |= Justicia carnea ndl. |. Nees, in Martius, Fl. Bras. 9: 107. 1847. Leonard, Justicia serico- Seric vara is TIP — — ^ MT л (lectotipo, designado por -= — E o his V. А. W. Graham]. —— Nees, in к tius, Fl. Bras. 9: 113. 1847. TIPO: (lectotipo, — por Graham, 1988) Sarotheca elegans Nees |= Justicia glutinosa (Bremek.) V. A. .G сапап], Seric — rigida Nees |= оаа — . in Moric., Pl. Nouv. Amér.: 156. 1847. ТІР — lepida Nees |= Justicia lepida ( е; Wassh. + Fellium О. Kuntz „ Rev. Gen. 2: 486. 1891, non S. Kurz 'inónimo nomenclatural de Justicia L. 1753 * isc totipificación). TIPO: Justicia Farc gen L. Chaetochlamys Lindau, Bull. Herb. Boiss. 3: 490. 1895. PIPO: (lectotipo, designado por Graham, 1988) Che- tochlamys mac rosiphon Lindau [= Justicia macrosi- phon (L indau) V. Graham |. Cyphisia Rizzini, Rev, Bios; Biol. 6: 521. 1946. TIPO: ‘YP hisia venusta. Rizzini [= Justicia venusta (Rizzi- ni) V. A. W. Graham]. Аал ше — . Kon. Ned. Akad. We- tensch., Afd. e de Sect. 45: 55. 1948. TIPO: poen ae comata (L.) Bremek. [= Justicia comata L..]. ME a pon Arch. Jard. Bot. Rio de Janeiro 8: . 1948. TIPO: — paniculatum Rizzini islicia л. Lindat Acelica Rizzini Arch. Jard. Bot. Ro de Janeiro 9: 55. 1949. TIPO: Acelica cydoniifolia (Nees) Rizzini sada en Adhatoda cydoniifolia Nees |= Justici cydoniifolia (Nees) L — dec Vie Arch. Jard. tio de Janeiro 9: 56. . TIPO: Pupilla isis (Nees) Rizzini ba- rid en Leptostachya heterophylla Nees |= Justicia (Nees) Lindau |. stris Rizzini, Dusenia 3: 189. 1952. TIPO: Thalestris pile er iie Rizzini [= Justicia comata L.]. Thales Hierbas perennes, sufrátices o arbustos erectos, procumbentes o ascendentes, glabros o pubescen- les, con cistolitos. Hojas opuestas, sésiles o pecio- ladas, generalmente enteras, raramente crenadas o crenuladas. Inflorescencia terminal o axilar, las flo- res sésiles o cortamente pediceladas, con dos brac- téolas, solitarias, fasciculadas o cimosas, en las axi- las de brácteas, éstas frecuentemente dispuestas en espigas o lirsos, que a su vez a veces se agrupan en racimos o panojas; brácteas y bractéolas varia- das. Cáliz profundamente 4- o 5-partido, los 4 o 5 segmentos generalmente angostos y subiguales. Co- rola marcadamente cigomorfa, roja, morada, lilaci- na o blanca, raras veces amarilla o naranja, el tubo más o menos angosto a campanulado, derecho o incurvado, con surco estilar en la parte posterior, a veces ampliado en la garganta y con marcas colo- readas en la fauce, del lado anterior, el limbo bi- labiado, el labio posterior interno en la preflora- ción, generalmente angosto, erecto o incurvado, a veces cóncavo, con el ápice entero, bífido o leve- mente bilobado, y el labio anterior más ancho, más o menos patente o recurvado, profundamente a le- vemente trilobado. Estambres 2, los filamentos ge- neralmente insertos abajo. cerca, o por encima de la mitad del tubo, filiformes o algo dilatados en la base; anteras exertas, bitecas, las tecas general- mente oblongas, a veces levemente curvadas o re- niformes, discretas, superpueslas a cercanamente paralelas en el conectivo, a veces la inferior redu- cida, múticas, o una o ambas agudas, y apendicu- ladas o espolonadas en la base; estaminodios au- sentes. Polen prolado, 2- o 3- o 4-porado o colporado, diversamente ornamentado, con la exina Disco nectarífero anular, reticulada. cupulado o ciatiforme, entero o sinuado-lobado. Estilo filifor- Volume 89, Number 2 Ezcurra 235 2002 Justicia en Sudamérica Austral me, exerto, diminutamente bilobado en el ápice. — damérica, donde probablemente habiten cerca « " Ovulos 2 en cada lóculo. Cápsula de contorno 300 especies. Algunas de Sudamérica austral se oblongo. elíptico u obovado, con la porción basal cultivan como ornamentales por sus flores lamati- sólida y lateralmente comprimida, y la porción su- vas, como Justicia carnea (= Jacobinia pohliana. "vara de la justicia”) y Justicia floribunda (= Ja- cobinia pauciflora o Justicia rizzinit, “banderita es- pañola”) (Parodi & Dimitri, 1980). aunque existen perior cilíndrica. ovoide o subesférica, con los sep- tos con retináculos. Semillas 4 o menos por aborto, generalmente comprimidas o lenticulares, a veces subglobosas. lisas o variadamente ornamentadas. . . : = чуе аз, Sas РЕА л! rnamentadas varias otras nativas de Sudamérica austral con valor rugosas, verrucosds o pilosas: retináculos CUPVOS V E | z d | ке 5 *.— ornamental que merecerían cultivarse, como J. ora- agudos o truncados. 2. "e : nensis y J. brasiliana. Algunas también tienen im- El genero Justicia comprende más de 600 es- portane la económica como forraje ras de emergen- pecies ampliamente distribuidas en ambos hemis- cia, en épocas de escasez en gramíneas, como ferios. especialmente en los trópicos, extendiéndose Justicia tweediana (= J. campestris, “alfalfilla”) у también a las regiones templadas del mundo. Pre- J. gilliesii (= J. echegarayi, “albahaca de vaca”) Su- (Burkart, 1943). senta un centro de diversidad importante e CLAVE ARTIFICIAL PARA LAS ESPECIES DE JUSTICIA DE SUDAMÉRICA AUSTRAL Flores de más de 2.5 em long. xS E 2a. s lilacinas, azules d blancas. 3a. Hojas axilantes de las flores o inflorescencias truncadas o cordadas J. lilloi 3b. Hojas axilantes de las flores o inflorescencias obtusas a redondeadas J. rusbyi 2b. Flores rosadas. rojas o mora a. Flores en de isos tirsos terminales: corolas de 5-6 cm long J. carnea tb. Flores en inflorescencias variadas. nunca en tirsos densos: уи menores. За. Espigas con brácteas ovadas, elípticas u obovadas de más de З mm lat. ба. Brácteas de 7-9 mm lat: cáliz 5-partido — — J. aequilabris Ob. Brácteas de 3—5 mm lat.: cáliz 4 o 5-partido. 7a. Cáliz 4- — flores rojo-moradas J. ramulosa 7b. Cáliz 5-partido, flores rojo-anaranjadas J. oranensis 5b. ES con brácteas más angostas. Flores con corola angosta, de 2.5-3 em long. por 2-4 mm diam. Oa. Espigas densas; cáliz d-partido J. tocantina Ob. Espigas laxas; cáliz 5-partido J. baenitzii 8b. Flores con corola más larga v ancha. Cáliz siempre 5-partido. 10a. Espigas secundifloras. lojas subsésiles, de 0.7-1.5 em lat; arbustos xerófilos. Flores con labio posterior de 1—1.5 em long. — J. x Mosteoides 11b. s pecioladas de 1.2-2.5 em lat.: balas o subarbustos higrófilos. Flores con labio posterior de 1.52 cm long. m J. mandoni 10b. — bilaterales o inflorescencias de otro tipo. a. Flores solitarias o inflorescencias de espigas fasciculadas mucho más cortas que las hojas axilantes. 13a. Inflorescencias axilares multifloras de flores rojas: brácteas espatu- ladas u oblance lola a lineales J. brasiliana 13b. Espigas axilares paucifloras o utellus de flores con el ápice am rillo: brácteas triangular-lanceoladas J foribunda 12b. Inflorese 'encias en espigas o racimos de espigas, más eh argas que las hojas la. Brácteas angostas, de са. | mm lat, que ennegrecen al secarse, Flo- res rojas J. chacoénsis П. Brácteas más anchas. de más de 2 mm lat.. verdes. con el margen blanquecino y hialino. Flores moradas а А dumetorum Ib. Flores de menos de 2.5 em long. 15a. Corolas de 3-6 mm long. J. comata 15b. Corolas de más de 6 mm long. l6a. Cáliz 5-partic Іта. Inflorescencias generalmente terminales (a veces axilares también presentes). é ha 18a. Espigas capituliformes. condensadas en los extremos de las ramas. Brácteas y brac- téolas largamente ciliadas. Semillas subesféricas —— J. squarrosa 236 Annals of the Missouri Botanical Garden 18b. Espigas alargadas. Brácteas y bractéolas no largamente ciliadas. Semillas aplanadas J. laevilinguis 17b. Inflorescencias generalmente axilares. 19a. Hojas pecioladas, de 2.5-5(-7) em lat.; plantas higrófilas. 20a. Inflorescencias pedunculadas mas largas que las — 21а. Corola de menos de 1 — ип long J. glaziovii 21b. Corola de más de 1.4 cm long. 22a. Flores blancas con mancha lila en la fauce | Flores lilacinas con mancha blanca en la fauce 20b. e "las к — longitud o más cortas que las | 23a. Flo J. corumbensis J. kuntzei a. Flores de ca. 1.5 em long. con segmentos del cáliz de menos de | em long. | | ‚ Jujuyensis 23b. Flores de ca. 2 em long. con segmentos del cáliz de más de | ст long. 19b. Paje de menos de 2.5 cm lat.; J. phy ое alyx plantas xerófilas Flores solitarias en las axilas de las айз superiores 25a. Cáliz ac — con los segmentos „йн. — lanceoladas a linea- res, — menos de 0.3 em lat. N Р Е Е Е J. riojana 25b. Cáliz c Vi segmentos libres; hojas lanceoladas a ovadas, de 0.5-1.5 (2. 5) « em lat. 24b. — 's re nidis en espigas con brác ‘teas diferenciadas. 26a. nillas tuberculadas. 27a. Hojas de 0.3-1.5 cm lat., cortamente pecioladas J. hunzikeri і Hojas de 1.8-2.5 cm lat., largamente pecioladas J. gilliesii 26b. Semillas lisas | - J. lilloana lob. Cáliz 4-partido. 2 a. Flores solitarias en las axilas de I: J. tweediana Q7 as la is superiores o de brácteas similares a las hojas 9 J. axillaris 28b. — 's en inflorescencias ; multifloras con brácteas marc adame nte diferenciadas, mucho me- no 292. — ncias generalmente de espigas simples. 30a. Hojas angostas, de menos de | cm lat. 30b. Hojas de más de | em lat. 3la. Hojas con lámina ovada a orbicular, de 2.5-5.5 em long 31b. J. polygaloides cuspidulata Hojas con lámina angostamente ovada, ovado-lanc olada o lanceolada, de 1-7 cm long. 32a. Espigas de 4—6 cm long.; corolas de ca. 0.8 em long. Hojas de 1-2 em lat., generalmente pubescentes J. pectoralis 32b. Espigas de hasta 12 ст long.; corolas de 0.8-1.2 em long. Hojas de 1.5-2.5 em lat., generalmente glabras . yhuensis 29b. Inflorescencias compuestas de espigas agrupadas en racimos (raramente espigas sim- ples en las axilas de las hojas). 33a. Infloresce ncias con los ejes y las brácteas glandulosos. a. — ovada o elíptica de 7-15 « nm long Ww — m long.; espigas densas; brácteas de 3— . glutinosa 34b. Lámina ovals de 3-10 cm long.; espigas laxas; brácteas de 2-3 mm long. ч J. saltensis 33b. Inflore scene las no glandulosas. cáliz de ca. 0.6 cm long.; anteras con la teca apical perfecta y la basal ob "o ta; reducida a un leve engrosamiento sobre el filamento J. lythroides 35b. Cáliz de más de 0.6 em long.: anteras con la teca apical perfecta y la basal reducida pero no obsoleta. 36a. Brácteas, bractéolas y segmentos del cáliz albo-marginados. Corola de ca. 2.5 em long. . hassleri 36b. Васе, bractéolas у segmentos de | cáliz totalmente verdes. Comla de 1-2 em long. 37a. Corola de 1.5-2 cm., con tubo largo de 1—1.5 em long. J. goudotii 37b. Corola de ca. 1 em, con tubo corto de 0.5-0.6 cm long J. oblonga l. Justicia aequilabris (Nees) Lindau, in Engler (Nees) Nees, in DC., Prodr. 11: 358. 1847. Ja- & Prantl, Nat. Pflanzenfam. 4(3b): 350. 1895 cobinia aequilabris (Nees) Lindau, Bull. Herb. Orthotactus aequilabris Nees, in Martius, Fl. Boissier 3: 486. 1895. TIPO: Brasil, s. loc., Bras. 9: 134. 1847. Amphiscopia aequilabris Sellow 174 (holotipo, B destr., fot. F 8902) Volume 89, Number 2 2002 Ezcurra 237 Justicia en Sudamérica Austral Orthotactus strobilaceus Nees, in Martius, Fl. Bras. 9: 133. 847. Amphiscopia strobilacea (Nees) Nees, in DC.. Prodr. 1 l: 358. 1847. Justicia strobilacea (Nees) Lin- i gler И риа. Nat. РИ; 1(3b): 350. 1895. “TIP Brasil. — fl. San Fran- cisco, in sylvi des Antonio el Alegre сеа Mar- tius — м! ) «ж Blanchet 2575 (sintipo, Wi; isosintipos, BM!, кз arnolttianus E . in M: artius, Fl. Bras. 9: 132. TIPO: Brasil. — Blanchet 2575 ex herb. D si sert (isotipo. e Orthotactus venosus Nees, in Martius, Fl. 1847. si s venosa (Nee Prodr. 11: . 1847. Justicia renosa (Nees Kew Bull. ps (ay 617. 1988. TIPO: p E ranhao: Pohl 18-49 (holotipo. d. fot. iu velutina Lindau. Bull. Herb. 189 ‚ише nebutina (Lindau) Lindau ex Hassler. 6 1917 Mi. PI. Hass TIPO: Brasil. Glaziou anzenfam. Bras. 9: 133. Nees, in DC. ¢) Graham. Rio Ma- Eu 3: 487, 076 бейти, 1 ! R! ои к “а Lind Bull. Herb. Boissier 3: 1806. cla durian (Lindau) Gra- . 43 (4): € 19 TIPO: Bolivia. : 000-800 m. و‎ san. (holotipo. В F 0900!: ! de str. fot. isolipo, NY Sufrútice de 0.3—0.5 m alt., con tallos subcilin- dricos, pubérulos. Hojas con pecíolo de 0.7-1.5 em long., pubérulo, y lámina ov е a anchamente elíp- пса. de 4-9 em long. por 2.5-5 em lat.. cuneada у decurrente en la base, — u obtusa en el ápice. levemente pubérula en el haz, velutino pubérula en е | envés. cm Inflorescencias en espigas de 5—7 subsésiles y alternas en las axilas de las ho- le cm — у 0.7-0.9 em lat., con la venación long.. imbricadas. ovadas. jas superiores; brácteas 1.3-1.7 marcada. muy levemente pubérulas, ciliadas: brac- téolas lanceoladas de ca. | em long. y 2 mm lat.. elabras, ciliadas. Cáliz profundamente 5-partido. con los segmentos lanceolados, de ca. 5 mm long.. generalmente glabros. Corola roja de 3—3.5 cm long.. con tubo basal angostamente cilíndrico. de 2.5 ст long. y hasta 4 mm diám.. el labio posterior bidenticulado. de | em long. por 3.5 mm lat.. anterior trilobado, de ea. 1 em long., con los lóbulos de 6-7 mm long. Estambres exertos con las anteras debajo del ápice del labio superior, las tecas a dis- tinta altura, ambas de ca. 2 mm long.. müticas. Cápsula claviforme de ca. 1 em long., con semillas aplanadas y (según Nees, 1847a) tuberculado-ve- rrugosas de ca. 2 mm long. Ilustraciones. Nees, 1847a: tab. 21 (sub Ori- hotactus venosus). Este de Bo- livia, Brasil y noreste de Paraguay, en la región de Distribución, hábitat y fenología. Amambay. Habita en bordes de bosque y florece en primavera, de octubre a noviembre. Justicia aequilabris se caracteriza por las espigas axilares con brácteas imbricadas, ovadas, anchas y papiráceas. y las flores de corola roja generalmente de más de 3 cm long. Es una especie variable y de distribución amplia (ver tipos citados en sinoni- mia), aunque aparece raramente v está relativa- mente poco coleccionada, como lo demuestra la fal- ta de ejemplares de Bolivia y Brasil en los herbarios. Los nombres que se consideran sinóni- mos presentan ejemplares tipo cuyas pocas diferen- clas parecen formar parte de la variabilidad mor- fol6gic ‘a de esla espec ie, Material seleccionado. PARAGUAY. Amambay: In dumetis Ponta Рога. Sierra de Amambay. Rojas 10929 (G). : Zwischen Rio Apa und Rio Aquidaban. E, G, P) Concepción Centurion, Григ 1103 (BM. in Engler & 1895. Justicia axillaris (Nees) Lindau. Prantl, Nat. МЭЬ: 350. Pflanzenfam. Rhytiglossa axillaris Nees, in Martius, Fl. Bras. 9: 122. 1847. TIPO: Brasil. Rio Grande Barbara da Eneruc il- D destr.. fot. do Sul: fl. Jacuy ad S. hada. Sellow s.n. 9011). (holotipo. Rhvtiglossa campestris Nees, in Martius, Fl. Bras. 9: 118. 1847. Justicia campestris (Nees) Lindau, in Engler & Prantl. Nal Рт nfi am. 4 (3b): 350. 1895. non Justicia campestris Griseb., Goell. Abh. 10: 225. 874. TIPO: Brasil. Rio — do Sul: Porto Alle- 1. (h m B destr.. fot. F !: probables latinos. EL Hierba pequeña, rizomatosa, de hasta 15(—30) em alt., con tallos erectos o algo decumbentes. sub- cilíndricos. delicados, con dos líneas longitudinales у opuestas de pubescencia fina y curva en cada entrenudo, surgiendo de las axilas de las hojas. 3 raíces engrosadas. Hojas subsésiles a cortamente pec ¡oladas. las inferiores orbiculares a ovadas, de 2. y 1-1.5 superiores ovadas o angostamente ovadas а elípti- 5 ст long. em lat., redondeadas. las cas (raramente orbiculares), de 1-2 em long. y 0.3— em lat.. agudas. todas glabras о hirsuto-pubéru- las. cilioladas en el margen. Flores bibracteoladas sésiles en las axilas de las hojas superiores o de brácteas algo más angostas y pubescentes, poco di- lerenciadas, a veces formando espigas terminales: bractéolas lanceoladas, de 0.7 a 1.5 em long. y ca. l mm lat.. pubérulas. садах. Cáliz profundamente I-partido con los segmentos lineales, de 0.5-1 cm long., generalmente ciliados. Corola lila. pálido-li- lacina a blanquecina de 1.4-1.7 em long.. pubé- rula. Labios de aproximadamente la misma longitud que el tubo basal, el posterior brevemente biden- tado, el anterior profundamente trilobado, con los lóbulos de 5-6 mm long. y 3 mm lat.. el lóbulo central más ancho y venoso-rugoso en la fauce. Es- tambres con las tecas divergentes. y estilo filiforme 238 Annals of the Missouri Botanical Garden con estigma bilobado. Cápsula claviforme, de 7 mm long. y 2 mm diám., glabra. Semillas algo aplana- 8 ) B go а das, de 1.5-2 mm diám., amarillentas, tuberculado- verrugosas; retináculos de 2 mm longitud. Ilustraciones. | Dawson, 1979: 573. Distribución, hábitat y fenología. Sur de Brasil, Uruguay, este de Argentina y Paraguay oriental, en campos abiertos arenosos o pedregosos. Florece en primavera y verano, de octubre a marzo. Justicia axillaris y Justicia campestris parecen corresponder a formas con distinta morfologia foliar y diferente cantidad de pubescencia de una misma especie polimórfica característica de campos abier- tos, por lo que el último nombre se trata como si- nónimo del primero. Por su cáliz 4-partido, brác- teas angostas y semillas tuberculadas, se clasifica en Justicia sect. Dianthera subsect. Saglorithys (Ta- bla 1). Justicia reitzii Leonard (Tipo: Brasil. Santa 11 ao Norte de Abe- lardo Luz, campo, 500—600 m, 25 Dic. 1956, Smith y Reitz 9230, holotipo, US!) es una especie rara del Catarina: Abelardo Luz, sur de Brasil muy afín a J. axillaris, que se dife- rencia por las flores siempre solitarias en las axilas de las hojas superiores y el tamaño generalmente menor y la textura más coriácea de sus láminas foliares. A pesar de que recientemente se citó como 1999), se consi- dera que las relaciones entre ambas deberían es- sinónimo de J. axillaris (Ezcurra, tablecerse sobre la base del estudio de más mate- rial del sur de Brasil. Material seleccionado. ARGENTINA, Corrientes: Concepción, — “е 906 (CTES); Ituzaingó, Ea. El Plata, Meyer 6414 (LIL); Santo Tomé, Gdor. Virasoro, Es- tabl. Timbauva, a más o menos 9 km de 14, monte próximo al casco, Romanczuk 188 (SI); San Miguel, 21 km al S de Loreto, Schinini 8327 (CTES); Ituzaingó, Playa- dito, 20 km W de Apóstoles, Schinini 21794 (CTES): San Roque, M. P. Mansilla, Ybarrola 2866 (LIL). Entre Ríos: Concordia, Ayuí, alrededores Hotel Salto Grande, flor lila, campos pedregosos, Troncoso 1158 (SD; Federación, Salto Grande, laderas pedregosas, Troncoso 27231 (S1); La | La Paz a siones: Capital, Posadas. — me x Feliciano, Troncoso y Bac — 3337 (SD. ab: Burkart 14171 (S * rt 14280 (SI); fon: Candelaria, Ruiz = es, campi pland, campos, Jörgensen 69 AL); ШШ 4876 (LIL). РАК, stan- Elena, Pira Pyta, Schinini y Caballero Marmori 27168 (С TES). ( — Pr. Caaguazti in campo : Hassler 9271 (G). Canindeyú: in regione Yerbalium de Maracayü, Paraguaria euro austral, Hassler 5108 (BM, С). Cordillera: in fluminis Y-acá, il in campo pr. Valenzue la, sicco, 62 (BM, G, NY): ») ‹ "Plantae Pos den А ' Hassler 1907 (G 3. Justicia baenitzii (H. Winkler) C. Ezcurra, Bol. Soc. Argent. Bot. 25(3—4): 348. 1988. Be- loperone baenitzii H. Winkl., in Fedde Repert. 7: 113. 1909. TIPO: Bolivia. San Carlos bei Mapiri, in Walden, 750 m, Ago. 1907, Buchtien 1409 (holotipo, B destr., fot. F 8922!; isotipo, US!). Nov. Spec Be lope: rone — Rusby, Mem. New York Bot. Gard. 7: 307. TIPO: Bolivia. Huachi, 1800 ft., 13 Ago. D. "his 550 (holotipo, NY!; isotipo, Us). Sufrátice perenne, a veces decumbente o apo- vante, de cerca de 1 m alt., con ramas cilíndricas, levemente 4-suleadas, geniculadas por encima de los nudos, glabras o pubérulas. Hojas con pecíolo de 1-3 em long. y lámina ovada o anchamente elíp- tica de 8-20 em long. por 3-7 cm lat., general- mente glabra, acuminada en el ápice, cuneada y algo decurrente en la base, lisa a ondulado-crenada en el margen, con las venas principales pubérulas y prominentes en el envés. Flores en espigas uni- aterales laxas agrupadas en racimos axilares y ter- minales que superan ampliamente a las hojas en longitud. Raquis pubérulo; brácteas de las espigas lanceolado-lineales de 2—4 mm long.. glanduloso- pubérulas; bractéolas en las bases de las flores li- neales, de 2 mm long., pubérulas. Cáliz de 3—4 mm long. con los 5 segmentos triangulares, agudos, de 2.5—3 mm long., glanduloso-pubérulos. Corola roja de 2.5-2.9 cm long., con tubo de cerca de 1.5 cm long.. angosto, la garganta de 4 mm diám., los la- bios de 1.5 em long., el posterior levemente biden- tado, el anterior trilobado y levemente reticulado- venoso en la fauce. Estambres con filamentos de 1 v anteras con tecas desiguales, una de 1.5 cm long. mm long., la otra de más de 2 mm long., la superior mútica, la inferior apendiculada en la base. Cáp- sula inmadura claviforme, de ca. 1 em long. por 2 mm diám., pubérula, con la mitad inferior angosta y aplanada y la mitad superior inflada y apiculada. Semillas maduras desconocidas. Ilustraciones. Ezcurra, 1993a: 342. Distribución, hábitat y fenología. Perú (fide Brako & Zarucchi, 1993) y Bolivia hasta el Norte a Argentina en selvas de la provincia fitogeo- — de gráfica de las Yungas. En Argentina ha sido encon- trada únicamente en el extremo NW de la provincia de Salta, y en el departamento de Ledesma, en la provincia de Jujuy. Es frecuente en el sotobosque de la selva entre los 600 y 1500 m s. mar, y florece en invierno y primavera, de julio a octubre. Justicia baenitzii se asemeja al material tipo de Jacobinia tenuistachys Rusby (Tipo: Bolivia. Be- tween Guanai and Tipuani, 1892, Bang 1441, ho- Volume 89, Number 2 239 2002 Justicia en Sudamérica Austral lotipo, МҮ!; isotipo. E!) en la morfología de la in- simples o compuestas, de hasta 3—5 em long.. más florescencia y el tamaño de las hojas. Sin embargo el tipo de J. tenusitachys no presenta flores desa- rrolladas y además tiene hojas sésiles, por lo que no parece tratase de la misma especie. Material seleccionado. Martín, PN Calile ARGENTINA. Jujuy: San egua, Río Las Piedras, Judica y aed Santa Victoria, PN Baritú. Nacientes del de я fini а de Y Үе er rvoorst y Cuezzo 7725 C (LIL): — fon а C — 3 km from Aguas Blancas to Fin- i 150 m. — 1955 (K): Orán. Vado e Willink 133 (LH 4. Justicia brasiliana Roth. Sp. Pl. Nov: 17. 1821. Beloperone brasiliana (Roth) Bremek.. Verh. Kon. Nederl. Akad. Wetensch. Amster- dam. Afd. Naturk. sect. 2, 45(2): 52. 1948 TIPO: Orig. de Brasil. cult. en Europa (no vis- Lo). Justicia nodosa Hook.. Bot. Mag. 56: tab. 2914. 1829. Dianthera — soot) Benth.. in Bentham & дн ег, Gen. Pl. 2 (1): 1113. 1876. TIPO: Brasil. ult. en Liverpool Bot. Gard. (holotipo, K!). виле —— Nees. in Wallich, Pl. Asiat. Rar. 3: 102. 18 TIPO: "Ex horto ill. Amherstiae speci- mina. a i. in Indam orientalem translata." in herb. Wallich (fide Nees, 1847b) (no visto). Beloperone amherstiae um var. debilis Nees, Fl. Bras. 9: 139, TIPO: Brasil. S... pr. prae de Mineiros,” Martius s.n. (sin- tipo. MP: Porto Allegre, Sellow s.n. (sintipo, B destr.: isosintip in. Martius. “In svlvaticis po Be рем — Nées var. — A es. in Mar- tius. Fl. Bras. 9: 139. 18 TIPO: Brasil. Montis Buta агау. Sellow 120 е В destr.: isotipos. ‚ K). in Martius, Sellow 145 Be ums amherstiae je es fo: ciliaris Nees, Fl. Bras. E 139, TIPO: Brasil. olotipo. B dest Beloperone — Ne "es و‎ sd Nees, in Mar- tius, Fl. Bras. 9: 139. TIPO: Brasil. In the woods of Hio Grande. — 752 (holatipa, K!). Sericoeraphis — а Nees, in DC 1847. TIPO: Bri er 10lot "ipo, я fot. US Jacobinia he Rizzini, Ted 5)6): . 5. 1975. TIPO: Brasil. Rio de Janeiro: Bus p жын Riz- zini y Ini ne 21 (holotipo. RB!). 361. Tweedie s.n. Prodr. Rio Grande de Sul: Arbusto o hierba sufruticosa robusta, de 1-2 m alt.. con tallos erectos o apoyantes, cilíndricos. gla- bros. Hojas cortamente pecioladas. con láminas lanceoladas a angostamente ovadas. de (2.5—)4— 1018) em long. por (12)2-3(4) em lat.. periores menores. acuminadas en el ápice y obtusas — as su- a cuneadas en la base, coriáceas, discolores. en- teras o algo crenadas. glabras a muy levemente pu- bérulas. Flores sésiles dispuestas en espigas densas cortas que las hojas, cortamente pedunculadas у opuestas en las axilas de las hojas superiores. Brác- teas axilantes de las flores oblanceoladas a espa- tuladas o lineales, de 8-12 mm long. por 1—2(—3) mm lat.. una fértil y la otra estéril en cada nudo. glabras a levemente pubescentes; bractéolas linea- es a lanceoladas de ca. 5 mm long. por 0.5 mm lat.. semejantes a las brácteas. Cáliz profundamente >-partido, de 4-5 mm long., con los segmentos lan- ceolados. de 3—4 mm long., agudos. Corola roja de 3-4.5 cm long.. con tubo de ca. 2.5 cm long.. es- trecho. de hasta 5 mm diám. en la fauce. y labios de 1-1.5 em long., el posterior erecto, angostamen- te ovado, levemente bilobado, y el anterior patente. trilobado. con los lóbulos de 3-5 mm long. por 2— 3 mm lat., redondeados. Estambres más cortos que el labio posterior, con las tecas a distinta altura. de Estilo fili- y ovario glabro. 1-1.3 con la mitad inferior sólida y lateralmente ca. 2 mm long.. la inferior sub-oblicua. forme con estigma bilobado. Cáp- sula claviforme castaño-brillante. de mm long.. estrechada. y la superior subesférica. de 4 mm diám. Semillas aplanadas de contorno orbicular, de З mm «ат... lisas, pardo-claras, con un reborde en su cara interna: retináculos de ca. 2 mm long. Ilustraciones. Wasshausen y Smith. 1969: 96 (fig. 14A). Sur de Brasil. norte de Uruguay. noreste de Argentina y Paraguay Distribución. hábitat y fenología. oriental, desde el nivel del mar hasta los 600 m. Es una especie higrófila que habita en claros y bor- des de bosques húmedos y en selvas. frecuente- mente en lugares alterados como márgenes de ríos y bordes de picadas. Florece principalmente en ve- rano y otoño. desde noviembre a mayo. Justicia brasiliana se caracteriza por sus flores angostas de color carmín de hasta 4.5 cm long. dis- puestas en inflorescencias axilares densas, con brácteas oblanceoladas a espatuladas o lineales. Es una de las Acantáceas más frecuente en bosques húmedos de Paraguay oriental y del sur de Brasil y este de Argentina. y por sus flores llamativas. más coleccionadas en la región. A pesar de haber sido introducida en cultivo en Europa tempranamente en el siglo XIX bajo los nombres de Justicia bra- siliana y Justicia nodosa, actualmente no se registra como planta cultivada en Argentina (Parodi & Di- mitri, 1980) ARGENTINA. Corrientes: ue 'Tón de. \ы rada, Berón de Astrada, Паре, Ruiz Huidobro Grande, — — 4249 desvío Km 395.8 FCNEA. Ruiz Hui- Capital, Ruta 5, 19 mn del Triángulo, Material seleccionado, Santo Tomé, Cuay ries an "Май, dobro 4301 (LIL): Annals of the Missouri Botanical Garden Schinini y Martínez Crovetto 12771 (SI); Mburucuyá, Ea. Santa Ana, Schwarz 94 (LIL); Concepción, Río Santa Lu- cía, Schwarz 9275 (LIL); CT, Dpto. Saladas, Ayo. Ambro- sio, Schwarz 9394 (LIL). Chaco: CH, Resistencia, Mar- garita Belén, Aguilar 566 (LIL); Libertad, ruta 16, a orillas del Río Negro, — ee 32 (LIL); Libertad, Popular, Schultz 16244 . Formosa: Mojón de Krapovickas 1086 (SI); — Ruta 11 vieja, pasando Arroyo F. C i |, Guaglianone 726 (SI); Pilco- mayo, Puerto Ramos, Morel 7182 (LIL); — Ma l Guayacán, Reales 94 (LIL). Javier, costas Río Uruguay, Bertoni 2513 ; Guaraní, Gob. Morales y Fracrán, — Sáenz 29211 (S . Alem, 7 Km E de L. Alem, lot Marufiak 678 (Li 1 " Oberá, Oberá, Villa Blanquita, е нет 242 (LIL); San Martin, San Martín, De Marco 10654 C (LIL, M); 5 A, Schwarz САГУ Iguazú, Ruta Nac. 101, 5 : Colonia Fierro. Misiones: San — — — entre an Ignacio, obraje EM 5 km del límite E del P. Nac.. Zuloaga 5270 (S1). PARAGUAY. Alte > Paraná: i in regione а Alto den Fiebrig 6174 (BM, E. К, LIL, SI, ; Pr. Escuela Técnica Fore sx km 12 Puerto Stroess- ner, Bernardi 18891 (BM, G, ), NY). Amambay: Par- que Nacional Cerro Cora, al in del río Aquidabán cerca del fortín — Brunner 1534 (G. MO, US); Pedro Juan Caballero, Bernardi 19330 (С. O, NY). a: Arroyo Yakare'i, along N side from route 2, Zardini y Aguayo 10730 (MO, US): Tebicuary Mí. in nemore collino Isla Alta, Be rnardi 18720 (6. MO, NY). Caazapá: Tavai, 500 m d 3894 (CTES): R. Pirapó, N of José Bassai, West dus 1). Canindeyú: — s Palos ad septentr., . Soria . Casas y Molero 5855 (MO, NY): in e yerbalium de Maracayu. Pa- raguaria euro-austra, Hassler 5207 (BM, G., K, P, W). Cen- tral: I Assomption, dans les lieux ombragés. Balansa 2459 (K, P); fig Centralis, Hes lacus Ypacaray, Hassler 11740 (BM, Е.С. K, LIL, MO, US). Concepción: entre Paso Horqueta у unes ión, Krapovickas et al. 14227 (MO, US); Zw. Rio ja und Rio Aquidaban, Villa Sana, Fiebrig 4662 (BM, E, K, M). Cordillera: in regione lacus Ypacaray, Hassler 11105 (BM, K, MO, US); Montes, San Bernardino, Hassler 1593 (С. Р, S1); on silvis pr. Cor- dillera de Altos, Hassler 1579 (С. К, P. US); Paraguaria ce entralis, in silva San Bern: silicio: Hassler 3016 (BM, G . MO, Р, W). Guairá: Cordillera Ybytyruzú, Villa San Pedro. — 1876 (MO); Road Melgarejo—Antena, 8 km f Melgarejo. Za rdini 11349 (MO, US): Tororo, ca- mino à v rro Polilla, Degen 1032 (G); Tororo, Cerro Асат, Soria 2588, 2637 (С); Villa-Rica. dans les bois, Balansa 2459a (P). Itapúa: Encarnación, Mboycaé, Bertoni 4024 LIL). Misiones: San Ignacio, Estancia Brusquetti, Mere- les 1198 (G). аа Pilar, Curupayty, Meyer 1595( (LIL); Villa Pilar, — 7876 (LIL). e Acahay Massif, on rocks on SE peak. Zardini et al. 12782 (MO, US); Cerro Mbatoví, in forest, Zardini y Soria x (MO. US); Cerro Palacios, Soria 2242 (G); Soria 2720 (G); Escuela Agricola, Ypoa, bosque wi borde ‘a Parque Nac e Ybyci quez 12094 (M 1189 (MO). Presidente Hayes: in regione rioribus pps Pilcomayo, Rojas 664 (G): Chaco, km . Mereles 3904 (US). San Porn 0 km US); Loma . — el lago. 0); Dau. N de San E тне d Krapovickas et al. 13€ Hoby, Woolston 193 (SD, Woolston 193C — Central Paraguay, Morong 200a (ВМ. С. К. MO, 5. Justicia carnea Lindley, Bot. Reg. 17: t. 1397. 1831. Ethesia carnea (Lindl.) Raf.. Fl. Tellur. 4: 63. 1838. Jacobinia carnea (Lindl.) Nich- olson, Ill. Dict. Gard. 2: 206. 1885. TIPO: Bra- sil. Rio de Janeiro: enviado a Inglaterra e in- ilustrado en Lindley troducido en cultivo, (1831). irr id magnifica Nees, in Martius, Fl. = 9: 100, 1847. Jacobinia n s, Vilm. 1. 3) 1: 81 PME, (ed. 3) 1894. Jacobinia mag- nifica (Nees) L — in & Prantl, Nat. Pflan- zenfam. 4(3b): 351. TIPO: Brasil. B de Ja- xeiro: Tijuca, — s.n. ш. V. elo: — Nees var. minor Nees, in Martius, Fl. Bra . 1847. TIPO: Brasil. Semidorio, Rie- del s.n. — BR, fot. US); Rio de Janeiro, Corcovado, — s.n. (sintipo, LE no visto). Conant мег эр ees, in Martius, Fl. 17. Je abia — (Nees) Voss, = шо (ed. 3) 1: 810. 1894. Jacobinia ane (Nees) L — in Engle r & Prantl, Nat. Pflanzenfam. 4(3b): 3 1895. TIPO: Brasil. S. est.: pr. Mideiros el Entre 2 Morros, Pohl 196 (sintipo, W!); Minas Gerais: in silvis primaevis ad Cabo d'Agosto et Villa do Principe, Martius s.n. _(sintipo, M!); Rio Grande: Porto Alegre, Serra de 7, 252 (sintipos, B destr.); Tres Irmáos, Sellou — y; Praesidio eu Joao Baptista, d 254 (sintipo, В destr., fot. F 8915!, AURA: K!). C ушш pohliana Nees var. in Martius, 01. 1847. Jain ohio (N ees) L. lew. — ntes — l: 136. 1923. TIPO: Brasil. * Raven (sintipo, M!); Ins. Catarit HOC SMS s.n. (sintipo, LE no visto). C йыга аала Nees ls velutina Nees, in Martius, Fl. Bras. 9: 101. 18 . Jacobinia — (Nees) Vilm. Blume — (ed. 3) 810. 1894. Brasil. Rio de Janeiro: — ‘rs a fine rose — s.n. (sintipo, K!); Argentina. Buenos Aires: “one of the . . . ornaments of the gardens of s Avs flowers all the year,” Tweedie s.n. (sintipo, K!). Orthotactus ciliatus Nees, in Martius, Fl. 1847. — Ep _ciliata (Nees) Prodr. 59. 1847. TIPO: Brasil. Monte С — e hnatt s.n. (holotipo, M!). = Н. Ba "Absque d o nala Voss, TIPO: colour,” Bras. Nees, in Rio de Janeiro: in herb. 35. - A Martius Arbusto o hierba sufruticosa robusta de hasta 2 m alt., con ramas subtetrágonas, glabras o corta- mente pubescentes. Hojas con pecíolo de 3-5 cm long. y lámina ovada o elíptica de 10-25 cm long. por 5—8 cm lat., obtusa y levemente acuminada en el ápice, cuneada en la base, parcialmente decu- rrente sobre el pecíolo, de textura herbácea, áspera, con el margen irregularmente crenado, con las ve- nas principales marcadamente prominentes y no- torias en el envés, las secundarias arqueadas y pa- ralelas, la nervadura en general pubescente. Flores sésiles dispuestas en tirsos terminales densos y bracteados, ovoides, de 10-20 cm long. y 5-8 cm didm., bractéolas floribundos. Brácteas folidceas; Volume 89, Number 2 2002 Ezcurra 241 Justicia en Sudamérica Austral lanceoladas, de aprox. 15 mm long. por 2 mm lat., obtusas, glabras y ciliadas. Cáliz profundamente 5- partido, con los segmentos lanceolados, de 10 mm long. por 2 mm lat., agudos. Corola rosada o roja, de cerca de 5-6 cm long., con tubo de ca. З cm long.. ensanchado a 3—4 mm diám. en la garganta, y labios de ca. 2.5 bidentado, el anterior patente y trilobado. con los lóbulos de 3 mm long. Estambres inclusos debajo cm long., el posterior curvo y del labio posterior, con filamentos de ca. 4 cn long.. y anteras con tecas a distinta altura: la su- perior separada por el conectivo y oblicua, de ca. Fruto desconocido. Ilustraciones. Nees, 1847a: tab. 14 Cyrtanthera magnifica). Wasshausen y 1969: 90, fig. 13B (sub Jacobinia carnea). Sur de Brasil. Habita 3 mm long. ambas. (sub Smith, Distribución, hábitat y fenología. noreste de Argentina y Paraguay oriental. en bosques húmedos, en claros y bordes de pica- das. frecuentemente en terrenos bajos y cerca de cursos de agua. Florece principalmente de prima- vera a otoño, entre noviembre y mayo. Nombre vulgar. “Vara de la justicia” en Argen- tina (Parodi € Dimitri, 1980), “Plume flower” Nueva Zelandia (Webb, 1995), “Brazilian plume en Estados Unidos (Daniel, 1995). Justicia carnea es una especie del sur de Brasil en у regiones limítrofes cultivada como ornamental en lugares de clima cálido de todo el mundo por sus grandes inflorescencias de flores rojas o rosadas, que aparecen durante todo el año, hasta en am- bientes sombríos. Su cultivo puede haber extendido su distribución natural original, ya que se ha asil- vestrado en regiones donde ha sido introducida. Por ejemplo ha sido citada para Ecuador y Colombia (Leonard, 1951-1958: & Smith, 1969). y como introducida y naturalizada en Nueva Zelandia (Webb, 1995: 154) seleccionado. ARG Wasshausen Material QENTINA. Misiones: Iguazú, Salto Iguazú, ши 789 (К, Sl): General Ma- nuel Belgrano, reserva estricta de San Antonio, Múlgura 1891 (S1): San Pedro, serva Esmeralda, Deginani 1320 (SD: General Manuel DN San Antonio. 9 km de San Antonio, Deginani 2080 (S1): San Javier, Arroyo Patino, Schwarz 4176 a IL). EA UAY. Alto Parani: in re- gione fluminis Alto Paraná, Fiebrig 5730 (BN LIL, SI), Fiebrig 5826 (G, LIL, SD: procedente puerto — rloni. Alto Paraná, culta Jardín puta o, Rojas 10688 (LIL). 6. Justicia chacoénsis Wasshausen & С. Ezcu- rra, Candollea 52: 178. 1997. Beloperone ri- paria S. Moore, Trans. Linn. Soc. London, Bot. 1: 432. 1895, non Justicia riparia Kameyama, Bol. Bot. Univ. Paulo 14: 204. 1995. TIPO: Brasil. Near Corumbá. Moore 1047 (ho- lotipo, BM!: isotipo, B destr.. fot. F 89441). Sáo Arbusto o sufrútice ramoso de 0.80-1.50 m alt., con tallos glabros, ramificados, de corteza pálida. Hojas con pecíolo de 0.5-1.5 cm long. y lámina ovada a anchamente elíptica, de 6-13 cm long. y 2.56 cm lat., decurrente en la base, algo discolor, glabra en el haz, velutino-pubérula y glabrescente en el envés. Inflorescencias en espigas densas, de 3-8 cm long.. acuminada en el ápice, cuneada | en las axilas de las hojas superiores, éstas a veces reducidas y entonces las espigas reunidas en un racimo apical denso, el conjunto oscureciéndose al secarse; brácteas angostamente lanceoladas, de 1— 1.3 em long. por ca. 1 mm lat., largamente acu- minadas, pubérulas, negras al secarse; bractéolas lineales, de 3-5 mm, más angostas, pubérulas. Cá- liz profundamente 5-partido, con los segmentos lan- ceolados, de 4 mm long. por menos de | mm lat., agudos, pubérulos. Corola de ca. 4.5 cm long., roja, con tubo angostamente cilíndrico de 2-2.5 em long. y labios de 1.5-2 cm long.. el posterior angosta- mente oblongo. agudo, el anterior de 0.8 em long., profundamente trilobado. con los lóbulos angosta- mente oblongos, de 0.7-1 ст long. y З mm lat., redondeados. Estambres exertos casi hasta el ápice del labio posterior, las anteras con las tecas mar- cadamente superpuestas, de ca. 2 mm long., las inferiores separadas de las superiores ca. 2 mm, algo menores y basalmente apiculadas. Cápsulas angostamente claviformes, de 1-1.3 em long. por 2-3 mm lat., con la base sólida y lateralmente es- trechada en el tercio inferior, muy levemente pu- bérulas. Semillas aplanadas, de 2 mm diám., ve- rrucosas, oscuras; retináculos de ca. 1 mm long. Ilustraciones. Figura 5. Distribución, hábitat y fenología. Sudoeste de Brasil, norte de Paraguay y regiones limítrofes de Bolivia. Habita en bosques xerófilos abiertos, y flo- rece en primavera y otoño, de septiembre a mayo. Justicia chacoënsis se caracteriza por sus flores rojas grandes, de cerca de 5 em long., en inflores- cencias espiciformes con brácteas pequeñas y an- Por sus caracte- gostas que oscurecen al secarse. rísticas ubicó en Justicia sect. Orthotactus (Wasshausen & Ezcurra, 1997). morfológicas se Material seleccionado. PARAGUAY. Chaco: Agua Dulce, Schinini y Bordas фе (GC. МО); —— lo * renza, Pique Histórico, ca. 10 km SE de P. Lageren- ‚ Charpin y Ramella 21610 ( 3); Cerro León, Schinini y Bordas 17794 (US); Cerro León, desde lomada al S (cam- — hasta meseta бешш, Charpin y Ramella 21742 ;); Fortín Gabino, Mendoza a Lagerenza, a 12 km. desvío a TE Cué, Degen 3280 (FCQ); Mayor Pedro Lageren- sque abierto de quebracho у samohá, Schinini у Bordas 14999 (US); Palmar de las Islas a Ravelo, Cerro san Miguel, Mereles y Ramella 2795 (ЕСО, G 242 Annals of the Missouri Botanical Garden / 1 сор Figura 5. Justicia chacoénsis. —A. Rama con flores. Flor con bráctea, cáliz, corola, estambres y estilo. —C. j). Corola abierta con androceo. Fortunato 3738 (G), Billiet y Jadin 3263 (G 4. Justicia comata (L.) Lamarck, Encycl. 1: 632. 1785. Dianthera comata L., Syst. Ed. 10: 850. 1759. — hya comata (L.) Nees, in DC., Prodr. 381. 1847. Ecbolium comatum (L.) s — Can. Pl. 2: 487. 1891. Stetho- ma comata (L.) Britton, Bot. Porto Rico 6: 218. 1925. Psac adoc — comatum (L.) Bre- . Wetensch.. Afd. 1948. TIPO: Jamaica. Brow ne s.n. (lectotipo, — por Graham (1988), LINN, microficha BM!). ‚ Ned. Tweede Sect. — 55. — martiana Nees, in Martius, Fl. Bras. 9: 152 1847. TIPO: Brasil. Bahia: — — М!), Salzmann s.n. (sintipo, B destr., ; isosin- tipo, ВМ!); s. loc., Sellow s.n. тке B кн иен prob- able isosintipo, BM!); Pará, Poeppig s.n. (sintipo. B destr.). Cuba. S. loc.: Poeppig s. n. (sintipo, B destr.). Leptostachya martiana Nees var. hisp artius, d à jw . 9: 152. 1847, TII PO: Guyana. Británica. urgh s.n. nibe, K no visto). e martiana Nees var. macrophylla Nees, in la Nees, in Martius, Fl. Bras. 9: 152. 1847. TIPO: Guyana Bri- lánica. Schomburgk s.n. (holotipo, К!; isotipo, E! er nya parviflora Nees, in Martius, Fl. Bras. 9: 1 51. 18 TIPO: Brasil. Bahia: Ilheus, Махат Vied s.n. (sintipo, BR no visto), — 125 (sintipo, 3 destr., probables isosintipos, K!, E!); Soteropolin, Manche s.n. (sintipo, G no visto; isosintipo, M!); ;astelnovo, Riedel s.n. (sintipo, LE no visto). ا‎ acuminata Nees, in DC., Prodr. 11: 354. 1847. Justicia acuminata (Nees) Lindau, in Engler & ет Nat. ae fa. A(3b): 351. 1895. TIPO: Mexico. asco: marais de Teapa, Finden 1633 (lec- totipo, Meiner por - Daniel — 5), K Justicia Swartz, Prodr.: 14. 1788. ж Jamai- са. Swartz s.n. (holotipo, Bu ів про, М!). Thalestris nai Pau Dusenia 3(2): 190. 1952. TIPO: Brasil. Paraná: Parque Nacional do Iguazú, Duarte 1705 (holotipo, ier isotipo, NY!). 3 m — Hierba perenne, con largos rizomas horizontales provistos de raíces adventicias en los nudos, y ta- llos erectos de 30-80 cm alt., glabros o muy levemente pubérulos. Hojas opues- cilíndricos, frágiles, Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral tas, sésiles, subsésiles o cortamente pecioladas, an- gostamente ovadas, oblongas o lanceoladas, de 4— (0.8)1-2 ecem lat.. redondeadas. cordadas o cuneadas en la base, fre- 8(-15) cem long. por cuentemente decurrentes sobre el pecíolo, agudas. levemente discolores. glabras o pubérulas. Flores muy pequenas dispuestas en espigas secundifloras laxas у delicadas que se agrupan formando. seu- doverticilos en inflorescencias paniculiformes la- xas, terminales o axilares, con los ejes filiformes. elabros o levemente pubérulos, y con algunos pelos elandulares. Brácteas axilantes de las flores lan- ceoladas, de 0.8-1.2 mm long., generalmente gla- bras: bractéolas algo más pequeñas. Cáliz profun- damente 5-partido, con los segmentos lanceolados. 1-2.5 mm long. por menos de 0.5 mm lat., agu- dos. glabros o levemente glanduloso-pubérulos. Co- rola blanca, lilacina o violácea. de 3-6 mm long.. elabrescente. con el tubo de 2 mm long., la gar- ganta de | mm diám., los labios de cerca de 2.5 mm long., el posterior emarginado, el anterior tri- lobado. reticulado-rugoso en la fauce. con las mar- cas moradas. los lóbulos de aprox. | mm long. por | mm lat. Estambres con filamentos de ca. 2 mm long.. y anteras con las tecas situadas a distinta altura. divergentes, de ca. 0.4 mm long.. la inferior algo menor, separadas por el conectivo. Cápsula pubérula de 3-47) mm long.. con el tercio infe- rior sólido y lateralmente comprimido, la porción superior engrosada, de contorno oblongo. Semillas 1. aplanadas, orbiculares. de 1 mm diám.. dimi- nutamente verrugosas о pusticuladas, pardas: reli- náculos pequeños. Ilustraciones. Ezcurra. 1993a: 352. Distribución, hábitat y fenología. Ampliamenle distribuida en América tropical desde el sur de México hasta Bolivia. Paraguay y norte de Argen- tina. En el sur de su distribución se encuentra en- tre el nivel del mar y los 600 m de altitud. Habita en suelos anegados de charcas. bañados, esteros y pantanos. o en lugares muy húmedos y soleados de bordes de bosques y selvas. También invade terre- nos removidos húmedos y adopta características de maleza. Florece casi todo el año, principalmente en verano, de diciembre a febrero. Justicia comata se caracteriza por las flores muy pequeñas dispuestas en inflorescencias paniculifor- mes tenues y laxas con ramificaciones verticiladas, de aspecto graminoide, y no tiene afinidades claras. Como varias otras especies higrófilas de ambientes anegados, es muy variable morfológicamente y tie- ne un área muy extendida, y eso ha dado lugar a la creación de numerosos nombres que resultan en su extensa sinonimia. Material seleccionado. ARGENTINA, Chaco: Orillas del Río Guaycurú, Meyer 2094 (LIL). Corrientes: Gen- eral Paz, Lomas de Krapovickas y Cristóbal 11812 (SD: Paso de los libres. Paso Troncón. Ruiz Hui- dobro 3714 (LIL); Alvear, Ea. Santa Ana. guay, Schinini 16905 (LIL » San Martin, peyd; costa Río Uruguay, Ybarrola 1891 ( mún en monte y esteros. a nsen 2832 a Vallejos. ~ ~ ph e en chacra з. Jorge nsen 3278 (Sl): Pirané. al S del pueblo — Mor 605 (LIL): Pilcomayo, Puerto ~ Моге! 7190 ( у Jujuy: 5 h anta — Ruta Prov. 1. Laguna i . Misiones: — Cata- San Ignacio. Jardín : i esa 5967 (LIL): Santa Ana, — Rodr rigues 755 (К, LIL); San 1944 = = 3 = Ф m эм `5 z м = = >ч Q Yı © — as ее Arroyo. ltacaruaré, camino vie jo. Zuloaga turi 5591 (К. PAR jet »n Fiebrig 1423 (K, G, M). Alto Dua: in regione fluminis Alto Paraná, — 6081 (G, K, LIL, SI, — Rio Mon- day. Bertoni 1378 (LIL). — Bella Vista, Río \ра. Schinini y Bordas 20628 (C N de Coronel Oviedo, Krapor ickas r 2 (CT S); 3 km S del destacamento, Rio (СТЕ. MO). Puerto Há Enram: — cano al Rio Paraguay, Se — y Bordas 1331-4 (CTES MO. US): San Lorenzo, Campus sS ts 1208 (CTES, С); Villa a dans les mares, Balansa 2.108 (6. К. Concepción: г. ‚Ара. on Rio — Villa Sana, Fiebrig 4688 "Ki lillera: ad ripam lacus Ypacarat, Hassler — (BM. С. K, LIL. MO Ny. P); Cercanías de San Bernardino, Mereles 3270 (G). Guai- Schinini y Bordas 25242 : Común en el agua. Villa Rica. s.f.. NY. SI, US). Itapúa: Encarnación. Río Colonia Encarnación. Montes 715 Isla ( fous Pe- ene ron: 20 km ES, l Tavaf, Te E пагу. Soria 3303 Central: arroyo | cer- га: С xw Independencia, (CTES 3796 (L n MO. Paraguay. nes: Santiago, pU ча La Soledad, dersen 5994 (K, MO, NY). Presidente Hayes: in regione cursus inferioris — Pilcomayo, Rojas 362 (BM К: Ruta trans- Chaco, km 219, Schinini y Palacios 25 K). San Pedro: Colonia Primavera. Woolston 770 (K, LIL, NY, SI), Woolston 1383 (K). S. dep.: Para guaria Seple ee — 1907 (G. К): 38 Hassler 5985 (С. K, LIL, MO, NY. Р) Jörgensen . Misio- G; To DL 8. — corumbensis (Lindau) Wasshausen & Candollea 52: 178. 1997. Belope- rone m Lindau. Bull. Herb. 2 5: . 1905. TIPO: Brasil. Mato Gros- so: pr. — Abr. 1903, Malme 3029 (ho- lotipo. B destr.. fot. F 392891). . Ezcurra, Boissier SGE. 2. ws + Hierba sufruticosa de 0.6-1 m alt, con tallos subcilindricos, cuadrisulcados, pubérulos a gla- bros. a veces con dos líneas longitudinales pilosas B | en su juventud. Hojas con pecíolo de 0.5-1 cm long. y lámina ovada de 6-12(-18) cm long. por 2.5-5(-1) em lat., aguda y acuminada en el ápice. a J obtusa a redondeada en la base, decurrente sobre el pecíolo, discolor, con cistolitos poco notorios en su superficie, subcoriácea. generalmente glabra. C Flores sésiles dispuestas en espigas pedunc e Is. Annals of the Missouri Botanical Garden A Figura 6. Justicia corumbensis. —A. Rama n flores. —C. Corola abierta con androceo. =). estilo. (SI). éstas solitarias o de a pares en las axilas de las hojas superiores. Espigas frecuentemente iguales o más largas que las hojas, densas, con pedúnculos de 2—4 cm long., y raquis a veces ramificado en la base; brácteas florales espatuladas a obovadas, de ca. 5 mm long. por 1.5-2.5 mm lat., acuminadas, foliáceas, generalmente glabras, desiguales en cada nudo, una fértil y la otra menor y estéril; bractéolas oblanceoladas, de menos de 5 mm long. Cáliz 5- partido, de 3—4 mm long., con los segmentos tri- angulares, de ca. 2 mm long., agudos, glabros, ci- liolados en el margen. Corola blanca con marca lila en la fauce, de ca. 1.5 cm long., con tubo de 5-6 mm long. ensanchado en una garganta acampanada de 4 mm diám., y labios de 5—7 mm long., el pos- terior brevemente bidentado, erecto, cóncavo. de 4. {/ oD E —RB. Flor con bráctea, bractéolas, cáliz, corola, estambres y "stilo y estigma. Е. Cápsula abierta. Е. Semilla. Jörgensen 2337 ~ el anterior de 7 mm lat., ampliamente tri- mm lat., lobado, patente, con los lóbulos de ca. 2.5 em lat., redondeados. Estambres con filamentos de 4-6 mm long. y anteras con las tecas a distinta altura, la superior de 1.5 mm long., la inferior oblicua, algo mayor y con un apéndice glandular en la base. Es- tilo glabro; estigma redondeado y ovario glabro. Cápsula glabra o pubérula de ca. 1 em long., con la mitad inferior sólida y lateralmente comprimida, y la superior subesférica y 4-seminada, de 3—4 mm diám. Semillas aplanadas, de contorno orbicular, de 2-3 mm diám., lisas, pardo-anaranjadas, con un reborde en su cara interna; retináculos de cerca de 2 mm longitud. Ilustraciones. Figura 6. Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral Distribución, hábitat y fenología. Sudoeste de Brasil, oeste de Paraguay, principalmente en la región del sudeste de Bolivia. noreste de Argentina y Chaco. Habita a lo largo de los bosques ribereños de los ríos de la región chaqueña. Florece en ve- rano, de enero a abril. Justicia corumbensis pertenece a la sección Pla- giacanthus y es una especie muy afín a J. juyuyensts . Ezcurra (Wasshausen € Ezcurra, 1997) con la que suele confundirse en herbario. Ambas se ase- mejan en el tamaño, color y forma de las brácteas y flores. y en las características de los frutos. Jus- ticia corumbensis se diferencia de J. jujuyensts por las flores en espigas pedunculadas frecuentemente más largas que las hojas axilantes, mientras que J. jujuyensis posee inflorescencias axilares más cortas. Justicia jujuyensis es característica de los bosques pedemontanos de transición entre las provincias fi- togeográficas Chaqueña y de las Yungas. mientras que J. corumbensis generalmente habita en los bosques en galería de la región del Chaco septen- trional. Material se ا‎ А ARGENTINA. Chaco: Maipú. ruinas km 75, ruta go 15846 (CTES); — Schulz 3235 (CORD, t TES 1082: Pirané: R. 90, 15 km al S de Pirané. pr Ade 680 (SI); Común en A monte. km 139 del F.C.. Jörgensen 2336 (LIL, SD, 2337 (SD: Patiño, гата (СТЕУ): Matacos, Ing. Juárez. Morello s.n. herb., Se 18213 — PARAGUAY. Alto Paraguay: ad meridiem Fuerte Olimpo. guay, Bernardi 20457 (MO). Presidente — Estancia M Wap a durs del río Siete Puntas, Degen y Mereles 3250 (FCO): Km 326 de Asunción, haud proc al ab Pozo Colorado. — i 12 (G. MO, NY, US): Santa Elisa, Hassler 2754 (BM. С. K, NY j Bartolomé de las Casas, © . Justicia cuspidulata (Nees) Wasshausen. Mo- Syst. Bot. Missouri Bot. Gard. 45: 1253. 1993. Rhytiglossa cuspidulata Nees, in DC., поют. Prodr. 11: 348. 1847. TIPO: Perú. Mathews 3152bis (holotipo, К!; isotipos. sub Mathews 3152 BM!, B destr., fot. F 184.12). Hierba sufruticosa en la base, rizomatosa, erecta, de hasta 50 em alt., con tallos angostos, subcilín- dricos, poco ramificadas, hirsuto-pubérulos. Hojas de 2.5- obtusa o aguda en el subsésiles con lámina ovada a orbicular. 5.5 em long. у 1.8-3 cm lat., ápice, truncada a levemente cordada en la base. — en el haz. y pubescente a pubérula en « — 'encias en espigas terminales laxas de 21 1 em long.. con el raquis glanduloso-pubes- cente; brácteas triangulares de 34 mm long. рог 1.5 mm lat., pilosas: bractéolas de la misma lon- gitud. algo más angostas. Cáliz profundamente 4- partido, con los segmentos lineal-lanceolados, de 6-8 mm long., acuminados, pilosos. Corola lilacina y 2 mm diam. en la fauce, y labios de 5 mm long., el de ca. 8 mm long., con tubo de 4 mm long. posterior oblongo. de 2 mm lat., bidentado, y el anterior trilobado, con los lóbulos de 3 mm long. por 1.5 mm lat., redondeados. Estambres exertos hasta la mitad del labio superior, con las tecas su- perpuestas, oblicuas, marcadamente separadas por el conectivo. Cápsula de 8 mm long. por 2 mm de grosor, glabra. Semillas aplanadas, de 1 mm diám.. diminutamente verrugoso-tuberculadas o pusticu- ladas, pardo-oscuras; retináculos de menos de | mm long. Ilustraciones. Figura 7. Distribución, hábitat y fenología. Originalmente descripta para Perú, se la ha encontrado en el este de Bolivia y en el noreste de Paraguay. en la región del Amambay. Probable en Brasil. Habita en bordes de bosque. Florece en primavera y verano, de agosto a marzo, El material de Paraguay que aquí se determina con el nombre de Justicia cuspidulata pertenece a Justicia sect. Dianthera subsect. Saglorithys, un erupo de especies afines muy diversificado en Amé- rica del Sur que se caracterizan por las inflores- cencias en espigas simples y las flores medianas, con cáliz 4-partido y semillas tuberculadas (Tabla 1). A este grupo también pertenecen por ejemplo Justicia rusbyana Lindau (Bolivia. Yungas. Bang 370, isotipo, M), Justicia boliviana Rusby (Bolivia. Vic. Cochabamba, Bang 1225, holotipo, NY!; iso- tipos. K!. US), y la especie descripta como Rhyti- glossa. piahuiensis Nees (Brasil. Piahui, Gardner 2290, holotipo, K!: isotipo, BM), especies muy afi- nes a Justicia cuspidulata. La delimitación entre las especies de este grupo no es clara, y eso hace es- pecialmente difícil asignarle un nombre al material de Paraguay. Por otro lado el tipo de Justicia cus- pidulata, aunque morfológicamente muy similar al material de Paraguay, es de Perú, y existen pocos ejemplares que claramente parezcan pertenecer a esta especie en el área intermedia entre este país y Paraguay. Sin embargo. la flora del bosque tro- pical estacionalmente seco presenta distribución discontinua en Sudamérica, y existen varios otros ejemplos de especies de esta flora que presentan Ч al., 2000). Por estas razones, antes de publicar un este tipo de distribución disyunta (Pennington ~ nombre nuevo para el material de Paraguay, prefie- ro identificarlo con el de Justicia cuspidulata, con el que está indudablemente relacionado. El nombre Justicia amamba yensis C. Ezcurra que utilicé para esta especie en determinaciones en etiquetas de herbario hace unos años nunca fue publicado. 246 Annals of the Missouri Botanical Garden M Zi duas : — — —— E 1 f x x V. seD Figura 7. Justicia cuspidulata. —A. Rama con flores. —B. Flor con bráctea, cáliz, corola. estambres ye 2 = C. Corola abierta con androceo. —D. Cápsula abierta sin semillas. —E. Semilla. Soria 2036 (G ). Schinini 20534 (С). Volume 89, Number 2 2002 Ezcurra 247 Justicia en Sudamérica Austral El tipo de Rhytiglossa cuspidulata (Mathews 3152bis) está mezclado con el de Rhytiglossa hoo- keriana (Perú. Chachapoyas, Mathews 3152). una especie con hojas más pequeñas y flores más an- gostas, como queda claramente explicitado en su descripción original, y los isotipos que se encuen- tran en BM tienen la identificación invertida. Material seleccionado. BOLIVIA. Santa Cruz: Bue- cras, ears h 6973 (BM). Beni: — Estación Biológica Beni, са. 1/2 hour by jen o R. Curiraba. Solomon 14633 (M). PARAGL ede Amambay. Ruta 3 y Río Aquidabán. Se ы y Bordas 20534 (CTES, US): in regione cursus py rioris fluminis or Hassler 8003 (BM, С. NY), 1 (BM, G, MO, NY. Р); orillas del Río ш: ruta 3. Kra- pávickas 15995 (С — Parque Nacional Cerro Cora, ori- llas + — Soria. Zardini y Ortiz 2036 (С. FC Q. MO re › Apa und Rio Aquidabán, he — 4406 (BM. 3 = P): Caballero Cué, Fiebrig 4473 (BM, E. G, K.F du Торак. Cerro Tobatí, in nas 287 (CTES): * batí, Sparre y Vervoorst 1412 (LIL). navista, desmontes chac 10. Justicia dumetorum Morong, Ann. New York Acad. Sei. 7: 193. 1893. TIPO: Paraguay. Río Pilcomayo, 1888-1890, Morong 1538 (ho- lotipo, NY! CE N. E. Brown, Trans. Bot. Soc. Edinburgh : 67. 1896. TIPO Río Pilcomayo 1890-189 l. патот моз Lindau, Bull. 2, 73. 1905. TIP о Santa f lisa, Fab. 1903. leg. Rojas, Н lotipo, B безі, fot. Я я 1201: LIL MO! NY! PL Argentina. x т = Seed BM! GL KL isolipos, Hierba sufruticosa de 0.8-1.5 m alt., con ramas erectas, cilíndricas. longitudinalmente estriadas, en- grosadas por encima de los nudos. Hojas con pecíolo de 0.5-2 em long. v lámina ovada a lanceolada de 4-12 cm long. por 1-3 em lat., aguda y acuminada en el ápice, cuneada o redondeada en la base. ge- neralmente glabra, con cistolitos pequeños densa- mente esparcidos en su superficie. Espigas simples. densas, apicales y en las axilas de las hojas supe- riores, de 3-10 ст long.. a veces agrupadas en ra- cimos foliosos en el extremo de las ramas. Brácteas lanceoladas, opuestas, de 1—1.5 em long. por 1.5-2 mm lat., glanduloso-pubérulas, con el margen blan- quecino y Майпо, cada una con una flor sésil en su axila y dos bractéolas de 7-9 mm long. en la base. Flores con el eáliz de la misma longitud que las bractéolas. profundamente 5-partido, los segmentos de 6-10 mm long. por 1.5 mm lat.. glabros y albo- marginados. Corola de color morado de 34. cm long. con tubo de 2-2.5 cm long.. poco ensanchado en la fauce. y labios de 1.3-2 em long.. el posterior apenas emarginado. el anterior profundamente trilobado. Еч con los lóbulos de 11 mm long. por 4—5 mm lat. Estambres con filamentos de 8 mm long. insertos en el ápice del tubo, v anteras con las tecas a distinta altura: la superior de 2 mm long., la inferior apenas más larga. v apendiculada en la base. Cápsula obo- voide, de 1.5 em long. por 0.6 em diám., sólida y lateralmente estrechada en la mitad inferior, inflada y 4-seminada en el ápice. Semillas subesféricas, de lisas y lustrosas: jaculato- —3 mm diám., castañas, res de 2.5 mm long., obtusos. Figura 8. Ilustraciones. hábitat Sudeste de Paraguay y norte de la Argentina. en Distribución, y fenologta. Bolivia, a región del Chaco oriental. Habita en bordes y abras de bosque chaqueño, en lugares soleados con inun- dación periódica, y florece en primavera, verano y otoño, de octubre a mayo. Justicia dumetorum se caracteriza por las flores rojas o moradas, de corola larga (3—4 cm). aparen- temente polinizadas por picaflores. Existen algunos ejemplares de Paraguay y Bolivia que se citan con asterisco (*) en la lista de material (Fiebrig 5074. Fiebrig 5.138, Beck 5373 y Troll 1688) que se di- ferencian solamente por las corolas menores (de ca 2 em long.), lilacinas y de tubo basal más corto, con aspecto de ser polinizadas por abejas о abe- jorros (melitófilas). Estos ejemplares se asemejan bastante a Justicia furcata Jacq., especie descripta para México, y en algunos casos fueron determi- n ejemplares ~ nados con ese nombre por Lindau de herbario (leg. Fiebrig 5074 y s.138). Justicia furcata no habita en América del Sur (Wasshausen, com. pers.). y además estos ejemplares se diferen- cian de Justicia furcata del sur de Nortemérica у de América Central por las brácteas generalmente más anchas v las inflorescencias con menos flores. Sin embargo estos ejemplares coinciden mucho con especies afines. descriptas para. América del Sur como Justicia peruviana Lam. (ilustrada en Cava- nilles, Icon. 1: 17, tab. 28. 1791, y Curtis Botanical Magazine 11—12: 430. y Justicia lithospermifolia Jacq. 1797), Justicia carthagine- nesis Jacq.. Todas estas especies afines son muy variables, y en ellas se han encontrado formas con flores de distinto tamaño y diferente colorido (Leonard. 1951-1958: Daniel, 1995). con Daniel (1995) en que el conjunto de estas es- Debido a ello coincido pecies de Justicia sect. Simonisia indudablemente aliadas a Justica carthaginensis Jacq. 1 Justicia furcata Jacq. deberían ser objeto de un estudio de material abundante de toda su área de distribución, que incluya un análisis detallado de la forma y el color de la corola además de la distribución y el ambiente habitan. que requisitos. indispensable para intentar diferenciarlas. Hasta tanto se realice 248 Annals of the Missouri Botanical Garden — 8. Justicia dumetorum. —A. Rama con flores. | А OF ) B. Flor con bráctea, bractéolas, cáliz, corola, estambres y estilo. —C. Corola abierta con androceo. —D. Estilo y estigma. —E. Cápsula abierta. —F. Semilla. Schulz 8673 (SI) Saravia Toledo 10779 (SI). ese estudio, trato al material de flores cortas y li- lacinas que aparece en Bolivia y Paraguay como parte de la variabilidad intraespecífica de Justicia dumetorum. Se aclara que el supuesto isotipo del nombre Justicia kerri, Kerr 108, que se encuentra en el her- bario de Edimburgo (E). pertenece a Justicia bra- siliana Roth y no concuerda con la descripcion original de Justicia kerri. Probablemente esté erró- neamente etiquetado. Volume 89, Number 2 2002 zcurra 249 Justicia en Sudamérica Austral ARGENTINA. Corrientes: rs Bernardello y Galetto 800 (SI); Dpto. San losme, 5 km E de Paso de la Patria, Krapovickas y Cris- E 11889 (SI). Chaco: — Resistencia, — ras, Meyer 16235 (LIL); Dpto. 1° de mayo, С‹ 1 Be- nítez. Schultz 8673 B. i r monte, Jörgensen 2332bts (SD): Clorinda a Formosa, Km 3, Morel 17 Tigre, Parodi 8511 (LIL). Salta: Rivadavia, 18 k de Dragones, Saravia Toledo 10779 (51); Dpto. Orán, Em- barcación, km 1317, Schreiter 10866 (SI). BOLIVIA. Beni: Ballivián, ерйп, еп la zona de influencia del Rio Yac ‘uma, и húmeda, Beck 5373 * (flor — и ep.: Cañamina, Río La Paz, Troll 1688* (M). Boquerón: Colonia Menno, ect Are- nas 1097 (LIL), Arenas 1457 (US); Dr. ro P. Peña, 10 km al W del pueblo, Spichiger et al. 2217 (G, US); Fila- delfia, Hahn 787 (US). Chaco: Mayor Pedr Material seleccionado. Casa Cué, c ro Lagere nza. kn Salado, de Limpio a Embose ada, Arbo (CTES): Рае Correo, montes húmedos, orilla Río Salado. Hassler 1650 (SI). Concepción: Zw. Rio 2 und Rio Aquidaban, V dla Sana, Fiebrig 5074* (G, К Presidente Hay Laguna Yaraguf, 7 km « . dep.: s. loc., Fiebrig s.138* (G). (BM); Colonia Menno, Я Arenas 143 (US); in regione cursus inferioris fluminis 'omayo, M 81 (G, MO); pozo € — eds les y gen 5563 (FCO): Villa Hayes, Zardini et al. 2592 (М‹ у). Zardini et p 2593 (FCQ, MO, US 11. Justicia floribunda (C. Koch) Wasshausen. Darwiniana 35: 151. 1998. Libonia floribunda C. Koch, Wochenschr. Ver. Bef. Gartenb. Kon. Preuss. Staat. 6: 266. 1836. TIPO: ilustración del ejemplar cultivado en el Jardín Botánico м de Bruselas sobre el que se basó la descrip- ción original en Morren, Belg. Hort. 14: 12. 1864 (lectotipo, — por Wasshausen en Peixoto et al., 1 Seric sige i ig Nees, in Martius, Fl. Bras. 9: 109. 1947. Justicia — (Nees) Griseb., Abh. Ko- i Wiss. Gottingen 24: 262. 1879, non Justicia а Vahl, Eclog. Am. 1: 2. 1796. Jacobinia pauciflora (Nees) Lindau. in Engler & Tus Nat. flanzenfam. asl. 1895. Wasshausen, Baileya 19: 3. loc., Sellow s.n. iim K!). — — Justicia. rizzinii 1973. TIPO: Brasil. 5. fot. F 89147; prob- В destr., p. ables isosintipos, El, Sericographis pauciflora — a — Nees, nr tius, Fl. Bras. 9: 109. TIPO: Brasil. S. loc., Tweedie s.n. (sintipo, oo Arbusto ramoso de 0.50—1(—2) m alt., cilíndricas. pubérulas. Hojas con pecíolos de 0.4— con ramas 1 em long. y lámina ovada, elíptica a obovada, muy variable en tamaño, de 1.5-7 cm long. por 0.8-2. mm lat., las de cada par desiguales, las superiores mayores, obtusas у cuneadas, enteras a levemente aserradas, discolores, glabras o leve y cortamente pubescentes. Espigas paucifloras frecuentemente reducidas a flores solitarias sobre pedtinculos dé- biles hasta de 3 em long. en las axilas de las hojas superiores, con brácteas y bractéolas triangular- anceoladas, ca. de 1 mm long. por 0.5 mm lat., glabras a cortamente ciliadas. Cáliz 5-partido de 5 mm long., segmentos lanceolados, de 3.5 mm long., acuminados, pubérulos. Corola de 2-3 cm long., roja y con la mitad al tercio superior amarillo, tubo ~ ‘a. de 20 mm long. por 5 mm diam. en la fauce, por fuera pubérulo, con algunos pelos glandulares, interiormente glabro; los labios oblongos, ca. de 5 mm long., el posterior erecto y obtuso o emargi- nado, el anterior erecto o algo patente y trilobado, con los lóbulos de ca. 2 mm long., triangulares. Estambres inclusos debajo del labio superior, fila- mentos pilosos en la parte media, anteras ca. de 2 mm long.. algo superpuestas, oblicuas, caudadas. Ovario glabro o glanduloso-pubérulo en la porción apical. estilo glabro. Cápsula claviforme de 12-17 mm long. por 4 mm lat. y 2 mm de espesor, glabra o glanduloso-pubérula; semillas aplanadas, subor- biculares. ca. de 2 mm diám., castañas, levemente rugosas: retináculos, ca. de 2 mm long., obtusos. Ilustraciones. Peixoto et al., 1998: 153. Distribución, hábitat y fenología. Sur de Brasil v noreste de Argentina (este de Misiones). Tal vez también se encuentre en Paraguay oriental. Habita en selvas y bosques húmedos, especialmente sobre suelos periódicamente inundados, entre el nivel del mar y 1400 m de altitud. Florece en invierno, de mayo a agosto, y sus flores rojas de ápice amarillo probablemente sean polinizadas por picaflores. Usos. Esta especie se cultiva como ornamental en regiones cálidas del mundo por sus llamativas rojas (Hay & Beckett, pauciflora). Las flores de esta especie se producen flores 1975. sub Jacobinia principalmente en invierno, y las plantas se repro- ducen fácilmente por gajos no floríferos y semillas (Parodi & Dimitri, 1980, sub Jacobinia pauciflora). Vombre vulgar. “Bandera española” en Argen- tina (Parodi & Dimitri, 1980, sub Jacobinia pau- ciflora). Justicia pauciflora se caracteriza por las flores ornitófilas con la corola roja de ápice amarillo. Pa- rodi y Dimitri (1980, sub Jacobinia pauciflora) ci- taron a esta especie frecuentemente cultivada en \rgentina como originaria de Brasil, por lo que cabe la posibilidad de que las poblaciones silves- tres de Misiones hayan resultado de escapes de cultivo. Esto ha sucedido con otras especies de Acantáceas neotropicales cultivadas, que se han asilvestrado en regiones alejadas de su área de dis- tribución original, como por ej. Justicia. carnea Lindl. (en Nueva Zelandia. Webb, 1995) y Ruellia breviflora (Pohl) Ezcurra (en Asia y Australia, Ez- 250 Annals of the Missouri Botanical Garden curra, 1989). Sin embargo, es probable que las po- blaciones silvestres de Justicia floribunda del este de Misiones se encuentren dentro del área de dis- tribución natural de la especie, ya que ésta fue descripta originalmente para regiones contiguas del sur de Brasil (Nees, 1847a, 1847b; Wasshausen & Smith, 1969) y el siglo XIX, como lo demuestra el ejemplar colec- ademas ya existfa en Misiones en cionado por Niederlein en 1887. Material seleccionado. ARGENTINA. Misiones: Ber- nardo de lri rigoyen, Hunziker, J. 950, 959 (LIL); Varana, territorio de Misiones, Niederlein 1776 (LIL); Guaraní. Predio Guaraní, Arroyo Soberbio, borde del arroyo en bar- ranca inundable, — et al. 5647 (SD; Gral. Manuel Belgrano, de Bernardo de Irigoyen a San Antonio pasando Salto Andresito, Zuloaga 2111 (SD. 12. Justicia gilliesii (Nees) Bentham, in Bentham « Hooker, Gen. Pl. 2(2): 1109. 1876. Adha- in DC., Prodr. 11: 395. . TIPO: Argentina. San Luis: Gillies s.n. toda gilliesii Nees, 1847 (sintipo, K!; isosintipo, E!); Santiago del Es- Tweedie 1260 p.p., mezcla con Justicia tweediana (sintipo, K!). tero: "St. lago de Cordova," Justicia е Hieron., Bol. Acad. Nac Córdoba 4(1): 61. 1881. TIPO: Argentina. oh duos De} zo. — cia Maradona, Ene. 1876, Echegaray n. (holotipo, CORD!) Sufrátice ramoso de hasta 0.60 m alt., con ramas ascendentes o rastreras, cilíndricas, estriadas, pu- bescentes. Hojas pecioladas con pecíolos de 0.5-2 cm long., ovadas a elípticas, de 2.5-3 em long. por 1.8-2.5 em lat.. agudas и obtusas y apiculadas, cu- neadas en la base, membranáceas, generalmente pubescentes. Flores sésiles o cortamente pedicela- das, con dos bractéolas en la base, solitarias en las axilas de brácteas, formando espigas de 4-12 cm long.; bractéolas elípticas de cerca de 1 cm long. por 2 mm lat.; brácteas de 9-12 mm long. por 4— 5 mm lat., ambas glanduloso-pubescentes, ciliadas en el margen. Cáliz profundamente 5-partido, con los lóbulos angostamente ovados a lanceolados, su- biguales, de 4 mm long.. foliáceos, generalmente glabros. Corola celeste-lilacina a blanca de 2 cm long., pubérula, con tubo corto de cerca de 1 ст long., ensanchado en una garganta de aproxima- damente 0.8 cm diám., los labios de 1—1.3 cm long., el anterior trilobado hasta la mitad, con la fauce blanca, convexa y transversalmente rugoso- venosa, con puntos morados. Estambres insertos en la base de la garganta, las anteras con las tecas subparalelas, una por encima de la otra, la inferior generalmente apendiculada en la base. Cápsulas obovoides, pubescentes, de 1.2-1.5 em long. por 5-7 mm ancho, con la mitad inferior sólida y la- teralmente comprimida. Semillas generalmente 4, suborbiculares, aplanadas, de 3—4 mm diám., par- das, diminutamente rugoso-tuberculadas; retinácu- los obtusos. Ilustraciones. Figura 9. Noroeste vy Distribución, hábitat y fenología. centro de la Argentina, desde Catamarca, Tucumán, La Rioja, San Juan y Mendoza hasta Córdoba, San Luis y Santiago del Estero. Se la encuentra en mon- tes áridos entre 300 y 1000 m s.m. Florece en ve- rano y otoño, de diciembre a abril. Nombre vulgar. “Albahaca de vaca” (Burkart, 1943, sub Justicia echegarayi). Justicia gilliesii ha sido citada como de impor- tancia como planta forrajera (Ariza-Espinar, 1971), al igual que varias otras Acantáceas del sur de Su- damérica (Burkart, 1943). Material seleccionado. ARGENTINA. Catamarca: La Paz, el Divisadero, 13 Abr. 1947, Brizuela 1132 (LIL); Ancasti, El Cajón, Brizuela 1266 (LIL); Ambato, entre La Puerta y Huaicama, Cristóbal 442 (LIL); Capital, Cata- marca, Nicora s.n. (SI 18656); Camino a El Juncal, Soria- no 628 (SI); Paclín, entre Catamarca y Cuesta del Totoral, Villa — 1119 (LIL). Córdoba: Tulumba, Salinas Grandes . V. Mansilla, Hunziker, А. 10985 (CORD); Cruz del Eje. — Cruz del Eje y Villa S mbre, Nicora 1793 (SI). oja: San Martín, Bajo s Campo Balde El Tala, uso 17, — 4044 (CORD); Chamical, Sierra de are C — ily la ipu Biurrun 1372 (CORD); Capital La Rioja, Burkart 12531 (SI); en las cercanías de arlos, camino a Córdoba, Hierony- mus y Niederlein p p ORD); Sierra de Velasco, — al Dique Los Sauces, Hunziker, А. 4721 a IL); A. V. a 29, al S de Punta de los Llanos, * el desvío a Tama у Carrizal, — А. 15096 (CORD); Famatina, Guanchin, Venturi 6 (SI). Mendoza: Paz, Desaguadero, Ruiz Leal Te (LIL). San Juan: Va- lle Fértil, N de Usno, Kiesling 8961 (SI); Valle Fértil, camino a La Majadita, Kiesling 5012 (SI). San Luis: Cap- ital, San. Luis, alrededores, Troncoso s.n. Pe — Bel- grano, Alto Pencoso, Bruch y Carette rm Sierra de Las Quijadas, Qda. del Alambre, Fi W de los Llanos, falda F & » 2. I — n Pierotti s.r mán: Vipos, Lillo 7894 (L n ); Dpto. — ‘as, Tapia, Ven- 1814 (SI). — — — y A oe — ~ D - — = ae ^ D DP ~ = n z ~ = - = ~ Y — = turi 13. Justicia glaziovii Lindau, Bull. Herb. Boiss. 483. 1895. TIPO: Brasil. Glaziou 13073 (lectotipo, designado por Graham (1988), K!). Rhytiglossa — Nees, in Martius, Fl. Bras. 9: 129. 7, non Justicia paniculata Burm. f. 1768. TIPO: Brasil. 8 ahia: Ilheos, Blanchet 2979 (isotipo, BM!). Lophothecium paniculatum Rizzini, An. 5 Bot. Hio de aneiro 8: 336, t. 5. 1948, ' “Minas ( ›е- rais: Ituitaba, Macedo 1123 (isotipo, BM!) Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral 1 mm Е мдү у М Y ` C Figura 9. Justicia gilliesii —A. Planta con flores; —B. Flor con bráctea. bractéolas. cáliz. corola, estambres v estilo. —C. C ў Cá ) Márquez s.n. (SI 45519) Justicia velascana Lindau. Bull. Herb. Boiss. 3: 184 1895. TIPO: Bolivia. “Prov. m, Ago., Kuntze s.n. (holotipo, B destr.. fot isolipo, NY!). ‹ elasco orientali", 200 p К 0070|. . OOO... Arbusto ramoso con tallos ramificados, cilíndri- cos, сопатепіе pubérulo-tomentosos. Hojas con pecíolo de 0.5-1.5 em long., y lámina ovada. de 5— 8(-15) em long. у 2.54.5 em lat. (la de las hojas apicales menor). aguda en el ápice, cuneada a re- dondeada en la base v algo decurrente. sobre. el pecíolo, glabra o pubérula en el haz. pubérula o cortamente pubescente en el envés, en especial so- Corola abierta con androceo. —D. Estilo y estigma. Е. Cápsula abierta. —F. Semilla. Kiesling 8961 (S1). bre las venas. Inflorescencias paniculiformes en las axilas de las hojas superiores, amplias, tenues. de espigas laxas reunidas en racimos compuestos ge- neralmente más largos que las hojas, con los raquis pubérulos: brácteas anisofloras. lineales o triangu- ares, diminutas. de 1-2 mm long. y 0.5 mm lat., elabras o pubérulas: bractéolas similares, más an- gostas. Cáliz profundamente 5-partido, de ca. 4 mm long., con los segmentos lineales, de 3-4 mm long. por 0.5 mm lat., glanduloso-pubérulos. Corola color crema de 0.8-1.3 em long., con tubo acampanado de 5-8 mm long. y 3-5 mm Фат. en la fauce. los Annals of the Missouri Botanical Garden labios de ca. 5 mm long. el posterior ovado, de 4— 5 mm lat., biapiculado, el anterior trilobado, con los lóbulos de ca. 2 mm long. por 2.5 mm lat. Es- tambres exertos hasta el ápice del lóbulo posterior, con las tecas superpuestas, oblicuas, de menos de | mm, separadas por el conectivo, la inferior más pequefia y con un apéndice glandular claviforme en la base. Cápsulas angostamente claviformes de ca. | cm long. y 2 mm lat., sólido y lateralmente comprimido y la porción su- perior engrosada, algo estrechada en su parte me- dia, pubérulas. Semillas de 1.5 mm diám., tuber- culado-verrugosas; retináculos de ca. 1 mm long. Ilustraciones. — Wizzini, 1948: 339 (sub Lopho- tecium paniculatum). Sur de Brasil, este de Bolivia, y noreste del Paraguay. Habita en Distribución, hábitat y fenología. bordes de bosque. Florece en verano y otoño, prin- cipalmente de enero a marzo. Justicia glaziovii se caracteriza por las inflores- cencias paniculiformes amplias y tenues, con flores color crema de cerca de 1 cm long., y es una es- pecie relativamente rara que ha sido poco colec- cionada. Material seleccionado. PARAGUAY. Amambay: in altiplanitie et declivibus "Sierra de Amambay”, Rojas 10864 (BM, G, K, M, NY, P, W); Sierra de Amambay, in via sylvatic a, Picada Yatobó, Cerro Рога, Rojas 104 ; Zw. Río Apa und Río Aquidabán, Ir а, Fiebrig 4811 (BM, G, K). Concepción: Estancia Potre- rito, Basualdo 3485a (FCQ). 14. Justicia glutinosa (Bremekamp) V. A. W. Kew Bull. 43: 613. 1988. Sarotheca glutinosa Bremek., Proc. Kon. Ned. Akad. Wet. C, 72: 426. 1969. TIPO: Bolivia. Brooke 5677 (holotipo, BM no visto, fot. US!). Graham, Sarotheca elegans Nees, in Martius, Fl. Bras. 9: 113. ‚ non Justict ia — Bêauv. 1810. Ju sa- — . W. — taham, gs — 43: 614. TIPO: brasil: Goiás . Felis E E = ciras, Pa 1989 holotipo, №, P F 327 Hierba sufruticosa de 0.5-1.5 m alt.. con tallos subcilíndricos, largamente hirsuto-pubescentes cuando jóvenes, glabrescentes a la madurez. Hojas con pecíolo de 1—4 cm long., y lámina ovada a anchamente elíptica de 7-15 em long. por 3-5 cm lat., aguda y acuminada en el ápice, cuneada y decurrente en la base, pubérula en el haz y más o menos hirsuto-pubescente en el envés, especial- mente sobre las venas principales. Inflorescencias en espigas densas reunidas en racimos peduncu- lados en las axilas de las hojas superiores: pedún- culos y raquis glanduloso-pubescentes; brácteas con el tercio inferior opuestas isofloras o alternadamente anisofloras, anceoladas, de 2.5-3 mm long. y | mm lat., glan- duloso-pubescentes; bractéolas lineales, algo me- nores. Cáliz profundamente 4-partido, con los seg- mentos lanceolados de 4—6 mm long. y 1.5 mm lat., densamente glanduloso-pubescentes. Corola de ca. | cm long., blanca con marcas lilas en la fauce, externamente glanduloso-pubescente, el tubo le- vemente obcónico, de 4-6 mm long., y labios de 4-5 mm long., el posterior ovado, de 2 mm lat., agudo, el anterior trilobado, con los lóbulos de 1.5— 2.5 mm long. redondeados. Estambres exertos hasta el ápice del labio posterior, las anteras con la teca superior oblicua, de ca. 1 mm long., la inferior algo menor y apendiculada en la base. Cápsula clavi- forme de 1.5-2 cm long. y 0.4 cm diám., densa- y cortamente glanduloso-pubérula. Semillas marca- damente aplanadas, de 3 mm diám., papilosas; re- tináculos de menos de 2 mm long. Ilustraciones. Nees, 1847a: tab. 18 (sub Saro- تس theca elegans). Distribución, hábitat y fenología. Perû, Bolivia, extremo norte de Argentina, sudoeste de Brasil y nordeste de Paraguay. En Paraguay habita en bor- des de selvas riberefias rodeadas de campos cerra- dos. y florece principalmente en primavera, de agosto a noviembre. El nombre Justicia sarotheca se reduce a la si- nonimia de Justicia glutinosa por primera vez. Esta especie se caracteriza por sus hojas grandes y flores medianas, de cerca de 1 em long., en inflorescen- cias glandulosas de espigas densas reunidas en ra- Existe una ilustración de esta es- Flora cimos axilares. pecie (como Sarotheca elegans) en Brasiliensis (Nees, 1847a). Material seleccionado. ARGENTINA. Salta: Dep. San Pu Quebrada Astillero, Schulz 5484 (CTES). PARA GUAY. Amambay: Cerro Corá, en arroyo interrum- pido en selva, Schinini y Bordas 20344 (CTES, G, US); Parque Nacional Cerro Cora, bosque húmedo de hasta 25 m de altura al lado = Arroyo Aquidaban-Ningiif, suelo Brunner 1442 (G, MO): Parque Nacional Cerro sora, cerca de la casa forestal, en una maleza algo meda al lado de | la carretera, Ferndndez Casas 4114 rque Nacional Cerro Cora, ape: 5 del rend Aduidaban- Aguí, Soria 1727 (FCQ, G, MO); Parque Na- ional Cerro Сога, selva marginal i — Aquidabán- N iguí, Ferrucci, Vanni y Ferraro 693 (CTES); Parque Na- cional Cerro Cora, Mereles 2054 (G). arenoso, Graham, Kew 1988. s — ylax umbrosus Prodr. 313. 1847, non Jus- 1841. TIPO: Argentina. Tucumán: San Javier, Twee- die 1262 (lectotipo, aquí designado, К!). 15. Justicia goudotii V. A. W. Bull. 43: 603. Nees, in DC., ticia umbrosa a 2: Hare: 79. Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral pcr boliviensis Lindau, Bull. Herb. Boiss. 3: 492. non Justicia boliviensis (Bremek.) V. А. W. 1988. b IPO: Bolivia. i. Kew Bull. 43: 613. anta Cruz: 1000 m, Mayo 1892, Kuntze s.n. (sintipo, p d fot. Е 88971; isosintipo. NY!): Brasil. Pa- í, s. col. 466 (sintipo, B. destr. lla vestitus Rizzini, Dusenia 3(3): 191. 1952, non Justicia — Schult., — 1: 145. 1822. TIPO: Brasil. Paraná: Parque Nacional do Iguazú. 3 Abr. 1949, Duarte y Pereira 1927 — RB no visto). Hierba perenne, rizomatosa en la base, con tallos erectos subcilíndricos, a veces decumbentes y ge- niculados, de hasta 80 cm long.. opuestas con pecíolo de 3-10 mm long. y lámina 2-9 cm long. por 1.5—3.5 pubérulos. Hojas ovada de 2 cm lat., acu- minada en el ápice, cuneada y algo decurrente en a base, entera, pubérula en el haz, pubescente so- bre la nervadura en el envés. Inflorescencias for- madas por flores sésiles con dos bractéolas en su base, dispuestas en el eje de la inflorescencia for- mando una espiga unilateral. Espigas terminales o en las axilas de las hojas superiores, a veces en- riquecidas por espigas derivadas de yemas super- numerarias. Brácteas y bractéolas lanceoladas, rí- gidas, de 4-6 mm long., ciliadas, con el nervio ее medio prominente en el dorso. Cáliz de cerca de € mm long., con los 4 segmentos rígidos. lineales, subulados, dorsalmente pubescentes sobre el ner- vio medio, ciliados. Corola pálido lilacina a violá- cea de 1.5-2 cm long., con tubo muy angosto de 1-1.5 em long., estrecho en la garganta, el labio superior erecto, de 4 mm long. por | mm lat., el labio inferior ancho, trilobado, de cerca de 5 mm long.. con estrías blancas en la fauce. Estambres con filamentos de aprox. 5 mm long.: teca apical perfecta, de contorno ovado, de 1 mm long., la otra inserta más abajo en el filamento, semiestéril, más angosta y basalmente caudada. Cápsula de contor- no obovado, de aprox. 7 mm long. por 2.5 mm lat., cortamente sólido-estipitada en la base. glabrescen- te y de paredes finas. Semillas 4, aplanadas. de mm diám.. con pelos cortos y gloquidiados: reti- náculos de 1 mm long., obtusos. Ilustraciones. Ezcurra, 1993a: 344. Distribución, hábitat y fenología. Colombia. Bolivia, Brasil, Paraguay y norte de la / Habita en el interior o en los bordes de selvas me- Arge ntina. sófilas y bosques xerófilos de tipo Chaqueño, fre- cuentemente en ambientes alterados de planicies y serranías hasta 1300 m aprox. Florece a fines de verano otoño e invierno, de enero a julio. “Uchu yuyo” en el noroeste de 1943, sub Chaetothylax um- Nombre vulgar. Argentina (Burkart, brosus). Esta especie se describió bajo el nombre de Chaetothylax umbrosus Nees sobre la base de dos sintipos, uno de Colombia (Goudot s.n.. K!) y otro de Argentina (Tweedie 1262, K!). Graham le puso un nombre nuevo al pasarla a Justicia, Justicia gou- dotii Graham, por ya existir el nombre anterior Jus- ticia umbrosa Benth., pero no designó un lectotipo. Debido a que los sintipos pertenecen a dos especies diferentes, una de Colombia y otra de la región subtropical de Sudamérica (norte de Argentina, su- deste de Bolivia, Paraguay y sudoeste de Brasil): 1) Selecciono el ejemplar de Tweedie de Argentina como lectotipo de Chaetothylax umbrosus (= Jus- ticia goudotii). La elección de este ejemplar se basa en que, ante la falta de otras evidencias en el pro- tólogo. el epíteto umbrosus que eligió Nees para la especie coincide con los datos del ejemplar de Tweedie “muy abundante en los bosques mas som- bríos de San Javier” (traducido). Por otro lado, la elección de este ejemplar preserva el uso actual del nombre, ya que la especie de la region subtropical de Sudamérica ha sido identificada con los nombres Chaetothylax umbrosus y Justicia goudotii en varias publicaciones (por ej. Ezcurra, 1993a, 1999), y está identificada con estos nombres en numerosos ejem- plares de herbario. (2) El ejemplar de Goudot de Colombia pertenece a la especie determinada por Leonard (1951-1958) como Chaetothylax huilensis, que se diferencia de Justicia goudotii por las inflo- bractéolas y rescencias más anchas, las brácteas, lóbulos del cáliz más largos, y las cápsulas mayo- res. Material seleccionado. ARGENTINA. Catamarca: del Alto. Balcagua, Venturi 7080 (SI). Corrie — Ca- mino a Sta. Cat: Soriano 1665 (SI). Chaco: Las Bre- ñas, legname y — 8704 (LIL): * е PN Cha Маша 927 (SI). Formosa: Km 39 de j se Lagun a rs Meyer Venturi 5301 (LID): dina, Calile gua, da, Legname 6857 (LIL); Capita Ruta 51, km 20, = km al E de Salta, Novara 3502 (SI); Anta, Tacuí, 15 Km al S de Macapillo, límite con Santiago, Novara y Terne 10284 (LIL): Orán, San Ignacio, Pierotti 83 (LIL); Rosario de la Frontera, pon 6184 (LIL). Santa Fe: Villa Guillermina, Meyer 6 (L IL. Santiago del Es- ero: Choya, La Punta, E Del Mogote, Qda. De Cala- pachin, Frenguelli 128 (SI); Villa La Punta, Sa. de Gua- saván. Ulibarri 1509 (SI). T та Rio Chico, Qda. de Marapa, — — s.n. (LH í, Parque Aconquija, de la Sota 1138 (LIL); Tucumán, 5 s.n. (К); Famaillá, La died "Ree 1724 . PARAG = Guairá: Cordillera de Ybytyruzú, Cerro My My, camino a Cerro Polilla, Zardini y Velásquez 13803 (MO). Р Жыр a — 1 Chaco, Santa Elisa, Rojas 2818 „СЕ, LIL, MO, NY, T in — cursus inferioris ША. Pil- 'omayo, Us jas 155 (G); 50. Soria 280 (G); Route — :haco, pk E Billiet y y Jadin 3057 (BM, G NY). їл = 254 Annals of the Missouri Botanical Garden 16. Justicia hassleri (Lindau) V. A. W. Graham. Kew Bull. 43: 615. 1988. Beloperone hassleri Lindau, Bull. Herb. (App. 1): 30. 1898. TIPO: Paraguay. "In дитей pr. Hacu- rubi,” Hassler 1089 (holotipo, С: isotipos. MO!, NY!, P!). Boiss. e marginata Lindau, Bull. Herb. Boiss. 3: 19 895, non Justicia marginata Lindau, 1894. puede rb cad A. W. Graham, Kew Bull. 3: O15 TIPO: Paraguay meridional, Se dm 1892, — ге s.n. кош B destr., fot. F 8% NY". isotipos, С!, — Hierba sufruticosa, rizomatosa en la base, con tallos erectos subcilíndricos, de hasta 40 em long., generalmente glabros. Hojas sésiles, opuestas, con Amina ovada de 3.5-6(-9) ст long. por 1.3-3 cm lat., entera, rígida, coriácea, aguda y acuminada en el ápice, redondeada a cordada en la base, con el margen escarioso y las venas prominentes en el en- vés, glabra. Flores sésiles dispuestas en espigas denm sésiles, terminales o en las axilas de las hojas superiores, con brácteas anisofloras. Brácteas y bractéolas angostamente lanceoladas a lineales, imbricadas, rigidas, de 8-15 mm long. por 1.5-2.5 mm lat., largamente acuminadas, escariosas, con el margen blanquecino. Cáliz 4-partido de 10-15 mm long., con los 4 segmentos rígidos. lineales, subu- lados, albo-marginados. Corola pálido lilacina a violácea de ca. 2.5 em long., con tubo muy angosto de 1.5-2 em long., estrecho en la garganta, el labio superior erecto, de 6 mm long. por 1.5 mm lat., el labio inferior ancho, trilobado, de cerca de 8 mm long., con los lóbulos de 5 mm long. por 4 mm lat.. el medio algo mayor. Estambres con filamentos de aprox. З mm long.: teca apical perfecta, de contorno ovado, de 1 mm long., la otra inserta más abajo en el filamento, más angosta y basalmente caudada. Cápsula de contorno elíptico, de 6 mm long. y 5 mm grosor, cortamente estipitada en la base, glabra y de paredes finas. Semillas 4, aplanadas, de 1 mm diám., corta y densamente papiloso-pilosas, pardo- oscuras; retináculos agudos. Ilustraciones. Figura 10. Distribución, hábitat y fenología. Sur de Brasil (Mato Grosso do Sul) y Paraguay central. Habita en bordes de bosques. Florece a lo largo del año, prin- cipalmente en invierno, de j junio a septiembre. Justicia hassleri se asemeja a J. alopecuroidea T. F. Daniel (Daniel, 1990), especie de México, lo que sugiere relaciones filogenéticas estrechas entre al- gunas especies de Justicia de Sud y Norteamérica. las que deberían estudiarse. Material PARAGUAY. seleccionado. Cordillera: 3 km desvío ruta 2 a Piribebuy, Ferrucci 758 (CTES); in — sto Hacurubi, Hassler 3053 (G, K, NY, P, №); in du- eto pr. Itacurubi, Hassler 1138 (G); San Bernardino, ori- llas montes, Hassler 138 (SI). Paraguarí: Cerrito, in du- metis, Osten y Rojas 8978 (С): Cerro Hu, pres de Paraguarí, Balansa 2456 (С. К); Cerro Mbatoví. camino al arroyo Mbatoví, Basualdo 2658 (FCQ); Cerro — forest along N slope, Zardini y — 9943 (MO); Cerro Paraguarí, Hassler 916 (BM, W), Fiebrig 916 (E, G, К, ); in re gione collium Cerros de Раа, Hassler 6476 (BM, G. K, LIL, NY, W). S. dep.: Paraguaria Centralis, Hassle 3057 (BM); Piura Centralis, reg. lacus Ypa- caray, Hassler 12530 (BM, E, G, LIL, MO, NY, US); Sed- paraguay, Auntze s.n. (G) Justicia hunzikeri Ariza Espinar, Kurtziana 6: 92. 1971. TIPO: Argentina. La Rioja: San Martín, Sierra de Ulapes, falda E, frente a Ula- pes, 25 Mar. 1958, Hunziker, A. y Caro 13560 (holotipo, CORD!). Arbusto de 0.50—1.50 m alt., con ramas erectas, cilíndricas, las jóvenes amarillentas y cuadrisul- cadas, con estrías marcadas entre las costillas, ge- neralmente glabras. Hojas cortamente pecioladas, angostamente ovadas a elípticas, de 2—4 cm long. por 0.3-1.5 cm lat., agudas, cuneadas en la base, generalmente glabras, pubérulas solamente sobre la vena media en el envés. Flores sésiles o cortamente pediceladas en espigas terminales y axilares de 3— 5 cm. long.; brácteas ovadas o e otic as de 1.4 cm em lat., bractéolas lanceoladas de cer- ca de 1 em long. por 0.2 cm lat. long. por 0.7 Cáliz profunda- mente 5-partido, con los lóbulos angostamente ova- dos a lanceolados, subiguales, de 3—4 mm long.. foliáceos, generalmente glabros. Corola lilacina de 2-2.5 em long., pubérula, con tubo corto de cerca de | cm long., ensanchado en una garganta de aproximadamente 0.8 ст diám., 1.3 em I el anterior trilobado hasta la mitad, con la fauce blanca, los labios de l- ong., convexa y transversalmente rugoso-venosa. Estambres insertos en la base de la garganta, las anteras con las tecas subparalelas, una por encima de la otra, la inferior generalmente apendiculada en la base. Cápsulas obovoides, ro- bustas, glabras, de 0.8-1 cm long. por 5-6 mm diám.. con la mitad inferior sólida y lateralmente ‚ suborbicu- diminutamente rugoso-tuberculadas: comprimida. Semillas generalmente 4 ares, retiná- culos obtusos. Ilustraciones. — Ariza-Espinar, 1971: 93. Endémica de Argentina, se la encuentra en La Rioja, San Juan Distribución, hábitat y fenologia. Habita frecuente- mente en matorrales serranos sobre suelo rocoso a aprox. 600—800 m s.m. y florece en primavera v verano, de noviembre a marzo. y San Luis, en regiones áridas. Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral и. tilo. 1. Cápsula abierta. —D. Semilla. Hassler 156 S Justicia hassleri. — А. Rama con flores. Justicia hunzikeri se asemeja a Justicia tweedia- na en las corolas lilacinas, pero se diferencia por las flores en espigas secundifloras y las semillas rugoso-tuberculadas sin engrosamiento en el mar- gen, en lo que se asemeja a Justicia xylosteoides. Justicia hunzikeri habita un área restringida al ex- tremo NW de 5 Rioja y posee distribución simpátrica con J. xylos- San Luis. E de San Juan y S de La teoides. Esto sugiere que ambas especies podrían tener un origen filogenético común, pero que ac- tualmente estén aisladas geográficamente y posean diferentes tipos de polinizador. Material seleccionado. ARGENTINA, La Rioja: San Martin, Sierra de U — falda E, frente a Ulapes. Hun- ziker. A. y Caro 13562 (CORD): San Martín. Sierra de арав, Cerro de Mona. Stuckert 17193 (CORD). San jc Caucete, cerros al E de Marayes, Ruta 20, Múlgura 675 (SD: Valle Fértil, Quebrada de Mesada. Chucuma. ы 738 (SI): Valle Fértil, 3 km al W de Maraves en dec ción a San Agustín del Valle Fértil, Fortunato 5880 . Flor con bráctea. bractéolas, cáliz. corola. estambres у (S1); Valle Fértil: de Marayes a Las Tumanas, Kiesling 3013 (S1). San е Ве Igrano, Sierra del Gigante, cerca үтел 1936 (CORD): Бета, Sierra del Sie en las i Hunziker. A. у — jones de La Calera. Maldonado 16245 (CORD). 18. Justicia jujuyensis C. Ezcurra, Bol. Soc. Ar- vent. Bot. 25(3—4): 350. 1988. Justicia leonar- dii De Marco & Ruiz, Publ. Especial Inst. Li- llo: 50. 1976. 1973. TIPO: \rgentina. camino al Bana- non Wasshausen Jujuy: Ledesma. nal. antes de llegar a Arroyo Quemado, en- trando a la izquierda 3 km, leg. Legname y Cuezzo 5970C (holotipo, LIL! — alt.. tallos subcilindricos, cuadrisulcados, pubescentes a casi Hierba sufruticosa de 0.5-1 n = con elabros, con líneas longitudinales pilosas en su ju- ventud. Hojas con pecíolo de 0.5-1.5 cm long. у lámina ovada de 10-16 em long. por 1—7 cm lat., 256 Annals of the Missouri Botanical Garden discolor, aguda y acuminada en el ápice, obtusa a redondeada en la base, decurrente sobre el pecíolo, con cistolitos poco notorios en su superficie, ge- neralmente glabra en el envés. Flores sésiles dis- puestas en espigas cortamente pedunculadas en las axilas de las hojas superiores. Espigas más cortas que las hojas, densas; brácteas florales lanceoladas, de 8-10 mm long. por ca. 1 mm lat. segmentos triangulares, de 3—4 mm long., agudos glabros, ciliados en el margen. to, patente. long. y anteras con las tecas a distinta altura, | superior de 2 mm long., mayor y con un apéndice glandular en la base. Es- tilo glabro o pubérulo en la base; estigma bilobado; ovario glabro y disco sinuado. Cápsula de ca. 1 em long., con la mitad inferior sólida y lateralmente comprimida y la superior subesférica y 4-seminada, de 4 mm diám., glabra o pubérula. Semillas apla- nadas, de contorno orbicular, de 3 mm diám., con un reborde en su cara interna; retináculos obtusos, de cerca de 2 mm long. Ilustraciones. Lillo, 1937, lam. 2 (sub Justicia nemoralis). De Marco y Ruiz, 1976: 51, 58 y 59 (sub J. leonardii). Ezcurra, 1993a: 357. Distribución, hábitat y fenología. Sur de Boliv- ia y norte de la Argentina, en las provincias de Jujuy, Salta, Tucumán y Santiago del Estero. Habita en lugares hámedos y sombreados de bosques de transición en la zona de contacto entre las provin- cias fitogeográficas Chaquefia y de las Yungas. Flo- rece en verano, de enero a marzo. Nombre vulgar. “Ichiyuyo” (Ezcurra, 1993a). Justicia jujuyensis se asemeja mucho a J. corum- bensis, pero habita una región más occidental y se diferencia por las inflorescencias cortamente pe- dunculadas más cortas que las hojas. Los nombres Justicia nemoralis Lillo y Justicia nemoralis Lillo var. tomentosa Lillo (Lillo, 1937) son nomina nuda que fueron utilizados ampliamente en esta pn cie en etiquetas de herbario. El nombre Justicic nemoralis S. Moore (J. Bot. 47: 296. 1909) utilizado para una especie de África es un homónimo ante- rior que no permite su utilización para la especie de América del Sur. Por esta razón se creó el nom- bre nuevo Justicia jujuyensis para esta especie (Ez- curra, 1988) , foliáceas, pubérulas; bractéolas lineal-lanceoladas, de 5-8 mm long. Cáliz 5-partido, de 4-5 mm long., con los Corola blanca de 1.5 cm long., con tubo de 7 mm long. ensanchado en una garganta acampanada de 5 mm diám., y labios de 7-8 mm long., el posterior brevemente bidentado, el anterior ampliamente trilobado erec- Estambres con filamentos de 6 mm p la inferior oblicua, algo Material seleccionado. ARGENTINA. Jujuy: Santa Bárbara, Palma Sola a El Fuerte, ca. 5 km, Kiesling 5562 (SI); Ledesma, Calilegua, toma del Arroyo del Medio, Ca- brera 31312 (SI); Valle G rande, Ledesma a Valle Grande, Meyer 16416 (LIL). Salta: Orán, — io, camino a la Cruz, Abbiatti y Claps 221 (LIL); Metán, Metán frente al pueblo, 2 km W de la ruta 9/34, Novara 5826 (M, SI); Rosario А la Frontera, El Naranjo, O'Donnell 5368 (LIL); Rosario de — — — Venturi 8121 (K, SI). — del Est Pellegrini, C° del eins, Ven- turi 585: “Tu — teas . Oda. de la Hui- guera, 55 la Sota 96 (LIL); Famaillá, Oda. de FA Río Lules, Herrera 232 (LIL); Tafí del Valle, Yerba ICON Venturi 88 (SI); Burruyacu, El Timbó, Venniri 2835 (LH 19. Justicia kuntzei Lindau, Bull. Herb. Boiss. 3: 483. 1895. TIPO: Bolivia. Santa Cruz: 1000 m, V 1892, Kuntze s.n. (B destr.; isotipo, NY!). Lophothecium boliviense Bremek., Verh. Kon. Ned. Akad. tensch. Afd. Naturk., Tweede sect. C, 72(4): 426. 1969. TIPO: Bolivia. — on kio dedi nis oil re- finement and commencement of the pipeline, 2500 ft., 4 Sep. 1949, Brooke 5586 (holotipo, BM: isotipo, NY!). Sufrátice o arbusto de hasta 1 m alt. con ramas cilíndricas, lisas, pubérulas en su juventud, gla- brescentes y engrosadas por encima de los nudos a la madurez. Hojas con pecíolo de hasta 1.5 em long. y lámina elíptica de 6-12 em long. por 2.5— 5 cm lat. (la de las hojas apicales menor), aguda o acuminada en el ápice, cuneada y decurrente en la ase, esparcidamente pubérula en el haz y sobre la nervadura en el envés. Flores sésiles, con 2 brac- téolas en la base, en las axilas de brácteas lineales y pequefias, agrupadas en espigas laterales reuni- das en racimos simples o compuestos, laxos, pa- niculiformes, generalmente axilares y más largos que las hojas. Brácteas de las flores de 1.5-2.5 mm long.; bractéolas menores. Cáliz de 3—4 mm long. con los 5 segmentos lineales, agudos, glabros. Co- rola lilacina manchada de blanco en la fauce, de 1.5-2 em long., con tubo basal de 3—4 mm long., garganta obcónica e inflada de 5 mm long. por 4 mm diám., labio posterior entero, de 10-12 mm long. por 6-8 mm lat., el anterior trilobado, de 8— 10 mm long. por 5-6 mm lat., con lóbulos de cerca de 1 mm long. por 2 mm lat. Filamentos de 10 mm long.; anteras con las tecas superpuestas, separadas por el conectivo de 1 mm long., la teca inferior basalmente apendiculada, con un apéndice glan- dular claviforme de 0.5 mm long. Estilo glabro; es- tigma levemente engrosado; disco sinuado. Cápsu- las de 1.5 em long. por 3—4 mm diám. con la mitad inferior sólida у lateralmente comprimida y la su- perior engrosada, de contorno oblongo, estrechada en su parte media. Semillas 4 o menos por aborto, Volume 89, Number 2 2002 Ezcurra 257 Justicia en Sudamérica Austral orbiculares, oscuras, de 1.5-2 mm diám., tuber- culadas. Ilustraciones. Ezcurra, 1993a: 355. Distribución, hábitat y fenología. | Bolivia y N a Argentina, donde habita en el sotobosque de de la selva en serranías de 600 a 1500 m de elevación. Florece principalmente en primavera, de agosto a noviembre. Justicia kuntzei se caracteriza por las hojas gran- des y las inflorescencias multifloras de flores me- dianas con corola lilacina manchada de blanco en la fauce ARGENTINA. Jujuy: Lede "s- o bosque alto, Venturi 5217 (SI, К); 40 (SI); Ledesma, 10 km NE 7. — Salta: Santa 2007 (SI); Anta, Material seleccionado. ma, Sa. Calilegua, baj Capital, Zapla, Huan 306 of Ledesma, Eyerdam y Beetle 22275 (SI, Victoria, PN Baritú, Arroyo Sidras, Brown 2 PN El Rey, camino a Pozo Verde, Novara y Charpin 10344 (LIL); Santa Victoria, camino de Baritú a — Mar- mol 8734 (LIL); Orán, Tartagal, Schreiter 11370 (SI); San Martín, Dique Itiyuro, Stange s.n. (LIL) 20. Justicia laevilinguis (Nees) Lindau, Engler Bot. Jahrb. 19, Beibl. 48: 20. 1894. Rhyti- glossa laevilinguis Nees, in Martius, Fl. Bras. 9: 120. 1847. Dianthera laevilinguis (Nees) Durand & Jackson, Ind. Kew. Suppl. 1: 132. 1902. TIPO: Brasil. Rio Grande do Sul: Porto Alegre, Sellow s.n. (sintipo, B destr.); San Ga- briel, Estancia dos Fideles, Sellow s.n. (sinti- po. В destr.); sin loc., Sellow s.n. (probables isosintipos, K!, E!): Isla Santa Catarina, Twee- die s.n. (sintipo, K no visto); Argentina. Cór- doba: Río Segundo, Tweedie s.n. (sintipo, K!). кее laevilinguis E s var. longifolia Nees, in Mar- Fl. Bras. 9: 120. 1847. TIPO: Argentina. Bue nos еи Rosle s.n. — G herb. DC no visto, microficha!; probable isotipo. K!). quon anagallis in Martius, Fl. Bras. 9: 119. 847. Nees, 1 Justicia anagallis (Nees) — in Engler & ae Nat. Pflanzenfam. 4(3B): 1895. TIPO: Brasil. Bahia: Caitaité et Caxoeira, Mart ius s.n. (sin- tipo, M!); entre Porto Alegre y Montevideo, L Ballons s.n. (sintipo, В destr., fot. F 88081): — San os llow s.n. (sintipo, B destr Rhytiglossa repens Nees, in Martius, Fl. Bras. 9: 119. 1847. Ганна UND p or in Engler & Р E Nat. Pflanzenfam. 4, Abt. 351. 1895. Sellow La TIPO: Brasil. (бойро, " фан, fot. К m T Rhytiglossa obtusifolia Nees, in Martius, Fl. Bras. 9: 120. isticia — (Nees) Lindau, Bull. Herb. Boiss. 2, 1903. TIPO: Brasil. Rio — do Sul: dia 2s gre; Sellow s.n. (sintipo, B destr. Arg жашай Buenos Aires, herb. Arnott s.n. (sint "uius оннара Nees var. hirsuticaulis Nees, artius, Fl. Bras. 9: 120. TIPO: Brasil. Río Negro, Riedel s.n. (holotipo, LE no visto). ше — S. Moore, Trans. Linn. Soc. 2, Bot. 132. )5. Justicia m (S. Moore) V. A. E uin Kew Bull. 43(4): 600. 1988. TIPO: Brasil. Moor re — eee бт Isotipos, МҮ! B destr.. fot. )!). MEM a gaming Rusby, Mem. New York Bot. Gard. . 19 TIPO: Bolivia. Rusby 1421 (holotipo, Y! Justicia ЕРИ —— Not. Mus. La Plata 13, Bot. 12: 92. 1948. TIPO: Argentina. Buenos Aires: Tigre, D d = aM LP no visto). Hierba palustre, con rizomas horizontales grue- sos y tallos ascendentes, simples o ramosos, radi- cantes en los nudos inferiores, subtetrágonos, cua- drisulcados, de hasta 80 cm alt. y 5 mm diám., glabros, a veces algo pubescentes. Hojas superiores sésiles. las inferiores con pecíolo de hasta 3 mm long.. oblongas a lanceoladas, a veces lineales, de 3=12 (0.2—)0.5-1.5(-2.5) em lat., tusas en el ápice, redondeadas en la base, enteras em long. y ob- o levemente crenadas, generalmente glabras. Flores en espigas unilaterales terminales o axilares de 5— 12 em long., con pedúnculos de hasta 8 cm long. y entrenudos de ca. 1 cm long.: brácteas triangu- lares de 1.5-3 mm long. por ca. 1 mm lat., y brac- téolas triangular-lanceoladas de 1 mm long., ambas tlabras. Cáliz 5- ceolados de hasta 1 mm long. y 0.25 mm lat., agu- 1.5-2 em long. con tubo partido con segmentos lineal-lan- dos. Corola lilacina, de basal de 6—10 mm long. por 3 mm diám., más an- cha en la garganta: labio posterior ovado, de ca. ! cm long. y ca. 4 mm lat., obtuso a bidentado, v labio inferior dá 13 mm long., lóbulos de 6 mm long. y ca. 5 algo más ancho y con dibujo pectinado-rugoso y trilobado. con los mm lat., el del medio con manchas en la fauce. Estambres exertos, inclu- sos debajo del labio posterior, con las anteras su- perpuestas de menos de 2 mm long.. separadas por el conectivo. Cápsula de ca. 2 cm long. y 6 mm lat., estipitada, muy aplanada en la porción supe- rior, apiculada y glabra. Retináculos de ca. З mm Semilla suborbicular, ong., bidentados. de ca. 5 mm diám., muy aplanada, glabra, con el margen delgado y lacerado. Ilustraciones. Dawson, 1979: 575. Distribución, hábitat y fenología. Colombia, Venezuela, Perú, Bolivia, sur de Brasil, Uruguay, noreste de Argentina y Paraguay. Habita en am- bientes palustres de regiones bajas. En el sur de su distribución florece en primavera y verano, de octubre a abril. Justicia laevilinguis es una especie higrófila muy variable en su morfología, con una distribución muy amplia en América del Sur. Presenta especies estrechamente relacionadas en América del Norte 258 Annals of the Missouri Botanical Garden como Justicia ovata (Walter) Lindau, que se ase- meja en la morfología de tallo, inflorescencia y flor, y cuyas relaciones deberían estudiarse. ARGENTINA. Buenos Aires: Material seleccionado. Ribera — 1 del Río de la Plata, Hicken 12911 (SD; La Plata, 7 — s.n. (К, SI). Corrientes: Saladas, Arbo y Tur 2 22 214 (LIL); San Luis del Palmar, Ayo. Riachuelito, 25 Km E de * Luis del Palmar, Quarín 367 s AL); Con- cepción, Ybarrola 324 (LIL); Gral. Paz, Gral. Paz, Ybarro- la 3504 (LIL); а, RN 12, 20 km de н һаста el E, Zuloaga 3230 . Chaco: Las Palmas, — 2335 (LIL); Resistencia, Margarita Belén, Meyer 379 (а T Puente de la Plaga, PN Chaco, Pajonal, Milgura 9 (SI). Entre Ríos: Uruguay, Concepción del Uru- guay, Arroyo La C hina; campos vec inos, Bacigalupo 1376 SI); ¢ à i, Meyer 10281 (LIL). Formosa: Pilcomayo, Estanci: ia Bouvier, riacho Araguay, Guaglianone 454 (SI); Pirané, Pirané, Morel 161 (LIL). isiones: Posadas, riberas del río, — ise 1168 (LIL); Coe. Pindapoy, Bertoni 243. s Santa Fé: Gral. Obligado, Villa Ana, — vg IL); San a Pto. Cabot. Morello SI); Las Colonias, Esperanza, Ruiz Huidobro — (LIL); Garay, Santa Rosa, Ruiz —* 3348 (LIL). I AGUAY. Alto Paraguay: Alto „ Chaco, PURO 1269 (К, М); Руй Casado (С hac o) Кој 2199 (SI). Amamb jay: in m in regione c oris f Apa, Hassler 816€ ) (K, W), 8160a (BM. G, regione cursus — шн — — 81604 (BM, G, NY). Ca zü: cerca y a ` de Yhú, junto al arroyo Yhú, Fernández Casas et al. Р (С. МО, NY) Canindeyú: in regione vicine — in regione Yerba- lium de Marac :ayú, Hassler 4889 (BM, G, NY, P, W). Cen- tral: L Assumption, sur le bord des marads, Balansa 2449 P); Man 5 del río Salado, camino de Limpio a Em- Arbo et al. 1657 (CTES, ; Patino, near Asun- м: Teague 572 (BM). Н on: Pr. Concepción, ену» r 7624 (BM, G, NY), Hassler 7560 (BM, G, К, МО, ЫР) Zw río Apa und río Aquidabán Centurión, Fiebrig PA 30 (BM, E, G, K). Cordillera: Cordillera de Altos, Fie- brig 356 (B ‚ К, M, US, W); Cordillera de Altos, in campis, Hassler 3508 (BM, ( К, NY, Р, W); Cordillere de Peribebuy, dans les ао s ——— Balansa 2451а (G, Р); San Bernardino, costa del Lago Ipacaray, — et al. 1510 (CTES, US); Tobaty, Arroyo Tobaty, Rojas 4847 (US). Guairá: “ meadows, Villaric: "d — (K, MO, NY, K. P). N cembucú: ра Tene a la boca del río Ber- mejo, Meyer 16102 (LIL). I ari: Loma Acahay, s.f., Chodat s.n. (G); Caballero, nius 435 (NY); Cerro Mba- tobf, forest along N slope, in rege neration, Zardini y Ve- lásquez 9935 ( 3M). Presidente Ha 164 de la ruta Trans e Schin, n а 1031 ‚ Mereles pris (C TE S. С). San Pedro: in inen ida vicine San Estanislao, Hassler 1 { San Es stanislao, Rio Tapira- cuay, Krapovickas et al. 13896 (CTES, US). S. « ATA occidentalem flum. Paraguay, Gran Chaco, Rojas 930 (С). Jerónimo, Riacho Coronda, frente O36 In minis = = ‚ке 21. жн lilloana Ariza Espinar, Kurtziana 6: 1971. TIPO: Fértil, campo de Ischialasta, 17 Kurtz 14223 (holotipo, CORD!). Valle 1907. San Juan: Ene. Argentina. $ Arbusto bajo, de hasta 0.50 m alt., con ramas erectas, cilíndricas, las jóvenes verdosas y tetrá- gonas y cuadrisulcadas, con estrías marcadas entre las costillas, pubescentes en los nudos, de corteza cenicienta a la madurez. Hojas cortamente pecio- ladas, angostamente ovadas a elípticas, de 0.5-1.5 cm long. por 0.3-0.5 cm lat., obtusas, cuneadas en a base, generalmente pubescentes. Flores sésiles o cortamente pediceladas, con dos bractéolas en la base, solitarias en las axilas de brácteas, formando en conjunto espigas laxas; bractéolas lanceoladas de cerca de 1 cm long. por 2 mm lat.; brácteas obovadas de ca. 9 mm long. por 4 mm lat., ambas glandulosas y ciliadas. Cáliz profundamente 5-par- tido, con los lóbulos angostamente ovados a lan- ceolados, subiguales, de 4 mm long., foliáceos, ge- neralmente glabros. Corola azul violáceo, lilacina a blanca de 2 cm long., pubérula, con tubo corto de cerca de І em long., ensanchado en una garganta de aproximadamente 0.8 em diám., los labios de 1— 1.3 em long., el anterior trilobado hasta la mitad, con la fauce blanca, convexa y transversalmente rugoso-venosa. Estambres insertos en la base de la garganta, las anteras con las tecas subparalelas, una por encima de la otra, la inferior generalmente apendiculada en la base. Cápsulas claviformes, muy fuertemente comprimidas, de 1.1—1.5 ст long. por 5—7 mm ancho, glabras y de brillo nacarado en su superficie, con el cuarto inferior sólido y late- ralmente comprimido. Semillas generalmente 4, su- borbiculares, aplanadas, de 6—7 mm diám., lisas; retináculos obtusos. Ariza-Espinar, 1971: 85. Ilustraciones. Distribución, hábitat y fenología. | Noroeste de Tucumán y Catamarca la Argentina, desde Salta, hasta La Rioja y San Juan. Se la encuentra en re- giones áridas entre 1300 y 2600 m s.m., en montes xerófilos bajos, sobre suelo rocoso. Florece en ve- rano, de diciembre a abril. Nombre vulgar. “Alfalfillo” (Ariza-Espinar, Justicia lilloana se caracteriza por sus hojas pe- queñas y sus inflorescencias glandulosas, sus cáp- y muy fuertemente comprimidas de sulas grandes y brillo nacarado, y sus semillas superficie glabra ' lisas. El nombre Justicia anchas, muy aplanadas y platicarpa Lillo es un nomen nudum que se utilizó para esta especie (Lillo, 1937). Material seleccionado. ARGENTINA. Catamarca: Belén, Los Nacimientos, Cabrera 16761 (CORD). La ۰ ja: Gral. Lavalle, Parque de Talampaya, Hunziker, J. Gamerro 11669 (51); Independencia, e — Pagancillo y El Balde (ruta Villa Unión a Patquía, km 96/97), Hunziker, 22441 (SI); Independencia, entre Puerta de Talampaya Volume 89, Number 2 2002 Ezcurra 259 Justicia en Sudamérica Austral Gualo, Kurtz 13260 (CORD). San Juan: Ischigualasto, Roig 8053 (CORD): Iglesia. Cuesta del C^ Negro. entre Palasto y Las Caldas Ruiz a Man (LIL). Tucu- s dg Tafí. Amaicha, Diers 225 (SI); Trancas. Lomas de Arcas. Schreiter 1105 (CORD. IL) 22. Justicia lilloi (J. L. Lotti) C. Ezcurra. in Ca- brera. Fl. Prov. Jujuy (República Argentina) 9: 349. 1993. Chaetochlamys lilloi J. L. Lotti. Publ. Especial Inst. Lillo: 63, fig. 1 y lám. |. 1976. TIPO: Argentina. Salta: Capital. 600 m SMi; Lillo 8087 (holotipo, LIL). Hierba sufruticosa, de 0.50-1.50 m alt.. con ta- llos subcilíndricos. cuadrisulcados, erectos. pubé- rulos. Hojas con pecíolo de 0.5-2 cm long. y lám- ina ovada o elíptica de 5-9 cm long. por 2-1 cm lat., aguda en el ápice y cuneada o redondeada en la base, verde clara, pubérula, con cistolitos den- samente esparcidos en su superficie. las venas principales prominentes en el envés. Flores sésiles en dicasios condensados y generalmente reducidos a una flor con dos bractéolas, agrupados en cortas espigas con brácteas foliosas. Espigas axilares у terminales, con el pedúnculo corto y el raquis re- ducido a uno o dos entrenudos: brácteas muy si- milares a las hojas, pero sésiles у algo menores. у truncadas o cordadas en la base: bractéolas linea- les. de hasta 15 mm long., pubescentes у corta- mente ciliadas. Cáliz profundamente 5-partido. con los segmentos ovado-lanceolados. de 10- 12 mm long. por 2-3 mm lat.. acuminados y densa rígidos. y cortamente ciliados, con el nervio medio promi- nente en el envés. Corola lilacina de 34 em long.. con el tubo de 2 em long. por 3—4 mm diám.. poco ensanchado en la garganta. los labios de cerca de 1.5 em long., el posterior de 7 mm lat.. bidentado. el anterior profundamente trilobado. con los lóbulos de 1.2 em long. por 6-8 mm lat. Estambres con filamentos de 12 mm long. y anteras con las tecas superpuestas y algo separadas por el conectivo. la superior de 2.5 mm long.. la inferior de 3 mm long. y apendiculada en la base. Estilo glabro con estig- ma levemente engrosado y bilobado: ovario glabro Cápsulas de v disco en forma de copa. ondulado. cerca de 1.5 em long. v 6 mm diam. con la mitad inferior sólida y lateralmente comprimida y la su- perior engrosada, ovoide y I-seminada. Semillas or- lisas. biculares subesféricas. de 2-3 mm diám., pardo-oscuras: retináculos de 3 mm long. Ilustraciones. Lillo. 1937: lám. 7 y 8 (2) (sub Chaetochlamys tucumanensis). Ezcurra. 1993a: 350. Distribución. hábitat y fenología. Sudeste de Bolivia. oeste de Paraguay y norte de la Argentina. Habita principalmente en facies húmedas y bajas del Chaco Serrano septentrional y en bosques de transición a las selvas montanas de las Yungas en- tre 500 y cuentra en laderas de cerros del Chaco Paraguayo. 1200 m de elevación. También se en- Florece de fines de primavera a otoño, de diciembre a mayo Chaetochlamys tucumanensis Lillo es un nomen nudum con el que originalmente se denominó i sla especie (Lillo, 1937: Lotti, 1976). Justicia lilloi es una especie muy afin a Justicia thunbergioides (Lindau) Leonard, descripta para Mato Grosso do Sul. Brasil (Brasil. Mato Grosso pr. Corumba. Mal- me 3026. sintipo В destr.. fot. F 89191). mites entre ambas no son claros (Ezcurra, "1993b). v los lí- Junto con Justicia allocota Leonard (Brasil. Goiás: Yale Dawson 15023, holotipo, US!) y Justicia ve- nusta (Rizzini) V. A. W. Graham (Brasil. Minas Ge- Brade 17563, isotipo. US!) forman un grupo de diferenciación compleja que parecen todas de- rais: rivadas de un antecesor común y adaptadas a di- ferentes regiones con distintas condiciones de hu- medad y precipitación. Justicia lilloi es la que presenta generalmente hojas más chicas y se en- cuentra principalmente en las selvas pedemontanas de las Yungas de Bolivia y Argentina y en serranías del Chaco Paraguayo, mientras que Justicia thun- bergioides y Justicia allocota se encuentran en re- giones más húmedas del sudoeste y centro de Bra- sil respectivamente. Material seleccionado. ARGENTINA. Catamarca: El Alto. entre Alijilán y El Alto, Cristóbal 340 (LIL) нута Santa — Abra de los Morteros. Cabrera 21719 (S San Pedro, San Pedro a El Milagro entrando hacia 2 M лени 5372 (LIL). Salta: Rosario de la Frontera, halle 4417 (LIL): Ор he Vie 03. Lomas al N — $532 (SI): Orán, Cerro Tartagal. / : Metán. Metán. Roc ha 108 (LIL J [iners Toledo 21 200 m antes de Vipos, | Vereoorst 1221 (SI): ( Barranca Colorada. Venturi 1150 de la finca Guillermina, camino a — Medina, Villa Carenzo 1750 (V. AGL "araguay: Cerro Pedrera. Rio Navileque. Fuerte Olimpo, — 20425 (С); Fuerte Olimpo, Cha- s 13842 (LIL): Fuerte Olimpo. in co- Е — Bernardi 20335 (©. desde lomada al S (cam- co Paraguayo. linis Tres Marías. MO). Chaco: Chaco. Cerro León, ы е. hasta mesela J entral, Charpin y Ramella 2173 (P) p-p- (G). S. dep.: Paraguay. "кады ALIZ 23. Justicia lythroides (Nees) V. A. W. Graham. Kew Bull. 43: 603. 1988. Heinzelia lythroides Nees. in Martius, Fl. Bras. 9: 154. 1847. TIPO: Brasil. Rio de Janeiro: 7 rahy.” Pohl s.n. (holotipo. W!). ad fluvium Pi- Heinzelia ovalis Nees. in Martius. Fl. Bras. 9: 154. 18 260 Annals of the Missouri Botanical Garden non — ovalis Ridley, J. Fed. Malay States Mus. 10: 150. 1820. TIPO: Brasil. Sáo Paulo: Sáo Paulo. Ypanema, in d Riedel 1983 (lectotipo, aquí de- signado, LE, fotogr. US!). Hierba perenne, rizomatosa en la base, con tallos tenues, subcilfndricos, decumbentes y geniculados, luego erectos, de hasta 40-70) em long., pubéru- los. Hojas opuestas con pecíolo de 3-10 mm long.. y lámina elíptica a angostamente ovada de 3—8 cm long. por 1.3-3 em lat., aguda o acuminada en el ápice, cuneada y algo decurrente en la base, entera, generalmente glabra, a veces pubérula sobre el ner- vio medio en el haz, verde clara y pubescente sobre la nervadura en el envés. Inflorescencias formadas por flores en espigas unilaterales tenues, simples o compuestas y ramificadas, de 4-8 cm long.. ter- minales o en las axilas de las hojas superiores, cor- tamente pedunculadas, con el eje arquado. Brác- teas y bractéolas lineal-lanceoladas, rígidas. de 1— 3 mm long., pubérulas, con el nervio medio pro- minente en el dorso. Cáliz de cerca de 6 mm long.. con los 4 segmentos de 5 mm long., rígidos, linea- les, subulados, dorsalmente pubescentes sobre el nervio medio. Corola blanca, pálido lilacina o vio- lácea de ca. 1 em long., con tubo de 0.5-0.6 cm long. por 1.5 mm diám., ampliado en la garganta, el labio superior erecto, lanceolado, de 3 mm long., agudo, el inferior ancho, trilobado, de ca. 4 mm long., con los lóbulos de ca. 2 mm long., redon- deados. Estambres con filamentos de aprox. 4 mm long.; teca apical perfecta, de 1 mm long., la in- ferior estéril, reducida a un leve engrosamiento so- bre el filamento. Cápsula de contorno angostamente elíptico, de aprox. 5 mm long. por 1.5 mm lat., cortamente sólido-estipitada en la base, glabrescen- te y de paredes finas. Semillas 4, algo aplanadas, de 1 mm diám., pardo-oscuras, con pelos cortos y gloquidiados; retináculos de | mm long., obtusos. Ilustraciones. Nees, 18; (a: Acosta-Castellanos, 1907: tab. 97. Ezcurra y Distribución, hábitat y fenología. Sur de Brasil. noreste de Argentina y Paraguay oriental. Habita en el sotobosque y en bordes de bosques htimedos, frecuentemente en lugares alterados. Florece en otofio e invierno, de abril a agosto. Justicia lythroides se caractriza por sus flores en espigas secundifloras tenues y delicadas y sus flo- res pequefias. Es muy afín a la especie descripta bajo el nombre de Chaetothylax hatsbachii Leonard (Brasil. Paraná: Cerro Azul, Turvo, Rio Ribeira, Hatschbach 5419, holotipo US!), que podría ser si- nónimo de esta especie. El nombre Heinzelia ovalis fue creado por Nees sobre la base de tres ejemplares: dos de Brasil (Sâo Paulo, Ipanema, Riedel 1983, y Minas Gerais, Bar- bacena, Riedel 88, LE) y uno de Argentina, Tucu- man (Tweedie s.n., K!). Estos sintipos comprenden una mezcla de dos especies. El ejemplar coleccio- nado en Argentina por Tweedie pertenece a Justicia goudotii, mientras que el material de Brasil colec- cionado por Riedel, del que he visto una — y un probable isosintipo (Brasil. Sáo Paulo, sylvula ad vias pr. Ypanema, 111-1834, Riedel s.n., NY!) pertenece a Justicia lythroides. Debido a que a diagnosis de Heinzelia ovalis coincide más con as características del material de Riedel que con el de Tweedie en tamaño de hojas y flores, designo como lectotipo de Heinzelia ovalis al ejemplar de Justicia lythroides de Brasil, Ipanema, Riedel 1983 (LE). Esta designación hace que el nombre Hein- zelia ovalis deba considerarse sinónimo de Justicia lythroides. Material seleccionado. ARGENTINA, Misiones: El- dorado, Eldorado, km 28, Burkart 14603 (SI); Frontera, reforestaci ‘ión Gral. М dez 144 (CTES): Cainguás, Arroyo Caña-Pirú, 8 km de A del Valle hacia Jardín de América, Hunziker, J. 10891 SI); San Ignacio, Ayo. Macaco, Schwarz 5687 (LIL); Mon- ; 3‹ : a 2188 (SI) Alo: Par araná Fer 6040 (G); in regione fluminis Alto Parana, Fisbrig 6173 (G, SI); Río Itabó, sotobosque, ribera n río, пай к 1001 (МО). Caazapa: Tavaf, Castor Cue, S, 55°20'W, Degen 1611 (MO); Tavat, Compañía ко. — 1522 (G, MO), Ortiz 1280 (G, MO). Canindeyú: Mbaracayú, comunidad Mbyá, Basua- lado 2546 (FCQ). 24. Justicia mandoni (Lindau) Wasshausen & C. Ezcurra, Candollea 52: 175. 1997. Beloperone mandoni Lindau. Bull. Herb. 5: 615. 1897. TIPO: Bolivia. La Paz: Larecaja, in vi- cinity of Sorata, Cerro Iminapi, Mandon 297 (holotipo, B destr.; isotipos, BM! p.p., K! p.p.). Boiss. Justicia — De * 0 & Ruiz, Publ. Espec. Inst. Lillo: 53, fig. 3 y láms. 5 y 6, 1976. TIPO: n ш. Chicli e) Puesto Las Pavas, 1060 1., Meyer 15125 (holotipo, LIL!). Sufrútice o arbusto apoyante de 1-2 m alt., con ramas cilíndricas, lisas o muy finamente estriadas, generalmente glabras. Hojas con pecíolo de 0.5-1 em long., pubérulo, y lámina angostamente ovada, de 4-8 cm long. por 1.2-2.5 em lat., redondeada y levemente decurrente en la base, acuminada en el ápice, levemente pubescente en el envés, pubérula en el haz, con nervaduras marcadas. Flores sésiles con dos bractéolas basales en las axilas de brácteas dispuestas en espigas generalmente terminales, las brácteas opuestas pero alternadamente anisofloras, dándole un aspecto unilateral a la espiga. Brácteas Volume 89, Number 2 2002 Ezcurra 261 Justicia en Sudamérica Austral 2.0 foliosas, pubescentes: bractéolas lineal- elíptico-lanceoladas, de 1-1.5 em long. por 1.5— mm lat., lanceoladas, de 1-2 mm lat., algo más cortas que el cáliz. Cáliz profundamente 5-partido, los seg- mentos — ovado-lanceolados, de 1.5-2 ст long. por 2-3 mm lat., generalmente glabros a pu- bérulos. — rojo-morada de 3-3.5 em long.. el tubo de aprox. la misma longitud que los labios v de 5 mm diám.. interiormente surcado en la parte posterior por dos pliegues longitudinales que alojan el estilo, el labio posterior bidentado, de 1.5-2 cm long. por aprox. 8 mm lat., el anterior trilobado, de 7 mm lat., reticulado-venoso y convexo en la base. con los lóbulos de aprox. | em long. por 3—4 mm lat. Estambres con filamentos de 2 em long. v tecas situadas a distinta altura, la superior de 2 mm long., mutica. la inferior separada del conectivo. algo más grande y apendiculada en la base. Cáp- sula de contorno oblongo, de 1.5-2 cm long. 0.5 aplanadas de 2 mm diám., verrugosas; retináculos por em didm., pubescente. estipitada. Semillas de 2 mm long. Ilustraciones. Lillo. 1937: lam. 6 (sub Justicia tucumanensis). De Marco y Ruiz, 1976: 54. 60 y Ol (sub Justicia odonellit). Ezcurra, 1993a: 337 (sub Justicia odonellii). Distribución, hábitat y fenología. Sudeste de Bolivia y Norte de la Argentina, en las provincias de Jujuy. Salta v Tucumán. Habita en las selvas de la provincia fitogeográfica de las Yungas en serra- nfas entre 1000 y 1700 m, generalmente apovada sobre otros arbustos o árboles. Florece en otoño. invierno y primavera, de febrero a septiembre. Justicia mandoni es un arbusto apoyante de sel- va que se caracteriza por las grandes flores rojo- moradas en espigas de aspecto unilateral. Se ase- meja a J. dumetorum, pero esta última es de porte menos elevado, sus inflorescencias no son unilate- rales, y sus flores tienen labios más cortos en re- lación a la longitud del tubo basal. Por un error involuntario, recientemente se trató a J. mandoni como especie endémica de Argentina (Ezcurra. 1999), a pesar de que su distribución abarca el sur de Bolivia. Justicia tucumanensis es un nomen nu- dum con el que originalmente se denominó a esta especie en Argentina (Lillo, 1937). ARGENTINA, Jujuy: Capital, Material selece we le antena, Cabrera Хара. Mina 9 de Octubre, 5 Calde ra, ca. Abra de Ruta 9. km 1 > Abra de Santa Laura, Novara 8055 (M. 51); Santa p Чопа, Camino de Toldos al Lipeo, a. 20 km de Toldo ате y Te 8163 (LIL): Anta, campo Las Heras, — 966 (LIL); Rosario de la Fron- Venturi 9338 (К). а Tafí, quebrada de los subida a la brera 3413. Caldera, N de Ojo de с 5 km < ^ lera, Arroyo Azucena, 10 1079 (St): Chicligasta. Piscicultura, Meyer 13853 (LIL): Barravacu, Barravacu, Türpe 40 (LIL); le y Villa Nogués, Venturi 1799 (SI). Sosa, 25. Justicia oblonga (Nees) Lindau. in Engler & Prantl, Nat. (3b): 350. 1895. atada oblonga Nees, in Martius, Fl. Bras. 9: 124. 1847. TIPO: Brasil. Rio Grande do Sul: Porto Alegre, Sellow s.n destr.. fot. F 8847): San Lorenzo, Sellow s.n. (sintipo, B destr.): sin Pflanzenfam. (sintipo. B Uruguay. Santa Lucía у loc., Sellow s.n. (probable isosintipo, K!). Hierba perenne, rizomatosa en la base, con tallos erectos, subcilindricos, de hasta 60 em long.. bérulos. Hojas opuestas con pecíolo de 0.2 long. y lámina oblongo-elíptica de 2-5 em lace por ].2—9 a cm lat., acuminada en el ápice, cuneada algo decurrente en la base, entera, coriácea, pu- bérula en el haz, pubescente sobre la nervadura en el envés. Inflorescencias formadas por espigas den- — sas, sésiles, terminales o en las axilas de las hojas superiores, a veces formando racimos de espigas en la porción terminal de los tallos. con las hojas re- ducidas. Brácteas lanceoladas. rígidas. de 5-7 mm long.. pubérulas, ciliadas. Cáliz de cerca de 5 mm long.. con los 4 segmentos rígidos. lineales. subu- lados, dorsalmente pubérulos, ciliados. Corola pá- lido lilacina a violácea de ca. | em long., con tubo de 0.5-0.6 em long., el labio superior erecto, de 3.5 mm long. por | mm lat., el labio inferior ancho. trilobado, de ca. 3 mm long., con estrías blancas en la fauce. Estambres con filamentos de aprox. 2 mm long.: teca apical perfecta, de contorno ovado. de ca. | mm long., la otra inserta más abajo en el filamento, más pequeña, angosta y basalmente cau- dada. Cápsula de contorno obovado, de aprox. 8 mm long. por З mm lat., cortamente sólido-estipi- tada en la base, glabrescente y de paredes finas. Semillas 4, eloquidiados: retináculos de ca. de ca. | mm diám., con pelos cortos у | mm long., se miobtusos. Ilustraciones. Ezcurra y Acosta Castellanos. 1997: 109 Distribución, hábitat y fenología. Sur de Brasil. Uruguay, noreste de Argentina y Paraguay oriental. Habita en ambientes húmedos. Florece principal- mente en verano y otoño, de noviembre a abril. Justicia oblonga se caracteriza por sus flores de aprox. | em long., en espigas densas con brácteas pequeñas. Por las características de sus inflores- cencias, anteras y sus cápsulas y semillas induda- blemente pertenece a Justicia sect. Chaetothylax Nees) V. A. W. Graham (Tabla sus corolas poseen un tubo corto, lo que no es muy 1). a pesar de que — 262 Annals of the Missouri Botanical Garden frecuente en esta sección (Ezcurra & Acosta-Cas- tellanos, 1997). Material se s ccionado. ARGENTINA. Misiones: campos, Burkart 14443 (Sl); Can- odriguez 288 (Sl); San Javier, San Javier, Krapovickas 15161 (CTES, SI); Leandro Paso Carreta, Krapovickas 15021 (CTES, SI); Candelaria, Berón Astrada, Yabebyry, Montes 841 (SI); San Ignacio, ico, Schwarz 5706 (LIL); Cainguás, ruta 1 ^i Km 952. : — 3065 (LIL). Corrientes: Ituzaingó, | z de Ruta 12, camino a San Carlos, CTES); Santo Tomé, Estancia Garruchos, potrero Puente, Krapovickas 21611 (CTES): Santo Tomé, Ea. Timbó, Avo. Ciriaco у Ruta 40, Schinini 23450 (CTES, LIL). PARA- GUAY. Alto Parana: in — fluminis Alto Parana, Fiebrig 5929 (BM, G, K, LIL, SI); Хас iei Montes 9806 ; Nacunday, Mantes 1093 3 (L P). Guaira: от Montes 16162 (LIL); Iti — Monies 15762 (LIL): Yegua Porá, Montes 16383 (LIL >? E 2 a — > = S © ч. = = = a > co m — — 26. Justicia oranensis De Marco & Ruiz, Publ. Especial Inst. Lillo: 46, fig. 1 y láms. 1 v 2, 1976. TIPO: Argentina. $ la Finca de Jakulica, a 8 km del puente del Salta: Oran, camino a Rio Bermejo, 600 m s.n.m., Legname y Cuezzo 8564c (holotipo, LIL). Arbusto ramoso de 1—3 m alt.. con ramas sub- cilíndricas, las jóvenes cuadrisulcadas, con dos bandas longitudinales pilosas a lo largo de los en- trenudos. Hojas con pecíolo de 1—4 cm long. y lám- ina ovada o elíptica de 5—12 cm long. por 3-8 em lat., aguda y levemente acuminada en el ápice, cu- neada en la base, parcialmente decurrente sobre el pecíolo, con las venas principales marcadamente prominentes y notorias en el envés, las secundarias arqueadas y paralelas, la nervadura en general pu- bescente. Flores sésiles dispuestas en espigas den- sas y bracteadas que se agrupan en racimos o pa- nojas laxos, pedunculados, más largos que las hojas, en las axilas de las hojas superiores. Brác- teas florales obovadas, de 5-8 mm long. por 3-5 mm lat., redondeadas, corta y densamente glandu- loso-pubescentes; bractéolas oblongas, de aprox. 6 mm long. por 2 mm lat., obtusas, glandulosas. Cáliz profundamente 5-partido, con los segmentos lan- ceolados, de 6-7 mm long. por 1-2 mm lat.. agu- dos, levemente glandulosos. Corola rojo-anaranja- da, de cerca de 2.5 cm long., con tubo de 1.2 cm long.. ensanchado а 3—4 mm diám. en la garganta, y labios de 1 cm long., el posterior bidentado, el anterior trilobado, con los lóbulos de 1 mm long. Estambres con filamentos de 1.5 cm long., y anteras con tecas a distinta altura: la superior separada por el conectivo y oblicua, de 2 mm long., la inferior algo más larga, apendiculada en la base. Cápsula de 1.3-1.5 em long., estrechada y sólida en la mi- tad inferior, la porción superior inflada y 4-semi- nada, glanduloso-pubérula. Semillas suborbiculares de 3 mm diám., tuberculadas. Jaculatores de 2 mm long. Ilustraciones. De Marco y Ruiz, 1976: 47, 56 y 57. Ezcurra, 1993a: 341. Sur de Boliv- Argentina, en las provincias de Distribución, hábitat y fenología. ia y norte de la Jujuy, Salta y Tucumán. Habita en selvas bajas de la provincia fitogeográfica de las Yungas entre 500 y 1600 m, y florece en invierno y primavera, de julio a noviembre. Especie llamativa por sus pa- nojas de flores rojas, según De Marco y Ruiz (1976) es fácilmente cultivable y se puede propagar por estacas. Justicia oranensis es una especie llamativa por sus panojas de espigas de flores rojo-anaranjadas, y según De Marco y Ruiz (1976) es fácilmente cul- tivable y se puede propagar por estacas. Justicia schreiteri Lillo es un nomen nudum que se utilizó para esta especie (Lillo, 1937). ARGENTINA. Jujuy: Ledes- — та 27954 (SI); Ledesma, PN Cali- 9 (SI). Salta: Santa Victoria, camino de Los Toldos al Lipeo, Legname y Cuezzo 9707 (LIL); Tabacal, Martinez Cet 3376 (LIL ); — José de San Martín, Tartagal, 5 1, Cedral, Ro- driguez 1045 (LiL): — Vado ное жене 7043 ( (1 Ма terial seleccionado. ma, Caimancito, legua, Cabrera 3214 AL). Tucumán: sin loc., Baer 80 (SI). 27. Justicia pectoralis Jacquin, Enum. Syst. Pl. 11. 1760. Dianthera pectoralis (Jacq.) Murr., Syst. Veg. Ed. 14, 64. 1784. Stethoma pecto- ralis (Jacq.) Raf., Fl. Tellur. 4: 61. 1838 (1836). Rhytiglossa pectoralis (Jacq.) Nees, in London J. Bot. 4: 637. 1845. Ecbo- lium pectorale (Jacq.) Kuntze, Revis. Gen. РІ. 2: 487. 18 Bentham, 1 . Psacadocalymma pectorale (Jacq.) Bremek., Verh. Kon. Ned. Akad. We- Afd. Natuurk., 45: 55. 1948. TIPO: Ilustración en Jacquin, Selec. Stirp. Hist., designado ante la ausencia de ejemplar tipo, 1995). tensch., Tweede Sect. Amer. t. 3. 1763 (lectotipo, aquí ver Daniel, Hierba pequeña, con tallos ascendentes, radi- cantes en los nudos inferiores, de hasta 30 cm alt. y | mm diám., pubescentes. Hojas con pecíolo de 3—4 mm long., lanceoladas a ovado-lanceoladas, de 4-6 cm long. y 1-2 em lat., agudas en el ápice, redondeadas en la base, enteras o levemente cre- nadas, levemente pubescentes sobre las venas. Flo- res en espigas unilaterales terminales y axilares, simples (en el material de Argentina y Paraguay) o ramificadas, a veces agrupadas en inflorescencias paniculiformes, de 4-10 cm long., con pedúnculos Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral de 1.5-2.5 em long. y entrenudos de menos de 1 ¢ em long.; brácteas lanceoladas de 2.54 mm long. por 0.5 mm lat., y bractéolas lanceoladas de 2 mm Cáliz 4-partido con seg- mentos de 3.5 mm long. y 0.25 mm lat., pubérulos. Corola blanca de 0.8-1.2 cm long. con long.. ambas pubérulas. ( obtusos, tubo basal de 5 mm long. por 2.5 mm diám. en la fauce: labio posterior ovado, de 3 mm long. y ca. 2 mm lat., obtuso. y labio inferior de 5 mm long., trilobado, con los lóbulos de 3 mm long. y ca. 2 mm lat., el del medio algo más ancho y con man- chas rosadas en la fauce. Estambres exertos, inclu- sos debajo del labio posterior. Cápsula de ca. Т cm long., superiormente inflada y 4-seminada. Semillas orbicular-oblongas, aplanadas, de ca. 1 mm diám., pusticuladas, pardo-oscuras. Durkee, 1986: 12. Ilustraciones. Distribución, hábitat y fenología. América trop- ical hasta el Sur de Brasil, y Paraguay oriental y regiones limítrofes de Argentina. En el sur de su distribución habita en el sotobosque de bosques es- tacionalmente secos, generalmente sobre suelo are- noso. Florece en otoño y primavera, de septiembre a abril. El material de Justicia pectoralis de Paraguay y Argentina se diferencia de la mayoría de los ejem- plares de esta especie que habitan más al norte por el porte bajo y las espigas terminales generalmente simples y no agrupadas en inflorescencias panicu- liformes. Esta forma empobrecida podría deberse al efecto de una temporada favorable más corta por efecto de su crecimiento a una mayor latitud geo- eráfica, con veranos de menor duración. Esta mor- fología se asemeja a la descripción de Justicia sar- (Nees) Lindau basada en Rhytiglossa Sellow 56, holoti- mentosa sarmentosa Nees (Brasil. S. loc., po, B destr.). La ausencia de material tipo asociado a este epíteto hace dudosa su identificación. ARGENTIN A. Formosa: Ma- Material selece — tacos, Ing. Juarez, chacra mataca, 3 km camino al R. Ber- mejo, Maranta 295 | SI): Ing. Juarez Toldería Toba. 1 km al N del pueblo, Arenas 2283 (CORD, SI). PARAGUAY. Alto INE San Carlos del Alto Paraguay, Molas ). Boquerón: Chaco Paraguayo, ruta Trans Cha- со. Schinini y Palacios 25788 (CTES). Chaco: Mayor Pe- dro Lagerenza, selva ber del Río Timane, Schinini y Bordas 14938 (CTES, С). ieva Asunción: Teniente Primero A. Picco, Spic — et we 2298 (G). 28. Justicia phyllocalyx (Lindau) Wasshausen & Poikt- Lindau, Engl. Bot. 1989. TIPO: Brasil. Cocal, Glaziou Ezcurra, Candollea 52: 175. | e lacanthus phyllocalyx 1 Jahrb. 25. Beibl. 60: 48. Goyaz: inter As Brancas et 21876 (holotipo. B destr.. K!). fot. F 87851; isotipo. — cien macedoana um TÉ Jard. Bot. Rio Ja- iro 8: 357, t. 7 y 8, f. 3. . TIPO: Brasil. Minas — Uberlandia, el (holotipo. RB fot. US». СЕК тас — Rizzini var. — — for- ma redacta Rizzi Dusenia 3: 189. TIPO: ras al. inas Cau Ituiutaba, — — (ho- lotipo, RB: isotipo, US!). Hierba erecta, de 20-70 em alt. con rizoma en a base, y tallos poco ramificados, cilíndricos, gla- bros a pubescentes o cortamente tomentosos. Hojas cortamente pecioladas, anchamente elípticas, de 3— 6(-8) em long. y 1.5-3 cm lat., agudas, cuneadas. glabras a densamente pubérulas o cortamente to- mentosas. Inflorescencias en espigas terminales o axilares de 3-8 cm long., sésiles a cortamente pe- dunculadas, alternadamente anisofloras (una sola de cada par de brácteas portando una flor con sus dos bractéolas en la axila); brácteas lineal-lanceo- ladas, de 10-15 mm, más o menos glanduloso-pu- bérulas y seríceas; bractéolas lineales, de 8-10 mm long., pubérulas o glabras. Cáliz profundamente 5- partido, con los segmentos subiguales, elíptico-ova- dos, de 1.2-1.5 em long. por 3—4 mm lat.. agudos o acuminados, pubérulos o glabros, trinervados, con el nervio medio marcado. Corola blanquecina, de ca. 2 em long., con el tubo de 1.3 cm long., acam- panado, hasta de 0.7 ст diám. en la fauce, el ló- bulo posterior cóncavo, anchamente ovado, de ca. у 0.7 trilobado, reticulado-venoso en la fauce, con los ló- | em long. em lat., bidentado, el anterior bulos anchamente oblongos, de ca. 1 em long. y 8 mm lat., redondeados. Estambres exertos por de- bajo del labio posterior, las tecas marcadamente su- perpuestas, separadas por un conectivo ancho. la inferior apendiculada en la base. Cápsula desco- nocida. Ilustraciones. Figura 11. Distribución, hábitat y fenología. Sudoeste de Brasil, y noreste de Paraguav en la región del Amambay. Habita en cerrados, en bordes de bosque. Florece en verano y otoño, de enero a abril. Justicia phyllocalyx se asemeja a otras especies de Justicia sect. Simonisia como J. lilloi v J. rusbyi, por lo que se la ubica en esta sección (Tabla 1). Al igual que estas especies, posee polen biporado, pero соп ínsulas sobre toda su superficie, no sola- mente en hileras a los costados de los poros. En esto se asemeja a especies de Poikilacanthus. El polen 4- u 8-porado y la exina con ínsulas en toda su superficie que le dan un aspecto facetado es característico (Daniel. 1991. 1998), pero el polen de esta especie no coincide de Poikilacanthus totalmente con el de Potkilacanthus. A pesar de las ínsulas en toda su superficie, debe ser tratado en 264 Annals of the Missouri Botanical Garden Figura 11. Justicia por poseer 2 poros y por la disposición de las ínsulas, y por eso fue transferido a este género (Wasshausen & Ezcurra, 1997). Material seleccionado. 'ARAGUAY. Amambay: Ar- a (G); in re — cursus superioris fluminis Apa, Hassler 8423 (BM, G, K, LIL, , W); Parque Nacional Cerro Corá, open — NE of lie 'adquarters, Solomon 7055 (MO, PY), 7056 (MO); Sierra de Amam- bay, ad marginem silvarum pr. m lla, Rojas 10152 (G, MO). Justicia phyllocalyx. —A. Rama con flores. —D. Corola abierta con androceo. —E. Cápsula abierta sin semillas. Hassler 8423 (BM, MO), (51) —C. Corola. be urra 1831, 1840 —B. Nudo con bráctea, bractéolas у — 29. э polygaloides (S. Moore) Lindau, Jull. Herb. Boiss. Ser. 2, 3: 633. 1903. Dian- кук, polygaloides S. Moore, Trans. Linn. Soc. Ser. 2, 4: 433. 1895. TIPO: Brasil. Mato Gros- so: Santa Cruz, Moore 667 (holotipo, K!; iso- tipo, B destr., fot. F 8855!). Hierba pequefia, con tallos ascendentes, radi- cantes en los nudos inferiores, de hasta 30 cm alt. E y 1.5 mm diám., glabros o pubérulos. Hojas sub- sésiles o cortamente pecioladas, lanceoladas o an- Volume 89, Number 2 2002 gostamente oblongas, de 3-6 em long. y 0.4-1 cm at., agudas en el ápice, cuneadas en la base, en- teras, generalmente glabras. Flores en espigas ter- minales simples, de 3-8 em long.. con pedúnculos de 1.54 em long. y entrenudos de menos de | cm long.: brácteas lanceoladas de 1.5—4 mm long. por 0.5 mm lat., y bractéolas triangulares de 2-3 mm long., glabras o pubérulas. Cáliz 4-partido con seg- mentos de 3-5 mm long. y 0.25 mm lat.. acrescen- te. Corola lilacina de 1 em long. con tubo basal de 5 mm long. por 2.5 mm diám. en la fauce: labio ca. 2 mm lat.. posterior ovado. de 3 mm long. v obtuso, v labio inferior de 3 mm long.. trilobado. con los lóbulos de 2 mm long. y ca. 2 mm lat., el del medio algo más ancho y con dibujo pectinado- rugoso, blanco. en la fauce. Estambres exertos. in- clusos debajo del labio posterior, con las anteras con las tecas superpuestas de ca. 1 mm long.. la inferior oblicua. Cápsula de contorno elíptico. de 0.7 mm long. y 2 mm grosor, glabra. Semillas 4. aplanadas, de ca. | mm diám., diminutamente ve- rrucosas. Figura 12. Ilustraciones. Distribución, hábitat y fenología. Sudoeste de Brasil y Paraguay oriental. Habita frecuentemente en prados ribereños. Florece en primavera y vera- no, de septiembre a febrero. Justicia polygaloides se caracteriza por las hojas lanceoladas o angostamente oblongas y las flores de aprox. | em long. en espigas terminales simples. Es una especie relativamente rara, que ha sido co- leccionada pocas veces, en general en prados ri- bereños. El ejemplar de Argentina, Entre Ríos. pre- senta una distribución anómala, pero se asemeja al resto del material. seleccionado. ARGENTINA. Ríos: alneario e Tortuga Alegre, al N de Concor- dia, li 997 AGUAY. Alto dorm Puert to Casado, Rojas ne (SI). € oncepción: Pr. Con- . Hassler 7653 (BM, G, laterial Entre = > =ч “~~ layes: Gran Chaco, Loma Clavel. Rojas 2631 (BM. G К. NY): in regione cursus inferioris fluminis Pilcomayo, T 669 (G). San Pedro: Colonia Primavera. Woolston SI). 30. Justicia ramulosa (Morong) C. Ezcurra, Bol. Soc. Argent. Bot. 25: 350. 1988 (Octubre). Be- — ramulosa Morong. Ann. New York Acad. 7: 194. 1893. Justicia ramulosa — A. W. Graham, Kew Bull. 43: 604. 1988 (Noviembre). TIPO: Paraguay. Asunción: — 706 peor NY isotipos, BML El. , MO!, NYL KL Û Beloperone tetramerioides Lindau, Bull. Herb. Boissier 3: Ezcurra 265 Justicia en Sudamérica Austral 488. "E — е AN Andau) V. Graham, ull. 603. TIPO: EN а. Yapacani, roe m, uy о — (holotipo, B destr., fot. F 89487; isotipo, NY!). Beloperone = Rusby, Mem. Torrey Bot. Club 6: 103. ›. Justicia coc habumbensis (Rusby) V. A. . TIPO: Bolivia. е p E!, K!, W. 58 Bull. 43: € 15 (holotipo, NY US!. Be lopero rone — — “inci w 1159. 27. Justicia Pow tue Mildbr. . W. Gra- ham, Kew Bull. : r^ un ioo Stein- 7137 bis — B destr., fot. F 8938!: iso- tipos, BM!, K!, NY!, UC fot. US!). ^ Hierba sufruticosa con rizoma leñoso en la base y tallos erectos, de 0.40—1 m alt., subcilíndricos. cuadrisulcados, con dos líneas pilosas longitudi- nales en su juventud. Hojas con pecíolo de 1-3 cm long. y lámina elíptica de 4-11 em long. por 2-5 em lat.. aguda en el ápice, cuneada en la base у algo decurrente sobre el pecíolo, adpreso-pubérula, con las venas principales prominentes en el envés. Inflorescencias en espigas apicales y en las axilas de las hojas superiores, densas, formadas por flores sésiles con dos bractéolas en las axilas de brácteas decusadas e imbricadas. Brácteas elípticas a an- gostamente obovadas, de 1—1.5 em long. por 3-5 mm lat., foliáceas, agudas y mucronadas, pubéru- las. ciliadas, con las venas principales muy pro- minentes en el envés; bractéolas lineales de 6-8 mm long., agudas, ciliadas, con la vena media pro- minente. Cáliz profundamente 4-partido, con los segmentos lanceolados, de 6-7 mm long.. ciliados. Corola de color rojo-vinoso a violácea. de 3 cm long.. con el tubo angosto, de 2 em long. por 2-3 mm diám.: labios de 1 em long., el superior angos- o. de 4 mm lat., levemente bidentado, el inferior trilobado, los lóbulos de 4 mm long. por З mm lat. Estambres con filamentos de cerca de 1 em long.. as anteras con tecas superpuestas, la superior de 1.5 mm long., la inferior más pequeña y espolonada en la base. Cápsula de contorno angostamente obo- vado, de menos de | em long., cortamente estipi- tada en la base, de paredes finas. Semillas apla- nadas. de menos de 2 mm diám., pardas, con pelos cortos y gruesos en toda su superficie; retináculos agudos. Ilustraciones. Ezcurra, 1993a: ЗАТ. Distribución, hábitat y fenología. Perú, Bolivia, sudeste de Brasil. Paraguay y norte de la Argentina. en ا‎ н htimedos a menos de 500 m de eleva- . Florece todo el año, principalmente en otoño. de. mayo a Junio. Justicia ramulosa se caracteriza por sus flores rojo-moradas en espigas densas con brácteas im- bricadas, elípticas. Es una especie variable que po- 266 Annals of the Missouri Botanical Garden Figura 12. Justicia polygaloides. —A. Planta con flores. —B. Flor con bráctea, bractéolas, cáliz, corola, estambres y estilo. —C. Corola abierta con androceo. —D. Cápsula abierta. —E. Semilla. Woolston 799 (SI), Jörgensen 4237 (SI). Guayculec, Jörgensen 2336 pp (LIL). Jujuy: Ledesma, Confluencia del Rfo Piedras y Arroyo Pantanoso, Fabris . А " | 7302 (LP). Misiones: Salto Iguazú, Lillo 10520 (LIL); mucho a la especie descripta para Guatemala bajo p ы М І Е пепо Aguirre, Rojas 8070 (LIL). Salta: Orán, Río Pes- el nombre de Chaetothylax cuspidatus Gibson cado. Pintascavo. Borsini 613 (LIL); Iruya, Pierotti 662( (Guatemala. Pittier 1792; isotipo, US!) y a Justicia (LIL); San Martín, Gri. Ballivián, El Монет Герпате teletheca T. F. Daniel (Mexico. Breedlove 56314: 10151 C (LIL). PARAGUAY. Alto Paraná: Reserva Ita- isotipo, К!), y las tres parecen estar muy estrecha- bó. Caballero Marmori 395 (CTES, MO), Amambay: in alta planitie et declivibus Sierra de Amambay, in silvis Punta Pora, Rojas 10463 (B * A K, NY, P, W); Parque see una distribución amplia en América del Sur, desde Perú hasta el norte de Argentina. Se asemeja mente relacionadas. Material seleccionado. ARGENTINA. Corrientes: Nacional Cerro Corá, Hahn 2481 (G, MO, US). Central: San Cosme, Paso de la Patria, Costa Toledo, Meyer 9067 V Assomption, dans les bois, Su 2448 (BM, G, К,Р). (LIL). Formosa: km 139 del FC, Jörgensen 2336bis (SI); Cordillera: ad. marg. laqus Ypacaray, Hassler 3174 (BM, Volume 89, Number 2 2002 Ezcurra 267 Justicia en Sudamérica Austral G. К, LIL, MO, NY. P, W): у de Altos, Hassler 2977 (BM. С. К, MO, NY. P, -araguaria centralis, in regione lacus Ypacaray, Hassler Bos (BM, К, LIL, MO. ¿SL US). Guairá: Tororo, Propiedad de Raul Alvaren- ga. Soria 2670 (МО). Paraguarí: Guarapi, dans les bois. Holt 3290 (BM. С. Р): Macizo Acahay, W side of E peak. Zardini 5882 (MO. US): Mbatoví. Montain Forest cks of cerrado type, Zardini y Menem: h 3145 (MO): „| km aro | San Pe ш Colonia — ra. Zardin area. Parque Nac ш, н : .U Sy Woolston 688 x sal SI, US). S Velásquez 12954 (MO). de ep.: s . loe., 31. Justicia riojana Lindau. Bot. Jahrb. 19(48 19. 1894. TIPO: de Famatina, Los Berros, derlein 588 (isosintipo. CORD!): entre La En- crucijada у Las Cuevas, 3 May 1879, Hierony- CORD!): Sierra de Velasco, Cuesta de la Puerta de Pie- dra. 8-11 Feb. 1879, Hieronymus y Niederlein 46 (isosintipo, CORD!). Argentina. La Rioja: Sierra Hieronymus y Nie- mus y Niederlein 546 (isosintipo, m alt, erectas, cilíndricas. las jóvenes verdosas y octocos- Arbusto ramoso, de 1-1.50 con ramas tadas. con estrías marcadas entre las costillas, ge- neralmente glabras. Hojas cortamente pecioladas, angostamente lanceoladas a lineares. de 2—4(—6) em long. por 0.1-0.3 em lat., agudas, cuneadas en la base, generalmente glabras, pubérulas solamente sobre la vena media en el envés. Flores sésiles o cortamente pediceladas, con dos bractéolas en la base. solitarias en las axilas de las hojas superiores de las ramas principales, o de ramitas laterales, formando en conjunto espigas laxas y foliosas: brac- j téolas lanceoladas de cerca de | em long. Cáliz acampanado, de ca. | em long., levemente 5-par- tido, con los lóbulos triangulares. subiguales, de З mm long.. foliáceos. generalmente glabros. Corola azul violáceo o lilacina de 222.5 em long., pubé- rula, con tubo corto de cerca de | em long. .. ensan- chado en una garganta de aproximadamente 0.8 em diám.. los labios de 1—1.3 em long.. el anterior tri- la mitad. con la fauce. lobado hasta convexa y transversalmente rugoso-venosa. Estambres inser- tos en la base de la garganta. las anteras con las tecas subparalelas. una por encima de la otra. la inferior. generalmente apendiculada en la base. Cápsulas obovoides, robustas, glabras, de 1.3-1.8 em long. por 5-6 mm diám.. sólida y lateralmente comprimida. Semillas gene- ralmente 4, suborbiculares. lisas. de 4-5.5 mm diám.. con un engrosamiento en el borde interno: retináculos obtusos. Ilustraciones. — Ariza-Espinar, 1971. Distribución, hábitat y fenología. | Éndémica de con la mitad inferior a provincia de La Rioja. Su distribución está res- tringida a las sierras de Famatina y Velasco, entre los 1300-3300 m s.m. fondos de riachos secos. Habita en barrancos y e Florece en verano. de no- viembre a febrero. ” (Ariza. 1971). Justicia riojana es muy afin a Justicia tweediana Vombre vulgar. “Pichanilla en las flores axilares solitarias y las semillas lisas con reborde engrosado, pero se diferencia por las hojas muy angostas y los segmentos del cáliz sol- dados cerca de dos tercios de su longitud. Es con- siderada forrajera. Vaterial seleccionado. ARGENTINA. La Rioja: Fa- matina, Los Corrales, Cabrera 27228 (Sl): Famatina. La Batea. Sa. Famatina, Иља 1. 1869 (SI); Felipe Sierra de Sañogasta, falda W, subiendo desde Aicuña ha- i as ther, А. 22854, 22863 (CORD); Sa. Fa- Varela, matina, camino a * Mina La Mexicana, Kiesling 6402 (SI): Chilecito, Cuesta de Miranda, Cabrera 27112 (S1): Gral. Lavalle de Cuesta de Miranda. Kiesling 6780 (SI 32. Justicia rusbyi (Lindau) V. A. W. Graham. Kew Bull. 43: 605. 1988. Chaetochlamys rusb- yi Lindau, Bull. Herb. Boiss. 3: 1895. TIPO: Bolivia. La Paz: Guanai, Rusby 1117 (isosintipo, BM!, K!, NY): Santa Cruz, 380 m. June 1892, Kuntze s.n. (isosintipo, NY! USD). New York Acad, Sel. ds ax gii Morong, Ann. € iul. . 1893, non Justicia lanceolata (Cha TER TIP О; Paraguay. Paraguarí: — Райи» һе we n Pirayú and Jaguarón, 8 Abr. 1989, s!) Morong бөг (holotipo. NY! isotipo. Be rene ce adas Bull. S Boiss. 6, 1: 30. 18 TIPO: Paraguay. Cordillera: in — pr. ( bn de Altos, Hassler 19.36 (lec жеш aqui designado. G!: E еза, К!, NY! m Hierba sufruticosa, de 0.70—1.50 m alt.. con ta- llos erectos, subcilfndricos, cuadrisulcados. engro- sados en los nudos, de color verde amarillento, pu- hérulos. glabrescentes. Hojas con pecíolo de 0.5-3 em long. y lámina ovada de 5-12 cm long. por 2— 6 em lat., aguda en el ápice y obtusa a redondeada en la base. algo decurrente sobre el pecíolo, verde clara, pubérula, con cistolitos densamente espar- cidos en su superficie, las venas principales pro- minentes en el envés, amarillentas. Flores sésiles. en espigas truncadas reducidas a flores solitarias acompañadas de varias brácteas formando un fas- cículo, opuestas en las axilas de las hojas superio- res, o en espigas laxas con entrenudos largos, ter- minales y axilares, con brácteas foliosas muy similares a las hojas pero menores: bractéolas lan- ceoladas, de hasta 15 mm long., pubérulas y cor- tamente ciliadas, con el nervio medio prominente. Cáliz profundamente 5-partido. con los segmentos o Annals of the Missouri Botanical Garden lanceolados, rígidos, de 15-20 mm long. por 2.5- 3.5 mm lat., acuminados y densa y cortamente ci- liados, con el nervio medio prominente. Corola li- acina de 3—4.5 cm long., con el tubo blanco, de 1.5-2 em long. por 3—4 mm diám., la garganta dor- 1.5-2 cm long.. salmente gibosa, de poco ensan- chada, y los labios de cerca de 1.5 cm long., el posterior erecto, de 6 mm lat., muy levemente bi- lobado, el anterior lila y profundamente trilobado, reticuloado-venoso en la fauce, con los lóbulos pa- tentes, em long. por 6-8 mm lat. Estambres con filamentos de ca. 15 mm long., y anteras con separadas por el conectivo, de las lecas superpuestas y ambas de 2.5 mm long., glabro con estigma levemente engrosado y biloba- basalmente agudas. Estilo do; ovario glabro y disco en forma de copa, ondu- lado. diám. con el tercio inferior sólida y lateralmente Cápsulas de cerca de 1.3 cm long. y 5 mm comprimido, y la porción superior engrosada, ovoi- de y 4-seminada, pubérula. Semillas subesféricas, de 3 mm diám., castañas, lisas; retináculos de 3 mm long. Ilustraciones. Figura 13. Distribución, hábitat y fenología. Sur de Perú, sudeste de Bolivia, sudoeste de Brasil y este de Paraguay. Habita en el sotobosque y en matorrales de bordes de selva a 200-800 m de elevación, y es frecuente en claros de selvas degradadas. Flo- ece principalmente en verano, de noviembre a abril. Justicia rusbyi se caracteriza por las flores gran- des, lilacinas y blancas con la garganta gibosa, fre- cuentemente solitarias en las axilas de las hojas. A veces se encuentran ejemplares de esta especie de- terminados erróneamente como Ruellia en los her- barios, e incluso fue descripta por primera vez bajo el nombre de Ruellia lanceolata Morong, tal vez por el color lilacino y el tamaño grande de sus flores que se asemejan superficialmente a algunas espe- cies de Ruellia de la región. El epíteto lanceolata no se puede utilizar para esta especie en Justicia por existir un homónimo anterior (Ezcurra, 1993c). Material se оо PARAGUAY. Alto диче in Alto Paraná, Fiebrig 5934 (G, US, SI). arque Nac ional Cerro С ога, Сегго indez Casas у Molero 6113b (MO, NY); in cam- po in regione cursus superioris — уз Hassler — (G); Pe iino те C — го, Sch warez 62 (LIL). E 2 km E de ( y pois d 2 (CTE ; Ruta Я km (NO) = — | ML). Central: Nemby, Vavrek 605 (MO, US); Ся pow rio Industrial, — 3748 (CTES): Villa Elisa, Pedersen 3132 (K. . S1). Concepción: Zw. Rio Apa and Río editus — 1433 (BM, К, G, К, P). Cordillera: Cordillera de Altos, Fiebrig 567 (E, G. К, W); Emboscada, Segunda Compania, па Trompo, Bordas y Schmeda 4093 (CTE Y, US); in regione lacus Ypacaray, Ha: ssler 12458 (BM. 1 UR Cc, К, MO, NY, US); deb vd centrali is, in silva pr. San Bernardino, — 3017 (B е 7): dads Cordillera de Altos, Hassler 3740 (BM. К, MO. . Guaira: Cerro Mumuy, Soria 2991 (G, MO); € olonia Inde penden- cia, Villarica, Pedersen 10148 (K, MO). Ihú, cerca ге — Arbo et al. ж (СТЕ, US); Cerro Palacios, 5 km of Paraguarí, forest on SW slope, e d A. (MO, US); deem Fiebrig 888 . С. К); Parque Nacional Ybycu’f, gallery forest along ^i rroyo Mina, 3 km N of administration area, Zardini y Agua yo 11981 (MO, US); Yaguarón, Cerro Corá, Krapo- vickas 12328 (CTES, US). Presidente Hayes: Paraguay expedition Capt. T. J. Page, Palmer s.n. (US). San Pedro: in regione cursus sup. fluminis Jejui Guazú, Hassler 5719 MO, NY, P); Primavera, Alto Paraguay, Wools- ton 451 (K, LIL, NY, o 1471 (К). S monte Mhune III, Jörgensen 4303 (L — z ‚© — dep.: Común cerca SI, U IL, MO, NY S). 33. Justicia saltensis T. Ruiz & De Marco, Lilloa 35(2) 13. 1980 (“1979”). TIPO: Argentina. Salta: Gral. San Martín, Piquirenda a Quebra- da de Yacuí, 600 m s.m., 3 Feb. 1925, Schrei- ter 3509 (holotipo, LIL 48919!). Hierba sufruticosa en la base, erecta, de 0.4-1 m alt., con ramitas tenues, subcilíndricas, levemen- te cuadrisulcadas, pilosas, glabrescentes a la ma- durez. Hojas con pecíolo de 0.5-3 cm long. y lám- ina ovada, de 3—10 cm long. por 1.5—4.5 cm lat., cuneada a redondeada en la base, aguda y acumi- nada en el ápice, pubérula en el haz, y esparci- damente pilosa a pubescente en el envés. Inflores- cencias axilares, en espigas unilaterales laxas o en racimos de espigas, más largas que las hojas, las flores sésiles en las axilas de brácteas lineales, con dos bractéolas basales, el raquis levemente hirsuto o glanduloso-pubescente. Brácteas lineales, de 3— 5 mm, glandulosas y pubescentes, y bractéolas de 2—3 mm long., glanduloso-pubescentes. Cáliz pro- fundamente 4-partido, con los 4 segmentos angos- tos, lineales, subulados, de 5-7 mm long.. glan- duloso-pubérulos. Corola rosada a morada, de l- 1.3 em long., el tubo basal de 4 mm long., ancho, la garganta de 4 mm long. y 3 mm diám., y los labios de 6 mm long., el posterior de З mm lat., levemente bidentado, el anterior de 5 mm lat., pro- fundamente trilobado y reticulado-venoso en la fau- ce, con los lóbulos de cerca de 2 mm long. y lat. Estambres con filamentos de 5 mm long., las tecas a distinta altura, la superior de 1.5 mm long., la inferior oblicua, separada por el conectivo, algo mayor y apendiculada en la base. Cápsula glan- duloso-pubérula de 1.5 mm long. por 4 mm diám., con el tercio inferior sólido y lateralmente compri- mido, y la porción superior engrosada, cilíndrica y i mm 4-seminada. Semillas comprimidas, de 2.5 Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral 269 Figura 13. — rusbyi. — con — —JD. ápsula — —E. Sem . Planta con flores. —B. Nudo con bráctea. bractéolas. cáliz y estilo. illa. Pedersen 3132 (SD. Woolston 451 (SD. —C. Gorola 270 Annals of the Missouri Botanical Garden diám., pardo-oscuras, densamente hirsuto-pubes- centes; retináculos de 2 mm long., obtusos. Ruiz y De Marco, Ilustraciones. 1980: 15. Norte de Ar- gentina, sudoeste de Bolivia y oeste de Paraguay. Distribución, hábitat y fenologta. Habita en el estrato herbáceo de bosques de regio- nes bajas (500-700 m) del Chaco Serrano de Ar- gentina, y de cerros y serranías del Chaco boliviano y paraguayo. Florece en primavera, verano y otoño, de septiembre a mayo. Justicia saltensis se caracteriza por las flores me- dianas de aprox. 1 cm long. frecuentemente dis- puestas en racimos de espigas secundifloras laxas, con brácteas pequeñas y glanduloso-pubescentes. Debido a sus características morfológicas, en este — trabajo la ubico en la sección Sarotheca (Nees Benth. (Tabla 1). Como enunciaron Ruiz y De Mar- co (1980), presenta afinidad con especies de Amé- rica del Norte como Justicia pringlei B. L. Robin- son. Material dies ipe оиы ARGENTINA. Salta: Yaquias- mé, a 200 n E camino al pie de la montaña, Galetto 186 (Sl); Me tan, 9. 27 km de Metán camino a Gral. Giiemes, Morrone pus SI); Anta, 7 km al 5 de las Ví- boras, Vervoorst 4244 (SI), 8448 (LIL); Orán, Piquirenda (SI). Tucumán: Burruyacu, El Morado, — Burruyacu, El Río, Peirano 10453 (LI AU U А, Chaco: Cerro León, Schinini y oe 17793 (CTES); Parque Nacional Defensores del Chaco, Cerro León, Mi- sión Tribu Nueva, Duré y Brunner 428 (CTES, MO, PY). 34. Justicia squarrosa Grisebach, Abh. Ges. Wiss. Goettingen 19: 226. 1874. Beloperone squarrosa (Griseb.) Lindau, Bot. Jahrb. 19(48): 21. 1894. TIPO: Argentina. Córdoba: Asco- chinga, Lorentz (sintipo, GOET fot. US!); Pa- saje, Lorentz y Hieronymus 301 (probables ran В destr., fot. F 8947!, G no visto . F 26579!). Jacobinia ciliata Nees, in DC., Prodr. 11: 333. — non Justicia ciliata Jacq.. Ноп. Vindob. 2: 47. 1772-3, nec Justicia ciliata (Ruiz & Pav.) Pers., Syn. 1: 23. 1805. TIPO: Argentina. 1261 (holotipo, K!). Santiago del Estero, Tweedie Hierba sufruticosa ramosa de 15—30 cm alt., con rizoma lefioso en la base y tallos ascendentes, al- gunos postrados y radicantes en los nudos, subci- líndricos, cuadrisuleados, glabros o con una línea longitudinal pilosa. Hojas cortamente pecioladas, con la lámina angostamente ovada a lanceolada, de 3-7 em long. por 0.5-2 em lat., aguda o acumi- nada, redondeada en la base y algo decurrente so- bre el pecíolo, generalmente glabra. Flores sésiles dispuestas en espigas muy cortas y densas, con- densadas en el extremo de las ramitas, cada una en la axila de una bráctea y provista de dos brac- téolas basales. Brácteas opuestas, lanceoladas, de cerca de 2 cm long. por 2 mm lat., acuminadas, ciliadas, con largas cerdas extendidas (de hasta 1.5 mm long.) sobre los márgenes; bractéolas lineales, mm lat., Cáliz profundamente 5-partido con los de 2 em long. por 1 largamente cerdoso- ciliadas. segmentos lanceolados, de 1 cm long. por 1 mm — aL, acuminados e hirsuto-pilosos hacia el ápice. Corola violácea o lilacina de ca. 2.5 cm long., con tubo algo más largo que los labios y de 2-3 mm diám., el labio posterior erecto, angosto, de 1—1.2 bidentado, y el anterior cm long. por 5 mm lat., profundamente trilobado, rugoso-venoso y mancha- do de blanco en la fauce, con los lóbulos de cerca de 1 cm long. por 5 mm lat. Estambres con fila- mentos de 7 mm long. y tecas a distinta altura, la superior de 2 mm long., la inferior algo más larga, separada por el conectivo y aguda en la base. Cáp- sula de 1.5 cm long. por 3—4 mm diám., con la mitad inferior sólida y lateralmente estrechada y la superior ovoide. Semillas 4, subglobosas, de 2.5 mm diám., castañas, lisas y lustrosas; retináculos obtusos, de 2 mm long. 90. Ezcu- Ilustraciones. rra, 1993a: 354. Ariza-Espinar, 1971: Distribución, hábitat y fenología. | Especie am- pliamente distribuida en el sudeste de Bolivia, oes- te de Paraguay, y norte y centro de la Argentina, característica de la provincia fitogeográfica Cha- quefia occidental. Habita en lugares secos y — dados entre vegetación de monte entre los 500 y 1000 m s.m. a junio. Florece en verano v otoño, de enero Justicia squarrosa se caracteriza por sus flores lilacinas grandes con el tubo angosto dispuestas en espigas capituliformes densas con brácteas angos- tas y ciliadas. En estos caracteres parece muy afín a una especie de América Central, Justicia. isth- mensis T. F. Daniel ( = Justicia panamensis (Lin- dau) V. A. W. Graham, Panamá. Chagres Valley. Pittier s.n.; isosintipo, BM), ambas deberían estudiarse. Es apetecida por el ga- y las relaciones entre nado. ARGENTINA. Catamarca: oreno, Brizuela 927 (L : L); Capayán, Capayán, Material seleccionado. a Paz, El M Muller 172 (LIL); Valle Viejo, Cuesta del Portezuelo, Hun- ziker, А. 15313 (CORD); El Juncal, — 845 (S 1). Cha- co: Napalpí, Campo Largo, Buratovic $ — (LIL). Cór- doba: Colón, Ascoc ie Giardelli 3 (SD); Pocho, Sierra de Poc Ap» falda W al noroeste de С inc anf, Hun- ziker, A. 9810 (CORD); Ischilin, Sierra de Copac 'abana, Hunziker, A. ous COH Tulumba, ca. de L. V. Man- Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral 271 del desvío a Turcal. Hunziker, A. 25229 (CORD): Capital, C^. de las Rosas, Pierotti 5131 LIL). Jujuy: San Pedro, San Pedro, Ahumada 4481 (SI); Santa Bárbara, Pto. Nuevo, Cabrera 26. 314 (S silla, en las inmediaciones — Km de 4 Caro 13618 (CORD): Chamical, — De Soria, 8 Km al SE de Chamical, Hunziker, A. 1 7 (CORD). Salta: La 5 de Cnel. Mc — p 5838 (SI); Ro- sario de la Frontera, 2.3 Km de la ruta 34, c — a San i Cristóbal 46675 (LIL): Metán, (LIL); Anta, n V. González. Morello s.n. (LIL). Santiago del Estero: Guasaván, falda oriental ER la Sa. de Guasayán, de S santa Catalina a La- Heo — 138 (SI); Dpto. Ojo de Agua, Cachi. Di Lullo 7 (5 . El Salve Фан 2. Legname s.n. igres, З Km al N de Monte Quemado. ;. Pellegrini, Cerro del Remate. ia man Tum — Santa Rosa, Qda. De La Higuerita, — 13958 (SI). Tucumán: Trancas. — de Astrada, Tapia, Diers 46 d — Trancas, Vipos, ickas y 49, idor, Venturi 2686 (LIL). PARAGUA querón: Campo Loro, limite con Alto Paraguay, Sc a 866 (CTES, US): Estación Experimental Filadelfia, Vanni 2577 (CT ES): Pi- "г cada entre Teniente Ochoa y Mariscal Estigarribia, а 10 km de Mariscal — en — del bosque, Soria 1308 (CTES. MO ueva Asunci ca. Estancia Co- pagro. 506 km de m бп, Bernard 20202 (( (G. US); Es- tancia La Madelón, piste Trans Chaco pk 635, — y Jadin 3123 (G); Fortín Teniente | iso, ruta Trans — y Bordas 16471 (CTES, G, US); ( . Garay, línea de Hito, frontera con Bolivia, Charpin Ramella 21485 (G); Ruta Trans-Chaco, Schinini y Bordas 6375 (CTES). zeneral Eu ugenio 35. Justicia prg (Nees) V. A. W. Graham, Kew Bull. 43: 604. 1988. Chaetothylax tocan- tinus — in Martius, Fl. Bras. 9: 153. 1847. TIPO: Brasil. Goyaz, Pohl (sintipo, W!; isosintipo, М: Minas Gerais, Claussen (sintipo, P no visto: isosintipo, K!). SM. Hierba perenne, rizomatosa y leñosa en la base, con tallos erectos subcilíndricos, a veces decum- bentes y geniculados, de hasta 50 em long., pu- bérulos o glabros. Hojas opuestas con pecíolo de 5-20 mm long. y lámina ovada de 4—10 cm long. por 2.5—5.5 em lat., acuminada en el ápice, cunea- da y algo decurrente en la base, entera, general- mente glabra, con la nervadura prominente en el envés. Inflorescencias formadas por flores sésiles con dos bractéolas en su base, dispuestas en es- pigas densas enriquecidas por flores derivadas de yemas axilares supernumerarias. Espigas termina- les o en las axilas de las hojas superiores. Brácteas y bractéolas lanceoladas, rígidas, de 4-6 mm long., cilioladas, con el nervio medio prominente en el dorso. Cáliz de 5-8 mm long., con los 4 segmentos rígidos, lineales, subulados, con pelos glandulares. 7-2.9 cm long.. exter- namente pubescente, con tubo angosto de 2 cm Corola rosado—violácea de 2. long., ampliado en una garganta de ca. 4 mm diam.. el labio superior erecto, de 7 mm long. por 2 mm lat., el labio inferior ancho, trilobado, de ca. de 8 mm long., con estrías blancas en la fauce. Estam- bres con filamentos de aprox. 8 mm long.: teca api- cal perfecta, de contorno ovado, de 1.5 mm long., a otra inserta más abajo en el filamento. semies- téril, más angosta y basalmente caudada. Cápsula de contorno obovado, de aprox. 7 mm lat., mm long. por 2.5 cortamente sólido-estipitada en la base. generalmente glabra y de paredes finas. Semillas 4 aplanadas, de 1 mm diám., con pelos cortos y glo- quidiados: retináculos de 1 mm long., obtusos. Nees, 1847a: 26 Ilustraciones. — sub Chaetothy- lax tocantinus). Distribución, hábitat y fenología. Este de Bo- livia, sudoeste y centro de Brasil, y extremo norte de Paraguay. Habita en bosques estacionalmente secos. Florece en otofio y primavera, principalmen- te en abril y mayo. Justicia tocantina es afin a Justicia goudotii. pero se caracteriza por sus flores rosado-violáceas grandes de aprox. 3 cm long. con tubo angosto le- vemente ampliado en la garganta, dispuestas en in- florescencias kae iformes densas similares a las de Justicia goudotii. Es la primera vez que se cita esta especie para Paraguay y Bolivia. — seleccionado. in (Lagerenza). Palese 2970 (G ); mento) hasta Mes Chaco: Cerro PARAGUAY. p-p- (G): Parque Na ional Defensores del Chace León, Hahn 1530 (G, MO, PY IVIA. Santa Cruz: Caballero, 8-9 km east of S a at on road to Pulgina. Wood 10919 (K); Caballero, ca. Rio Comarapa, Wood 10923 (K). 2 km above Comarapa by 36. Justicia tweediana (Nees) Grisebach, Abh. yes. Wiss. Goettingen 19: 225. 1874. Adha- ida tweediana Nees, in DC., Prodr. 395 1847. Justicia tweediana (Nees) Benth.. in Bentham & Hooker, Gen. Pl. 2(2): 1409. 1876. Ecbolium tweedianum (Nees) O. Kuntze. Rev. Gen. Pl. 1(2): 981. 1891. Poikilacanthus twee- dianus (Nees) Lindau, Bot. Jahrb. 1893. TIPO: Argentina. Córdoba: Río Segun- do. Tweedie s.n. (sintipo, K!); Parana (“Para- ma"), Tweedie s.n. (sintipo, К); Buenos Aires, Tweedie s.n. (sintipo, K!; isosintipo. BM): San Luis: Aguadita, Gillies s.n. (sintipo, K!; isosin- BM!, E!). tipos, Adhatoda tweediana Nees var. angustifolia Nees, in DC. Pr odr. 11: 395. 1847. Argentina. Parana ("Panama"): Tweedie s.n. (sintipo, К); Buenos Aires. Tweedie s.n. 272 Annals of the Missouri Botanical Garden e K!), Gillies s.n. (sintipo, K!); San Luis, Gil- lies s.n. (sintipo, K!). Justicia campestris Griseb., Abh. . Wiss. — 19; 225. 1874, non iis erie ris Nees, Martius, Fl. Bras. 9: 118. 1847. Justicia — Lindau, Bot. Jab — kie 1894. Ecbolium lo- rentzianum (Li ind: Kuntze, Rev. Gen. Pl. 3(: 248. 1898. ПРО: ү усен Córdoba. Las Teak Lorentz (holotipo, GOET no isolipo, CORD!, fragmento LI Justicia diamantina Lindau, Bull. Herb. Boiss. 5 (ser. 2): 371. 1905. TIPO: Argentina. Entre Ríos: Diamante, * 1903, Malme s.n. (sintipo, B destr.; no vistos). т visto; — I isosintipos Sufrútice o arbusto ramoso, de 0.50-1.50 m alt., con ramas erectas, cilíndricas, las jóvenes verdosas y octocostadas, con estrías marcadas entre las cos- tillas, generalmente glabras. Hojas cortamente pe- cioladas, angostamente ovadas a lanceoladas, de 2— 4(-6) em long. por 0.5-1.5(-2.5) ст lat., agudas, cuneadas en la base, generalmente glabras, pubé- rulas solamente sobre la vena media en el envés. Flores sésiles o cortamente pediceladas, con dos bractéolas en la base, solitarias en las axilas de las hojas superiores de las ramas principales, o de ra- mitas laterales, formando en conjunto espigas laxas y foliosas; bractéolas lanceoladas de cerca de | ст long. Cáliz profundamente 5-partido, con los ló- bulos angostamente ovados a lanceolados, subigua- les, de 5-8 mm long., foliáceos, generalmente gla- bros. Corola azul violáceo, lilacina a blanca de 2- 2.5 em long., pubérula, con tubo corto de cerca de l em long., ensanchado en una garganta de apro- = ximadamente 0.8 ст diám., los labios de 1—1.3 сп long.. el anterior trilobado hasta la mitad, con la fauce blanca, convexa y transversalmente rugoso- venosa. Estambres insertos en la base de la gar- ganta, las anteras con las tecas subparalelas, una por encima de la otra, la inferior generalmente apendiculada en la base. Cápsulas obovoides, ro- bustas, glabras, de 1.3-1.8 em long. por 5-6 mm diám., con la mitad inferior sólida y lateralmente comprimida. Semillas generalmente 4, suborbicu- lares, lisas; retináculos obtusos. Ilustraciones. Justicia campestris). Dawson, 1979: 1971: 96 (sub 573 (sub Jus- ticia campestris). Ezcurra, 1993a: 348. Ariza-Espinar, Distribución, hábitat y fenología. Norte, centro y E de la Argentina, desde Jujuy hasta San Luis, Córdoba, Santa Fé, Entre Ríos y N de Buenos Ai- res. Vegeta en regiones áridas y semiáridas entre el nivel del mar y 2500 m (según la latitud). Es apetecida como forraje por cabras y ovejas en lu- gares con escasez de gramíneas. Florece principal- mente en verano, de septiembre a mayo. Nombres vulgares. “Boca de conejo,” “palomi- llo," “а Ша,” “ааа,” dura” (Burkart, 1943, sub Justicia campestris). Justicia tweediana se caracteriza por las flores “quiebrarao,” “escoba medianas, solitarias, axilares, de color celeste-li- lacino a blanco, y las semillas aplanadas, lisas, con engrosamiento perisférico. Tiene un alto valor como forrajera (Burkart, 1943). A RG ; s NTINA. Buenos Aires: 1 Nicolás, San Nicolás, barrancas, — — (SI); E Baradero, barrancas, Burkar K, SI). Ca- tamarca: Andalgalá, Andalgalá, hae en or 31); Ca- payán, c camino al dique nivelador € án, Legname 7264 LIL); Santa María, El Sarcito, Reales 1055 5 (LIL); Belén, Pozo de Piedra as, Sleumer y Vert — e П IL). Chaco: CH, Pampa del Infierno, Meyer 8570 (LIL); Campo del Cielo, fn, Schulz 1095 (LIL). b ;apilla del Mon- te, а нра s.n. (LIL); Punilla, Sierra С s ‚ al N del С° Unpico, Hunziker, А. 8947 (CORD); S n Justo entre Concepción del Tío y El Tio, Hunziker, A. 10068 CORD); Totoral, alrededores del agua en el arroyo Macha, 5 Km W Hunziker, A. 21843 (CORD); Río Cuarto, entre Berrotarán y Santa Isabel, Ragonese y Pic- Sur sila seleccionado. — de Las Pefias, > Р. — 9408 (SD; Tercero Arriba, Río Tercero, Burkart 3 (SI). Entre Ríos: Paraná, Barr s de Paraná, rior 20647 (Sl); Victoria, Vic toria, Мезе er r 10 2198 (LIL); 1 (LP); Tumbaya, Laguna de Volcán, Kiesling 4345, 5823 (SI). La Rioja: Chilecito, cuesta de Mi K 540, — 4. 22674 iren Da 10 (SD. Salta: Camino a Cachi, pasando Es- "agnum 9005 (LIL); Candelaria, Unquillo, Schrei- ler "935 (LIL); Rosario de Lerma, Campo — Venturi 5060 (51). San Juan: 0 Juan, Pa p (LIL). Fé 3813 (SD. San Luis: 1 auis, Covas 1102 (L AL ); Pe- = -= = 5 AA — — оо р; کب‎ derson 1811 (CORD): Dpto. Pringles, Volcán, Kiesling 4718 (SI); Chacabuco, s arán, 605 m, — 495 (LIL). at del F a 16, Legname 7521 (LIL); Copo, Pampa de los Gua- nacos, Malvárez 675 (LIL). на 1 ntr fiernillo у Amaicha del », Legname y Cuezzo 5514 (LIL); Tafí, Sa. Del Cajón, Venturi 4438 (SI). 37. Justicia xylosteoides Grisebach, Abh. Ges. Wiss. Goettingen 19: 225. 1874. Ecbolium xylosteoides (Griseb.) O. Kuntze, Rev. Gen. Pl. 3(2): 248. 1898. TIPO: Argentina. Santiago del Estero: “antes que se pasa el Saladillo”, 4/5 Dic. 1871, Lorentz 12 (holotipo, GOET no vis- to; isotipo CORD!). - Beloperone scorpioides Nees, in DC., Prodr. 11: 422. 1847. Justicia 1 scorpioides (Nees) Griseb., Ae Ges. Wiss. Goettingen 24: 262. #45 — — scorpioides L.. Sp. Pl. ed. 2: 21. 1762-63. TIPO: Argentina. “Tucumán”, Saladillo, — pe ы K!, fot. US!). Arbusto de 1-2 m alt., mas erectas, 4-sulcadas en su juventud, general- robusto, ramoso, con ra- Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral mente glabras: ramitas laterales con entrenudos cortos en la base, alargados superiormente. Hojas subsésiles o muy cortamente pecioladas, con lám- ina angostamente elfptica, ovada, obovada u oblon- ga de 2-4(—6) ст long. por 0.7-1.5 em lat.. glabra o muy levemente pubérula, glanduloso punteada, obtusa a redondeada en el ápice, cuneada en la base. Espigas sésiles, densas, terminales o axilares, a veces agrupadas en racimos, con brácteas lan- ceoladas a ovadas, opuestas y decusadas. a veces alternas por desplazamiento y concaulescencia. siempre una fértil y la otra estéril de cada par, las fértiles mayores, de hasta 8(-10) mm long. por 2— l mm lat., las estériles menores, de 2 mm long. y | mm lat., el raquis arqueado: bractéolas triangu- lares, de 2-3 mm long. y 1 mm lat., escabroso- pubérulas. Cáliz de 3-4 mm long., con los 5 seg- mentos ovados. agudos, pubérulos. Corola roja. de 34.5 cm long.. con tubo de 2-2.5 cm long.. sub- cilíndrico. poco ensanchado en la fauce, y labios de 1-1.5 em long., el posterior apenas bidentado, el anterior profundamente trilobado. Estambres con filamentos de 1.3-1.7 em long., y anteras con las tecas a distinta altura, la inferior oblicua y apen- diculada en la base. Cápsula de 1.3 ст long. por 0.5 cm diám., sólida y lateralmente estrechada en la base, superiormente inflada y 4-seminada. Se- 3-4 tuberculado-rugosas. más o menos oscuras. Ariza-Espinar, 1971: 81. millas suborbiculares, aplanadas. de mm diám.. Ilustraciones. Distribución. hábitat y fenología. Regiones ári- das del sur de Bolivia, noroeste de Paraguay y no- roeste de Argentina. Habita en claros de montes xerófilos bajos y en matorrales áridos entre 700 y 2300 m s.m. Florece principalmente en verano, de septiembre a mayo Tal como lo señaló Lindau (1894). Justicia xylos- tevides es una especie de afinidades poco claras. tal vez por sus modificaciones en adaptación a la aridez. Esto la hace difícil de ubicar en el sistema de Graham (1988). Por sus espigas simples con brácteas generalmente ovadas, flores grandes у se- millas tuberculadas se ubica provisoriamente en la sección Orthotactus (Tabla 1). Es una especie Па- mativa por sus grandes flores rojas. ornitófilas, que merecería cultivarse en climas áridos. Es apetecida por el ganado. — seleccionado. ARGENTINA, Catamarca: La Paz, Ramblones, Brizuela 456 (LIL); Fray M. n Polos ruta 62, Hunziker, A. 22457 (CORD): C apital, ( ia- amar Vicora s.n. (SI 18623); Santa María. El Sarcito, Re bai 1043 (LIL); Pachín, entre € catamarca y Cuesta del Córdoba: en el camino de Villa Bañado de Paja y El Chocolate. Dolores a Chanc ‘anî, — А. 14694 (CORD): Sobre- monte, 6/7 Km : oeste de la plaza de San Francisco del Chafiar, rumbo 9 ansilla, Hunziker, A. y Subils 24951 ( CORD): C ruz z del Eje, Camino a Soto, Nicora s.n. (SI 17632). La Rioja: Chamical, КРА, 5 Km al N de la Salina La Antigua, Biurrun 584 (CORD): Capital, Dique Los Sauces, falda de la Sierra de Velasco, Hunziker, A. 1708 (LIL); Gral. Belgrano, Sierra de Los Llanos, ns Olta y El ое Hunziker, А. 13915 (CORD): Gral. La- iranda, Kiesling 6783 (SI). Salta: La Vina, 5 km als de Coronel Moldes, Kiesling 5844 (%1): . Lomas altas del Cerro Santa Victoria, Legna- LIL): € tachi, Bajada de Tin Tin, Cabrera 30759 (SI); Cafayate, camino de Santa María a Cafayate, a 2 Km de Cafayate, Legname 8893 (LIL): Metán, entre Gas del Estado y Río Juramento, Meyer s.n. (LIL) Molinos, Se- clantás a Brealito, Meyer 12195 (LIL). Santiago del Es- suasayán, Sa. de ( suasayán, Avesling 1262 (Sl): Pe- Sol de Mayo a Los Baños, Peirano 9021 (LIL): as Termas, Meyer 12789 (LIL); Choy valle, Cues a de M - = A. уч MC © - ~ lero: llegrini, Rio Hondo, | ба, desde Puerta Chiquita, yendo hacia Alto Bello, Subils 5666 (CORD). Tucumán: Tafí. Amaicha del Valle, Burkart 5328 (SI): Trancas, Tapia, Rodriguez 597 (SI). PARA- GUAY. Boquerón: Colonia Menno, Estancia Fe 2 68 km | 5 і 2: 208 ic S) Made lon; ruta Transe ‘haco, o E ешп. Û ki N g the — por la picada Boliv- ‚ Vavrek 334 (MO, US). 38. Justicia yhuensis Lindau, Bull. Herb. Boiss. Ser. 2, 4: 411. 1907. TIPO: Paraguay. In re- gione fluminis Yhú, Oct. 1905, Hassler 9566 G!; isotipos, BM!, K!, NY! Р!). (holotipo, Justicia hylobates Leonard, Sellowia 9: 83, fig. 2. 1958. TIPO: Brasil. Santa Catarina: prios Popi, pelo caminho a Sant Antonio, mata, 200-35( p Feb. 1957. Smith, Klein y cia 2» 7 (holo- lipo. US! isotipo, NY!). Hierba débil enraizadora en la base, de hasta 60 em alt., con tallos erectos, subtetrágonos, ramifi- cados. longitudinalmente surcados, pubérulos a elabros. Hojas cortamente pecioladas o subsésiles, con la lámina angostamente ovada, de 4—7 cm long. por 1.2-2.5 ст lat., redondeada a levemente cordada en la base, algo largamente aguda en el ápice. discolor, glabra o muy laxamente pilosa. Inflores- cencias en espigas terminales y laterales de hasta 12 em long., laxas, delicadas, con pedúnculos de 2 em long.: brácteas angostamente triangulares, de 2-3 mm long. y 0.5 mm lat., agudas, aquilladas, elabras o pubérulas; bractéolas de la misma lon- Cáliz profundamente 4-partido, con los seg- . por 0.5 mm lat., оиа. € mentos lanceolados, de 3 mm long agudos. glabros. Corola blanca de ca. 9 mm long. con tubo obcónico de 3—4 mm long. y labios de 4— 5 mm long.. el posterior triangular, de 2 mm lat. en 274 Annals of the Missouri Botanical Garden la base, obtuso; el anterior trilobado, con los ló- bulos de 1.5-2 mm long. y 2.5 mm lat. (los laterales algo menores), redondeados. Estambres exertos hasta la mitad del labio posterior, con las tecas su- perpuestas, de ca. 1 mm long., separadas por un conectivo ancho, la superior mayor y oblicua, am- bas agudas en la base. Cápsula delicada, angosta- mente claviforme, de 7-8 mm long. por 2 mm diám., sólida y lateralmente estrechada en el tercio inferior, glabra. Semillas aplandas, de 1 mm diám., con papilas cortas; retináculos de menos de 1 mm long. Пизїгасї iones. Wasshausen y Smith, 1969: 96 (fig. Sur de Brasil, Habita Distribución, hábitat y fenología. noreste de Argentina y Paraguay oriental. en bordes de selvas de regiones bajas y florece en primavera y verano, de octubre a mayo. Justicia yhuensis se caracteriza por sus hojas lan- ceoladas y sus espigas simples de flores medianas a pequefias. El nombre Justicia hylobates se con- sidera sinónimo de esta especie. Material se — ARGENTINA. Misiones: Ori- . Deginani 1135 is Guaraní, RP 2 inci | de El Soberbio ca- ming al Parque Prov. Moconá, ion ie (51); San Ja- i — Schwarz 4310 (LIL). PARAGUAY. Amam- ra de Amambay, ad marginem silvarum ad fl. М. Быр, Rojas 9718 (G); Zw Río Apa und Río Agu dabán, Caballero Cué, Fiebrig 4 481. 3 (BM, E, G, K). Cor- dillera: inter rupas in dumetis Cordillera de Altos, Hass- ler 2135 (G). Cone реш Río Aquidabán, Paso Horqueta, Palacios 1873 (MO) E Especies DUDOSAS O EXCLUIDAS Justicia flexuosa (Nees) Wasshausen & L. B. Smith (citada por Wasshausen & Smith, 1969, para Santa Catarina, Brasil). Justicia umbrosa (Nees) Lindau (citada por Wasshausen & Smith, 1969, para Santa Catarina, Brasil). Estos dos nombres son sinónimos de Poikila- canthus glandulosus (Nees) Ariza (Ariza-Espinar. 1983; Ezcurra, 1999). Poikilacanthus Lindau es un género superficialmente muy similar a Justicia, con polen con características algo diferentes (Daniel, 1991, pecies de Justicia del Nuevo Mundo a nivel molec- ular (Mc Dade et al., 2000), por lo que eventual- mente podría ser tratado dentro de Justicia s. 1998), pero muy relacionado con algunas es- خر al. Pero hasta tanto se estudien las relaciones morfo- lógicas y moleculares de un número mayor de es- pecies de Justicia del Nuevo Mundo prefiero man- tener a Poikilacanthus como género separado de Justicia. A continuación se da la sinonimia com- pleta de esta especie. Poikilacanthus glandulosus (Nees) Ariza, Kur- tziana 17: 157-161. 1984. Orthotactus glan- dulosus Nees, in Martius, Fl. Bras. 9: 132. 1847. Adhatoda umbrosa Nees, in DC., Prodr. 11: 406. 1847, non Adhatoda glandulosa Nees, 1847. Ecbolium umbrosum (Nees) Kuntze, Rev. Gen. Р]. 1(2): 981. 1891. Justicia umbrosa (Nees) Lindau, Bot. Jahrb. Syst. 19(48): 20. 1894, non Benth. 1841. TIPO: Brasil. Sellow 164 (sintipo, В destr., fot. F 8877!). Nov. Act. Nat. Cur. 11: Justicia — Nees & Mar 55. 182: 1 Jap: ), non Justicia ed gs Thunb., 22. 1784 ЕЕ — a — | Mar rt.) P р Mar Fl. Bras. 9: 136. . TIPO: Bra- Шош s.n. ‘(holotipo B de 'slr.; — K!). Maid eun Nees, in وا‎ 9: 148. 17. Ecbolium flexuosum (Nees) ( intze, Rev. : 980. 1891. Poli anthus Ер (Nees) . Bot. Jahrb. 18: 57. . Justicia flexuosa — d & Smith, in R. Reitz, Fl. lustr Catarinense ACAN: 113. 1969. TIPO: Brasil. Grande do Sul: e ntre 5 Borin Alegre y Montevideo, “ad А DM 997 (sintipo, B destr., fot. i ; Minas ( Villa Rica, de str.), Martius s.n. (sintipo, . — Taubaté et Mogy, Riedel s.n. (sintipo, LE no visto; isosintipo, NY!); Sáo Paulo, ad . Carlos, Riedel s.n. (sintipo, LE no vis- — ripas fl. Urugua Ft Gerais, oppidulum S to). Ilustraciones. Wasshausen y Smith, 1969: 110 F) (sub Justicia flexuosa, fig. 15 Literatura Citada . Las especies de Justicia (Acan- Argentina. Kurtziana 6: 77— Ariza-Espinar, L. 197 thaceae) del centro a la 101. м 19 Notas sobre Acanthaceae ТУ. El género Paikilas anthus I. Bol. Soc. Argent. Bot. 22: 259-261. Bailey, 1 949, Manual of Cultivated Plants. Macmil- lan, т p rk. Bentham, G. 1876. Acanthaceae. In ho = р & J. D. — Genera Bison 2(2): 1 Brako, L. & J. L. Zarucchi е К ee Cate of the Hose Plants and Gymnosperms of Peru nogr. Syst. Bot. ao А al Garden 45. Bremekamp, С. B. 1969. An е list of the Acanthaceae ы ted by › W. М. A. Brooke on her Ned. FIN Wetensc trave ^ 10 Bolivia. Proc. Коп. ch.. Ser. C, 72(4): 420—430. [Reprinte n in — Bot. Mus. VEN к. Univ. Utrecht 331: —430. 1969, | Burkart, A. 1943. Acantáceas ® como forrajeras de emergenc iniana 6: 192—202. ~ м Cabrera, A. L. . Willink. — Biogeograffa de Amé- rica Latina. Mong 13. O. E. A., Washington, D.C. Daniel, T. E * ). New and sete Foch Mexican Ac — hac eae | — Contr. Univ. Michigan Herb. 17: 133-13 1991. A synopsis of Poikilacanthus (Acantha- me in Mexico. Bull. Torrey Bot. Club 118(4): 451— 458 1993. Mexican Acanthaceae: Diversity and dis- Volume 89, Number 2 2002 Ezcurra 275 Justicia en Sudamérica Austral tribution. /n T. P. e R. Bye. A. Lol & J. Fa (editors), Biological Dive Distribution. Oxford Univ. | : Origins and 5, New York & Oxford. — ‚ Ac ca ‘eae. ; Breedlove, Flora of Chianas : — 58. Calif. | , San Francisco. —— 19 qo morphology of Mexican pis Calif. Em ^ сеае: — rsity ane systematic significance. Proc. Acad. — 1. 50(8): 217-256. 00. Additional _ numbers of Amer- ican в a ‘eae. Sys : 15-25. ————4& T I. Chuang. о Chromosome New World no 'anthaceae. Ea Bot. 18: 28 B. arfitt & M. A. Baker. 1984. A their systematic implications in some | 9: 346—355 numbers of 3-289 — Ts anc ud North * rican Acanthaceae. Syst. Bot. | L. Chuang € M. A. Baker. 1990, Chromosome — rs of — ‘eae. Syst. Bot. 15: 13-25. Dawson, G. 1965 \canthac ‘eae. In A. L. Cabrera (editor), Л. Buenos — 5: 312-335. Colece. Ci. INTA, Buenos Aires. . 1979. Ас — ‘eae. Entre Rios 5: 545- In ^. du (editor), F . Colece. INTA, Buenos je De dm N. & T. Ruiz. 1976. Tres nuevas especies del género Justicia Bow ex L. (Ac — ш рага Argen- . Esp. Inst. Lillo 1187: Durkee, L. H. 1986. Acanthaceae. * Flora болайык, Family 200. Fieldiana, Bot., n.s. 18: 1-87. — (editor), Ezcurra. €. 1988. Novedades nomenc е еп especies de Justicia Goes ie) del 1 de la Argentina. Bol. Soc. — Bot. 25(3—4): : Sn 1989. Ruellia sanguinea (Acanthaceae) y espe | cies — ionadas en Argentina, Uruguay y sur de Bra- sil. Darwiniana 29: 269-287. 1993a. Acanthaceae. In la LER ne la de Jujuy 9: L. Cabrera — Flora de 278-359. Colecc. INTA, donde Aires. . Notas sobre la posición sistemática y dis- geog ográfi ica — lilloi (Acanthaceae). Argent. Bot. 29(3—4): 187—190. 1993 © Systematics к Ruellia (Acanthaceae) in aiher South America. Ann. Missouri Bot. Gard. 80 787 ibus ión Bol. x 1-845. 1999, Acanthaceae. In F. O. Zuloaga & O. Mo- rrone (editors), Catálogo de las aes Vasculares de la Re RA a Argentina. Monogr. Syst. Bot. Missouri Bot. Gard. 74: 1-14. & й sta-Castellanos. 1997. — — yJ. ae front, Acanthaceae), 1 vas citas para Argentina y bon. y notas E su posición. sistemática. Bol. Soc. Argent. Bot. 33(1— 107—111. 07— Gentry, A. H. 1982. Neotropical floristic diversity: Phy- кее ‘al connections between central and South America, Pleistocene climatic fluctuations, or an acci- dent of yere orogeny? Ann. Missouri Bot. Gard. 69: 551—593. Gibson, D. 912. е plants Ш. Fein. _ y Graham, V. A. W. — and infra-generic classification. of — (Acanthaceae). Kew Bull. 34: 991-624. In Studies in. American 4. Grant, W. F. 1955. A cytogenetic study in the Acantha- ceae. ion 8: 121-149. Hay, R. & К. A. Beckett. 1975. Headers Digest Encyclo- — of Garden Plants and Flowers. The Reader's Di- est Association, London. Hite heock, A. S. & M. L. Green. 1930. Standard species of Linnean genera of Phanerogamae (1753-54). Int. Bot. Congr. PELA (Е ngland) 1930, Mount Prop. Brit. Bot.: 110 Holmgren, P Ps N. H. Holmgren & L. C. Barnett. 1990. Index Herbariorum. Part I: The He of the World. 3th ed. New York Bot. Gard., Bron. New eón-Yánez. 1999, antas Vasculares del Ecuador. Monos Syst. Bot. Mis- York. Jörgensen, P. M. & S. Le Catálogo de las Jot. Mi souri Bot. Gard. 75. E C. 1995. Cipó, Minas Gerais, 14: 18 l- ne | — Acanthaceae. In Flora da Serra do Brasil. Bol. Bot. Univ. Sâo Paulo . 1951-1958. The Ac + of Colombia. Contr. U. ^y Natl. Herb. 31(1-3): 1-781. . М. & C. Ezcurra. 1991. ment as an aid in the study in Justicia and Ruellia (Acanthac eae). Bot. J. Linn. Soc. 105: 285-288 r nzyme etching treal- of seed surface sc нея 59 Catálogo de las Acantáceas Argentinas. : 21-66. 9 bare. 1893. Beitrage zur Systematik. der. Acantha- . Jahrb. Syst. 18: 36-04. 894. Beitrage zur argentinischen Flora. Bot. 1. aah, Beibl. 48: 8 — 895. Acanthaceae. . Engler & Die Natûrlic hen st r n 43b): 274— rantl. Linné 1753. Species Plantarum, ed. 1. ES niii TrA уп, жи otti 976 nueva espec ie de Acanthaceae: isa ina —— е Lotti. Publ. Esp. Inst. Lillo 1187: 336€ "i Dade, a A. 198 — — eid Costa Rica, with taxonomic notes. Syst. Be n. : 489—497. & T E Daniel. . New species of Justicia and Razisea é 1998. ер the Justicieae (Acanthac — Amer. J. Bot. 85: —144. (Sy E. Masta & К. » pe 2000. Phy- logenetic relationships within the tribe Justicieae (Acanthaceae): Evidence from molecular morphology, and cytology. Ann. Missouri Bot. 35—458 sequences, Gard. 87: Nees, C. G. Acanthaceae. /n C. Martius (editor). Flora Brasiliensis 9 1-164. 47b. Acanthaceae. In A. P. de Candolle (edi- tor). p m Systematis Naturalis Regni Vegetabilis Ll: Parodi, L. M. J. Dimitri. 1980. Las plantas cultiva- das. hs ic alon edia — de Agricultura y Jardinerfa | (2). — ed. ACME. Buenos Ai Peixoto, M. 1998. pecies de — (Acanthaceae) del noreste Darwiniana 36(1—1): 13—28. ————, C. Ezcurra & D. Wasshausen. 1998. а Jlo- FEM (Acanthaceae), nueva combinación y nueva cita Darwiniana 35(1—4): 15 * 54. a 3 D iudi morfo-anatómicos en es- Argentino. зага la flora argentina. Pennington, R. T., D. E. Prado € €. Pendry. 2000. Neotropical seasonally dry forests and Quatemary veg- elation changes. J. Biogeogr. 27: 261— Petriella, B. 1968. El polen p» las Ac e 'eae = nas. Revista Mus. La Plata, Sece. Bot. 11: 51-6€ Piovano, М. & 1. M. 1991. numbers in Argentinean Acanthaceae. Syst. Bot. 16: Bernardello. C — 89-97. Profice, S. R. 1993. Estudos em Acanthaceae Novas 276 Annals of the Missouri Botanical Garden — s e novos sinonimos, Justicia L. Eugeniana 20: Raj, B. 1961. Pollen — studies in the Acan- thaceae. Gran na Palynol. 3 ‘urther pei ESA to the pollen mor- phology of the Acanthaceae. J. Palynol. 9: 91-141. Rambo, 1964. Acanthaceae Riograndenses. Bot. т 36. Rizzini, C. T. 1948. Disquisito circa Acanthacearum ali- quot genera Brasilensis. Arch. Jard. Bot. Rio de Janeiro 95-372. — Iheringia . 1949. Contribuicáo ao conhecimiento da tribo Justicieae (Acanthaceae). Arch. Jard. Bot. Rio de Ja- neiro E 37-€ 951. Sinopse parcial das Acanthaceae brasilei- ras. —* ‘nia =i 45-180. Ruiz, T. & N. De Marco. 1980. Una nueva especie de Justicia a dedo 'eae) en la Argentina. Lilloa 35(2): 13. 1979-1980 Sell, Y. 1969. Te 's complexes inflorescentiels de quelques Acanthacées. Etude particulière des phénomènes de condensation, de race misation, d homogé néisation et de tronc — s Sei. Nat. Bot. sér. 12, 10: 225—300. . Tendences e Sale 's parmi les complexes — sce a ‘ls. Rev. _ n. Bot. 83: 247-267. ›. Acanthaceae. In Lundell (edi- pecies and new combinations of Jus- ticia (Acanthaceae) from the Venezuelan Guayana. No- von 2: 62—80). & L. B. Smith. 1969. ч ne In R. Reitz, Fl. Il. Catarinense | (ACAN): 3-134. & C. Ezcurra. 1997. n. names and new com- binations in Justicia sects. Simonisia, Plagiacanthus and Orthotactus a from southern South rit ш а. Candollea 52: 171—179. Webb, C. J. 1995. Checklist of dicotyledons, sperms, and pteridophytes naturalized or casual in New Zealand: additional records. New Zealand J. Bot. 33: 151-182 gymno- APÉNDICE | LISTA DE. ESPECIES Justicia pi unde (Nees) Lindau Justicia axillaris (Nees) Lindau Justicia baenitzii (H. Winkler) С. Justicia brasiliana Roth Justicia carnea. Linc Justicia chacoénsis Wessbausen & C. Justicia comata (L.) Lamar Justicia corumbensis (L паш) Wasshausen € C. Ez- № – Ezcurra > Y — Л Ezcurra ^I oen ^ TU curra 9. Justicia cuspidulata (Nees) Wasshausen 10. Justicia dumetorum Morong 11. Justicia floribunda (С. Koch) Wasshausen 12. Justicia gilliesii Asel Bentham 13. Justicia glaziovii Lin 14. Justicia ш (re me kamp) V. A. W. Graham 15. Justicia goudotii : Graham 16. Justicia hassleri a indan) V A, W Graham 17. Justicia hunzikeri Arizi 20. Justicia laevilinguis (Nees) Lindau 21. Justicia lilloana Ariza Espinar 22. Justicia lilloi (J. L. Lotti) C. Ezcurra 3. Justicia lythroides (Nees) V. A. W. Graham 4. Justicia mandoni (Lindau) Wasshausen & C. Ezcurra 5. Justicia oblonga (Nees) Lindau 6 7 — . Justicia oranensis De — o & Ruiz Justicia pectoralis Jac 28. Justicia phyllocalyx (1 DNE Wasshausen & C. Ez- curra 29. Justicia polygaloides (S. Moore) L — 30. Justicia ramulosa (Morong) €. Ezcu 3l. Justicia riojana Lindau : Вицо ta rusbyi (1. — V. A. W. ш z & De Маг 3 A. W. Graham Зб. Justicia tweediana ( lees) саан 37. Justicia xylosteoides Grisebach 38. Justicia yhuensis Lindau APENDICE 2 ÍNDICE DE EXSICCATAS Lista de ejemplares examinados con — de cole "cc io- nista (en orden alfabético) y número de colec mero entre paréntesis corresponde a la especie a la que pertenece cada eje mplar según la numeración en la lista de especies. Los ejemplares tipo se indican con asterisco * Abbiatti y Claps 221 (18). Aguilar 566 (4). Ahumada 4481 (34); 8234 (3). Anderson 1811 (36); 1936 (17); 4044. 12). Arbo y Tur 2214 (20). Arbo et al. 906 (2); 1652 (10); 1657 (20); 2664 (32). Archer y Rojas 4847 (20). Arenas 143 (10); 287 (9); 716 (32); 1097 (10); 1189 (4); 1457 (10); 2283 (27). Arnott s.n.* (20). Bacigalupo 997 (29); 1376 (20). Bacle s.n.* (20). Baer 80 (26). Balansa 2448 (30); 2449 (20); 2451, 2451a (20); 2456 (16); 2459, 2459a (4); 2468 (7); 3296 (30). Balegno 1191 (37). Bang 1215* (30). Basualdo 1876 (4); 2546 (23); 2658 (16); 3485a (13). Beck 5373 (10). Bernardello y Galetto 800 (10). Be — 18720 (4); 18891 (4); 19330 (4); 20142 (8); 20202 (34); 20335 (21), 20425 (22); 20457 (8). Bertoni 1168 (20); 1378 (7); 2433 (20); 2513, 4024 (4). Billiet y Jadin 3057 (15); 3123 (34); 3126 (37). Biu- rrun 584 (37); 1372 (12), Blanchet s.n.* (7); 2575* (1); 2979* (13). Bordas 3748 (32). Bordas y Schmeda 4093 (32). Borsini 613 (30). Botta 138 (34). Brizuela 456 (37); 927 (34); 1132, 1266 (12). Brooke 5586* (19); 5677* 3). Browne s.n.* (7); 2007 (19). Bruch y Carette 24 (12). eee І т (14); 1534 (4). Buchtien 1409* (3). Bura- 34). Burkart 5328 (37); 8498 (36); 12528 (34): : J 13368 — 19998 (355 14171 (2); 14280 — = 27112, 27228 (31): 27954 (26) * ) (37); 31312 (18); 32149 (26); 32659, 34155 (24). Calne та y Sáenz 29211 (4). Carter s.n. (10). Castellanos s.n. (15); (36). Claussen s.n.* (35). Covas 1102 (36). € ristóbal a (22); 442 (12). Cuezzo y de la Sota 1527 (23). Chamisso s.n.* (5). Char- pin y Ramella 21485 (34); 21610 (6); 21733 (22); 21742 (6). Chodat s.n. (20). de la Sota 96 (18); 1138 (15). De Marco 10654 С (4). Degen 1032 (4); 1522 (23); 1611 (23); 3280 (6). Degen y Mereles 3250 (8). Deginani 1135 (38); 1320 (5); 2080 (5). Descole 3187 (7). Di Lullo 27 (34). Díaz 10191 (33). Diers 46 (34); 225 (21). Duarte 1705* (7). Duarte y Pereira 1927* (15). Duré y Brunner 428 (33). ~ Volume 89, Number 2 2002 Ezcurra Justicia en Sudamérica Austral Echegaray s.n.* (12); s.n. (36). Eyerdam y Beetle 22275 (19). d (302 (30). Fe rnánde az 144 (23). Fernánde Z he asas 4114 (14). Fernández С \ 113b (32. Fernández Casas et a 7516 (20). Pernice; a (16). E 4 (10); 5730, 50206 (5); 5929 (25): 5 E (2: " 6081 (7); 6173 (23); 6174 (4). Fortunato 5880 17): Fortunato y Palese 2970 (35). Frenguelli 128 (14). Galetto 186 (33). Giardelli 1173 (34). Gillies s.n.* (12): s.n.* (36). Glaziou 13073* (13); 13076* (1); 21870* (28). го 154 (20): 680 (8); 726 (4). Hahn 787 (10); 1530 (35): 2481 (30). — 138, 916, 1080%, 1138 (10): 1579, 1593 (1): 1650 (10): 1907 (2): 32): 2135 (38); 2754 (8): 2755* (10); 2 2071 (30): 3016 (4); 3017 (32); 3053, 3057 (16 : 3508 (20); 3740 (32); 3920 (7); 4889 (20): 5 5207 (4): 5719 (32): 5984 (20); 5985 (7); 6476 (10); 6813. 6813a, 7106 (2): 7560. 7624 (20): 7633 (29); 7749 (32): 7834, 8003 (9): 8160, 8160a (20); 8423, 8423a (28); 9271 (2): 9508* (38): 11105 (4); 11701 (80); 11740 (4): 12458 ;: 12530 (16). Hayward 1444 (20). Herrera 232 (18). Hicken 12911 (20). Hieronymus y — in 46* (31): Hunziker, ; 4721 (12); 8947 pee 9810 (34): 12322 (12); 13915, 15096 (12); 15313. 160647 (34): 2 : 22457 (37); 22074 (36); 22854. een (31): 34). Hunziker. А. y Caro 13560*, 13562 (17): 118 (34). Hunziker, A. у Cocucci 16388 u2) Hunziker. A. y Maldonado 16245 (17). Hunziker, A. y Subils 24951 (37). Hunziker, J. 950, 959 (11); 10891 (23). Hunziker, J. y Gamerro 11669 (21). ludica у Ramadori 354 (3). Jórgense n 698 (2); 982 (360); 2332bis (10); 2335 (20): 2336 pp (30): 2336 pp. 2337 (B); 2336bis (30): 2338 (15): 2832, 3278, 3796 (7); 4237 pp (20); 4237 pp (29): 4303 — 14792 (34); (32). Kerr 108% (10). Kiesling 1262 (37): 3013 (17); 4345. 1718. 5823 (36): 5012 (12); 5502 (18); 5618 (7); 5838 Н 58H (37): 6402 (31): 6410 (30): 67! 30 (31); 6783 12328 (32); 12525 15161 (25); 17978 (25); ‚ Krapovickas y Cristóbal 11812 (7 14290 (37); 46075 (34). — kas et al. EN 3890 (20); | 1227 00), Kuntze s.n. * (1): s.n. (16). s.n.* 19), s.n.* (29), s.n.* (32). de * 1 1 300% (20). — y Cuezzo 5514 (36) 8403 H 1): 8564c* (26): 15372 (22): 5907 (7): 6052. 6857 (15): 726 1 (36): 7 7521 (36); 8893 (37); 9005 (36): | (30). Lillo 4417 (22): 6082. 7894 (12): 8087* : 10520 (30). Legname у López 8704 (15). Linden 1633* (7). Lorentz s.n.* - t); 12* (37); 118% (36). Lorentz y Hieronymus 301% (34). Luschnath s.n.* (5). Macedo 952* ya [үә (13): 1617* (28). Malme s.n.* (36); 3029* (8). Malvárez 675 (36). Mandon 297* 1). Maranta 295 у ). Marmol 8734 (19). Мае 7 Cro- ES 3376 (26). Martius s.n.* (1), s.n.* (4), s.n. * (5), s.n. (19), s.n.* (39). rig 678 (4). Mathews 3152bis* (9). л As e eda a ыл E ч ^^ © I rg Vied* Mereles 1198 (4); 1208 (7); 2054 (11): 2430 (20); A (6): 3270 (7); 3904 (4). Mereles у Degen и y: 5563 (10). Mereles y — 2795 (6). Meyer s.n. (37); 2004 (7): 2998 (15): 3197 (20): 8570 (36): 8669 (15); 9067 (30); 10198 (36); 10281 (20); 12195, 12789 (37); 13853, 15125* (24); E (20); 16235 (10): 10416 (18); 16492 (34); 21829 30 Montes 841 (25); 7157 AU 9800, , 16383 (25). Moore ps * (29); 941 * (20); 1047* (6). 161 (20): 605 (7): 1719 (10): 7182 (4): 7190 (7). Morello s.n. (34); 966 (24); 16936 (20). Morong 200a (4); 435 (20); 667* (32): 706* (30); 1031 (20); 1538* (10). Morrone 695 (38); 3087 (33). Mroginski 2: 4). Múlgura 675, 738 (17): 927 (15); 1039 (20); 1079 (24); 1891 (5). Muller 172 Nicora s.n. (SI 17632), s.n. (SI 186023) (37); s.n. (SI Charpin 10344 y nell 5368 (18). — 1280 (23). Osten y Rojas 8978 (16). Palmer s.n. (32). Parodi 8511 (10). Pedersen 3132 (32); 5994 (7): 10148 (32). Peirano 9021 (37); 10453 (33). Pie- rotti s.n. (12); 83 (15); 4204 (4); 5131 o 6020 ~ 1172 (22). Pire 3813 (36). Poeppig s.n.* ( ohl s (23); s.n. (35); 196% (5); 1849* (1); 1989* a ч Porta 39 (23). Quarín 367 (20). Quarín et al. 1510 (20). y Piccinini 9408 (36). Re ie 94 k еч Ragonese w أ‎ = Y л o ~ е, ~ c g =. = w Y * — -—— > — — ~~ /; cd N s.n. — 58^ at 1): lf © > hae т 41 2198 (29): 2199 (20); 2 2930 (20): 8070 (30): 97 18 (38): 3): 10403 (30); 10688 (5); 10864 (13): 10929 (1); 2 (22). Romanczuk 188 (2). Ruiz Huidobro 2128 (4); im 50 (20); 3348 (20); 3714 (7); 4249 (4): 4301 (4); — a Leal 11819 (12); 22103 (21). Rusby 1117* (32); 14 20). 166* (14). S. E Herb. Martii 1047* (4). Salz- mann s.n.* (7). Saravia Toledo 2138 (22); 10779 (10). Schinini 8327 (2); 21794 (2); Aq 23450 (25). Schi- nini y Bordas 13314 4 ); 14938 (27); 14973 (6); 16375 (34); 16471 (34); 17793 (33); 17791 (0) 18090 (6); 20344 (14): 20534 (9); 20628 (7); 25242 (7). Schinini y Caballero Marmori 27168 (2). Se — y — ( ros Schinini y Palacios 25788 (27); 25857 Se cum da 866 (34). Se homburgk : s.n.* (7). — o Schreiter 1105 (21); 3509*, 4 192 D 9358 (30): 10866 (10); 11228 Gy 2 (26). Schultz 1095 (36); 3235 (8); 7 (10 " 16244 (4). Schulz 5484 (14); 15846 (8); 18213 (8). Schwarz 94 (4); 1639 (4); 4176 (5); 1310 (38): 5087 (23): 5706 (25); 9275 (4); 9394 (4); 11702 (32). Se hwindt 3065 (25). *ellow s.n. us. s.n.* (4), s.n.* (7), п. (39), 97% (3): 120* + 174* (1); 214%, 252%, y Vervoorst 2425 (30). Smith. Kla in y Sehnorrenberger 11727* (38). Solbrig 51 (36). Solomon 7055. 7056 (28); 14633 (8). Soria 280 (14); 1308 (34); 1727 (14); 2242, 2588, 2037 Vi 2070 (90); 2720 (4); (32); 3303 (7); 3310 (32); 3894 (4). * Zardini y Ortiz 2036 (9). Soriano 828 (12); 845 (34); 2); e (7 i 10152 —* S. col. Y Е <> (37122717 (10 jè 27). Stange s.n. (1 i 6973 eS y 7137 hist Agel Stuckert 17193 (17). Subils 3000 (37). — вл (Th Teague 572 (20). Tressens et al. 5047 $ 1). Troll 1688 (10). Troncoso s.n. vs 60286) (12); | à 58, 27231 (2). Tron- созо у Es — EIE (2). s 10 m Tweedie s.n. (4). (11). s.n. (15). (20). s.n.* (20), s:n." 278 Annals of the Missouri Botanical Garden (36 * (37); 752 * (4); 1159* (37); 1260 p.p.* (12: — Beloperone squarrosa 270 1261: > К (34): 1262* (15). Beloperone tetramerioides 205 Ulibarri 1509 (15). Beloperone viridissima 238 lanni 2577 (34). Vanni y Ferraro 693 (14). Varela 495 Boca de conejo 272 (36). Wasshausen 1955 (3). Vavrek 334. (37); 605 (32 Brazilian plume 241 Venturi 88 (18); 1150 (22); 1724 (15); 1799 (24); 1814 Cyphisia 234 (12); 2686 (34); 2835 (18); 4438, 5060 (36); 5217 (19) Cyrtanther 234 5 (15); 5591 (7); 5853 (18); 5866 (33); 5868 (34); — magnifica 240 6184, 7080 (15); 7826 (12); 8121 (18); 9338 (24); 9631 Cyrtanthera — var. minor 240 (15). Ve — 4221 (22); 4244, 8448 (33). Vervoorst у Cyrtanthera pohlia 240 Cuezzo, 7725 С (3). Villa С e nzo 1119 (12); 1750 (22). Cyrtanthera — var. obtusior 240 Villa Carenzo y Le е 1133 (37). Cyrtanthera pohliana var. velutina 240 lasshausen 1955 We dde] 3117 (22). West 8537 — Chaetochlamys 234 (4). White 550% (3), S ink 133 (3). Wood 10919 (35); Chaetochlamys lillo 259 10923 (35). Woolston 193, 193C (4); 451 (32); 688 (30); Chaetochlamys marginata 254 770 (7); 799 (29); 1383 (7); 1471 (32). Chaetochlamys rusbyi 267 Y barrola 324 (20); 1891 (7); 2866 (2); 3504 (20). Chaetoc — tucumanensis 259 Zardini 4973 (30); 5882 (30); 11349 (4). Zardini y Chaetothy 234 Aguayo 9806 (32); 10622 (32); 10730 (4); 11981 (32). Ce boliviensis 253 Zardini y Soria 4442 (4). Zardini y Velásquez 9935 (20); Chaetothylax cuspidatus 266 9943 (16); 12094 (4); 12954. (30); 13143 (30); 13803 (15). Chaetothylax hatsbachii 200 Zardini et al. 2592, 2593 (10); 12782 (4). Zuloaga 1944 — Chaetothylax huilensis 253 (7); 2111 (11); 2188 (23); 3230 T 5270 (4). Chaetothylax tocantinus 271 Chaetothylax umbrosus 252 APÉNDICE 3 — vestitus 253 ; | PEE Е A Dianth 234 ÍNDICE DE NOMBRES CIENTÍFICOS (NOMBRES VÁLIDOS EN UL — 242 NEGRITA, SINONIMOS EN BASTARDILLA) Y DE NOMBRES Dianihera rraminifolia 257 8 Tu Dianthera laevilinguis 257 Acelica 234 — Dianthera nodosa 239 Adhatoda 234 — Dianthera — 257 Adhatoda flexuosa 274 Dianthera pecto 262 Adhatoda gilliesii 250 Dianthera rice 264 Adhatoda glandulosa 274 Ecbolium 234 Adhatoda tweedianc 27] Echolium comatum 242 Adhatoda twee diana var. angustifolia 271 Ecbolium flexuosum 274 Adhatoda umbro 274 Ecbolium lorentzianum 272 Albahaca de vaca 235.250 Ecbolium pectorale 262 Alfalfillo 258 Echolium tweedianum 271 Alfalfita әтә Ecbolium umbrosum 274 filla 335.272 — Ecbolium xylosteoides 27 Amphiscopic 234 Escoba dura 27: Amphisc opia ae quilabris 2360 thesia 234 Amphisc ора ciliata 240 Ethesia carnea 240 Amphise ора strobilacea 237 Нешгейа 234 Amphiscopia venosa 237 Heinzelia lythroides 259 Bandera espa añola 235, 249 — Heinzelia ovalis 259, 200 Beloperone 234 — Iehiyuyo 256 Be ын гопе =: 247 — Jacobinia 234 Beloperone amhe 239 — Jacobinia — 236 Be e rone ne did "rstiae var. ciliaris 239 Jacobinia « 240 Beloperone amherstiae var. debilis 239 — Jacobinia ciliata 270 Beloperone amherstiae var. graciliflora 239 Jacobinia fest 239 Beloperone amherstiae var. lanceolata 239 Jacobinia — teata 237 Beloperone baenitzii 238 Jacobinia magnifica 240 Beloperone brasiliana 239 Jacobinia obtusior 240 Beloperone cochabambensis 265 Jacobinia pauciflora 249 Beloperone corumbensis 243 Jacobinia pohliana 240 Beloperone hassleri 254 — Jacobinia tenuistachys 238 Belope rone kerrit 247 Jacobinia velutina 237, 240 Beloperone mandoni 200 Justicia 234. Pii matthewsit 267 Justicia acuminata 242 Beloperone pseudociliata 205 Justicia aequilabris 236 Beloperóns ramulosa 265 Justicia alopecuroidea 254 Beloperone riparia 241 Justicia allocota 259 Beloperone scorpioides 272 Justicia amambayensis 245 EX spathulata 274 Justicia anagallis 257 Volume 89, Number 2 Ezcurra 2002 Justicia en Sudamérica Austral Justicia ascendens 257 Justicia riparia Justicia axillaris 237 Justicia rizzint Justicia baenitzii 238 Justicia rusbyana 2 Justicia boliviana 245 Justicia rusbyi 207, : Justicia brasiliana 235, 2309 Justicia saltensis > Justicia campestris 237,272 Justicia sarmentosa Justicia carnea 235, 240 Justicia sarotheca Justicia carthaginenesis 247 Justicia scorpioides Justicia ciliata 270 Justicia schreiteri Justicia cochabambensis 2605 Justicia spathulata Justicia comata 242 Justicia squarrosa Justicia corumbensis 243. 256 — Justicia strobilacea Justicia cuspidulata 245 Justicia teletheca Justicia chacoénsis 241 Justicia tetramertoides Justicia diamantina 272 Justicia thunbergioides Justicia dumetorum 247, 261 Justicia tocantin: Justicia echege i 250 Justicia tucumanensis Justicia flexuosa 274 usticia tweedia 235, 271, 3 Justicia floribunda 235, 249 Justicia umbrosa 252, Justicia furcata 247 Justicia velascana Justicia gilliesii 235, 250 Justicia velutina Justicia glabribracteata 237 Justicia venusta Justicia glaziovii 250 usticia venosa Justicia glutinosa 252 Justicia vestita Justicia goudotii 252.211 Justicia xylosteoides Justicia hassleri 254 usticia yhuensis DTP Justicia humifusa 212 Le ptostach ya 234 Justicia hunzikeri 254 Leptostachya coma 242 Justicia hylobates 273 — Leptostachya martiana 242 Justicia isthmensis 270 Leptostachya martiana var. hispida 242 Justicia jujuyensis 255. 245 Leptostachya martiana var. macrophylla 242 Justicia kuntzei 250 leptostachya parviflora 242 Justicia laevilinguis 257 Libonia floribunda 249 Justicia lanceolata 267 Lophothe cium 234 Justicia leonardii 255 Lophothecium boliviense 256 Justicia lilloana 258 — Lophotecium paniculatum 250 Justicia lilloi 26. Orthotactus 234 Justicia lithospermifolia Orthotactus aequilabris 230 Justicia lorentziana Orthotactus arnottianus 237 Justicia lythroides Orthotactus ciliatus 240 Justicia mandoni Orthotactus lanl 274 Justicia marginata 1 Orthotactus strobilaceus 237 Justicia — 250 Orthotactus venosu. 237 Justicia nemoralis var. tomentosa 256 Palomillo Justicia nodosa 239 Pichanilla usticia oblonga 201 Plume p Justicia obtusifolia 257 Poikilacanthus 203, Justicia odonellii 200 Poikilacanthus glandulosus Justicia oranensis 235. 202 Poikilacanthus phyllocalyx Justicia ovalis 260 Potkilacanthus flexuosus Justicia ovata 25 Poikilacanthus tweedianus Justicia paludosa 257 Psacadocalymma k Justicia panamensis 27 Psacadocalymma comatum 242 Justicia paniculata Psacadocalymma pectorale 202 Justicia paraguayensis Pupilla 234 Justicta pauciflora Quiebrar: 212 Justicia pectoralis Ruellia ee eolata 267 usticia peruvian Rhytiglossa 234 Justicia phyllocalyx Rhytiglossa acuminata 242 usticia platicarp Rhytiglossa anagallis 257 Justicia polygaloides Rhytiglossa axillaris 237 Justicia pringlei Rhytiglossa campestri 272 Justicia pseudociliata Rhytiglossa cuspidulata 245 usticia ra Rhytiglossa hookeriana 247 Justicia reitzii Rhytiglossa laevilinguis 257 Justicia repens Rhytiglossa laevilinguis var. longifolia 257 Justicia riojana Кар oblonga 201 280 Annals of the Missouri Botanical Garden Rhytiglossa obtusifolia Rhytiglossa sei ey var. hirsuticaulis Rhytiglossa panicu Rhytiglossa pector alis Rhytiglossa piahu шеп Rhytiglossa гере Rhytiglossa sarmentosa Ruellia = 'eolata Sarothec ак са elegans Sarothec "ca. glutinosa Sect. Dianthera subsect. Dianthera Sect. Dianthera subsect. Saglorith Sect. Dianthera subsect. Strobiloglossa 267, : 234 Sect. PABA ae r Se ricographis Sericographis macedoana Sericographis macedoana var. elegans Sericographis pauciflora dira didi pauciflora var. speciosor ericographis squarrosa Жей ыша comata Stethoma pectoralis Thalestris Thalestris graminiformis Jehu yuyo Vara de la justicia 25: 235, 241 PLANT DIVERSITY OF THE Peter Goldblatt? and John C. Manning? CAPE REGION OF SOUTHERN AFRICA! ABSTRACT Comprising a land area of ca. 90,000 km”, less than one twe mtieth (5%) the land area of the southern African subcontinent, the Cape Floristic Region (CFR) is. for its size. one of the world’s richest areas of plant species diversity. А new synoptic flora for the Region has made possible an accurate reassessment of the flora, which has an estimated 9030 vascular plant species (08.7% endemic). of which 8920 species are flowe n plants (69.5% endemic). The number of species packed into so small an area is те emarkable for the te mpe rate zor and compares favorably with species richness for areas of similar size in the wet tropics. The Cape region consists of a mosaic of sandstone and shale substrata with local areas of limestone. It has a highly dissected, rugged topography, and a diversity of climates with ү all mostly — in the winter _ and varying from — mm locally to less than 100 mm. Ecological gradients e steep as a result of abrupt differences in soil. altitude, aspect, and precipitation. These factors combine to form an ана шек — of local habitats for plants. * —— rived soils have characteristically low nutrient status and many plants present on such soils have low seed dispersal « spies a jae Pd promoting localized dienten. , Proteaceae, Restionaceae, An unusual family ea inc dune Iridaceae. Aizoaceae, Ericac Rutaceae. and Orchidaceae among the 10 largest familhes i in the flora. following Asteraceae and Fabaceae, as the most spec iose families. ا‎ radiation has resulted in over spec ies falling in the 10 largest families and 77.4% in the largest 20 families. Twelve genera have more than 100 species and the 20 largest genera contribute some 31% of the total species. Spe cies ric — of the Cape flora is — sized to be the result of ge —— and parapatric radiation in an area with a mosaic of different habitats due to local soil, climate, and Ken differences that A combine to produce steep ecological gradients. Also c contributing to the diversity has been a relatively stable geologic al history since the end of the Miocene that saw the establishment of a semi-arid and extreme se asonal climate at the southwestern part of southern Africa Key words: floristics, Mea type climate, phylogeography. plant diversity, southern Africa, spectation. Situated at the southwestern tip of the African subcontinent (Goldblatt, 1978, 1997), the Cape continent between latitudes 31° and 34°30'S (Fig. Floristic Region (CFR) is one of the world’s most 1). the area that has come to be called by biologists — botanically diverse regions. An estimated 9030 the Cape Region has a flora. and to a lesser species of vascular plants (ferns and other vascular extent a fauna (Stuckenberg, 1962), that is so cryptogams, gymnosperms, and flowering. plants), sharply distinct from that of the land immediately the majority of which, some 8920 in total, are flow- adjacent to it that it has impressed naturalists from ering plants, occur there, almost 69% of which are the time of its discovery by European explorers in endemic (figures based on Goldblatt & Manning, the 16th century. Indeed, the floristic characteris- 2000, but reflecting taxonomic changes made since ties of the Cape Region are so unusual that it has the completion of that work). Thus, the flora of the sometimes been regarded as one of the world’s six Cape Region comprises 44% of the estimated floral kingdoms (e.g. Good, 1974; Takhtajan. 20,500 species that occur in all of southern Africa 1986). There are. however, no objective criteria for (Arnold € de Wet, 1993; de Wet, pers. comm.). The distinguishing such “floral kingdoms,” and recog- level of species richness is notable, particularly in nition of a Cape Floral Kingdom is not universal. Africa, the tropical flora of which is relatively de- We use the neutral term “floristic region” here sim- paupératé. but is remarkable for the world's tem- ply for convenience. perate zone, comparing favorably with species rich- Comprising a land area of ca. 90.000 Кар, less ness for areas of comparable size in the wet tropics than 5% of the total area of the southern African rather than for areas of temperate climate. ! Fieldwork in southern Africa was supported by the National Geographic Society (grants 5408-95 and 5994-97), the Missouri Botanical Garden. and the National Botanical Institute. South Africa. We thank Richard Cowling for reviewing the manuse — and for pro yr numerous helpful comments 2 В. rukoff Curator of African Botany, Missouri НИ Garden, Р.О. Box 299, St. Louis, Missouri 63100, LS.A. peter.g асро org. * Compton Herbarium, National Botanical Institute, P. Bag. X7. Claremont 7735, South Africa. manning@nbict.nbi. ac.Zza ANN. Missouni Bor. GARD. 89: 281—302. 2002. Annals Missouri B Garden | NS. T i i ani e M Xe Figure 1. a 2000). Abbreviations: Southwestern; SE, Southeastern. AP, Not only is the plant species richness of the Cape flora exceptional, but its familial and generic composition is remarkable (Bond & Goldblatt, 1984; Goldblatt, 1997; Goldblatt & Manning, 2000). Unexceptionally for a region of fairly dry climate, the largest families are the Asteraceae and Fabaceae, together comprising some 20% of the total species. The families next in size, how- ever, are not matched in any other flora—nowhere else in the world do Iridaceae, Aizoaceae, Erica- ceae, Proteaceae, and Restionaceae assume such numerical significance, except in some parts of Australia where Proteaceae and Restionaceae are also unusually well represented. Other peculiari- ties of the Cape flora are the dominance of fine- leaved sclerophyllous shrubs, the paucity of trees, and a remarkably large number of geophytes, here defined as seasonal herbaceous perennials with bulbs, corms, tubers, or prominent rhizomes (thus subshrubs that excluding shrubs and resprout from a woody caudex, usually after fire). Such geo- phytes, especially numerous among the monocots (notably, in order of importance, Iridaceae, Orchi- daceae, Hyacinthaceae, and Amaryllidaceae), also include many species of Oxalis (Oxalidaceae) and Pelargonium (Geraniaceae) as well as other eu- dicots. Geophytes as so defined comprise slightly more than 17% of the species in the CFR. Con- versely, the Cape flora has a surprisingly low pro- portion of annuals for an area of largely semi-arid climate. Approximately 6.8% of the species are annuals, which is a striking contrast to California Get aN HEIGHT ABOVE SEA LEVEL 20 0 20 40 60 80 100km 22°E 23° 24" 25* 26" The Cape Floristic Region, showing relief, with the phytogeographic centers marked (from Goldblatt & Agulhas Plain; KM, Karoo Mountain; LB, Langeberg; NW, Northwestern; SW, (30% annuals) Kalin Arroyo et al., latitude and climate. Both geophytes and annuals are primarily adapted to seasonally dry climates Chile (nearly 16% annuals) 1994), areas of comparable — and escape the time of year unfavorable for growth by retreating to underground storage organs or by ensuring continued survival only by the produc- tion of seeds. The compilation of a synoptic flora of the Cape Region (Goldblatt & Manning, 2000) replaces ear- lier vegetational analyses based on the work of Bond and Goldblatt — now very much out of date. Statistics presented here taken from Goldblatt and Manning — with minor modi- fications reflec ‘ting taxonomic changes in press or published since its completion. Familial and or- dinal taxonomy is that recommended by the An- giosperm Phylogeny Group (APG, 1998). Changes in the CFR made since this synoptic classification was published are recognition of Veronicaceae (now including Plantaginaceae and several genera previously of Scrophulariaceae), enlargement of Stilbaceae to include non-Cape genera, the reduc- tion of Achariaceae in Flacourtiaceae, and Prion- iaceae in Thurniaceae, and the transfer of Hyaen- anche from Euphorbiaceae to Picrodendraceae (= Pseudanthaceae) (Olmstead et al., 2001, and pers. Chase et al., 2000; Savolainen et al., 00). Tamaricaceae, represented by one species comm.; of Tamarix, was omitted in error from the account of the flora. These changes are represented in the revised familial statistics. Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity PHYSICAL CHARACTERISTICS LANDSCAPE AND CLIMATE Mountain belts of the Cape Region are not par- ticularly high, generally 1000—2000 m in elevation, and although the peaks are well below a truly al- pine zone at the latitude of the Cape region. they are high enough for winter freezing to be a factor affecting the vegetation. The mountains are rugged, and cliffs and exposed rock are evident everywhere. The rugged topography and vertical landscape am- plify the effects of local climatic variation with the result that the mountains offer a greater diversity of habitats than are present in the lowlands. The climate is largely mediterranean, and strictly so in the west. although the eastern half of the CFR receives substantially more summer precipitation, Rain thus falls mainly in the winter months and while summers are hot and dry, they are relatively less so in the east. In areas of low total rainfall, the average monthly precipitation distributed in the summer months may be higher, notably in the Little Karoo, evaporation (P/E) ratios effective rainfall is still but due to more favorable precipitation/ mainly in the winter. South-facing mountain als benefit from summer moisture in the form of rain or fog from the southeast trade winds. The narrow coastal plain in the Knysna area, which has a com- paratively equable climate and high rainfall. sup- ports an evergreen, broad-leaved forest. Likewise. sheltered valleys and locally wet sites throughout the region, affording higher P/E ratios, support for- esl vegetation, Local variation in rainfall is particularly pro- nounced in mountainous areas, and this is impor- tant when precipitation is orographic. Mountain slopes facing prevailing winds receive considerably more precipitation than those in the lee. Rainfall patterns in the Cape Region show dramatic varia- tion in quantity, dropping from 2000 mm per year on the high mountains of the ranges immediately facing the coast, to less than 200 mm on the lee- ward slopes of the interior ranges. Mosaic effects of soils across the entire region are thus compli- cated by variation precipitation from the coast to the interior, as well as changes in seasonality from the west to the east. In addition, elevation and aspect affect precipitation depending on the direc- tion of moisture-bearing winds. The eastern and western parts of the CFR are considered to be un- der fundamentally differing climatic controls, prob- ably a long enduring pattern that has affected the evolutionary histories of the areas and, hence, re- sulting in their distinctive floras and faunas (Cowl- 1999). ing et al., The number of ecological niches available to plant life is multiplied by soil differences, and this is particularly pronounced as precipitation. levels drop. With ample rain the effect of soil on vegeta- Rainfall is limiting almost throughout the region, however, and tional composition is less prominent. vegetation varies conspicuously with soil and avail- able moisture. Climatic gradients are steep, al- though perhaps not more so than in most other ar- eas of mediterranean climate, but the effect may be compounded in the Cape Region by soil diversity. Different soil types in the Cape Region support characteristic vegetation types depending on asso- ciated levels of precipitation. Forest vegetation is typical of deeper soils and sites where precipitation is high and fairly evenly spread. throughout. the year. As soil qualities change and precipitation be- comes lower or more seasonal, forest gives way to shrubby or herbaceous vegetation types. On sandy soils, forest yields to a sclerophyllous vegetation (fynbos) in which species diversity decreases and composition changes until rainfall minimums reach about 300-250 mm p.a. when a succulent: shrub- and becomes dominant. On clay soils forest gives way to fynbos and then to the characteristic renos- terveld, a shrubland dominated by shrubby, micro- phyllous Asteraceae. At precipitation levels below 100 mm p.a. renosterveld is increasingly dominated by succulent perennials. The dissected landscape ensures that broad sweeps of one vegetation type are isolated from others by habitats that will not support their growth. The mosaic of different soil types alone contributes to increasing diversity, but the peculiar nature of nutrient-poor soils may result in more pronounced effects on plant diversity, plant dispersal, and hence gene flow. GEOLOGY AND SOILS Over most of its surface the Cape Region is cov- ered by soils derived from rocks of pre-Carbonif- 1962), 400 mya (Fig. 2). Most of these rocks comprise part of the erous age (King. thus more than Cape System, an ancient Devonian—Ordovician se- ries of sediments consisting of alternating lavers of largely quartzitic sandstones (the Table Mountain and Witteberg Groups) or fine-grained shales (the Bokkeveld Group). surface was folded and warped as Antarctica sep- During the Jurassic the land arated from the south coast of southern Africa and The folds consistently run parallel to the coasts, result- South America rifted away from the west coast. ing in a series of east-west trending mountain rang- es in the southern half of the Cape Region and north-south trending ranges in the west. 284 Annals of the Missouri Botanical Garden Nieuwoudtville Ba let. (Sil Bokkeveld shale (Devonian) [| Superficial sands and limestones (Quaternary) ; Uitenhage sandstones and shales (Cretaceous) Witteberg sandstone (Devonian) Cape ЖЕ, d а eae ee PE on Cape Agulhas Figure 2. Differential weathering of the components of the Cape System has yielded two fundamentally differ- ent soil types, coarse-grained sandy soils, poor in essential plant nutrients, and richer, clay soils of nutrient-intermediate status (Groves et al., 1983). At the low precipitation levels that are usual in the Cape Region, these factors become so limiting that these soils support markedly contrasting vegetation. Apart from variation in nutrient status, the soils depart significantly in their structure and water-re- tention properties. Erosional patterns differ on the two rock types, and the result is that the mountains consist primarily of sandstone rocks and the valleys of shale. Where folding or faulting have been se- vere, more ancient rocks of the Precambrian Mal- mesbury Group are exposed. These are largely shales that give rise to clay soils of the same type as do the shales of the Cape System. Granitic schists are locally exposed in deep valleys and along the west coast, and limestones, mainly of Ter- liary age, are exposed near the coast where they are extensive only along the southern coast from the Agulhas Peninsula east to Mossel B e coastal plain includes areas with la ш, soils derived from reworking of Cape Sandstones. Since these major episodes of folding and rifting in the Mesozoic, only erosional forces have had a ma- jor impact on the Cape landscape, modified only slightly by minor uplift, associated with coastal downwarping in the mid to late Tertiary (King, 1962). Moving from the coastal plain to the interior, Geology of the Cape Floristic Region (adapted from Cowling, 1992). the resultant landscape is a mosaic of coastal lime- stones and deep sands, or valleys with clay soils alternating with mountain ranges of nutrient-poor sands. Local faulting has added a secondary com- ponent of islands of one rock type embedded in another. Both the nutrient-poor and nutrient-inter- mediate soils favor the development of a fairly uni- form, sclerophyllous, shrubby vegetation that is fire-adapted (see discussion under Diversity). WINTER-RAINFALL AND CLIMATIC STABILITY How important to the flora is the current medi- terranean climate that prevails over most of the Re- gion? The vegetation of the Cape Region prior to the establishment of a winter-rainfall climate was very different from that now found there. Evergreen forest has been decreasing since the middle of the Tertiary, and its diversity has dropped dramatically since the mid Miocene, some mya, as well (Coetzee, 1993). Families such as Arecaceae, Cas- uarinaceae, Chloranthacaeae, Sarcolaenaceae, and interaceae, no longer found on the African main- land but still extant in Madagascar, were present in the Cape Region at least until the mid Miocene (Coetzee & Praglowski, 1984; Coetzee & Muller, 1985). In addition, early to mid Miocene deposits on the Cape west coast indicate a fauna adapted to forest and woodland (Hendey, 1982). The establishment of the cold Benguella Current along the west coast of southern Africa in the Mio- Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity cene, with its cooling and drying effects, was prob- ably the single most important influence affecting vegetation change in the subcontinent. Summer drought became increasingly severe in the west as this current strengthened as a result of the spread of the Antarctic ice sheet at the end of the Miocene, ca. 5.5 mya. Even in the later mid Miocene, how- ever, there was a fairly rich subtropical flora replete with palms (Coetzee & Rogers, 1982) near Saldan- ha Bay. on the west coast of the Cape Region in an area that today supports a largely treeless. succu- lent or sclerophyllous shrubland. No palm species occur today in the CFR. Faunal remains suggest that by the late Miocene the once widespread forest and woodland were being replaced by more open savanna (Hendey, 1982). It was probably not until after the beginning of the Pliocene, i.e.. less than 6 mya, that the present Cape flora т be distin- guished, although elements of that flora are recog- nized in Oligocene pollen cores taken within the Cape Region and nearby (Scholz. 1985). It seems that summer drought and lower overall rainfall. was the clear climatic change, including increased driving force for vegetational change in southern Africa into the Pliocene. Although post-Pliocene changes in climate of the Cape Region are poorly documented, the climate appears to have been relatively stable. In compar- ison with southern Europe. North America, and southern South America, all of which experienced cycles of extreme climatic fluctuations with periods of mild climate alternating with extreme cold and dryness (Villagrán, 1994), the climate of the Cape Region appears to have been relatively stable throughout. the Quaternary. Whereas. in central Chile. southern Europe, and North America moun- lain. glaciers developed and winter temperatures f their floras could not tolerate, the Cape appears to = must have fallen to levels that large portions « have merely undergone cycles of drier and cooler The ameliorating affects of large oceans to the south and alternating with wetter and warmer conditions. west would have prevented the extreme conditions that result in major extinction events. Although no glaciers developed, there is evidence of colder cli- mates in the past (Deacon, 1979) consistent with a temperature depression of the order of 5°C at the atitude of the Cape Region. The data of Meadows and Sugden (1991) are among few documenting the history of the Cape flo- ra over the past 20,000 years. Their pollen profiles from the Cedarberg Mountains in the northwestern part of the CFR show no vegetational changes com- parable to those known for Chile (Villagrán. 1994) 1978). or California (Raven & Axelrod, Instead, there seems to have been a series of subtle shifts in conditions that favored one community type over another in the 14,600 years covered in their sam- pling. The Cedarberg data are especially notable because that range lies at the northern, more arid end of the Cape Region, an area that is therefore particularly sensitive to climatic change. In con- trast, the somewhat older fossil wood assemblages from Elands Bay Cave on the west coast of the CFR (Cowling et al., 1999) document a shift from xeric to mesic thicket and fynbos and then to afromon- tane forest at 18,000 BP, indicating higher moisture conditions in the west during the Last Glacia In the southern Cape the lowering of sea €— as much as 120 m below present levels at times during the Pleistocene resulted in the extension of the coastal plain off the southern coast of Africa. The vegetation along the coast at this time was probably grass-dominated, and it supported the dominant alcephaline and equid fauna (Klein, 1977). this area has a large grass component. Even today, the vegetation on clay soils of FLORISTIC COMPOSITION MAJOR FAMILIES AND GENERA The Asteraceae, usually the largest family in flo- ras of arid to semi-arid regions, are also the most speciose family in the Cape flora, with 1030 species Table since the flora was last documented (Bond & Gold- blatt, 1984) now make this family the second larg- Additions to several genera of Fabaceae — est in the flora (previously believed to be fourth argest), which is also unexceptional, as Fabaceae are well developed in most parts of the world. How- ever, the huge contribution made by Iridaceae, Ai- zoaceae, and Ericaceae, next in numbers of species (Table 1), is a unique aspect of the flora (and con- sequently of the southern African flora as a whole). Scrophulariaceae, Proteaceae, and Restionaceae follow in importance, showing a pattern, described in more detail by Goldblatt and Manning (2000). Aizoaceae, with its huge southern African. repre- sentation of Mesembryanthemoideae, are probably the second largest family in the southern African flora (Goldblatt, 1978) and appear to be the largest family in the Namaqualand—southwestern Namibia region (R. Cowling, comm.) that lies to the north of the CFR and also has a winter-rainfall and summer-dry climate. This arid zone has such strong floral affinities to the Cape flora that the inclusion of Namaqualand-southwest Namibia in the CFR to pers. comprise a Greater Cape flora has been given se- rious consideration (Bayer, 1984: 1991, 1997). The relationships of the extended Namaqua- Jürgens, 286 Annals of the Missouri Botanical Garden Table | Ranking of the 20 largest families in the Cape Flora Region as indicated by species number. Family circumscriptions reflect the recommendations of the Angiosperm Phylogeny Group (1998). These families contribute 6989 species to the flora, or 77.4% of the total. Number endemic (% of total of tot Total g a Family Total species species) (number endemic) Species/genus l Asteraceae 1036 655 (63.2) 121 (33) 8.6 2 Fabaceae 761 629 (82.7) 37 (6) 20.6 3 Iridaceae 677 540 (79.8) 28 (6) 24.2 1 Aizoaceae 659 524 (79.5) 76 (18) 1.5 5. Ericaceae 657 637 (96.9) 1 (0) 651 6. Scrophulariaceae 414 297 (71.7) 33 (7) 12.5 da Proteaceae 329 319 (97.0) 14 (9) 23.5 8. Restionaceae 318 294 (92.5) 19 (10) 16.7 9. Rutaceae 273 257 (94.1) 15 (6) 18.2 10. Orchidaceae 227 138 (60.8) 25 (2) 9.1 11. Poaceae 207 80 (38.6) 61 (3) 3.4 12: Cyperaceae 206 101 (49.0) 29 (4) 7.1 13. Hyacinthaceae 191 83 (43.5) 14 (0) 13.6 14. Campanulaceae 183 140 (76.5) 13 (6) 14.1 15. Asphodelaceae 157 81 (51.6) 8 (0) 19.6 16. Geraniaceae 157 9] (58.0) 3 (0) 52.3 17. Polvgalaceae 141 122 (86.5) 3 (0) 47.0 18. Rhamnaceae 137 127 (92.7) 5 (1) 27.4 19. Crassulaceae 134 35 (26.1) 5 (0) 26.8 20. Thymelaeaceae 125 94 (75.2) 4 (1) 31.3 Y = 6989 5244 (77.0) 514 (112) land flora (or Succulent Karoo Region) seem un- questionably closer to the Cape flora in its tradi- tional sense than to the flora of the summer-rainfall karoo (Nama-Karoo Region) although their common boundary is not clearly fixed and there is evidence for its east-west shift in the past (Jürgens, 1991). The large numbers of species of Proteaceae and Restionaceae, seventh and eighth largest families, respectively, are another striking feature of the flo- ra. The importance of Ericaceae, Proteaceae, and Restionaceae both in terms of biomass and species diversity is widely appreciated, but the huge num- ber of species of Iridaceae, predominantly a family of herbaceous, seasonal geophytes, is especially no- table. Nowhere else in the world does this family comprise such a significant floristic component. In- deed, the adaptive radiation of Ericaceae and Iri- daceae in the Cape flora is one of its most striking features. The diversification of Ericaceae, Protea- ceae, Restionaceae, and even Cyperaceae, is close- ly associated with the impoverished sandstone soils of the Cape mountain ranges, and these families are poorly represented on other soils. The massive radiation of Fabaceae and Iridaceae shows no such correlation. Although the wealth of Scrophulari- aceae, sixth largest family in the Cape flora, seems remarkable in a world context, the family is also well represented across Africa, especially in the flo- ras of drier areas (Maggs et al., 1998). In the Cape flora Scrophulariaceae contribute 166 species to the annual flora, far more than does the next most important family, Asteraceae, with 138 annual spe- cies. As circumscribed for the Cape flora (Goldblatt & Manning, 2000) Scrophulariaceae do not include the parasitic and e genera now referred to Orobanchaceae (A 98) but, nevertheless, it remains a major family in the flora. Rutaceae, ninth largest family, are also a surprising aspect of the Cape's floristic composition. Over 95% of the 273 species of Rutaceae there are small shrublets belonging to the tribe Diosmeae, most members of which are restricted to the Cape flora, and reflect the large numbers of shrub species in the flora. The wealth of Polygalaceae, Rhamnaceae, and Thyme- laeaceae, 17th, 18th, and 20th in size, respectively (Table 1), likewise exemplify the importance of the shrubby habit in the Cape flora Poaceae are comparatively poorly represented in the Cape flora. Although they are the third largest family in the flora in number of genera (Table 1), they are only llth in size in total species, with fewer representatives than Restionaceae and barely more than Cyperaceae, the two other families gen- erally that occupy similar habitats. This situation is Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity families of the CFR. Note that the characteristic and nearly endemic Table 2. Endemic family Bruniaceae has 11 genera and 61 endemic species of a total 64 in the family and is not included here, its range extending east- ward to southe rn KwaZulu-Natal. In addition, Olmstead et (2( = = ~ — ) have proposed enlarging Stilbaceae to include non Cape genera rendering it no longer endemic. Number of genera Number Family of species Penaeaceae 7i 23 Sulbaceae s. str. 5 14 Grubbiaceae 1 8 Roridulaceae l 2 Geissol l l l Totals 15 13 paralleled only in southwestern Australia but con- trasts sharply with adjacent southern Africa where Poaceae are prominent and diverse. Poaceae are the largest family in the flora of Namibia (Maggs et al., 1998) and one of the five largest in southern Africa excluding the CFR (Goldblatt, 1978). Although 150 families of seed plants and another 23 of ferns and other vascular eryptogams are rep- resented in the flora, remarkably few account for the bulk of the species. While 23 families in the flora have over 100 species, only 12 have over 200 species. In contrast, 38 families have just one spe- cies each. The largest 10 families account for 5351 species, well over half the total 9030 in the flora, and the largest 20 families account for 6989 spe- cies, over 77% of the flora (Table 1 ENDEMIC FAMILIES The unique floristic composition of the Cape Re- gion, with its high representation of Ericaceae, Ir- idaceae, Proteaceae, and Restionaceae (Table 1). is emphasized by the presence of several families that are endemic or nearly зо. The endemic families are all eudicotyledons of diverse affinity and relatively low evolutionary specialization (Table 2). The larg- est is Penaeaceae (Myrtales), followed by Grubbi- and aceae (Cornales). Roridulaceae (Ericales). Geissolomataceae (together with Ixerbaceae and Strasburgeriaceae, sister to Crossosomatales) (clas- sification following APG, 1998). Based on a molec- ular clock calibrated using 135 mya for the diver- gence of the eudicot lineage (V. Savolainen et al., 2000. have diverged 20 mya from its sister clade, the Af- and unpublished data), Penaeaceae may rican Oliniaceae plus the Neotropical Alzateaceae, while Roridulaceae diverged from Ericaceae, its il., 2000), са. 48 closest relative (Savolainen et a mya. Geissolomataceae appear older, having di- verged perhaps 55 mya from Ixerbaceae plus Stras- burgeraceae. Grubbiaceae may have diverged from the Tertiary, 63 mya. We have no comparable data for Cornaceae plus Hydrostachyaceae in earliest the monotypic, near endemic monocot family Lan- ariaceae. Stilbaceae (including Retziaceae) (Lami- ales) are provisionally regarded as a Cape endemic eg) amily, but if the changes to its circumscription sug- gested by Olmstead et al. (2001, and pers. comm.) are accepte d, the family occurs across sub-Saharan Africa, (see below). Arabia, the Mascarenes, and Madagascar Bruniaceae, one of the distinctive families of the Cape flora, are almost endemic. Of an estimated 64 species in ll genera, just 3 species in 2 genera extend outside the confines of the Cape Region, 2 locally, and 1 as far east as southern KwaZulu-Na- tal. Bruniaceae are now thought to be the sister group to the order (Savolainen et al.. 2000). rank. Dipsacales perhaps meriting recognition at ordinal The discovery of pollen matching modern Bruniaceae in early Tertiary and late Cretaceous (?Senonian) deposits in northern Namaqualand (S. E. de Villiers, the CFR, attest to considerable age for the pers. comm.), well to the north of — amily in southern Africa. The pollen record also accords with Savolainen’s preliminary early ‘Tertiary dating of the divergence between Bruniaceae and Dipsa- cales at about 57 mya (V. Savolainen, unpublished data). Retziaceae (1 genus: | species) have often been onsidered an endemic Cape family (e.g.. Bond & Goldblatt, been in dispute (Goldblatt et al.. 1984), although its affinities have long 1979). DNA se- quence analysis shows the genus nested in Stilba- 1994; Savolainen et al., 2000). Floral specialization for bird pollination appears to The familial status of Stilbaceae is not in question, but Olmstead et al. (2001) included the Afro- Arabian Vuxia in tribe Stilbeae and added the Afro-Mada- Halleria the African Bowkerieae to the family (previously Scrophulari- ceae (Bremer et al., be the source of most of its distinctive features. gascan and southern tribe aceae) rendering Stilbaceae no longer endemic to the CFR or even to southern Afric In contrast to the Cape Region, which alone has 5 (or more likely 4) endemic families, all of south- ern Africa perhaps has just 10 endemic families (9 according to the revised concept of Stilbaceae). In addition to those absolutely restricted to the Cape Region, the southern African endemic families are the eudicots Bruniaceae (ordinal position sister to Dipsacales, 11: ca. 64), Greyiaceae (Geraniales, З), and Rhynchocalycaceae (Myrtales, 1:1); the 288 Annals of the М Botanical Garden Table 3. 9030 species occur in the flora, of which 6208 are endemic Ranking by numerical size of the 20 largest ger vera in the CFR (endemic species number). An estimated (68.7% Erica 657 (637) spalathus 272 (258) Pelargonium thosme Total in largest 10 genera = 1935 spp. (21.4% of flora) Combined total in largest 20 genera = 2851 spp. (31.6% of Muraltia 106 (100) Gladiolus 106 (86 Selago 100 (77 Crassula 95 (26 Disa 92 (78 Ruschia 88 (79 estlo 85 (82 Leucadendron 82 (79 Helichrysum 81 (34 Thesium 81 (35) flora) monotypic monocot genus Lanaria is also currently regarded as comprising its own family, Lanariaceae (Asparagales) (APG, 1998; Chase et al., 2000); and the cycad family Stangeriaceae (1:1). Lanaria is widespread in the Cape Region and extends outside Aitoni- Curtisiaceae, and Oftiaceae have at times its confines a short distance to the east. aceae, been accorded recognition but they are no longer considered to be separate families. They are readily referable to Meliaceae (Pennington & Styles, 1975), 1993), and Scrophulari- 1979), respectively. Prioniaceae Cornaceae (Xiang et al., aceae (Goldblatt, (Poales), treated as a family by Munro and Linder (1998) for the monotypic and largely Cape Prion- ium, and recognized by Goldblatt and Manning (2000) for the Cape flora, are now regarded as be- longing to Thurniaceae (Chase et al., 2000). Behn- iaceae, described for the monotypic southern Afri- can Behnia, is an endemic southern African family (1997), although it is reported to occur in Zimbabwe. However, the status according to Conran et al. of Behniaceae is in question and it is likely to be subsumed in a more widely circumscribed Agava- ceae (M. W. Chase, pers. comm.). Likewise, Achar- iaceae (Malpighiales, current name for Flacour- 2000), two species of which also occur in the Cape Region, are often tiales, 3:3) (Savolainen et al., regarded as a southern African endemic family. The genera are, however, nested in Kiggelariaceae (Sa- volainen et al., 2000) and are regarded here as members of that family. GENERA Some 944 genera of seed plants (or 990 genera of vascular plants), about half of those occurring in southern Africa, are represented in the Cape flora, of which some 160 or 16.3% are endemic (Gold- blatt & Manning, 2000). The level of generic en- demism is modest and reflects little of the unusual nature of the flora. The number of near-endemic genera (those of moderate size with just one or two species extending locally outside the Cape Region) is, however, dramatically high. The largest genus in the flora by far is Erica, with some 657 species lable 3). Changes in the circumscription of Erica —. Oliver, 2000) have now resulted in the inclusion — of all the minor genera of southern African Erica- ceae: Ericoideae, leaving Erica with over 7% of the species in the entire flora. It is unclear whether this remarkable pattern of speciation without generic diversification is associated with the relatively re- cent arrival of ancestral ericaceous stock in the CFR, or with adaptive radiation following the es- tablishment of a semi-arid climate there. By com- parison, the smaller families Proteaceae and Res- tionaceae appear to belong to old African (or even Gondwanan) groups, now poorly represented else- where in Africa, and they show the highest levels of endemism at the generic level. These two fami- lies plus Bruniaceae are the only non-endemic fam- ilies that show greater than 50% generic endemism. Aspalathus (Fabaceae) is the second largest ge- nus, with followed by gees Phylicc Rhamnaceae), Lampranthus (Aizoaceae), and * 2 species, — Geraniaceae), Agathosma (Rutaceae), — alis (Oxalidaceae), each with between 118 and 148 species (Table 3). Thirty-six genera have over 50 species and 13 genera have over 100 species. The 1935 The next 10 largest genera 10 largest genera contribute over 21%, or species, to the flora. contribute an additional 922 species. The 20 larg- est genera in the Cape flora thus contain over 31% of the total species. There is no one unifying ecological pattern evi- dent in the species-rich genera. Agathosma, Aspa- lathus, Cliffortia, Erica, Phylica, and the two larg- est genera of Proteaceae, Leucadendron and Protea, are best developed on sandy soils and are most Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity Table The d represent the percentage of the total flora. The Selected statistics for the CFR and various comparable regions (emended from Bond & Goldblatt, 1984). ten largest genera were determined by species number. Mean sp. number 10 largest Percentage of total flora Annual Region per genus genera Monocots Asleraceae species Cape flora 9.] 21.4 24.5 11.5 0.7 Southern Africa (incl. Cape flora) 9.6 15.1 23.0 11.0 7.0 Cape Peninsula 1.2 17.5 34.6 11.5 9.6 Nata 3.9 17.0 27.1 11.4 ca. 0.5 Eastern North America 5.2 21.8 28.2 12.7 8.7 Europe 1.0 14.0 183.0 12.0 ? California Flora 5.2 15.2 19.2 13.6 27.4 Sonoran Desert 3.3 12.8 12.1 15.0 21.4 Texas 3.9 10.2 24.4 13.4 20.4. Hawaii 1.4 81.0 11.8 15.9 0.04 New Zealand 5.1 26.3 27.9 12.5 6.0 diverse in montane habitats. In contrast, species of ing that elsewhere in the world, is thus no more Lampranthus, Moraea, Pelargonium, Oxalis, Glad- iolus, and Crassula appear to occur with equal fre- quency on nutrient-poor, nutrient-intermediate, or comparatively rich soils and favor lowland habitats. Lampranthus and Crassula are succulents, and Disa, Oxalis, Gladiolus, geophytes, as are some species of Pelargonium. The and Moraea are seasonal remaining genera among the largest 20 comprise mostly shrubs or small trees. A few species of Pel- Crassula, and Helichrysum are argonium, Senecio, annual species. Thesium species are hemiparasitic shrubs. The most obvious shared factors in diversified genera in the Cape flora seem to be either a shrub- by habit or seasonal geophytism. Over 17% of the total species in the flora are geophytes with bulbs. corms, tubers, or rhizomes. The number of species with underground perennating buds would be even higher if plants with woody caudexes were regarded as geophytes. The ratio of species per genus, 9.1 (9.3% ex- cluding ferns) in the Cape flora (Table 4). is partic- — ularly high compared to other floras and is one o the highest in the world (Fenner et al., 1997). al- though southern Africa as a whole. including the Cape Region, has a comparable ratio (9.6%, fide Goldblatt, 1997). A ratio of three dicots to one monocot species in the Flora is close to the average for floras across the world. Although the proportion of monocots does not seem unusual, the monocot families that are represented in the Cape flora are most. distinctive. Some half of the species in the monocot families are geophytic and belong in the petaloid monocot families, notably Iridaceae, Or- and Amaryllidaceae. chidaceae, Hyacinthaceae, The proportion of monocot to dicot species, match- than coincidence. PALEOENDEMIC GENERA The endemic and near-endemic families of the Cape Region are all small and contribute relatively few genera and species to the flora, and they are best regarded as paleoendemics. With the excep- tion of the rhizomatous perennial Lanaria, members of these families are all evergreen, sclerophyllous shrubs. They are often summer-flowering. usually have small flowers, and typically grow on sand- stone-derived soils. We speculate that these plants are relicts of an ancient temperate southern African flora adapted to nutrient-poor soils. Among non-endemic families, these paleoen- demic genera are often members of the geologically oldest communities, tropical thicket and evergreen forest. Especially notable are Platylophus, a mono- typic genus of Cunoniaceae and one of two conti- nental African members of this family. The other, Cunonia capensis, is common in the Cape Region Africa. Other monotypic genera such as Heeria and Lauroph yllus but extends into eastern. southern (Anacardiaceae), Hartogiella and Maurocenia (Ce- lastraceae). Lachnostylis (Euphorbiaceae), Smelophyllum (Sapindaceae) also exemplify the pa- leoendemic component of the depauperate tree el- and ement of the flora and mostly have ranges restricted to the southern or eastern portion of the Cape Re- gion (Goldblatt & Manning, 2000) In non-forest habitats there are few paleoendem- ic genera or species apart from members of the en- demic (and near-endemic) families of the Cape Re- include the monotypic shrublets. gion. These Empleuridium (Celastraceae) and Lxianthes (Scro- Annals of the Missouri Botanical Garden Table 5. flora, central Chile, and the CFR. Figures are percentages Comparison of life forms in the California of total species number; perennials below include geo- phytes and ее нта Data for California and Chile are from Kalin Arroyo et al. (1994). Percentage of total flora Region Trees Shrubs Perennials Annuals Cape Region 2.5 53.3 37.5 6.7 California 4.6 11.0 56.2 30.2 Central Chile 2.9 17.8 63.4 15.8 phulariaceae or Stilbaceae, Olmstead et al., 2001). The small tree, Hyaenanche (Euphorbiaceae or Pi- crodendraceae, Savolainen et al., 2000), also mono- typic, and Metrosideros angustifolia (Myrtaceae), a member of an otherwise Australasian genus, show an odd pattern for the tree flora. Their ranges are restricted to the western half of the Cape Region where there are few tree species. Metrosideros an- gustifolia, the only African member of Myrtoideae: Metrosiderinae, seems as geographically isolated from its closest relatives as Cunonia and Platylo- phus. Like Metrosideros, Bulbinella is also a Cape— Australasian disjunct, but in this case the radiation within the genus has occurred largely in the Cape Region. The dwarf, tree-like monocot Prionium (Thurniaceae), « sandstone soils, is widespread in the Cape region The other f watercourses in nutrient-poor ~ but extends some distance to the east. member of the family is the Brazilian shield genus Thurnia, the pair thus exhibiting an unusual dis- junction. The taxonomically isolated Oldenburgia (Asteraceae-Mutisieae), with three Cape species and one occurring a short distance beyond its east- ern limits, is perhaps another example of that dis- tribution pattern, for its closest allies occur in the Guyana Highlands. The small number of paleoen- demic taxa emphasizes the huge contribution that recent speciation in a narrow range of families and genera has made to the total species diversity in the Cape Region. LIFE FORMS In contrast to other mediterranean floras, the Cape flora has relatively few trees, and this life form only accounts for some 220 species, about 2.596 of the flora. California flora ac- count for 4.696 of the species, but in Chile some 2.996 of the species are trees (Table 5). Most of the Cape flora are shrubs Trees in the remaining elements of the and perennial herbs. The shrubby habit is the most common life form in the Cape Region, accounting for an estimated 4797 species, 53.3% of the total flora. Shrubs are diverse in form, but typically in- clude species with sclerophyllous, and mostly mi- crophyllous leaves, the characteristic that has given rise to the word fynbos, an Afrikaans word describ- ing fine-leaved vegetation. Shrubs also include large numbers of species with succulent leaves (es- pecially Aizoaceae). Stem-succulents include spe- cies of Apocynaceae and Euphorbiaceae, some of which are so reduced in size that the term shrublet hardly seems appropriate. The Cape flora stands out when compared to that of both California and Chile in the overwhelming proportion of shrubs (Ta ble 5). which is largely explained by the nutrient- poor soils that favor this life form. The Cape flora shows some notable differences with other mediterranean floras. One of these is a surprisingly low proportion of annuals—only some 609 species, about 6.796 of the total flora, as com- pared to 30% and 15% respectively for California and central Chile (Table 5) (Kalin Arroyo et al., l 1990). No comparable figures are available for Basin. Al- though the proportion of annual taxa in the CFR is ow, the annual flora is quite species rich. The total » Nx 4; Cowling et al., the Mediterranean number of annual species is actually almost twice as high as the 378 annuals recorded in central Chile, an area of comparable size, and although the Cape has less than half the 1279 annuals in Cali- fornia, its area is only about one fourth that of Cal- ifornia. For its geographic area then, the Cape flora is not depauperate in annuals, but its wealth of other life forms makes the annual habit appear un- der-represented. A small annual flora is also char- acteristic of southwestern Australia, an area that has a recent geological history comparable to that of the Cape Region and a similar pattern of nutri- ent-poor sandstone soils and richer clay. The low proportion of annuals in the Cape flora has re- mained without a satisfactory explanation since it was first noted by Bond and Goldblatt (1984) but the answer may simply lie in the disproportionate numbers of other life forms, especially microphyl- lous shrubs that are particularly well adapted to the nutrient-poor soils. Two families, Scrophulariaceae and Asteraceae, are especially important in their contribution to the annual flora. Scrophulariaceae, with 166 species, contribute the largest number of annuals, and not, as might be expected, Asteraceae, which have some 138 annual species (Table 6). The Aizoaceae, Bras- sicaceae, Campanulaceae, Crassulaceae, Cypera- ceae, Fabaceae, and Poaceae, each contribute be- 20 and to the annual flora. mlaceae, in particular, tween 35 species need taxonomic pal Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity 291 Table 6. Families with more than 10 annual species in the CFR. Total: 609 species (6.7% of the flora). Table 7. Total species (percentage in parentheses) of the different life forms in the Cape flora. Family Species number Scrophulariaceae 166 Asteraceae 138 Campanulaceae 35 Brassicaceae 33 Poaceae 31 \izoaceae 25 Fabaceae 25 Crassulaceae 23 Cyperaceae 21 Gentianaceae 18 Molluginaceae 16 study and our estimation of the number of species in the family, including its annual component, 15 subject to significant revision. In contrast to the low proportion of annuals, the Cape Region has perhaps the highest proportion of geophytes of any part of the world, and is four to five times richer in this life form than is docu- mented for other mediterranean floras (Esler, 1998: Esler et al., 1998; Goldblatt € Manning, 2000). At least 1550 species, over 17% of the total, have spe- cialized underground organs. including bulbs. corms, rhizomes, or tubers and are seasonally dor- mant (Table 7). The overwhelming number of geo- phytes are monocots, with over 1300 geophytic spe- cies, 662 of which belong to one family, Iridaceae. Most of these geophytes are seasonal and lie dor- mant underground in the dry season, but we have Species Species % of Life form number total flora *Perennials 1025 (11.4) ‘Trees 221 (2.4) Shrubs 1805 (53.2) Geophytes 1575 (17.4) Graminoids 795 (8.8) Total, all perennials 8421 (93.3) Annuals 609 (0.7) Total 9030 (100) * Perennials excluding trees. shrubs, geophvtes. and graminoids. included in the geophyte category the few more or less evergreen species (e.g.. Agapanthus, Kniphofia) with similar underground organs. The other main category of the monocots are the graminoids. that is, the perennial species of Cyperaceae, Juncaceae. Poaceae, Restionaceae, and a few other families. which account for 795 species (8.8% of the flora). VEGETATION Far from having uniform vegetation, the Cape Region encompasses five biomes (Fig. 3) and sev- eral distinctive vegetation types. each with its own suites of species and physical characteristics (Rutherford & Westfall, 1994; Cowling & Holmes, 1992a). The most common and distinctive biome is =, heathland, locally called fynbos, an analogue « E (0 Fynbos Karroid shrubland X ES Renoster Shrubland E Dune thicket м Afromontane forest Figure 3. Major vegetation communities of the Cape Floristic Region (adapted from Louw & Rebelo, 1996). 292 Annals of the Missouri Botanical Garden Table 8. Comparison of species richness (indicated by — number per area), endemism, and the proportion of life forms in the floras of the six phytogeographic centers of the e flora (n/a = not available). For comparison, data are listed for the Cape Peninsula, a small area within the — Center. Floristic center Physical Total % inm Geophyte (listed from west area species Species/ species Tree species to east 10% km? number 10% km? endemic species Nein (90) number (9€) Northwest Center 22 4066 184.8 26.3% 69 415 (10.3) 55 (21.2) Southwest Center 23 4661 202.7 32.0% 95 312 (6.8) 846 (18.2) Agulhas Plain 3 1374 158.0 14.9% 24 92 (6.7) 202 (14.7) Karoo Mountain 19 2151 113.2 15.5% 47 130 (6.1) 330 (15.5) Langeberg Center 7 2363 337.6 11.7% 100 127 (5.4) 389 (16.4) Southeast Center 18 2832 157.3 9.7% 163 156 (5.5) 427 (15.1) Cape Peninsula 4.7 2250 178.7 1.59€ n/a n/a n/a Cape Region 90 9030 100.3 08.7% 221 609 (6.7) 5 (17.4) Californian chaparral and Mediterranean maquis. Shrubs with ericoid or short, narrow, often needle- like leaves predominate, but most species of Pro- teaceae, a family common in this vegetation, have broad sclerophyllous leaves. Fynbos typically oc- curs on sandstone soils. A second distinctive veg- etation type, renosterveld, is usually restricted to richer fine-grained soils. It shares few species with fynbos although they often grow adjacent to one another. Microphyllous Asteraceae are common in renosterveld, which consists of a dense shrubland with a rich herbaceous understory that becomes ev- ident after fire or clearing but is often suppressed under a mature shrub cover. Dry sites with rainfall less than 200 mm p.a. support a vegetation of small succulent-leafed shrubs, including many Aizoaceae called succulent karoo. Thicket (a dense, semi-succulent and spinescent and Asteraceae, in a biome evergreen shrubland to low forest) and evergreen forest make up the remaining biomes. Fire is an integral part of the ecology of the Cape Region and accounts for several aspects of the veg- etation (Cowling, 1987). Growth form in mature fyn- bos and renosterveld is a relatively uniform, closed, low canopy of twiggy and microphyllous to sclero- phyllous shrubs. These vegetation types are highly prone to periodic fire. Fire itself has a disruptive effect on the vegetation. It has obviously been a feature of the ecology for so long that there is a large flora of ephemerals, geophytes, other peren- nials, and short-lived shrubs that appear in the years following a fire, often flower profusely, and subsequently disappear, as they are succeeded by longer-lived shrubs. The long-term ecological con- sequence of fire on the flora is the existence of a niche for species that grow rapidly after fire to per- sist and bloom in the immediate post-fire years. This fire-adapted suite of species contributes sub- stantially to the overall diversity in the flora. Ma- ture vegetation is affected by fire in more subtle ways, but fire may cause local perturbations in spe- cies Composition and the elimination of some taxa. DIVERSITY REGIONAL DIVERSITY The patterns of endemism within the Cape Re- gion are fairly consistent, and examination of these patterns in selected genera that have diversified largely on sandstone substrates has resulted in the recognition of several regional centers of endemism (Fig. 1, Table 8). Weimarck (1941) pioneered this field, which has now been refined by Cowling and his coworkers (e.g.. Cowling & Holmes, 1992a Cowling & McDonald, 1999). The presence of these centers suggests that exchange between them is limited because of effective geographic isolation or because different microclimates in each center fa- vor local species at the expense of migrants. One aspect of our account of the Cape flora (Goldblatt & Manning, 2000) has been the formal recognition of phytogeographic centers so that dis- tributional data for species can be analyzed within е ius, we have been able to estimate for the first time the floristic diversity for each center. To some extent the statistics are approximate be- cause some centers are under-collected (or under- cited in taxonomic accounts). We suspect that the Karoo Mountain (KM) and the Agulhas Plain Center (AP) have more species (and thus low- er levels of endemism) than our data suggest Center — Fig. E — At the geographic center of the Cape Region, the Southwestern Center (SW) has the largest number of species (4661) and the highest level of endemism (32%). The Northwestern Center (NW) follows in taxonomic size (4066 species) and endemism (26.3%). The Karoo Mountain and Southeastern Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity Centers have much smaller floras and substantially lower levels of endemism. Much smaller in extent, the Langeberg (LB) and Agulhas Plain (AP) Centers understandably have smaller floras. The Southeast- ern Center (SE), of almost the same physical size as the NW and SW Centers, has a markedly smaller flora (2832 species), This may be explained by its more equable and and only 9.7% endemism. apparently less diverse climate or simply by higher levels of extinction during colder and drier periods Differences in levels of ende- The SW and NW Centers each have about twice the proportion of the Pleistocene. mism across the Centers are striking. of endemic species as the others, which may be a reflection of their greater climatic diversity. The different life forms are unevenly distributed across the Cape Region (see Table 8), and the num- bers of species of the two most distinctive Ше forms, annual species and geophytes, drop dramat- ically moving from west to east. The summer-dry NW and SW Centers have the largest numbers of geophytes and the highest proportion of geophyte species in the Cape flora, 21.2% and 18.2%. re- spectively. These two centers each have over 50% of all the geophytes in the entire CFR. In compar- the other centers have between 14.7% and 16.4% geophytes. Annual species are more. com- mon in the west, and the NW Center has 10.3% annual species. while the remaining Centers each The NW Center alone has 65% of the total annual species in the ison, have no more than 6.8% annuals. Cape flora. The distribution of trees shows the con- verse, with relatively few tree species in the NW Center and the highest numbers by far in the SE Center, which may be explained by its less pro- nounced seasonality and more predictable rainfall. Cowling (pers. Comm.) suggests that proximity to the source pool of species for recolonization during warmer (e.g.. Holocene) periods is probably the most important factor in explaining tree distribu- tions. These patterns seem directly related to climate. Both geophytes and annual species seem best adapted to a seasonally extreme climate with a wet winter and dry summer. A climate with higher, sea- sonally more evenly distributed rainfall, character- of the LB and SE favor a tree flora and fewer geophytes and annuals. Istic Centers, seems likely to Like annuals. geophytes are adapted for survival in semi-arid, seasonal habitats. This is reflected in the grealer representation of annuals and geophytes in the western half of the Cape Region, where summer precipitation is lowest. The unexpectedly high numbers of trees in the SW Center reflect the dis- sected landscape with the presence of fire-sheltered valleys and the regular occurrence of rainfall in the summer along its southern coast and interior. COMPARISONS WITH OTHER FLORAS An aspect of the Cape flora that is of particular interest is the high level of species diversity, both For its size (ca. 90.000 kn). the number of species of vascular plants in the CFR. 9030 (8918 seed plants plus 112 pterido- phytes), is comparable with areas of the wet Neo- tropics (Table 9). Thus Panama (75.000 km?) has 7300 seed plant species and Costa Rica (54,000 Table 9). In fact, southern Africa as a whole has a particularly regional and local. km?) may have over 9000 species (see rich and diverse flora for a predominantly temper- ate region. The area customarily treated for floristic purposes as southern Africa (Botswana, Lesotho, Namibia, South Africa, and Swaziland) has about 20.500 native vascular plant species in an area of 2.074.000 km?, and South Africa alone may have 18.275 seed plants plus 1.221.000 km? (€. de Wet. This is striking compared. with an estimated 19,000 species in all of North America north of Mexico (19,341,000 km). or the estimated 800 seed plants plus ca. 700 pteridophytes) recognized for Peru. an area of 1.285.000 km? (Brako & Zarucchi. 1993: gional context, all of tropical Africa may have about 26.500 species (Lebrun & Stork, 1997), in an area nearly 10 times larger than that of southern Africa some 18,500 species (ca. 245 pteridophytes) in pers. comm.). 16.500 native vascular plant species (15. Zarucchi, pers. comm.). To put this in a re- and about 250 times as large as the Cape Region. Africa and North Africa have mately 21,500 additional species (those not shared Southern approxi- with adjacent tropical Africa). making a total of ca. 47.000 species for the entire African continent. The tiny Cape Region, less than 0.5% of the total area of Africa, then has about one fifth of all the species on the continent. Subtropical southern Africa, ex- cluding the Cape Region, has only about 14.300 species, a figure comparable with that for Tropical For the Af- rican continent then, not only is the Cape Region — T last Africa (Polhill, unpublished data). remarkably rich in species, but southern Africa has a higher species diversity than would normally be predicted given the general trend that species di- versily increases toward the equator. The species richness in the Cape flora is. by any measure, remarkable. Moreover, some 0208 or about 68.7% of the species are endemic there (Та- ble 9). The high degree of species endemism in the Floristic Cape flora compared to the California Province. for example, with some 4240 species. Annals of the 294 issouri Botanical Garden M 9'£0I cel Qr вә OOLZ e2 -— = 97 ALIS єз él 9 “Bo TS0c — — VE втшрлес и 48! él Gol 2 00? — — Ic asauuodo[aq 90681 él OT F2 9OLT 2 — = 6 F2 al) spue[st ивәившәџрәр Vol 11 08 F2 OOS LT F2 = OOOT F2 ТӨС 1воѕедереу VL € 18 9661 ОТ 666 897 рив[294 мәм 9°89 IT “OL 98 8811 SI L9G 9'9[ емен 5рир[в1 әјвләйшәј 10 perdoj, CLO 6 cr F2 0062 == 0081 SL вшвивд L991 8 0006 "BJ = 1181 YS BOTY 81509 sa1do1] 19M о] ISIOJA 9'6c L SL TD 0008 F2 10 Cel 9 89 0€9€ Og F2 COV OLE biſpasnv шәјәмпос Vol © VES 9? LESS = = SSI c O'S Y CGE ВӘ COLE — [6S TOI 1 91149 189997) Vel € EL Over G9 908 VEE 22ut^01d әпѕцор erojen £001 1°89 0£06 rol 066 06 eaujy YINdg ‘uoIsay ade?) әјешцә URVURLA}IPeul JO ввәлү gel c TIE OOS 9T Uc Olcc €8cI Mod 97 I €708 2960 [ША OFT? T29c eoujy шәщшпос searP [eJuaunuor), UY OT 32u219J29M Y satoeds ж в1әиәЗ Jo ¿MY «0T /sotoadg 2muapuo [F10], әїшәриә 19QUINN] Bale satoeds в1әиә‹) [Bossy (2661) 122200 Y [EPIN “ZI (9661) 1° 19 z `6 їсшшоә "srod) шпАрлсу y јәшшең ^g :(z66[) 19ddog ^7 :(0761) P94 `9 (7661) ѕәләтлвә) y ooy uey С YS “LL 40661) TP 19 19U (1661) 192€ “OT (2861) AAV. “(P661) [E 19 олошү uey ^t (8/6[) родәхү Y USAPY ^e ‘(wwo ләй) rgoonaez *(e66[) nqoonaez y oyeg `g (wuwo *s1ad) Jay әр 77) `1 ‘SQOUIIIJIY "ppro« әці jo ѕиог8әл p3122[as ut sjue[d re[nose4 әлцви јо wsiwəpuə jo uosueduio") сб әү4еј, Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity 1978). under- scores the peculiarities of the Cape flora (Table 9). 47.7% endemic (Raven & Axelrod, Such levels of endemism are usually associated with islands that have been isolated for long periods of geologic time, or with areas that have very sharp boundaries limiting direct plant migration. In a bi- ological sense, the Cape Flora Region is virtually an island, not surrounded by ocean, but by a zone of dry climate or sharply different soils, or of sea- sonal rainfall distribution. The flora of southwestern Australia shares with the Cape Region an unusually high endemicity for a local continental flora (Beard, 1970: Hopper, 1992), and so does that of southern Africa as a whole (Table 9). Why the latter region should have such a high level of endemism is not at all clear. The high level of diversity and local endemism of the CER is starkly emphasized by comparison of the phytogeographic centers of the region with other The SW 4000 species and over and NW Centers each have over 31% or 26% endemism, re- spectively, compared with about 2400 species and areas. an estimated 22-23% endemism (depending on the geographical definition of the region) for the entire mediterranean flora of Chile (Table 9), an area over five times greater than either of these. phytogeo- graphic centers of the Cape flora. Likewise, impor- tant centers of local endemism (hot spots) within the Mediterranean Basin, including the islands of Sicily, Sardinia, and Crete, or the Peloponnese Pen- insula, all of more or less comparable size to the NW or SW Centers, have approximately half or less than half the number of species and between 6% and 12.5% endemism (Médail & Quézel, 1997). ALPHA DIVERSITY Aspects of plant species diversity have been ad- dressed on several levels. and it has been shown that at the local level selected areas within the unusual on a world scale 1992; Goldblatt, 1997) and are less species-rich than many areas sampled in the New 1988a. 1988b). Patterns of alpha diversity (the number of Cape Region are not (Cowling et al., and Old World lowland tropics (Gentry, species in a homogeneous community, e.g., Cowling et al., 1992) in a range of vegetation types in the Cape Region, including fynbos, renosterveld, forest thicket, and evergreen forest, are surprising. Fyn- bos sites (with seasonal species not included in species counts) have a mean alpha diversity of 68 species рег 1000 nv, with 121 the highest number of species recorded at any site (Cowling et al., 1992). and data indicate that diversity is higher in the west Fynbos diversity is by no means uniform, than the south and in sites of intermediate produc- tivity than in either more mesic or drier sites (Bond, 1983). Non-fynbos sites have been less well studied. Figures in the literature for renosterveld include means of 66 (Tilman et al., 1983) and 84 (Cowling 1992) species per 1000 nr. Forest thicket sites have a mean of 59 species per 1000 nr, forest et al.. sites have ranges of 44 to 52 species, and succulent karoo shrublands a mean of 43 species in the same area (Tilman et al., 1983). Comparisons made by Cowling et al. (1992) in- dicate that California chaparral communities have alpha diversity levels around 34 species per 1000 m? but the more comparable southwestern Austra- lian region has an average of 69 species per 1000 m? in heathland (kwongan), a vegetation type sim- ilar to fynbos. While these figures are consistent with higher total species richness in the CFR and southwestern Australia compared with the Califor- nia Floristic Province, they do not explain the com- parable regional diversity of the CFR and the low- land wet tropics. In the wet tropics mean alpha diversity of trees (including woody lianas) alone has been found to range from 129 species (Africa) to 140 species (Neotropics), to 193 species (Asia) per 1000 m? (Gentry, 1988b), about twice the alpha di- versity found in mediterranean communities on nu- trient-poor soils. Inclusion of epiphytes and other herbaceous plants raises alpha diversity in tropical sites. Gentry and Dodson (1987) have shown that a major component of the plant species diversity in tropical forests actually lies with the epiphytes. Similarly the inclusion of seasonal geophytes would increase the figures for the Cape flora. and until more Comprehensive surveys are made any com- parisons can only be tentative. BETA AND GAMMA DIVERSITY Beta diversity (species turnover along a habitat Cowling et al., 1992) is Cowling (1990), for example, has reported nearly complete replace- or environmental gradient, relatively high in the Cape Region. ment in sites along the Agulhas plain that differed - = | soil features but were climatically and topo- — similar. Differences in composition be- tween Communities on sandstone versus clay soils are so nearly complete that the plants on these two soils are treated as belonging to different vegetation types. Gamma diversity (species turnover in equivalent eradients, also called 1992) is likewise ex- tremely high in the Cape Region, and is reflected habitats along geographic delta diversity, Cowling et al.. 296 Annals of the Missouri Botanical Garden in the high levels of regional endemism. Species replacement values of 46-70% have been reported by Kruger and Taylor (1979) for sites 25 km apart, and Linder (1985) has suggested that geographic replacement may account for 3046 of the differenc- es in species composition along geographic gradi- ents in similar habitats. These figures are, however, lower than some estimates for lowland Neotropical sites (B. Boyle in Goldblatt, 1997). REASONS FOR CAPE FLORISTIC DIVERSITY Richer in species than any other temperate flora and most tropical ones occupying comparable phys- ical area (Table 9), the Cape flora is also highly distinctive. One of five regions in the world with a mediterranean climate, the Cape has substantially more species than do either California or central Chile, which are substantially larger in land area. Although southwestern Australia has a flora that may have about the same number of species as the Cape Region (Table 9), it is at least three times larger in area. Only the Mediterranean Basin, ap- proximately 25 times larger in area, has a flora that is larger than the Cape flora, with about 2.5 times as many species. A formal comparison of these pat- terns is provided by Cowling et al. (1996). The rea- sons for the substantially higher species diversity of the Cape Region compared with these floras are several, and include a range of factors, both phys- ical and historic. A diversity of soils, a rugged landscape, and ex- tremely variable and complex rainfall patterns have combined in the Cape Region to produce a mosaic of sharply different habitats that lie in close prox- imity to one another in a pattern repeated across its entire area. This high physical heterogeneity, al- though striking, is not unique to the region, and may even be greater in other regions. For example, the California Floristic Province has a wide range of soils, including serpentine substrates not present in the Cape Region, diverse climates, a rugged to- pography with higher mountains than those in the Cape, plus a wider latitudinal extension. Likewise, the Mediterranean Basin, orders of magnitude larg- er than the CFR, has a wide diversity of soils and a rugged landscape, with the mountains also higher than those found in the CFR. Both areas are often cited as being species-rich compared to neighbor- ing regions, yet both are substantially poorer in species than the Cape Region, California in abso- lute number, and the Mediterranean area in sub- stantially less alpha diversity per unit area. South- which has a flora western Australia, in contrast, approaching that of the Cape in size, noticeably lacks the rugged topography of other Mediterranean regions, although it exhibits some of the ecological features of the Cape Region. Physical heterogeneity alone cannot therefore account for the richness of the Cape flora, and edaphic factors as well as his- torical biogeography may be more significant. Of the five mediterranean regions of the world only the Cape and southwestern Australia have soils that include large areas of nutrient-poor quartzitic sands, and at least part of the explanation for the higher species numbers here has been thought to relate directly to the particular effects of this substrate on plant life. High levels of local di- versity have been considered characteristic of nu- trient-poor soils (Tilman, 1982, 1983) and if this is correct, then the mere existence of large areas of such soils should account for the comparatively high alpha diversity of heathland vegetation in both South Africa and southwestern Australia compared with that in California or central Chile unless high levels of local and regional richness are not cou- pled. This hypothesis is, however, not supported either in the Cape or southwestern Australia, where alpha diversity levels on nutrient-poor and nutri- ent-intermediate soils appear to differ only mini- mally (Goldblatt, 1997). Although nutrient-poor soils may not support a flora significantly richer than that occurring on soils of other nutrient status in the Cape, they are able to maintain particularly high beta diversity in the associated fynbos vege- tation, both on geographical gradients and on slightly different soils under the same climatic con- ditions. Comparable data for nutrient-intermediate and nutrient-rich sites are not yet available. Nutri- ent-poor soils in mediterranean climate zones have a higher proportion of reseeding versus resprouting shrubs (Wisheu et al., 2000), and these authors ar- gue that this directly contributes to higher diversity because the high frequency of fires that destroy re- seeders would lead to their shorter generation times and thus higher speciation rates. The high frequen- cy of fire in areas of low-nutrient soils is thus an- other aspect that may contribute to diversity in the Cape Region. Fire is also significant in southwest- ern Australia and California but not in the medi- terranean zone of Chile (Kalin Arroyo et al., 1994). The combination of edaphic and topographic di- versity, steep local climatic gradients, peculiar nu- trient-poor soils, and frequent fires, although un- doubtedly important in promoting species diversity in the Cape Region, is still inadequate to explain the presence of the unusually rich flora in the Cape Region, particularly when compared to southwest- ern Australia. A notable and perhaps crucial dif- ference between the Cape and other areas of med- Volume 89, Number 2 2002 Goldblatt & Manning 297 Plant Diversity iterranean climate, possibly excluding southwestern lies in their Pliocene-Pleistocene cli- Available data indicate a history for Australia, matic southern Africa very different from that experi- enced in the Northern Hemisphere and to a lesser extent Chile (Villagrán. 1994). cold and aridity alternating with warm wet phases history. Cycles of extreme made these areas largely uninhabitable by anything resembling their current floras, elements of which either became extinct or were locally restricted to sites of relatively mild climate. A similarly dynamic history for the Cape flora has been hypothesized in which Pleistocene glacial cycles caused a north- ward shift in the winter-rainfall zone, which in turn caused a northward expansion of the flora during the pluvial periods but local extinction and its re- striction to refugia during dry periods (Axelrod & Raven. 1978). The limited evidence available for the Cape Region, however, indicates a more mod- erate climate without changes of such cataclysmic dimensions. modest. shifts in the flora (Meadows & Sugden, 199] sitive Cedarberg mountains. Pollen cores reflect. comparatively . even in the semi-arid and ecologically sen- Changes there might — be expected to have been more severely influenced by the drier and colder climatic conditions that are postulated to have occurred during glacial periods. when belts of vegetation adapted to mediterranean climates contracted away from the dry zones that lay toward the equator. The absence of any evi- dence of major changes in the vegetation of the Cape Region during the Pleistocene suggests that the glacial cycles did not have the catastrophic ef- fects on plant life in southern Africa that they did in the Northern Hemisphere or Chile. In the CFR relatively drier and wetter cycles may simply have induced changes in the local composition of vege- tation, perhaps causing limited extinction, which in turn created opportunities for speciation. The unusually high species richness of the Cape flora is thus. in all likelihood, a consequence of sustained climatic stability and reliability, enabling a more or less uninterrupted evolution and diver- sification of the flora to occur in a region of high physical complexity. The history of this evolution can be traced to some extent by considering the Although there are few such studies for plant groups centered modes of speciation evident in the flora. in the Cape flora, the available evidence suggests that parapatric speciation linked to substrate or mi- croclimatic differences is an important mode of speciation in some families (Linder & Vlok. 1991: Cowling & Holmes, 1992b: Goldblatt & Manning. 1996). Vicariant species exhibit differences in ecol- ogy such as edaphic, microclimatic, seasonal, or pollination characteristics. For example. vicariant species of Rhodocoma iBastionae eae). a genus re- stricted to nutrient-poor sandstone soils in montane habitats, favor different habitats and are not signif- icantly isolated geographically (Linder & Vlok. 1991). Parapatric speciation also appears to have geographic isolation in 1992b) and in the genus Lapeirousia (Iridaceae) in the CFR been more important than Agulhas Plain shrubs (Cowling € Holmes. 5 8 and adjacent parts of the southern African west coast (Goldblatt & Manning, 1996). High levels of both beta and gamma diversity likewise support the hypothesis that microgeographic speciation has played a major role in speciation in the Cape flora. Nearly adjacent habitats with the same climatic and topographical conditions, which differ only in their substrates (coarse sand or fine sand or limestone). can support plant communities that differ radically in their species composition while still being broad- ly similar in family and generic composition (Cowl- ing & Holmes, 1992a) Parapatric or microgeographic speciation may actually be the rule not the exception in plants 1993), and is favored by reduced gene flow f contrasting substrates that characterizes the CFR (Levin. across strong selection differentials. The mosaic appears to provide such a strong selective differ- ential. Although the Cape flora may not differ from other floras in mean pollen dispersability (Linder, 1985). have low seed dispersal distances. The majority of many of its most characteristic elements species in the Cape Region show no evident ad- aptations for seed dispersal and are regarded as passively dispersed, with estimated seed dispersal distances under 5 m (and most likely much less than this) (Linder, 1985). Dispersal in most Aizo- aceae is by rain drops falling on hydrochastie cap- sules, and this mechanism, although an active one. results in very small dispersal distances (Desmet & Cowling, 1999). disproportionately well represented in both the Cape Region (Bond & Slingsby, 1983) and Austra- lia (Berg, 1975). Some 1000 Cape species, notably Proteaceae. Restionaceae. Active seed dispersal by ants is in the families Fabaceae., Rhamnaceae. and Rutaceae, produce seeds with lipid bodies (elaiosomes) that are attractive to ants, and an undetermined additional number are trans- ported to underground nests by harvester ants. In vegelation types prone frequent fires, such as fynbos, the burial of seeds is a valuable adaptation 1992p). distances for ant-dispersed seeds are also short. up 6 m (Linder, 1985). strategies are relatively restricted in their occur- (Cowling & Holmes, However, dispersal More effective dispersal rence. Wind-dispersal is characteristic of many As- 298 Annals of th Missouri Botanical Garden teraceae and Orchidaceae (which have compara- tively low levels of endemism in the Cape flora), while long-distance dispersal involving flying ver- tebrates (birds and bats) is least common, and is especially rare in plants on nutrient-poor sub- strates. There is an assumption that plants on such soils cannot afford to allocate resources to protein- rich berry or drupaceous fruits (Bond & Slingsby, 1983). The low frequency of fruits with burs and spines, adapted for exochory, suggests that dispers- al by non-flying vertebrates has long been unim- portant in the flora, perhaps because the fauna has historically been a small one in terms of numbers of species and individuals. This may be due largely to the unpalatability, low nutrient status, and low productivity of the flora as a whole. Indirect evidence of the importance of reduced gene flow distances in stimulating local species di- versity Comes from a comparison of the number of species and their level of endemism between taxa with widely dispersed seeds and those with reduced dispersal distances. Genera with fleshy diaspores or those that are well adapted for wind dispersal tend to have wide ranges, few species per genus, and low levels of local endemism. Compare the ber- ry-fruited Nylandtia (Polygalaceae). which has two species, with its relative Muraltia, which has dry fruits and over 100 species, most with narrow rang- Similarly, Chasmanthe (Iridaceae), which has fleshy or deceptive (brightly colored) seeds, has two widespread species and one localized one, whereas its close relative Tritonia, which has dry seeds, has es. 16 species in the flora, most of them with narrow ranges. Another striking example is Chrysanthe- moides (Asteraceae), which has seeds enclosed in a fleshy pericarp. The two species of the genus ex- tend throughout the Cape Region and one far be- yond it into tropical Africa. The numerous species of the closely related genera Osteospermum and Tripteris have dry seeds and mostly have smaller geographic ranges. This comparison is also instruc- tive at the family level. Low seed dispersability is typical of many of the larger and most characteristic families in the flora with high ratios of species to genus and high levels of local endemism. Erica- ceae, Iridaceae, and Fabaceae, which largely lack highly developed mechanisms for long-distance seed dispersal, have ratios of above 20 species per genus and higher than 80% endemism at the spe- cies level. Levels of local and regional endemism for Asteraceae (63% endemic species) and Orchi- daceae (60.8%), mostly with wind-dispersed seeds, and Poaceae (38.6%) and Anacardiaceae (32.1%), many with fruits adapted to wind and/or animal dis- persal, show levels of local and regional endemism below the mean for the entire flora (68.7%). The massive speciation in the Cape flora is, we suggest, most likely explained by a model of local speciation in the absence of catastrophic climatic or topographic perturbations. It appears that a rel- atively stable climate prevailed in the Cape during the Pleistocene and that local parapatric speciation across steep environmental gradients may account for a considerable proportion of the speciation events that occurred in the CFR. Because of this relative stability it seems reasonable to postulate that extinction rates in the main vegetation zones, fynbos, renosterveld, and succulent shrubland, were low and more than compensated for by local speciation events. The nutrient-poor soils scattered in a mosaic across the CFR must have stimulated local speciation rates because of the characteristic low vagility of the seeds in the great majority of the plants adapted to these soils. Likewise, the low va- gility of many of the species in Succulent Karoo, although likely a result of different selective forces, has the same consequences, with high levels of lo- cal speciation, and thus high levels of diversity across geographic, environmental, and edaphic gra- dients (e.g.. Cowling € Holmes, 1992}; Cowling et al., 1998; Desmet € Cowling, 1999). The unusually high diversity of the Cape flora is matched by its extraordinary composition of fami- lies and genera. Instead of a balanced flora with relatively small numbers of species per genus there has been massive local radiation in a series of un- related genera. This is so pronounced that almost 22% of the total species in the Cape Region fall in just 10 genera, while the 20 largest genera account for over 30% of the total species (Table 3). Typical examples of these genera are Erica (over 650 spp.). Aspalathus (272 spp.). Agathosma (143 spp.), Phy- lica (133 spp.). and Cliffortia (114 spp.). Signifi- cantly, none of the genera that display a pattern of massive local radiation in the Cape Region are en- demic there, but rather they extend northward into Namaqualand, eastward into southern KwaZulu- Natal, or even further to the northeast into tropical Africa. They are, however, primarily restricted to nutrient-poor soils wherever they occur. In sharp contrast to these examples are the en- demic families of the Cape flora, which are without exception depauperate in species, although they may contain several genera. These families are typ- ically restricted to montane habitats in acidic sand- stone soils, and many of their constituent species are highly local endemic plants of particular moun- tain chains or peaks. They display the character- istics of paleoendemism and probably represent el- Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity ements of a previously more widespread southern temperate flora adapted to nutrient-poor soils in a summer-rainfall regime. It is probably no coinci- dence that most of these species flower in summer, and are thus out of phase with the spring flowering peak of the flora. With the development of a winter- rainfall climate in the Pliocene (e.g... Coetzee, 1993), it is reasonable to infer that these pre-Cape flora elements were gradually restricted to mesic sites in which some moisture was present over the summer months. Concomitantly it appears that oth- er elements of the flora were able to radiate into emerging niches, thereby establishing the huge neo-endemic element of the flora. The highly scler- ophyllous or microphyllous habit developed by taxa adapted to nutrient-poor substrates can thus be re- garded as an important pre-adaptation to the med- iterranean climate. The hi the mediterranean climate, as well as its regularity, would also have favored families and genera with a geophytic habit, for example, Amaryllidaceae, Hyacinthaceae, Iridaceae, and Oxalidaceae. The rapid and extensive radiation of plant taxa in the CFR must have been favored by both the emergence of new habitats through climatic change and the exposure of the coastal plain as sea levels fell at times during the Pleistocene (Coetzee, 1993), At the same time the flora was increasingly isolated by the as well as by the stability of the climate. winter wet and summer dry climate regime from recruitment from the summer-rainfall-adapted flora of adjacent parts of southern Africa, which largely lack the low- or nutrient-intermediate soils so char- acteristic of the Region. of the ( diversity of flower form final characteristic Cape flora is the great, and often extreme, that is a feature of many of the genera. This is linked to a diversity of pollination strategies, many of which are poorly exploited outside the region. In particular, pollination by sunbirds, long-proboscid flies, hopliine beetles, rodents, and the butterfly Aeropetes аге more extensively Haw ni in the a (Manning & 1998; Goldblatt All these strategies favor dif- Cape flora than. elsewhere in Goldblatt. 1996; Goldblatt et T & Manning. 1999). ferently shaped. large. brightly colored flowers. Pol- linator diversity is likely to be part of the expla- nation for the high species diversity in the CFR, particularly for certain families in the flora. Both Iridaceae and Ericaceae, for example. have adopted a range of pollination strategies not evident or only weakly expressed elsewhere across their range (Vo- gel. 1954: Goldblatt & Manning. 1996, 1998; hardt & Goldblatt, 2000), and are often particularly striking when in bloom. ; Bern- In Gladiolus, a prime ex- ughly seasonal nature of ample of a genus with diverse pollination systems and represented in the CFR by 106 species. the genus exhibits no less than 27 shifts in pollination system in southern Africa, most of these in the CER (Goldblatt et al., 2001), and repeatedly evolved ad- aptations for long-proboscid Ну, sunbird. moth, and butterfly pollination, which must have played a ma- jor role in its radiation. The short season favorable for both plant and insect growth is probably the overriding factor responsible for the diversification of pollination strategies and, more than anything has made the else. Cape flora so extremely ap- pealing to human sensibilities. CONSERVATION Like other parts of the world with species-rich floras and an expanding human population, there are varied and serious threats to the Cape flora. Expanding agricultural activity for growing food or to supply other human needs has transformed low- and areas near population centers to the extent that little or 1992). Moreover, the peculiarly local distribution no native plants remain locally (Rebelo, of many Cape plants (high beta diversity) means that a higher proportion is imminently threatened with extinction than would be the case with wide- spread species. The number of species known to be lost forever is relatively low, but the number of spe- cies represented by single, reduced, and diminish- Taylor (1978 species in the ing populations is alarmingly high. ' estimated that some 500 endemic CFR were threatened, were known to be extinct (0.07%). while at least 60 species Red Data Book numbers for southern Africa in 1985 include 1320 Cape plants (14.67% of the total) (Hall & Veldhuis. 1985). Criteria for classification of threatened spe- cies vary: Rebelo (1992) estimated 218 threatened species (extinct, endangered, or vulnerable) in the CFR (ca. 4.8%), slightly less than half the figure provided by Taylor, but nevertheless, considerable. In 2001 these figures are certain to be higher as ^ result of the expanding human population. de- spite ever more sophisticated conservation activity. \ peculiar threat to the Cape flora is the spread of alien vegetation, largely woody species of Austra- lian Acacia and Hakea and European Pinus. These species rapidly invade montane and lowland areas not currently suitable for agriculture, thus com- pounding the threats of agricultural activity and ur- ban growth. The threat posed by this alien vege- — ation is being vigorously countered by biological control methods and manual removal, but the prob- lem is immense and requires constant monitoring. Kruger (1977) estimated that 60% of fynbos had 300 Annals of the Missouri Botanical Garden been replaced by alien vegetation, agriculture, or urban development. ‘Twenty-five years later, this fig- ure is bound to have increased. Concrete examples may better illustrate the sit- uation. Rebelo (1992) estimated that of 306 endem- ic species of Proteaceae 65 were threatened (de- vulnerable) (ca. 21%), while 131 were treated as Red Data species. The pattern in other families is similar, although fined as extinct, endangered, or slightly less serious, with 10.5% of endemic Iri- daceae, 996 of endemic Rutaceae, ca. 396 of en- demic Ericaceae, ca. 2% each of endemic Astera- sae threatened (Rebelo, 1992). The number of species of each of these families includ- ceae and Fabac ed in the Red Data Book is substantially higher (Hall & Veldhuis, 1985 A detailed analysis of reserves and other — con- servation areas in the CFR is presented by Rebelo (1992). substantial areas for preservation of diverse vege- Recent conservation efforts have secured tation types in the CFR and a framework for a con- servation plan to establish reserves to cover por- tions of all vegetation types has been developed 2001). Although such reserves cannot include populations (Cowling & Heijnis, 2001; Cowling et al., of all endangered species because of their erratic distributions, future loss of distinctive vegetation types will be limited and preservation of substantial tracts of each will in the future be enhanced. Literature Cited еа Phylogeny — (APG). 1998. An ordinal sification for the families of flowering plants. Ann. Missouri Bot. Gard. ч 531—553. “Н. & B. К de Wet (editors). 1993. Plants of Africa. Names and Distribution. Mem. Bot. Surv. S. Africa 62. National Botanical Institute, Preto- ria. Axelrod, D. I. & P. H. Raven. Tertiary vegetation history of / . A. Werger — ta and Ecology of Southem Africa. W. Junk, The Hague. Bayer, M. B. 1984. m Cape flora and the karoo—A win- ter — اا‎ versus a fynbos biome. Veld & Flora 70(1): Beard, — pro^ (editor). An Annotated Checklist of the Plants of bó stern Australia. Kings Park Board, Perth. Berg, R. Y. 1975. Myrmecochorous «кү in — and their dis pera by ants. Austral. J. Bot. 23: 475-508. Bernhardt, P. & P. Goldblatt. 2000. The aeui of pol- lination mec inns in the — of southern Afri- ave ые Cretaceous and 1 Pp. 77-130 in М ca. Pp. 301- 1 K. Wilson & P. Weston (editors). Systematics e Bios of the — ‘ots. Royal Botanic ardens, Sydney, Australia devis of the ( в : lora: African Bot., Suppl. 13 On — dive rsity and the ric [ав of the Cape боа: А study in ugs rn. Cape — P — 13 in F J. Kruger, D Mitchell & J. mate Ecosystems. Springer-Verlag, Berlin. ———— & E. Jarvis (editors), Mediterranean-Type oe ms. The Role al pias nts. Springer- Verlag. Be — —— & P. Slingsby. Seed need by ants in — a the Cape и and its evolutionary implications. S. African J. Sci. 79: 231-233. Brako, L. & J. E Zaruce hi. 1993. Catalogue of the Flow- i eru. Monogr. Syst. and e of . Missouri Er Gard. ‹ B B., R. job (& J. Sweere. 1994. rbcL sequences support in exclusion p Retzia, Desfontainia, and s fox from the Gentianales. Pl. Syst. Evol. 190: 2 123 0. Chase, M. W., D. E. a. P. S. Soltis, P. J. Rudall, M. К. jet W. H. p 5. Sullivan, J. е h, М. Molvray, P. J. Kores, T. J. Givnish, K. J. S a & J. C. Pires. 2000. Higher- but са s of he monocotyledons: An assessment of current cigar and a new clas- sification. Pp. 3-16 in К. Wilson & P. Weston (editors), Systematics and Biology of the Monocots. Royal Botanic Gardens, Sydney. Coetzee, J. A. 1993. African flora since the terminal Ju- rassic. Pp. 37-61 in P. Goldblatt (editor), Biological Re- ae Between’ Africa and South America. Yale Univ =ч ress, New Hav игы — 7 . Rogers. 1982. Palynologic val and — Mu nce for the Miocene — nvironment in the danha — — Africa). Palaeogeogr. — Palaeoecol. 71-85. & — 1985. The phytogeographic signifi- cance of some extinct Gondwana pollen types from the Tertiary of the و‎ — (South Afric а). Ann. Missouri Bot. Gard. 71: 1088-10 — —— & lL. Pra (ues 1984. Pus evidence for the occurrence of Casuarinaceae and Myrica in the Tertiary of South Afric "ul. 23: 23— Conran, Chase & P. I Rudall. 1997. Two new monoc йу айо families: Anemarrhenaceae and ie — ‘eae (Lilianae: Asparagales). Kew Bull. 52: 995— Grana ~ Cooling, R. M. 1987. speciation in Gondwanan shrublands. S. African J. Sci. 83: 106-112. 1990. Diversity compone nts in a spec ies-rich area of the Cape Floristic Re холоп. J. Veg. Sci. 1: 699- Fire and its role in coexistence and 7 1992. Press, с Tow & ( Heli ijnis. 2001. The identification of broad habitat units as biodiversity entities for systematic con- se — — in the Cape floristic region. S. Afri- : 15 сап E bei Ecology of Fynbos. Oxford Univ. — 1992a. Flora and vegetation. Pp. 23-01 2 M. Cowling (editor), The Ecology of Fyn- bos. Oxford Univ. - 'ess, Cape Town & )2b. Endemism айд speciation in a — lowland flora s hei Cape Floristic Region. Biol. Linn. Soc. 47: 307— & D. J. Me Donald, plant dede i 17 (е Ae 1999. Local endemism and ^W Ru ndel, G. Montenegro & F. J bandi Degradation in Mediterranean-Cli- — T. E Witkowski. 1994. Convergence and non-convergence of plant traits in climatically and — ‘ally matched sites in saree an Australia and S ree Africa. Austral. J. Ecol. 19: 220-232. .H olmes & A. С. Ки 1992. Plant di- Volume 89, Number 2 2002 Goldblatt & Manning Plant Diversity versity and endemism. Pp. 62-112 in R. M. Cowling — — Ecology of Fynbos. Oxford Univ. Press, Cape p / ——., P. ndell, B. B. Lamont, M. Kalin Arroyo & M. St. 1996. Plar climate — s. p — Evol. 362— ————, P. W. Rundell, P. C. Desmet ^ ^ Tea 1998, шоо d T -scale plant diversity in southern African aid lands: Subcontinental and global Diversity & Distributions 4: си. —— ‚ Cartwright, J. E. Parki Allsop. 1999. Fossil wood charcoal assemblages ов. ed ie Cave, South. Africa: Implications for Late Quater- ary vegetation and climates i in the winter-rainfall fyn- x Biogeogr. 26: 307—378 = ssey, A. T. Lombard, C. E. Heijnis, D. Mie еа & N. Cole. 2001. Framework for a Con- servation Plan for the Cape Floristic Region. Institute : Plant Conservation Report 9902 submitted to WWF: it r in Сва c отрап isons. yo | hos Бае: p nm W. D. 1987. Flora of Panama. Checklist and In- dex. Missouri Ке 19 ee ology. Pp. 58- 66 ‚ Day, . Sici. À N. Louw & M. L. Jarman (editors), A Preliminary — S. Africa Sci. foar kep 40. CSIR, Pretor P. G. & R. M. Cowling. 1999. Biodiversity. hab- itat and range-size — of a flora from a winter-rain- fall desert in north-western Namaqualand, South Africa. 23-33. D Garden, St. Louis at. Desmet, Esler, K. J. 1998. Unusual geophytes of the succulent ka- roo. з & Flora 84: 6-9 › W. Rundell & P. ete. 1998. Bioge ography of тк leaved geophytes in semi- i South Africa: Hypotheses | on functionality. Pl. Ecol. : 105-120. Fenner. M., W. G. Lee & J. Barstow ane 1997, comparative slt idy of the alistribulion of genus size in twenty angiosperm floras. Biol. J. Linn. Soc. 62: 225- 287. — Gentry H. 1988a. Tree species ric — of uppe r — azonian species. Proc. Natl. Acad. Sci. U.S.A. ¢ 56- 159 1988b. Changes plant community diversity and floristic composition on environmental = geo- graphic val gradients. Ann. Missouri Bot. Gard. 75: 1-34. — & C. Dodson. 1987. Diversity and ее — у ШП у: vascular epiphytes. Ann. Missouri Bot. ard. 233. быды, А oe lts characteristics, ir a and origins. Au Bot. Gard. 65: 369— 1979. Miscellaneous кыны! counts in an- An analysis of the flora of southern P ncluding new family and generic. re- Gard. 66: 865-86 giosperms, Missouri Bot. Floristic — sity Conservation 6: cords. Ann. 1997. South — a. 359-311. & J.C. ›. Phylogeny and speciation in Layers — Laer — eae subfam- ily BE deae). Pu Bot. ( Gladiolus in eben Fe e Press, Cape Тип п. & ————. 1999. The long-proboscid fly pollina- Ga ud. € 516—301. Africa. : ion — m in Gladiolus | (Iridaceae). Ann. Missouri Bot ard. Es — 2000. Cape Plants. National Botanical Institute, Cape Town. — ———, К. Dahlgren, B. J. Nielsen & J. P. Rourke. 1979. in the Cane Flora of Further notes of Retziaceae: Chemical contents and af- finities. Ann. Missouri Bot. Gard. 66: 545-550. ‚ P. Bernhardt & J. C. Manning. 1998. Pollination of ita aa geophytes by monkey beetles (Scarabaeidae: Rutelinae: Hopliini) in southern. Africa. Ann. Missouri Bot. Gard. 85: 215-230. . €. Manning & P. Bernhardt. 2001. d of pollination systems in Gladiolus (Iridace: de — in southern Africa. Ann. Missouri Bab 34. : Crocoi- Gard. 88: б R 1974. The Geography E the Flowering Plants, ed. 4. Longmans Green, — or Groves, R. H., J. S. Bea H. E Deac on, J. J. Lam- N. Specht & W. D. origins and character- A Rabinovite de Vin, Introduction: e brechts, A. Stock. 1983. istics of mediterranean ecosystems. Pp. Day (editor), Mine ral Nutrients in Mediterranean Eoo- can Nat. Sci. Progr. Rep. 71. CSIR - - 1 in J. syste ms. S. Afr Ha & H. A. Veldhuis. 1985. South Mrican Pe Data As — a karoo biomes ep. SIR, Pretoria. m >: Au te m s A Record of Past Life. South hacen Museum, Cape Town. Hopper, S. D. 1992. Patterns of plant diversity at the pop- ulation and species levels in south- west Part E báo subdivision. Vegetatio 97: 21— 38. ————,‚ 1997. arid regions. Floristic biodiversity and history of African Biodiversity & Conse — 6: 195-514. The mediterra- What do we Notic. Biol. Kalin Arroyo, M. T. & L. Cavieres nean type-climate flora of — nc know P how e 48-5 an we assure ils protection, . С. Marticorena € M. Munoz-Schick. 1994. с onvergence in the Mediterranean floras in Cen- tral Chile end California: Insights from comparative bi- ology. Pp. 43-88 in M. T. Kalin Arroyo, P. H. Zedler & M. D. Fox (editors), Ecology and Biogeography of Med- California Austra- lia. w Stud. 108. Springer-Verlag, New Yo 02. The Geology of the Earth. Oliver & Boyd. ‚ 1977. The ecology of early man in southern Africa. E тепсе 197: 115-126. Kruger, F. J. 1977.1 Toward a strategy for conservation. S. African J. Sci. 73: 81-85. & H. C. Taylor. 1979. Plant species diversity in Le ologic al reserves in the C ape fynbos: TO В fynbos: Gamma and delta diversity. Vegetatio 41: 8593. Lebrun, J.-P. & A. L. Stork. 1997. tes à Fleurs d? Nie Trogi ale. 5: Conse din Botanique de Levin, D. A. 1993. Loc e spec — in — The rule d O Syst. Bot. 18: 197 85. Gene flow, spec ialion pem species di- Enumeration des Plan- ervatoire el Jar- a Ville de Genéve, Geneva. L inder, ve rsity pev in a species-rich area: The Cape flora. Pp. 53-57 in . Vrba on Species and Specia- tion. Transvaal A Monogr. Transvaal Museum, Pretoria. & J. H. J. Vlok. 1991. The morphology. taxonomy 302 Annals of the Missouri Botanical Garden „Р — and — of Rhodocoma (Restionaceae Syst. Evol. 175: 139-1 — Louw, ^ B. & A. G. Rebelo. 1996. Vegetation of South Africa, Lesotho pe Swaziland. De ‘partment of Environ- mental Affairs oa Tourism, Pretoria Maggs, G.L., P. ¢ n & Н. Kolberg. 1998. Plant species richness, uU M: and genetic resources in Namibia. Biodiversity a Conservation 7: 435—440. Manning, J. €. & P. Goldblatt. 1996. The Prosoeca pr ingueyi (Dipte ra: Nemestrinidae) pollination syndrom in southern Africa: Long- — sd flies and their tu epa lowers. Ann. Missouri Bot. Gard. 83: 67-86 Meadows, M. E. & J. M Sue n. 1991. A vegetation his- € of и last 14 00 ) years on baie — south- Cape Province. 5. African J. Sci. 87: 3 Médail, E & P. Quézel. 1997, Hot ш analysis for cc servation of plant biodiversity in the Mediten rranean Ba- sin. Ann. y Bot. Gard. 84: 112-127. —— . P. Linder. 1998. The phylogenetic po- sition of Prionium (Juncaceae) within the order Juncales based on —— ‘al and rbcL sequence data. Syst. Bot. 23: 43-5 Oliver, E. G. н. 30 JO. Systematics of Ericeae (Ericaceae: Ericoideae), species with indehisc - and partially de- hiscent m Contr. Bolus Herb. 19: 1-483. Olmstead, R. G., C. . D. Wolfe, N. D. Nelson, W. J. Elisens & P. , „Жаа. 2001. Disinte- E e the Scrophulariaceae. Amer. J. Bot. 88: 348— 361. — T. D. & B. T. Styles. 1975. graph of y bs Es ‘eae. Blumea 22: Raven, P. H. . Axelrod. 1978. — relation- ships of de к flora. Univ. Califumia Publ. Bot. 12 W. TS generic mono- 540. 12. Rebelo, А. С. 1992, Preservation of biotic diversity. Pp. 308—344 in R. M. “айан, (editor), The Ecology of Fyn- A — Univ. Press, Cape Town. Rutherford, C. € R. H. Westfall. 1994. southern ve а: An objective categorization. Mem. Bot. Africa ӨЗ. V., M. К. Fay, D. €. Albach, A. Backlund, M. .. M. Cameron, S. j Johnson, M. D. Pintaud, М. Pow Sag Soltis, Р S s ж Wes LC e 4 neat rly a te E gene E., Biomes of ‚К. J. 2000. Phyla ny of the eu- familial analysis based on sequences. Kew ui 55: 257-309 Schatz, G. P. P. Lowry, M. Lescot, A. E. Wolf, S. An- driambololonera, V. Randrinasolo &J. jos ысый ona. 1996. Conspectus of the vascular plants of Mada- gascar: A taxonomic and | conservation electronic data base. Pp. 10-17 in L. +. van der Maesen et al. (ed- itors), The Biodiversity a African Plants. Proceedings of the XIV AETFAT Congress, Wageningen, The Neth- erlands. Kluwer Academic Publications, Dordrecht. | upper lacustrine — — Banke, Namaqualand. African pre : 1-10 Sine erie — B. R. “The — of the montane — enic ele in the South — invertebrate . Ann. Cape Prov. Mus. 2: 190-205. тшш. A. 1986. Floristic Re ol as of the World. Univ. — Press, Berkeley. [Translated by Y Crovello. | Taylor, 1‹ 171-229 in М. J. A. Wer “reer n Biogeography = Ко ology of Southern . 1. W. Junk, The Hag 18. Capensis. Pp. 7 - > — Afr Tilman, jo 1082. Resource Compe * dd Community eto Structure. Princeton Univ. Press, Princ 1983. Some thoughts on — "competition and — in plant communities. Pp. 322-336 in F. J. En uger, D. E Mitchell & U. M. J: editors), Vpe » Ecosystems, The Role c nf Nantes анан i Berlin. — ———, V. J. Bond, B. M. Campbell, F. J. Kruger, н. J Linde >, A. Scholz, H. C. Taylor & M. Witter, 1983. igin and maintenance of plant species diversity. Pp. 125-135 in J. A. Day — Mineral Nutrients in Mediterranean Ecosystems. S. African Nat. Sci. Progr. Rep. 71. CSIR, Pretoria Villagran, С. М. 1994. Quarternary history of the medi- lerranean vegetation of Chile. 3-20 in M. T. Kalin Arroyo, P. H po & M. D. ы Гете Ecology and Biogeography f Mediterranean Ecosystems in Chile, таат. fornia and ain s Ecol. Stud. 108. Springer- Ver- lag — Vork. W ishc u, I. C., . Rosenzweig, L. Olsvig-Whittaker & A. Se — 2000. What naka 's nutrient-poor mediter- ranean heathlands so rich in plant diversity. Evol. Ecol. Res, 2: 935-955. Vogel, S. 1954. Blüte — pen als Elemente der Sippengliede s Bot. Stud. -338. Wagner, W. L. 1991, E — еу son of the avai and M floras: A compari- Marquesan Ar 267-284 in E. C. Dudley (editor), ' lutionary Bue Vol. 1. Dioscorides Press, Portland, айы 1990. Manual of Hawaiʻi Press, Herbst € S. Н. Sohmer. the Flowering Plants of Hawaiʻi. Univ. Honolulu. Weimarck, Н. 1941. Phy roups, centers ss — rar the — flora. Аш ta Univ. Lund. ne үз 5): -M3 Xiang 0.) — Ой» D. R. Morgan & P. S. Soltis. 293. TS e N O e L. sensu lato om lr cata relatives inferred od — sequence ale . Missouri Bot. Gard. 80: 723—7 Volume 89, Number 2 Crosby 303 2002 Statistical Summary STATISTICAL SUMMARY OF SOME OF THE ACTIVITIES IN THE MISSOURI BOTANICAL GARDEN HERBARIUM, 2001 Vascular Bryophyte Total Acquisition of Specimens Staff Collections 31.270 1.494 38,704 Purchase 201.004 35.000 236,004 Exc ang 23.088 1.812 24.900 Gif 7,837 1.705 9.542 Total acquisitions 263.199 46.011 309.210 Mountings Newly mounted at MO 70,177 23.093 93.270 Specimens mounted when acquired 14.004. 0 14.004 Repairs Specimens repaired 13.482 n/a 13.482 Specimens stamped 1.566 n/a 1.566 Total repairs 15.048 0 15.048 Specimens sent On exchange 22.102 160 22.262 As gifts 19,242 028 19.870 Total 11.344 788 12.132 Loans sent Total transactions 359 38 397 Total specimens 22,749 2,519 25.020 To U.S. institutions ‘Transactions 192 23 215 Specimens 13.028 1.948 14.976 To foreign institutions "Transactions 167 I5 182 Specimens 9,721 931 10,652 To student investigators Transactions 50 13 03 Specimens 6.772 1.207 7.979 To professional investigators Transactions 309 25 334 Specimens 16.161 1.672 17.833 Loans Received Transactions 233 28 261 Specimens 23.980 2.882 20.862 From U.S.A. From abroad Total Visitors 234 92 326 During 2001, 108,840 — were accessioned into the herbarium: 93,270 mounted at MO, 14. 004 mounted when acquired, and 1566 old MO specimens in the herbarium on 1 January 20( EE 5,219,2 Purchases in 2001 included а Missouri Bot. Сага. repaired as needed, stamped, and accessioned in the near future. bryophytes packeted, and all accessioned as re Herbarium (see specimens — — Bog red). Bull. 89(6): hs 2001). sources allow The 4,832,175 vascular . In addition, the bryophyte collection from LA total number of mou plants and 387,041 bryophytes). 100 vase plants and 35,000 stil in the Clyde E Reed The estimated 162,000 mounted vascular plants should be The unmounted vascular plants will be mounted. the inte xd. accessloner К, approx- imately 30.000 specimens, accumulated and curated by William Dean Reese (1924-2002). was acquired on permanent loan in late 2001. 304 Annals of the Missouri Botanical Garden 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 20( and as year-end totals. Note that the specimen records in TROPICOS are primarily based on MO specimens, meaning that about thirty-seven percent of the bryophytes (an increase of about six percent over 2000) and thirty-two percent of the vascular plants (an increase of about one percent) in the herbarium are now computerized, with an overall total of about thirty-two percent (an increase of about one percent). TROPICOS records—Calendar Year 2001 Additions Bryophytes Vascular Plants Total Specimens 30,709 94,277 124,986 Names 931 25,453 26,384 Synonyms 872 12,097 12,969 Distributions 369 31,376 31,745 Types 152 19,040 19,192 Bibliography 1,170 3,605 4,775 TROPICOS records—Year-End 2001 Totals Bryophytes Vascular Plants Total Specimens 143,298 1,548,109 1,691,407 Names 100,431 795,401 895,832 Synonyms 63,415 392,242 455,057 Distributions 39,028 798,705 837,793 Types 7,983 278,617 286,200 Bibliography 23,274 04,654 87,928 Specimens in herbarium 387,041 4.832.175 5.219.216 Percent of specimens computerized 27 22 32 In TROPICOS, literature-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 synonymy may be derived from these direct links, e.g., all other combinations of a basionym treated as a synonym of a given name are also synonyms of it. —Marshall R. Crosby Volume 89, Number 2, pp. 125-304 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on July 10, 2002. ANNALS OF THE MISSOURI BOTANICAL: GARDEN ATUDMISSOCUREP BOTANICAE GARDEN ANNUAL REPORT ARE NOW AVALTLABLE UN: J STOR! 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 — Review of Ecology and Systematics Conservation Biology Diversity and Distributions Ecological Applications Ecological Monographs Ecology Evolution Functional Ecology Global Ecology and Biogeography Letters Information regarding JSTOR is available at http://www.jstor.org 120 Fifth Avenue, New York, NY 10011 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 Botanical Society Limnology and Oceanography Missouri Botanical Garden Annual Report Quarterly Review of Biology Systematic Biology Systematic Botany viiam — | | | I | | i | | | || | | CONTENTS Conservation, the 47th Annual Systematics Symposium of the Missouri Botanical Garden Introduction: Conservation and the Future of Life George E. Schatz & P. Mick Richardson The Seven Great Questions of Systematic Biology: An Essential Foundation for Conservation and the Sustainable Use of Biodiversity 2 Joel Cracraft Taxonomy and Herbaria in Service of Plant Conservation: Lessons from Madagascar's Endemic Families _ $ George E. Schatz Chicago Wilderness: A New Force in Urban Conservation Debra K. Moskovits, Carol J. Fialkowski, Gregory M. Mueller & Timothy A. Sullivan Safeguarding Species, Languages, and Cultures in the Time of Diversity Loss: From the Colorado Plateau to Global Hotspots - Gary Paul Nabhan, Patrick Pynes & Tony Joe The United States Naturalized F he — se He of Deliberate Introductions hard N. Mack & Marianne Erneberg The Dodo Went Extinct (and Other Ecological ph Stuart L. Pimm The Global 200: Priority Ecoregions for Global Conservation avid M. Olson & Eric Dinerstein El Género bias (Acanthaceae) en Sudamérica Austral — _ Cecilia Ezcurra Plant Diversity of the Cape Region of Southern Africa — Peter Goldblatt & John C. Manning Statistical Summary of Some of the Activities in the Missouri Botanical Garden Herbarium, 2001 . Marshall R. Crosby — Cover illustration, Breonia richardsonii Razafim., drawn by Barbara Alongi. Annals of the Missouri Dotanical Garden 2002 Y Volume 89, Number 3 Summer 2002 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 Committee 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, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2002 is $140 per volume U.S., $150 Canada & Mexico, $175 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 O Missouri Botanical Garden 2002 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. PosrMASTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. 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. € AKICI/ATIC[(. 740 40 IRAN РТУ. : ` © Thi: ең at Dríinmor) Ls / olume 89 Annals umber 3 of the 002 Missouri Botanical Garden PHYLOGENETIC Ana María Cialdella? and RELATIONSHIPS OF THE Liliana Mónica Giussani? GENUS PIPTOCHAETIUM (POACEAE, POOIDEAE, STIPEAE): EVIDENCE FROM MORPHOLOGICAL DATA' ABSTRACT A systematic treatment of the entire genus Piptoc — — h complements ce of Cialdella and Arriaga for the South American spec les, 18 presente d * теш. A synopsis of : spec ies and 2 varieties, a key to ident tify these 5| ECTES, as well as synonymy and geographic ‘al distribution, are given. Hlustrations and E тсн of di gnostic characters are presented for the North American species = the first time. We propose a phylogenetic hypothesis for the genus based on morphological and anatomical characters. Monophyly " Piptochaetium was supported by the combination of three synapomorphies: a grooved palea. aaa lemma margins that fit into a ол Шла palea groove, and the palea protruding from the lemma. Two sections were previously recognized in Piptochaetium: Podopogon and Piptoc haetium, The cladistic analysis showed that only South (American species of Pipt vi opc sect. Piptochaetium, together with / brevicalyx, form a monophyle lic group (Obovoid а while Piptochaetium sect. Podopogon is polyphyletic. Therefore, the taxonomic recognition of sectional taxa cannot be supported. Phylogenetic relationships of two varieties of Pipto- chaetium stipoides (var. stipoides and var. echinulatum) were nol Kn resolved. Furthermore, based on cladistic analyses, taxonomic observations, and the study of type material. Piptoc haetium tovarii subsp. pilosa is raised to specific rank (= P. pilosum). Key words: phylogenetic relationships, Piptochaetium, Poaceae, Stipeae. Rest MEN * presenta un tratamiento sistemático de todo el género Piptochaetium, que e omple menta el propuesto por Cialdella dia para las especies sudamericanas. En este — se presenta una sinopsis de las 36 e * ‘cies y 2 variedades, una с have para identificar las especies, su sinonimia. distribución geográfica y observaciones. Se incluy or primera vez ilustraciones у fotograffas de los caracteres di agnóstic os para las « "spec les norteamericanas. Se propone una hipóte SIS ' This research was supported by CONICET, PEL: 0000/97. We thank O. Morrone, F. O. Zuloaga, M. Barkworth, J. Maze, and two — reviewers who kindly reviewed versions of the manuscript and guided us with hair construc- tive comments, and G. Davidse and R. Soreng who helped us to solve nomenclatural queries. We are also grateful to the curators and staff di the herbaria mentione di in the text, who allowed examination of specimens in their care, whether through loans or during personal visits. Thanks to Daniel Rodriguez for assistance in the use of the scanning microscope, and Vladimir ro Dudás for preparing the line drawings and the composition of the photographic — F Sem we thank the staff » the Darwinion, who helped us — *— ? Instituto de Botánica calera m Labardén 200 CC 22 (B1642HYD) San Isidro. Buenos Aires. Argentina. acialdella@darwin.educar ANN. MISSOURI Bor. GARD. 89: 305-330. 2002. 306 Annals of the Missouri Botanical Garden filogenética basada en nofilia de Piptochaetium sobre la base de tres sinapomorfías: una pea surcada entre los lemma involutos insertos en el surco longitudinal de la pálea y la pálea caracteres morfológicos y anatómicos. Mediante el análisis filogenético, fue confirmada la mo- nervios, los márgenes de la asomando por entre los márgenes de la lemma. Dos secciones fueron previamente reconocidas para Piptoc haetium: Кол. у — haetium. Como resultado del análisis cladfstico, * miento de categorfas formales infragené ricas. En e sti ae y var. ec Шш im) у la posición а де aw variedades no fue clara e (= P. pile p tovarii subsp. pilosa fue elevada al rango de espec vaciones taxonómicas y el estudio del — * ólo las especies sudamericanas de la sección un grupo — © (grupo Obovoide). Consecuentemente, en este tratamiento, no se quani ni vá 1álisis, se incluyeron 2 variedades de Piptochaetium stipoides (va ‘on P brevicalyx, conformaron álido el reconoci- Piptochaetium, junto a. Piptoc hasten жит), sobre la base del estudio filogenético, las obser- mente resuelta Piptochaetium J. Presl is an American genus dis- tributed mainly in temperate areas of both hemi- spheres and scarcely represented in tropical re- gions of the Andes (Parodi, 1944; Barkworth, 1986; Sánchez Vega, 1991; Cialdella & Arriaga, 1998). The genus belongs to the tribe Stipeae and established by Presl in 1830, based on Piptochae- tium setifolium J. Presl |= P. panicoides (Lam.) E. Jesv.|. It includes perennial, caespitose plants, generally less than 1.5 m . with loosely or densely flowered panicles of one-flowered spikelets: membranous glumes, frequently longer than the flo- ret, enclose the anthecium. Three characters are useful for recognizing the genus: the grooved palea, involute lemma margins that fit into a longitudinal palea groove, and the palea protruding from the lemma. The South American species of Piptochaetium have been studied by Parodi (1944), who treated 21 species in 2 sections: Piptochaetium sect. Po- dopogon (Raf.) Parodi and Piptochaetium sect. Pip- tochaetium (= Piptochaetium sect. Eupiptochaetium Parodi). Sanchez Vega (1991) treated the 6 species for Peru, and grouped them into the sections pro- posed by Parodi. Cialdella and Arriaga (1998) pub- lished a more recent treatment of the South Amer- ican species, in which they recognized 27 species, 2 subspecies, and 2 varieties for both sections, con- sidering leaf anatomy and micromorphological characters of the lemma in their analysis. T North American taxa were studied by Valencia and Costas (1968) with emphasis on eytological char- A compari- acters, and they recognized 7 species. son of these treatments, including the infrageneric is shown in Table 1. In a revision of Barkworth (1986) and classification, the North American. Stipeae, North without grouping them into sections. Mesoamericas =] treated 13 species ir The aim of this work is to complete the morpho- logical account of the genus by considering the spe- cies not included in the previous paper by Cialdella and Arriaga (1998). opsis of 36 species and 2 varieties, a key to identify This study comprises a syn- them, and their associated synonymy and geograph- ical distribution. A phylogenetic analysis of Pipto- chaetium was performed to explore the relationship among species, using cladistic methodologies with the morphological data. MATERIALS AND METHODS MORPHOLOGICAL STUDY This study was based on the analysis of almost 1000 herbarium specimens, including type material for most of the analyzed taxa (Appendixes 1, 2). The specimens studied were from BA, BAA, BAB, BAL, C, CAS, CEN, CORD, CPUN, CTES, G, ICN, K, LE, LP, LPB, MO, MEXU, MVM. MVFA, NY, P, PR, SGO, SI, SRFA, UNR, US, USM, UTC and XAL (Holmgren et al., 1990). Micromorphological structures of the lemma were analyzed using a scanning microscope Zeiss DSM 940 A, located in the — de Botánica Darwinion and as summarized in Cialdella * Arriaga (1998: 108). Epidermal ndis (ma- crohairs, prickles, hooks, and papillae) follow the classification proposed by Ellis (1979). CLADISTIC ANALYSIS Twenty-nine exomorphological and epidermal characters were used in the cladistic analysis (Ap- pendix 3); all multistate characters were treated as unordered. Only two autapomorphies were included in the analyses: characters 20 and 27 (Cialdella & Arriaga, 1998: figs. 3D and 9F, respectively). Char- acters were directly observed from herbarium ma- terial totaling nearly 1000 specimens. The data ma- trix is presented in Table 2. Two different computational strategies, using min- imum parsimony, were used to analyze phylogenetic relationships among species of Piptochaetium. In both cases, polymorphic characters were included Table 2). One strategy assigned all characters equal 3 weight. This analysis was performed using NONA version 1.6 (Goloboff, 1993a). In the second strategy, a concavity function of the homoplasy (K = 3) was Volume 89, Number 3 Cialdella & Giussani 2002 Phylogenetic Relationships of Piptochaetium [| Figure 1. SEM photos showing epidermal charac ter rs of the lemma of Piptochaetium.—A,. Lemma without any epider- mal — s of P. brevicalyx (Rzedowski 35863 ‚ Macrohairs of P. fimbriatum (Davidse et al. 29775). —C. Prickles and hooks of P. avenaceum (Hart ey 1654). —D. Pric kle s, hooks, and macrohairs of P angustifolium (Palmer 720). y used to apply weight to all characters, a method for nus from its allies (Spegazzini, as Oryzopsis Michx. estimating implied weight (Goloboff, 1993b). This sensu Speg.. 1901: Parodi, 1944: Barkworth, 1990: analysis required Pee-Wee version 2.8 (Goloboff. Peñailillo, 1996; Cialdella & Arriaga. 1998). Other 1997). In both analyses, settings amb- (clades re- characters were also used to uniquely identify Pip- solved only if they have unambiguous support). and tochaetium, such as a palea longer than the lemma. poly = (polvtomies allowed) were used to resolve protruding from the lemma apex, and the lemma phylogenies. The order of taxa was randomized. cre- margins involute and fitted into the palea. groove. ating a Wagner tree and submitting it to branch- The 37 entities recognized within the whole genus swapping by means of tree-bisection reconnection were considered as terminal taxa. (TBR) using the sswap* command. To evaluate the relative support of clades, branch support. with in- OUTGROUPS struction "bsupport" (bs) of NONA (Bremer. 1988. 1994), was calculated. Vassella E. Desv. and Hesperostipa (Elias) Bark- worth were selected as outgroups, based on their — similarities in morphological and anatomical char- acters to Piptochaetium (Thomasson, 1978b: Bark- Since Desvaux (1853), Piptochaetium has been worth, 1990) as well as previous phylogenetic re- recognized by the presence of a grooved palea, and lationships proposed by Barkworth (1990). this character was diagnostic to distinguish the ge- A detailed history of Nassella can be found Annals of the 308 Missouri Botanical Garden 11 unmqní y wnyognl q wnyognl 4 чо av miri d итп d итү d sn AN n9 wnyouquy y UNIDUQUAS y 11 штр]үпәпә Y ungpjmono y 1n uy SUdISAYDI q SUdISAYDI Y SUIIAYDI J тү WNUDIADYING Y WNUDIADYING + WNUDILDYANG q aN x4qD21424Q. 41 XA]DINIALG f xona qd yu) әѕиәјодир 4 asuajozun + ISUIJOTUD J aN umpofusnduo y dnoa:) pro^oq() WALL V HOOLdld I ALEAV HOO. dld WALLAV HOOLdld I OLET V HOOTdld 9A “OW “ney SUAISALVA d SUIISALTA d SUIISIL1A d aN ‘Nd мә]ә$ ү ag mndojspavs J nnZojsp3ps y INIASDIDS 4 1Q ug чү wnupiyyooidns q unupj224dn4. y unupnp[224dna J 1 N 1ә#ина q 19 alsniod q ənsnpd y ү asuapnjsodpu y asuapjsodpu J ISUIDISOÍDU j 1Q ag чү штрәш + штрәш y 10102 4 = 3g 53 09 чү unmpu 4 штпри q unmpu 4 шптри ¢ үчү е а 3 Шет А PY qq sa AN по) uunjpaquiaf 4 әд 19U01819YIDOS J 121018124102 ү] 1әио]кәуттә 41 1 uy wnsnfuod q wnsnfuor x] umsnfuoo y IV 2D4240D2 J 2D421QD2 d 2012192 4 ү vNULIASAYIDAQ 4 umuladsíyopaq Y штица f 11 Чочу Dd орол M озн О sq ѕәр1отәриәпр d sapio1oDuaap d P sapioi0DuodpD d Sn N 1unaopuaan qq UNIIDUIAD + unaonuoap yJ aN umnofisnzup q 2 ungofiusnzup y 19 unuidp q und. q пол ons[sudsqoq NODOdOGOd NODOdO(GOd NODOdOGOd NOODOdOGOd uonnquisip sisA[eue orouozo| sud (8661) (1661) (8961) (rr61) !poreq jeorydeis0a+) əy} woj} e3eLLIV Y еәрүегә sdnois [ешлоуи PBA zaqoueq 581507) DY EIOUS[EA nsuas suog 'pajuasaud OS[B SI 591 ээ әш jo uornqugjstrp orgde1302^) ‘SISA[BUE 294330] &ud 34] шол 51891 Чим uosueduio» ut umapyooidid ишцим SUOTJBOYISSP]) опәпәЗр1уи] “B[PNZIUDA “OA S81 Pau] “Sn ‘Aenanı 1] "un 1d “Əd ‘Aene qf “ed :OJIXIJA “OW `РТРШӘ1РП С) пс) чорепоу EX 1214шо[07) *07) U) чэ: 11219 ug :BIAIOH "og teurnuaz1y AY :0] puodsaunioo ѕиоцет, A2Jqq y т 1L Volume 89, Number 3 Cialdella & Giussani 3 2002 Phylogenetic Relationships of Piptochaetium Continued. Table 1. a 3 | Barkworth (1990). She expanded Nassella, includ- = ec C Р - ing a total of 79 species, the majority of which were = ESS = б previously considered in Stipa L. The cireumscrip- E — — p 5 ye 5 " 2 tion proposed by Barkworth (1990) was later ac- = Sx gua pS 4 со CP а iu cepted by Peñailillo (1996) and Torres (1997). In gost ee به‎ E SHP ې‎ ы this sense, Nassella includes species with obovoid | oia or terete florets, a short and blunt to long and sharp 3 : callus, strongly overlapping lemma margins, a lem- , A = + ma apex fused in a crown, and a highly reduced, = E — nerveless and glabrous palea. All species consid- be 3 y е ie ered in Nassella have a characteristic lemma epi- as t Eg 5 © 2 2 » dermal pattern: short fundamental cells, with irreg- 57 Ет 8 = £ Š i "E ularly sinuate sidewalls and abundant short cells. = E * = E E B. Ё & & E E Hesperostipa was one of the nine sections of Stipa ш р E Bas * bie = A = -— formally proposed by Elias (1942). Thomasson hdc co M M, (19784) noticed a very different epidermal pattern within this section in comparison with any other E Stipa: long fundamental cells with strongly sinuous 3 Fe E sidewalls, and absence of short cells. After studying = Е $ E North American Stipeae, Barkworth (1993) raised 2. = T2 а Stipa sect. Hesperostipa to the generic. rank, in T 2a H8 . E Su 4 which she included four species: Hesperostipa co- 28 232 i в Е ti р : mata (Trin. & Rupr.) Barkworth. H. curtiseta Bc E à چ‎ E E E 3 A (Hitehe.) Barkworth, H. neomexicana (Thurber) = ¿72 S x ст а § Barkworth, and H. spartea (Trin.) Barkworth. Hes- E паа. E oO А АРЕ ll a; perostipa 1s endemic to North America and includes plants with narrowly terete florets, a sharp callus, and a persistent, twice-geniculate awn. The lemma m is indurate, with flat margins, slightly or not over- 3 S lapping: the palea is subequal to the lemma, prow & z lipped and 2-nerved (Barkworth, 1993). p z Š i z е KEY TO PIPTOCHAETIUM, NASSELLA, AND HESPEROSTIPA 5 = х © = = Jl Palea longitudinally grooved, 2-nerved, longer B = = a x than the lemma, and SEU from the lemma x x \ \ apex; lemma margins involute, fitting into the palea groove Piptochaetium 1'. Palea flat, 2-nerved or nerveless, shorter than or equal to the lemma, usually not protruding beyond the lemma apex; lemma margins fl: 2 Lemma margins strongly overlapping; pa- n lea nerveless, less than one-third the E length of the lemma lassella em а 2'. Lemma margins not or slightly sonia»: ge E E ping: palea. 2-nerved, always as long as E в) = the lemma, apex prow tipped or © * “pine A Hesperostipa Е: xX 5 E RESULTS 3 E Seven equally parsimonious trees were obtained — = = when the characters were weighted equally. These z= - 8 : Ы were 58 steps long. with a С — y Index (CI) = 2 = E 5 = i ёв : < 0.55 and Retention Index (RI) = = 0.87. When the | چ‎ s 3 ES ГРЕЕ So analysis was performed without autapomorphies F & ÉSE = E X E 3 E (characters 20 and 27), the CI was 0.53. When the eee A ECA x characters were weighted, five equally parsimonious 310 Annals of the Missouri Botanical Garden Table 2 Data matrix used in the cladistic analyses for Piptochaetium. Numbers in the first row represent the characters described in the text, together with their codification, Numbers within brackets represent a polymorphism for the particular character of the species involved. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Nassella 0 0 2 ооо [OI] [01] [01] [O I] [01] [023] [01] [01] Hesperostipa 0 0 0 1] 1 2 l 0 1 0 [0 1] 0 [0 1] Piptochaetium alpinum l d 1 1 l 1 l 0 1 0 1 0 0 l P. angolense 1 1 1 1 1 1 0 | 1 1 0 0 2 1 0 P. angustifolium Il dd d 1 1 1 0 0 ] 0 0 0 0 1 P. avenaceum |I 1 | 1 .) !] 1 0 1 0 JO 1] 0 0 0 0 P. avenacioides | 1 d 1 1 !] l 0 1 0 0 0 0 0 P. bicolor 1 1 | 1 | ! 1 0 1 0 1 0 1 0 0 P. brachyspermum Г l l 1 l 1 1 0 1 0 l 0 l 0 0 P. brevicalyx Г 1 1111 0 l 0 1 ) 1 2 0 0 P. burkartianum I 1 1111 0 l 1 l [01] 0 2 0 0 P. cabrera 1| 1 l 1 | 1 l 0 l 0 0 0 0 0 P. calvescens |I] 1 1 J ! 0 l 0 l 0 0 2 0 0 P. confusum 1 1 | 1 1 ! 1 0 1 0 1 0 0 0 0 P. cucullatum l 1 |1 1 |1 1 0 l 0 l 0 0 2 0 0 P. featherstonei l 1 | 1 1 ! 1 0 1 0 0 0 0 0 1 P. fimbriatum | 1 | 1 |1 1 0 l ] 0 1 0 0 0 1 P. hackelii |o 1 | 1 |1 1 l 0 1 0 l 0 0 0 0 P. hirtum 1 1 d 1 l | 0 І 0 | 0 0 2 0 0 P. indutum 1 1 l l 1 1 l 0 1 0 0 0 0 0 1 P. jubatum 1 d 1 og) l 1 0 | 1 | JO 1] 0 2 0 0 P. lasianthum l l 1 1 1 | 0 | l l l 2 0 1 P. lejopodum 1 1 | |1 1 1 0 l 0 | 0 0 2 0 0 P. medium | 1 | l 1 1 l 0 1 0 l l 0 0 P. montevidense | Ж 1 ]) ].» I 0 1 0 l [01] 0 3 1 0 P. napostaense | 1 |1 1 11 | 0 l 0 0 0 0 0 P. palustre | 1 1 1I 1 !]| | 0 1 0 1 0 0 0 0 P. panicoides | dd 1 |) l | 0 1 0 l 0 0 3 1 0 P. pilosum | i 1 1 1 d 0 | l l 0 0 2 0 1 P. pringlei 1 d gd | l | l 0 l 0 l 0 0 0 1 P. ruprechtianum 1 dd 1 |1 l | l 0 l 0 l 0 l 0 0 P. sagasteguii 1 l | 1 1l I l 0 1 0 1 0 0 0 1 P. seleri l l l gd l | 0 1 l 0 ] 0 0 0 1 P. setosum 1 1 1 | 1 | 0 1 1 1 0 2 0 0 P. stipoides var. echinulatum | | 1 1 | | 0 l l l [01] 0 2 0 0 P. stipoides var. stipoides Lor 1 |1 1 !] 0 l l l [01] 0 2 0 0 P. tovarii Lot d | 1 l 0 1 0 1 0 2 0 0 P. uruguense I t 1 | l l 0 1 l l [01] 0 2 0 0 P. virescens l i 1 1 1] 1 0 l 0 0 0 1 trees were obtained. These trees were also 58 steps their fit was 190.7, and the Although the trees obtained with the two strat- long, rescaled fit was 0.68. egies differed in some respects, both yielded the same strict consensus tree as well as the unambig- uous Only volved P. calvescens, P. cucullatum, P. jubatum, P. synapomorphies that support clades (Fig. 3). the relationships within one clade, which in- lejopodum, P. stipoides var. stipoides, and P. stipoides var. echinulatum, differed between the strategies. Relationships among these species are not resolved on the consensus tree (Fig. 3). Figure 4 shows one of the seven hypotheses found with NONA. This tree is the most similar to one of the five trees found with Pee-Wee. which, They differ in the position of P. alpinum in Pee-Wee's results, is sister to the clade that includes P. pringlei to P. palustre; and they also differ in the order of P. uruguense and P. angolense, which are inverted in Pee-Wee analysis. Cladistic analysis supports Piptochaetium as a monophyletic genus [Bremer support (bs) = 3]. It is clearly defined by three synapomorphies (Fig. 3): lemma margins involute (1), a boat-shaped palea (2), 2A, B). and a palea longer than the lemma (3; Fig. Volume 89, Number 3 2002 Cialdella & Giussani 311 Phylogenetic Relationships of Piptochaetium Table 2. | Extended. l6 17 18 19 20 2] 22 5:23. 24 25 20 27 28 29 Vassella [0 1] [0 1] [0 1] 0 O [01] [01] O [01] 0 [01] O [01] 0 Hesperostipa 0 | l 0 0 0 1 0 1 0 1 0 0 0 Piptochaetium alpinum 0 l | 0 0 | l 0 1 0 1 0 0 0 P. angolense 0 1 l 0 0 l 1 0 0 0 1 0 | 0 P. angustifolium 0 | | 0 | l 0 l 0 l 0 0 0 P. avenaceum 0 1 [O 1] [01] 0 | l 0 1 0 l 0 0 0 P. avenacioides 0 1 [0.1] [OT] 0 | l 0 1 0 l 0 0 0 ? bicolor 0 1 0 l 0 l l 0 l 0 1 () 0 0 P. brachyspermum 0 1 0 1 0 | l l О 0 1 0 0 0 P. brevicalyx 0 0 0 0 0 | | 0 0 0 0 0 | 0 P. burkartianum 0 0 0 0 0 | | 0 0 0 1 0 | 0 P. cabrerae 0 l | 0 0 | l | 0 0 1 0 0 0 P. calvescens [0 1] | | 0 0 | 0 | 0 l l 0 | | P. confusum 0 l 0 | 0 | | 0 1 0 | 0 0 0 P. cucullatum 0 1 0 l 0 | l | 0 0 1 1 | 0 P. featherstonei 0 1 | 0 0 | | 0 1 0 1 0 0 0 P. fimbriatum 0 0 0 0 0 | | 0 [01] O [OT] 0 | 0 P. hackelii 0 1 0 | 0 0 | ( 1 0 1 ) О 0 P. hirtum 0 0 0 0 0 | | 0 0 0 | 0 | 0 P. indutum 0 1 | 0 0 0 | 0 0 0 | 0 0 0 P. jubatum 0 l 0 | 0 | ОТ | 0 | 0 І | P. lasianthum 0 1 l 0 0 | 0 l 0 l 0 | 0 P. lejopodum 0 1 0 | 0 | 0 | 0 0 1 0 | 1 P. medium 0 1 0 | 0) | | 0 0 0 1 0 | 0 P. montevidense l 0 0 0 0 | | 0 0 l 1 0 | 0 P. napostaense 0 1 0 l 0 0 | 0 1 0 1 0 0 0 P. palustre ) 1 0 | 0 | | 0 l 0 l 0 0 0 P. panicoides [Ol] 0 0 0 0 | | 0 0 l 1 0 | 0 P. pilosum 0 l l 0 0 0 | 0 0 0 1 0 | 0 [ pri slei 0 0 0 0 0 l l 0 1 0 1 0 0 0 P. ruprechtianum 0 | 0 | 0 | | 0 l 0 l 0 0 0 P. sagasteguii 0 l l 0 0 0 | 0 1 0 1 0 0 0 ? seleri 0 1 1 0 0 | | 0 l 0 1 0 [0 1| () P. setosum 0 1 0 | 0 | | 0 0 0 1 ) l 0 P. stipoides var. echinulatum 0 | | 0 | l 0 І 0 l l 0 | | P — var. stipoides 0 [0 1] [O 1] [OI] 0 | [Ol] 1 [01] [01] [Ol] 0 | 101] Р. tova О 0 0 0 0 0 l ( 0 0 0 0 l 0 P. uruguense 0 l 1 0 0 l l 0 0 0 1 0 | 0 P. virescens 0 1 l 0 0 0 | 0 1 0 l 0 0 Q Hesperostipa is the most closely related group to Pip- tochaetium, as they share a 2-nerved palea (4) and long fundamental cells of the lemma epidermis (5). South American species of Piptochaetium sect. Piptochaetium, together with P. brevicalyx from Mex- 2), which is now recognized herein as the Obovoid ico, are gathered in a monophyletic clade (bs eroup (Fig. 3). It is supported by two synapomor- phies: a deciduous awn (10) and obovoid florets (13, Fig. 5H). Three species are closely related to the Obovoid clade: P. fimbriatum (Fig. 5G). P. seleri (Fig. 5F). and P. angustifolium (Fig. 5E). Piptochaetium — fimbriatum and P. seleri present a blunt or truncate o = 1) and, together with P angustifo- bs = 1). These three species are not included in the Obovoid clade. callus (8 lium, they all have a short callus (7: as they do not have a deciduous awn as well as a North can species; P fimbriatum and P. seleri were also well-defined obovoid floret. They are Ameri- found in Guatemala, with intermediate forms ob- served between the Obovoid clade and the rest of the species (Figs. 3 and 4: see discussion). Species of the Obovoid clade present the relation between length and width of the floret always less than 3.5 312 Annals of the Missouri Botanical Garden Figure (Rzedow = 35863). —B. P. fimbriatum (Davidse et al. 2 (Palmer 726 (28). However, this character is ambiguous for the clade that gathers the Obovoid group with P. seleri and P fimbriatum, and consequently, this synapo- morphy is not shown in Figure 3. An additional anal- ysis was performed to test the influence of the in- termediate species (P. fimbriatum, P. angustifolium, and P. seleri) on the results. Without these three spe- cies, the major clades and relationships among the species did not change. All species of the Obovoid clade, except P la- sianthum, have densely flowered inflorescences (11), although for some species of this clade this character is variable. These species also have a crown without macrohairs (24), which reverts in P jubatum and is stipoides. Species polymorphic P. stipoides var. from P. uruguense to P. stipoides var. echinulatum form a clade defined by a glabrous lemma (15, Fig. LA, 5H; bs = 1), and P. pilosum (= P. tovarii var. SEM photos showing distal portion of * — with different crowns in Piptoc — —A. 5). —C. P. avenaceum (Harvey 185 P. brevicalyx —p. P a pilosa Sánchez Vega) is consequently apart from its counterpart P. tovarú (= P. tovarú var. tovaril), sug- gesting this rank is not appropriate. Species with a few prickles on the lemma (18) are gathered in a clade that includes species from P. burkartianum to Р. stipoides var. echinulatum (bs = 1). This grouping is divided into two major clades. The first clade, Р burkartianum, P. montevidense, P. panicoides, P. hir- tum, P. brevicalyx, and P. tovarii (bs = acterized by a lemma without prickles (17, Fig. 1A), and, except for P. burkartianum, the other species also have a glabrous callus (9; bs = 1). Piptochae- ). is char- tium brevicalyx and P. tovarii are related by the ab- sence of prickles in the crown (26; bs = 1), while P. montevidense and P. panicoides are gathered in a clade (bs = 3) and share three synapomorphies: lens-shaped (13) and laterally compressed florets (14), and presence of papillae in the crown (25). Volume 89, Number 3 Cialdella & Giussani 313 2002 Phylogenetic Relationships of Piptochaetium Nassella Hesperostipa Piptochaetium sagasteguii A сл E 2 P. virescens * 7 1 P. alpinum 1 11 1 P. pringlei 0 hackelii N w = napostaense bicolor о = о ruprechtianum 15 19 brachyspermum medium confusum avenaceum avenacioides palustre featherstonei cabrerae indutum angustifolium fimbriatum seleri lasianthum \ pilosum uruguense angolense E OD DAD DOT DDO TD DOT DOV OD TO YG ODO ® burkartianum hirtum brevicalyx a) „ tovarii O-O<000O0 P. montevidense y N v .2 € © © . a = 3 TCO Q0 24 stipoides 1 HT P. jubatum 1 20 P. calvescens —{— P. stipoides var. ) 1 echinulatum Figure 3. Strict consensus of seven equally parsimonious trees resulting from phylogene tic analysis of "e haetium (NONA program). A similar tree resulted from the strict consensus of five equally parsimonious trees under implied weight (Pee-wee program). Optimization of characters are shown on the tree: solid bars — unambiguous abel vn S, empty bars = homoplastic character transformations, solid crosses = reversals. Character numbers (see Appendix 3) are above symbols, and character state changes are below q or Crosses. Tae сав show evolution of floret shape in Piptochaetium (see discussion). 314 Annals of the Missouri Botanical Garden IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIS, Nassella LPAI NTE PE PENDAN Hesperostipa Piptochaetium sagasteguii virescens featherstonei cabrerae indutum pringlei hackelii napostaense ruprechtianum bicolor vuvuwy yD озо DHA NV brachyspermum . medium . avenacioides . avenaceum . confusum P. palustre d alpinum angustifolium seleri References #444. Species found all over North and South America Species with — M in North and South Am — found in — or n NA and Guatemala Nabe A Species found in South America Figure 4. One of the seven equally parsimonious trees of 58 fimbriatum lasianthum montevidense — دن panicoides hirtum brevicalyx bh эээ о о э э озозҗоҗз зл ул» tovarii setosum cucullatum jubatum lejopodum stipoides var. stipoides calvescens stipoides var. echinulatum steps resulting from the phylogenetic analysis of Piptochaetium (NONA program), based on morphological and epidermal characters. Distribution of species is mapped on the tree as shown in the references. Numbers below branches represent the Bremer support value for each clade. ateral compression of the floret (14) is a derived character that occurs also in P. angolense (Cialdella & Arriaga, 1998: fig. 9C). The second major clade comprises P. setosum, P. calvescens, P. cucullatum, P. jubatum, P. lejopodum, and two varieties of P. sti- poides (bs = 1). This clade is characterized by the presence of prickles in only one-third of the lemma (19), except P. stipoides var. echinulatum, which has prickles on almost all the lemma surface. With the exception of P setosum, all other species have a rev- Volume 89, Number 3 2002 Cialdella & Giussani 315 Phylogenetic Relationships of Piptochaetium olute crown toward the outside (23; bs = 1): and 19, 22, 25, and 29 is ambiguous on this clade. а among optimization of characters 9, 18, species are not resolved within this clade. Only / brachyspermum and P. cabrerae also have a — crown, a character that appears independently these taxa. Species traditionally included in Piptochaetium sect. Podopogon (Cialdella & Arriaga, 1998) result- ed in a polyphyletic group. Species of this section are linked to some North American species, not in- cluded in any previous infrageneric classification (Table gathered 1, Fig. 3). Some species of this section are in stable clades based on unambiguous synapomorphies: Р, featherstonei, P. cabrerae, and P. indutum have loosely flowered inflorescence branch- 1; bs = 1). while P. cabrerae and P. indutum bs = 1). Piptochaetium cabrerae is characterized by a gla- brous lemma (15) and a revolute crown (23). A clade including P. pringlei (Fig. 5D) to P. palustre is dis- tinguished by the absence of prickles over almost the entire lemma surface (18: bs = 1). es (1 do not have macrohairs in the crown (24: All species of this clade (bs = 1), except P. pringlei, have a glabrous lemma (15). with prickles restricted to the distal third (19). Piptochaetium hackelii and P. na- postaense are identified by a lemma as wide as the bs = 1). This character appears as re- versals in P. indutum, P. pilosum, and P. tovarit. Four crown (21: species, P. bicolor, P. brachyspermum, P. medium. and P. ruprechtianum, are gathered in a clade (bs = 1) supported by the shape of the floret: all of them have obconical florets (13), rather than a terete to fusiform floret. Piptochaetium brachyspermum and P. medium are distinguished by the absence of macro- hairs in the crown (24; bs = 1). Only Р medium has a wide floret with respect to its length (28. Cialdella & Arriaga, 1998: fig. 8B). Only P. cucullatum (Cialdella & Arriaga, 1998: fig. OF) presents a cone-shaped awn (27). and P stipoides var. echinulatum (Cialdella € Arriaga, 1998: figs. 3D, 10J) has prickles with their bases covered by adjacent epidermal cells (20). Pipto- chaetium brevicalyx and P. alpinum present the up- per glume as long as the lemma (12). although this character has an independent origin in both taxa. DISCUSSION Piptochaetium is a monophyletic genus clearly defined by reproductive characters related with lemma and palea structures (Fig. 3). The origin of Piptochaetium was supposed to be in North Amer- ica according to Thomasson (1978a. 1980). who re- lated this genus with Berriochloa Elias as its an- cestor. This fossil grass was first characterized by Elias in 1932. 1985) found some similarities between Piptochaetium and Be- Thomasson (1978a, b. rriochloa based on floret shape. which can be cy- lindrical, obovoid, obconical, or lens-shaped in both genera. Although he emphasized. similarities between epidermal patterns of the lemma, Berrioch- loa, Piptochaetium, and Hesperostipa (treated as Sti- pa sect. Hesperostipa by Thomasson) share a similar lemma epidermis with long fundamental cells with 1978a, b. 1982). This study shows Hesperostipa to be more closely related to Piptochaetium than Nassella. A comprehensive evaluation of sinuous sidewalls (Thomasson, more ص the relationship « Piptochaetium to other genera in the tribe requires a more inclusive study. Within Piptochaetium, species were grouped in two sections: Podopogon and Piptochaetium (Table 1) according to the revision carried oul by Parodi (1944) and as later accepted by Valencia and Cos- tas (1968). Sánchez Vega (1991). and Cialdella and Arriaga (1998 As suggested by cladistic analysis, Piptochae- tium sect. Podopogon is not a natural entity. Con- sequently, infrageneric ranks cannot be supported. The species traditionally included in Piptochaetium sect. Piptochaetium (Parodi, 1944). with the excep- tion of P. fimbriatum and P. pringlei. were assem- bled called the Obovoid group. also including P. brevi- — here 1 - a monophyletic group. informally calyx. This clade is supported by two synapomorp- The Ob- ovoid clade shares other synapomorphies with three hies: a deciduous awn and obovoid florets. North American species: a short callus with P. an- gustifolium and a short and blunt callus with P. fimbriatum and P. seleri. Short and blunt calluses are derived states within the genus, while acute and Fig. 3). This who found that all the floret shape. always had an acute callus. He suggested long calluses are plesiomorphic states — agrees with Thomasson (1985), species of Berriochloa, regardless of that the evolution of the blunt callus in Piptochae- tium is a post-Miocene event. Transition in floret shape was observed by Tho- 1978b) Ne- braska. He found a change in the floret shape from — i=) masson fossils of Berriochloa in the oldest to the youngest stratigraphic levels: cy- lindrical or evlindrical-fusiform shapes were found in the oldest levels, while the youngest levels pre- sented all possible forms within the genus: eylin- drical. evlindrical-fusiform, obovoid. as well as spheroid florets. The present study supports Tho- masson’s findings, showing that obconical, орохо. and lens-shaped florets are always derived states (Fig. 3). Species with obconical florets form a 316 Annals of the Missouri Botanical Garden Л, М | W ==, mae. — eS — ڪڪ а LAE Florets of North American species of Piptoc haetium. — P. avenacioides (Fredholm 5725). Note that the D. P Е 1. P. virescens (King & Soderstrom 5151). Figure 5 scale of this drawing is 2 mm. —B. P avenaceum (Harvey pringlei (Pringle 1410). —E. P. angustifolium (Машаа 29776 a). LE 2 seleri (Nee & Diggs 24775). Pringle 8595). —H. P. brevicalyx (Rzedowski 35863). G. P. fimbriatum — Volume 89, Number 3 2002 Cialdella & Giussan Phylogenetic Relationships of Piptochaetium monophyletic group derived from species with cy- lindrical florets. Species with lens-shaped florets, P. montevidense and P. panicoides, are derived spe- cies within the Obovoid group (Fig. 3). The pres- ence of papillae is another autapomorphy within this small monophyletic group of lens-shaped flo- rets, which is only present in living grasses and never found in fossil taxa studied in this group of Stipeae (Thomasson, 1985). Piptochaetium angustifolium, P. fimbriatum, and P. seleri (Fig. 5E, G, and F, respectively) have both plesiomorphic (persistent awns and cylindrical flo- м rets) and apomorphic (short and blunt callus) states. As shown by the phylogenetic analyses, these three North and Central American species are intermedi- ate in the evolutionary trend through the Obovoid group. Distribution patterns of species in Piptochaetium suggest that these species found new and different habitats to colonize in South America, with а con- sequent higher diversification than in the Northern Hemisphere. The monophyletic Obovoid group is mostly South American, except P. brevicalyx and P. montevidense, which are also present in Mexico, and P. setosum, and P. uru- guense, North and South America (Fig. 4). According to Thomasson’s suggestions (1980), disjunct species were first wide- spread along animal migration routes and, later, due P. stipoides var. stipoides, with disjunct distributions ir =] to severe climatic fluctuation in the late Tertiary, their distribution became gradually reduced, result- ing in relict islands in both hemispheres. Whether species with obovoid florets are derived from species with cylindrical florets in North America or in South America is difficult to determine, as there is no cer- tain evidence for this fact. INFRASPECIFIC LEVELS Results from parsimony analyses showed a differ- ent origin for subspecies of P. tovarii, first recognized by Sánchez Vega (1991). Both subspecies of P. to- varii are separated in all cladistic analyses (see NONA and Pee-Wee results, Fig. 3). erigió tovarii subsp. tovarii, included herein within the burkartianum to P. panicoides clade. is characterized by the absence of prickles on the lemma surface. while P. tovarii var. pilosa (= P. pilosum herein) is the sister group of the P. uruguense to P. stipoides var. echinulatum clade and does have prickles and macrohairs. Based on these results, the study of type material, and the distribution of characters across the entire genus, there is no reason to keep these entities within the same species. Therefore, Р tovaru subsp. pilosum is elevated to the rank of species (see the taxonomic treatment herein). Piptochaetium stipoides is included in the P. cu- cullatum to P. stipoides var. echinulatum clade (Fig. 1). stipoides were not resolved as much as among the Relationships between the two varieties of P. — other species of the clade. A possible source of variation within this clade is the polymorphic state of P. stipoides var. stipoides for several characters principally related with the floret. Moreover, tran- sitional shapes and epidermal characters of the lemma may be variable within a single specimen of this species. TAXONOMIC TREATMENT Piptochaetium J. Presl, Reliq. Haenk. 1: 222. 830, nom. cons. Urachne Trin. sect. Pipto- chaetium (J. Presl) Trin. € Rupr.. Mém. Acad. Imp. Sci. St. Pétersbourg sér. 6, Sci. Nat. 5: 22. 1842. TYPE: Piptochaetium setifolium J. Presl [= P. panicoides (Lam.) E. Desv.]. Podopogon Raf., Neogenyton: 4. 1825. nom. rejic. TYPE: Stipa avenacea L. (lectotype, designated by Clayton, м 1983, not seen) Caryoc — Spreng., Syst. Veg. 4(2): 22. 30. 1827, leg.. non Campanian Trinius. 826. TYPE: yochloa montevidensis Spreng. hom. Car- Caespitose, perennial plants. Culms herbaceous, unbranched above, erect or somewhat decumbent, terete or slightly flattened, longitudinally striate, gla- brous; young shoots intravaginal. Nodes compressed, occasionally thickened, the basal nodes slightly ge- niculate. Leaf sheaths generally longer than the in- ternodes at the base, becoming shorter toward the apex. Ligules membranous, blunt, usually unfringed. Leaf blades linear, convolute to flat, longitudinally striate, frequently glabrous. /nflorescence a panicle with branches densely to loosely flowered. Spikelets fusiform, laterally compressed to terete, usually long pedicellate, the pedicels unequal in length, generally in pairs, rarely alone toward the apex of the panicle. Flowers cleistogamous or chasmogamous. Glumes two, persistent, subequal, usually longer than the flo- et, membranous, lanceolate, acuminate, 3- to 7(8)- nerved, glabrous. Florets easily deciduous, terete to aterally compressed. Lemma indurate, awned, lon- gitudinally striate, generally with epidermal append- ages: lemma margins involute, fitting into the lon- eitudinal groove of the palea; lemma apex fused in a crown, frequently contracted to the base of the awn and with different epidermal appendages. Awn usu- ally twice-geniculate, twisted and hispid on the basal югпоп. straight and shortly scabrous toward the — apex, deciduous or persistent. Callus acute, subacute or blunt and obliquely truncate, usually hairy, rarely glabrous or subglabrous. Palea laterally compressed, 318 Annals of the Missouri Botanical Garden indurate, membranous toward margins, glabrous, shiny and inconspicuously striate, boat-shaped, bi- keeled, the apex of the keels projecting as a minute point above the summit of the lemma. Lodicules 2 or 3, membranous, blunt or acute, glabrous. Androeci- um of З stamens always with short filaments, ca. 0.2 mm long, and small anthers, ca. 0.5 mm long in the KEY TO THE SPECIES OF PIPTOCHAETIUM cleistogamous flowers, and longer anthers 1-5 mm long in the chasmogamous ones. Ovary obovoid, glo- bose to laterally compressed, glabrous, with 2 styles free to their bases; stigmas 2. Caryopsis terete to globose or lens-shaped, surface dull, rugose and gla- brous, with linear hilum and a small embryo; en- dosperm hard, without lipid. la. Florets terete, fusiform or obconical, occasionally fusiform, slightly obovoid; awn robust and persistent; callus usualy acute or subacute, rarely blunt emma glabrous 3b. Lemma with inconspicuous ho Lemma with hooks and prickles за. F ib. Lemma With hooks all over surface, decreasing in density toward the callus and the — ле with hooks and pric — on | the upper F ponini, oks 10. P. cabrerai — 25. P. palustre always conspic uot lo rel terete, not gibbous, slightly narrowed bi the crown, (б—)9—20 mm long, 1— — ume 7(8)-nerved; glumes 20-26(—35) mm long; lemma surface yr with onspicuous longitudinal striae - » P. hackelit 6b. ie glume 3(4)-nerved; glumes 15-23 mm long; lemma surface ЫЫ with conspicuous longitudinal striae Ta. Floret (14-)15-20 mm * RC 5. P. avenacioides 7 Floret n ge Tae )mmlong |... 24. P. napostaense 5b. Floret — slightly gibbous, or obconical, strongly narrowed below the crown, (3.5-)7 8a. ie гаа, Ja. ^ 2 ОЬ. етта; — obe onical. 8b. (—12) mm long, (0.8—)1.2-2 mm diam. pus et 1.2-1.5 mm wide; lemma with hooks and prickles only below the 1; callus 3.5—4.8 mm long, occasionally shorter Floret | mm wide; lemma with hooks and prickles on the distal half of the callus 2-3 mm long P. confusum — . avenaceum own with ascending macrohairs up to 0.6 mm * conspicuously ex- pem the crown in length. sometimes also with a = — llb. oo 10b. бый йш macrohairs, with coke and puciles up to 0.2 mm — tly — than Crown 0.5-4 2 mm wide 12b. кю 1За. "ua is sharp and nar a. Callus 1-2 mm long; floret (3.5— I C rown 0.8-1 mm diam.: de — with macrohairs — — all o over surfac e. w hooks and p ee )4.5-6(—7) mm long > bicolor © Callus 2.5—4 mm long; floret (5.5-)7.5-9.5 mm long — D ruprec — long, crown vs mm dans 4 callus 1.2-1.5 mm long; floret (1.2—)1. 22. P. — ‘callus 1.9-2 mm long; floret 1.2-1.5 mm 7. P. brachyspermum nfk orescence 3-9. em long, with 4 to 15 spikelets; spikelets 9713.5 mm long; Bolivia and Arg 14b. Попе ‘ence longer, 10-30 cm long, with 15 to 80 spikelets; spikele ts 5-10 mm long. PREND DERE P. indutum Lower glume shorter than the floret; upper — as long as the floret; Brazil 15b. Both — — than the flore l6a. Plant spike ; Peru — 1. P. alpinum 250.65 m long; — les 8—11.5 em long; panicles with 15 to 22 NR IA 30. Д у ш 3 16b. Plant ^ 70-1. 30 m Jong; peduncles 40-44 cm ae panic les with 2 25 to 80 13b. Callus subacute or paie always broad. l 7a. ) mm long: Flo ret: long; callu ets; Mexico, Guatemala, and Venezuela Lemma — sometimes a very few hooks present near the crown. 18а. К 7 callus subacute is blunt Р. virescens . 28. P pringlei 15. Р in 17b. Lemma longitudinally striate, with hooks and pric ев all over surface 19а. Macrohairs of the lemma, including those of the crown, 1 mm е floret fusiform ' angu solum 19b. Macrohairs of the * тта, inc йш аве of the crown, 05 5 mm ions: floret terete, slightly obovoid. Volume 89, Number 3 Cialdella & Giussani 319 2002 Phylogenetic Relationships of Piptochaetium 20a. Peduncles 5-14 em long; inflorescences 3-6.5 cm long, dense: Peru - 14. p featherstonet 20b. Pedunc — 22-26 cm бан dali 'scences 57-220 cm i long. loose, not dense: н xico and Guatemala: e ôç . P. seleri Florets obovoid or lens-shaped: awn weak and deciduous; callus blunt. 21а. Lemma with many macrohairs 22a. Macrohairs of the елй up to 4 mm long, conspicuously longer than the crown 20. Р lasianthum 22b. Macrohairs of the lemma up to 2 mm long, the upper hairs slishily longer than the crown — — 27. P. pilosum — 21b. Lemma она, 23a. Aw n * eanie-baped, base as wide as the crown (0.9-1 mm), glabrous, or occasionally with very fev habs. САРЕ GARE ASA 3. P. cuc — 23b. en thread-like. shortly hispid, minutely scabrous toward the ap 2 ‘rown 1-1.8 mm diam., not contracted toward the base of e awn. 25a. Callus glabrous. or with few short hairs as long as the callus. 26a. Hooks of the lemma inconspic uously pointed, eel disposed all over the surface, decreasing in size and quantity toward the callus and near =й crown , и сеп м 26b. Lemma with very few licel. these confined to the distal part ian ig CTOW e d › le — 25b. Callus densely hairy, the hans xe 'eeding the callus. 27a. Hairs of the callus longer than the floret 19. P. jubatum 27b. Hairs s the callus one hind or half as long as the floret. 28a. Lemma with pric ‘kles partially covered by the edges of — ent epidermal s, generally in groups of 2 or 3, unifor mly disposed all over the lemma surface, the biggest ones on dua айай portion of the floret — 33b. P. stipoides var. echinulatum | e. 28b. — with hooks and pric ке not үш by the adjacent cells, uni ormly disposed on the distal half of the floret, or lemma — epider ا‎ ndages sans 33a. P. stipoides var. stipoides 24b. C rown 0.4—0. 6-0. 9 mm diam., contracted toward the base of the awn. a. Lemma with hooks ni wickles, densely and uniformly disposed on the distal % of the floret; papillae absen 30a. Floret globose, sometimes slightly compressed laterally, (1.6—)2.4—3(-3.5) mm long, 0.8-1.2 mm diam.; Paraguay, Uruguay, and northern Argentina "——— P ee 35. P — 30b. Floret е and conspicuously Тами аренда, | (2 5-)3.8-4-5) 1 long, 1-2 1 ide; Се 2. и nse 29b. Lemma ib pay illae or with hooks and prickles only below the crown, or lemma without epidermal appendages 3la. Crown revolute жану. the outside |... 2 83a. P. stipoides var. stipoides 31b. Crown not revolute. 32a. Floret lens-shaped, conspie ‘uously laterally compressed: lemma and crown with or without papillae 33a. Lemma with — dense and uniformly — р рн, in ul size and density toward the crown and the callus . P. montevidense 33b. Lond without papillae, or confined to the region near the crown, on the dorsal side of the floret 26. P. panicoides 32b. Floret globose to slightly compressed, — г obconical; lemma and crown with or аш hooks, prickles, and macrohi р. 34a. Lemma with a few hooks toward the crown . — 82. P. setosum 34b. Le m oce epidermal appendages. 35a. Leaf blades glabrous: floret 2.2-2.5 mm long: callus hairy. the upper hairs % the length of the floret › burkartianum 35b. Leaf blades hispid on both surfaces, sometimes aba Шу gla- brous: floret 2.5-3.5 mm long; callus with or without short hairs, the hairs equaling ilis le meth of the callus. 36a. Floret obovoid, not gibbous, 0.8-1.2 mm — — . P. tovarit 36b. Floret conspicuously — 1.5-1.9 mm — Glumes exceec ding t the length of the floret: lemma dull, with conspicuous еи ıl striae; callus 0.4—0.5 mm long. crown with ] or 2 rows of hooks and prickles on margin . P. hirtum 37b. Glumes as long as or slightly shorter than da floret: lemma shiny, "amos callus 0.150.25 mm long: crown without epidermal appendages _ 8. P. brevicalyx 320 Annals of the Missouri Botanical Garden Phyto- Santa Ca- B. Smith, Brazil. 1. Piptochaetium alpinum L. logia 22: 89. 1971. TYPE: tarina: Bom Jardim da Serra, Fazenda da La- ranja, 1400 m, Reitz & Klein 7710 (holotype, US not seen; isotype, HBR not seen). Distribution. Southern Brazil in the states of Rio Grande do Sul and Santa Catarina, up to 1400 m elevation. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 2. Piptochaetium angolense Philippi, Anales Univ. Chile 93: 734. 1896. TYPE: Chile. “An- gol, Nov. 1887," Philippi s.n. (holotype, SGO 057391!). Distribution. Endemic to central Chile (IV Re- gión). For a description, illustrations, апа specimens examined, see Cialdella and Arriaga (1998). 90 . Piptochaetium angustifolium (Hitchcock) Valencia & Costas, Bol. Soc. Argent. Bot. 12: 177. 1968. Stipa angustifolia Hitchc., Contr. U.S. Natl. Herb. ». 1925. TYPE: Mex- ico. Coahuila: on summit slope of Sierra de L: Puebla in sight of Saltillo, 25 July 1905, Palmer 726 (holotype, US 570290 not seen: isotypes, BAA Herb. Parodi 2955!, MO!, US 993385 not seen, US 906342 not seen ures 2D, 5E pau) — Fig- Plants perennial. Culms 10-30 em tall, glabrous: nodes 2, compressed, glabrous to sparingly pubes- cent; internodes 3.5—4.5 cm long. Leaf sheaths tightly embracing the culms, shorter than the in- 1-2 cm long. blunt, margins entire, glabrous. Leaf blades linear, convolute, 9-10(-20) em long, 0.5-0.6 mm wide, longitudinally striate, glabrous. Peduncles te- ternodes, 3—4 cm long, glabrous. Ligules rete, slightly flattened, 7-10 cm long, longitudinally striate, glabrous. Inflorescences 6.5-7.5 cm long, few-flowered (8 to 18 spikelets), loosely disposed; pedicels 2-7 mm long, scabrous. with branches Spikelets fusiform, 6—6.5 mm long, 1 mm diam. Glumes longer than the floret, subequal, 6—6.5 mm greenish, hya- 4.5-5 long, 0.9-1 mm diam. Lemma 4.5-5 mm long, con- tracted densely pilose, with macrohairs 1 mm long, hooks long, shortly acuminale, 5-nerved, line toward the margins. Floret fusiform, mm below the crown, longitudinally striate, and prickles all over surface; crown contracted at the base of the awn, 0.4-0.5 mm diam., straight, not revolute, with densely disposed prickles and macrohairs, the latter | mm long, persistent; callus 0.6-1 mm long, subacute, densely pilose, the upper hairs slightly longer than the callus; awn bigeni- culate to almost straight, 10-11 mm long, persis- lent, scabrous along length. Palea 4 mm long. Lod- icules 2, 1 mm long, blunt. Caryopsis not seen. Distribution, This species is known only from northeastern and eastern Mexico, in the states of Coahuila, Nuevo León, and Mexico, up to 3500 m. куу, specimens examined. MEXICO, México: elon de Nado, 2700-3500 m. Maruda e et D 29776 um. Nuevo León: Hac inda Jista Herm 35 mi. 5 of Monterrey, 810 m, 25 June 1939, Hanes 1038 (US). Piptochaetium angustifolium is similar to P featherstonei and P. seleri, but these species differ in having short, 0.5 mm, macrohairs on the lemma and crown, and a terete, slightly obovoid floret. 4. Piptochaetium avenaceum (L.) Parodi, Re- vista Mus. La Plata, Secc. Bot. 6(25): 229. 1944. Stipa avenacea L., Sp. Pl. 1: 78. 1753. TYPE: U.S.A. Virginia: Clayton 621 (lecto- type, designated by Clayton (1983), LINN- 94.5 not seen; isolectotype, BM not seen). Fig- ures 2C, 5 Stipa barbata Michx., Bor.-Amer. 1: 53. 1803, ie g. non Stipa barbata — Fl. At Stipa virginica Pers., Syn. Pl. 1: 99, Papen barbatus Ral. « ex as D. Jac ks.. Index Kew. . 1894, nom. inval., as syn. of Stipa avenacea E m U.S.A “In sylvis Virginiae ( «arolinae,” Mi- 1. (holotype, P not seen; isotype, US not nom. er Stipa а avenacea var. bicolor Eaton & J. Wright, Man. Bot. (e d. 8): 444. 1848. TYPE: “United States of Amer- ica” (not locate ad). Stipa teiaha Hitche., Contr. U.S. Natl. Herb. 24: 236. 925. Piptochaetium и (Hitche.) Beetle, 4. 1983. TYPE: : hytologia 54: l Mexico. Puebla: 2600 m, Hitchcock EA (holotype, US 993443 not seen). Piptoc — avenac es Barkworth, Syst. Bot. * 3: 196. 1988. TYPE: Mexico. erc ар 4 km W of on bes 23% 2 N, №, on ime — itle open pine forest, 31 0 ‚ 4 Aug. 1 Stanford e el al. 645A n US | iol seen; vata es MO not — Plants perennial. Culms (30—)50—80 cm tall, gla- brous; nodes 2-3, yellowish to reddish, com- pressed, glabrous; internodes 5-25 cm long. Leaf sheaths embracing the culms, shorter than the in- (4—)6—-16 cm long, glabrous. Ligules 2-3 mm long, subacute to blunt, margins entire, gla- ternodes, brous. Leaf blades linear, convolute, occasionally flat, 8-30 cm long, 0.8-0.9 mm wide, adaxial side conspicuously striate, glabrous or scabrous, mar- gins shortly scabrous. Peduncles terete, slightly flat- tened, 8-10 cm long, longitudinally striate, gla- Volume 89, Number 3 2002 Cialdella & Giussani 321 Phylogenetic Relationships of Piptochaetium brous. Inflorescences 15-22 cm long, few-flowered (10 to 15225) spikelets), with branches loosely dis- posed; pedicels flattened, Spikelets fusiform, 10—12 Glumes subequal, longer than the floret, 1.5-5 em long. hispid. mm long, 1-1.2 mm diam. 10-12(— 15) mm long, acute, lower glume 3(5)-nerved, up- per glume 5-nerved. Floret terete-fusiform, slightly compressed, (7—)9-12(—13) mm long, 1 mm diam. Lemma contracted below the crown, 9-12(-13) mm long, thinly striate, with hooks and prickles toward the distal portion (1/3-2/3 of the surface): crown contracted at the base of the awn, straight, not rev- olute, 0.5-0.6 mm diam., with prickles and macro- hairs up to 0.5 mm long, occasionally up to | mm long; callus ca. 2 mm long, sharp, hairy, the upper macrohairs reaching 4-4 the length of the floret: awn bigeniculate, 5-7 cm long, persistent, hispid, scabrous toward the apex. Palea 7-8 mm long. Lod- icules 2, 1.5 mm long, acute. Caryopsis terete, 5.5— 6 mm long, 0.5-0.6 mm diam.; hilum linear: em- brvo Y, the length of the carvopsis. Chromosome numbers. 2n = 22, 28 (Gould, 1958, as Stipa avenacea L.); n = 11 (Valencia € 1968). Distribution. Costas, Widespread in the eastern and southern United States and northeastern Mexico. Alabama: Additional specimens examined. — ng Vani: 240, Marshall Co.. margin of a woodland, а about ТО mi. E of Morgan City, Hen der — Il~ (MQ): Lee Co., Auburn, Earle & — 8. 73). ES insas: Garland Co., 30 Apr. 1939, Den maree ped (US); Coastal Plain Renan E maree 63483 (MO Haven Co., dry Weatherby 5 Wildlife . Connecticut: New yi PE es toward im of ML. Carmel, Florida: Wakulla Co.. Marks А СК 67795 е Co.. \ of Sebring. . Deam 64413 ( ; Gaine 8- ue iS 99248 ,eorgia: Ben n Py 16.3 mi. ENE d Fitzgerald. Fokel 5079 (MO): Glynn o., N end of Jekyll Island, Strong 1342 (US); Walker Co., near summit of ridge near Maddox Gap, between Villanow and La Fayette; Cronquist 4432 (US). Indiana: N side of the Pidgeon River about 3 mi. SE of Mongo, Deam 40701 (US). Louisiana: Pipeline off La. 507, 3 mi. W of Bien- ville, Larrick 249 (ВАА); roadside about 2 mi. NE of Mansfield, — 28558 (1 ы” т aryland: region of Marl- : 12616 (US). sachusetts: Barnstable C 20.. Hoffmann s.n. (MO): гн КО ul re , near junction with hwy. 138, Blue Hill Re ervation, SE of Boston. piri & Reeder zo (US). Mie h- igan: 6 mi. W of Richland at Spring Brook and railroad, Harvey 1764 (Sl); Kalamazoo Co.. Sect. 36, 2 mi. S of Alamo. — 1854 (ВАА. US). Mississippi: Lowndes j W of Mayhew junction on ridge, Me т 1811 a Co.. Bucatunna. Jacob 815 (US). New Jer- ear Je kins, > Barrens. Harkin 809 (МО); we 7312 (US). New ‘ork: Eastport, Long 1. land, Chase 7401 (US). North Caroli na: Wake Co., Lake Johnson, 4 mi. SW of Raleigh, — 3686 (US); Durham Co., wooded slope. Simpson s.n. 2479778). Oklahoma: Le Flore Co.. near state line on hwy. 63, E Pennsylvania: s.l., Providence Co., Collins s.n. ( 1 of US of Big Cedar, Waterfall 14854 (US). Scribner 266 (US). Rhode Island: us oe South Caro- . just SE of Ora. Миў $ of Wale Be "ach; Griscom 16409 ina: Laurens Co., side 23032 (MO); (US): MeCormick Co.. lr f Savannah River on SC hwy. 28, Hos eman & Rin 8642 (US). Tennessee: Marion Co.. 9 mi. SE of Tra City, slo pe of ч wns Plateau, W of Rock- Horry Сох Roane Co.. wood, Wesker 6234 xas: Jasper Co., 8 mi. SE of Jasper, hw 90, E н 7029 (SI); Shelby Co., к scrub oak 3 7 mi. S of Center. Correll 16170 (US). Virgi ia: picta Co., Chincoteague Island, шаи 8537 ( x Hampton, dry open woods, Miller Jr. s.n $ lT We уппа: on wooded — near Rey- mon Memorial farm. Wardensville, Berkle ) (MO). MEXICO. Tamaulipas: 3 mi. N of i p бй Hare et al. 2482 (US, paratype of Piptochaetium avenacellum). Piptochaetium avenaceum resembles P. confusum and P napostaense, as they share the shape of the floret and the appendages (prickles and hooks) on the lemma and crown, but they are distinguished as follows: in P. avenaceum, the prickles and hooks are distributed on the distal portion. sometimes reaching half the length of the floret. In P? confusum and P. napostaense such appendages occur only im- mediately below the crown. Both P. avenaceum and P. napostaense have florets that are | mm in di- but the avenaceum and ameter, callus length is 2-3 mm long in P 3.56 mm in P napostaense. Pip- tochaetium confusum has florets 1.2-1.5 mm in di- ameter with a callus 3.5-4.8 mm long. 5. Fiptochadtium avenacioides (Nash) Costas, Bol. Soc. Argent. Stipa y ere . Club 22 T Lake on bi pine dn near Cassia, 16/ 30 June 1895, Nash 2051 (holotype, NY not BAA!, MO!, US!). Figure 5A. Valencia Nash. TY PE: U.S.A. Flor- seen: isotypes, Plants perennial. Culms 0.70-1.20 m tall, gla- brous or pubescent below the basal node ы nodes yellowish. glabrous: internodes long. Leaf sheaths ¢ embracing the culms, shorter than the Ligules 1-3 glabrous. — internodes, em long. glabrous. (-2 mm long, blunt or acute, margins entire, Leaf blades linear, convolute, 15-30 em long. | mm wide, longitudinally striate, glabrous. Peduncles te- 22-25 cm brous. Inflorescences 10-30 cm long, few- to many- rete, long. longitudinally striate, gla- flowered (10 to 50 spikelets), with branches loosely disposed: pedicels 1.5-2 em long. scabrous. Spike- lets Glumes longer than the floret, 18-22 mm long, acu- fusiform, 18-22 mm long, 1-2 mm diam. minate, (3)5-nerved. Floret terete, (14-)15-20 mm long, 1-1.2 mm diam. Lemma 15-20 mm long, con- tracted below the crown, longitudinally striate, with 322 Annals of the Missouri Botanical Garden small hooks on the distal portion (1/3-2/3 of the surface); crown contracted at the base of the awn, 0.6-0.7 mm diam., crohairs 0.4—0.5 mm long, with hooks, prickles, and ma- mm long, persistent; callus 6—8 acute, densely hairy; awn bigeniculate, persistent, 8-12 mm long, hispid at the basal por- tion, scabrous toward the apex. Palea 12 mm long. Lodicules 2, ca. 2.5 mm long, acute. Caryopsis not seen. Distribution and ecology. Only known from the United States, in Florida. This species grows in dry and sandy places. Additional specime ns. "red ed. U.S.A. Florida: Bre- vard Co., Fredholm 5725 (BAA, MO, US); Eau Gallie, Curtiss 5834 (MO, US); н Co., Mermiack, Baker 115 US); Manatel Co., Palma Sola, Tracy 7031 (MO, 05); Du- val Co., Fredholm 5092 (US); W of Tavares, Hitchcock 812 (US); Braidentown, Hitchcock 967 (US); s.l.. Dowell 7 7309 (BAA, US); Orange Co., Clarcona, Meislahn 191 (US — 6. E bicolor (Vahl) E. Desvaux, Fl. Chil. E 1853. Stipa bicolor Vahl, Symb. Bot. — "id Oryzopsis bicolor (Vahl) Speg.. "iia us. Nac. Montevideo 4(2): 6. 1901. Piptoch — bicolor var. typicum Pa- Mus. La Plata 6(25): 252. 1944, . TYPE: Uruguay. “Habitat in Bra- siliae Mons Video. Dn. Thouin," Herb. Vahl, IDC microfiche photo Vahl, nr. 73 Ш, 4-5 (ho- lotype, С!). Stipa — Trin. & Rupr., étersbourg sér. 6, Sci. Nat zl. "Brasilia me топа — 9 784 (holotype, LE not 3 not see d Stipa Lt Sag ke Steud., Syn. Pl. Glumac. 1: 124. 1854. * Chile. Valparaiso, Concon, Nox 1829, Herb. s pri > Oryzopsis bicolor (Vahl) S jus Acad. Imp. Sci. St. 26. 1842. TYPE: Bra- seen; isotypes, US!, “ minor Speg., Anales . 1901. Piptoc от us. Nac. onte — ^0 102 y < bicolor (Vahl n . Desv. var. minor opa g.) Parodi, Re- vista Mus. La Plata, Secc. Bot. 6(25): 256. 1944. TYPE: Argentina. Buenos Aires: Bavio, Estancia Elizalde, Spe gaz ie s.n. — LPS 12513! sheet A; isotype, LPS 12513! sheet В). Distribution. Central Chile, in V Región, west- ern and southern Uruguay in the departments of Canelones, Florida, Montevideo, and San José, and central eastern. Argentina in the provinces of Bue- nos Aires and Entre Ríos. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 7. Piptochaetium brachyspermum (Spegazzini) Parodi, Revista Mus. La Plata, Secc. Bot. 6(25): 241. 1944. Oryzopsis napostaensis Speg. var. brachysperma Speg., Anales Mus. Nac. Montevideo 4(2): 17. 1901. TYPE: Argentina. Buenos Aires: Carmen de Patagones, Spegaz- zini 42 b (holotype, LPS 2471!) Distribution. Endemic to Argentina, where it is only known from the southern province of Buenos Aires. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 8. Piptochaetium brevicalyx (E. Fournier) Ric- er, Contr. U.S. Natl. Herb. 17: 286. 191 Mexic. Pl. 2: 150. = Stipa brevicalyx E. Fourn., 1886. TYPE: Mexico. San Luis Potosf: cerca de San Luis Potosí, 6000-8000 ft., 1878, Par- ry & Palmer 959 (holotype, US not seen; iso- types, BAA!, K not seen, MO not seen, NY not seen). Figures 1A, 2A, 5H Piptochaetium brevicalyx (E. Fourn.) Ricker subsp. flexuo- um Barkworth, Syst. Bot. = 200. 1988. TYPE: i of Durango City on road to La Flor, 2620 m ale, Breedlove 44215 (holotype, CAS 622475 not seen). Culms (6—)10—50 cm tall; nodes 3 or 4, compressed, reddish, glabrous; inter- Plants perennial. nodes 2-12 cm long. Leaf sheaths tightly embracing the culms, shorter than the internodes, 2-5 cm long, glabrous to scabrous. Ligules 0.6-1.2 mm ong, blunt to subacute, margins entire or fringed, glabrous. Leaf blades linear, convolute or flat, 5—30 cm long, 0.6-1.2 mm wide, glabrous or hispid. Pe- duncles terete, 2.5—7 cm long, longitudinally striate, glabrous. /nflorescences 3—10 cm long, few-flowered (5 to 20 spikelets), with branches densely disposed; pedicels = mm long, glabrous or hispid. Spikelets globose, 2.5-3.5 mm long, 1.5-1.7 mm diam. Glumes subequal, shortly acuminate, violet-colored, hyaline toward the margins, 5- or 7-nerved; lower 2.4-3.3 mm long: upper glume shorter than or as long as the floret, 2.4—3.5 compressed, gibbous, 2.5-3.5 mm long, 1.5-1.9 glume slightly shorter than the floret, mm long. Floret globose, slightly mm wide. Lemma contracted below the crown, 2.5— 3.5 mm long, smooth, shiny, without epidermal ap- pendages: crown contracted at the base of the awn, inconspicuous, straight, not revolute, 0.5-0.7 mm diam., without epidermal appendages; callus 0.15- 0.25 mm long, blunt, obliquely truncate, with few, occasionally many, short hairs; awn slightly genic- ulate to almost straight, 6-18 mm long, hispid, sca- brous toward the apex, deciduous. Palea 2.5-3.7 mm long. Lodicules 2, 1 mm long, acute. Caryopsis not seen. Distribution. tely 3000 m. Central Mexico, up to aproxima- Additional specimens examined. MEXICO. Durango: 5.3 km W of (Ejido) Los Mimbres turnoff, litis et al. 214 (US); 41 km NE of El Salto, Mick & Roe 87 (US! paratype Volume 89, Number 3 2002 Cialdella & Giussani 323 Phylogenetic Relationships of Piptochaetium of P. brevicalyx subsp. flexuosum). Guanajuato: 10 km al sur de Ibarra. Ocampo, sobre la carretera a León, Rze- dowski 50768 (MEXU). Hidalgo: 8 km al noreste de Pa- ica, sobre la carretera a Real del Monte, Rzedowski (MEXU, XAL): Cerro Ventoso. above Pachuca. Pringle 7606 (MEXU): in oak woods near Real del Monte, Rose et al. 8680 sack Pachuca, Hitchcock 6733 (US). México: rocky roadcut, 27 mi. S of Juan del e Gould 9211 (US 5). San pus Potosí: Cerro La Bolsa, ' Cerrito s Dolores, Villa de Arriaga. B el di 132 (CAS, MO): near Cerro Prieto, Sohns 1535 (US). Tlaxcala: js. San Cristóbal and Calpulalpan. Sole 553 (MEXU. US). Piptochaetium brevicalyx is similar to P. hirtum: both species share the globose floret shape. the ab- sence of epidermal appendages on the lemma, and generally few hairs on the callus. They can be rec- ognized because P hirtum presents conspicuous longitudinal striae on the dull lemma surface. | or rows of hooks or prickles in the crown, and both glumes 5.5-6 mm, longer than the floret. Barkworth (1988) recognized Piptochaetium bre- vicalyx subsp. flexuosum based on the relatively long and widespread branches of the inflorescenc- es. Although there are specimens with dense or ex- tended panicles, the difference between the length of the branches is slight. and, within the studied material, there are specimens with both types of inflorescences, such as Banda et al. 132 * ону burkartianum Parodi, Revis- ta Mus. lata, Secc. Bot. 6(25): 291. 1944. TYPE: пан Corrientes: San Martin, La Cruz. Nov. 1936, Parodi 12329 1/2 (holotype. BAAS). Distribution. Endemic to Argentina; only found in the eastern province of Corrientes, at the type locality. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 10. Piptochaetium cabrerae Parodi, Revista Mus. La Plata. Secc. Bot. 6(25): 243. 1944. TYPE: Argentina. Buenos Aires: Puan, Villa Iris. 7 Nov. 1940, Parodi 13771 (holotype, BAAN. Distribution. Endemic to Argentina: only col- lected in the southwestern province of Buenos Ai- res. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 11. Piptochaetium calvescens Parodi. Revista Mus. La Plata. Sece. Bot. 6(25): 278. 1944. TYPE: Argentina. Prov. Buenos Aires: Sierra Currumalán, 600 m s.m., 10 Nov. 1932, Pa- rodi 10343 (holotype, BAAN. Distribution. Central and eastern Argentina in the province of Buenos Aires, and southern Uru- guay in the department of Montevideo. ‘or a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998) 12. Piptochaetium confusum Parodi, Revista Mus. La Plata, Secce. Bot. 6(25): 246. 1944. TYPE: Argentina. Entre Ríos: Concordia. 3 Nov. 1921, Parodi 3950 (holotype. BAA!). Distribution and ecology. Uruguay in the de- partments of Maldonado, Montevideo, and Paysan- dú, and central eastern Argentina in the province of Entre Ríos. This species grows in dry grasslands and rocky places. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 13. Piptochaetium cucullatum Rosengurtt & = Univ. Rep. Fac. ‚ 1960. TYPE: Uru- guay. Dpto. Salto: — del Arapey. próximo 9 Dec. 1962, МУКА not US 2951786 not seen). Izaguirre de Arturio, Agron. Montevideo 90: a la Estación, 1512 (holotype, BAA!, Arrillaga et al. seen: 180ty pes. Distribution. Endemic to northern Uruguay in the departments of Salto and Paysandú. ‘or a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998) (Hitchcock) 14. Piptochaetium featherstonei Tovar, Opusc. Bot. Pharm. Complut. 4: 104. 1988. Stipa featherstonet Hitche., Proc. Biol. Soc. Wash. 36: 196, 1923. TYPE: Peru. Rio Blanco, 4500 m, 8-19 May 1922. J. Е Mac- bride & W. Featherstone 803a (holotype, F 517 í 331 seen). pro parte not seen: isotype, US not Piptoc ноне Мети Tovar & Gutte, egni Veg. 91(4): 205. 1980. TYPE: auli, La Oroya, Berg oberhalb Pachacayo (Puya н Vorkommen). Steiniger Abhang, Hohe, 3750 m. Gutte 2130 (holotype. LZ not seen: vendi SME not seen, USM!). Feddes Repert. Distribution. Southwestern Peru in the depart- ments of Ancash, Ayacucho, and Junín at eleva- tions between 3000 and 4500 m. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). Vote. Hitchcock (1923) that the holotype is mounted with plants of Stipa hans- mentioned meyeri, but, according to Soreng (pers. comm.). it is not clear if that specimen was subdivided into parts a and b by the collectors or by Hitchcock 324 Annals of the Missouri Botanical Garden himself; the US isotype has only Stipa feathersto- net on it. 15. со fimbriatum (Kunth) Hitch- cock, J. Wash. Acad. Sei. 23(10): 453. 1933. Spa pu Kunth, Nov. Gen. Sp. 1: 126. 1816. Oryzopsis fimbriata ee Hemsl., Biol. Cent.-Amer., Bot. 1885. TYPE: Mexico. Guanaxuato, Billalpando, Humboldt & Bonpland s.n., Herb. HBK 4224 (holotype, P not seen; isotype, BAA!). Figures 2B, 5G. Piptoc — — (Kunth) ү s Eo — M. Johnst., J. A : : Ў JO. г! — ‘bor . TYPE: Mexico. ca uila: 'a del jm » La n. ‚ John- ston & Mueller 186 үкөм GH по! seen; — LL not seen). Plants perennial. Culms 35-80 cm tall, glabrous, sometimes pubescent below the nodes; nodes 2—3, slightly compressed or enlarged, reddish, glabrous; internodes 6-11 em long. Leaf sheaths tightly em- bracing the culms, shorter than the internodes, (3—)5-8 cm long, glabrous. Ligules 1.5-2 mm long, acute, glabrous. Leaf blades linear, convolute or flat, 6-26 cm long, 0.9-1.1 mm wide, adaxial side con- spicuously striate, glabrous on both faces, margins scabrous. Peduncles terete, 12—28 cm long, longi- tudinally striate, glabrous. /nflorescences 14-20 cm long, 60 spikelets), with branches loosely disposed; pedicels flattened, 4—12 many-flowered (20 to mm long, hispid. Spikelets fusiform to slightly ob- ovoid, 3.5—5 mm long. | mm diam. Glumes sube- qual, longer than the floret, 4-6 mm long, shortly acuminate, 5- or 7-nerved. Floret fusiform, slightly obovoid and gibbous, 3.5-5 mm long, 1.2-1.8 mm diam. Lemma contracted below the crown, 3.5—5 mm long. hairs, densely disposed; crown contracted at the smooth, shiny, with deciduous macro- base of the awn, inconspicuous, straight, not revo- lute, ca. 0.8 mm diam., ages; callus 0.2-0.5 mm long, truncate, hairy, the upper macrohairs reaching V, the length of the floret; awn bigeniculate, 10—20 mm long, persistent, hispid to scabrous. Palea 3.5— 3.6 mm long. Lodicules 2, ca. | mm long, subacute. Caryopsis fusiform, 2.5 mm long, 0.6 mm diam.; the length of the cary- without epidermal append- blunt, obliquely hilum linear; embryo 1/5 opsis. Chromosome numbers. 2n n = 21 (Reeder, 1968); 2n = tas, 1968). Distribution, southeastern United States and Mexico, at eleva- tions between 1900 and 2500 m. This species has — 44 (Brown, 1951): 44 (Valencia & Cos- Widespread in the southern and also been cited for Guatemala, in the department of Huehuetenango (Swallen, 1955; Beetle, 1977). ap iud — ns examined. U.S саве Сапе elo Hills, along me 15 ni. N of & McHargue 2222 (MO); Patagonia ey € Peobles 10114 (US); Cochise Co., Creek Chiricahua Mountains, Peterson & Annable 1010 — Lh New M v. 104, ex © Y 722 © М = № 355 cky \ . Correll 13099 (US tains, Brewster Co., l mi. SW of Boot Spring, 2120 m, Moore & Steyermark 3172 (MO, US). MEXICO. Baja Cal- Sur: granitic — surrounding long interior val- ley (La Laguna), S of Pico La Aguja on the Sierra La Laguna, Breedlove & — 43293 (MO). Chiapas: Teo- = Chihuahua: foothills MEXU); Miñacam, Hitchcock 7741 (US navista, Saltillo, Cañón San — nzo. 5 cia Zacatecas, Madrigal A. s.n. (XAL); Canon San Loren- zo, 5 km de Saltillo hacia Zac 'atecas, Serrato Sánchez s.n. (MEXU 309021). Distrito Federal: Cima de Toluca, Lyonnet 48 (MO). e on top of rugged volcanic plateau 5 5 »3 km W of (Ejido) Los Vibes turnoff al km 54.2 WSW of — and ca. halfw El Salto, /ltis el al. 213 (XAL). Guanajuato: ca. 8 km W of San Felipe, Sohns 424 (MO, US). Hidalgo: open pine woods near Ocotillos between — 0: and Huasca, ca. 2200 m, Moore & Wood Jr. 4164 ). ТТ о: Mezquitic, 8 km SE del Rancho El Mortero, — ski 17704 (US). Méx- ico: Tezer oco, lado sur de la Cañada de Aguas, 13.5 km s E de Tezcoco. 11 km al ESE de Coatlinchán, Koch 77110 (MEXU); — old hwy. 190 between turnoff to c haleo (Hwy. 115 ita Bárbara ca. 30 m above Azotla, ] . Mie оаа 0.6 mi. E of Poblado ca. 20 mi. W of Morelia, . Nuevo Leon: 28.6 mi. N La Ascen- Brunken & Perino 234 (MO). Oaxaca: 7 km N de быш. Nochixtlán, Mendoza is Pia 2640 MO). Puebla: Tepeyahualco, Calzada et al. 4645 (XAL); vicinity of San Luis Tultilanapa, Purpus 3589 (US). San uis Potosí: Cerro La — 2.5 km S de La Ama- pola, Escalerillas, Zavala Ch. 862 (XAL); in canyons in the Sierra de San Miguelito, ca. : dus ы of Te srrero, 1850— 2200 m, 8 Sep. 1954, Sohns 1147 (U de los Ajos, dro ho de los Ajo т уа et al, 92-794 (MO). Tamaulipas: Marcella. Stanford et al. 2604 (MO). Tlaxcala: between San Cristóbal and Capu- lalpan in open, lightly wooded areas, Sohns 587 (MEXU) Contadero, Pringle 8595 (MEXU). Zacatecas: 9.5 mi. W of Sombrerete, Taylor & Taylor 6254 (US) — Ф 5 m ES oc N EX pu The epidermal surfaces of the lemma in Pipto- chaetium fimbriatum and P. pringlei are similar, as both species present smooth and shiny floret sur- faces with the macrohairs deciduous. However, P. pringlei differs in having bigger florets, from 7 to 10 mm long, and a subacute callus. Volume 89, Number 3 2002 Cialdella & Giussani 325 Phylogenetic Relationships of Piptochaetium 16. Piptochaetium hackelii (Arechavaleta) Pa- . Revista Fac. 7(1): 162. Anales Mus. 1895. Anales 4(2): 10. 1901. Pipto- Agron. Veterin. Stipa hackelii Arechav., Nac. Hist. Nat. Oryzopsis hackelii (Arechav.) Speg., Mus. Nac. Buenos Aires 4: 179 Montevideo chaetium hackelii (Arechav.) Herter. Estud. Bot. Reg. Urug. 4: Florula Urug.: 34. 1930, comb. superfl. TYPE: Uruguay. Cerro de Mon- Arechavaleta 39 isotype, LPS 2480! tevideo, cerca de la cumbre. a (holotype, not located: US not seen). Stipa — Kuntze, Rev. Gen. Pl. 3(2): 373. 1898. TYP Nov. 1892, Argentina. Buenos Aires: Tandil. Otto Pea s.n. (holotype, NY! Distribution. This species grows in the grass- lands of Uruguay in the departments of Lavalleja. Maldonado. Argentina in the provinces of Buenos Aires and La and Río Negro. as well as in central Pampa. For a — illustrations, and specimens examined, see Cialdella and Arriaga (1998). Note The holotype of Stipa hackelii Arechav. could not be located. According to Alonso Paz (pers. comm.), Arechavaleta did not choose types. so in many cases a lectotype must be designated. 17. — ‘haetium hirtum Philippi. Anales Univ. Chile 43: 559. 1873. TYPE: Chile. Cerro Dec. 1864. Philippi s.n. (lectotype. designated by Cialdella & Arriaga (1998). SGO 45078!; isolectotypes, K!, SGO 57422!). de Renca, — haetium З Phil. Anales Univ. Chile 93: 33. 1896. — Cerro de Renca, 11 Nov. 1877. 20 s.n. (le . designated by Cial- della & Arriaga (1998), pr 0 45077!; SGO 45078!, SGO 57421). isolec lolvpes. Distribution. Known from Cerro de Renca it Chile and San Martin de los Andes in the province of Neuquen in Argentina (Parodi, 1961). For a description and illustrations, see Cialdella and Arriaga (1998). 18. rapes tium indutum — Revista lus. La Plata, Sece. Bot. 6(25): 258. 1944. TYPE: Argentina. Salta: Rosario de Lerma. Puerta Tastil, 2700 m, Venturi 8414 (holotype. US not seen: isotype, BAAD. Distribution. Originally known from northern Argentina from the provinces of Salta and Jujuy. distributed in Bolivia in the departments of La Paz and Potosí, central western Peru in the departments 2500 it is also cited from Ecuador in the La Libertad, and Lima. between 1500 m: of Ancash, and province León-Y ánez. 19991. of Azuay (Jorgensen « PERU. Ancash: Re- Río Pachacoto, mw & . Huascarán National Additional specimens артлап. cuay Prov., Huascaran irk, Torres 11794 ; (MO); Tie Prov Park. Llanganuco Sector. Smith 10488 (MO): Recuay Prov., Quebrada Huanca, Smith: & Buddensiek 10964 (MO). La Libertad: Sánchez machuco road, 20 km quez M. 3327 (MO W of и ену ан Smith & Vás- For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998 19. Piptochaetium jubatum Henrard, Recueil Trav. Bot. Néerl. 36: 537. 1939. TYPE: Uru- guay. Canelones: Río Santa Lucía. Paso Cue- По. Gallinal et al. 2198 (holotype, Herb. Lugd. Bat. n. 938,280-383, 11: 1723425 not seen, BAA!) isotypes, US Distribution. Known only from the department of Canelones in Uruguay. For a examined, description, illustrations, and specimens see Cialdella and Arriaga (1998). 20. Piptochaetium lasianthum — Grisebach, Symb. Fl. Argent.: 297. 1879. TYPE: Argen- Entre Ríos: Quinta del Colegio, en praderas. Lorentz s.n.. tina. Concepción del Uruguay. Flora Entrerriana n. 1157 (holotype. not lo- CORD). cated; isotype, — s erianthum Balansa. Bull. Soc. Bot. France i і. 1885. TYPE: Uruguay. Cerro de Montevi- s "0. “187 4. Balansa 9 (holotype. not located). — Distribution. Southern Brazil in the states of Rio Grande do Sul and Santa Catarina, central and southern Uruguay in the departments of Flores, La- valleja, and Montevideo, and widespread in central eastern Argentina in the provinces of Buenos Aires, Córdoba. Corrientes, Entre Rios, Misiones. and Santa Fe. Additional specimen examined. ARGENTINA. Misio- nes: Candelaria. Villa Venecia, Renvoize et al. 3025 (MO). For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 21. Piptochaetium la. (Spegazzini) Henrard, Recueil Trav. Bot. Néerl. 36: 530. 1939. Oryzopsis — Speg., Anales Mus. Nac. Montevideo 4(2): 19. 1901. TYPE: Ar- gentina. Sierra de la Ventana, Valle de Nov. 1895, C. L. Spe- баат s.n. LPS 12660). Buenos Aires: las Vertientes, (holotype. Distribution. Southern Uruguay in the depart- © A 326 Annals of the Missouri Botanical Garden ments of Lavalleja and Maldonado, and central eastern Argentina in southern Buenos Aires. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 22. Piptochaetium medium (Spegazzini) Torres, Bol. Soc. Argent. Bot. 11(4): 251. 1969. Ory- zopsis bicolor (Vahl) Speg. var. media Speg.. Anales Mus. Nac. Montevideo 4(2): 9. 1901. TYPE: Argentina. Buenos Aires: Sierra de Cu- ramalal, Dec. ‚ C. L. Spegazzini s.n. (lec totype, designated by — (1969), LPS 12517!). Distribution. Southern Brazil in the state of Rio Grande do Sul, southern Uruguay in the department of Florida, and eastern Argentina, widespread in i Córdoba, Entre Ríos, La Pampa, Misiones, and Santa the provinces of Buenos Aires, Additional specimens examined. BRAZIL. Rio Grande do Sul: Esmeralda, 29 Nov. 1988, M. Sallés 141, 142 (MO) For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 23. Piptochaetium montevidense (Sprengel) Parodi, Revista Fac. Agron. Veterin. 7(1): 163. 1930. Caryochloa montevidensis Spreng., Syst. eg. 4(2): 30. 1827. Oryzopsis montevidensis (Spreng.) Hauman, p us. Nac. Hist. Nat. Buenos Aires 29: 116. 1917. montevidensis (Spreng.) Speg., Revista Argent. Bot. 1(1): montevidensis (Spreng.) Speg. f. typica Speg.. Revista Argent. Bot. 1(1): 11. 1925, Piptochaetium montevidense (Spreng. Estud. Bot. Reg. Urug. 4: Florula І 930, comb. superfl. TYPE: Uru- guay. Montevideo: Sellow s.n. (holotype, B not seen; isotype, MO!) Oryzopsts 1925, comb. superfl. Oryzopsis nom. in- — Urachne panicoides Trin. var. brasiliensis Trin. & Rupr., — Acad. — Sci. St. dis t. 5: 23. ar. basili ensis (Trin. & Rupr.) Bot. 13(10): 40. 1913. TYPE ad fines regni Paraguayani,” 10t located). Piptoc haetium tuberculatum E. Desv., in Gay, Fl. Chil. 6: 272. 1 ryzopsis tuberculata (E. Desv.) Speg., Anales Mus. Nac. Montevideo 4(2): 26. 1901. TYPE: Chile. Valdivia: Gay s.n. (holotype, P!). в — Steud., Syn. Pl. Glumac. 1: 123. 1854. Chile. Rancagua: Bertero 456 (holotype, P!). Piptoc * ie verrucosum Phil., Linnaea 33: 280. 1865. zopsis verruculosa (Phil.) Speg., error for verru- Anales Mus. Nac. Montevideo 4(2): 28. 1901. Oryzopsis verrucosa (Phil.) Speg., Revista Argent. Pétersbourg sér. 6, Sci. 342. Caryochloa montevidensis Spreng. ) Doll ex Ekman, Ark. "A Montevideo usque ° Sellow s.n. (holotype, cosa, Bot. 1(1): 12. 1925. TYPE: Chile. Valdivia: San Juan, — 5. з fnoletype, SGO 57334'; probable isotype, SGC 350!). ae и бите Phil., Anales Univ. Chile 93: 730. TYPE: Chile. colo on: médanos, Philipp s.n. е 6 45087 Piptocl haetium diiit Phil, Anales Univ. Chile 93: 732. 1896. TYPE: Chile. Concon: Philippi s.n. (ho- lo otype, SGO heats ШЕ moelleri Phil., Anales Univ. Chile 93: 734. 896. TYPE: Chile. Renaico, Möller s.n. iu 7332. ) 578: Oryz zopsis montevidensis (Spreng.) Speg. f. оо .. Revista Argent. Bot. 101): 11. 1925. TYPE Argentina Misiones: Posadas, praderas de * alre- dedores, Spegazzini s.n. (holotype, LPS 13166! PENNE montevidensis (Spre ng) Speg. f. ase levista Argent. Bot. з ): 11. 1925. ТҮРЕ: Brasil Rio Grande do Sul: Uruguayana, campos se- cos, Spegazzini s.n. eee LPS 13175). Distribution. Found in Venezuela in the state of Trujillo, and widespread in southern South America in Bolivia, departments of La Paz and Cochabamba, central Chile, IV, VI, and УШ Regions, southern Paraguay, departments of Misiones and Paraguarf, southern Brazil, states of Paraná, Rio Grande do Sul, and Santa Catarina, southern Uruguay, depart- ments of Colonia, Lavalleja, and Maldonado, and northern and central Argentina, provinces of Bue- nos Aires, Chaco, Cérdoba, Corrientes, Entre Rfos, Misiones, Salta, and Santa Fe. This species, which has been considered disjunct in Mexico, northern South America, and southern South America (Tho- masson, 1980), has also been cited for Peru (Brako & Zarucchi, 1993) and southern Ecuador, province of Chinchipe (Jørgensen € León-Yánez, 1999). This distribution pattern makes it probable that it also occurs in Colombia, and, together with P. pa- nicoides, would confirm one of the possible migra- tion routes proposed by Thomasson (1980). This species is the most frequent of the genus, growing on riversides, grasslands with rocky places and modified soils, up to 3700 m elevation. Additional specimens examined. BRAZIL. Paraná: Ruta BR- 280, hacia — nte, a 1 km de ruta BR-153, Rúgolo et al. 1611 (MO). VENEZUELA. Trujillo: Cara- che, 220 m, Rivero & Díaz 1384 (MO). For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 24. Piptochaetium A dup n (Spegazzini) Hackel, Anales Mus. Nac. Hist. Nat. Aires 11: 103, 1904. Ts napostaensis . Anales Mus. Nac. Montevideo 4(2): 15 . Oryzopsis napostaensis Speg. var. ma- crophylla Speg., Anales Mus. Nac. Montevideo 4(2): 17. 1901. TYPE: Argentina. Buenos Ai- res: alrededores de Carmen de Patagones, Feb. 1898, C. L. Spegazzini 41 a (lectotype, desig- nated by Parodi (1944), LPS 2483!). Buenos Volume 89, Number 3 2002 Cialdella & Giussani 3 Phylogenetic Relationships of Piptochaetium Oryzopsis napostaensis Speg. var. — Speg., Ana- les Mus. Nac. Montevideo 4(2): 17. 1901. TYPE: Arge ntina. Бас Aires: Sierra de Мр Dec. 899 . Spegazzini 41 A (holotype, s 69!). Stipa — Hack., Anales Mus. Nac ie nos Aires 11: 95 1904. TYPE: presea е Dep. Rio Т, — Villamonte, Fa. — 13803 type. nol seen; Isosynlypes, ORD!, LPS!), aie 14082 (syntype, not located). Stipa ton Hack., Aires 13. 1911. TYPE; Pu Buenos ación он ЕС... M. Estrada s.n., kert 17397 Бе ada W not 3168626 not San Teodoro, Anales Mus. Nac. Bu Hab. Siue: isotype, US seen; seen). Distribution. Endemic to central Argentina in the provinces of Buenos Aires, Catamarca, Córdo- ba, La Pampa, Mendoza, Río Negro, and San Luis, up to 1900-2000 m elevation. This species grows in dry and arid grasslands and is a frequent species of the tree” (Prosopis caldenia Burkart), in the “caldenal.” an open wild forest of “calden “Espinal” biogeographic province. ‘or a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 25. Piptochaetium зки Mujica-Sallés & Longhi-Wagner, Candollea 48: 15. 1993. TYPE: Brazil. Santa Catarina: Urupema, junto Retransmisora del Morro de Campo Novo, 1680 m s.m., 25 Nov. 1984, J. F. M. Valls et al. 8083 (holotype, ICN!; isotype, CEN. la Estación Distribution. Known only from southern Brazil at the type locality. For a description and illustrations, see Cialdella and Arriaga (1998). 26. Piptochaetium panicoides (Lamarck) E Chil. 6: 270. 1853. Stipa . Genr. 1: 158. 1791. Speg., ‚ 1901. without locality, Desvaux, in Gay, Fl. panicoides Lam., Anales — Oryz — panicoides (Lam. Mus. Nac. Montevideo 4(2): 31 TYPE: Uruguay. Montevideo: 1767. Commerson s.n. (isotypes, P!, MPU not seen, US not seen). Hist. Nat. 12: 445. 1827. Oryz > — A. Rich., Dict. : (holo- : Uruguav. Montevideo: iba TSON S.H. type, probably P not seen). Piptoc — setifolium J. Presl, Reliq. Haenk. 1: 222. 1830. Stipa vil (J. Presl) — وا‎ PL 1: 182. br 833. Urachne simplex Trin. & Rupr. e = ruviana Trin. К ao m Pétersbourg sér 842. Ory ZODsIS setifolia (J. P a He — — ш rb. 40: 57. 1921. TYPE: Peru. Sin localidad, T. Haenke s.n. (holot . PR). Oryzopsis e arpa Speg., Anales Mus. Nac. Montevideo 42) 33. 1901. Piptochaetium leiocarpum (Speg.) Hack., Anales Mus. Nac. s Nat. Buenos Aires 13: 463. 1906. TYPE: Arge . Tucumán: пава, Lorentz & Hieronymus 608 ( (holotype. ‘ated; isotype, LPS!). — — — Phil., Anales Univ. Chile 1896. * : Chile. Río Itata, — s.n. D m SG * not seen; isotypes, SGO 57374!, US not een). iu — leiocarpum (Speg.) Hack. دغ a‏ ilé- not lo- f. Up want lack. ex Stuck., Anales Mus. Nac. Hist. . Bue- nos Aires 13: 463. 1906. Piptochaetium „л, (Lam.) E. Desv К. ex Stuck.) f. subpapillosum (Hac Parodi, Revista Mus. La Plata, Secc. Bot. 6(25): 302. 1944. Piptochaetium ралбай (Lam.) E. Desv. var. Miri AMAA (Hack. ex Stuck.) Petetin. FI. ónica, Colecc. Ci. Inst. Nac. Tecr : Argentina. without locality, Lillo 4314 (h ody pe not located; isotype, Herb. Arg. Stuckert 15418, CORD!). Stipa — Mez, Feddes Repert. Spec. Regni Veg. . 1921. TYPE: Argentina. Tucumán: La Cié- naga. — & Hieronymus 608 (holotype. B not seen). Oryzopsis pis arpa Speg. var. major Speg., Revista Argent. Bot. 1(1): 10. 1925. TYPE: n нв, Jujuy: Sierra de Santa Barbara, Sta. Spegazzini 2490 (осете: LPS 126411). еј Tucumán: Cornelia, Distribution. This species has a disjunct distri- bution within South America: an Andean area, in- cluding mountain regions of Venezuela, Colombia. and Peru (Brako & Zarucchi, 1993), widespread in Ecuador (Jørgensen € León-Yánez, 1999), Bolivia. — Chile, and Argentina, in the provinces of Córdoba. Jujuy, Salta, and Tucumán, and a coastal zone lim- ited to sandy soils of southern Brazil, Uruguay, and Argentina, in the provinces of Buenos Aires and Entre Ríos. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 27. Piptochaetium pilosum (Sánchez Vega) Cialdella & Giussani, stat. nov. Basionym: Pip- tochaetium tovarii Sánchez Vega subsp. pilosa Sánchez Vega, Arnaldoa 1(1): 29. 1991. TYPE: Peru. Cajamarca: entre Llacanora y Namora, 2750 m, Sánchez Vega et al. 2459 (holotype. CPUN!: isotypes, MO!, USM! The florets of this species have lemmas with prickles, hooks, and macrohairs all over the sur- face, and a callus that is densely pubescent. These characters clearly distinguish Piptochaetium pilo- sum from Р, tovarti, with florets without any epi- dermal appendage on the lemma and callus. These differences, together with the results of the cladistic analysis, make it reasonable to raise Piptochaetium tovarii subsp. pilosa to the rank of species. Distribution. Peru in the departments of Caja- marca and Junín. 328 Annals of the Missouri Botanical Garden For a description and illustrations, see Cialdella and Arriaga (1998 28. Piptochaetium pringlei (Beal) Parodi, Re- vista Mus. Plata, Secc. Bot. 6(25): 230. 1944. Oryzopsis pringlei Beal, Bot. Gaz. 15: 112. 1890. Stipa pringlei (Beal) Scribn., in Va- sey, Contr. U.S. Natl. Herb. 3(1): 54. 1892. TYPE: Mexico. Chihuahua: dry ledges, Sierra Madre, 5 Nov. 1887, 8500 ft., Pringle 1410 (holotype, US!; isotype, MO!). Figure 5D. Plants perennial. Culms 0.5—1 m tall, terete, gla- brous, pubescent below the nodes; nodes 2—3, red- dish, compressed or enlarged, glabrous or slightly pubescent; internodes 7-20 cm long. Leaf sheaths embracing the culms, shorter than the internodes, 6-18 ст long, scabrous, hispid to the base. Ligules 1-2.5 mm long, blunt or subacute, margins entire, hispid at the abaxial side. Leaf blades linear, flat, 10-30 cm long, 1-3 mm wide, glabrous, scabrous on the nerves and margins. Peduncles terete, slight- ly flattened, 30-50 cm long, longitudinally striate, hispid. Inflorescences 15-16 ст long, few-flowered (10 to 25 spikelets), with branches posed; pedicels flattened, up to 1 mm long, hispid. Spikelets fusiform, 9-12 mm long, 1.2-1.9 diam. Glumes subequal, longer than the floret, 9— loosely dis- mm 12 mm long, acuminate, glabrous; lower glume 5- or 7-nerved; upper glume 7-nerved. Floret terete- fusiform, 7-10 mm long, 1—1.5(-1.9) mm diam. Lemma contracted below the crown, 7-10 mm long, smooth, shiny, with yellowish macrohairs, decidu- ous, and few prickles toward the crown; crown con- tracted at the base of the awn, inconspicuous, straight, not revolute, 0.4-0.8 mm diam.. of long macrohairs 0.8-1 mm long and prickles; callus 0.8-1 mm long, subacute, obliquely trun- cate, hairy, the upper hairs slightly longer than the callus; awn bigeniculate 1-2 (—3.5) em long, persistent, hispid, scabrous toward with rows to almost. straight, the apex. Palea 6-7 mm long. Lodicules 2, « mm long, acute. Caryopsis not seen. Chromosome numbers. 2n = 42 (Myers, 1947, as Stipa pringlei (Beal) Scribn.); n = 21 (Reeder, 1977, as Stipa pringlei (Beal) Scribn.). Distribution. Southwestern United States and northern Mexico, up to 2000-2500 m elevation. Additional specime ns examined. U.S.A. Arizona: Co- arr Canyon, Huachuca Mountains, Williams, Hitchcock. 1903 near Tucson, So- Silveus 3470 — hcock 3 (US). E “Cube T- chise Co., 6 mi. up Gould elt des 2456 (B. АА, US); (US); Mt. Lemmon, Santa Catalina Mts., derstrom — (US). California: Santa Rita, (TEX). New Mexic :0: Ranger Station, Queen, 813 (В. АА); Guadalupe Mountains, Hitchcock 1 Texas: Black Mountains, Tharp 4175 (TEX, l son Co., S McKittrick Canyon, top of Guadalupe Mts., Warnock 12082 (TEX); Lincoln Co., Hinckl Jr. 629 (US). Chihuahua: Sierra Madre reverie Sierra Gazachic 35 km SW of Minaca, Pennell 18929 (US); Si- erra La Viga, 6 km E de Jame, Puerto Maravillas, Villa- rreal et al. 1975 (XAL). Sonora: Morelos, El Rancho de Robles, Vera Santos 1959 (MO, US); E of Cananea, Sierra de Los Ajos, Beetle M-7856 (MO). — Piptochaetium pringlei is slightly similar to P. fimbriatum in that both species share the smooth surface of the lemma, only with macrohairs and few hooks near the crown, Piptochaetium fimbriatum differs in having a shorter floret, 3.5—5 mm long, and a blunt callus. 29. Piptochaetium ruprechtianum К. Des- vaux, in Gay, Fl. Chil. 6: 274. 1853. Oryzopsis ruprechtiana (E. Desv.) Speg., Anales Mus. Nac. Montevideo 4(2): 12. 1901. Stipa ruprech- tiana (E. Desv.) Herter, Revista Sudamer. Bot. 6(5—6): 141. 1940. TYPE: Without lo- cality, Sellow s.n. (holotype, LE-Trin not seen; Brazil. isotype, US not seen; drawing of US isotype, Herb. Parodi 4127 sheet b, BAAN). Oryzopsis — (Vahl) Speg. var. major Speg., Anales Mus. Nac. Montevideo 4(2): 9. 1901. TYPE: tina. с. de Tandil, DAMM s.n. (holotype, LPS 2515!; isotype, BAA! Argen- Distribution, Southern Brazil in the states of Rio Grande do Sul and Santa Catarina, Uruguay in the departments of Lavalleja, Maldonado, and So- riano, and Argentina in the provinces of Misiones and Buenos Aires. This species is frequent in rocky prairies, and in Brazil it grows in savannas. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). 30. Piptochaetium sagasteguii Sánchez Vega, Arnaldoa 1(1): 1991. TYPE: marca: Cerro El Guitarrero, ladera occidental del valle de Cajamarca, 2800 m, 15 Jan. 1983, I. Sánchez Vega 2914 (holotype, CPUN!; iso- types, AAU not seen, CHAPA not seen, F not seen, HAO not seen, K not seen, MO!, US 3232467 not seen, USM not seen). Peru. Caja- Distribution. ments of Cajamarca and La Libertad at elevations between 2700 and 3500 m. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). Known from Peru in the depart- Volume 89, Number 3 2002 Cialdella & Giussani 329 Phylogenetic Relationships of Piptochaetium 31. Piptochaetium seleri (Pilger) Henrard, Blu- 1940. Oryzopsis seleri Pilg., Verh. 192. 1909. mea 3: 452 Bot. Vereins Prov. Brandenburg 51: TYPE: Guatemala. Huehuetenango: between Todos los Santos and Chiantla, 3000 m. Seler 3238 (holotype, В not seen; isotypes, BAAL! US 2767420 not seen). Figure 5F. Sep. ?, Plants perennial. Culms 20-90 em tall, terete, glabrous; nodes 2 or 3, reddish, compressed, gla- )10-15 ст sheaths embracing the culms, generally shorter than brous: internodes (4.5— long. the internodes, 8-9 em long, glabrous. Ligules 0.8— 2 mm long. blunt to subacute, membranous, mar- gins entire, glabrous. Leaf blades linear, convolute, 5.5-12 cm long. 0.4-0.5 mm wide, abaxial side scabrous on the nerves, adaxial side glabrous. Pe- duncles terete, 22-26 cm long, longitudinally stri- ate, glabrous. Inflorescences 7—22 cm long. few to many flowered (25 to 60 spikelets), with branches loosely disposed: pedicels angled, 0.4—1.5 ст long. glabrous. Spikelets fusiform, 4-5 mm long. 1-1.4 mm diam. Glumes subequal, longer than the floret. shortly acuminate, 5-nerved, viola- Floret 34.5 mm long. 1-1.2 mm diam. 1-5 mm long. ceous, hyaline toward the margin. terete, slightly obovoid. Lemma narrowed below the crown, 34.5 mm long. striate. with short macrohairs and hooks all over the surface: crown contracted at the base of the not revolute, 0.4—0.5 mm diam., with callus 0.4-0.6 mm hairy, awn, straight, prickles and short macrohairs; long. blunt, obliquely truncate, the hairs slightly longer than the callus; awn bigeniculate, 1-1.5 em long, persistent, shortly hispid. Palea ca. 3.5 mm long. Lodicules 2. 0.8-1 mm long. acute. Caryopsis terete, slightly obovoid, 3-3.2 mm long. 0.8—0.9 mm diam.: length of the carvopsis. hilum linear: embryo %4 the Widespread in Mexico and south- between 2000 and 3600 m. Distribution. ern Guatemala, Selected specimens examined. MEXICO. Coahuila: Cañón de San Lorenzo, Sierra de Zapalinamé, 3 km S of Saltillo, Snow et al. 6721 (MO). Distrito Federal: De- sierto de los Leones, Lyonnet 2719 (US); lava fields ca. 2 km SSW of La Cima, R.R. station on — side of old hwy. 95, 14 Aug. 1960, Ticis et al. 958 (US). Hidalgo: woodlands near Trinidad Iron Works. eee 13249 (MO, S) México: Sierra de Las Cru Pringle 5200 (МЕХ. US): 10 km al E de Amecameca. sol tera a Tlamacas. 27 km SW of Toluca on road to Temaxcaltepic. Roe 180 (US): 19202 (МЕХ): Llano Grande, near Río Río. Sharp 44144 (US): Chapingo. Ixtapaluca, faldas del Cerro El Papayo. Campo Experi- mental Zoquiapán, Zavala Ch. 245 (XAL); al SW del Cen- tro Experimental Zoquiapán, Velazco Torres & J. Trinidad s.n. (XAL); Telapon, Lyonnet & Elcoro 1948 (US): Parque Nacional Lagunas de Ze "mpoala, Traylor 61 (МЕХ, US). ces, re la carre- Leaf Laguna de San Gre- san Andre Beaman 4273. нени EXU). Santa Clara del Cobre, nit of Cerro S Michoacán: ` gorio, Escobedo 1476 (XAL): sumn km N of Ciudad Hidalgo. elos: Le mpoala. Lyonnet 2511 ( zi > ca. 12 1529 (US). Mor Oaxaca: 18 mi. SW of city of Oaxaca, Nelson 1373 US). Tlaxcala: ca. 3 mi. NE of Tlaxco, v Sols 604 (U 5), Veracruz: Acajete, Rzedowski 11940 б XU): Calca- hualco, end of passable portion of road from i Cos - pec-Esc ola-Jacal to Miguel Hidalgo and Tlac hichuca, Nee & Diggs 24775 (MO, XAL); La Joya, Mejra S. 218 (XAL) Las Minas, vereda Cruz dar anca a Rinconado. india E. & Burgos 498 (XAL): igas, pasar la presa del ij Fo rumbo a Te pes газ. Chazaro & — 3792 (XAL) Ventura A. 18506 (XAL): Sandoval. 68 AL); Perote. por la brecha que va a Tonalaco en las faldas del Cofre de Perote. Castillo et al. 1980 (XAL): Soledad Atzompa, Wer" gr * Acosta r 1063 (XAL); Xico. Tonalaquillo, Arriaga C. AL). GUATEMALA — huetenango: Sierra de vn a uc vri 8, . immediate ly of Tojiah at km 322, on ruta nacional 9 392 Ў (US); near Tunimá, Sierra de los Cue matan ss, mark 48274 (US); trail between Soloma and 5 Sierra de los Cuchumatanes, Steyermark 484, 5 WS). To- tonicapán: Desconsuelo, potrero nature flora alpina, de Koninck 114 (US); on the Tecum Uman Ridge at km 154 ca. 20 km E of Totonicapán, Bea- km despues de Manzanares, Las Perote. Escobillo. “>< — . Beaman on ruta nacional nro. 1. man 4155 (US). Beetle (1977) and Hitchcock (1951. under Ory- zopsis seleri) placed Piptochaetium seleri in synon- ymy with P. fimbriatum. Both species have terete. slightly obovoid florets and lemma wilh macrohairs. but P. fimbriatum has smooth or thinly striate lem- mas, without prickles or hooks, and a crown without epidermal appendages. 32. Piptochaetium setosum (Trinitus) Arecha- valeta, Anales Mus. Nac. Montevideo 1: 330. 1896. Urachne setosa Trin., Mém. Acad. Imp. Sci. St. Pétersbourg sér. 6, Sci. Nat. 6(1): 124. 1834, nom illeg. superfl. TYPE: Chile. With- out locality, Cumming s.n.. TRIN 1473.1 (lec- totype. designated by Parodi (1944). LE not seen: Isotype, US not seen). е — dye ex Steud., Syn. Pl. Glum l: 123. . TYPE: C E pt te — 153 (iso- Iv pe. o iare i purpuratum Phil., Linnaea 29: 1857 TY Chile. Valparaíso, Nov. 18547 — ^» lotvpe. SGL 574191). Piptoc haetium — ex Griseb., Symb. Fl. Ar- зат TYPE: Chile. Cerro San С ristóbal, "a اس‎ re 8 Philipp s.n. designated here, SGO Anales Univ. Chile 93: Pi iptoc Jn macrocarpum Phil.. 896. TYP : А | V. Briones s.n. (ho- (35. ). ше. Nuble, lotype. SGO 45082!). Distribution. This species presents, at the mo- ment, a disjunct distribution in southern North America (California) and South America (Bark- 330 Annals of the Missouri Botanical Garden worth, 1986, 1993), where it grows in central Chile from Valparafso to Valdivia. Observation. This species has a strikingly pros- trate habit in some specimens collected in Califor- nia (Barkworth, pers. comm.). r a description, illustrations, and specimens 199 -rj ‘or examined, see Cialdella and Arriaga Note. Piptochaetium setosum (Trin.) Arechav. was based on Urachne setosa Trin., a nomenclatur- ally superfluous name when published, as Trinius cited Stipa panicoides Lam. and Oryzopsis setacea A. Rich. as synonyms, both earlier published. Nev- ertheless, Р, setosum is the correct name in Pipto- chaetium for Urachne setosa when Stipa panicoides and Oryzopsis setacea are considered synonyms of — another species, Piptochaetium panicoides (Lam. E. Desv 33. Piptochaetium stipoides (Trinius & Ru- precht) Hackel, Anales Mus. Nac. Montevideo 1(4): 328. 1896. 33a. Piptochaetium stipoides (Trinius & Ru- precht) Hackel var. stipoides. Urachne — Mém. Acad. Imp. 8 Pétersbourg, sér. 6, Sci. Nat. 5: 25. 1842. ra zopsis stipoides (Trin. & Rupr.) Speg., Anales Mus. Nac. Montevideo 4(2): 23. 1901. Pipto- chaetium stipoides (Trin. & Rupr.) Hack. var. genuinum Parodi, Revista Mus. de La Plata, des Trin. € Rupr., есе. Bot. 6(25) 266. 1944, nom. inval. TYPE: Brazil. Brasilia meridionalis, Sellow n. (holotype, LE not seen). Symb. Fl. Argent.: Piptoc 'haetium oe Griseb., 298. 1879. Piptochaetium ov — Desv. var. chae- Anales Mus. Nac. Hist. е 1904. TYPE: Argentina. Córdoba: Dpto. Punilla, in de Azúcar, Hieronymus s.n., Herb. Grisebach m — GOET not seen; isotype, CORD!). „шш — — Phil., Anales Univ. Chile 93: ‚ 1896. TYPE: Chile. e. de Conchalí, Philippi s.n. s holotype . SGO 57401!; isotype, US 1819510 not о (Grise ij Hac n). Piptoc — ovatum E. Desv. var. purpurascens Hack., Anales Mus. Hist. Nat. Aires 21: 86. 1911. Piptoc — — Trin. & Rupr.) Hac o var, n ens (Hack.) Parodi, Revista Mus. de Plata, Secc. 6(25): 272. 1944. TYPE: Arge — Chaco: — 1 de Mayo, Margarita — ‘n, Stuckert 19189 (lectotype, here ds 'signated, ( Piptoc — ovatum E. Desv. f. atrata Hu Anales Mus. Nac. Hist. Nat. p поз Aires 21: 85. 1911. ina. Chaco: Dpto. | de Mayo, Colonia 18304 (holotype, W not seen; Buenos TY РЕ: Argent Benitez, Stuckert type, CORD!). Stipa verruculosa Mez, Feddes Repert. Spec. Nov. Regni leg. 17: 206. 1921. Piptoc — — ulosum (Mez) Henrard, Meded. Rijks-Herb. 380. 1927. Piptochaetium verruculosum (Mez) i. Revista iso- Bot. 6(5-6): 141. 1940, comb. superfi. Argentina. Buenos Aires: without locality, Balansa s.n. (holotype, B not seen; isotype, US not seen Sudamer. TYPE: Piptochaetium — Gpeg. ) Herter, Revista Sudamer. * e ). Oryzopsis E Speg.. us монеті 4(2): 4. . TYPE: mentia: E ntre e Río : Concepción del * lay, Lo- rentz s.n., a Entrerriana 1691 — С ORD!; isotype, | Pst), Pichi stipoides (Trin. & Rupr.) Hack. var. parvi- florum Parodi, Revista Mus. de La Plata, Secc. Bot. 6( 75. 1944. TYPE: Argentina. Buenos Aires: Pdo. Balcarce, Balcarce, Martinez Crovetto 950 1/2 (holotype, BAA!). Distribution. This variety presents, at the mo- ment, a disjunct distribution (Thomasson, 1980; Barkworth, 1986, 1993) in the southwestern United States and Mexico (Beetle, 1977), and in South America, where it grows in southern Brazil in the states of Rio Grande do Sul and Santa Catarina. It is also widespread in Uruguay, in Canelones, Co- lonia, Flores, Florida, Lavalleja, Montevideo, Pay- sandú, Río Negro, Salto, Soriano, and Tacuarembó; in Chile, in УШ and IX Regions and in Región Metropolitana; and in Argentina in the provinces of Buenos Aires, Catamarca, Chaco, Córdoba, Corri- entes, Entre Ríos, Jujuy, Misiones, Río Negro, and Santa Fe. This variety is known up to elevations of 2000 m in arid regions. ARGENTINA. Cha- Additional pm — | Nov co: Margarita én, оу J08, Stuckert 19208 (CORD). MEXIC О. Tamaulipas: 5i de San — Armadillos, 10 July 1930. Bartlett. 1022: Cerro de los — For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 33b. Piptochaetium stipoides (Trinius & Ru- precht) Hackel var. echinulatum Parodi, Re- vista Mus. La Plata, Bot. 6(25): 271. TYPE: Argentina. Buenos Aires: Pdo. Villa Parodi 13759 (holotype, Secc. Puán, Iris, ВАА!). The florets of this variety present the lemma with prickles partially covered by adjacent n s Cialdella & Arriaga, 1998: — cells at the base 3D, 10J) Distribution. central Argentina in the provinces of Buenos Aires, La Pampa, and San Luis, and in northern Uruguay This variety has been found in the department of Salto. For a description, illustrations, and specimens examined, see Cialdella and Arriaga (1998). Volume 89, Number 3 2002 Cialdella & Giussani 331 Phylogenetic Relationships of Piptochaetium 34. Piptochaetium tovarii Sánchez Vega, Ar- naldoa 1(1): 25. 1991. TYPE: Peru. Cajamar- ca: Cajamarca, entre Cajamarca y Cumbema- yo, Fundo Universidad, 3450 m, 22 May 1971, I. Sánchez Vega & N. Vilhena 678 (holotype. CPUN!: isotypes, MO!, US 3232468 not seen). Distribution. Southern Ecuador in the limited region between provinces of Loja and Azuay, and northwestern Peru in the departments of Piura, Ca- jamarca, and Ancash, at elevations between 2700 and 3450 m. ECUADOR. Azuay- Loja: Nudo de cordillera occidental y cordillera oriental, entre Ofia y Rane 5 Ovejero, 1-2 Aug. 1959, Barclay & Juajibioy 8455 ( Additional specimens examined. For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 35. Piptochaetium uruguense | Grisebach. Symb. Fl. Argent.: 297. guayensis (Griseb.) Speg., Anales Mus. Nac. Montevideo 4(2): 29. 1901. Piptochaetium uru- var. genuinum Parodi, Revista 1879. Oryzopsis uru- guense Griseb. Mus. La Plata. Sece. Bot. 6(25): 290. 1944. nom. inval. TYPE: Argentina. Entre Ríos: Dpto. Uruguay, Concepción del Uruguay. 13 Nov. 1875. P. G. Lorentz 471 (holotype, prob- GOET not. seen; BA! ably in isotypes, CORD!). — —— ш ксн var. microcarpum Parodi, evista Mus. de ita, Secc. Bot. 0(25): : 290. ۳ М. TY "we ntina. godes ntes: Dpto. San Mar- tín. s ú, Parodi 12637 (holotype, BAA!). = = Distribution and ecology. For the time being. this species presents a disjunct distribution in Mex- 1980; Barkworth, 1986) and South America, in Paraguay, departments of Misio- ico. (Thomasson, nes and Guairá, in western Uruguay, departments of Cerro Largo, Flores, Florida. Paysandú, and Ro- cha, and in northern and northeastern Argentina, provinces of Chaco, Corrientes, Entre Rios, Misio- Santa Fe. Piptochaetium uruguense nes, Salta. and * also grows in southern Brazil, state of Rio Grande do Sul, and it has been cited for (Smith et al., ' ries. Santa Catarina 1982). This species grows in wet prai- Additional examined. BRAZIL. Rio Grande do Sul: Fazenda E ns rimental de Criacao Bage. | Dec. 1945, Swallen 7561 (MO) spec ime n For a description, illustrations, and other speci- mens examined, see Cialdella and Arriaga (1998). 36. Piptochaetium virescens (Kunth) Parodi, Revista Mus. La Plata, Secce. Bot. 6(25): 230. 1944. Stipa virescens Kunth, Nov. Gen. Sp. 1: 126. 1816. TYPE: Mexico. Guanajuato, Santa Rosa, Cuesta de Belgrado y La Buffa, Hum- boldt & Bonpland s.n. (holotype, P not seen: isotype, US 00141712 not seen). Figure 5C. — Repert. Spec. Nov. Regni Veg. 8: 515 (Kunth) Parodi var. Stipa arsenii Hack., 1910. Piptochaetium virescens arsenii (Hack.) Beetle, Phytologia 54: 4. 1983 TYPE: Mexico. Michoacán: Morelia, 2200 m. Fr. Ar- sene 3211 (holotype. W not seen: US nol MO 845899 not seen). isotypes. seen, Plants perennial. Culms 0.70—1.30 m tall. terete, glabrous, occasionally hispidous below the nodes: nodes 3 or 4, brownish, compressed, glabrous: in- ternodes 12—45 the culms, shorter em long. Leaf sheaths embracing than the internodes, 9.5-15 cm long. glabrous. Ligules 1-2.5 mm long. blunt, mar- gins entire, glabrous, sometimes hispidous at the 10-35 cm long, 0.5-1 mm wide, longitudinally striate, gla- abaxial side. Leaf blades linear, convolute. brous, scabrous at Ше margins. Peduncles terete, 40-44 em long, longitudinally ribbed, glabrous. /n- florescences 10-30 cm long. many-flowered (25 to 80 spikelets), with branches loosely disposed: ped- icels slightly angled, 2-18 mm long. hispid. Spike- 6-8 mm long. 1.2-1.4 Glumes subequal, longer than the floret, 6-8.5 mm lets. fusiform, mm diam. hyaline toward &(—9) mm long. | mm diam. Lemma slightly contracted below 5.5-8 hooks and macrohairs uniformly disposed: crown long, shortly acuminate, 5-nerved, the margins. Floret К ed 5.5— the crown, mm long. thinly striate, with contracted at the base of the awn, straight, not rev- olute, 0.5-0.7 mm diam., with prickles and rows of fa) macrohairs 0.5-0.6 mm long. occasionally 1 mm long; callus 0,9—1 mm long, acute. occasionally subacute, densely hairy, the upper macrohairs slightly longer than the callus; awn bigeniculate. 13-20 mm long, persistent, hispid, scabrous toward the apex. Palea ca. 5 mm long. Lodicules 2. 1.4— 3—3.2 mm embryo Y, the 1.6 mm long. acute. Caryopsts terete. long. 0.9 mm diam.; hilum linear: length of the caryopsis. Chromosome numbers. 2n = ca. 60 (Gould, 1966, as Stipa virescens). Distribution. Guatemala and central and south- ern Mexico. This species was cited also for Vene- zuela, Barinas. in TROPICOS based on J. Brunken & C. Perino 355 (MO), which we did not see. GUATEMALA. Hue- huetenango: Cumbre Papal, between summit and La Li- bertad, Steyermark 50955 (US). Quezaltenango: La Es- Selected specimens examined. Annals of the Missouri Botanical Garden peranza, De Koninck 73 A and B (US). San Marcos: Barrancos 6 mi. S and W of town of Tajumulco, NW slopes of Volcán Le жеты Steyermark 36700 (US). MEXICO. Chiapas: 5 exican hwy. 190 near Rancho Nuevo, about 9 mi. of San Cristóbal las Casas, Bre a 14171 (US). Distrito dense) е de Mexico, С. de Ajusco, Matuda 2575: XU): S. de Ajusco, Pringle 6588 (MEXU, MO, US) ен У Serranfa de Ajusco, Lyonnet 1877 (MO); 4 km W of San Andreas, Pedregal de San Angel, Sohns 177 (US). sa Tepea- pulco, Cerro de Xihuingo, Ventura . А. 376 (MO); Tulancin- go, open pine woods near reservoir at — between Acaxochitlan and Puebla, Moore Je 2843 (US); Real de Monte, Matuda 18907 (US). Jalise — km SW of Tequila on Volcán de Tequila, Век оге 39241 (MO); Za- potlán, Hitchcock 7169 (US); Sierra del Tigre. 5 mi. S of Colim: de Mexico, D nac ho. Amecameca, Matuda 25724 (ME т О); Ix- aluca, 8 km S de Río Frío, Koch 762 (MO); Popo Park, pr — 5965 (US); Cerro de Pinal, Otzoloapan, Matuda el al. 31875 (US). Michoacán: in open pine forest anc in and around the NE side of the Volcán de Р aric utin, Sohns 819 (MEXU); ca. 18 mi. S of Patzcuaron, ing & Soderstrom 5151 (MEXU): Las Cañas, Rzedowski «€ McVaugh 623 (MO); Las Cañas, estribaciones inferiores noroccidentales del Cerro Patamban, Tangancicuaro, Rze- dowski & McVaugh 623 (US). Morelos: Sierra de Ajus Pringle 62. * و‎ XU, MO, US); Lagunas оен ж пе! 2498 . Oaxaca: 23 km de Tlaxiaco, rumbo a Chalcotongo, — M-4743 (MO); 0.5 km : de Las Huer- tas, Nundichi, Manzanero M. 450 (M l En al S de San Andrés, carr. Oaxac "a Puerto m huatlan, Torres & Cedillo 875 (MO). М de Tlaxco, al S de Cerro La Paila, Chimal et al. 9 (MO); 5 mi. NE of Zac 'atepec, on hwy. 40 from Jalapa, Soderstrom pass, 46 mi. n open areas Tlaxiaco, . to Mexico City, (US). San Luis Potosi: on the NE slopes of hills near Aguaje de García in the Sierra de Guadalcazar, 1500 (US). Tlaxcala: Huamantla, 500 m al SW tamirano Guadalupe, Guerrero et al. 632 (MO); 3 mi. NE of Tlaxco, Sohns 601, Sohns bis IS). Veracruz: Aya- XN 228 (MO). Zacatecas: collected near Plateado, Rose 2750 (US). Sohns Piptochaetium virescens is similar to P. indutum: they share the floret and callus shape and the epi- dermal appendages on the lemma. Piptochaetium indutum can be recognized because it includes smaller plants, 15-35 em tall, with smaller florets, (4—)5-6.25 glumes, (7-)8-13.5 mm. mm long and 1 mm diam., with longer Piptochaetium virescens and P. sagasteguii are also related species, with similar floret characters: P. sagasteguii may be recognized because it in- cludes smaller plants, 25-65 em tall, with few-flow- ered inflorescences of 15 to 22 spikelets. The geo- graphical distribution of these species is also different: P. sagasteguii is only known for Peru, while P. virescens can be found in Mexico and Gua- temala. EXCLUDED SPECIES FROM PIPTOCHAETIUM Piptochaetium mexicanum (Hitche.) Beetle, Phytologia 54: 4. 1983 = Nassella mexicana (Hitche.) Pohl, Taxon 39: 611. 1990. [Basionym: — mexicana Hitche Contr. U.S. Natl. Herb. 24: 247. 1925,] NOMINA NUDA Е avenaceus Raf. ех B. D. Jacks., Index Kew. E 894, nom. inval., as syn. of Stipa avenacea E ochaetium avenaceum (L.) Parodi. Е dan Raf. ex B. D ). Jacks., ui 894, nom. vg .„„ as syn. of Stipa avenacea L. iptochaetium avenaceum n (L .) Parodi. Stipa sia Willd. ex Steud., Nomencl. Bot. (ed. 2) 2: 341, nom. inval., as syn. of Stipa avenacea Stipa if Walter (1788) = Piptochaetium avenaceum (L.) 1 li Avena — Willd. ex Ste ii 146. 1841, nom. inval., "MH = Hitche — panicoides Trin., in Martius, Fl. id Enum. pl. 2(1): 376. 1829, Mém. Acad. ms Sci. St. Péters- bourg s sér. а Sci , pro syn. nom. nud. — Piptoc haetium montev idense (Spre ng.) Paro- di. Index Kew. 2 a rodi, Bot. (ed. 2): 2: Milium mexicanum Piptoc hun fimbriatum (Kunth) Nomencl. syn. of i Agrost.: 173. 1812, Piptatherum elegans P. Beauv., ud. P — a DE E. от. n ) esv. Ыш гит panicoides | Gilbert, Enum. Pl. Monte 17. 873, nom. nud. = Piptoc ает pu s а ) Е Desv. ТУ — Gilbert, Enum. Plant. Montev.: 117. 73. nud. — Piptochaetium stipoides (Trin. & n ) var. — Literature Cited Barkworth, 1986. — и Presl (Gramineae, Stipeae) in — and Mesoameri Sans and dis- tributional observations. Brenesia 25- 26: 169-178. 1988. New taxa in Piptoc — Sipe ae, Gra- pa from Mexico. Syst. Bot. 13: 196-201. ——— 1990. Nassella (Gramineae, рене) Revised in- terpretation and nomenclatural changes. Taxon 39: 7 . 1993. North American Stipeae (Gramineae): ' onomic c line 's and other comments. Phytologia 74: 25. ax- Beetle, A. А. pns 37: pps Brako, 1 ла 1993. Catalogue of the Flow- ering — and Gymnosperms of Peru/Catálogo de las eg siis y pn Yospermas del Perú. Monogr. Syst. Missouri ua wd. 45. Bremer, К. 198€ The limits of amino acid sequence data in — rm phylogenetic reconstruction. Evolution 42: — grasses from Mexico V. -- EC Brane h support and tree stability. Cladis- tics 10: 295- Brown, №. V. ps ¢ hromosome numbers of some Texas grasses. Bull. Bot. Club 78: 292-299, Cialdella, A. M eee 1998. Revisión de las es- pecies — ricanas del género Piptoc haetium (Poa- ceae, Pooideae, Stipeae). Darwiniana 36: 107-15 Clayton, W. D. 1983. 212 Pipto- Proposal to conserve Volume 89, Number 3 2002 Cialdella & Giussani Phylogenetic Relationships of Piptochaetium chaetium К. B. Presl over Podopogon Rafi. (Grami- ‘axon 32: 649. 1853. ун In Ea fisic ч у Ж a de Chile 6: 233— Elias, M. K. 1932. Grasses = other — from the Ter- tiary rocks of Kansas and Colorado. Univ. Kansas Sci. Bull. 20: — 1942. ‘Tertiary Prairie Grasses and Other Herbs from ‘he Bi Plains. Special Pap. Geol. Soc. Ame M. Ne а Geological Survey, The University of Ne- neae). 1 Desvaux, E. Gay (editor), Historia . Lincoln. Ellis. R. P 1979. A procedure for и compar- alive leaf anatomy in the Poaceae. in epidermis as seen in surface view. ое 12: Bl. 71. Goloboff. Р. A. 1993 `2 puter prn a al [Distributed by the author. | ONA version nit). Com- 93b. Estimating character weights during tree search. gs — 9: 83-9] 1997. Pee-Wee version 2.8 (32 bit). жараш — ме manual. [Distribute d y the author Gould. )58. C hromosome numbers in lua "n grasses. н J. Bot. 45: 757-768. 1966. С hromosome numbers of some Mexican grasses. — Le ч 4: 1683-1006 Hite hcock. A. ‚ New species "d grasses from South America. da roc. [^ Soc. Wash. 36: 195-198 . 19 — of the cal of ih Unite di States, ed. in A. Chase. Misc. Publ. U.S. Dept. Agric. — Holmgren, P. K.. N. H. Holmgren & L. C. Barnett. 1990. ndex Herbariorum. Part 1: The Herbaria of the World. cara Veg. 120. Jacobs W. L., J. Everett € M. Barkworth. 1995 C erui of morphological terms used in the Sti- peae (Gramine * and a reassessment of Nassella ii 11. Australia. Taxon 44: 33— Jorgensen, P. M. & León-Yánez (editors). 1999. Cata- logue of the Vascular Plants of Ecuador. Monogr. Syst. Bot. Missouri Bot. Gard. 75. Metcalfe. C. R. 1960. Anatomy of the Monocotyledons. I. Gramineae. Clarendon Press, Oxfor Munoz-Schick. M. 1990. Revisión del género Nassella — К. De sv. (Gramineae) en Chile. Gayana. Bot. 47: ue TS, E M. 1947. € — and genetics of forage grass- s. 319-422, Parodi, L. R. . Revisión de americanas de fa género Piptochaetium. Revista Mus. La 10 c las gramíneas australes Plata, sece. Bot. 6: 213— 1961. Gramineas argentinas nuevas o críticas. . Rev с Argent. Agron, 28(3—1): 100— Репай е, . P. 1996. Anatherostipa. un nuevo género de Poaceae. irene Bot. 53: 277-28 Presl, C. B. 18: Prague I: x y ээ. Reeder, J. R. 1968. Notes on Mexican grasses VIH. Mis- cellaneous Chromosome numbers—2. Bull. Club 95: 6 92 30. 1977. € ОЛЕГ numbers in western grasses. Amien J. Bol. 04: M )2 Sánchez Vega, |. 19 del género Piptoc — J. 8. [= V ). Reliquiae Haenkeanae. J. C. Calve. Torrey Bot. S visión de las espec ies peruanas Presl (Gramineae). — i . Wasshausen & R. M. Klein. 1982. Gê meros: 45. Deschampsia até 84. Pseude- In R. Reitz (editor). Flora Hustrada Cata- rinense | (Gram.): 443-906 Gramineas. chinolaena. Spegazzini, C. 1901. em Platenses. Anales Mus. Nac. Montevideo 4: 1-173 Swallen, J. R. 1955. Flora af Guatemala 2: Grasses of и . Fieldiana Bot. 24: 1-390 Thomasson. + R. 19 = 78a. E its К patterns of the lemma in some fossil and living grasses and their phylogenetic каа ance. Science 199: 975-977. 1978b. Te rs fossil plants — in Nebraska. Res. Rep. Natl. Geogr. Soc. 1978 Projects: 553-501 280). — in — — (Gramineae Stipeae): Paleophytogeograph Iselva ignificance. А ‚ Fossil grass anthoecia and other plant fos- sils from arthropod burrows in the Miocene of western ) silo adaptation in ы live ла, ts (er mma Ар A аў. Апп. Missou- ri Bot. Gard. 72: 843-851. Torres, M. A. 1969. Nota sobre una i de Pipto- alatum de Argentina у Uruguay. Bol. Soc. Argent. Bot. 11: 251-254. 997. Nassella (Gramineae) del noroeste 1 — nlina, Stipa (Gramineae) del noroeste de la "ed ne tina, América del Sur. Vicoraella —— un nuevo género para : Ministerio de la produc ción v el empleo. provinc ia ie Buenos Aires, Comisión de ves- tigaciones Científicas (С IC) — 13 16. Valencia. J. 1. & M. Costas. 1968. E — citotaxonó- micos wipe — him a ae ). Bol. Soc. Ar- gent. Bot : 168-179 APPENDIX ] Examined material of doe — Each specimen ts cited by the last name of the f more than one collector. Species number is indicated be- st collector when there is tween parentheses. 1. Piptoc ман alpinum 1. B. wa 2. Piptochaetium angolense Philip} 3. Piptoc adum angustifolium ( Mite heock) Valencia & Costas V. Piptochaetium avenaceum (L.) Parodi 5. Piptochaetium avenacioides (Nash) Valencia & Cos- tas Piptochaetium bicolor (Vahl) К. Desvaux 7. Piptochaetium se ы — — 8. Piptochaetium brevicalyx (E. Fournier) Ric 9. Piptochaetium burkartianum Parodi 10. Piptochaetium cabrerae Parodi Piptochaetium calvescens Parodi 12. Piptochaetium confi i 13. Piptochaetium cuc ара Rosengurtt & Izaguirre de Arturio 14. Piptochaetium featherstonet (Hitchcock) Tovar 15. ан (Kunth) Hitchcock a nd hackelii (Arec havale 1а) Parodi эз т 4 N N =] pes "E 24. Piptochaetium napostaense (Spe 25. Piptoc haetium palustre Mujica-Sallés & agner ». Piptochaetium panicoides (Lamarck) E. Desvaux Longhi- 334 Annals of the Missouri Botanical Garden 7. Piptoc haetium pilosum (Sánchez Vega) Cialdella & Giussani Piptoc haetium pringlei (Beal) Parodi ruprechtianum E. Desvaux N 28. 29. 30. 31. 32. 33. 33a. Piptochaetium seleri (Pilger) Henrard Piptochaetium setosum (Trinius) Arechavaleta Piptochaetium stipoides (Trinius & Ruprecht) Hackel Piptochaetium stipoides (Trinius & Ruprecht) Hackel var. stipoides Piptochaetium stipoides (Trinius & Ruprecht) Hackel var. echinulatum Parodi Piptochaetium tovarii Sánchez Vega 35. Piptochaetium uruguense Grisebach 36. Piptochaetium virescens (Kunth) Parodi Aleman 267 (26); Alonso J. 151 (20); — ows s.n. (4): Arriaga C. 351 (31); Arzivenco 349 (20), 352 (28). 5 (5), 1155 (5); Banda 132 (8); Barclay 8455 (34); Barkley E A 660 (15); Bartlett 10225 (33a); Bartley 2011 (4); Beaman 3928 (31), 4155 (31), 4273 (31), 4329 ‚ 4387 (36); Beetle M-141 (36), M-4743 (36), M-7086 . M-7338 (15), M-7742 (15), M-7856 (28), М-7864 ; Berkley 1579 (4); Biswell 94408 (4); Boechat 160 , 161 (20); Bonner 354 (4); Boott s.n. (4); Boufford 182: 12 (4); Bozeman 8842 (4); ipid — 31 (15), 34 (15); Brant 2239 (15); Braun 1422 ( J42 (4): Breedlove 14171 (36), 39241 (36), 43293 a 44133 (15); Bro. Ar- sene 2699 (36), 5305 (36), 5375 (36), 6749 (36); Brunken 2 м (15), 247 (36), 387 — gis ss 221 (4). Calzada 4645 (15); Carter 2396 (15); Castillo ae (31); Chapman 1037 (4); Chase 7220 (4), 7312 (4), 7 (4); Chazaro 3792 (31); С — 9077 (15); С DURAM 9 B "ae Churchill 123 (4), s.n. (4); Cole s.n. ( .n. (4); Combo 876 (4); vens 2069 (4); Correll 353 (4), 1 3699 (15), 16170 (4); € ronquist 4311 (4), 4432 (4). 5014 (4), 5075 (4); Curtiss 5834 (5 Darrow 1287 (15); Davidse 9287 a 5), 9485 (15), 9832 (36), 10031 (15), 11137 A (33a), 29775 (15); De Chalmot s.n. (4); De Koninck 73 (36), 114 (31); Deam 20709 (4), 4), 40704. (4), 64413 (4); Demaree 14586 (4). 19003 (4). 63483 (4); Dowell 7309 (5); Duncan 9485 (4); 31). 33b. 34. = Earle 165 (15); Early s.n. (4); Eggert 36 (4), s.n. (4); Egler 41-136 (4); Ellis 1113 (31); Escobedo 1478 (31). Faircloth 5079 (4), 5798 (4), 7070 (4); Felger 92-794 (15); Fernald 7760 (4); Fernández 734 (28), 735 (20); Fer- ris 2736 (15), 2812 (15); Fisher 108 (31), s.n. (31). s.n. (4); Fogg Jr. 2224 (4); Fredholm 5092 (5). 5725 (5); Free- man 54117 (4). Garber s.n. (4); Gentry 8492 (15); Gleason 8537 (4): Godfrey y 3686 (4), 8085 (4), 67795 (4); Gould 2456 (28). 4419 (15), 9211 (8), 9602 (15). 9661 (15). ‚1 5409 (4); Griffiths 94 (15), 107 (28), 4836 , 7032 (15); Griscom 16409 (4); Guerrero 2 (36): € ;utte 185 (36) p^ 164 (4); Harkins 809 (4); Harvey 1038 (3), 1458 (15), 1764 (4), 1854 (4); Havard 26 (15), s.n. (15); Hen- derson 91-03 (4), 94-29 (4); Henrickson 15143 a (15), 15226 (15); Hermann 3201 (4), 4198 (4); Hernandez Xo- locotzi X-2443 (15), X-2828 (36), X-2836 (36); Hill 14632 (15); Hinckley 245 (28), 493 (28); Hinton 1336 (36), 1893 (36), 2315 (36), 2316 (36); Hitchcock 520 (4), 812 (5), 813 (28), 837 (15), 967 (5). 1271 (4). 1318 (4). 1341 (4), 1346 (4), 1903 (28), 3715 (15), MES (15), 5947 (36), 5965 (36), 5973 (36), 6019 (36), 6266 (36), 6733 (8). 7151 (36), 7169 (36), 7677 (28), 7681 (28), 7741 (15), 12616 (4), 13511 (15), 13553 ү“ s.n. (4); Hoffmann s.n. 4); Hoge 260 (15); Hoogstraal 991 (36); House 2101 (4). Iltis 213 (15), 214 (8), 958 О, 23032 (4); Innes 1130 D (15). Jacob 815 (4); Jones 479 (36), 24157 (4). Kearney 10114 (15), 10541 (28); Kearney Jr. 1102 (4); Kellogg 82 (4); King 3378 (31), S 32 (36), 5151 (36); Kneucker 564 (4); Knieskern s.n. (4); Koch 762 (36), 71 10 (1: 5): Kral 2065 (4). 25586 (15), 34987 45934 (4), 46012 (4), 46479 1). 55030 (4); Krauc h 6495 Larrick 249 (4); Latham 2755 a. 4087 (4), 22391 (4); Le Doux 2086 (31); Le Sueur 143 (15); Leavenworth 1077 36); Lemmon 2923 (15), 4678 (2 8); 1 — 11 м (26); ecc pa 2121 (1), 2497 (1); Lundell 13153 (15); Luz 39 (26), Lyonnet 48 (15), 62 a (36), 1826 (36). 1877 (36), 1948 (31), 2498 (36), 2511 (31), 2512 (31), 2527 (36), 2684 (36), 2719 (31). Mackenzie 3109 (4); Madrigal A. (15); — — TO M. 450 (36); Marsh Jr. 629 (28); — 1063 (31); thews s.n. (4); Matuda 18907 (36), 19202 (31), 24 (36). 25753 (36). 25755 (36), (36), 29473 (36), 29776 (3), 31875 (36); McCarthy (4); McDougall Ws (4); Me- Ge p 57 (15), · (15); MeVaugh 12268 (36); Mearns 2573 (28): Nu i (5); Mejia S. 218 (31); Mendoza ses | 5); Metcalfe 746 (15); Metzler 2 (36); Mez 3164. (15). 3166 (15); Mick 87 (8), 180 (31). о үе Y — f B Miller Jr. s.n. (4); Mohr s.n. (4); Moo 72 (15), : (15). 4164 (15); Moore Jr. 1233 (31). bus : 36). pene A en Moorgen s.n. (29); Morong s.n. (4); Mueller 2342 (15), 2442 (15), 7948 a (15); Mujica Sallés 2 (20), 22 (28), 42 (26), 46 (26), 47 (26), 63 (33 a), 72 (26), 74 A (26), 81 (20), 83 (29), 130 (35). 141 (22). 142 (22), 144 (35), 151 (20). 155 (22), 166 (20). Nealley s.n. (4); Neally 129 (15); Nee 22928 (36), 24775 (31): Nelson 1373 (36), 1373a (31): Noyes s.n. (4). O'Neill 491 (4); Ortiz 1132 (36), 1194 (36). Palmer 339 (15), 5154 (4), 5248 (4), 24541 (4), 31887 (15); Pedersen 5188 (35), 13333 (24); Peebles 3389 (15); Pennell 17704 (15), 8929 (28), 18940 (15): Peterson 4010 — T 25777 — (15); Piedmont 4311 (4); Plank 53 (4); Pringle 18 (28), 502 (15), 639 (15), 3035 (15), 4759 (36), 5200 (31), 6236 (36). 6588 (36). 7606 (8), 8595 (15). 9574 (36). 13249 (31), s.n. (15); — 3589 (15). Quarín 3163 (33a). Redfield 4317 (4); Reeder 233 (4), 2221 (31), 3295 (15); Renvoize 3025 (20). 4143 (23); Reverchon 4128 (4). s.n. (4); Rivero 1384 (23); Robinson 1 (4); Rodriguez 1284 (31); Roivainen 481 (11); Rolfe 43 (4); Rose 2750 (36), 5611 (86), Soon 8), 8700 (36), 8744 (15); ), 11940 (31). 16904 (15), 17547 5) 17704 15) : 35555 (31), 35863 (8), 50768 (8). Sánchez Vega 2458 (34), 2769 (34), 2935 (30), 4984 (30); Sánchez-Ken 141 (15); Sandoval 68 (31); Schaack 2881 (4); Schinini 19229 (35); Schultz 41 (4), 14705 (23); Scribner 138 (4), 266 (4); Scully 652 (4); Serrato Sánchez s.n. (15); Seymour 109 (4), 1060 (4); Sharp 44144 (31), s.n. (4); Shinners 7029 (4); Shreve 7707 И 5); Silveus 727 (15), 785 (15), 3470 (28); Simpson 1); Smith 926 (36), 3327 (18), 8302 (14), 10109 B (18), 10331 (26), 10488 (18), 10960 (34), 10964 (18), 11038 (18), 11794 A (18); 5now 6721 (31); Soderstrom 141 (28), 493 (36), 752 (15); Sohns 177 (36). 289 (15). 300 (15), 424 (15). 505 (15), 507 (15), 533 (15), 534 (15), 538 (36), 553 (8), Volume 89, Number 3 2002 Cialdella & Giussan 3 Phylogenetic Helal bitine of Piptochaetium 581 (15). 601 (36). 603 (36). 604 (31). 782 (15). 790 (36), 799 (36). 804 (36). 819 (36). 977 (36). 982 (36). 1012 (31). 1016 (31). 1022 (15). An (15). 1147 (15). 1500 (36). 1535 (8); Solomon 1535 (4), 1710 (4). 11342 (23); Sperry 370 (15); St. Pierre 888 (36); Standley 40512 (15). 10699 (15). s.n. (15); Stanford 2482 (4), 2604 (15): Ste- vermark 3172 (15). 36700 (36). 48274 (31). 48452 (31). 50955 (36); Strong 1342 (4); Svenson 10246 (4); Swallen 951 (4). 1109 (15), 7561 (35). Talbot 8 (4): Tavlor 276 (15). (15): Tharp 4175 (28): — 28558 (4): — ye (4). ; Toolin 2222 (15): Torres 875 (36), 1226 (29); e s * ). 4544 (4). 7031 (5). s.n. (4): s 61 r Tuckerman 848 ( ji Valencia 442 (22), s.n. (23): Valls 1233 (20). 2 2805 (20). 2821 (33 a). An (20), 2939 (33 a), (22). 12259 (35); Van Eseltime 127 (4): Van Schaack (4): Van Sickle s.n. (4); Vasey s.n. ( (31); Ventura A. 376 (36). 2153 (15). 18506 5). 1921 (15). 1922 (28), 1959 Ов, (15); Villarreal I1: (28). Warner s.n. (4): Warnock 12082 (28), 46717 (15): War- nock T 598 "e Pies: 6322 (15). 12627 (15). 14854 (4). 16410 (36): oo 1122 (4). 5831 (4). 6234 (4): Weatherw rways E (4). 37 (4); West 26673 (4); W — 1 (15). 88 W "y ‘ox s.n. (15); Wolff 4974 (4): Wooton $; E (15): ou 643 Yacolucei 843 ( та Zanin s.n. (1). 18 (1); Zavala Ch. Zöllner 8672 (32). 6254 (15): Tenorio 1197 үз ee (4). 22502 (4). — 215 (20). 12188 3575 2 15 55 — 245 (31). 862 (15): APPENDIN 2 cimens of Hesperostipa and Nassella used for mor- phological and cladistic analyses, as outgroups of Pipto- chaetium. Hesperostipa comata (Trinius & Ruprecht) Barkworth (= Stipa comata Trin. & Rupr.). U.S.A. Colorado: El Paso Co., Ehlers 7526 (UTC vada: Elko Co., E side of Ruby Mtns., 0.7 road mi. Shantytown, Williams et al. 84—49—8 (UTC). Utah: field Co.. located at the head of South Fork, Dugout Creek on the Mes AM in the Henry Mountains. Velson (UTC Hesperostipa neomexicana (Thurber) Barkworth (= Stipa و‎ nnata L. var. neomexicana Thurber). U.S.A. Arizona: without locality, dec 3051 (BAA). Texas: San Antonio, Silveus 1215 (BA 1). Ne- S of Gar- Hesperostipa spartea (Trinius) Barkworth (= Stipa spar- te K — ). S.A. Colorado: Bookvale. Clear Creek Co.. Chur- PR s.n. (BAA 15155). Hlinois: Stark Co., Wady Petra, Chase 808 (BAA). Indiana: — Co.. sandy soil along Pennsylvania railroad and hwy. 29, 2.3 mi. S of Winamac. Potzger 4243 (B: Nassella charruana (Arechavaleta) Barkworth (= Stipa charruana Arechav.). ARGENTINA. Entre Rios: Gualeguaychú. Burkar 25865 (SI); Gualeguaychú, ruta 12, km 180, Burkart & Crespo 22895 (SI). Nassella leucotricha (Trinius & Ruprecht) R. W. Pohl (= Stipa leucotricha Trin. & Rupr.). XICO. С i ‚ Sierra de Arte: aga. cañón E de Jame. — 249 (UTC). U.S.A. Ca near Sonoma. Heller 5350 (Sl). Tex- de Jame, : ifornia: Sonoma Co.. Karnes Co., [sel Texas 2154 (S1): S of Karnes City, as: Bee Co., Herb. Univ. roadside wet ve immediately 10634 (UTC). Nassella meyeniana (Trinius & Ruprecht) Parodi (= Urachne meyeniana rin. & Rupr.). ARGENTINA. Jujuy: Coe T а. Ahn a Pampa. Venturi 9385 (SD. Humahuaca: Tres Cruces, Parodi 9609 (S1): Sierra del Aguila, Venturi € (SI). Nassella pubiflora (Trinius is ape cht) E. Desvaux (— Urachne pubiflora Trin. & Ruj ARGENTINA. Jujuy: ipa CRM ‘a, camino a Pale а de Aparzo, Kiesling & López 3624 (SI). Tucumán: Tafí, km 82. Quebrada del Barón, Diers 282 (S1). Nassella pulehra (Hitchcock) Barkworth (= Stipa pulch- ra Hitche. S.A. California Marin Co., 5 mi. W of Mill Valley. Beetle 2722 (SI): Contra Costa Co.. Orinda, in sandy soil of open cut, Beetle 1718 (SI); Marin Co.. San Rafael Hills. Howell 16251 (UTC Nassella Tungane E. CHILE. IV Región: — Yiles 472 (SI). Nassella tenuissima (Trinius) Barkworth (= Stipa tenuis- sima Trin.). MEXICO. Veracruz: between Tziutlán and Beetle M-581 (UTC). ARG ENTINA, Córdoba: odoro, Stuckert 198 (SI). om Toay, Ea. A Rúgolo 1060 (51). Tucumán: Tafí del Valle. Hueck 9 (S1). Perote. Nassella tric 'hotoma (Nees) Hackel ex Arechavaleta (= Stipa АИ Nees). ARGENTINA. "ui os Aires: Bolívar, Pirovano. Bur- kart E (SD: Tandil. Troncoso 1290 (У). Entre Ríos: E. Carbó. orillas de vías férreas. Burkart 18129 (S1). APPENDIX 3 List of characters and character states. the characters and character states and their range of var- iation are presented as considered in the cladistic m \ description of l. Lemma margins: not involute = 0: involute = 2. Palea shape: flat on abaxial surface = 0: bikee : d on the ab os surface = 3. Palea length: half the length of the lemma to equal ше length of the lemma = 0: longer than the lemma l: s E than half the le — of the le ‘mma = 2 (Fig. 2 2A. 1. Palea ve ^ nerveless = 0; 2-nerved = 1. 5. Long fundamental ce à of the lemma e Pm rmis: ab- sent = Û; prese 6. Fundamental cel e walle irregularly sinuate = 0: regularly sinuate = 2. Epi- dermal characters (5, 6). were taken from Thomasson (1978a). Barkworth (1990). and Torres (1997). Callus length: shorter than 0.8 mm = 0; equal to or longer than 0.8 mm = 1. Callus length is variable within Piptochaetium; it is usually long in species with cylindrical or obconical florets but short in ob- ovoid or lens-shaped florets. In Nassella the callus is variable in length, while it is always long in Hespe- regularly dentate = l; rostipa. 8. Callus shape: acute or — ute = 0 (Fig. 5 blunt or truncate = 1 (Fig. 5F-H). The callus. vias n short, is generally trunc die or blunt; if long, it is acute or subacute. Callus shape varies also within Nassella s.l.. while in Hesperostipa it is always acute. Charac- 336 Annals M — Garden Du On ^ ~ О. 1. wa ters 7 and 8 were traditionally considered as diag- nostic to identify Piptoc haetium sects. Podonoson and Piptochaetium (Parodi, 1944; Sánchez Vega, — Cialdella & Arriaga, 1998). Piptoc haetium sect. — generally included species with a long md ‘ute or subacute callus, while Piptochaetium sect. Ppt haetium presented species with a short and blunt callus Callus — e: glabrous or slightly hairy = 0; densely hairy = 1. Most Piptoc — ы ies have florets with a densely pubescent callu 5A-G). It is usually glabrous or only alinir о іп Piptochaetium brevicalyx (Fig. 5H), P. calvescens, P. > hirtum, P. don p montevidense, panicoides, and P. tovarii. In Nassella the « glabrous or pubescent, with trichomes densely to slightly disposed, while in Hesperostipa the callus is always densely pubescent. cucullatum, P allu us is Awn: persistent = 0; deciduous = 1. йс п e branches densely flowered = 0; not densely flowered = 1. This character varies among species of Piptochaetium = Nassella, in which the branches may be denseh loosely diene’ while in Hesperostipa branches are loosely flowerec Upper glume le ngth: — than the floret = 0: equal to the floret = 1 s of Piptochaetium, Nassella, and ени" ban upper glumes longer than the ‘ly or )ecie floret. Only in P alpinum "id = Drei eal is the upper — as ine as the floret. Floret — rete to fusiform = 0 (Fig. 5А—С); ob- conical = — & Arig, 1998, dh вв. п) obovoid to eh bo = 2 (Fig. 5H): lens-shaped = (Cialdella & Arriaga, 1998; fig. 10C, D). Floret — varies among spec ies of Piptochaetium, being terete, fusiform, obconical, obovoid, or lens-shaped. Nassella presents terete, obovoid, or lens-shaped florets, while in Hesperostipa florets are always terete Lateral compression of the floret: — or slightly compressed = 0 (Fig. 5A-H); conspic — com- — = 1 (Cialde lla & Arriaga, 1998: figs. ip 10C, This character is variable among species of Piptoc htm and Nassella, while in —— * lorets are never laterally cor mpressec Lemma pubesc ence: absen ) (Figs. ТА. 5H); pres- ent = | (Figs. 1B, 5 Papillae of the lemma: then = Ө; present = 1 (Cial- Arriaga, 199 . 2D. 10C). Papillae are ew uan ies — — ^aetium and Nas sella, but are alway ipa Prickles of the lemma: absent ). s absent in Hesper = 0; present = | (Fig. 1 Prickles over b all the lemma surface: absent — 0; present = 2D). Prickles on is 2s third. of the lemma: absent. — . B). Hooks are similar to pric — although these appendages differ slightly (Ellis, 1979). These characters mologous by Metcalfe (1960). Intermediate forms can also be found, making it a difficult chari acter to c ‘las- sify. 0; present = | (Figs. 2€, 5A were rec ognized as ho- some species of Piptochaetium, Nassella, and all spe- cies of He sperostipa. Distribution of prickles (с пагас - ters 18, 19) varies among species, although it is ug ways constant within species. Only the быш. of 2 22. 23. N = 20. N =J — sin P — var. stipoides (Cialdella & лара, 1998: fig. LOF-I) is highly variable, even within a specimen. Base of prickles: surrounded but not covered by adja- cent epidermal ce a = 0; partially covered by — epidermal cells = 1 (Cialdella & Arriaga, 1998: 3D). Partially covered prickle bases occur only in Pip. toc — stipoides var. echinulatum Parodi. Lemma width j the crown: as ed as the crown = 0; narrower than the crown = 1. This char- acter is related to floret shape: lemmas of cylindrical florets are usually as wide as the crown or slightly nar- rowed, while in obconical, obovoid, and lens-shaped florets lemmas are conspicuously narrowed below the crown, being more conspicuous in gibbous florets. Crown contracted to 2 base of the awn: absent — 0; present = 1 (Fig. The crown concept was dis- — by diffe E ator (Spegazzini, 1901; : Parodi. Вн, 390; Muñoz-Schick, 1990; Jacobs el D 95; Cis dell & Arriaga, 1998). As inter- preted м ге, it refers to the distal portion of the lem- ma, which is fused to the base of E awn (Barkworth, 1990; Cialdella & Arriaga, 1998). In Piptochaetium the crown can be contracted or when not con- — it exceeds the — of the awn (Cialdella Arriaga, 1998: figs. OF, . Nassella and Hes- pones. present a contracte T crown, bei ing me mbra- nous, — parted, and, ing at tli base of es aw wn bos — not revolute = Arriaga, 1998: fig. ТОЕ); re — toward the outside Cialdella & Arriaga, 1998: fig. ӨК). Some van cies of Piptochaetium show a s e crown like a ring. generally covered by pri əks, and some- times macrohairs, In РАТА and Hesperostipa the crown is always straight. Macrohairs in p crown: absent = 0 (Fig. 2A); е ent = | (Fig. 2C). This character is only variable P. stipoides var. stipoides and P. fimbriatum. Thes 'se appendages are frequently associated with other types of epidermal appe ndages such as и ‘kles ч hooks. Papillae in the crown: absent = 0; present = 1 (Cial- della & Arriaga, 1998: fig. 3C). These penus are found 1 in a few j W 8P ecies of — — while еу ust. below ^ — ickles, hoc x. 2A); present . This lype of Pere appendage eral species of Piptoc Dan а an id a few sy is frequent in sev Nassella, and complete sly — nt in Hesperoaiina it is always pres nt. / ei e thread-like (E 5H); ja d Aaldella & Arriaga, 1998: ig. ӘП). In Nassella, di and almost all spec ies of Piptoc — A awn is thread-like, bigeniculate, shortly hispid, nd twisted in its basal portion. Only in Piptochae: tium — is the awn cor me-shaped, 1 ight or ес ie =" = slightly curved but not twisted, its base as wide as * crown p ).9-1 mm wide). Floret length/width ratio: > 3.6 = 0; < 3.5 = Crown width: narrow [0,4—0 6(-0.9) mm diam. үз wide (1-1.8 mm diam.) = 1. crown is narrow when contracted awn, except in teclea — which pre- sents a wide and contracted crown (Cialdella & Arria- ga, 1998: fig. ӨК) due to the conic ical she ape of the awn. In Nassella — Hesperostipa the crown is always nar- row in shap SYSTEMATIC REVISION AND Silvia S. Denham, Fernando O. Zuloaga, PHYLOGENY OF PASPALUM and Osvaldo Morrone? SUBGENUS CERESIA (POACEAE: PANICOIDEAE: PANICEAE)! ABSTRACT y-five species are treated in this work, in which exomorphological characters are analyzed cladistic ally. Species of ei subg. Ceresia are characterized by their rigid. filiform to lanceolate blades, inflorescences with one to several racemes, rachis of the racemes winged and hy: dw to membranous, spikelets pilose, occasionally glabrous, with the upper anthecium pale, hyaline to membranous, occasionally chartaceous, and the upper le — not enclosing the tip of the upper palea. Species grow in South America from Mexico to Argentina and Uruguay. A cladistic — of subgenus Ceresia was conducted to test its monophyly, and to establish its relationship with о vila groups of Paspalum. A key to the species in subgenus Ceresia is given. as well as morphological description and illustration, and distribution maps. Key words: America, Ceresia, cladistics, Paspalum, Poaceae. Paspalum L. includes approximately 330 species um subg. Ceresia, considering vegetative and distributed in tropical and subtropical regions of productive characters, analyzing habitat and distri- America, with a few taxa growing in the Old World bution range. A cladistic analysis was further per- (Clayton € Renvoize, 1986). Due to its large num- formed. in order to test the monophyly of the ber of species. and the morphological variation subgenus and to establish the relationships. of its present within the genus, Paspalum has been di- species. vided into subgenera. sections, or informal groups mainly on the basis of morphological characters, HISTORY such as inflorescence type or features of the spike- let (Nees, 1829: Doell. 1877: Chase. 1927: Pilger, 1929; Clayton & Renvoize, 1986; Morrone et al 1995. 1996, 2000; Cialdella et al.. 1995). Chase (1929) divided Paspalum into subgenera Paspalum and Ceresia (Pers.) Rehb., characterizing the Ceresia was established at the generic level by Persoon (1805—1807), on the basis of C. elegans (= ^ Paspalum membranaceum Lam., non Walt.. 1788 Subsequently, Reichenbach treated Ceresia as a subgenus of Paspalum (Chase, 1929). Several spe- cies were described in Ceresia in the 19th century atter by its foliaceous rachis, with one to sev- (Flüggé, 1810; Trinius, 1826. 1828-1836. 1834: eral racemes per inflorescence and spikelets dense- Nees. 1829; Doell, 1877) ly pilose, with long white hairs, mainly along the Nees (1829) considered, within Paspalum. sec- margins of the upper glume and lower lemma. tion Digitaria, including P. pectinatum and species Taxonomic studies i Ceresta are fragme ‘магу; = currently treated in Axonopus; section Lanigeri, in- cluding P. eucomum, P. guttatum, P. erianthum, P? sanguinolentum, P. ammodes, P. polyphyllum, and P. blepharophorum (= P. polyphyllum); and sections with respect to the subgenus, these Paspalum spe- cies have been analyzed in regional works or floral treatments (Nash, 1912: Chase 1927, 1929: Pilger, 1929: Hitchcock, 1951: Burkart. 1969: Rosengurtt— Cristati, Genuini. Axonopodes, and Ceresiae, with Р. et al., 1970: Sendulsky € Burman, 1978, 1980: da pyramidalis (= P. repens) and P. stellatum. Silva et al.. 1979: Pohl, 1980: Judziewiez. 1990: Doell (1877) Renvoize, 1998: Rodríguez. 1998). The present nen treatment revises Paspal- grouped Paspalum into three sec- lions: sect. Eremachyrion, sect. Emprosthion, and sect. Opistion, including in section Opistion the fol- ! We thank Vladimiro Dudás for preparing the excellent — and the staff of Instituto de Botánica си T — rn cially Alejandra Garbini. We also thank the Consejo Nacional de Investigaciones Científicas y Teen (CONICE T for a research grant (PIP 4440) and the National Ge — Society for research grants #6042-97, 6698. A АЕ 68 th Istituto ga Botanic a Darwinion. Labardén 200, Casilla de Correo 22, San Isidro B1642HYD. Argentina. sdenham larwin.edu.a = ANN. MISSOURI Bor. GARD. 89: 337-399. 2002. 338 Annals of the Missouri Botanical Garden lowing species: Paspalum aspidiotes (subg. Ceresia), P. pectinatum (subg. Ceresia), P. eucomum (subg. Ceresia), P. blepharophorum (= P. humboldtianum, subg. Ceresia), P. ammodes (group Eriantha), P. gut- tatum (group Eriantha), P. sanguinolentum (group Eriantha), P. erianthum (group Eriantha), P. stella- tum (subg. Ceresia), P. membranaceum (= P. cere- sia, subg. Ceresia), P. heterotrichon (subg. Ceresia). P. trachycoleon (subg. Ceresia), P. lanciflorum (subg. Ceresia), and P. carinatum (subg. Ceresia). Bentham (1881) considered three sections in Paspalum: sect. Anastrophus, sect. Cabrera, and sect. Eupaspalum, the first two characterized by the presence of distichous spikelets with the lower lem- ma facing the axis, while in section Eupaspalum spikelets are unilateral with the lower lemma out- ward from the axis of the raceme. Within section Eupaspalum, Bentham (1881) distinguished four subsections: subsect. Anachyris, as an artificial group with a single bract below the upper anthe- cium; subsect. Ophistion, with species without a fo- liaceous rachis; subsect. Seudoceresia, with a rachis more or less foliaceous, green, and concave with glabrous and small spikelets (including such spe- cies as Р, stoloniferum and Р. repens); and subsect. Ceresia, characterized by its foliaceous rachis, with the margins membranous and colored, spikelets cil- iate and larger. Within subsection Ceresia, Bentham included many tropical species and also Р cymbi- forme. Pilger (1929) divided Paspalum into eight sec- tions: sect. Eupaspalum, sect. Anachyris, sect. Pter- olepidium, sect. Erianthum, sect. Cymatochloa, sect. Ceresia, sect. Eriolepidium, and sect. Moen- This author treated in section Ceresia such species as P. membranaceum (= Р. ceresta), P. stel- latum, P. heterotrichon, P. carinatum, P. eucomum, P. pectinatum, Р. cordatum, P. humboldtianum, and P. trachycoleon. Chase (1929). species of Paspalum, related subgenus Ceresia with in her study of North American her group Dissecta, both with the rachis of the ra- cemes foliaceous and winged, separating Dissecta by the presence of glabrous — Among the species included by Chase in her treatment are: Paspalum pectinatum, P. contractum (= P. lanciflo- rum). P. stellatum, P. heterotrichon, P. trachycoleon, P. cymbiforme, and P. (1929) stressed that P sanguinolentum (excluded humboldtianum. Chase here) did not belong to subgenus Ceresia but to group Eriantha. Later, Chase (ined.) plac ‘ed in sub- genus Ceresia P. carinatum, P. ceresia, P. heterotri- chon, P. stellatum, P. trachycoleon, P. phyllorhachis, P. soboliferum (= P. humboldtianum), P. polyphyl- lum, P. humboldtianum, P. humboldtianum var. P. humboldtianum), P. buchtienii, P. 2 stuckertii (= malmeanum, P. eucomum, P. splendens (= Р. eu- comum), P. guttatum, and P. ammodes. On the other hand, Paspalum aspidiotes, P. pectinatum, P. cor- datum, P. setiglume, P. contractum, and P. lanciflo- rum were grouped in section Pectinata (see Table 4 for groups considered here). Clayton and Renvoize (1986) recognized eight sections within Paspalum: sect. Diplostachys, sect. Pterolepidium, sect. Anachyris, sect. Erianthum, sect. Eriolepidium, and sect. Moenchia with a tri- quetrous, not membranous, rachis: sect. Paspalum and sect. Ceresia with a membranous or foliaceous, winged rachis. Section Ceresia was distinguished Clayton & Renvoize, 1986) by including species — with a membranous, colorful rachis, and spikelets with the upper glume and lower lemma scabrous, usually ciliate, sometimes winged. Rodríguez (1992) established Paspalum sect. Pectinata and distinguished it from section Ceresia by including species with spikelets winged and broadly ovoid, cordiform or not at the base (1994) dese ире, within Paspalum subg. Ceresia, section Biaristata, includ- Filgueiras and Davidse ing two species: P. biaristatum and P. longiarista- tum, both defined by conspicuous awns on the up- per glume and lower lemma. Most recently, Rodríguez (1998) made an ac- count of nine species of Paspalum subg. Ceresia for Venezuela: P. setiglume, P. pectinatum, P. lanciflo- rum, P. carinatum, P. humboldtianum, P. polyphyl- lum, P. stellatum, P. trachycoleon, and P. heterotri- chon, all with a winged rachis and pilose spikelets. MATERIALS AND METHODS MORPHOLOGICAL ANALYSIS Specimens were examined from the following herbaria: B. BAA, BM, COL, CONC, CTES, G, — LIL, LPB, MA, MEXU, MO, P, R, ‚ US, W (Holmgren et al., 1990; Ap- pendix PS | list of specimens examined for the outgroup taxa is provided in Appendix 3. Scanning electron micrographs (SEMs) were prepared of the upper anthecium of six species, utilizing the pro- cedures described by Soderstrom and Zuloaga 1989). The marked with an asterisk in Appendix 2: the type specimen of P. reticulinerve, Solomon 17003, was vouchers for this SEM study are — also used. The specimens were viewed on a Zeiss 940 A scanning electron microscope of the Dar- winion institute, operating at 10—20 kV. PHYLOGENETIC ANALYSIS Characters. Fifty-eight exomorphological char- acters were used in the cladistic analysis: 15 char- Volume 89, Number 3 2002 Denham et al. 339 Paspalum subg. Ceresia Table 1. Morphological characters and character states used for the cladistic analysis of Paspalum subg. Ceresia. Life form 1. Life cycle: perennial (0), annual (1). This character is a sy napomorphy for — Racemosa in Paspalum. es in section — are Paspalum lon- uum ang P. cac —* uatic or — aquatic plants (1). ıs with spongy Sith: absent (0), present (1). Both characters 2 and 3 are helpful to solve relationships within the outgroup. Blades 4. Shape: linear-lanceolate (0), linear (1). э. Basal leaves: not. pseudopetiolate (О), shortly pseu- dopetiolate (1), markedly pseudopetiolate (2). Paspal- um buchtienti, P. cymbiforme, anc have s Annual specie ві aristi 2 — shortly pseudopetiolate leaves, with the pseu- dopetiole a narrowed zone at the base of the blade. Markedly pseudopetiolate leaves is an autapomorphy a long narrowed zone between the — and the flattene d blade: both about the same leng for P. imbricatum, where leaves > Inflorescenses First-order branching of species here treated in Par cum and Anthaenantiopsis are considered аа to unilateral racemes in Paspalum. 6. Spikelets b arranged on the rachis: absent (0). present (1). l inilaterally arranged spikelets are present in all studied species alum. Truncate inflorescences: oni (О), present (1). The inflorescence is truncate whe i end in a spikelet, and the most distal raceme is a first order branch 8. Raceme number per inflorescence: 1 (0). 2 (1), 3 or more racemes (2). 9. ннан, of racemes: solitary (0), conjugate (1), subdigitate (2), alternate (3). Racemes are conjugate 1 the internode between two contiguous racemes is almost absent; when these internodes are short, up lo finally, when the internodes are well developed, racemes are con- ст long. racemes are subdigitate: sidered as alternate 10. Rachis of the racemes: triquetrous (0), winged with longitudinal veins at — wings (1), winged with anas- tomosed veins at the wings (2), winged and nerveless, with hyaline to кр Мокен wings (3) scored as triquetrous when the ventral side has the e rachis was same development of the lateral margins. while it was s winged when the lateral margins are Nerveless and haine to m ашин wings i» occurs in some species of considered clearly developed and flat. ‘the racemes: duni ina spike let (0), in a duis du (1). in both a apike ‘let or in a naked point let or in a naked boit in several species of * sub- genus, while in P. goyasense it always ends in a developed spikelet 12. Rachis disarticulating at maturity: absent (О), present (1). This character is a synapomorphy of the Race- mosa group. Table 1. Continued. 13. Pedicels laterally inserted in the spikelets: absent (О), present (1). This — i erotrichon, P. petr rachycoleon, and P. phyl- lorhachis, where ar is — at the lateral base of the upper ane (Fig. ТЕ, Spikelets 14. Shape: biconvex (0), plano-convex (1). Plano-convex sp vikelets are a EM within the studied spe- cies of Paspalur 15. Disposition — (O), paired (1), lower spikelet reduc ed or r aborted paired with the 2); with 3 or more ГС le are present on one а h in п Pi icum = and P. obtusum. 16. Spikelet with an annular thic kening at the — ab- sent (0). p esent (1). This character is present in Pas- palum — thon, P. petrense, P. — and P. phyllorhachis. Lower Glume 17. Presence: present (0), absent (1). The lack of a lower glume is a synapomorphy of the studied. species of p aspalum Upper glume 18. Texture: hyaline (0), ии We consider Ше de ght passes through it. present. (1). winged upper glume ST flat and extending margins. This character is shared by n Pectin- ata, with — exception és Paspalum paper Md and ? cachimboe 20. Base — als n (0), prese nt (1). As in the pre- vious character, this feature is present in section Pec- tinata (but mot: in Ты — нун * P. cach- imboense). species of sectio — N Apex: nol awned (0). awned (1). Awned glumes are present in Paspalum longiaristatum and P. biarista- tum. 22. Corky margins: absent (0), present (1): usually found in Pospalum buchtienti, P. суть P. humbold- tianum, P. polyphyllum, P. niquelandiae. and P. bur manii. iforme, ). We consider when they 23. Ciliate margins: absent (0), present (1 that the margins are ciliate rigid, and setose hairs in only one series, Paspalum reticulinerve and Р. aspidiotes. 24. Pilose margins: absent (0), present (1). Mu eee I have short > long, not rigid, and arranged in s èveral s 25. Radiate marginal hairs: absent (0), present m A hen marginal hairs are radiate, they look like a crown, as in Pom buchtienii, p homo. P. poly- phyllum, P. аша а and P. burmanit. 26. Surface pilosity: absent (0), hairs — arising from і basal point (1), bus present in the lower half of the жые glume (2), hairs all over the surface of th upper EM (3). Hairs are usually found in Paspalum 78 g. Ceresia in the basal portion and margins of the übt glume; they are less frequent in the distal por- ion. Annals of the Missouri Botanical Garden Continued. Table 1 Table 1. Continued. 2 (0), 3 (1), 5 (2), 7 (3). equidistant (0), lateral nerves near the margins (1). This character is helpful to re- solve relationships within outgroups. 29. Position of lateral nerves: marginal (0), submarginal (1). 30. Lateral extension of the margins: flat (0), curved after the nerve (1), plicate along the nerve (2), absent (3). We considered a lateral extension present when the margin continues after the lateral nerve. 31. Marginal nerves only present in the lower half of the . This is a com- s on Pectinata, with the —— of Passo bn iflorum and P. cach- 27. Numbe в of nerves: = = - = = к ч ys = = = = = - = = > ы У. А = — — <~ => w ~ 4 ج‎ 2 элш; — imboens 32. id veins in the upper glume: absent (0), present (1). This is a character observed in Paspalum reticulinerve and P. aspidiotes. 33. Base ко шар, absent (0), present (1). ' is observed in Paspalum buchtienii and P. а. tianum, both species with basal margins extended. Lower floret 34. Lower floret: present, staminate (0), absent (1). The absence of a lower floret is a synapomorphy of Pas- patum Lower lemma 35. Texture: hyaline (О), membranous (1). 36. Wings: absent (0), present (1). Wings on the lower lemma were considered present when the margin of the lower lemma is flat and extended, as in rene reticulinerve, P. imbricatum, and P. aspidiot 37. Base: rounded (О), subcordate (1), cordate (2). Ad date lower lemma was observed in Paspalum imbri- catum and P. aspidiotes, while it is subcordate in P. pectinatum and P. cordatum. 38. Apex: not awned (0), awned (1). awned apex o the lower lemma is found in ae longiaristatum and P biaristatum. 39. Ciliate margins: absent (0), present (1). As in char- acter 23, the lower lemma margins are considered ciliate when these appendages are short, rigid, setose, and arranged in one series along the margins, as in Paspalum reticulinerve and P. aspidiotes. 40. Margin pilosity: absent (0), present (1). 41. Tuberculate, marginal hairs: absent (0), present (1). This is a character observed in / ie lane — P. cachimboense, P. pectinatum, and P. cordat 42. Surface pilosity: Absent (0). ent 43. Number of nerves: 2 (0), : . 5 (2). 14. Disposition. of the nerves: — nt (0), lateral nerves near the margins (1). This character is helpful to resolve relationships within outgroup 45. Disposition of lateral nerves: marginal (0). submargin- al (1). 46. Lateral extension of the margins: flat (0), curved after the nerve (1), plicate along the marginal nerve (2), absent (3). See character 30. 47. Marginal nerves only present in the lower half of the One or two mar- ginal nerves, › not reach the apex, are present in Paspalum imbricatum and P. aspidiotes. fob - -U ۹ 48. Relative length of upper glume and lower lemma: both bracts of the same length (0), upper glume longer 1), lower lemma longer (2). Upper anthecium 49. — — to lanceolate (О), i 0 Stipe: abse 0), present (1). p aspalum ——— P. cac мй р aspidiotes, and P. stellatum. — i ne apex: not gaping, when the upper lemma encloses the apex of the palea (0), g ing, when the apex of the palea is free (1). All species of Paspalum subg. Ceresia have an open upper an- thecium. 53. Relative length of the anthecium and the spikelet: equal (0), upper anthecium shorter than the spikelet 54. Papillae at the upper lemma: absent (0), sin ple a pillae (1), verrucose papillae (2). A simple papilla i an epidermic prominence including the cellular lu- men, and a verrucose papilla e a papisa with addi- tional prominences on the surface (Arriaga, 1987). Verrucose papillae are found in — guttatum, in the Eriantha outgroup, and in some specimens of P. stellatum within Paspalum subg. Ceresia. 55. * ‘a bodies at the upper lemma: absent (0), present 56. Unicellular macrohairs at the upper lemma: absent (0). present (1). 57. Bicellular microhairs at the upper lemma: absent (О), present 58. Lodicules: present ( (0), absent (1). Lodicules are ab- sent in biaristatum. Paspalum ceresia, P. longiaristatum, and P. acters are multistate and were treated as unordered; 11 of these characters were uninformative since they did not show variation within the ingroup. Ta- ble 1 provides the list of characters used in the analysis. Autapomorphies were excluded from the analysis. The data matrix is shown in Table 2. Some potentially informative characters were left out of the analysis, either because of the large amount of intraspecific variation, the lack of enough material, or because it was extremely difficult to score them due to overlapping states in different species, e.g spikelet length and width, rachis width, and blade pilosity. Cladistic analysis. based on maximum parsimony using NONA version 2.0 (Goloboff, 1997a) under equal weights. Due to he large number of terminals, a heuristic analysis The cladistic analysis was was run following command sequence: “hold 10000; rseed 0; hold/40; poly = mult* LOO; max*; sswap*." One hundred subsearches were performed, each constructing a Wagner tree using a random ad- dition sequence. swapping the initial tree with TBR > amb-: Volume 89, Number 3 Denham et al. 341 2002 Paspalum subg. Ceresia Figure 1. 'anning electron mic uerographs of the upper anthecium of Paspalum.—A. P. — apex of the upper lemma (Dusén ge SI). —B. P. reticulinerre, apex oi M = per palea (Solomon 17003, MO). —C. P. trachycoleon, upper palea with simple papillae — 2046, US P. biaristatum, upper palea "D bicellular mic rohi vrs, macrohairs (Oliveira 750, SI. E. KV. P. trachycoleon. 3 pus let base. dorsal view (Tamayo 2046, US). —F. S pikelet base. ventral view (А — 3004. Pa (tree. bisection-reconnection) and retaining a maxi- swapping cutting two sister nodes simultaneously. mum of 40 trees in each replication. The resulting The default option “amb-” retained only nodes that trees were swapped using one round of TBR (max* are supported by unambiguous optimizations, while sswap*.” which performed TBR poly = treats trees as polytomous. we and one round of 342 Annals of the Missouri Botanical Garden Table 2. Data matrix used in the cladistic analysis. Polymorphic character states are scored as: A = [01]; B = [02]; € = [03]; D = [12]; E = [23]; F = [012]; G = [013]; Н = [023]; 1 = [123]. Dashes indic ‘ate character inapplicable for a taxon. 1991). First, command “jump*3” was used, which performs TBR swapping on trees up to 3 steps lon- ger than initial trees (Goloboff, 1993). The second strategy was carried out using “hold/5; mult*400,7 which performed 400 independent searches begin- ning from different Wagner trees. Bremer support (Bremer, 1994) and jackknife Lor gd 1 1 1 1 |) 1 | Character | 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 Panicum laxum 0 0 0 0 0 1 0 2 3 0 0 0 0 0 G 0 0 1 O0 Panicum obtusum 0 0 0 0 0 0 0 2 3 0 0 0 0 0 G 0 0 I O0 Anthaenantiopsis fiebrigii 0 0 0 0 0 1 0 2 3 0 0 0 0 0 I 0 о | 0 Paspalum acuminatum 0 | 1 0 0 |! | F H 1 0 0 0 | 0 0 1 0 O0 P. almum 0 0 0 0 0 1 l D 1 1 0 0 0 l 0 0 l 1 0 P. candidum I1 0 0 0 0 1| 1 2 3 2 1 1 0 1 0 0 ıl — — P. cromyorrhizon 0 0 0 1 0 1-1 D 2 0 0 0.0 1.0 0 1 1 0 P. dissectum 0 l 1 0 0 l 1 2 3 2 0 0 0 1 0 0 1 A 0 P. erianthum 0 0 0 0 0 1| 1 2 E 1 0 0 O 1 1 ото O0 P. guttatum 0 0 0 1 O 1.1 F F 1 0 0 0 1 0 0 1 1 O0 P. notatum 0 0 0 0 0 l 1 D D 0 0 0 0 1I 0 0 | 1 0 P. prostatum 1 0 0 0 0 1 1 2 E 2 1 1 0 1 0 O0 | 1 0 P. racemosum 1 0 0 0 0 1 1 2 E 2 1 1 0 1 0 0 1 | 0 P. sanguinolentum 0 0 0 0 0 1 !|! D 3 1 0 0 0 1 2 0 1 I 0 P. aspidiotes 0 0 0 0 0 1 1 2 331 0 0 1 0 O 1 | l P. biaristatum о 0 0 1 0 | | F C 3 1 0 0 1 0 0 1I 0 O0 P. buchtienii о 0 0 0 1 l | 2 E 1 2 0 0 1 2 0 1 A O0 P. burmanii 0 0 0 0 0 1 1 2 3 2 2 0 0 1 1 0 | 1 0 Р. cachimboense | 0 0 0 0 |! | F 2 3 1 0 0 1 0 0 1 ] 0 P. carinatum 0 0 0 1 0 | 1 A € 3 1 0 0 1 0 0 ! ] 0 P. ceresia 0 0 0 A 0 1 1 F € 3 l 0 0 l 0 0 ] A 0 P. cordatum 0 0 0 1 0 | 1 2 2 1 1 0 отоо 1 1 Р. cymbiforme 0 0 0 0 1 1 IFC 3 1 0 0 1.) 2 0 1 A 0 Р. eucomum 0 0 0 1 0 1 1 D D 3 1 0 0 1 0 0 ! ] 0 P. Ke 0 0 0 0 0 1 1 F € 3 0 0 0 1 0 0 1 1 0 P. heterotrichon 0 0 0 0 0 1 l F € 3 l 0 1 1 0 1 1 0 0 P. ern о 0 0 0 1 1 1 D 3 1 2 0 0 1 2 0 1 0 O0 P. imbricatum 0 0 0 0 2 | 1 D E 3 1 0 0 1 0 O 1| 1 1 Р. lanciflorum 0 0 0 0 0 | | F B 3 1 0 0 1I 0 0 | | 0 P. longiaristatum | 0 0 1 0 1 1I F C.3 1 0 0.1 0 0 1 0 0 P. malmeanum 0 0 0 ] 0 | 1 A A 3 1 0 0 1.0 0 1 | 0 P. niquelandiae 0 0 0 0 0 1 1 2 3 1 2 0 0 | 1 0 1 1 0 P. pectinatum 0 0 0 0 0 1 1 D D 3 1 0 0 1] 0 0 1 A | P. petrense 0 0 0 0 0 1| г 2 S 2 1 0 1 | 2 ] 1 A 0 P. phyllorhachis 0 0 0 0 0 1 71 2 3 2 | 0 1 111 1 A 0 P. polyphyllum 0 0 0 A 0 1 1 F € 2 2 0 0 1 2 0 1 0 0 Р. reticulinerve 0 0 0 0 0 1 ID D 1 1 0 0 1 1 0 |! l P. stellatum 0 0 0 1 0 ! 1 A A 3 1 0 0 1 0 0 ! 1 0 P. trachycoleon 0 0 0 0 0 | 1 2 3 2 1 0 1 l l l | A 0 Tests to evaluate homoplasy were consistency in- group frequencies (Farris et al., 1996) were con- dex (CI) and retention index (RI) (Farris, 1989). ducted in order to calculate branch support with Two strategies were carried out in an attempt to NONA. Commands used for Bremer support were find possible islands for MP trees (Maddison, — "subN: hold1500; find*”; bsupport performing TBR swapping on preexisting trees, saving trees up to * 4 steps longer at each round. Jackknife frequency values were calculated using the available instruc- tion file for NONA, the JAK.RUN file. One thou- sand iterations were performed, randomly deleting 36% of the characters, with the search options “mult*5,” and the FQ.EXE file was used to cal- Volume 89, Number 3 2002 Denham et al. 343 Paspalum subg. Ceresia Table 2. Extended. 2 2 2 2 2 2 2 2 2 2 j j 3 3 3 j 3 j j 3 0 1 2 3 1 5 6 7 8 9 0 1 2 3 4 5 6 T 8 9 0 0 0 0 0 0 0 D 0 l 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 3 0 l 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 3 2 0 1 0 0 0 0 0 l 0 0 0 0 0 0 0 0 0 0 0 2 l l 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 2 О О 0 l l 0 0 0 0 — — — — — — — — — — — — — — ] 0 0 0 0 0 0 0 0 0 0 0 0 2 l | 2 О 0 0 ] | 0 0 0 0 0 0 0 0 0 0 0 2 l | 1 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 j 2 l l l 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 j 2 | 1 1 0 0 0 l l 0 0 0 0 0 0 0 0 0 0 0 2 l l 2 0 0 0 1 l 0 0 0 0 0 0 0 0 0 0 0 1 0 | l 0 0 0 1 l 0 0 0 0 0 0 0 0 0 0 0 1 0 | l 0 0 0 1 | 0 0 0 0 0 0 0 0 ] 0 3 2 | l 1 0 0 0 1 1 0 0 0 0 l 0 0 l 0 0 0 3 0 l 0 l l О 1 1 l 2 0 l 0 l 0 0 1 0 1 1 0 0 3 0 0 0 1 1 0 0 | 0 0 0 l 0 l | 2 | 0 1 2 0 0 l 1 1 0 0 0 0 0 0 1 0 1 1 2 1 0 l 2 0 0 0 l l 0 0 0 0 0 0 0 0 0 0 l l 0 | | 0 0 0 1 1 0 0 0 0 0 0 0 0 l 0 2 ] 0 | 2 0 0 0 1 | 0 0 0 0 0 0 0 0 1 0 l 1 0 0 3 0 0 0 1 0 0 0 0 0 l 0 0 0 0 0 0 2 0 | 0 1 0 0 1 1 0 l 0 0 0 0 | 0 l 0 2 l 0 | 2 0 0 0 1 1 0 0 0 0 0 0 0 0 l 0 2 l 0 l 2 0 0 0 l l 0 0 0 0 0 0 0 0 l 0 2 | 0 | 2 0 0 0 1 l 0 0 0 0 0 0 0 0 1 0 l 1 0 | 2 0 0 0 1 | 0 0 0 0 0 0 l 0 1 1 3 1 0 | 2 0 0 1 1 1 0 0 0 0 l О О 0 0 0 0 2 0 l 0 1 0 0 1 1 1 2 0 0 0 0 0 0 0 0 | 1 0 1 l 0 0 0 1 1 0 0 0 0 0 l 0 0 l О l l 0 0 3 0 0 0 ] 1 0 0 l 0 0 0 0 0 l 0 2 l 0 | 2 0 0 0 1 1 0 0 0 0 0 0 1 0 1 1 0 l 0 | 2 0 0 0 ] 1 0 0 0 0 0 ) 0 0 0 0 D 0 l 0 l 0 0 1 l 0 1 0 0 0 0 0 0 l 0 2 l 0 | 2 0 0 0 1 l 0 0 0 0 0 0 0 0 0 0 0 1 0 | 2 0 0 0 l 1 0 0 0 0 0 0 1 0 l | 3 1 0 | 2 0 0 0 1 1 0 0 0 0 1 0 0 l 0 0 0 j 0 | 0 l | 0 1 l l 0 0 l 0 0 0 0 1 0 1 \ 0 0 3 0 0 0 l l 0 0 0 0 0 0 0 0 1 0 2 l 0 | 2 0 0 0 l l 0 0 0 0 culate the majority rule consensus trees of the out- put (Goloboff. 1993). Separate analysis using implied weights (Golo- boff, 1993) was run in PeeWee version 3.0 (Golo- boff, 1997b) using the same search strategies as in NONA. PeeWee weights characters according to how well they fit a specific tree. Homoplastic char- acters were downweighted in proportion to their number of extra steps. We used a medium concave function, K = 3. In the present. treatment we fol- Outgroups. lowed Nixon and Carpenter (1993) for selecting the outgroups. Species of Paspalum that share potential synapomorphies with those of subgenus Ceresia were included as outgroups: taxa of the Dissecta. Eriantha, Racemosa, and Моа groups in Paspal- um were chosen because they share with subgenus Ceresia a wide rachis in the inflorescences (Chase. 1929, ined.: Morrone et al., 1995; Zuloaga & Mor- rone. in prep.). Therefore, P acuminatum and P. dissectum, of group Dissecta, P erianthum, P. gut- tatum, and P. sanguinolentum of group Eriantha, Р, Annals of the Missouri Botanical Garden Table 2. Continued. 4 4 4 4 4 4 1 4 4 5 5 5 5 5 5 5 5 5 Character 0 1 2 3 4 5 7 8 9 0 1 2 3 4 5 6 7 8 Panicum laxum 0 0 0 ро 1 0 0 0 0 D 0 0 0 O0 1 0 O0 Panicum obtusum 0 0 0 2 O0 l 0 2 0 0 2 0 0 0 1 0 1 0 Anthaenantiopsis fiebrigti 1 0 1 2 0 | 0 2 0 0 1 1 O 1 ! 1 | O0 Paspalum acuminatum 0 0 0 2 I| 1 0 0 0 0 2 0 1 1 1 1 1 O0 P. almum 0 0 0 2 | l ооо 0 2 0 l и 0 0 0 0 P. candidum о 0 0 1 0 1 0 — 0 0 D 0 0 1 1 0 0 0 P. cromyorrhizon 0 0 0 2 l l о 0 0 0 D 0 1 1 1 01 0 P. dissectum 0 0 0 2 l 1 0 0 l 0 2 0 0 l 1 0 0 0 P. erianthum 1 0 | 1 01 0 0 0 0 2 0 1 1 0 0 1.0 P. guttatum lO ! ]! 0 | 0 0 0 0 2 0 1 2 1 0 1 O0 P. notatum 0 0 0 2 | l 0 0 0 0 2 0 1 1 0 0 1.0 P. prostatum 0 0 0 1I 0 1 0 0 0 0 2 0 0 1 0 0 то P. racemosum 0 0 0 1 0 |! 0 0 0 0 2 0 1 O т 0 0 O0 P. sanguinolentum 1 0 1 2 | I 0 0 0 0 2 0 0 I т O I O0 P. aspidiotes 0 0 0 2 0 1 1 1 O 1 |! 1 1 1 1 I |! O0 P. biaristatum 1 0 | 1 0 0 0 1 0 0 ! l 1 0 0 1 1 1 P. buchtienii 0 0 0 | 0 | 0 0 0 0 1 1 1 1 1 l | 0 P. burmanii 0 0 0 1 0 | 0 0 0 0 1 1 0 1 1 1 | 0 P. cachimboense | 1 0 | 0 1 0 0 0 | | 1 1 1 0 1 | 0 P. carinatum | 0 1 | 0 l 0 1 0 0 1 1 1 1 0 1 | 0 P. ceresia | 0 1 0 0 l 0 0 0 0 0 l 0 0 1 1 l 1 P. cordatum | 1 l 1 0 |! 0 1 0 0 ! 1 l 1 l 1 | 0 P. cymbiforme 0 0 0 1 0 1 0 0 0 0 | 1 dd 1 1 1 | 0 P. eucomum 1 0 1 1 0 |! 0 |! 1 0 |! 1 | 0 0 0 0 O0 P. goyasense 1 0 | 1 0 1 ол 0 0 D 1 1 1 0 1 0 0 P. heterotrichon 1 0 0 1 0 1 0 0 0 0 1 1 1 01 1 | 0 P. humboldtianum 0 0 0 +t 0 | 0 0 0 0 1 1 1 1 1 1 | 0 P. imbricatum 0 0 0 2 0 1 | B 0 0 ! 1 l 1 0 1 | 0 P. lanciflorum 1 1 0 | 0 | 0 1 0 1 lol 1 1 1 O ]|! O0 P. longiaristatum | 0 | | 0 O0 0 ] 0 O 1 1 1 O I 1 | | P. malmeanum l 0 l l 0 1 0 0 | 0 l l 0 0 0 0 0 0 P. niquelandiae 0 0 0 1 0 I 0 0 0 0 D 1 0 1 0 0 I O0 P. pectinatum 1 1 1 1 O |! 0 1 0 0 1 1 1 1 1 1 I ©0 P. petrense 1 0 | 1 01 0 0 0 0 1 1 1 1 l 1 | 0 P. phyllorhachis 0 0 0 +t 0 ! 0 0 0 0 1 1 0 1 0 | | 0 P. polyphyllum 1 0 | | 0 1 0 0 0 0 1 1 0 1 l l | 0 P. reticulinerve 0 0 0 l 0 1 0 l 0 0 1 1 1 1 0 l l 0 P. stellatum 1 0 1 0 0 0 0 0 l 1 1 1 1 B 0 0 0 0 P. trachycoleon 0 0 0 l 0 l 0 0 0 0 l 1 0 1 0 l l 0 almum, P. cromyorrhizon, and P. notatum of group Notata, and P. candidum, P. prostatum, and P. ra- cemosum of group Racemosa were included in the ма present analysis (Table 4 Panicum obtusum Kunth and Anthaenantiopsis fiebrigii Parodi were also included as outgroups due to their relationship with Paspalum, stressed by a recent molecular analysis of the tribe (Giussani et al., 2001). Finally, for the purpose of rooting, Pan- icum laxum Sw., a C, representative of the x = 10 clade of the Paniceae (Giussani et al., 2001) was also considered in this analysis. MORPHOLOGICAL CHARACTERS Habit. perennial, rarely annual (in P. longiaristatum and Species of Paspalum subg. Ceresia are Р. cachimboense). Culms are herbaceous, occasion- ally lignified (in P. buchtienii and P. humboldtian- um), unbranched or branched at the lower and mid- dle nodes, hollow, and erect or leaning on the vegetation. Blades range from filiform to linear or lanceolate; a short pseudopetiole is present in P buchtienti, P. humboldtianum, and P. cymbiforme, while P. imbricatum is characterized by its long pseudopetiolate blades. Volume 89, Number 3 Denham et al. 345 2002 Paspalum subg. Ceresia F Figure 2. Scanning elec icu micrographs of the upper anthecium 9 —— —A. Р. pectinatum, apex of the upper palea (Dusén 16057, SI). —B. ralmeanum. upper palea (Killeen 2024, SI). —C. P. heterotrichon, upper palea with silica bodies and macrohairs — 4176, US). —D. Р. cordate, apex ol f the upper palea (Weddell 1699, P). —E. Р. pee spikelet base, ventral view (Ailleen 247 1, SI). US | P heterotrichon, spikelet base, ventral view (Tovar 4176, Annals of the Missouri Botanical Garden Inflorescences. Inflorescences are usually ter- minal; axillary inflorescences are common in Pas- palum buchtienii and P. polyphyllum. The number of racemes varies from a single one, e.g., P. stel- latum and P. carinatum, to two conjugate racemes, e.g., P. malmeanum, P. eucomum, P. reticulinerve, and P. pectinatum, while two alternate racemes are characteristic of P. goyasense. When several ra- cemes are present, these can be alternate and dis- tant, e.g., P. trachycoleon, P. heterotrichon, P. cere- sia, P. imbricatum, P. humboldtianum, | P. polyphyllum, P. cymbiforme, P. aspidiotes, and P. niquelandiae, or they can be subdigitate (with a short internode, up to 1 em long, between the ra- cemes) as in P. cordatum, P. lanciflorum, and P. cachimboense. Pulvini are pilose, and one or two bracts are usually present in P. stellatum, P. eucom- um, P. goyasense, P. malmeanum, P. carinatum, and P. pectinatum. Rachis of the racemes. Species of Paspalum subg. Ceresia are typically distinguished by a winged, membranous to foliaceous rachis, 0.8—10 mm wide; the rachis is usually glabrous, occasion- ally pilose (in Paspalum polyphyllum) or ciliate at the margins (in P. reticulinerve); hyaline to mem- branous, and nerveless margins are present in sev- eral species, e.g., P. lanciflorum, P. cachimboense, P. imbricatum, P. pectinatum, P. aspidiotes, P. het- erotrichon, P. cymbiforme, P. ceresia, P. stellatum, P. eucomum, P. malmeanum, P. carinatum, P. goy- asense, P. biaristatum, and P. longiaristatum. Although the rachis usually ends point, in several in a species, such as P. humboldtian- um, P. buchtienii, P. polyphyllum, P. niquelandiae, and P burmanii, the rachis finishes in a naked point or in a spikelet; in P. goyasense the rachis always ends in a developed spikelet. Spikelets are imbricate and arranged in two or four series. Spikelets. siventrally compressed, ranging from ovoid to ellip- Spikelets are solitary or paired, dor- soid to lanceolate. The size of spikelets varies from 1.4 to 8 mm in length; they are usually covered with indument, otherwise ciliate; glabrous spikelets are found in Paspalum phyllorhachis and P. imbri- P. lon- giaristatum and P. biaristatum. The lower glume is always absent, while the upper glume is usually as long as the spikelet and 3- to 5-nerved, 7(—9)- nerved in P. aspidiotes and P. reticulinerve, occa- sionally 2-nerved in P. stellatum. catum. Awned bracts are present only in Also, the upper glume is cordate and winged in P. imbricatum, P. reticulinerve, P. aspidiotes, P. cordatum, and P. pec- tinatum. The lower lemma is winged in P imbri- catum, P. reticulinerve, and P. aspidiotes. A lower palea and lower flower are always absent. Upper anthecium (Figs. 1, 2). The upper anthe- cium is planoconvex, ellipsoid, ovoid or obovoid, usually hyaline to membranous, but membranous to chartaceous in Paspalum niquelandiae and Р. goyasense. Margins of the upper lemma are flat, not enclosing the apex of the upper palea. A short stipe is present in P. stellatum, P. lanciflorum, P. cach- imboense, and P. aspidiotes. Regarding the upper anthecium ornamentation, simple papillae regularly distributed, bicellular mi- crohairs and macrohairs, more densely placed to- ward the apex of the upper lemma and palea, are common features of most species in subgenus Cer- esia. A smooth and glabrous upper anthecium is present in Paspalum malmeanum, P. eucomum, and P. stellatum. Bicellular microhairs, 50-70 рт long, have a short basal cell and an elongated apical cell. Unicellular macrohairs, present or absent in sub- genus Ceresia, can be short and hooked, ca. 40 jum long, to long and thin, 60-300 рт long. Rounded or elongated silica bodies are frequent in several species; these silica bodies have usually 2 to 4 con- strictions; 6 or more invaginations are present in P. lanciflorum. DisTRIBUTION AND. HABITAT Species of Paspalum subg. Ceresia grow in America, ranging from Mexico, Mesoamerica, and the Caribbean (Haiti, the Dominican Republic, and Trinidad and Tobago), to South America (Colombia, Venezuela, Guyana, Surinam, Peru, Brazil, Ecuador, Bolivia, Paraguay, central Argentina, and Uruguay). Brazil hosts the largest number of spe- cies, 22, and endemics, 10, in the subgenus. Taxa are usually found in limestone, sandy, or rocky soils, in open fields or mountain slopes, from sea level to 3000 m elevation. Several species, such as P. carinatum, P. lanciflorum, and P. pectinatum, are found in open areas subjected to periodic fires; these species have culms and sheaths adapted to fire. Paspalum malmeanum and P. cordatum grow in flooded savannas. Paspalum stellatum, P. pectinatum, and P. cari- natum are the most widespread species in subgenus Ceresia. Paspalum stellatum grows from Mexico to southern Brazil, Paraguay, northeastern Argentina, and Uruguay: Paspalum pectinatum is found from Mexico to southern Brazil, and P. carinatum from Nicaragua to Brazil and Trinidad and Tobago. Re- garding the biogeographical distribution of subge- nus Ceresia, species are common in cerrados, At- lantic forests, and savannas; they are also present in caatingas of northeastern Brazil and in the Guy- ana highlands of Guyana and Venezuela; they are Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia Table shown in ion 4. with character List of synapomorphies of the — state changes in the nodes and within taxa. Characters in bold are common synapomorphies for all 40 most parsimonious trees. Nodes 1-37 Node 1: character 55: 0 2 1, character 57: 0 > | Node 2: character 54: 0 е l Node 3: character 7: durs . character 14: 0 > I. character 17: 0 > 1, chr ter 34: 0 > 1 Node 4: character 1 : , character 27: 2 > |. character 28: | > 0, chara ter 43: 2 > 1 Node э: charac ter 40: 0 > l, character 51: 2 > 1, character 52: 0 > 1, — 56:0 > | Node 6: aie 9: 3 2 Node 7: charac id 24: 0 => l, character 26: 0 > 2 character 30: 1 > 2 Node 8 — 26: 0 > : character 50: 0 => | Node 9: character 19: 0 = I. character 20: 0 > 1. character 27: | > 2, athens 30: | => 0. od iter 31:0 Node 10: character 42: 0 > 1 Node 11: character 36: 0 = 1, character 40: | — 0 Node 12: character 2 > 0 2 1, character 27: 2 > 3. e — ter 39:0 => 1 „ character 55: 0 > | character 32: 0 = Node 13: character A | 4: character 4: => Node 15: character 46: 1 > 0, character 54: Node 16: character 49: 0 = 1, character 56: character 57: | => 0 Node 17: character 26: 2 > |, character 30: 2 > 3 Node 18: character 45: ] e ] 260, character 29: 1 > 0, 1 = 0, character 46: y >3 — 0, charac Е 21:05 I, character 38: 0 = 1. бт ви ter 48: bri Node 20: character 18: | > 0 Node 21: character 13: : > 1. character 16: 0 > | Node 22: character 10: Node 23: character T ! = 0, character 53: | > 0. character 55: | > character 40: | => 0 Node 24: character * 02 1. | : | => 2, character 25: 0 > | : character 15: 2 > 1, | character 18: 0 > 1 Node 29: character 1: 0s I. > character 12: 0 > | Node 31: character 28: 0 2 1, character 44: 0 > | Node 32: character 9: 3 > 2, character 30: 1 > 2. — tet A Node 33 | | . character 3: 0 > | Node 36: character 24: 0 = 1. character 26: 0 > 3, :h l, character 42: 0 => 1 Node 37: character 43: 2 = 1, character 44: | 0 Outgroup taxa Panicum obtusum: character Ө: 1 — О, character 27: 233 ые кл, fiebrigii: — ter 24: Ө — 1. charac- r 26: 0 > 3. chi — => 0. — 10: 0 > l, character 42: ( : не ter 51: 2 > 1, character 52: 0 — | Paspalum cromyorrhizon: character 4: 0 — | P. notatum: no autapomorphies Table 3. Continued. P. almum: character 10: 0 > 1, character 57: | = 0 P acuminatum: character 30: 1 > 0. character 35: | > 0, character 46 =>» 0, * ter 56: Ө => | P. dissectum: oneracios 10: 1 = 2, character 49: 0 > 1. character € = 0 P. sanguinolentum: character 15: e P. erianthum: character 15: 0 > 1. character 18: 0. character 35: | — 0, — дег 55: 1 => Г. — character 4: 0 = 1, character 46: char 54:12 R ee character 55: 1 > 0 P candidum: character 35: | > Ө, character 46: P. racemosum: character 54: | => 12690. ] ze Ingroup taxa: Subg. Ceresia P. lanciflorum: character 56: | = 0 P. cachimboense: character 1: 0 — 1, 1 = 0 P. pectinatum: no ene PAPE: 5 P. cordatum: character 4: А 2a ter 10: 3 > 1 P. imbricatum: character 5: 0 = 2 P. reticulinerve: character 10: 3 > 1, 01 P. aspidiotes: P. goyasense: character 11: | e P. carinatum: character 48: 0 > | P. eucomum: character 48: 0 > 1 P. malmeanum: character 53: | > 0 P. ceresia: ае 35:1 > 0, character 51: character 53: 0. PM E r55:0 = stellatum: characte ‘ter 49; ( . charac ter 50: 0 > 1. О. AN => 0 0 = 1. character 55: character 55: character 15: character 9: 2 => 3. character 50: 0 => 1 1 > 0, character 5 ] шә 0, Р character 56: P. longiaristatum: habitos l: 0s] P. biaristatum: no auté pop S Р S terotrichon: character 26: => 0 P — character 42: 0 > P. trachycoleon: no mgro an 5 Р. жү жө E chang : 1 0. character 26: 2 charact Р, Bd pa no a йан ы ТРИ P. buchtienii: no autapomorphies P. humboldtianum: character 26: 2 — p د‎ character ч @ а. = 1, character 42: 0 = р ОЛ? character 26 2 => 0, character 55: ). character 56: | => ( . character 54: 3 3, character 40: 0 less frequent in the Amazonian forests. Several spe- cies have an Andean distribution: P. humboldtian- um. from Mexico and Mesoamerica to northwestern and central Argentina; P. buchtienii is restricted to mountain slopes of Bolivia and Peru, while P. cym- biforme is found in Mexico and Mesoamerica. A disjunct ынана was observed in the following species: Р. polyphyllum, present in southern and central Brazil, in the states of Rio Grande do Sul and Santa Catarina, southern Paraguay, northeast- ern Argentina, and Uruguay, and also in Colombia 348 Annals of t Missouri — Garden Table 4. List of species of informal groups of Paspalum Eriantha, Dissecta, Notata, and Racemosa used as outgroups in the cladistic analysis and species of Paspalum subg. Ceresia and its sections Ceresia and Pectinata. Outgroups Dissecta Group (sensu Morrone et al., 1996) Paspalum а ? dissectun Eriantha Group (sensu Chase. Ined.) P. erianthum P. guttatum P. sanguinolentum Notata Group (sensu Canto-Dorow et al., 1996) P. almum P. cromyorrhizon P. notatum Racemosa Group (sensu Morrone et al., 1995) P. candidum P. prostatum P. racemosum Paspalum subg. Ceresia section Pectinata P. aspidiotes P. cachimboense F. cordatum P. imbricatum section Ceresia P. lanciflorum P. pec tinatum P. reticulinerve Р. — P. buc P. — P. carinatum P. ceresia P. cymbiforme P. eucomum 1sense P i hon P. humboldtianum P. longiaristatum P. malmeanum P. niquelandiae P. trachycoleon and Venezuela. Paspalum cordatum has been col- lected in southern Brazil, Amazonia, and Colombia. Paspalum ceresia is common in the Andean region from Ecuador to northwestern Argentina, and iso- lated populations grow in central Brazil, in the states of Goiás, Maranhao, Minas Gerais, and Para. Finally, P. trachycoleon grows from Mexico and Me- soamerica to Colombia and Venezuela and reap- pears in southern Brazil. As previously mentioned, a significant concen- tration of species is found in the cerrados of central Brazil. Paspalum biaristatum, P. longiaristatum, P. burmanii, and P. niquelandiae are restricted to ser- pentine soils of the state of Goiás; P. petrense, P. imbricatum, and P. goyasense are endemic species of this state; P. cachimboense was found only in the on the Serra do Cachimbo, state of Mato Grosso. border with the state of Рага, and P. phyllorhachis in Minas Gerais. Another species of subgenus Cer- esta endemic to central Brazil is P. eucomum, pres- ent in Minas Gerais, Sáo Paulo, Paraná, Goiás, and in the Distrito Federal. It is noteworthy to mention this high number of endemic species in serpentine soils, characterized by high concentrations of nick- el. chromium, cobalt, and other minerals (Brooks, 1987), combined with a low percentage of nutrients in these soils. Other endemic grasses were also pre- viously described for this region (see Filgueiras et al., 1993, for a detailed analysis of the area). PHYLOGENETIC ANALYSIS The cladistic analysis using NONA resulted in 40 most parsimonious trees, each of 196 steps, found in 84 of 100 replications, with CI = 38 and RI = 68. Searches to find possible islands yielded no additional trees. Volume 89, Number 3 Denham et al. 349 2002 Paspalum subg. Ceresia Panicum laxum Panicum obtusum E Anthaenantiopsis fiebrigii Paspalum erianthum P. guttatum | Eriantha Group P. sanguinolentum 1 P. acuminatum . | Dissecta Group 1 P. dissectum P. cromyorrhizon n P. notatum | Notata Group — — —1 > P. almum 75 Р. candidum ; | Р. racemosum ; Racemosa Group A P. prostatum P. lanciflorum 7] P. cachimboense P. pectinatum - 2 Р. cordatum a Y 22 P. imbricatum Fd ; Р. reticulinerve Р. aspidiotes = P. goyasense = 2 Р. сагіпаќит Р. сегеѕіа Р. stellatum - P. eucomum » B ов Р. таітеапит a = Ф Р. longiaristatum 5 c l 3 P. biaristatum & P. heterotrichon © Ф | — —— Р. petrense 2 & | P. trachycoleon | 1 Р. phyllorhachis Р. cymbiforme [99] 63 Р. buchtienii Р. humboldtianum Р. polyphyllum Р. niquelandiae 1 ii | | Р. burmanii ال‎ d o Figure 3. ct consensus tree from 40 most parsimonious c — ams found by NONA. Numbers below branches indicate Bremer E numbers above branches indica Figure 4 shows one of the 40 most parsimonious trees. Unambiguous synapomorphies of each node are listed in ' state change occurs just in that cladogram or in all equally parsimonious trees. З. with Bremer support values and jackknife frequencies. The strict consensus is shown in Figure In this consensus tree, Paspalum is monophyletic with Anthaenantiopsis fiebrigii or Panicum obtusum as sister taxa. Next within Paspalum, there is a Table 3. which indicates if a character alues. te jackknife major clade including the Racemosa group and subgenus Ceresia. Relationships of the remaining studied species of the genus, of Eriantha, Notata, and Dissecta groups, are not solved; they are placed in a basal polytomy in the tree. Two different to- pologies are equally parsimonious when solving this polytomy. according to the placement of species of the group Eriantha. The clade comprising groups Racemosa and Cer- esta (Fig. 4, see node 4) is supported in all most 350 Annals of the Missouri Botanical Garden Panicum laxum Panicum obtusum Anthaenantiopsis fiebrigii P. prostatum igure 4. parsimonious trees by character 11 (apex of the ra- chis ending in a naked point) and character 27 (up- per glume 3-nerved). In some of the most parsi- monious trees, character 28 (nerves of the upper glume equidistant) and character 43 (lower lemma 3-nerved) also support this clade. Paspalum subg. Ceresia forms a clade with Ra- cemosa as ils sister group. At node 5 (Fig. 4), char- acters supporting the monophyly of subgenus Cer- esta are: 51 (membranous upper anthecium) and 52 P. acuminatum P. dissectum P. sanguinolentum P. candidum Р. racemosum P. cachimboense P. goyasense One of the 40 most parsimonious trees found by NONA (196 steps le ын Cl = number of nodes in boxes. Synapomorphies of each node and taxa are listed in Table | | Notata Group | P. almum | Dissecta Group | | | P. erianthum | Eriantha Group P. guttatum | | Racemosa Group P. lanciflorum P. imbricatum o 9 a E Se 3 ejeunoeg әс̧ P. reticulinerve P. aspidiotes P. carinatum P. eucomum P. malmeanum P. ceresia P. stellatum еѕәзәо ‘Bans unjedseg P. longiaristatum P. biaristatum P. heterotrichon ‚ Р. petrense | eIsaja) ‘Pes Р. trachycoleon Р. phyllorhachis Р. cymbiforme Р. buchtienii Р. niquelandiae P. burmanii 38, RI = 68) with (upper palea gaping at the apex). Characters 40 (lower lemma with pilose margins) and 56 (upper lemma with unicellular macrohairs) also support subgenus Ceresia in some of the most parsimonious trees. Within Paspalum subg. Ceresia, two major clades appear in the analysis: one containing species of section Pectinata, and the other containing the re- maining species of the subgenus. Section Pectinata (see node 6, Fig. 4) is supported by character 9 Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia P. stellatum P. eucomum P. malmeanum P. ceresia P. longiaristatum P. biaristatum #9, 49, 56, 57 Figure 5. Figure 4. —B. Node 20 of cladogram of Figure (presence of approximate racemes); within this sec- tion, P. pectinatum, P. cordatum, P. imbricatum, P. reticulinerve, and P. aspidiotes are grouped together and are very well supported by the presence of characters 19, 20, 27, 30, and 31 (upper glume winged, cordate, 5-nerved, with flat margins, and marginal nerves not reaching the apex) in all of the most parsimonious trees Fig. cluding species here considered to be in section Cer- The other main — (see node 7, 1). in- esia, is supported, in all shortest trees, by characters 24 (upper glume with pilose margins). 26 (upper glume hairy on the lower half), and ЗО (margins of the upper glume folded at the nerve). Within this clade, two subgroups (Figs. 3, 4) are recognized: clade A, including Р. goyasense, P. carinatum, P. eucomum, P. malmeanum, P. ceresia, P. stellatum, P. longiaristatum, and P. biaristatum, with the following characters as synapomorphies (node 13, Fig. 4): 42 (lower lemma pilose) and 55 (upper anthecium with- out silica bodies). In all of the most parsimonious trees, P. longiaristatum and P. biaristatum constitute a well-supported terminal group (node 19. Fig. 4). both species with an awned upper glume and lower lemma (chars. 21 and 38). Clade B includes P. het- erotrichon, P. petrense, P. trachycoleon, P. phyllor- hachis, P. cymbiforme, P. buchtienii, P. humboldtian- um, P. polyphyllum, P. niquelandiae, am burmanii, and is defined by character 18 (hyaline upper glume) in all trees (node 20, Fig. 4). Two equally parsimonious topologies were found in clade А: one shown in Figure 4 (node 18), where P. stellatum is the sister species of P. longiaristatum P. heterotrichon P. petrense P. trachycoleon P. phyllorhachis P. cymbiforme #40 P. buchtienii P. humboldtianum P. polyphyllum P. niquelandiae P. burmanii Alternative most parsimonious topologies within Paspalum sect. Ceresta.—A. Node 15 of cladogram of and P biaristatum, due to support from characters 15 and 46 (lower lemma with marginal nerves and extension of the margins absent). In the other reso- 5A, P. stellatum, um, and P malmeanum constitute a group supported lution, shown in Figure P. eucom- by characters 9 (two racemes per inflorescence). 49 (obovoid upper anthecium), and 56 and 57 (upper anthecium without macro- and microhairs), with P ceresia as the sister species of P. longiaristatum and P. biaristatum. On the other hand, two equally parsimonious to- pologies were obtained for clade B: in one of them, shown in Figure 4, P heterotrichon, P. petrense, P. trachycoleon, and P. phyllorhachis are monophyletic (node 21), being sustained by characters 13 (spike- lets laterally inserted in the pedicels) and 16 (spike- lets with an annular The alternative topology (Fig. 5B) places P trachycoleon and P. phyllorhachis as sister group to a clade con- sisting of P. cymbiforme, P. buchtienii, P. humbold- thickening at the base). tianum, P. polyphyllum, P. niquelandiae, and P. bur- manii. This clade. with eight species, shares a elabrous margin of the lower lemma (char. 40). Two most parsimonious trees (fit = 413.7 and 199 steps) were obtained using implied weights with PeeWee (trees not included). Regarding the Paspal- um ingroup, the strict consensus of both trees has the same topology as the tree, obtained with NONA (Fig. 3). where group Racemosa is the sister to sub- venus Ceresia. Characters supporting clades in the two shortest trees are the same as in the equal weight analysis, except character 24 for section Ceresia 4). In both most parsimonious trees (node Fig. 352 Annals of the Missouri Botanical Garden from Pee Wee, groups Eriantha, Dissecta, and Notata are more closely related to each other than to the clade group Racemosa + subgenus Ceresia. DISCUSSION Monophyly of Paspalum subg. Ceresia is sustained by the present analysis. Two additional steps are re- quired for breaking down the subgenus Ceresia clade, which has a Bremer support of 2 and a jack- knife value less than 50; similar low values are com- mon in different cladistic treatments using morpho- logical data (Karis, 1995). Only homoplasic characters support this clade including subgenus Ceresia: a membranous upper anthecium (char. 51) is present in nearly all species of the ingroup, but P. niquelandiae and P. goyasense have a membra- nous or chartaceous upper anthecium, whereas in P ceresia it is hyaline. Regarding the outgroups, the upper anthecium is membranous in Anthaenantiopsis fiebrigii, and membranous to chartaceous in P can- didum and P. cromyorrhizon. Character 52 (upper palea gaping at the apex, not covered by the upper lemma) is a synapomorphy of the subgenus but also present in A. fiebrigii. The winged rachis, widely used to recognize the subgenus, is present in many of the outgroup taxa, evidently arising many times, or being lost, in the evolution of the genus. Within subgenus Ceresia two monophyletic sec- tions are recognized: section Peetinata, including Paspalum lanciflorum, P. cachimboense, P. pectin- atum, P. cordatum, P. imbricatum, P. reticulinerve, and Р. aspidiotes, and section Ceresia, including P. goyasense, P. carinatum, P. eucomum, P. mal- meanum, P. ceresia, P. stellatum, P. longiaristatum, P. biaristatum, P. heterotrichon, P. petrense, P. tra- chycoleon, P. phyllorhachis, P. cymbiforme, P. buch- tienii, P. humboldtianum, P. polyphyllum, P. nique- landiae, and P. burmanii. Section Pectinata is weakly supported herein (Bremer support 1; jackknife value less than 50); a similar delimitation of section Pectinata was recently shown by Rua and Aliscioni (in press). Within this section, the group including P. pectinatum, P. cordatum, P. imbrica- tum, P. reticulinerve, and P. aspidiotes is well sup- ported (Bremer support = 4; jackknife value = 81), all species sharing a winged and cordate, 5- or more nerved upper glume, with the marginal nerves not reaching the upper third of the glume. Paspal- um lanciflorum and P. cachimboense are basal spe- cies within section Pectinata, both taxa without a winged and cordate upper glume, but with a lower lemma pilosity similar to other species of section Pectinata. Section Ceresia is also weakly supported herein (Bremer support = 1; jackknife value less than 50). Within the section, Paspalum biaristatum and P longiaristatum are well-supported sister species (Bremer support = 3; jackknife value = 91) but are not distinct from section Ceresia. Therefore, we consider section Biaristata (Filgueiras & Davidse, 1994) as included in section Ceresia, because re- taining section Biaristata within our section Ceresia will render section Ceresia monophyletic. TAXONOMIC TREATMENT Paspalum subg. Ceresia (Persoon) Reichenbach, Consp. Regn. Veg.: 49. 1828. Ceresia Pers., Syn. РІ. 1: 85. 1805. Paspalum sect. Ceresia (Pers.) Nees, Fl. Bras. Enum. Pl. 2: 76. 1829, [as "Paspalus sect. Ceresiae"]. Paspalum sect. Ceresia (Pers.) Schltdl. ex Müll. Stuttg., Bot. Zeitung (Berlin) 19: 326. 1861. nom. — Paspalum subsect. Ceresia (Pers.) Benth., Gen. 3: 1098. 1883. Paspalum sect. Ceresia (Pers.) Pilg., Repert. Spec. Nov. Regni Veg. 26(15): 230. 1929, nom. illeg. TYPE: Paspal- um ceresia (Kuntze) Chase [= Ceresia elegans Pers. ]. PH sect. d F Es iras & Davidse, Novor 1994. Syn. YPE: Paspalum cin le & Davidse Paspalum sect. Pectinata Chase ex Rodr.-Rodr., Ernstia 2(1-2): 22. 1992. TYPE: Paspalum pectinatum Nees ex Trin. Plants perennial, erect or leaning on the vege- ation. Blades rigid, filiform to lanceolate. Inflores- cences with 1 to several racemes, rachis of the ra- — cemes winged, hyaline to membranous, 0.8-10 mm wide. Spikelets 1,4—8 mm long, pilose, occasionally glabrous; lower glume absent; lower palea and low- er flower absent: upper anthecium pale, hyaline to membranous, chartaceous in some species, the pa- lea free at the apex. ‘Twenty-five species growing from Mexico to Ar- gentina, Uruguay, and southern Brazil. Cerea Schltdl., Bot. Zeitung (Berlin) 12: 820. 1854, is a superfluous name for Ceresia. KEY TO DISTINGUISH THE SPECIES OF PASPALUM SUBG. CER l. Upper a НР | 2 1". Ipper glume not winged 6 2(1). Spikelets glabrous 14. P imbricatum 2: Spikelets pilose 3 3(2). Spikelets paired, arranged in four series 3. P. retic iculinen е 3. Spikelets solitary, arranged in two series 1 4(3. Lower lemma she nd ciliate on the maris, not papillose-pilos 1. P. aspidiotes 4”. ie wer lemma денш uously papillose- тов se the margins ме, 5 5(4). Cite short rhizomatous, 35-100 cm tall; sheaths red or brown, shiny at the outer sur- face; racemes 2(3 to 5), 2-8 em long; spike- Volume 89, Number 3 Denham et al 353 2002 Paspalum subg. Ceresia er lanceolate, 5-8 mm long, 2-3.3 mm Б Perennial; upper anthecium not — 16 Wide aaa 19. P. — 16(15). Spikelets solitary; the margins not corky, 5, Culms with long, curved rhizomes, 100— with hairs up to 3 mm — appressed or cm tall: sheaths stramineous or green, not ascendent, irregularly distributed on the shiny at the outer surface; racemes 5 to es margins í 12-16 cm np spikelets ovoid, 4—5(—6) n lo’. еј pi aired, occ asionally the lower one long, 2.5-3.5 mm wide „Р can — ed or aborted; the margins corky, with 6(1) Spikelets with the upper glume and the lowe radiate marginal hairs (ascendent in P cym- lemma awned — А d biforme) 20 Y. Spikelets not awned 0 8 170106). Racemes 2, conjugate; spikelets .1.6-3 mm 1(6) Annual, culms 15-36 « m tall: spikelets 1. E long. the margins densely pilose on the up- 2.2 mm long; awn of the upper glume 6-12 per portion, with hairs up to 2 mm long. mm long. that of the lower lemma 0.3-2 1 shortly pilose toward the base: upper anthe- long : 16. P. P longiaristatum cium as long as the upper glume and lower T! Perennial, culms 70-125 cm tall: spikelets lemma We a H 18 3.8-4.5 mm long; awn of the upper У 1 17’. Racemes 1 to 3, not conjugate: spikelets 3— 7.1 mm long, that of the lower 6 mma 3.8 5 mm * with long hairs on the basal mar- 1.5 mm long |... . P. biaristatum gins, sparsely and shortly pilose on the upper 8(0) Rachis (3.5-)4—10 mm wide __ © 9 portion, occasionally glabrous; upper anthe- 8”. Rachis 1-2.5-3) mm wide 15 cium shorter than the upper glume and lower 9(8) Spikelets solitary — — 10 lemma 19 o". Spikelets — 13 18017). Spikelets | 6-2 mm long 17. P. malmeanum 10(9). Spikelet (4.8—)5 6-7. 2 mm long; anthiecium 18. Spikelets 2.5-3 mm long 10. P. eucomum № to Y length of the spikele 19(17). Racemes l, occasionally 2: rachis of the = 15 P. ыа branches 1.8-2.5(-3.5) mm wide. straight to 10’. Spike let 2.3-3.9 mm lane: аа ‘ium as long subfalcate; upper glume densely pilose on as or slightly shorter than the spikelet 11 the lower third, ra scaberulous: — 11(10). Inflorescences with l (rarely 2) racemes, if 2 glume and lower lemma flat in the upper por conjugate; pulvini with | or 2 papyraceous tion: blades filiform 6. P. carinatum bracts: pedicels with a crown of hairs toward 19”. Racemes (1)2(3), rae dis e the branches 1— the apex: upper anthecium shortly stipitatez 1.6 mm wide, straight; lower glume densely blades filiform 24. P. stellatum pilose in the lower half, otherwise glabrous; П”. Inflorescences with 1 to 7 alternate racemes: upper glume and lower lemma slightly con- pulvini ebracteate; pedicels hispid, without a dades linear Е 11. P. govasense crown of hairs toward the apex: upper anthe- 20. Rac iua of the branches 2-3 mm wide 2] cium not stipitate; blades linear to linear-lan- 20". Rachis of the branches 0.8-1.541.8) mm eolate (occasionally filiform in P ceresia) 12 wide 277 22 12(11). ie his 3.54 mm wide; blades 1.5—1 mm 21(20). — ‘ences 8-12 em Jong, with і ra- wide; spikelets 2,3—3 mm long, not early de- emes; floriferous culms D. em tall: ciduous, with an annular thickening at the blades linear, 5-19(-24) x 0.2—0.5(—0.8) base: upper glume with papillose-pilose em: pseudoligule a ring of hairs hairs toward the — a few of them lon- 0 —— — —— 9. P. cymbiforme ger than the rest; lower lemma with long 21^. Inflorescences 20-25 cm long, with 9 to 11 hairs at the upper margins, otherwise & racemes; floriferous culms 150 cm tall; brous: lodicules : 12. P. НАУ сее hon blades * ‘eolate, 30-35 X 1-1.2 em; pseu- 12’. Rachis 6-10 mm wide blades 3-15(220) doligule absent - v 4. P. burmanti mm wide; spikelets (333.4-3.6 mm long. 22(20). Plants — 120-185 em tall, blades 20— early a without an annular thick- 15 0.8-2.2 ст, inflorescences with 0 ening al the base; upper glume with + equal. ae racemes .....................-- 18. Р — silky, not papillose-pilose, marginal hairs: 22'. Plants (ode 'eous, 30-100 cm tall; blades lower lemma ee all over the surface ez lod- 1-21 em long, 0.1-1.8 cm wide, inflores- icules ahsent P. ceresia cences with | to 11 racemes 2: 13(9). Plants нта spikelets зн 23(22). Rachis pilose on one or both surfaces; blades 2]. кт n * аг to linear-lanceolate, 4-13 cm long, 13’. Plants herbaceous: apie lets pilos .1—0.6 em wide; upper glume with une ja 14(13). Culms unbranched, 70-90 cm tall: — "s 3 marginal hairs; lower lemma with — 6 mm wide; rachis (6.5—)7—7.5 mm wide: up- cal margins |... 22 P plum per glume long-acuminate; lower li mma pi- 23"; Rachis дарш blade 's lanceolate, 5-21 cr lose on the — 'r half, not sulcate long. 0.5-1.8 cm wide; upper Jin = — 20. P. mum marginal hairs evenly distributed’ all of the 14”. Culms ене hing, — m tall: blades 0.5-1.5 same length; lower lemma with glabrous api- cm wide; rachis 4-6 mm wide; upper pus cal ma carinatum. nous, brown; pseudoligule absent, collar pilose. Blades lanceolate, 30-35 em long, 1-1.2 cm wide, rounded, the apex acuminate, mostly basal, the dis- al ones small, flat or with involute margins, the basal ones papillose-pilose, midnerve conspicuous. Peduncle subincluded, cylindrical, sparsely pilose Inflorescences terminal, 20-25 cm long. 5-12 cm wide; main axis 15-20 em long, glabrous: racemes 9 to 11, ascendent and divergent, alternate, the basal ones 7-9 em long; pulvini densely pilose with short and long hairs; rachis of the racemes 2.7-3 mm wide, flat, foliaceous, ending in a naked point, occasionally in a spikelet, glabrous, green, the mid- nerve Conspicuous and with anastomosed veins, the margins scabrous: pedicels unequal, terete, sca- brous; spikelets paired, imbricate and arranged in . series. Spikelets ellipsoid, 3.3—4 mm long, 1-1.2 mm wide, pilose. Upper glume 3-nerved, acumi- nate, membranous, pilose, more densely so toward the base. the margins corky, hairy, with spreading hairs up to 2 mm long. Lower lemma as long as the A 360 Annals of the Missouri Botanical Garden P. ceresia m P cymbiforme e P cordatum ж Р burmanii А Р cachimboense + Р. goyasense — — — i-e 20) \ \ \ \ \ \ | \ | \ \ \ \ à 200 400 600 800 1000 km о Ra 200 300 400 - Ын miles \ d by Hendrik K \ \ — жане A MRNA динанын... анаа m эрин Figure 10. Distribution of Paspalum burmanii, P. cachimboense, P. ceresia, P. cordatum, P. cymbiforme, and P. goyasense. upper glume, 3-5- 0.8 mm wide, plano-convex, papyraceous, shiny, with simple papillae all over the surface; upper lemma 5-nerved, with silica bodies and prickles on the upper portion, and bicellular microhairs and macrohairs all over the surface; lodicules 2, 0.2 mm long, conduplicate, hyaline; stamens 3, anthers 2 mm long; stigma plumose. Caryopsis obovoid, mm long, 0.7 mm wide, brown; hilum punctiform, basal; embryo less than Y the length of the cary- opsis. Distribution and habitat. Endemic of Nique- lándia, in Goiás, central Brazil, on serpentine soils Paspalum burmanii is related to P. niquelandiae, which can be separated by its spikelets 2.3—3.3 mm long, with the upper glume almost glabrous or shortly hispid on the surface, racemes 8 to 30, ra- chis of the racemes up to 1.5 mm wide, and pseu- nerved, acuminate, membranous, scabrous. Upper anthecium ellipsoid, 3 mm long, doligule a ring of hairs 5-8 mm long. It is also related to P. humboldtianum, which differs by its plants 30-100 em tall, culms branching at the low- er and middle nodes, otherwise unbranched, blades 5-21 em long, 0.5-1.8 em wide, evenly distributed along the culms, and inflorescences smaller with 2 to 6 racemes, the rachis of the racemes 1.2-1.5 mm wide. Representative specimen. BRAZIL. Goiás: Niquelán- dia, a 10 km, Filgueiras 2477 (SI). э. Paspalum cachimboense Davidse, Morrone & Zuloaga, Novon 11: 389, fig. 1. 2001. TYPE: Brazil. Mato Grosso: Mun. Colider, es- trada Santarém-Cuiabá, BR-163, km 762, Serra do Cachimbo a 30 km de Guaranta, 92355, 54°55'W, 19 Apr. 1983, M. N. Silva, I. L. Amaral, J. Lima, O. P. Monteiro & J. Coélho 24 (holotype, MO!). Figure 10. Volume 89, Number 3 Denham et a 361 2002 Paspalum subg. Ceresia Caespitose annual: floriferous culms 20-70 cm кин — Hochst. ex Steud.. PI. Glumac. tall, 1.5 em diam., red-tinged, geniculate, branch- 1855 [1953]. — carinatum Humb. А ing at the middle and upper nodes; internodes hol- И) a bes fs o^ 2) m E ~ — öll, ın Mart., А ras ). y к r x 7 5 ‹ А y ho: a € 2 Кек * low. glabrous: nodes brown, glabrous. Sheaths 3-8 rinam. Without locality, E ji Hostmann 1 306 em long, usually shorter than the internodes, dense- (holoi; ype, Pt isotypes, B!, BM! ! photo, SIL ly pilose toward the distal portion, otherwise gla- brous, the margins membranous. Ligules membra- nous, 2—3 mm long; pseudoligule absent: collar papillose-pilose. Blades linear to linear-lanceolate. 14-19 em long. 0.4-0.6 cm wide, attenuate at the base, the apex acute, densely papillose-pilose on both surfaces and along the margins, midnerve manifest. Peduncles exserted, cylindrical. glabrous. Terminal and axillary inflorescences present, 5-8 em long, 2-5 em wide: main axis absent: terminal inflorescences with | to 4 approximate racemes, ax- Шагу inflorescences usually with a single raceme: pulvini pilose, usually with a membranous bract гир to 5 mm long: rachis of the racemes winged, Y —§ 9 em long, 2-3 mm wide, ending in a naked — glabrous, the midnerve manifest and the margins membranous. nerveless; pedicels flat. glabrous: spikelets solitary, imbricate and arranged in 2 se- ries. Spikelets long-ellipsoid, 4—4.3 mm long. l- 1.1 mm wide, pale, pilose, occasionally glabrous. Upper glume membranous, rounded at the base. the apex acute, 3-nerved, with one central and two sub- marginal nerves, pilose at the base, the apex shortly ciliate. the rest of the surface glabrous. Lower lem- ma glumiform, as long as the upper glume, rounded at the base, 3-nerved, the margins with mixed pa- pillose and not papillose hairs. the apex shortly cil- iate. Upper anthecium ellipsoid, 2.2-2.5 mm long. 0.8 mm wide, shorter than the spikelet, stipitate. membranous; upper lemma with simple papillae. macrohairs, and bicellular microhairs toward the distal portion: lodicules 2. conduplicate. hyaline: stamens 3. anthers 1.5 mm long: stigmas 2. plu- тозе. Caryopsis not seen, Distribution and habitat. Endemic to Serra do Cachimbo in central Brazil, where it is known only from the type collection. Paspalum cachimboense is a member of section Pectinata and closely related to P. lanciflorum, which differs by its perennial habit, with simple and erect culms. rachis of the racemes (4—)5—7 mm wide, spikelets (4.8-)5.6-7.2 mm long. and upper anthecium 3.2—1.7 mm long. 6 Paspalum carinatum Humboldt & Bonpland ex Flüggé. Gram. Monogr., Paspalum: 65. 1810 [as “Paspalum ipsi TYPE: Co- lombia. Without locality. Н. A. von Hum- boldt & A. J. A. am ae s.n. (holotype, B-W not seen: isotypes. BM! photo. SI. US- 2942176!). Figure 9. O0-102047!. P!, US- Aen т де» spissum = Swale "n. е 14: 358. 1967. уп. TYPE: Brazil. Магапһао: tonio de Balsas, 20-25 Mar. 4050 (holotype, US-1612651! photo. SID. Carolina to San An- 934 R. Swallen xo Short-rhizomatous perennial; floriferous culms (25-)40-70 ст long, finely striate, glabrous: nodes brown, glabrous to shortly pilose. Sheaths 5-13 cm long. usually shorter than the internodes, glabrous to hirsute, the margins membranous. Ligules mem- branous, ca. 0.6 mm long, brown, glabrous: pseu- doligule absent or present, when present a ring of white hairs. Blades filiform, 5-20 cm long, 0.5-1(— 3) mm wide. involute, erect, mostly basal. rounded at the base, the apex setaceous, papillose-pilose on both surfaces, more densely so toward the base of the adaxial surface, the margins involute, ciliate. Peduncles up to 35 em long. subterete. glabrous. Terminal inflorescences 6-12 em long, with a mem- branous bract often at the basal node: main axis absent or up to 3 em long, flat, glabrous: pulvini glabrous: racemes 1(ог 2). alternate when 2 ra- cemes are present, straight or subfalcate. 5-12 ст 1.8-2.5(—3.5) ending in a naked point, glabrous, the long: rachis of the racemes winged. mm wide, midnerve conspicuous, the margins membranous to hyaline, partially covering the spikelets: pedicels short. flat: spikelets solitary, densely imbricate in 2 series. Spikelets long-ellipsoid. 3.3-5 mm long. 1— dorsiventrally compressed. pale. 1.2 mm wide. densely pilose at the basal portion. the apex acute r rounded. Upper glume as long as the spikelet. membranous, 3-nerved, densely pilose on the lower portion. with white hairs up to 3 mm long. shortly pilose on the apex, otherwise scabrous. Lower lem- ma glumiform. slightly shorter than the upper glume, 3-nerved, pilose on the lower portion, the hairs shorter than those of the upper glume, shortly pilose and scabrous at the apex, the rest of the surface glabrous. Upper anthecium long-ellipsoid. 2.7-3.6 mm long. | mm wide, 0.6-1.8 mm shorter than the upper glume, plano-convex, flattened to- ward the distal portion, membranous, shiny, finely papillose with simple papillae all over the surface; upper lemma with macrohairs, bicellular micro- hairs, and prickles at the upper portion; upper pa- lea with bicellular microhairs and macrohairs at the lodicules 2. 0.2 apex: mm long. conduplicate, hy- 362 Annals of the Missouri Botanical Garden aline; stamens 3, anthers 1.6—1.8 mm long; stigmas 2, plumose, of lateral emergence. Caryopsis ovoid, 1.6 mm long, 0.8 mm wide; hilum elliptic, embryo less than half the length of the caryopsis. Iconography. Smith et al.. Fl. Il. Catarinense: 918, fig. e-g. 1982. Renvoize, The Grasses of Ba- hia: 214, fig. 79, A. 1984. Judziewicz, Fl. Guianas: 469, fig. 83, D. 1990. Rodríguez, Ernstia 8(2-3): 37. 1998 Common names. oasis ciel (Brazil); Pa-mac” (Guyana). Chromosome number. = 10 (Davidse & Pohl, 1978); n = 40 (Davidse & Pohl, 1972). Distribution and habitat. Trinidad and Tobago, "Garcita" (Venezuela); * Nicaragua, Colombia, Venezuela, Guyana, Surinam, Brazil, and Bolivia. It grows in humid savannas on sandy, rocky, or limestone soils, in open places with periodical fires, between 100 and 1600 m. Paspalum carinatum is related to Р. goyasense and P. stellatum. Paspalum goyasense has inflores- cences with two distant racemes, rachis of the ra- cemes 1-1.6 mm wide, ending in a developed spikelet, spikelets with the upper glume densely pilose in the lower half, otherwise glabrous, upper glume and lower lemma slightly convex, and blades linear, 8-20 cm long, 0.2-0.4 ст wide. Paspalum stellatum has rachis of the racemes wider [(4—)5— 10 mm]; when two racemes are present, they are conjugate, upper glume and lower lemma with long ciliate margins, with hairs up to 4 mm long, and the upper anthecium is 0.4—1 mm shorter than the upper glume and lower lemma. A rudimentary membranous bract is common in the basal portion of the racemes in Paspalum car- inatum, as was pointed out by Sendulsky and Bur- man (1978). Swallen (1967) related P spissum with P. cari- natum, characterizing the former by its densely pa- pillose-pilose blades 2-3 mm wide, and appressed to the culms. After examination of many specimens and the type collections, we conclude that there is a notorious variation in these characters: filiform blades are mixed with linear blades in several specimens, such as Valls & Silva 8340, Irwin et al. 27667, and Swallen 3687, 4078; others have fili- form blades with papillose-pilose hairs. e.g., Dav- idse & Huber 15436 and Blydenstein & Saravia 1082. Therefore, P. spissum is here considered a synonym of P. carinatum. Although blades are predominantly basal, cau- line blades were present in several specimens. This is probably due to the absence of periodical fires, which are associated with its savanna habitat. Representative specimens. BOLIVIA. Beni: Prov. Vaca Diez, Riberalta, Beck 20564 (K). Santa Cruz: Prov. Ve- lasco, San Ignacio, Bruderreck 2 (K, LPB, SI); Velasco, 5 km E of Comunidad Carmen del Ruiz, Killeen 2823 (LPB, MO, US). BRAZIL. Amapá: Macapá, Black & Froés 51— 12291 ips Mun. Calcoene, Mori & Cardoso 17298 (MO). Bahia: 15 km S of intersection of ue BR-020 and Rio Roda Ve lha, Davidse et al. 12 M MO); 51 km E of Bar- reiras along Hwy. BR-242, ia et al. 12124 (MO). Distrito Federal: Brazilandia, Allem & Vieira 1558 0); km from Brasília on — road, Clayton 4804 (K, US). Goiás: Serra pc ca. 15 km S of Goiás Velho, Anderson — (COL, К, MO); Barra da Lagoa For- mosa, Glaziou 22547 (G, K, P, W); Formosa, Glaziou 22556 (К, Р, US, W). Maranhão: Barra do Corda to Gra- jahú, d 3643 (Р); Barra do Corda to Grajahú, Swal- pie 3 (US); Carolina a San Antonio de Balsas, Sw а — Mato Grosso: Mun. Amambaf, 12 бу -Iguatemi, Allem & Vieira 2007 (МО); са. "270 km N of Xavantina, Ramos & Sousa 182a (K, US). Minas Gerais: Metallurgica, Serra de Ouro Branco, Chase 10306 S): Serra do Espinhaço, 8 km E de Diamantina, /rwin et al. 27667 (MO). Para: Belterra, beira do Lago Jurucuf, — — . Ponta Grossa, (К, P, SI, US). Rio Grande do Norte: Estremoz to Natal, Swallen 4762 (US). Roraima: along Boa Vista—Bonfim road (BR-401), km 40, Coradin & — 696 (MO). Santa Catarina: Rio Capinsel, Dusé 3 IS). Sao Paulo: Casa Branca, — 10595 (MO, US). To- cantins: Пћа do Bananal, Mun. Lagoa da — Re- serva indígena — da Ely et al. 4082 (IBGE, S COLOMBIA. Arauca: Estación de Cravo Norte, Llano 4 (US). Boyacá: | llanos Orientales, al sur de El Yopal, Bh- denstein & Saravia 1082 (COL). Cundinamarca: savan- na of San Martín, 100 mi. SE of Bogotá, Shaw s.n. (US- 1343796). Guaviare: Mun. San José del Guaviare, serranía La Lindosa, Ciraldo.Cafias & López 2566 (MO). eta: Llanos Orientales, entrada de la Serranía San Mar- PE ). Vichada: — ls El Tuparro, entre , El Tapón, Daniel 60 (COL). . Weruni- luni. savanna, Abraham 83 (К, US); Naroni District, Lethem, 4 mi. N of — >m, Irwin 739 (US). NICARAGUA. Zelaya: ca. 5 km S of Bilwaskarma on road to Puerto Cabezas. Pohl & — 12274 (MO). TRINIDAD AND TOBAGO. Erin savanna, S of Point For- tin, Soderstrom 1124 (US). VENEZUELA. Amazonas: 3— 6 km N de S sien sobre ruta a Puerto Ayacucho, Dav- idse et al. 16775 ( Anzoátegui: Distr. Bolívar, just S of El а Fila El Purgatorio, Davidse & Gonzdlez 19317 (MO); Distr. Freites, Burro trail betw. San Durrial & Los Pajaritos, Davidse & Gonzdlez 19802 (MO). Apure: S of San Fernando de Apure. Alba 53-96 (US). Bolivar: Dpto. Heres, margen izquierda de la Quebrada Aguas Ne- gra, 10 km al E del tepuy El Zumbador, Aynard 6019 (MO). Guárico: Estación Biológica Los E anos, del M. | C. Calabozo, Aristeguieta 4219 — 05 NWN of San Felix & Rio Orinoco along Hess turín, Davidse et al. 4615a (MO). Zulia: DE. — 9 km S of the Machiques intersection along the Maracaibo- La Fria Hwy. (Hwy. 6), Davidse et al. 18355 (K ) los ríos Tomo y El Tupa ANA Volume 89, Number 3 2002 Denham et al. 363 Paspalum subg. Ceresia 7. Paspalum ceresia (Kuntze) Chase, Contr. U.S Natl. Herb. 24: 153. 1925. Paspalum mem- branaceum Lam., Tabl. Eneyel. 1: 1791, hom. Шер. Ceresia elegans Pers., Syn. Pl. 1: 85. 1805. Ceresia membranacea wd P. Beauv., Ess. Agrostogr.: 9 y 171, t. 5. f. 4. 1812. Paspalum elegans (Pers.) — & Schult., Syst. Veg. 2: 290. 1817, comb. Шер. Panicum ceresia Kuntze, Revis. Gen. Pl. 3: 360. 1898. TYPE: Peru. Without locality. without collector (holotype, P-LA! photo, SI; S-2855815!). Figures 10. | isotypes, Р!, US Каре — eum Lam. var. pir iia Doll. ‚ Fl. Bras. 2(2): 94. 1877. TYPE: Bolivia. La Par айй, 2600-2700 т. Feb.-Apr. 1861. б. Man- don 1255 — designated here, №1: isolecto- types, BM!, СІ, КІ, P! — dona м ин Lain var. не Doll. i 2: 94. 1877. TYPE: Brazil. campos between Natividade and aoe eição, 3. Gardner 4029 (lectotype, designated here В!; isolectotypes, BM!, G!, K!, PL. US-80085!. WD). .. lart.. . bras. Golds: ‘eb. D d — м ~ Caespitose, short rhizomatous perennial; rhi- zomes arcuate, covered by pilose cataphylls: florif- 10-80 cm tall, 0.1-0.2 cm diam., branched or unbranched; internodes 4—12 cm long. erous culms terete, hollow, glabrous; nodes glabrous. Sheaths 3— 12 cm long, glaucous, glabrous to sparsely pilose on the distal portion, the margins membranous, gla- brous. trun- cate, glabrous: pseudoligule absent, collar glabrous. Ligules 1-2 mm long, membranous, Blades filiform, linear-lanceolate or lanceolate, 7— 21 em long, 0.3-1.5(-2) em wide. flat or with in- volute margins, rigid, slightly divergent, glaucous, glabrous to sparsely papillose-pilose toward the base of the adaxial surface, the margins glabrous to sparsely papillose-pilose. Peduncles subincluded to exserted, up to 35 em long. cylindrical, glabrous. Terminal inflorescences exserted, 4-12 cm long. 2— 10 em wide; main axis absent or up to 7 em long. triquetrous, glabrous; pulvini glabrous or shortly pi- lose, without papyraceous bracts; racemes (1)2 to 7. ascendent, divergent and alternate, distant: ra- chis of the racemes 2-7 cm long, 6-10 mm wide. ending in a naked point, glabrous, green or purple at the middle portion, hyaline at the margins, these brown or ferruginous; pedicels up to 0.4 mm long. flat. hispidulous; spikelets solitary and imbricate in 2 series. Spikelets long-ellipsoid, (3—)3.1—3.0 mm long, 1.2-1.4 mm wide, acute, villous, early decid- uous: upper glume and lower lemma subequal. Up- per glume hyaline to membranous, 3-nerved with one central and two marginal nerves, densely pilose at the base and along the margins, the hairs silky, ascendent, up to 4 mm long. Lower lemma glumi- form, hyaline, pilose, with 2 submarginal nerves. Upper anthecium long-ovoid, 2.6-3.6 mm long. 1— 1.2 white, scabrous and pilose toward the apex: upper mm wide, not stipitate, hyaline and shiny. lemma with prickles, macrohairs, bicellular micro- hairs, and silica bodies at the apical portion, upper palea with bicellular microhairs and macrohairs at the apex: lodicules absent; stamens 3, anthers 2.4— 2.8 mm long. Caryopsis obovoid, 1.2-1.6 mm long. embryo Y the 0.8 mm wide; hilum punctiform., length of the caryopsis. Distribution and habitat. Ecuador, Peru. Brazil, Paraguay. Bolivia, and Argentina. in rocky moun- tain slopes and open fields on sandy and rocky soils. between 300 and 3000 m. Paspalum ceresia is related to P. stellatum, a spe- cies with filiform leaves 8-27 em long, 0.1—0.3 em wide, 1 (occasionally 2) racemes, | or 2 papyra- ceous bracts in the basal pulvini, radiate hairs on the pedicels, lodicules present, and upper anthe- cium stipitate. Paspalum ceresia is also similar to P. trachycoleon, a species from Central America, Venezuela, Colombia, and southern Brazil. which is characterized by the rachis of the racemes 4—0 mm wide, spikelets paired, with an annular thickening al the 3- nerved. А single collection was examined from Paraguay, base. and lower lemma membranous. Isabell s.n. at Р, a specimen without any information about collection date or locality; it is possible this specimen. was mislabeled, considering the distri- bution range of the species. There is considerable variation in the leaf shape, from filiform to linear and rigid, up to 0.3 cm wide. in the specimens Oliveira et al. 305, 598, Pinheiro & Carvalho 142. ‚ 625, Gardner 4029, Brooks et al. 125, Eiten & Biten 4418. 3777, 10789, and Swallen 3772, single specimen, e.g., Souza et al. 5548, Filgueiras 2280, as in filiform and lanceolate in a or with lanceolate leaves, up to 2 cm wide, Doell (1877) described variety aequiglume and variety inaequiglume of P. membranaceum on the basis of variation of the upper glume and lower lemma length, a variable character according the degree of maturity of the spikelets. This author cit- ed the specimens Rhodé s.n. and Mandon 1255 as syntypes of variety aequiglume, of which Mandon 1255 agrees with the protologue and is here se- lected as lectotype of this taxon. Also, Doell (1877) cited Pohl s.n. and Gardner 4029 as syntypes of variety inaequiglume, of which Gardner 4029 is here chosen as lectotype of the variety. 364 Annals of the Missouri Botanical Garden Figure 1 l. Paspalum ceresia.—A. Habit. —B. Detail of ligule. С. — of the rachis. —D. Spikelet, — view. — pikelet, ventral view. =. Upper anthecium, dorsal view. —G. Up pper anthecium, ventral view. —H. Upper lemma. —1. C: агу ро, scutellar view. —J. Caryopsis, hilum view. (A-H based on Hunziker et al. 10354. Sl: 1, J based on i et al. 22135, SL.) Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia ARGENTINA. ee re Santa Bárbara, El F . Cabrera & Fabris 22730 (US Dpto. Que ‘a, Sier 2 Zenta, Venturi 8320 4 LIL. MO, P, SI, W). Salta pt Candelaria, Unquillo, Schrei- ter deso "a IL): Dpto. ( Guac hipas, Alemania, Venturi 9905 ( n Pedro de Colalao, Dinelli 12876 (1. Т ). BOLIVIA. € анча : 15 km S of Padilla, Ren- voize & Cope 3870 (K, | Cochabamba: Prov. Quil- ac 45 km А Oruro, Beck 4025 (US). Potosí: Prov. Charcas, — L ama — Mostacedo 314 (MO Santa Cruz: Prov. Caballero, 15 km arriba de San — de Portero, Killeen ‘ iuh 4152 (MO); Prov. F lorida Es Río Mew 3 km N of hwy. at Yerba ns na, Nee 403 (К, MO arija: d Fiebrig 2664 (SI). B RAZI. Goias: ES . Buracáo, ca. 35 km (bs road) N of Goianésia, Brooks et al. 125 (К, MO, NY): Niquelândia, ca. 18 km W e Niquelándia, Filgueiras 2280 (MO, US): Niquelan- dia, 2 km da e y , : de Chao para o Macedo Velho, Oliv- Mun. Niquelándia, Macedo. ca. 0. ша, Oliveira et al. 508 y S). le Eales region, Kite n ¢ Representative — W). Tucumán: 5 eira et al. 30: km abaixo da mina de Maranhao: Mun. — K, P). iia Gerais Retiro ca. 16 km и» { 5548 (К ara: Estreito-Maraba, km | " 19 ‚ре ro & Carvalho 142 (US); km 2, Pinheiro & Со, 2 517, 025 ECUADOR. Azuay: valley of the Río Paute. be- tween Paute and Cuenca, between Ríos Azogues & Gual- aceo, Camp 2332 (K, P, US, W) soja: road San Pedro de la Bendita (W of C O asne, Ollgaard et al. 90721 (MO): 6 km W of Colaisaca on road to Sozoranga. Peterson & Judziewicz eon (К. MO, US). PARAGUAY Without de ‘partment: without locality, Isabell s.n. (P). PERU. Apurimac: Prov. Abancay, Cunyac, Chávez 3217 (MO). Cajamarea: Prov. kl Ti la carretera. Coc nO (GSL US); Dpto. Cajamarca, 4 km f Cele ndin on road to оны & Wright 519 © )). Cuzeo: Pep. unc Mollepata, Vargas 19 (M A). Huancaveli- — 1020 (US). ea: entre Izcuchaca y Acoria, 8. Paspalum cordatum Hackel, Ark. Bot. 9(15): 1910. TYPE: Brazil. Paraná: der Nähe des Stüdtchens Рома Grossa, am Rande eines Sumples, 7 Ene. 1904, P. K. H. Dusén 3248 (holotype, W!: isotypes. Gl. US- 2942131!). Figures ТО, 12. Rio Tibagy in Robust, rhizomatous perennials with long. arcu- 100—130 cm unbranched: internodes finely ate rhizomes: floriferous culms erect, tall, 0.4—1 em diam., glabrous; nodes compressed, gla- striate, hollow, brous, brown. Sheaths usually longer than the in- ternodes, the lower ones overlapping. loose, densely hirsute on the basal portion or glabrous. the inner surface shiny. brown to. red-tinged, the margins membranous. Ligules membranous, 1-1.3 mm long. brown, glabrous: pseudoligule a ring of dense. rigid hairs up to 6 mm long; collar densely pilose. Blades linear, 10-50 cm long. 0.5-1.1 em wide, flat. the margins involute, ascendent and rigid, densely vil- lous on both surfaces, attenuated at the base. the apex beaked. Peduncles exserted, up to 45 em long. subterete, glabrous. Terminal inflorescences exsert- ed. 7-18 cm long, 6-18 em wide, subdigitate: main axis up to 4 em long, wavy, glabrous to sparsely pilose; pulvini pilose; racemes 5 to 10, 12-16 em long, ascendent, alternate to subopposite, ending in a naked point; rachis of the racemes flat, 1.8-2.4 (-3.2) mm wide, glabrous and green, the margins shortly winged, scaberulous; pedicels short. gla- brous; spikelets solitary, densely imbricate in 2 se- ries. Spikelets ovoid, 4—5(—6) mm long. 2.5—3.5 mm wide, dorsiventrally compressed, pale, winged, cor- date, the apex acute. Upper glume as long as the spikelet, membranous, glabrous, cordate, 5-nerved, with 3 central nerves reaching the apex, and 2 lat- eral nerves only notorious at the base. Lower lemma 3.4—3.6(—5 nerved, the margins with rigid and thick papillose- mm long, | mm wide, subcordate, 3- — pilose hairs, not winged, shortly pilose. Upper an- thecium ellipsoid, membranous, 3-3.2 mm long. 0.8 mm wide, shorter than the spikelet, plano-con- vex, pilose on the upper margins, pale: upper lem- ma with simple papillae all over the surface. and bicellular microhairs, macrohairs, and silica bodies toward the apex: upper palea with simple papillae. bicellular microhairs, and macrohairs on the upper portion: lodicules 2, са. 0.4 mm long. conduplicate. hyaline. glabrous: stamens 3, anthers 1.6 mm long: lateral. Caryopsis not styles 2. stigma plumose, seen. Distribution and habitat. This species grows in southern Brazil. in the states of Amazonas. Paraná. Goiás. Minas Gerais, and Sáo Paulo. also in Colom- bia. in the states of Meta and Vichada. in open areas, lowlands, and swamps. Rodríguez (1998) considered Paspalum corda- tum as a synonym of P pectinatum without indi- cating the criteria for such a decision. Both species share spikelets with the upper lemma not winged and with conspicuous rigid, papillose-pilose hairs on the margins. Nevertheless, we consider P. pec- tinatum a different entity by including specimens —8 cm long. per inflorescence. Additionally, plants, which with 2. occasionally 3, conjugate racemes, grow in savannas or open fields on sandy or rocky soils. are smaller and more densely caespitose than those of P cordatum. with lower sheaths burnt by fire. Paspalum cordatum is also related to P. imbri- catum, this species with glabrous spikelets and low- er lemma winged. Note that in the specimens Blydenstein 939 and Cabrera 2278, the rachis of the racemes is 3—3.2 mm wide. with crenate margins, and spikelets are bigger, 5.2-6 mm long. eno 366 Annals of the Missouri Botanical Garden 4 — ZZ — — > ЕРІ FZ LIES OS P". IA мер, — > N ^ 29 — Da SR Paspalum cordatum.—A. Habit. —B. Detail of ligule. —C. Portion of the rachis. —G. Upper anthecium, ventral view. — US Figure 12. —D. Spikelet, — view. —E. Spikelet, ventral view. —F. Upper anthecium, dorsal view. . Up lemma, lodicules, and gynoecium. (Based on Hatschbach & Renvoize 5276, US.) Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia Representative specimens. BRAZIL. Amazonas: ا‎ Hwy.. 53 km W de Aripuana River, i “ = 8 (MO, US- AR 34). Goiás: Jataí, cedo T — Gerais: without locality, с 1699 "à 1 Turma, in paludosis, Dusén 9109 (BM. US): н Dusén 16174 (BM, G, SI, US); Mun. Campo — Serra S. Luis, Hatschbach et al. 12097 (К. MBM): V а En Bul Serra * Fur rnas, Smith et al. 14579 (K. PS ; Mun. - Rio Funil. Smith et Py 11858 (К. hne Brade 19638 (US). Sio Paulo: Estação Flo- tal de Paraguaqu Paulista, 6 km N of city, Clayton 4651 (К. US). COLOMBIA. Meta: Llanos orientales, Hato Po- zones, Blydenstein 939 M ~ MO). Vichada: Caño Ur- imica, Cabrera 2278 (Cl 9. Paspalum еа E. Fournier, Mexic. m. 2:5. TYPE: Mexico. México: San Pablo, А "s jose. 226 (lectotype. desig- nated by Chase (1929: 20), C not seen). Figure Shortly rhizomatous perennial; floriferous culms 60-90 cm tall, branching at the middle nodes or simple: internodes terete, striate. glabrous: nodes brown, covered by ascendent white hairs. Sheaths longer than the internodes. the margins ciliate at the lower portion. Ligules brown. membranous, 1-1.2 mm long: pseudoligule a ring of hairs. Blades linear, 5-19(-24) cm long. 0.2— 0.5(-0.8) em wide, flat, glabrous or shortly pilose on the adaxial surface. the margins with tuberculate striate, glabrous, — hairs up to 5 mm long, shortly pseudopetiolate, pseudopetiole with short, white hairs. Peduncles exserted, Terminal inflorescences 8—12 cm long, 3—4 cm wide: axis absent or up to 5 em long, shortly hispid or 3-8 cm long. alternate. ending in a naked point: rachis of the racemes 2-2.8 mm wide, green and tinged with purple. the midnerve manifest, wings short, hya- 15-20 cm long, terete, pilose. main glabrous: pulvini pilose; racemes 1 to 4, line, nerveless: pedicels ca. 0.3 mm long. pubes- cent, spikelets paired and imbricate in 2—4 series, usually the lower spikelet aborted. Spikelets long- | mm wide, acute. pilose. ellipsoid, 3 mm long. pale. Upper glume hyaline to membranous, 3- nerved, with one central and two submarginal con- spicuous nerves, the margins corky, with long, + ascendent hairs. up to 3 mm long. and with small hairs on the basal portion. Lower lemma as long as the upper glume, membranous. 3-nerved, shortly pilose toward the apex. otherwise glabrous, de- pressed on the basal portion. Upper anthecium long-ellipsoid, 2.2 mm long. 0.8 mm wide, membranous: upper lemma 3-nerved, finely papil- pale. lose and with prickles, silica bodies, and bicellular microhairs at the apex: upper palea papillose. with macrohairs and bicellular microhairs at the apex: lodicules 2, 0.2-0.3 mm long, conduplicate, hya- line: stamens 3, anthers 1.5 mm long: stigma plu- mose, lateral. Caryopsis not seen. Iconography. Chase, Contr. U.S. Natl. Herb. 28(1): 21, fig. 4. 1929 Distribution and habitat. Mexico to Nicaragua, on mountain slopes on rocky soils. between 750 and 1300 m. This species differs from P. humboldtianum and P. polyphyllum by the rachis 2-2.8 mm wide (vs. 1.2-1.5 mm in Р. humboldtianum and 1-1.2 mm in P. polyphyllum), and by the presence of hairs at the apical portion of the lower lemma. The specimen Davidse et al. 31637 has 2 to 7 Found from southern racemes per inflorescence. еги — EL SALVADOR. La Lib- ertad: Eggler 639 (US). Santa Ana: Mun. Candelaria de la Frontera, — El Vupe, Linares 3663 (MEXU). GUA- MALA. Zacapa: lower slopes of Sierra de las Minas, along trail above Río Hondo. Steyermark 29550 (US). HONDURAS. Distrito Federal: Llamapa, SW of Talan- ga. y um Santa Regina de Archaga. Pohl & Gabel 13531 (MO). Francisco Morazán: 20 km N of Pespire "pie Hwy. | to Tegucigalpa, Davidse & Pilz 31637 (MO): Mesas, Swallen 10809 — US). 10747 S). Olancho: El Espino, 8 km NE de Catacamas, En- — 23 (MO). MEXICO. Consoquitla. Liebmann 2 (MO): Frontera dg 6-8 km E C cea aa pose road to Ciudad Cuauhtemoc, 39106 (MEXU). Chiapas: "Mir xtapa, along Mexican Hwy. 190 in the т атап Paraje of Muctajoc, Breedlove 13822 (MEXU, ay bake 532 (BA. * b. NICARAGUA. Chinandega: Vol- sita. Montaña del Cielo, 12742' N. 06758 W. Rueda e d 1200 (MO). 10. dene eucomum Nees ex Trinius, Sp. Gram. l: pl. 110. 1828. Pas c eucomum Nees, "Y ie Enum. РІ. . hom. illeg. TYPE: Brazil. W — me Ай» r ‘alo isotypes. BM!, US-285 1605). s.n. (holotype, LE not seen: photo, SI!. P!, US-2854694!, Figures 13. 14. Paspalum — Nees ex Trin. var. — Doll, Mart.. Fl. Bras. 2(2): 65. 1877. TYPE: Brazil. Minas Gerais: Lago Santa, Ё. н аа s.n. jest nol ated; isotype, US-28546% Paspalum plende ns Hack.. oin n Z. 51: 238. 1901. azil. Goiás: without locality, A. КОМ. Gla- ziou p 55 51 (lectotype, designated here, W!). Caespitose, shortly — I florif- erous culms 60-110 cm tall, —0.3 cm diam.. erect, ascendent, unbranched, — branch- ing at the middle nodes: internodes 6-15 cm long. nodes glabrous. Sheaths usually striate, glabrous: longer than the internodes, 8-25 em long. overlap- ping. glabrous or papillose-pilose all over the sur- Annals of the 368 Missouri Botanical Garden 1mm — — = — © == == = — ===> A === RI SS — NS WA | INS — 9 1mm y M SSS — — Portion of the rachis. —D. Spikelet, dorsal abit. B. Detail of ligule. —C. Paspalum eucomum.—A. H Figure 13. ». Upper anthecium, ventral view. —H. Upper view, E. Spikelet, ventral view. —F. Upper anthecium, dorsal view. —( palea, lodicules, and anthers. (Based on Chase 9097, US.) Volume 89, Number 3 2002 Denham e al. 369 Paspalum ds Ceresia € P. humboldtianum ж P reticulinerve W P eucomum | \ o 200 400 600 800 1000 km AA) \ — — 0 UM 200 300 400 0 500 eno miles \ 7 repared by Hendrik R ma \ i — — — —— — 100 36 Г 70 — — L— — ——————— Е — Figure 14. face or only at the distal portion, the hairs up to 5 mm long, with one margin ciliate. Ligules membranous, 0.2— 0.5 mm long, truncate, brown, glabrous; pseudoli- gule a ring of white hairs up to 2.5 mm long. Blades 10-30 cm long, 0.2-0.5 em wide, flexuous. rigid and involute, narrow, the apex se- filiform, erect or taceous, adaxial surface with conspicuous papillae and tuberculate hairs, abaxial surface glabrous or both surfaces densely hirsute. Peduncles exserted, up to 30 ст long, subterete, pale or purple, gla- brous. Terminal inflorescences conjugate; main axis absent; pulvini pilose, with a bract usually present: racemes 2. (4-)6-17 cm long. occasionally with a third approximate raceme, ascendent, divergent, ending in a naked point: rachis of the racemes 1.5— 2.54) mm wide, shortly winged, glabrous, green to purple, the margins hyaline, nerveless; pedicels short, flat, and sparsely pilose; spikelets solitary, densely imbricate in 2 series. Spikelets ellipsoid, the margins membranous, glabrous or Distribution of Paspalum eucomum, P. humboldtianum, and P. reticulinerve. 2.5-3 mm long. 1.4-1.6 mm wide, dorsiventrally compressed, acute, densely pilose toward the base and the margins. pale or tinged with purple. U glume as long as the spikelet, nerved, the pper membranous, 3- base and margins covered with white, rigid hairs, the rest of the surface with short and appressed hairs. Lower lemma glumiform, slightly shorter than the upper glume, 3-nerved. Upper an- thecium obovoid. 2.4— mm long, 1.2 mm wide, membranous, pale, shiny, rounded at the apex. smooth: lodicules 2, ca. 0.3 mm long, membranous, conduplicate; stamens 3, anthers 1.5 mm long; stig- ma plumose, lateral. Caryopsis not seen. Distribution and habitat. Restricted to central and southern Brazil in the states of Goiás, Minas Gerais, São Paulo, Paraná, and in the Federal Dis- trict. It grows in cerrados, in open fields on sandy between 700 and 1100 m. Paspalum eucomum is closely related to P. mal- soils, 370 Annals of the Missouri Botanical Garden meanum. The latter species mainly differs by the 20 em long, 0.1 cm diam., terete, hollow, glabrous: spikelets 1.6-2 mm (vs. 2.5-3 mm long): also, P. malmeanum grows in Bolivia and Mato Grosso, on flooded soils, while P. eucomum occurs in central and southern Brazil, usually in drier settings. lt should be pointed out that spikelet size, 2.3-2.5 mm long, is intermediate between both species in Macedo 351, Sendulsky 1239, and Glaziou 17406. These species are here considered as different en- tities until more material becomes available. Chase (ined.) distinguished P. eucomum from P. splendens by the spikelet size (2.5 mm long vs. : mm), rachis width (1.5-2 mm vs. 2-3.5 mm), and general aspect of plants (delicate vs. robust). Upon — study of many specimens, we conclude that there is a gradation in these characters, which are not useful to distinguish both species. Chase (ined.) and Sendulsky and Burman (1980) pointed out that Glaziou 22555 is one of the syn- types of Paspalum splendens. Nevertheless, Hackel (1901) cited, as syntypes, the specimens Glaziou 22552, 22553, and 22554, of which Glaziou 22554 agrees with the original diagnosis and is here des- ignated as lectotype of the species. The rachis of the racemes is exceptionally wide, up to 4 mm, in the specimens Dias et al. 67 and Weddell 2556. Chase (ined.) studied, at B, two specimens col- lected by Sellow, 1238 and 1336, amined by Nees and probably represent isotypes of which were ex- P. eucomum. Duplicates of these collections from US were identified by us as P. eucomum. Representative specimens. BRAZIL. Distrito Fed- eral: Brasília, de Jesus 3 (US); Brasflia, Dias et al. 67 E : re of Lagoa Paranoá, Irwin et al. 15328 (K, MO, US). Goias: without н lity, Glaziou 22551 (B, G. К, E. Quo 22555 5). Minas Gerais: be- ween Sucupira and Ome а, У » U berlandia Uberabinha), — 1200 (BM, ; Serra de S. José, — 17406 (P. US, W); — Tos ie Macedo . 351 (US); Con- ceigáo ni Mato Dentro, Rodovia Belho Horizonte, == ч — = 3otucatú, 18 km N of Botucatu, | 1 km Sáo Manuel, Cobra 1043 (US). Without state: with- out locality, 1844, Weddell 2556 (P). 11. Paspalum goyasense Davidse, Morrone & 391, fig. 2. 2001. TYPE: К de Cabeceiras (ca. 4 Zuloaga, Novon 11: Brazil. Goiás: 25 km km E of Goiás—Minas Gerais), cerrado, 1000 19 Nov. 1965, H. S. Irwin, R. Souza & R. Reis dos Santos 10524 (holotype, MO!; isotype, US-2529124!). Figure 10. 40—05 internodes 2-3. 6— Caespitose perennial; culms cm tall, erect, unbranched, few-noded: nodes brown, densely pilose. Sheaths 3-10 cm long, mostly basal, papillose-pilose or hispid, the margins membranous. Ligules 0.5—0.9 mm long, membranous, brown, glabrous; pseudoligule absent; collar pilose. Blades linear, 8-20 cm long, 0.2-0.4 cm wide, mostly basal, flat or with involute margins, ascendent and rigid, rounded at the base, the apex acute, densely papillose-pilose on both surfaces, the margins ciliate. Peduncles long exserted, up to 30 cm long, filiform, pale or tinged with purple. Inflorescences terminal, exserted, with 2 racemes, occasionally with a single raceme or with a third raceme present; main axis up to 3 cm long, flat, smooth, glabrous, occasionally absent; pulvini gla- brous, with a squamiform bract; racemes alternate, ascendent and divergent, ending in a developed spikelet; rachis of the racemes flat, straight, (2—)3— 7 em long, 1-1.6 mm wide, glabrous, the midnerve pale to green, the wings hyaline, nerveless, pale to purple: pedicels short, glabrous: spikelets solitary, imbricate and arranged in 2 series. Spikelets long- ellipsoid, 4—5 mm long. 1—1.1 mm wide without the acute, hairs, plano-convex to slightly biconvex, smooth, pilose; upper glume and lower lemma densely pilose on the lower half with white hairs, otherwise glabrous, the margins with papillose hairs, up to 2.5 mm long, reduced toward the apex. Upper glume as long as the spikelet, acute at the apex, with or without a short tuft of hairs, membra- nous, 3-nerved, with one central and two submar- ginal nerves. Lower lemma glumiform, as long as the spikelet or slightly shorter, 3—5-nerved. Upper anthecium long-ellipsoid, 3.2—4 mm long, 0.9 mm wide, 0.7—1 mm shorter than the spikelet, plano- convex, membranous to chartaceous; upper lemma with small papillae regularly distributed, and a tuft of macrohairs at the apex; lodicules 2, ca. 0.3 mm long, conduplicate; stamens 3, anthers 2 mm long: stigma plumose. Caryopsis long obovoid, 2 mm long, 0.9 mm wide; hilum elliptic; embryo % as long as the caryopsis. Distribution and habitat. Endemic to central Brazil, where it grows in the Federal District and in the state of Goids, in cerrados between 900 and 1200 m. Paspalum goyasense is a close species of P. car- inatum, the latter with filiform leaves, inflorescenc- es usually with a single raceme, rachis of the ra- cemes 1.8-2.5(-3.5) mm wide, spikelets with long — hairs rising from the lower third, conspicuously sca- brous on the rest of the surface, and apex of the upper glume and lower lemma dorsiventrally com- pressed and flat Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia —“ specimens. BRAZIL. Distrito Feder- in et al. 10257 (MO, US); 10 km W of Taguatinga, ruta a Braslandia, Irwin et al. 10654 (MO, UB, US): Paros da Contagem, ca. 20 km | Irwin et ч 618 (MO. US): Brasília. Irwin et al. 9658 (MO, US — — Sul, Zuloaga 3835 (SI). — ae los ality, Glaziou 22441 (G, US, W), — 22443 а US, W): Serra dos Pirineus, 14 lis S de E — de ‚ 1678, 49?W, Irwin et al. 10737 (MO, US): ca. 5 km N a pe rd. W to Pires do Rio by BR D s el al. 75939 (MO). N de — 12. Paspalum heterotrichon Trinius, Sp. Gram. 3: pl. 285. 1829-1830. TYPE: Brazil. Without locality, G. H. von Langsdorff 1829 (holotype. LE! isotype, K! photo, SI!, US-2855297!). Fig- ures 15. 16 Paspalum n Schltdl., Linnaea 26: 134. 1854. hom. illeg. TYPE: Venezuela. Silla de Caracas, — 97 ee Paspalum — hon Trin. var. dew үте Hack.. Notizbl. . Gard. Berlin-Dahlem 1(8): 328. 1897. TYPE: ен Monte Furey, 1896, — — (ho- not seen; isotype, US-2 .. Paspalum ceresioides Carrillo, Revista E. vi 1. Е шу. Nac. Ante 08. Syn. nov. TYPE: 9 Cuz- co: Prov. Convención the — 920 m. 14 Abr. 1966. C. Vargas C. 15 (holotype. CUZ not seen: Sip “U SD. ntonio 2: 145. 19 Caespitose, short-rhizomatous perennial, cata- phylls pilose; culms (27-)45-90 em tall, 1-1.9 mm nodes or un- =] diam., branching at the upper branched; internodes 2-3 cm long on the lower third and 6-12 cm long on the upper portion of the culms, terete, hollow, glabrous: nodes glabrous or pilose, with hairs white, 1-1.5 mm long. Sheaths (3—)5-12 ст long. glabrous or with the upper mar- gins pilose. Ligules 0.5 mm long, membranous, truncate: pseudoligule present. a ring of hairs 3—1 mm long. Blades linear to linear-lanceolate, (4—)5— 15 cm long, 1.5-4 mm wide, the abaxial surface glabrous. the adaxial one glabrous to sparsely pi- lose, the margins scabrous, attenuate, the apex acute. Peduncles (6-) 10-16 em long, terete, pilose. Terminal inflorescences 6-12 cm long. 2-2.5 cm wide: main axis (1—)2—4 em long, 0.5 mm diam.. pilose; racemes (1 to 2)3 to 7, (27)3.5—6.5 cm long. alternate, ending in a naked point; pulvini pilose: rachis of the racemes 3.54 mm wide. acuminate. glabrous, green at the middle portion and ferrugi- e nous al the wings, wings nerveless; pedicels 0.4 mm long. hispid, laterally inserted in the spikelets: spikelets solitary, arranged in 2 series. Spikelets ellipsoid, 2.3-3 mm long, 1-1.2 mm wide. pilose. upper glume and lower lemma subequal, with an annular thickening at the base. Upper glume as long as the spikelet, hyaline, acuminate, 3-nerved, with basal hairs Y the length of the spikelet and papillose-pilose hairs at the margins, up to 3 mm long. Lower lemma membranous, acuminate, 3- nerved, with long hairs at the upper margins, oth- erwise glabrous. Upper anthecium ellipsoid, 1.5— 1.8 mm long, 0.5 mm wide, glabrous. papyraceous: upper lemma with a tuft of macrohairs at the apex, and bicellular microhairs and silica bodies all over the surface; upper palea with prickles and silica bodies on the upper portion; lodicules 2, 0.1 mm long. rounded at the apex; stamens 3, anthers 1.2 mm long: stigma plumose. Caryopsis not seen. Iconography. Chase, Contr. U.S. Natl. Herb. 28(1): 18, fig. 2. 1929. Swallen, Ann. Missouri Bot. Gard. 30: 93, fig. 18. 1943. Rodriguez, Ernstia 8(2— 3): 38. 1998. Distribution and habitat. Paspalum heterotri- chon grows in Mexico, Central America (Honduras and Panama), and the Caribbean (Haiti and the Do- minican Republic); in South America it is found in Colombia, Venezuela, Peru, and central Brazil, usu- ally in open fields on limestone or rocky soils, be- tween 300 and 2550 m elevation. Paspalum heterotrichon shares with P. petrense. P. trachycoleon, and P. phyllorhachis ellipsoid spikelets, thickened at the base. upper glume hy- aline, and rachis of the racemes winged. Paspalum petrense differs by the long-acuminate spikelets and lower lemma with a similar pilosity pattern of the upper glume, i.e.. with long hairs at the base of the surface and the margins. Paspalum trachycoleon is distinguished by the paired spikelets, while P. phyl- lorhachis has glabrous spikelets. The specimen Macedo 1121, probably a poorly developed plant, presents smaller spikelets than typical, 1.7-2 mm long, and culms 25-30 cm tall. Internodes are occasionally markedly branching, З em long and thinner, up to 0.6 mm in diameter. In P and US (2855297) there is a specimen, Rie- del s.n., that probably belongs to the type collection 1978, ments on Langsdorff and Riedel collections). of P. heterotrichon (see Renvoize. for com- BRAZIL. Distrito Estar: al: Reserva Ecológica do IBGE. Oliveira 719 (SI). without locality, Glaziou fe (G, К,Р, W). Mir las Gerais: Mun, ltu- Representative — Mato Gros- so: Cuiabá, bitaba. cality, Riedel 966 (G, 51). s.n. к Ol — A. Magdalena: Santa Marta, (P. US-2 855297). Smith 143 (BM, LOL, G, МО, P). DOMINICAN REPUBLIC. Azua: San т тө La Vieja, Cordillera Central, Ekman 13404 (US). Santo Domingo: Cordillera Central. Prov. de la Vega, Costanza, Bajo de La Angostura. Ekman 14018 372 Annals of the Missouri Botanical Garden == m 2 ] W 7 WS \ ш SX RRS ү au я] ^ М Figure 15. Paspalum i ва Habit. —B. Detail of ligule. —C. Portion of the rachis, —D. Spikelet, dorsal view. —E. Spikelet, ventral view. Е Upper anthecium, dorsal view. —G. Upper — ventral view. — H. Upper palea ih lodic m з. —l. Ci aryopsis, scutellar view. —J. байбы, hilum view. (A, В, D-J based on Malme 1562b, US; € based on Tovar 4176, US.) p — ocotzi X-62.3 (US); Mun. m, 24 Mayo 1989, Dias et al. 3493 (MO). idse 3082 (К. MO); prope Colonia Tovar, Fendler 1698 (G, taume V0530 (MO Volume 89, Number 3 Denham et al. 2002 373 Paspalum subg. Ceresia | | | | | | | | | A P. heterotrichon | e 7 lanciflorum | | | | \ | с \ \ \ \ \ | \ \ \ \ \ \ \ \ \ | \ \ \ | \ y \ \ p pi | \ \ non \ \ Q0 200 400 600 son Bs km \ \ | — E не ý u 100 200 — A00 St ТҮ miles \ \ Prepared by Hem K Rypkema J | — — — e = Figure 16. Distribution of Paspalum heterotrichon and P. lanciflorum МО. US). HAITI. du Nord: vicinity of St. Michel Че 13, Paspalum humboldtianum Flüggé, Gram í ; - V er, Paspe : . . as "Paspalus Ennery, . Leonard 8953 (MO), 8957 (US): Frouin, eruptive Monogr | "id P 61 1810, da "1 sis ni hills above C hapelle Jubilée, Ekman 2401 (G, US); Massif humboldtianus." Panicum humboldtianum des Matheux. 26 Oct. 1924, Ekman 2274 (US). HON- (Fliiggé) Kuntze, Revis. Gen. Pl. 3(2): 361. DURAS. Comayagua: 4 km from San Isidro on the road 1898. TYPE: Ecuador: Puembo, А W. H. A. to Pane, S side of rio Tula Gorge, Moran 5505 (MO). El von Humboldt & A. J. A. Bonpland 3104 (ho- Paraíso: near km 2, road to Yuscarán, 5 Nov. 1951. Swal- | . B-W! pl SI!: is BMI. PLUS len 11354 (MEXU, MO, US). Francisco Morazán: re- огуре, : photo, 515; isotypes, Mq gion of Las Mesas, Swallen 10810 (MO, US). 1 1401 (US). 601340! photo, SI!). Figures 14, 17 XICO. Chiapas: El Pozo to Oxchuc, Hernández Xol- Villa Corzo, above Colonia Vi- Guerrero on road t Finca Cuxtepec, Breed = Paspalum distic к Kunih, in H.B.K., Nov. Gen. P te love & Sp. 1: 86. 1816 TY ;olombia: Mesa de WI pe 54596 (MEXU, US). PANAMA. Chiriquí: near et Ibagué, Oct 1. A. von Humboldi & A. J. / El Boquete, Hitchcock 8297 (SI, US). PERU. — Bonpland s.n. ‚ Mee P!; isotypes, B!, — Luya Province, C spa и Cedro, 2450-2550 2855210!) Huanc s em Paspalum ciliatum Kunth, in H.B.K., Nov. Gen. Sp. Huancavelica, Distrito de v * 'olpa, entre Marcavalle pl. 24. 1816, hom. Шер. Paspalum Bepharophorum y Quintabamba, Tovar 4176 (US). VENEZUELA. Aragua: Roem. & Schult., Syst. уы 2: 292. 1817 km S of Alto de Choroní along road to Maracay, Dav- Colombia: Ibagué у, Valle e Caravajal, Monte Quin- diu, Oct. Н. A. von Humboldi & A. J. A. Bonp- d s.n. holotype PL isotypes, B!, ся 2855275!). Panicum obtectum J. Presl, — Hae nk. 1: 301. 1830. Tricholaena obtecta (J. Presl) Ё — ex Hemsl., Lara: Distr. Palavecino, Quebrada La Mata, towards side of the Parque Nac. Terepaima, Burandt & Gan- 374 Annals of the Missouri Botanical Garden = — — | ES Figure 17. Paspalum лшн gi n —A. Habit. —B. Detail of ligule. —C. Portion of the rachis. —D. Spikelet, dorsal view. —E. Spikelet, ventral view. —F Upper anthecium, — view. —G. Upper anthecium, ventral view. — H. Upper palea, lodicules, and gynoecium. (Based on Zuloaga et al. 5887, SI.) Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia 375 iol. Centr.-Amer., Bot. 493. 1885, comb. illeg. fluens obtecta (J. — Fourn., Mexic. Pl. 1886. TYPE: Mexico. Without locality, Z P. X. Hank x s.n. 1. (holotype, PR not seen; isotypes, US- 2942530!. Paspalum ul Hack.. in Stuck., Anales Mus. Nac. juenos / 11: 63. 1904. Paspalum ps um Flüggé var. stuckertii (Hack.) Hack.. Allg. Bot. Z. Syst. 12: 97. 1906. TYPE: Argentina. Córdoba: Dpto. Punilla, Sierra Chica de Córdoba, en la punta de la — del Salto, del Ayo. Ochoa, 1000 1 m, l T. J. V. Stuckert 13023 (holotype, ires -BA AL CORD!, G!, MO, US-80012! — SI! US-557978)). Paspalum — rum yc Contr. U.S. Natl. Herb. 24: 443. 1927. Syn. TYPE: nd Loja: — n Loja ме Sim — 2100 m, 6 Sep. 1923, 4. d 21495 (holotype, Us. 1164798! photo. B vpe. 5 not s fadus humboldrianum ir var. elegantissima Bee- Пе. Phytologia 52: 15. 1982. — nov. TYPE: Mex- ico. коне ) km — Los wr os, San José de Andrade, 29 Sep. 1980, A. Бойе M-5962 AS (holotype, not located). - trj Short rhizomatous perennial, with leafy rhizomes: culms (30-)50-100 em tall, 0.2-0.3 cm diam.. erect or decumbent, lignified at the base, branching at the lower and middle nodes or unbranched: lower internodes 0.51 cm long, the upper ones 4-15 cm long. terete, hollow, glabrous or shortly pilose at the distal portion; nodes glabrous to pilose. Sheaths 4— l cm long, longer than the internodes, overlap- ping. densely to sparsely papillose-pilose. with ca- ducous tuberculate hairs at the upper portion, the margins ciliate. Ligules 1.2-2.5 mm long, membra- nous, brown, glabrous; pseudoligule absent: collar 5-2] cm glabrous Blades lanceolate, flat, to sparsely pilose or hirsute, with long, tuberculate. glabrous or pilose. long. 0.5-1.8 cm wide, herbaceous, caducous hairs, green glaucous, shortly pseu- dopetiolate, the apex subulate. Peduncles exserted, up to 30 em long, terete, glabrous. Terminal inflo- rescences exserted, 8—12 cm long. 3-5 cm wide: main axis 2-7 cm long, flat, with longitudinal nerves on the wings, scaberulous; racemes 2 to 6 ascendent, alternate and distant, slightly divergent from the main axis, ending in a naked point, oc- casionally in a spikelet; pulvini with a long tuft of white, rigid hairs; rachis of the racemes (3—)5-9 cm long, 1.2-1.5 mm wide, flat, shortly winged, green to purple. glabrous, scaberulous on the margins: pedicels unequal, up to 1.2 mm long, scaberulous: spikelets paired, and imbricate in 4 series, occa- sionally the lower spikelet aborted. Spikelets long- ellipsoid, 3—4 mm long, 1.2-1.4 mm wide, acute. plano-convex, pale or tinged with purple, the mar- gins with long hairs: upper glume and lower lemma as long as the spikelet. Upper glume 3-nerved, hy- aline, with plicate, corky margins covered with rig- 2-3.2 mm long. the rest of the surface sparsely pilose. Lower lemma id. radiate and tuberculate hairs. 3-nerved, membranous, glabrous, scaberulous at the upper portion, depressed at the base. Upper anthecium ellipsoid, 2.6-3.2 mm long, 1-1.2 mm wide, membranous, pale, shiny, 0.4—0.6 mm shorter than the upper glume and the lower lemma: upper lemma 5-nerved, finely papillose with silica bodies. bicellular microhairs and prickles at the apex: up- per palea papillose, with bicellular microhairs at the distal portion; lodicules 2, 0.6 mm long, trun- cate, stamens 3, anthers 1.6 conduplicate: mm long. Caryopsis not seen. Common name. “Canutillo” (Venezuela). Chromosome number. n = 10 (Hunziker et al.. 1998); п = 20 (Saura, 1948); 2л 40 (Tateoka. 1902). Distribution and habitat. Found from Mexico and Central America to Bolivia and central and northwestern Argentina. |t. grows on mountain slopes. on loose. rocky to sandy soils: it is also common at margins of highways and in forest edges or modified areas. between 400 and 2900 m ele- valion. Paspalum humboldtianum is related to P. poly- phyllum and P. buchtienii by the upper glume with corky margins and covered with radiate hairs, the rachis usually ending in a spikelet. Paspalum po- lyphyllum differs by the narrow (0.1-0.6 em wide). to linear-lanceolate not pseudopetiolate, linear blades. rachis pilose and with anastomosed veins. spikelets with the upper glume with unequal mar- ginal hairs. and lower lemma pilose on the distal margins and the upper portion of the surface. Dif- ferences with P. buc htienii v were summarized in the treatment of the latter spe Paspalum — жанан has spikelets dark brown: however, pale spikelets and anthers were tinged with purple and anthers purple or observed in several specimens, such as Zuloaga et al. 5885. The type material of P. humboldtianum var. ele- gantissima was nol located; it should be mentioned that no herbarium was cited by Beetle (1980) when However, describing the species. the diagnosis al- lowed us to conclude that this variety is a synonym of P. humboldtianum. hen describing P. soboliferum based on a sin- 1927) distin- guished the species from P. humboldtianum mainly gle collection from Ecuador, Chase ( by the shorter, more slender, and ascending ra- cemes, spikelets 2.8-2.9 mm long, with short hairs toward the margins. Therefore, we considered the 376 Annals of the Missouri Botanical Garden type collection a poorly developed plant and placed P. soboliferum as a synonym of P. humboldtianum. Spikelets with proliferous bracts were observe in several specimens from Mexico, such as Müller 2036, Arséne 2813, and Schaffner 136. Representative — ARGENTINA, ооа чү Andalgalá, Las Payvas, bu 1765 (LIL, US). Córdoba: Ascochinga, Burkart 10211 (SD); rae Send Sierra Chica, Stuckert 14159 (CORD, G, MO, SI). Jujuy: Río Perico, Burkart & Troncoso 11014 (SI); Dpto. Volcán, — : = 5885 (SI). La Rioja: а: Dpto. € al San Luis, ees Cuesta de . ‚ US). San Luis: bod sdernera, San José del Morro, Ea. La Morena, Boel- ске & Moore 16646 (SI). Tucumán: Taff del Valle, Bur- kart 5335 (SI); Dpto. Burruyacú, camino a Cerro Medina, Villa Carenzo et al. 1818 (LIL). BOLIVIA. Chuquisaca: Sucre by the exit to Tarabuco, Wood 7852 (US). Cocha- bamba: Morochata, Cárdenas 3410 (LIL); W of Cocha- bamba, — rie ) (US). La Paz: viciniis Sorata, Mandon 125: M, C, P. US, W). Santa Cruz: 10 km E of Samaipa, ^ ^nvoize « Cope 4038 (MO); Buena Vista, Steinbach 6618 (BM, G, US). Suere: Sucre, Cárdenas 515 US). Tarija: Camargo, Fiebrig 3082 (G, LIL, SI, W). CO- LOMBIA. Antioquia: San Jerónimo, Aralar 629 (MA, US). Boyacá: without locality, Saravia 4006 (COL). Cau- ca: Chisquío, Finca Los Derrumbos, Asplund 10498 (G, US). Cundinamarca: La Esperanza, Cuatrecasas 3001 (MA). Huila: 6 km SE of Altamira along road to Florencia, Davidse et al. 5597 (COL, MO). Norte de Santander: vicinity of Pamplona, Killip & Smith 19770 (US). Prov- idencia: Pico-casabaja, /drobo 11636 (COL). Santander: Zapatoce: Fassett 25502 (COL, US). Without departa- ment: without locality, Mutis 5537 (US). — Valle de San Juan, Cerro del Real Mina del Sapo, Echeverry 1251 (COL). Valle: near Sevilla, Barclay ‘ Juajibioy 5703 (MO, US) Cordillera Occidental, E ope, : > 1, Loma Los Cristales, Cuatrecasas et al. 25702 (€ OL. US). COSTA RICA. Cartago: zu Juan Norte, Pohl et uil 11434 (US). Guanacaste: 2 km E of Liberia, Pohl et al. 11327 (US). ECUADOR. Azuay: vicinity of Cue nea, along Río Milchichi, about 5 km N of Cuenca, Camp E- 2743 (G, US). Chimborazo: Chunchi-Alausi, Acosta Solís 21454 (US). Loja: halfway between Loja and Catamayo, Grignon 84257 (MO). Oro: between Portovelo and E Tambo, Hichcock " 306 (US). Pic hincha: Tumbaco, As- plund 6531 (G, P, US). Tunguragua: Ambato, ue 21710 (US); UR Asplund 7593 (G, P, US). 7 VADOR. Ahuachapán: P.N. El D Sermeño s.n. (MO). San Salvador: San Salvador, Calderón 1152 (US GUATEMALA. Alta Verapaz: 5 mi. S of Cobán is Hwy. CA 14, Thomas s.n. (MO). Santa Rosa: Cerro Re- dondo, Heyde & Luz 6271 (6). Sololá: Volcán San Pedro, Steyermark 477177 (US). HONDURAS. Francisco Mor- azán: vicinity of El Zamorano, Swallen 10776 (US). Olancho: Mun. La Unión, 5 mi. E of La d along road to Olanchito, Davidse et al. 35448 (MO). M Chiapas: 22 km NE of Motozintla on hwy. to с nango. Stevens & Martínez 25701 (SI). Colima: Alzada, Hitchcock 7055 (US). Guanajuato: ca. 8 km W of San Felipe on Cerro del Fraile, Sohns 427 (U: : Montes de Oca, Vallecitos, Hinton 11400 (US). Jalisco: jomulco de Zuñiga, Cerro Viejo region, 3: SSW of Guadalajara, Machuca Nuñez 4640 (MO). Mi- Tumbaya: choacán: Mun. Apatzingan, Leavenworth & Leavenworth 1649 (US). Morelos: Cuernavaca, Hitchcock 6837 (US). arit: La Yesca, 23 km al SE Tenorio & Flores 16710 (MO). Oaxaca: Mun. Santa María Ecatepec i : Santa María Zapotitlán, Flores Flanco et al. 3643 (MO). W of San Luis Potosí, Reeder et 3 (US). Sonora: Ta popa, Sierra Saguaribo, — etal. 19307 (US). Tamaulipas: Mun. Villa de Casas, cen- tral e of the Sierra de Tamaulipas, Martínez & ee F-1953B (US). Veracruz: Zacuapán, — 2002 (US). Zacatecas: Mun. Valparaiso, 1 km a San Francisco, Rzedowski 176. 36 (US). NICARAGUA. Nueva Segovia: Santa María de los Pinos, 5 km NE de Dipilto, Stevens & Montiel 24761 (MO). PANAMA. Chiriquí: Partch 69-91 (MO): roadside near El Hato Stockwell & Aiello 893 (MO, SI). PERU. Gasca Prov. Chachapoyas, 5 km W of C bac ‘hapoyas on road to Caclic, Hutchison & — — (G, MO, P, US Challhuanaca, Ayma . Núñez 7175 (M Chumbes, Balls 6023 (BM. US). Cajamarca: Prol Chota, Yamaluc, a 11 km sobre la carretera Cochabamba—Huam- bos, Sánchez Vega 2341 (SI). Cuzeo: Prov. Convención, Amaybamba, Marin 851 (US). Huancavelica: entre Iz- cuchaca y Acoria, Tovar 1005 (US). Lambayeque: Prov. Chiclayo, Cerro Reque, — Quiroz 2322 (MO). Lima: entre ap dope y Cañete, Ferreyra 5373 (US). VENE ZUELA. Aragua: Distr. Ric 'aurte, ca. Colonia. Tovar along the Colonia Tovar-La Victoria Hwy., Davidse & Miller 28159 (MO). Distrito Federal: 5.5 km down poe to Carayac a, Davidse & lI ^tt 4073 (MO). Lara: Jim S of 9°42 09°38 ү pons үз — 21370 (M a Mérida: Se- 4 (US). Miranda: Los Teques, Bur- (S1). Táchira: Mun. Andrés Bello, Loma Aldea Salomón, Tamayo 2281 (US). > — = > - 7 5 «d = — 3 = = 2 Blanca, 14. Paspalum imbricatum Filgueiras, Bradea 3: 153, fig. 2. 1981. TYPE: Brazil. Mato Grosso: Mun. Sidraléndin. Rod. BR-163, orla de brejo, 27 Oct. 1970, G. Hatschbach 25281 (holotype, US-2705820! photo, SI!). Figure 25. 0.3— un- Caespitose perennial; culms 1-1.5 m tall, 0.5 em diam., herbaceous 1 » rigid, erect, branched: internodes 15-30 cm long, slightly com- pressed, hollow, glabrous, pale or tinged with purple; nodes brown, glabrous. Sheaths 15-25 em long. densely hirsute, with papillose-pilose hairs at the distal portion, glabrous at the base, the margins membranous, shiny on both surfaces. Ligules mem- 1-1.5(-2) mm long, pseudoligule a ring of hairs up to 1.3 mm long. 35-60 cm long, 0.6-1 cm branous, brown, glabrous: Blades linear-lanceolate, wide, acute, pungent, mostly basal, flat, rigid, as- cendent, the basal ones long pseudopetiolate, densely hirsute, with tuberculate hairs, on both sur- faces. Peduncles up to 40 em long, terete, glabrous, pale or purple. Terminal inflorescences exserted, 10-15 cm long, 8 em wide; main axis 1.5-5.5 cm long, flat; racemes (2)5 to 7(1 1), (5-)10-14(-16) em long, ascendent, slightly divergent from the main Volume 89, Number 3 2002 Denham et al. rs Ceresia 377 axis, alternate, ending in a naked point: pulvini shortly pilose; rachis of the racemes 2.4—2.8(—3) mm wide, flat, the margins membranous to hyaline, nerveless, glabrous, green, shortly winged. wings purple; pedicels short; spikelets solitary, densely imbricate in 2 series. Spikelets broadly ovoid, 5.5 mm long, 2.8-3 mm wide, plano-convex, gla- brous, pale, upper glume and lower lemma sube- qual, or the lower lemma slightly longer, cordate, the apex acute to almost rounded, the margins winged. Upper glume papyraceous, corrugate in the lower half, 5-nerved with 3 central nerves reaching the apex. Lower lemma glumiform, 5-nerved, cor- rugate in the lower half. Upper anthecium long- ellipsoid, 3.8-4 mm long, 1.2 mm wide, plano-con- vex, membranous, pale, pilose toward the apex. finely papillose; upper lemma with small papillae evenly distributed all over the surface. with macro- hairs and bicellular microhairs at the apex, hispi- dulous toward the margins and in the lower portion: upper palea hispidulous toward the apex: lodicules 2. ca. 0.4 mm long, conduplicate, hyaline: stamens 3. anthers 2-2.5 mm long; stigmas 2. plumose. lat- eral. Caryopsis nol seen. Brazil, where it is found in cerrados in the Distrito Federal Distribution and habitat. Endemic to and the states of Mato Grosso and Goiás, between 950 and 1250 m elevation. Paspalum imbricatum is a member of section Pectinata, sharing cordate upper glume and lower easily distin- lemma, the margins winged; it is guished by the glabrous spikelets. 15. Paspalum lanciflorum Trinius, Sp. Gram. 3: pl. 286. 1829-1830. TYPE: Brazil. M alo Gros- Cuiabá, 1829, H. von Langsdorff s.n (holotype, LE!; isotype, US-80075!). Figure 16. uc contractum Pilg., E^ — st. 29: 700. 898. TYPE: Colombia. M Janos de San жй — ү 68, M. A. pu po A ue В!; isotype, US-2942134! Р ке ес chino Mez, Bot. Jahrb. Syst. 50, Beibl . 1921. TYPE: Brazil. fis Bra iy ern — мш 1909, E. H. Ле — designated by Toda WICZ (1990: йч BL isolecto- types, K!, US-2854675! photo, US). ын — Swallen, Fieldiana, Bot. 28: 24. fig. 19 TYPE: ека Bolivar: Gran Sabana, e Kun г ara—Paru, in d of Rio Kuke пап, S of Mount Ro aima, 1065— ym A. Steyer mark 59090 — US- 19116601: Isotype. Р pun aureolatum Sells n, Fie Шала, Bot. 28: 22, fig 1951. TYPE: l'erritorio Federal An zonas: summit of Cerro Duida, 1025—1200 m, 2 Ben. 1 n Venezuela. 1944, Steyermark J 58234 (holotype, US- 1911652! photo, SI! — Caespitose, shortly rhizomatous perennial; culms 45-120 cm tall, 14 cm long, glabrous, brown: nodes glabrous to pi- erect, unbranched; internodes 9— lose. brown, compressed. Sheaths 9-17 cm long, densely hirsute on the distal portion, otherwise sparsely pilose or glabrescent, the adaxial surface 1.5-3.2 mm long, glabrous, brown; pseudoligule a ring of hairs shiny, brown. Ligules membranous, up to 0.8 cm long. Blades linear-lanceolate, 7-35 em long, (0.3—50.5-1(-1.5) em wide, flat or with involute margins, densely papillose-pilose, with tu- berculate hairs up to 7 mm long, the midnerve con- spicuous on the adaxial surface. Peduncles long- exserted, up to 40 cm long, terete, glabrous. Terminal inflorescences subdigitate, yellow, occa- sionally with a papyraceous bract, up to 4 mm long, at the base; main axis absent or up to 1.5 em long, 33 to O(11), ascendent, 5—15 em long, subconjugate and ending — glabrous, flat; pulvini pilose; racemes ( in a naked point; rachis of the racemes (4—)5—7 mm wide, winged, green at the center, the midnerve white, with purple, membranous and glabrous mar- gins: pedicels short, glabrous. flat: spikelets soli- ary, imbricate in 2 series. Spikelets long-ovoid, — (4.8—)5.6—7.2 mm long, 1.2-2.5 mm wide, plano- convex, acuminate, pale, pilose. Upper glume as long as the spikelet, membranous, rounded and densely pilose at the base, the basal margins pilose, with hairs up to 2 mm long, otherwise glabrous, 3- nerved. Lower lemma slightly shorter than the up- per glume, lanceolate, 3-nerved, densely pilose at the basal margins, the median portion with con- spicuous, tuberculate hairs 2.5-3 mm long. Upper 3.24.7 mm long, 0.9-2 wide, membranous, pale, stipitate, the stipe 0.3 mm anthecium ovoid, mm long: upper lemma finely papillose, with bicellular microhairs and silica bodies at the apex, otherwise lodicules 2, cale; stamens 3, anthers 1.6 mm long; stigma lat- elabrous: ca. 0.4 mm long, condupli- eral. Caryopsis not seen. Iconography. Chase, Contr. U.S. Natl. Herb. 28(1): 26. fig. 7. 1929 (under P. contractum). Jud- ziewicz, Fl. — 469. fig. 83, E. 1990 Common name. “Pama” (Venezuela). Distribution and habitat. Panama, Colombia, Guyana, Surinam, Venezuela, Brazil (Roraima, "J зага, Maranhão, Mato Grosso, Minas Gerais, Goiás, and Distrito Federal), and Bolivia; it grows in sa- vannas subjected to fire, on limestone, sandy, or rocky soils, between sea level and 1200 m Paspalum lanciflorum belongs to section Pectin- ata; within this section it is related to P. cachim- 378 Annals of the Missouri Botanical Garden with branched, geniculate culms, rachis of the racemes boense, which differs by the annual habit. 2-3 mm wide, spikelets 4—4.3 mm long, and upper anthecium 2.2-2.5 mm long. Also, P. lanciflorum is distinguished from P. pectinatum and P. cordatum by the glume densely pilose on the lower portion, not cordate, and upper anthecium stipitate, 4—% the length of the spikelet. This species is quite variable in its overall mor- phology. The specimens Wurdack & Guppy 27, Tate 157, and Hahn et al. 5592b are approximately 40— 50 em tall, with all blades basal. On the other hand, Kuhlmann 1674, Glaziou 20084, 22544, and Zu- loaga et al. 4413 are more robust specimens, ca. 120 cm tall, with blades evenly distributed along the culms. It is noteworthy to mention the presence of densely papillose-pilose specimens, e.g.. Zuloa- ga et al. 4413, Tamayo s.n., Egler & Raimundo 1262, and others, as sparsely pilose, such as Wil- liams 13088 and Maguire et al. 32038. Three to six racemes per inflorescence are fre- quent in this species; nevertheless, one to thr racemes are found in Swallen 4029, Williams 13088, Wurdack & Guppy 27, Gifford G-101, Oliv- eira & da Silva 700, Tamayo 3198, Ramirez 961, and Heringer et al. 4615; the specimen Echeverry et al. 2: Spikelets are small, 4.8—5 mm long. 1.7-2 mm wide in Davidse et al. 22647 and Blydenstein 1468. Chase (ined.) distinguished P contractum. from P. lanciflorum by the number of racemes (1 to 3 vs. 3 to 6) and the relative length of the lower lemma (as long as the spikelet vs. shorter than the spike- let). Our analysis of herbarium material did not al- is 11 racemes per inflorescence. low us to distinguish between both taxa. According to Judziewicz (1990), the specimen Pohl 1454 (seen in B and W) is a syntype of the species; nevertheless, we examined at LE a speci- men collected by von Langsdorff (see type materi- al), and there are no elements that suggest Trinius had cited two syntypes when describing the spe- cies. VIA. Santa Cruz: Noel Representative specimens. BOLI Vel — elasco, Parque Nacional Kempff Mercado, Campane nto Las Gamas, Killeen et al. 4929 (MO). BRA- ZIL. Distrito Federal: с Baci ча do Rio S. Barto- lomeu, He — et al. 46 erra de Arruda, Glaziou 4 (K, P, US); Serra Dorada Pohl 1454 (G). Goiás: Mur n. Couto Megülhaes, 16 km W of Piqizeiro along GO- 70, Plowman et al. 19132 (US). ode — Mor- ro antes do Rio Tem, Oliveira & da Silva 700 (S1): Caro- lina to San о ve Balsas. Swallen 402 ds (US). Mato Grosso: ca. 270 de Xavantina, Gifford G-101 (US): entre Bardo de ee ma e Utiarity. Kuhlmann 1674 (US). Minas Gerais: Ribeiráo Taquaruc u, Glaziou ( ‚Р, US). Para: Tapajós, Rio Cururú, Egler & R mi 1262 ( US). Roraima: Aldo ча Tucham Paulo, para a base do Roraima, Rondon 5. A ea 2454780); Mt. Ro- raima, Paulo, Tate 157 (US OMBIA. Cravo Norte Ara uca, al S del Río Meta, a 1468 (US). Meta: O de La Macarena, vía a C Е — & Ma- ld 6886 (Sl): carretera de Bella Vista a Piñalito, Echeverry & Taranto 2312 Mun. Puerto Carreño, ' Amat & — 3 (COL). GUYANA. Potaro- Siparun ni Ra egion, ahn et al. oe (US); Annai Pus — 923 (US). PANA- A. a: ca. 7 mi. of Minas, D'Arcy & Antonio 13527 (MO). SURINAM. Sipaliwini savana area on Brazilian анвар Oldenburger et al. 5 (US). VE zonas: Canaima, R. Carrao, García Barriga (COL): Santa Barbara savanna, m of Ríos Ventuari and. Orinoco, Maguire А al. 8 (US); alre- dedor del bajo Río Cataniapo, P жее ис cho Williams 13088 (US). Anzoategui: Dist. Bolívar, S of El Zamuro, Fila El Purgatorio, Davidse & — 19314 (MO). Bo- livar: El Pauji, Liesner 1935 0 (SI); pié de la Roca to Guayaraca, Davidse & Huber 22647 (МО); Gran Sabana, Vía Kavanayen, El ч — 961 (MO): Hato Santa Teresa, Tamayo — Hato La Vergareña, Wurdack & Guppy 27 (US 1 al S de San Ignacio de Yuruaní, Gran Sabana, Zuloaga et al. 4413 (K, SI, US). Herrer Z` la 16. Paspalum longiaristatum — & Fil- gueiras, Novon 3: 129, fig. TYPE: Brazil. Goiás: Mun. Ode ‘Me edo, 14°18'5, 48?23'W, 13 Арг. 1992. T. S. Fil- gueiras 2277 (holotype, IBGE!; — B not seen, BM!, FLAS not seen, ICN!, ISC not seen. ‚ MEXU-62.8764!, MO!, NY not seen, Р not SI!, RB not UB not seen, US not Figure 23. seen, SP not seen, R not seen, seen, UFG not seen, seen). Caespitose annual; floriferous culms 15-36 cm tall, 0.8—1 mm diam., branching at the lower nodes: internodes terete, hollow, glabrous; nodes dark, pi- lose. Sheaths 3-7 cm long. papillose-pilose, the margins smooth. Ligules membranous-ciliate, the membranous portion 0.5-0.8 mm long, cilia 0.2— 3.5-8.2 cm long, 1-2 mm wide, attenuate at the base, papillose-pilose on 0.8 mm long. Blades linear, both surfaces, more densely so toward the base. Peduncles terete, 0.5 mm diam., glabrous or with papillose-pilose hairs. Terminal inflorescences ex- serted, 5-8 cm long: main axis 1-2.5 cm long, pa- pillose-pilose, ending in a naked point; racemes | to 2(4), 2-7 cm long, alternate, arcuate, ascendent and divergent; pulvini pilose; rachis of the racemes membranous, 4—6 mm wide, winged, hyaline, and ciliate at the margins, the adaxial surface with pa- pillose-pilose hairs, the abaxial surface glabrous; pedicels 0.1-0.2 mm long, hispid; spikelets soli- tary, imbricate in 2 series. Spikelets ellipsoid, 1.8— 2.2 mm long. 0.5-0.6 mm wide, pilose, brown or tinged with purple, shiny. Upper glume hyaline, 3- nerved with one central and two marginal nerves, pilose at the base, the hairs reaching % the length Volume 89, Number 3 2002 Denham et al. Paspalum sib Ceresia of the spikelet. the margins pilose with ascendent hairs, awned, awn 6-12.2 mm long, scabrous. Low- er lemma glumiform, slightly shorter than the upper glume, membranous, 3-nerved, awned, awn 0.3-2 mm long, scabrous. Upper anthecium ellipsoid, 1.5-1.8 mm long, 0.7-0.8 mm wide, membranous. shiny: upper lemma 5-nerved, with prickles, bicel- lular microhairs and silica bodies at the upper por- tion; upper palea 2-nerved: lodicules absent: sta- mens 3, anthers 1.4-1.8 mm long. purple al maturity: stigma plumose. Caryopsis 1-1.2 mm long, 0.1—0.5 mm wide; hilum punctiform. basal, embryo more than half the length of the caryopsis. Distribution and habitat. Endemic to central Brazil. cerrados. in serpentine, rocky soils at 1000 m. in the state of Goiás, where it is found Paspalum longiaristatum differs from P. biaris- tatum by the perennial habit, spikelets 3.8—1.5 mm long. awn of the upper glume 4—7.1 mm long. and awn of the lower lemma 3.84.5 mm long of the latter taxon. Representative specime ns. BRA Goias: а. 15 km N of Niquelándia, E el v 144 ( gee а 18 km de Niquelándia. тая б т SI. y Filgueiras 2780 (MO, . Oliveira et al. 638 M A 23 ao 3 da estrada e 3 rra a direita da mina de niquel. Fonseca et al. 227 (MO, SD): Mun. Alto Paraiso. 14°10'S, 98°47 i Filgueiras & Oliveira — ‘do. О); P» ^ FS Vale da Lua. 3292 (US) 17. Paspalum malmeanum Ekman. Ark. Bot. 10: 12. tab. 4. fig. 3. 1911. TYPE: Brazil. Mato Santa Ana da Chapada, 31 May 1903. A. Malme s.n. (holotype, S not seen: iso- -80089!). Figure 23. gw * Us Caespitose, shortly rhizomatous perennial: culms 60-90 cm tall. long. purple. terete, hollow, striate; nodes glabrous. few-noded: internodes 6—10 cm Sheaths usually longer than the internodes, slightly compressed, glabrous, pale or tinged with purple, shiny. the margins ciliate or glabrous. Ligules mem- branous, 0.2 mm long. brown, glabrous: pseudoli- gule a ring of hairs up to 1.2 mm long. Blades linear, 10—45 cm long, 0.1 em wide. involute. pa- pillose on the adaxial surface. glabrous and smooth on the abaxial surface, the margins papillose-pilose toward the base with hairs up to 6 mm long. oth- erwise glabrous, the apex subulate. Peduncles ex- serted, terete. up to 35 em long. glabrous. Terminal inflorescences exserted; main axis absent: pulvini with a tuft of short hairs and occasionally with 2 small bracts on the basal pulvini; racemes 2, con- jugate. 7—14 em long. occasionally with a single raceme present; rachis of the racemes 1.2-2 mm wide, ending in a naked point, purple or with the margins red-tinged, glabrous, the wings flat: pedi- cels subterete, glabrous to short pilose at the apex: spikelets solitary, densely imbricate in 2 series. 1.6-2 mm long. 0.8—1.2 mm wide, villous, early deciduous, upper glume and Spikelets ellipsoid, lower lemma subequal. Upper glume membranous. 3-nerved, the margins rigid, densely pilose at the basal portion with short and appressed, tuberculate hairs, the margins ciliate, silky with hairs up to 1.5 mm long, the rest of the surface glabrescent. Lower lemma glumiform, 3-nerved, pilose at the margins and % of the surface. Upper anthecium obovoid. 1.5 mm long, 0.8-1 mm wide, plano-convex. mem- branous, pale, shiny, rounded at the apex: upper lemma smooth: lodicules 2. 0.1 mm long. condu- plicate: stamens 3, anthers 1 mm long. Caryopsis ellipsoid. 1 mm long: hilum punctiform, embryo less than % the length of the caryopsis. 2n = 20 (Killeen, 1990). Paspalum malmean- Chromosome number. Distribution and habitat. um grows in Bolivia and Mato Grosso, Brazil, where it is usually found in flooded savannas on sandy soils between O and 500 m elevation. Paspalum malmeanum is related to P. eucomum. which differs by the spikelets 2.5—3 mm long. Oth- er features of its distribution and habitat are point- ed out in the treatment of P. eucomum. Representative — LIVIA. Santa Cruz: Ve- lasco. Parque Nacional Noel En M.. ju wise nto Los Fierros, Gutiérrez et i? 55 (MO); Nuflo de Chavez, 3 km SW of Concepción on road around reservoir. Killeen 2024 (G. LPB. MO. SI. US); Ñuflo de Chávez. Est. Las Madres 2076. (MO. SI, US); Ñuflo de Chávez, Estanci O, SI. fa. Killeen 2478 (M US). f São Lourenço. E > F Ibo, Chase 11962 US): Juruena, Kuhlmann 7 (U =) us Linha Telegrafica, на, p^ (1 — da 18. als Sage niquelandiae Filgueiras, Novon : 30. fig. 1. 1995. TYPE: Brazil. Goiás: Mun. Niquelandia, ca. 14?18'S, 48°23'W, ca. 2 km este da localidad de Macedo. 19 May 1993, 7. S. Filgueiras & F. C. de A. Oliveira 2401 (ho- IBGE! isotypes, F not seen. G K!, MEXU sr, lotype. ICN not seen, not seen, MO!, R not seen, Figure 23. seen. ее SP see nol seen, ` not seen, US not seen). Rhizomatous perennial; culms 120-185 em tall. 0.1—0.6 cm diam., elabrous. hollow: nodes glabrous, dark. Sheaths unbranched; internodes 5-8. glabrous or pilose toward the base and papillose- hispid toward the distal portion. keeled, the mar- Ligules 0.2-0.3 mm long. membranous: pseudoligule a ring of hairs 5-8 mm long. Leaf blades lanceolate, 20-45 cm long. 0.8- gins ciliate or glabrous. 380 Annals of the Missouri Botanical Garden 2.2 cm wide, acuminate, flat, glabrous to papillose- hispid, the margins scabrous or ciliate. Peduncles exserted, 12-22 cm long, glabrous. Terminal inflo- rescences 14—18 em long, 6-9 ст wide; main axis 9-15 em long, glabrous, terete or wavy; racemes 8 to 30, 2.5-8.5 cm long, alternate, ascendent, end- ing in a naked point, occasionally in a spikelet; pulvini pilose, with 2-2.5 mm long hairs; rachis of the racemes 1-1.5 mm wide, foliaceous, flexuous toward the apex, glabrous or scabrous, midnerve manifest and wings with thin, longitudinal nerves; pedicels 0.1-0.3 mm long, paired, arranged in 4 series. Spikelets ellipsoid to scabrous; spikelets ovoid, 2.3—3.3 mm long. 1-1.4 mm wide, pilose, plano-convex. Upper glume membranous, 3- nerved, the margins corky, with radiate, tuberculate hairs 0.3-1 mm long, the rest of the surface gla- brous or shortly hispid. Lower lemma flat, membra- nous, J-nerved, glabrous, as long as the upper glume, depressed the base. Upper anthecium ovoid, 2.2-2.7 mm long, nous to chartaceous, glabrous; upper lemma 5- 1-1.1 mm wide, membra- nerved, with simple papillae regularly distributed all over the surface and sparse bicellular micro- hairs; lodicules 2, 0.1—0.2 mm long; stamens 3, an- thers 1.2-2 mm long; stigmas plumose. Caryopsis not seen. Distribution and habitat. Endemic to the state of Goiás, Brazil, where it grows on outcrops of ser- pentine, rocky soils, near edges of gallery forests. Paspalum niquelandiae is related to P. burmanii and P humboldtianum. Differences with P bur- manii were summarized in the treatment of that species. In. Paspalum humboldtianum plants. are (30—)50—100 cm tall, with branched or unbranched culms, two to six racemes per inflorescence are pre- sent, and spikelets are long ellipsoid, with radiate hairs, 2-3.2 mm long, at the margins of the upper glume. 19. Paspalum pectinatum Nees ex Trinius, Sp. Gram. 1: pl. 1828. Paspalum pectinatum Nees, Fl. Bras. Enum. Pl. 2 1829, hom. illeg. Anastrophus pectinatus (Nees ex Trin.) Schltdl. ex B. D. Jacks., Index Kew. 1: 118. 1893. TYPE: Brazil. Without locality, А Sellow s.n. (holotype, LE not seen; isotypes, B!, G!, K! photo, SI!, LE not seen, US-2942525!, WN. Figures 18, 19. Caespitose, shortly rhizomatous perennial; florif- erous culms 35-100 ст tall, erect, few-noded; in- ternodes 2 to 4, 7-14 cm long, 0.2 em diam., hol- low, slightly compressed, striate, glabrous; nodes brown. glabrous, Sheaths usually longer than the internodes, the basal ones shiny, red-tinged to brown, glabrous; upper sheaths green, keeled, vil- lous to hirsute all over the surface or only toward the upper portion, the margins membranous. Lig- ules membranous, 0.3-0.5 mm long. brown, gla- brous; pseudoligule a tuft of white hairs 3-4 mm 10-30 cm long, 0.2-0.8 em wide, mostly basal, flat or with involute margins, long. Leaf blades linear, densely pilose on both surfaces, the base attenuate, the apex acute. Peduncles long-exserted, up to 30 cm long, finely striate, glabrous. Terminal inflores- cences with 2(3 to 5) conjugated, ascendent ra- cemes, 2-8 cm long; main axis absent or up to 1 cm long; pulvini shortly pilose; rachis of the ra- cemes shortly winged, 2-2.5 mm wide. glabrous, membranous, green, with light-brown to purple, crenate, glabrous or pilose, margins, the midnerve manifest, ending in a naked point; pedicels up to 0.6 mm long, flat, glabrous; spikelets solitary, densely imbricate and arranged in 2 series. Spike- lets lanceolate, 5-8 mm long, 2-3.3 mm wide, dor- siventrally compressed, slightly plano-convex, cor- date, the apex acute, pilose, pale or tinged with purple. nerved, winged, papyraceous to hyaline, cordate, Upper glume as long as the spikelet, 3-5- slightly convex, glabrous, the base cordate, corru- gated at the middle portion. Lower lemma 4.5-7.2 mm long, 3-nerved, not winged, sparsely pilose, subcordate, the margins papillose-pilose with rigid, tuberculate hairs 1-2.5 mm long. Upper anthecium ellipsoid, 4-5 mm long, 1.4-1.6 mm wide, plano- convex, papyraceous, pale; upper lemma finely pa- pillose, with bicellular microhairs, short macrohairs and isolated silica bodies at the distal portion and along the margins; palea smooth: lodicules 2. ca. 0.4 mm long, hyaline, conduplicate: stamens 3, an- thers 2 mm long, ferruginous; styles 2, free, stigma plumose, lateral. Caryopsis not seen. "Grama-das-pedras" (Brazil). Chromosome number. 10 (Davidse & Pohl, 1978; Pohl, 1980); 2n — 40, 60 (Gould & Soder- 1967). Distribution and habitat. Common name. strom, This species grows from southern Mexico, Mesoamerica, Colombia, Ve- nezuela, Guyana, and Surinam to Bolivia, Paraguay, and central and southern Brazil, mainly in savan- nas on sandy soils with rocky outcrops. It is fre- quent in areas with periodical fires, occasional in margins of streams, between 180 and 1800 m ele- vation, Paspalum pectinatum is a member of section Pectinata by its winged and cordate upper glume. This species is related to P cordatum, a species distinguished by its robust plants, 100—130 cm tall, Volume 89, Number 3 Denham et al. 381 2002 Paspalum subg. Ceresia Figure 18. Paspalum pectinatum.—A. Habit. —B. Detail of ligule. —C. Spikelet, ventral view. —D. Spikelet, dorsal view. Е. Upper anthecium, dorsal view with lateral stigmatic plumules. —F. Upper anthecium, ventral view with lateral stigmatic plumules. —G. Upper palea, lodicules, anthers, and gynoecium. (Based on Filgueiras 2459, Sl.) 382 Annals of the Missouri Botanical Garden MEN 1 | ¢ P pectinatum e P phyllorhachis ж Р petrense — \ \ \ \ O 200 400 600 800 1000 km \ \ — ای یچ و‎ \ ° E 200 300 406 ме 600 т \ | red bı Hendrik K "e \ \ — — — Figure 19, with 5 to 10 racemes per inflorescence, each ra- сете 12-16 em long. The basal portion of the plant is usually covered with remains of burnt sheaths, which probably act as a protection for the renewal buds Phe specimens Fonseca et al. 153, Oliveira et al. 763, and Filgueiras 2459, collected in Goiás, Ni- quelándia, Brazil, have 8 to 9 internodes with leaf blades arranged along the floriferous culms. There is a papyraceous bract at the base of the inflorescence, as in Rivera 3, Pinto & Sastre 1066, Rivera L-L41, Davidse 5321, and 7649-۸ " Cuatrecasas Representative specimens. Manatee (or Coastal) Hwy Point, BELIZE. Belize: along y, 7.4 km S of Turnoff to Gales Atha 46904 (MO). BOLIVIA. Beni: Prov. Vaca Diez: Riberalta, 65 km hacia Santa Rosa, cerca del desvío a Cobija, Beck 20552 (K). La Paz: Prov. Ren i Luisita, Haase 630 (К, MO). Santa Cruz: Chiquitos, 4 m W of Santiago de Chiquitos, Killeen 2782a (MO, US BRAZIL. Distrito Federal: Brasília, lago Sul, Zuloaga 3839 (SI). Goiás: Mun. Niquelándia, Macedo, ca. 20 km Nee & Distribution of Paspalum pectinatum, P. petrense, and P. phyllorhachis de Niquelandia, dere 2455 (MO): Campo de Picho- ‚ Glaziou 2242 MO, P, US). Mato Grosso: Expedition Base Ca un. — 6766 (К.Р, US): « ca. 70 km N of Xavantina, /rwin & Soderstrom 6758 (K, MO, SI) Minas Gerais: Serra do Espinhaço, ca. 10 km SW o iir ir Anderson et al. 35213 (MO); Diamantina, Serra San Antonio, Chase 10350 (MO, US). Parana: Ponta Grossi, — 104 (SI); Jaguariaiva, Dusén 16057 (MO, P, SI). ta Catarina: Campo Novo, Mafra, Klein 3916 (US). ом Paulo: Boa Vista, 4-5 km SE da Estação Eng. Hermilo, Machado de Campos 111 (US). Without — Sellow 1232, 1239 (B). Pe ea Arauca: N del río Tame, Blydenstein 734 (COL). Boyacá: near Ais ué, Haught 2709 (COL). Caquetá: Mun. San Vicente del Caguán, Hato Caquetania, Bentacur & Porras 1510 (MO). Cauca: Chisquito, Asplund 10633 (US); Río — йш, San José del Guaviare, Cuatrecasas 7649-A (С IS). Cundinamarca: savanna of San Martín, SE of * gotá, Shaw s.n. (US). Guaviare: Mun. San José del Gua- лаге, Inspección de La Fuga. Giraldo-Canas & López 2631 (MO). Meta: 20 km SE of Villavicencio, Alston 7569 M. COL, US); Mun. Puerto Gaitán entre Carimagua y El Porvenir, Rivera 3 (COL). Santander: La Laguna, car- retera Lebrija-Pantano, Rivera L-141 (COL). Vichada: Parque Nacional Natural Tuparro, Barbosa € Zarucchi AA Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia 3032 (MO); ca. 45 km F between Las — and Santa Rita, 5321 (COL). COSTA RICA. Guanacaste: E of Cumaribo along dirt road Davidse & Llanos Comelco, W of Bagaces, Heithanus 163 (MO). San José: Cantón de Acosta, Fila road Morales 4395 (MEXU. MO) Puntarenas: Buenos Aires, Pohl & со 13922 (МО). ТЕМА f j a H AS. — Haci че с Espíritu Santo, — 7445 (US). MEXICO: Chiapas: Mun. ку nque, 2—4 km N of La Vic NWN toria and * km Palenque, Dandi 20544 (MEXU. MO). Oaxaca, — Sta. Marfa Chimalapa, El Ocotal Grande, ca. 5 km al E de Sta. María, Hernández 2283 (MO). Tabasco: km 64 rumbo бе q rim Rueda al O de Huimanguillo, Hill & Cowa 1548 (N Veracruz: a 11 km « Choapas con la carretera Cardenas-Coatzacoaleos, Orozco 182 (МЕХА). NICARAGUA. Chontales: 15 km SE of Juigalpa, along Hwy. 7. Pohl & Davidse 12361 (MO). Zelaya: near Cano mba, — 5 km SE of El Empalme | ay road to Lim- baika, Kral 69245 (MO). PANAMA. riqui: Cerro Vaca, eastern C hiriquí, Pittier 4351 (US): C. erro Campana, Lazor 3350 (MO). SURINAM. Sipaliwini savanna, Rom- PARAGUAY. Canendiyú: Mbaracayú tural Reserve ар red by Fundación Moisés Ber- toni, 242115 'W. Zardini & Ramirez Benitez 51117 (MO). VENE "n E : i Amazonas: E of Parucito and N of its tributary Caño Majagua, — 10153 (MO). Apure: Distrito Pedro Camejo, 22 km W de Galeras de Cinaruco, Davidse & González 15652 nn Bolívar: 2" of La Vergareña. 5 of Ciudad Bolívar, Alba 53-23 ( Cano 20. Paspalum petrense A. C. Burman, Kew Bull. 35: 297. 1980. TYPE: Brazil. Goiás: Ser- ra de Е . 20 km E of Pire nópolis, 8 Apr. 1979, T. . Filgueiras & A. G. Burman 430 aud К; isotypes, IBGE! SP not seen, RB-217458!). Figures 19, 20. Caespitose perennial; culms erect, 70-90 cm tall, unbranched: internodes terete, hollow, gla- brous or sparsely pilose toward the summit; nodes glabrous. Sheaths 7—10 cm long, papillose-pilose toward the apex, overlapping, keeled. Ligules ca. 1 mm long, membranous, truncate; pseudoligule ab- senl. 10-19 em long, 3-6 mm wide, pubescent, the margins sca- Blades linear to linear-lanceolate, brous, attenuate at the base, the apex acute. Pe- duncles 17-30 cm long, pilose at the apical por- tion. Terminal inflorescences 8-15 em long, 4—7 cm wide; main axis glabrous; pulvini pilose; racemes 3 to 5. 3-6 em long, alternate, ending in a naked pines rachis of the racemes (6.5—)7—7.5 mm wide. foliaceous, glabrous. green to purple. with anasto- mosed veins on the wings, the margins slightly sca- brous; pedicels 0.4—0.6 mm long. hispid, laterally inserted in the spikelets; spikelets paired, the basal one sometimes reduced or aborted, arranged in 2 › 4 series. Spikelets ellipsoid, 3-3.5 mm long. l- A | mm wide, long-acuminate, pilose, pale, upper glume and lower lemma subequal. with an annular thickening at the base. Upper glume membranous the lemma membra- to hyaline, 3-nerved, hairy on the lower half, margins papillose-pilose. Lower nous to hyaline, acuminate, 3-nerved, pilose on the lower half with hairs shorter than those of the upper glume. Upper anthecium ellipsoid, 2.8-3 mm long. 0.6-0.8 mm wide, membranous, shiny, with a tuft of hairs at the apex, otherwise glabrous: upper lem- ma and upper palea with bicellular microhairs and papillae all over the surface and silica bodies at the apex; lodicules 2, 0.1 mm long; stamens 3; stig- ma plumose. Caryopsis not seen. Distribution and habitat. Endemic to central Brazil, Goiás, where it is present in rocky soils on mountain slopes, at an elevation of 1200 m. Paspalum petrense is related to P. heterotrichon (see observations under this species) and P. trachy- coleon. Paspalum petrense differs from P. trachy- branching coleon by its unbranched culms (vs. culms), rachis (6.5—7.5 mm vs. 4-6 mm wide), the upper glume long-acuminate (vs. acute to acumi- nate), and lower lemma long-acuminate, pilose, not suleate, and with inconspicuous nerves (vs. acute, glabrous or with short, apical hairs, sulcate, and with conspicuous nerves). 21. Paspalum phyllorhachis Hackel. Oesterr. Bot. Z. 51: 240. 1901. TYPE: Minas Gerais: without locality, 1872, A. К.М. Glaziou 20078 (holotype, W!; isotypes. B!. K! photo. SI!, US-2855708! photo, SI!). Figures 19, 21 Brazil. Shortly rhizomatous perennial: culms 70-170 em tall, 2-7 mm diam., erect or climbing, bambusi- form, branching at the lower and middle nodes: in- ternodes 5-8 cm long, glabrous, terete, hollow. lig- nodes glabrous, Sheaths 5-8 long, glabrous or pilose at the upper portion. keeled, membranous, nified: brown. cm margins smooth, glabrous. Ligules 1.5-2 mm long. truncate to rounded ihe — at the apex, decurrent; pseudoligule absent. Leaf blades linear-lanceolate, 7-17 em long. 0.4—1.3 cm wide, rounded, long-acuminate at the apex, with the abaxial surface glabrous to pilose and the adaxial surface densely pilose, margins smooth to scabrous. Peduncles exserted, glabrous or pilose toward the distal portion. Terminal inflorescences exserted, 10-13 cm long, 5 mm diam., glabrous or pilose; racemes 1.5-3 em wide: main axis 4—9.5 em long, 0.5 7 to l - (: a naked point; pulvini pilose; rachis of the racemes 5-)3.5-5 em long, alternate, ending in 1-6 mm wide, foliaceous, with anastomosed nerves on the wings, glabrous, brown: pedicels 0.8 mm laterally inserted in the spikelets: long, hispid, spikelets paired, imbricate in 4 series. Spikelets 384 Annals of the Missouri Botanical Garden — — — — V — — — Ja 7 de br y MERE Я — "fL xL Y — < 17 — — 3.4 = — 3 = 1mm AA ERA TAX — т — ЖЧ NN COM 9 Л A VÀ UNS АШАА. Ч ANA M NW ^ D ' V \ Figure 20. Paspalum petrense.—A. Habit. —B. Detail of ligule. —C. Portion of the rachis. —D. Spikelet, dorsal view, —E. Spikelet, ventral view. —F. Upper anthecium, dorsal view. —G. Upper anthecium, ventral view. Н. Jpper palea and lodicules. (Based on Filgueiras & Burman 430.) 385 Denham et al. Volume 89, Number 3 2002 Paspalum subg. Ceresia D. Spikelet, >, Upper anthecium, ventral view. Figure 21. dorsal view. Portion of the rachis. Paspalum phyllorhachis.—A. Habit. —B. Detail of ligule. С. E. Spikelet, ventral view. —F. Upper anthecium, dorsal view. —( H. Upper palea, lodicules, and anthers. (A based on Chase 8975, Us: B-H based on Glaziou 20070, Р.) 386 Annals of the Missouri Botanical Garden ellipsoid, 2.3-2.8 mm, 0.8-1 mm wide, glabrous, upper glume and lower lemma subequal, scabrous at the upper margins. Upper glume as long as the acuminate, nerved. Lower lemma glumiform, acuminate, mem- hyaline to membranous, 3- spikelet, branous, 3-nerved. Upper anthecium lanceolate, 2.1-2.7 mm long. 0.7-1 mm wide, membranous, scabrous at the apex, otherwise glabrous; upper lemma and palea with conspicuous, simple papillae and bicellular microhairs all over the surface, and prickles at the apical portion; stamens 3, anthers 1.5 mm long; stigma plumose. Caryopsis not seen. Distribution and habitat. Endemic to Brazil, Minas Gerais, where it is found in open fields be- tween 1200 and 1400 m elevation. Paspalum phyllorhachis is related to P petrense and P. trachycoleon by its paired spikelets. It differs from both species by these spikelets being gla- brous. Glabrous and pilose spikelets were occasionally observed in the specimen Magalhães Gomes 908, with sparse hairs toward the apical margins of the upper glume. Representative — BRAZIL. Minas Gerais: Serra do Curral, S « d Horizonte, Chase 9325 (US). 6975 (US) Rios d Pedras a Matto Grande, Glaziou 20070, 20076 (P); Belo Horizonte, Magalhães Gomes 908 (US-1762607). 22. се ева Nees ех Trinius, . Paspalum polyphyl- lum Nees, Fl. — Pl. 2: 41. 1829, hom. illeg. TYPE: Brazil. Chapada, “in campis siccis pr. 1824, С. Н. von Langs- dorff s.n. (holotype, LE!; isotypes, K! photo, SI!, US-2855764! photo, SI!). Figures 7, 22. Gram. Panic. 4. m Mariannam," Paspalum —— ue m. & Schult. var. tenue Doll, in Mart., Fl. Bras. 2(2): 67. 1877. TYPE: zl. Without lac ality, L Riedel d (lectotype, — nated here, W!; isolec totype, ( liri M Hack., бенен Во! 5 196. 1901 с 3razil. Rio de Janeiro: sion locality, T oh rec — D 41 (holotype, W!; type, US- с photo, SI!). Paspalum bic T Mez, Re ре rt. Spec 15: 27. TYPE: Brazil. Goiás: laco, casc se do Rio Macaco, 12 July 1895, А. E M. Glaziou 22563 (lectotype, oa дем В!; isolectotypes, BR not seen, K!, P! photo, SI!, US- 42158! 29421: L . Nov. Regni Veg. Faze Pe do Sir- Caespitose perennial, with long, arcuate rhi- zomes, cataphylls coriaceous, villous; culms (10—) 25-90 cm tall, 0.1-0.2 em diam.. rigid, branched or unbranched, erect, occasionally decumbent; in- ternodes 1-10 cm long, terete, hollow, glabrous: nodes brown, glabrous or shortly pilose. Sheaths 1.5-8 cm long, usually longer than the internodes, glabrous to hirsute, with tuberculate hairs, the mar- gins glabrous or shortly ciliate. Ligules 0.4—2.5 mm long, membranous, brown, glabrous; pseudoligule, when present, a tuft of rigid hairs 2.4—5 mm long; collar glabrous to shortly pilose. Leaf blades linear linear-lanceolate, 4-13 cm long, 0.1-0.6 cm wide, flat, rigid, ascendent, glabrous to papillose- pilose on both surfaces, narrow or rounded at the base, the apex subulate, the margins short to long papillose-pilose. Peduncles exserted, up to 30 cm long, terete to filiform, glabrous to hirsute in the distal portion. Terminal and axillary inflorescences exserted; terminal inflorescences 4-12 cm long. 0.4—3(—4) cm wide; main axis absent or 1—4(—7) em long, sulcate, glabrous to hirsute; pulvini shortly pilose, with white hairs up to 0.6 mm long; axillary inflorescences 4—7 cm long, 0.5-3 cm wide; ra- cemes 1 to 4(8), ascendent, slightly divergent from the axis, straight, alternate, ending in a nake point, sometimes in a spikelet; rachis of the ra- cemes 3-10 cm long, 1-1.2 mm wide, pilose or with adaxial surface glabrous, green or tinged with purple, with anastomosed veins, the margins sca- brous and covered with long hairs; pedicels un- equal, up to 1.7 mm long, triquetrous, hispid; spikelets paired, densely arranged in 4 series, sometimes lower spikelet aborted. Spikelets long- 1.4-3.8 mm long, 0.6-1.2 mm wide, acute, pilose, the margins rigid, purple, covered ellipsoid, with long, ciliate spreading hairs; upper glume and lower lemma as long as the spikelet, hyaline. Upper glume with corky margins covered with radiate, white hairs 2—7 mm long, alternating with shorter hairs 0.8-2 mm long, otherwise shortly pilose, 3- nerved. Lower lemma hyaline, the margins shortly pilose, hairs up to 0.5 mm long, the rest of the surface glabrous to sparsely pilose, convex and de- pressed toward the base, 3-nerved. Upper anthe- cium long-ellipsoid, 1.2-3.6 mm long, 0.6-0.8 mm wide, acute, membranous, pale, with bicellular mi- crohairs and simple papillae all over the surface and macrohairs and silica bodies toward the apex of the upper lemma and palea, upper lemma 3- nerved; lodicules 2, ca. 0.2 mm long, conduplicate, hyaline; stamens 3, anthers 0.8-2 mm long. Cary- opsis long-obovoid, 0.8-1.3 mm long, 0.4—0.6 mm wide; hilum linear, % the length of the caryopsis; embryo Y to Y the length of the caryopsis. Common name. — "Capim-lanoso" (Brazil). Distribution and habitat. From Venezuela and Colombia to central and southern Brazil, Bolivia, Paraguay, northeastern Argentina, and Uruguay. This species grows in open fields and savannas, on Volume 89, Number 3 Denham et al. 387 2002 Paspalum subg. Ceresia nu ; ЦР | VA 7 Ss Don f Figure 22. — plan —A. Habit. —B. Detail of p E Portion of rachis. —D. Spikelet. = view. —E. Lemma, ventral у Е Upper anthecium, dorsal view Upper — ventral view. — Caryopsis, scutellar view. —I. С ‘aryopsis, hilum view. (Based on Smith y — 12106, US.) Annals of the Missouri Botanical Garden rocky, sandy soils, occasionally in swamps, between О and 2500 m. Paspalum polyphyllum shows a high degree of plasticity in its vegetative characters, such as plant height, branching, blade pilosity, and size of inflo- rescences and spikelets. This species has two dif- ferent growing patterns; one of them, with simple culms bearing larger inflorescences and spikelets of different size, from 2.6 to 3.6 mm long (e.g.. Swallen 8772, Smith & Klein. 12106, Hassler 11561, Schulz 17275, Archer & Barreto 4995, and Fiebrig 6222) more than 4 13933, which also have inflorescences more than mm, e.g.. Rojas 19 em long, racemes 11-13 cm long, and spikelets 4—4.2 mm long. The other phase is characterized by its culms profusely branching at the middle nodes, with smaller inflorescences and spikelets ca. 2 mm long (as in /rwin et al. 16232a, 17256, Kil- leen 2795, Hunt & Ferreira Ramos 6012, Hunt 6246, and Chase 9148). Also, both phases can be Swallen 9150, where simple, erect culms up to 40 cm tall, with found in a single collection, e.g., spikelets 3 mm long, are mixed with more delicate culms in which spikelets are 2 mm long. It is note- worthy to mention that the type material of P. po- lyphyllum belongs to the second phase of devel- opment. When studying type material of Paspalum bici- lium and P. macroblepharum it became evident that they belong to the branching phase of P. polyphyl- lum, with spikelets up to 2 mm long. Doell (1877) cited, following the concept of Tri- nius (1828-1836, figs. 134, 144). P. ble- pharophorum and P. distichophyllum what here is under considered P. polyphyllum, including in P. blephar- ophorum specimens of the unbranching phase and in P. distichophyllum specimens of the branching one. Both species are here considered synonyms of P. humboldtianum (see synonymy under this spe- cies). Doell (1877) cited three syntypes when describ- ing P. blepharophorum var. tenue, of which Riedel s.n. was here selected as lectotype of the species; of the remaining syntypes, Sellow s.n. also belongs to P. polyphyllum, and D'Orbigny 168, from Chu- quisaca, Bolivia, probably belongs to P. humbold- tianum, The number of racemes is higher, ranging from 4 to 12, in the specimen Chase 9369. Mez (1917) cited two syntypes when describing P. bicilium, Glaziou 22562 and 22563, of which the latter has been selected as lectotype of the species. According to Rosengurtt et al. (1970) this spe- cies is a low-vield forage grass. Representative specimens. ARGENTINA. Chaco: Dpto. San Fernando, Fontana, Meyer 2090 (LIL); Dpto. 1° de Mayo, Colonia Benítez, * 17275 (G, MO, SI). Corrientes: Dpto. Paso de Los Libres, Campo Militar . Carnevali 3200 (CTES, LIL). Misiones: Dpto. San Ignacio, sobre Ruta Prov. 210 m, a 7 km Ayo. Horqueta, Honfi 98 (LIL, MO); Puerto Nuevo, Schwarz 2214 (B, LIL, US). BOLIVIA. La Paz: Puente Villa-Yun- gas, Cárdenas 3606 (US). Santa Cruz: Prov. Chiquitos, S slope of the M d _ Santiago, Daly et al. 2171 (MO, ) iqui ia de Santiago, Killeen 2 (MO); Buena Vista, St 6 (BM, x K, MO, W, US). BRAZIL. Bahia: Mun. Palmeiras, Pai Inácio, e 242, km 232, Mori 13303 (К. MO); Dpto. Palmeiras, Morro do Pai Inacio, Zuloaga & Morrone 6944 (IBGE, SI). Distrito Federal: a 32 km do CENARGEM, Borges Dias Vieira 189 (MO); Parque icipal do Gama, ca. 20 km S of Brasília, Irwin & Soderstrom P (MO, SI, US). Goiás: Serra do Caiapó, ca. 33 S of Caiaponia on road to Jataí, Irwin & Sode "rstrom 7069 (К. MO, SI); between Jataf and C aiaponia. 45 km fre КО иу нсә Нит 6246 (К, US); Serra Azul, about 7 km W of the Barra do Garcas, — & Ramos 6012 (К, US). — Grosso: ca. 6 km S of Xavantina, Argent et al. 6460 (P, US); Serra do Roncador, tal. 162: Za (MO); summit, Serra Azul, — Glaziou 20082 (P, US, W); Serra do Cipó, Chapeu de Sol, Santa Luzia, ris & Barreto 4995 (US); Serra 110 km NE of Belo Horizonte, Chase 9148 (M0); Preto, Villa Rica, Chase 9369 (MO, US). Paraná: I: NE of Atuba (which is n outside of — along Hwy. ‚ Davidse et al. 10967 (MO); Po ^R Gros "azenda de Chacon, Swallen 8772 (US). Rio le Janeiro: а rade & Vianna 20377 (US); environs de Rio de Jane Glaziou 17940 (G. MO, W). Rio Grande do Sul: Mo n- tenegro, Araujo 144 (US); Pelotas, Swallen 9150 (US). Santa Catarina: Mun. Mafra, campo 10 km NW on the road to Barrac ‘as (20 km), Smith & Klein 12106 ж ) Sao Pauli = ~ — íchica, Sarai 4186 (COL). Santander: , Robinson « Beltran 3189 (US). PAR- AC SU, AY. Alto Parana: pi re — ои Alto Р агапа, Fiebrig 6222 (BAA, BM, US). Ama — in campis arenosis Estrella, — nod (G, US). uaz: Caaguazú, Balansa 90a (G, P). Cer iral: E re- gione lacus Ypacaray, Hassler 11. 501 (MO). Cordillera: ед Colonia Pedro Caballero, Rojas 13404 (LIL- 61328). Guairá: Borja, Montes s.n. (LIL-561328). Par- i: Cerro Peron, prés de deca Balansa 90 (BM, G, P. US, W). URUGUAY. Art — (BAA, W). Cerro Largo: — Herter US). Flores: Río Yí y Ayo, Matanzas, — s.n. a 1723957). Paysandú: Río Uruguay y Chapicuy, Rosen- — + n. (LIL, US). Rivera: Tranqueras, Castellanos 17922 (LIL). Tacuarembó: without locality, Osten 6632 (SI, ic 5). VENE A. Lara: Distr. Carretera de Humocaro Bajo via Buenos Aires, 2 er W erff & Rivero 7863 (MO). Trujillo: | a C riste ur — 1867 (US). 23. — pa ) ~ Paspalum reticulinerve Renvoize, Kew Bull. 50: 339, fig. 1 A-G. 1995. TYPE: Boliv- ia. La Paz: Iturralde, Siete Cielos, Río Manupare, ca. 2.5 km al E, 12°27’S, 67?37' W, 180 m, 8 June 1987, J. C. Solomon 17003 (holotype, LPB not seen; isotypes, K!, MON. Figure 1 Prov. Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia 389 Shortly rhizomatous perennial; floriferous culms (70—)90—120(—140) ст tall, 0.3 em diam.. branched or unbranched: erect. internodes finely striate, glabrous, hollow; nodes 5 to 7, compressed, brown, glabrous. Sheaths 15-24 cm long, usually longer than the internodes, keeled toward the upper por- the inner surface red- tion, glabrous to hirsute. tinged-brown, shiny, the margins membranous, gla- brous. Ligules membranous, 1-2.5 mm long. brown, glabrous: pseudoligule absent or a tuft of 17- 40 em long. 0.3-0.8 em wide, flat, glabrous to hir- hairs 5-7 mm long. Blades linear-lanceolate, зше, the base attenuate and the apex acute, the margins involute, with long hairs toward the lower portion, otherwise scaberulous. Peduncles up to 40 em long, terete, glabrous. Terminal inflorescences exserted, 5—15 em long: main axis absent or up to l em long, flat, glabrous: pulvini shortly pilose: ra- cemes (1)2(3). i point; naked 1-1.8 mm glabrous 1 5-15 cm long. ending in ¢ rachis of the racemes winged, wide, green or tinged with purple, ) sparsely pilose, the midnerve scabrous and margins scaberulous to papillose-pilose; pedicels unequal, flat. scaberulous: spikelets paired, densely imbri- cate in 4 series. Spikelets broadly ovoid, 4.8-6 mm long. 3—4 mm wide, dorsiventrally compressed, pla- no-convex, winged, cordate, ciliate on the margins of the upper glume. Upper glume as long as the spikelet, 7(—9)-nerved, winged, with just 3 central nerves reaching the apex, membranous, ciliate, cor- date, the apex acute. Lower lemma 3.5—1.4 mm long, 2 mm wide, 3-nerved, glabrous, coriaceous, corrugated and rounded at the base, the margins winged, shortly ciliate. Upper anthecium ellipsoid, 2.8-3.6 mm long, 0.8-1.5 mm wide, plano-convex. papyraceous, pale; upper lemma with simple pa- pillae, bicellular microhairs, and macrohairs toward the apical portion; upper palea with papillae and bicellular microhairs; lodicules 2, ca. 0.4 mm long, hyaline, conduplicate; stamens 3, anthers 1.6 mm long; styles 2. stigma plumose, lateral. Caryopsis obovoid, 1.6 mm long, 0.8 mm wide: hilum elliptic. embryo Y the length of the caryopsis. Iconography. Renvoize, Gramíneas de Bolivia: 450, fig. 97 Distribution and habitat. Present in Bolivia, in La Paz and Santa Cruz, and Brazil. in the states of Pará, Tocantins, and Mato Grosso. It grows in open fields and “campos rupestres," between 100 and 900 m. Paspalum reticulinerve is a member of section Pectinata, where it is related to P. aspidiotes and the latter spikelets and inflorescences with (2)5 solitary to 7(11) ra- Р. imbricatum; has glabrous, cemes. When describing P. reticulinerve, Renvoize (1995) related this species to P. aspidiotes, from which this author distinguished P reticulinerve by the densely pubescent, attenuate leaf blades, a cil- iate rachis, and smaller spikelets. Later, Renvoize (1998) segregated Р. aspidiotes from P. reticulinerve by having glabrous and cordiform leaf blades, the rachis scabrous, and a smooth upper anthecium. The study of type material of both species, and specimens from Bolivia and Brazil, allowed us to conclude that the real identity of P aspidiotes had been misunderstood: this taxon is clearly separated from P. reticulinerve by having solitary spikelets 6.5—7.5 mm long. Also we concluded that the pi- losities of leaf blades and inflorescences are unre- liable characters to segregate both species: in the specimen Kuhlmann 1671 the rachis of the ra- cemes is sparsely pilose, with hairs up to 0.4 mm long, while in Killeen 6541 pilose and attenuated leaf blades are present together with a glabrous ra- chis of the racemes. A similar situation is present in the specimen Plowman et al. 8949, with the ra- chis of the racemes sparsely pilose. Representative specimens. BOLIVIA. Santa Cruz Prov. Velasco, Parque Nacional Noel Kempff M., Serranías de Caparuch, Killeen 6541 (MO); жа Nacional Noel kempff M.. Killeen et al. 4824 (SI). BRAZIL. Mato Gros- so: Serra y d estrada do Di: amantina, Kuhlmann 1671 (US). rabá, Serra dos Carajás, Cavale — 2131 (US): e fone eicáo do Araguaia, about 4 km V r along hwy. PA-150 pb (MO, US). Tocantins: Mun. Сошо M: — 5; Maus iro along GO-70, Plowman et al. 9133 (MO, US of town cente Plowman et al. 24. Paspalum stellatum Humboldt & Bonpland ex Flüggé. Gram. Monogr., Paspalum 62 1810, as “Paspalus stellatus.” TYPE: Colom- bia. Without locality, А W. H. A. ron Humboldt A. J. A. Bonpland s.n. (holotype, B-W not BM!, US-80051!). Figure 23. & seen; — TE stellatum Humb. & p ex Flügg ke mon- / Fl. Br РІ. 2: 78. 1829, as a " TYPE: achyum Nees, ras. Enum. I 5 2 'aspalus stellatus var. — һу zil. Without locality, £ Sellow 2 esp е. M not seen; isotypes, B not seen, P!, 052! photo, 511). Paspalum ste — — oe ex и var. dis- . Enum. PL 2: 78. 1829, as TYPE: Brazil. G. H. von Langs- — ме pi var. ye ds” “In apricis arenosis prope — x dorff s.n. (holotype, LE not se Gram. = Paspalum cujabense Trin., Sp. 3: pl. 284. 1829— 1830. TYPE: Brazil. s Grosso: Cuiabá, in gra- minosis — 1829 Н. von Langsdorff s.n. (holotype, LE! isotypes, - photo, SI!, US-80050!). m o Schltdl., Linnaea 26: 13. 133 854. TYPE: zuela. Distrito Federal: Surface- . Wagener . 396 tà ype not seen). Paspalum ас Нас К. var. sphacelatum Hack., Oes- 51: 239. 1901. TYPE: Brazil. Goiás: 390 Annals of the Missouri Botanical Garden ж P malmeanum У Р niquelandiae W Р stellatum | e Г longiaristatum | | km \‏ 1000 800 600 400 200 0 \ لہ مایم کا ق | Шы 200 300 400 500 600 miles |‏ ° d bs Hendrik М Ks phem \ Figure 23. Santa Luzia, 8 Abr. 1895, A. КОМ. — т — W!: US isoly ypes P!, US-2856 ene ш. Ий, & le ex ¢ Flüggé f. hirsuta ack., in Stuck., Anales Mus. Nac. Buenos Aires . TYPE: Argentina. Chaco: Resistenci ча, ‚ Н. Mercenaro s.n., leg. T. J. a 4 holot type W!: isotypes, CTES!, SI! | a photo, SI!). Caespitose, shortly rhizomatous perennial, cata- phylls pilose; culms 25-100 ст tall, 0.1—0.2 em diam., unbranched, erect to geniculate, fasciculate at the base; internodes 3-18 cm long, terete, hol- low, glabrous: nodes 3—6, brown, glabrous to shortly pilose. Sheaths 4—8 cm long, striate, glabrous to hirsute, with tuberculate, deciduous hairs, the mar- gins membranous; basal sheaths overlapping. Lig- ules membranous, short, 0.2 mm long, brown, gla- brous; pseudoligule a tuft of short hairs intermixed with long hairs up to 0.5 mm long; collar glabrous. Leaf blades filiform, 8-27 cm long, 0.1-0.3 cm wide, mainly basal, ascendent, rigid, narrowed at Distribution of Paspalum longiaristatum, P. malmeanum, P. niquelandiae, and P. stellatum. as the spikelet. the base, apex acuminate, the margins involute, the basal ones ciliate, both surfaces sparsely papillose- hirsute covered with long, tuberculate hairs, or the adaxial surface glabrous. Peduncles up to 25 em long, glabrous to shortly hirsute. Terminal inflores- cences exserted; main axis absent; pulvini shortly pilose, with (1—)2 papyraceous bracts, up to 4 mm long, at the base; racemes 1, rarely 2 conjugated, falcate, ending in a naked point; rachis of the ra- cemes broadly winged, 2-15 ст long, (4—)5—10 mm wide, glabrous to shortly hispid on the adaxial sur- ace, green to purple, midnerve prominent, the mar- gins membranous to hyaline, nerveless, brown, flat to plicate and enclosing the spikelets; pedicels 0.1— 0.2 mm long, shortly hirsute, with radiate hairs at the apex: spikelets solitary, arranged in 2 series. Spikelets ovoid to ellipsoid, 2.4—3.9 mm long. 1.2— 1.6 mm wide, plano-convex, acute, pale or tinged with purple toward the apex or all over the surface, villous, the upper glume and lower lemma as long Upper glume membranous, 2-3- — Volume 89, Number 3 Denham et al. Paspalum subg. Ceresia nerved, the central nerves inconspicuous or absent, with a tuft of hairs toward the base, glabrous on the rest of the surface, the margins papillose-pilose. with white hairs up to 4 mm long. Lower lemma glumiform, 2-nerved, the nerves marginal. Upper anthecium narrowly obovoid, 1.8-2.9 mm long, 0.8—1 mm wide, 0.4—1 mm shorter than the upper glume and lower membranous, pale. smooth, glabrous, shortly stipitate, easily deciduous at maturity; upper lemma 5-nerved, smooth, with lemma, sparse bicellular microhairs; lodicules 2, 0.2 mm long: stamens 3, anthers 1.8-2 mm long. Caryopsis obovoid, 1.6 mm long, 0.8 mm wide; hilum elliptic; embryo Y the length of the caryopsis or slightly shorter. U.S. Natl. Herb. Las Gra- Iconography. Chase, Contr. 28(1): 16, fig. 1. 1929. mineas del Distrito Federal: Burkart, Fl. Il. Rosengurtt et al., Luces de Febres, 115, 65. 1963. ‚ fig. 156. 1909. Gramíneas uruguayas: 348, fig 151. 1970. Pohl. Fieldiana, Bot. 4: 454, » 171. 1980. Smith et al., Fl. Il. 922, fig. 185. 1982. Rodríguez, Ernstia 8(2—3): 46. 1998. Renvoize, Gramíneas de Bolivia: 452, fig. 98. D. 199 — name. fig. Entre Ríos: 377 n.s. Catarinense: “Capim-estrela” (Brazil). Chromosome number. 2n = 32,52 (Honfi et al.. 1991[1990]); 2n = 20, 32 (Killeen, 1990). Distribution and habitat. From southern Mexi- co, Central America, and the Caribbean to South America, where it is found in Venezuela, Colombia. Bolivia. Brazil. Paraguay, and Argentina: cited for Uruguay by Rosengurtt et al. (1970). It grows in fields or savannas on sandy or rocky soils, or lat- eritic outcrops, between 0 and 2000 m elevation. Paspalum stellatum is related to Paspalum cer- esia, P. eucomum, and P. malmeanum by its winged and nerveless rachis of the racemes, with spikelets densely pilose. Paspalum ceresia differs by its glau- cous plants, with filiform to linear-lanceolate leaf blades, 7-21 em long, 0.3-1.522) em wide: inflo- rescences with two to seven racemes (rarely a sol- itary raceme), alternate, without a bract at the basal pulvini; and pedicels hispidulous and upper anthe- cium not stipitate. Paspalum eucomum is distin- guished by having two conjugated racemes (rarely a third one) per inflorescence, rachis of the racemes 1.5-2.5(—4) mm wide, and upper anthecium as long as the upper glume and lower lemma. Paspalum malmeanum differs by its rachis of the racemes 1.2-2 mm wide, spikelets 1.6-2 mm long. and up- per anthecium as long as the spikelet. Compound papillae are occasionally present in the upper anthecium of specimens from Santa Cruz, Bolivia. According to Rosengurtt et al. (1970) and Smith et al. (1982), this species 15 a low-yield forage orass grass. Representative specimens. ARGENTINA. Chaco Dpto. 1° de Mayo, Colonia Benítez, Schulz 300 (LIL). Cor- . Goya, Paraje San Isidro, 46 km S de Goya, 1 E TES, K, LIL). Entre Rios: Dpto. , barrancas del Río Uruguay, Bur- (SI). Formosa: — we with- y. Joergensen n 2882 (C ‚ MO, SI, . Misiones: ija, Ekman МУ (US); Dpto. a, Montes 1953 3 (G, LIL, P. US, W). Beni: Prov. Iténez, Magdalena, Moraes et al. 1713 (L PB, SI). La Paz: Chaquimayo-A polo trail ca. 15 y & Foster 71186 (MO). Santa O rientes: Dpto Ahumada etal. 354 — Cruz: Pr dase San Ignacio, Beck & Seidel 123 (MO. SI); Nuflo de Chávez, Concepción, Killeen 2474 (MO, SI, US). BRAZIL. Amazonas: Uraricuera. Myers 3404 (US). Distrito Federal: Brasilia, campus da UNB, de Jesus 16 (US). Goiás: Rio Cristal, 44 km by road SE of Cristalina, Anderson E (ВМ. MO, US): without lo- cality, Glaziou 22546 (G, K, P, W). Maranhao: Mun. Lor- eto, Ilha de Balsas region, * & Ейеп 4417 (K, US). Mato Grosso: Mun. Terenos, Faz. Modelo (IPEAO). Al- MO); ca. 270 km N of Xavantina, Ratter уз en 1915 (К, US). Minas Gerais: Uberaba. Chase ), US). Parana: Jaguariaiva. Dusén 18006 (BM. No Es Rio Grande do Sul: São Gabriel, Araujo (US). Roraima: Campos Gerais ahi Regiáo de € x. Fidei 23626 (US). Santa Catarina: Chapecó. 9 km W of Campo Ет, Smith & Klein 11529 (US). Sao campos d'Ipanema, — 13334 (P, US). CO- . Arauca: 13 km al 5 de Arauca, laguna El Ve- pe rgensen 52 (C OL). үн ен Cerro San. Lucas. uie E-747 (US) »yaeá: Llanos ш ntales, El Yopal al S del Hato Matapantano, Blydenstein & Saravia 1186 (COL). Magdalena: Santa Marta, Smith 142 (BM, K, MO, US). Meta: Llanos Orientales, — de la selva de Arau- ca, Blydenstein 736 (COL). e de Santander: Los Estoraques, La Playa, Balick (C OL). COSTA RICA: MA : km » Carretera Interamericana along the road Boruea, . 11594 (US MINIC AN RE PU BL IC. Dajabón: Cordillera Central, 6 km de Partido al W en la carretera a Dajabón, Zanoni et Cara- oen LOME al. 31952 (US). Monte Cristi: Sabaneta, Leonor, Valeur 198 (MO, US). HAITI. Circa Carrajal, Holdridge 1858 (MO, US). du Nord: vicinity of = Michel de PAtalave Leonard 7537a (US). EL SALVADOR. San — Volcán de San Salvador, Calderdn 2272 (US). JATE- MALA. Jalapa: between Monjos and — a 10 mi. S of Jalapa. Steyermark 32212 (US). HO Co- E — Francisco Шеше, near mayagua: Siguatepeque, Clewell 3514 MO). road to are Swallen EE F (MEXU, US). km by road SE of Rio Agua Cz ‚ Pohl € a do (MO). MEXICO. Chiapas: near Caucua, Nelso (US) Mun. Villal Corzo, above Colonia Vincente c uerrero on road to Finca Cuxtepec, Breedlove & — 54586 (MEXU). Oaxaca: Distr. Inquila, Conzali 4364 (MEXU, US). PANAMA. F clé: between Paso de eh and Ola, Pittier 5020 (US Panama: ca. 10 km SW of San Carlos along the ns T- American Hwy., Davidse & D'Arcy 10126 (MO). Vera- guas: on Santiago—Santa Fe rd., 18 km S of Santa Fe, Nee Annals of the Missouri Botanical Garden 8166 (МО). PARAGUAY. Alto Paraguay: Fuerte Olim- po, Anisits 2047 (Р). Amambay: — y al sur de Bel- avista, Fernández Casas € Molero 6335 (G); in regione calcarea cursus * rioris — pes Hassler 11058 (BM, ( ч Р, US, W). Caaguazú: С aaguazú, Hassler 9259 , LIL, “MO > W. "Ca aazapa: Tavaí, Basualdo 2162 — MO, SI). € entral: ' larumandy, Schinini 6207 (G, LIL, SI, US). Cordillera: Barrerito, Anderson 1120 (US). Guairá: Villa Rica, Joergensen 4568 (BM, MO, SI. US). Itapúa: gr — 6 km de la Ea. Melgarejo. Quintum et al. SI). Misiones: San Ignacio, Burkart Paraguarí: Cerro Peron, prés de Paraguarí. — — G, K, P, US); C ordillera de Altos, Fiebrig 1 (BM, G, K, P, SI, US, W). VENEZUELA. Aragua: М stween S. don de los — and U berito, Pittier 11326 I5). Barinas: н km SW of the Mé He intersection just E of Barinas, Davidse 3165 (MO). Bolivar: Divina Pastora, Timaya 2881 (US). Carabobo: V — near Guacara, Pittier — (US km W of San Carlos along Hwy. Migne Federal: Caracas and vicinity, Bailey & Bailes 2 (US). Lara: Dpto. Palavecino, between the quebrada La а “Mata and the E side of the ы А Nacional Terepaima, Burand Jr. & а b > (MO). Miranda: Colonia Tovar, Escalona et al. 1 (MO). Portuguesa: Dpto. Guanare, terrenos de la Ho Ramírez 1789 (MO). 18242 (SD. 25. Paspalum trachycoleon Steudel, Syn. РІ. Glumac. 1: 28. 1855 [1853]. TYPE: Venezue- la. Valencia, 9 Oct. 1843, Н. C. Funck 742 (holotype, P!; isotypes, BM!, K!, G!, P!, US- 80059!, W!). Figures 24, 25. Shortly rhizomatous perennial; culms erect to de- cumbent, (0.8—)1—2 m tall, 0.3-0.6 mm diam., branching at the middle and upper nodes; inter- nodes 5-15 cm long, terete, glabrous, rigid, hollow; nodes pilose to glabrous, brown. Sheaths usually longer than the internodes, overlapping, pubescent to papillose-pilose at the upper portion. Ligules 2— 3 mm long, membranous, glabrous; pseudoligule absent. 10-20 en long, 0.5-1.5 em wide, flattened, pubescent to pa- Leaf blades linear-lanceolate, pillose-pilose on both surfaces, otherwise glabrous, the base attenuate and the apex acute. Peduncles terminal, partially exserted, pilose. Terminal inflo- rescences 5-17 cm long, 3-5 cm wide; main axis 3-11 cm long, pilose; racemes 4 to 12, ascendent, divergent, alternate, ending in a naked point; pul- vini pilose; rachis of the racemes 3-10 cm long, 4— 6 mm wide, foliaceous, green to purple, with a pu- bescent keel and anastomosed veins on the wings, glabrous, ending in a naked point; pedicels un- equal, hispid, laterally inserted in the spikelets; spikelets paired, densely arranged in series. Spikelets ellipsoid, 2.3-3.5 mm long, 1-1.1 mm wide, pale, pilose, with an annular thickening at the base. Upper glume as long as the spikelet, the margins pilose with white hairs, a few of them up acute to acuminate, hyaline to membranous, to 3 mm long, densely pilose toward the base, oth- erwise glabrous, 3-nerved. Lower lemma as long as the upper glume, acute, membranous, sulcate and with conspicuous nerves, 3-nerved, pilose at the upper portion, otherwise glabrous. Upper anthe- cium ellipsoid, 2-3.1 mm long, 0.8 mm wide, pla- no-convex, membranous; upper lemma with simple papillae and bicellular microhairs all over the sur- face and а tuft of apical macrohairs; upper palea with simple papillae and bicellular microhairs to- ward the distal portion; lodicules 2, reduced; sta- mens 3, anthers 1.3—1.4 mm long; stigma plumose. Caryopsis not seen. Chromosome number. = 20 (Davidse & Pohl, 1974). Distribution and habitat. co, Guatemala, Honduras, and El Salvador to Ve- nezuela, Colombia, and central and southern Brazil. From southern Mexi- It grows in mountain savannas, on dry mountain slopes, between 500 and 2000 m. Paspalum trachycoleon is close to P. phyllor- hachis in its similar habit and inflorescence type, which is distinguished by its glabrous spikelets. It is also related to P. heterotrichon, but the latter has the rachis of the racemes 3.54 mm wide and spikelets solitary. The specimen Killip & Smith 19038 has spike- lets 3.8-4 mm long. Representative — BRAZIL. Distrito Feder- al: Reserva Ecológica do IBGE, da Silva 2659 (MO); ipa Ribei irão Torto, NE of Lagoa Paranoá, Irwin et al. 15. | Goias: Niquelándia, Filgueiras 3567 (SD; sed E * — Burchell 7061-2 (K), Gardner 4390 (BM, G, К, Minas Gerais: without locality, Glaziou 20079 (К). loc — without Cordil- ЕШМШ, Cundinamarea: without collector 80 (US-1574261). Huila: este de Neiva, lera Oriental, Rusby & Pennell 1014 (US). Estoraques, La Playa, Balick T (C icini › & Smith ; vicinity of Charta, Killip & Smith 19038 (US). EL SAI VADOR. Copán: 3 km NW of Copán, Williams et al. 42957 (US). San Salvador: Volcans of San Salvador, Hitchcock 8946 (US). GUATEMALA. Zacapa: slopes of qee Virgen, Steyermark 42586 (US). HONDURAS. El m — to Yusc aran. Swallen 11353 А со Morazan: Rio de Gallo, near El Jicarito, Swallen. 10993 (MEXU. MO, US). Olan- cho: Jutiapa — Camp, Pohl & Gabel 13761 (MO). XICO. Chiapas: Mun. Cintalapa, 23 km of Las Cruces — ey to La Mina Microwave Ballon; Breed- love & Davidse s.n. (MEXU-830597); Mun. Villa Corzo, above б, Vincente Guerrero on road to Finca Cux- Breedlove & Davidse 54564 (MEXU). VENEZUE- nzoategui: Distr. Libertad, road from El Vigia to Bue nos Aires, Davidse & González 19487 (MO). Aragua: de Maracay a Alto, Croat 21469 (MO). Distrito Federal: Caracas, Bailey & Bailey 447 (US). Jimenez, e ntre — del Viento y Cerro Pando, Davidse & González 21162 (SI). Miranda: Sebastopol, Badilla 100 tander: Los lepec, Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia Portion of rachis. —D. Spikelet, dorsal 'pper anthecium, ventral view. Н. amayo ycoleon.—A. Habit. —B. Detail of ligule. С. i iew. —G. 1 lse & Tillet 4514, MO; В—1 based on 7 ispalum trach ventral view. =. Upper anthecium. dorsal Caryopsis, scutellar view. —L Caryopsis, hilum view. (A based on Dari let Figure 24. Pe view. —E. Spike 1084, US.) 394 Annals of the Missouri Botanical Garden | Ф P imbricatum | e Р trachycoleon | | \ à 200 400 600 800 1000 km \ \ — — — о ады 200 300 400 500 600 m: \ pared by Hendrik R. K DERE. Es 7-96 — — Figure 25. 5 — Distribution of Paspalum imbricatum and (US); San н de Morros, Davidse & Tillet — (MO). Mona : S-faci oe mountain slope direc Piele et gu ithout depar cas, Tamayo d MES Los Venados, amaro. 2046 (US). EXCLUDED SPECIES Paspalum sanguinolentum Trin. was placed in the subgenus Ceresia by Chase (1929), but this spe- cies belongs to subgenus Paspalum, group Erian- tha. Literature Cited Arriaga, М. О. 1987. Interpretación rA ornamento de antec io de pm ‘hloa (Poaceae). Bol. : 13 xs Beetle; A. — Noteworthy grasses from Mexico X. ai 52: 11-17 . Argent. Bot. Paspalum, Notes on Gramineae. J. ‚19: 34—3t — h support and tree stability. Cla- Brooks. R. R. ps m ntine and Its Vegetation. Dios- corides Press, Portland. Р, trachycoleon. Burkart, A. 1969. Gramíneas. /n: A. ir — ra de Entre . Agropecu. 6(2a): 1 Canto-Dorow, T. S., Flo- . Nac. Tec- — ч i ;olecc. pm Wagner & J. F. Monte- negro Valls. 1996. Re visão taxonómica das espécies de Paspalum L. grupo Notata (P o i Rio — do Sul, Brasil. Iheringia, ic 17: Chase l In: tcheoc A "s — ue bn Peru, and — a U.S. Natl. Herb. 24: 434—455. . 1929. The North * rican = ies of Paspalum. U.S. Natl. Herb. 28: 1—31( . Ined. Paspalum of eu о a. Hitchcock апа 227. Paspalum. Contr. Chase Library, — Departme nt, Smithsonian Insti- tution, a on, D. Cialdella, A. M., O. — & F О. Zuloaga. 1995. Re- visión A las especies del género Paspalum (Poaceae deris Paniceae), Grupo Bonplandiana. Darwin- iana 33: 67-95 Clayton, W. i & S. A. Renvoize. 1986. Genera Gramin- um. Grasses of the World. Kew Bull.. Addit. Ser. 13: 1-38 la Silva, T. S., A. G. Burman & T. Sendulsky. 1979. Es- pe ecies de Paspalum L. da Ilha do Cardoso, estado de Sao Paulo. Hoehnea 8: 11—28. ا Volume 89, Number 3 2002 Denham et al. 395 Paspalum subg. Ceresia Davidse, & R. W. Pohl. 1972. Chromosome numbers and notes on some M entral American grasses. Canad. J. Bot. 50: 273-283. — — . Chromosome numbers, meiotic be havior and notes on — ‘al American grasses. Can- ad. J. Bot. 52: 317—328. & . |“ 978. C ра — of tropi- — — * cal American grasses (Gramineae). . Missouri Bot. Gard. 65: 637-64 Doell. J С. pu Graminae. /n: C. F. P. Martius к Fl. Bras a 1-358. Munich, — Leipzig Farris, J. 3 989. The retention index and hee rescaled consistency — x. Cladistics 5: 417—419 ‚ V. A. Alberte, M Kalle sjö, D. — & А. G. РА 1996. — jackknifing outperforms ne ор Cladis es 12: 99-124. Filgueiras, T. S. & ( Davide 1994. Paspalum biarista- tum (Poaceae: — ае), from Goiás, Brazil, and the second awned species in Novon 4: 18— & «E new serpe ntine е ndemic E the genus. А О. c 1993. Ophiochloa. a new endemic serpentine g Pani- ceae) from the Brazilian cerrado vegetation. Novon 3: 310—317 erass genus (Poaceae: Fliigeé, J. 1810 Graminum Monographiae. . .Pars |. Pas- palum. Reimaria. V. Perthe es < J. Н. Besser, Hamburg. Giussani, ү . J. H. Cota-Sánchez. F. O. Zuloaga « К. . Ke — \ шейге ular phylogeny of the grass А ub — ‘ade sae (Poaceae) shows multiple origins of ( ооа Amer. J. Bot. 88: 1993-2012 A. 1993. Е anaing с — ‘ter — — m n. Re ^^ т w | .C€ Mica : 07a. NONA version 2.0 for Windows. ,Onm- — r pogran and doet 'ume ntatión — by the au- thor, eee site: htt 97 H Pee - Wee version 3.0 ‘for > Windows. Com- — r program and doc — eens by the au- thor, website: idistic Gould, F. W. & T. n. e 1967. КО num- A of tropic ‘al American grasses. Amer. J. Bot. 54: 7 die » Hac ke. - 1901. Neue Gräser. Oesterr. Bot. Z. 51: 233— Manual of the Grasses of the Unit- 'er Publications, Ma w Yor Holmgren, Р, K.. | . Holmgren & L. Eo tt. 1990. Index He Pd Part | : The ا‎ of the World. New York Botanical Garden: m York. Honfi, A. J. C. Quarín & J. Valls. cariológicos en gramíneas T ALS. Ii" ed States, ed. eng 1991 [1990]. sudamericanas. Darwiniana Ж 87-94. ` О. . Zuloags ga, Û. Morrone & А. Ese obar. canas (Poa eae Panic ol na) Darwiniana 35 —30. Judziewicz, E. J. 1990. Family 187. Poaceae be А R.A Gorts-van Rijn (editor), Flora of the Guianas. Serie P ing sad Koeltz Scientific Publications. ues nig- stein, Germany Karis, P. О. 1995. Cladistics of the subtribe Ambrosiinae (Asteraceae). = Bot. 20: 40-5 — T. J. 1990. The grasses e My cie Santa + via. fee Missouri Bot. Gard. 25-201 — D. 1991. The discovery and lon d of mul- deg islands of most parsimonious trees. Syst. Zool. 40: 315 Mez. C. 1917. pert. Spec. VIII. Generis Paspali species novae. Nov. Regni Veg. 15: 27-133 Re- Morrone, O., ). Zuloaga & E. Carbonó. 1995. Revisión del grupo zw emosa del | inue Paspalum (Poaceae: Panicoideae: Paniceae). Ann. Missouri Bot. Gard. 82: 82 -116 ‚ A. Vega & F. O. Zuloaga. 1996. Revisión de las especies del género Paspalum L. (Poaceae: Panicoi- deae: Candollea 51: Paniceae). grupo Dissecta (s.str.). 103-138. Jenham, S. S. Aliscioni & F. O. Zuloaga. 2000. Revtalón de ES especies de Paspalum (Panicoi- deae: Paniceae), subgénero Anachyris. Candollea 55: 105-135. Nash, G. V. 1912. P ae e © Tribe 5. Paniceae. North American Flora 1 a 1-196 Nees ab Esenbeck, С. Еч 1829, Кога — ‘NSIS seu ) 1 Martius (editor) 2(1): 1-608 “М. a nter. 1993. On demie Cla- distics 9: — 1805-1807. Synopsis Plantarum, 2 Vol. Pilger, R F. 1929. Bemerkungen zur Systemalik der Gathing Paspalum L. Repert. Spec. Nov. Regni Veg. 26: : — Pohl. | . 1980. Gramineae. In: W. и г (editor), Flora Costaricensis. Fie — Bot. 4: 350—392. Renvoize, S. А. 1978. The genus — um Group Lorea Segre — in the Gramineae: XLIII. Kew Bull. 32: 419-428 1995. А new species — Mos ا‎ (Gramineae) from Bolivia Kew Bull. 50: ЗА 998. Gramineas de мена Сайна Print- ing. н Rodriguez Ж aL nueva sección en el género Paspalum L. 1992. Pectinata Chase ex Rodríguez (Gramineae Ernstia 2: 21-2 1998. El — Ceresia (Pers.) Reichenb.. del género Paspaten L. (Gramineae) en Venezuela. Ernstia 8: 7-50. . B. R. Arrillaga de Maffei & P. Izaguirre de Artucio. 1970. Gramíneas Uruguayas. Publicaciones de : a Unive n de la Repüblic a, Montevideo. Rua. H. У. 5. Aliscioni. In pre A morphology- — dud a analysis of РЕ sect. Pectinata (Poaceae). Syst. Bot. К. 1948. Cariología ded gramíneas argentinas. vista Кас. Agron. Veterin. 12: 51-67 Sendulsky, T. & A. G. m 197 8. Pasgalión species of the Serra ds Ci Apó (1): A Bp to s study of the Brazilian Poaceae. Revista Brasil. : 1-15 280. Paspalum species ni the Ser гга А tan to bip study of the Brazilian Bot. 3: 23-3: йо к? Saura. Re- do Cipó (H): Poaceae. Rev ista Brasil. Smith, E. B... D. shades n wd Ң. M. Klein. 1982. — as. In: R. iss (editor), Flora Hustrada Catar- i Fas scículo Gram. 85. Paspalum até 115. Zea: м inense 910-1 10 оло. T. R. & the genus Olyra at s J olyra — ae: ies E Оте). Smithsonian : 1-79 — 1989. A revision of w segregate genus Parodi- 1962. A cytological study | E some Mexican Torrey Bot. Club 89: Tovar, O. 3. Las gramineas — Perú. Ruizia 13: 347-378 Trinius, €. B. 1826. De Graminibus Paniceis. Dissertatio 396 Annals of the Missouri Botanical Garden botanica altera. Impensis Ac a miae Imperialis Scien- tiarum, Sa oli (St. Petersbu — [8 18: 30. Species — iconibus et des- c iptonibus illustravit. 3 vols. Impensis Academiae Im- perialis Scientiarum, Petropoli. 1834, Panicearum genera retractavit specieb- usque ае illustravit Mé е Acad. St.-Pétersb. Sér. € . Phys 3(2): 89-35: — 0. & 0. ы. In prep. Ңеуїзїбп de las especies de Paspalum para América del Sur Austral (Argentina, Bolivia, sur de Brasil, Chile, Paraguay у Uruguay). APPENDIX | List of accepted species of Paspalum subg. Ceresia. 1. da en aspidiotes Trin. 2. P. biaristatum Filg. € Davidse 3. buchtienii Hac 4. P. burmanii Filg., Morrone & Zuloag 5. P. cachimboense Davidse, Morrone & EL 6. P. carinatum Humb. & Бор}, ex Flugg 7. P. ceresia (Kuntze) Chas 8. Pe 9. P. cymbiforme E. Fourn. 10. P. ees ex Trin. П. P Morrone & Zuloaga 12. P. hete 13. P. biis ciis Flüggé . P. imbricatum Filg. 15. P. lanciflorum Trin lo. P. D un Davidi & Filg. 17. P. malmeanum Ekman 18. P. ае Fi 19. P. 20. Р 21. P. phyllorhachis Hack. 22. P polyphyllum Nees ex Trin. . P. reticulinerve Renvoize 24. P. stellatum Humb. & Bonpl. ex Flügg 25. P. trachycoleon Steud. eucomum goyase nse Dav — trichon APPENDIX 2 Index to exsiccatae. Each : listed by the first collector. The number in parentheses refers to its associated species (see Appendix 1). An as- terisk (*) indicates a voucher for the o study. ; Acosta Solís 21454 (13); Ahumada 3541 (24); Alba 53-23 (19), 53-96 (6 M Allem 1558 (6), 2007 (6), 2236 (24); Almeida 572 (19); Alston 7569 (19); Amat 3 (15); Anderson 1120 (24), 8286 (24), 9413 (22), 9421 (24), 9969 (6), 10304 (22), 35213 (19), 35798 (22): Anisits 2047 (24); — 104. (19), 144 (22), 345 (24); Archer 629 (13), 3843 (24), 4995 (22); Arechavaleta s.n. (22); Argent 6460 (22), a (19); Aristeguieta 238 (25), 4219 (6); Arsène 2813 Asplund 6531 (13), 7593 * 10498 (13), 10633 no 12014 (13); Aynard 6019 ail 100 (25); Bailey 112 (2: h — (25); ۰ (24), 90 (22), 90a (22), 126 (24), 127 (25): Balick 1 (25): i 6923 (13); Bang 265 (: E B: on sa aa 3032 (19); Barclay 5703 (13); Barlow 30/1173 (19); Bar- reto 46 (24), 529 (22), 1483 (22). 1577 (24), 2952 (19): Basualdo 2162 (24); Beck 4025 (7). 8714 (3), 12330 (24). 19948 (3), 20552 (19), 20564 (6); Bentacur 1510 (19); Berro 2680 (22); Bertoni 5406 (24); Black 47-1777 (6). 51-12029 (19), 51-12148 (22), 51-12291 (6); Blake specimen is alphabetically TS 7 T — 22 7445 (19); Blydenstein 734 (19), 736 (24), 847 (6), 856 (19), 900 (19), 939 (8), 1037 (19), 1082 (6), 1188 (24), 1468 (15), 1562 (19); Boeleke 16646 (13); Borges 189 (22); Brade 19638 (8), 20377 (22); Breedlove 3313 (13), 13822 (9), жүр (9), 39106 (9), 42481 (25), 53954 (24), 54302 (12), 54468 (25), 54564 (25), 54586 — 54596 (12), s.n. (2: 5), Brooks 70 (24), 125 (7), reck 2 (6); Buchtien 447 (3), 3629 (3), < an. (i " B 3 V0495 (24), V0530 9— Bure hell 2 (22). 4574 (22), 4081 (22), 5769 (22), 7061—2 (25), 7124 (25), 7295 (25); Burkart 5335 (13). 10211 (13), 11014 (13), 16984 (13), 18242 (24), 21610 (24); Burman 393 (24). Caballero Pardo 77 (7); Cabrera 2278 (8), 22730 (7), 28254 (24); Calderón 2272 (24), 2708 (8). 2760 (1); Cal- lejas 6886 (15); Calzada 17051 (19); € ramp E-2743 (13), 2332 (7); Capell s.n. (22); Cárdenas 515 3s 3410 (13), 3606 (22); Carnevali 1083 (24), : 3200 (22), = *— = (24), 9148 (22). 9212 (19), 9250 Ts 9953. (24), 9205 (24), 9308 (24), э (21), 9352 (22), 9367 (22), 9369 (22), 9559 (22), 9917 (22). 10230 (19), 10249 (19), 10276 (19), 10306 (6), 10850 (19), 10595 (6), 10655 (19), 10740 (22), 10872 (24), 10907 (24), 11200 (10), 11543 (24), 11549 (22), 11602 en 11768 (24), 11962 d 12014. oe 12028 (24), 12085 (2 5), 12088 pies y 2292 (25), 12294 (13); Chávez 3217 үз Claussen s.n. (24); Clayton 1631 (8), rm (6); Clewell 3514 (2: 4: C onzali 4364 (24); Coradin 696 (6); Costa Sac- co 133 — 21469 (25); Cuatrecasas 3001 (13), 25702 (13). p D'Arey 141 (22), 2271 (22), 2659 (25), 4082 (6), 13527 (15); Daly 2171 (22); Daniel A-2 (19), 60 (6); Davidse 3065 (25), 3082 (12), : 4). . А 4514 (25), 4615a (6), 4708 (1), 5109 (19), 5242 (19), 5321 (19), 5597 (13), 10126 (24), 10655 (22), 10967 (22), 10985 (22), 11028 (22), 11123 (22), 11148 (22), (22). 11316 (22), 11380 (22), 11401 (22), 11407 (22), 12087 (6), 12124 (6), 12158 (25), 12159 (24), 12276 (24), . 15652 (19), 16117 (19), 16775 (6), 55 (6), 19314 (15). 19817 ө) 19487 (25), (6), 20544 (19), 21162 (25), 21312 (25), 21370 (13), 22047 (15), 22944 (1), 28159 TEN po 37 o. 35448 (13); Dawson 14142 (24), 14215 (24), 14863 (2: 192 (22); de Jesus 3 (10), 16 (24); de Lima. s.n. (22); de Paula 65 (24); Dedecca 302 (22); Dias 67 (10); Dinelli 12876 (7); Dombrowski 780 (19), 4570 (22), 5984 (22), 6488 (19), 6715 (19), 9037 (19), 9784 (19); Drew E-747 (24); Dusén 26a (22), 283 (22), 2770 (19), 2844 (19), 2844a (19), 3657 (22), 3659 (22), 4023 (24), 8011 (24), 9109 (8), 10412 (6), 10611 (19), 10613 (19), 13256 (19), 1: 3805a (22), 14489 (22), 15727 (19), 16057* (19), 16174 2), 17407 (19), 17934 (6), 18006 (24), (10), s.n. (22); Dutra 464 (22). Echeverry 1251 (13), 2312 (15); ene 639 (9); Egler 1262 (15): Eiten 3777 (7), 4417 (24 8 (7), 6746 (19), 6760 (19), 7626 (22), 9151 (19): а 594 (24), 596 2274 (12), 2302 (24), 2401 (12), 2484 (24), 13404 (12). 14018 (12); Ellenberger 1369 (22); Enamorado 23 (9); Escalona V274 (24); E (22). Fassett 25502 (13); Fe Тено 6462 (22); Fendler 1697 (25), 1698 (12), 2533 ( st Fernández 89 (24); Fernández Casas 6335 (24); Ferre 3 (13): Fiebrig 664 (24) 2061 (7), 3082 (13), 497 О (24), 4978 (24), 6222 (22); Filgueiras 2131 (14), 2184 (1: 4), 2280 (7), 2342 (16), 2408 (2), 2409 (2), 2455 (19), 2477 (4), 2780 (16), 3260 (22), 3392 (16), 3317 (14), 3567* (25); Filipovich 398 (13); Volume 89, Number 3 2002 Denham et al. 397 Paspalum subg. Ceresia Flores Flanco 3643 : 3); — 224 (2), p — 600 24 Ape (1); Freytag s.n. 6 (24 ); Fune ke 7 14 García Barriga 4993 (19), Dos п» ji ‚ 4390 (25); Gentry 15081 (19), 7 G-101 (15); Giraldo-Cañas 2560 (6), 2260 (22), 4309 (22), 6775 (22), 7980 (22), 13334 (24), 16548 (22). 17406 (10), 17410 (22). 17411 (22), 17414 (24). 17940 (22). 18607 (22), 18644 (22), 18674 (22), 20070 (21), 20076 e 20078 (21). 20079 (25), 20082 (22). 20083 (22), 2 4 (15), 22427 (19). 22437 (22 22441 (11). 22443 a n 22544 (15), 22544b (14), 22545 . 22547 (6), 22549 (24), 22551 (10), 22553 p.p. (24), 22555 (10). 22556 (6). 22551 (24). 22558 (22), 22559 (22), 0 — 22561 (22), 22562 (22), 22576 (12), 22577 (25), 22949 (24), s.n. (24); Goés s.n. (22); Goodland 923 (15): — rger 1043 (10); Gra- ham 520 (19); Grignon 84257 (13); Guala 1320 (24), 1450 (22); Guánc A > 3018 (1); Gutiérrez 655 (17). Haase ) (19); Hahn 5592b (15); Handro 97 (22): Harle "y T ^ (19). 10480 (19), SH209 (22); Hassler 1039 . 1058 (24), 9259 (24), 9261 (22), 926la (22). 9380 . 10209 (22). 11058 (24), 11501 (22), 11561 (22). 11916 (22); Hatschbach 5461 (22), 6864 (22). 7151 (22). 8965 (22). 11110 (22), 12097 x 2h 16134 (22). 25048 (19). 25264 (19), 32479 (19), 33152 (19), 44562 (22). 15097 (19), 47650 (24), 48811 (19), 64890 (22), 66690 (22): Haught 2709 (19); Hazlett 994 (24); Heithanus 163 (19) Heringer 4615 (15), 7249 (22), 7382 (14). 10432 (22): Hermann 10995 (19), 11089 ^а Hernández 2283 (19); He Du Xolocotzi X-623 (12); Herter 1755 (22): Heyde 6211 (1 Hicken s.n. (24); Hill 1548 (19); Hinton 1440 (13). mad ө, 11400 (13); Hitchcock 6355 (13). 0837 7297 (13). 8186 (13), 8297 (12). 8946 Gardner 4030 Gifford — (24); Holway 98 (22), Hostmann 1306 (6): dun 9643 (1 b (24). 6012 (22). 6246 (22); Hutchinson es (13). (7). 1); Hunt 5550 5194 (24); Idrobo 11636 (13); Hus 342 (13): inen 7 739 (6), 5163 (24). 5171 (22). , 6205 (22), 6579 (19), 6758 (19). 7219 (22), 8001 (24). 9495 (22). Ibarrola 2834 Imaguire 605 (22); 56 6800 (19). 7069 (22 2), 10737 (11). LUD. 13241 a D. р 15328 (10). 15347 ee 15928 (24), 15929 (24), 16232a (22), 16465 (24), 17256 (22), 17314 (24), 17617 (24), 17854 (22), 17967 (22), 19486 (22). 27667 (О), 28353 (22); Isabell s.n. (7). Jangoux 10153 (19); Joergensen 1765 (13). 2882 (24). 4568 is 1); Joly 1456 (24): Jorgensen 52 (24): Jürgens G- 299 : E 087 (24), 904 (24), 1795 (22), 1952 (24). 1996 (24). 201 1B (24), 2024* (17), 2076 (17). 2458 (24), — ). 2474* (24), 2477 (24), 247 (24). 278 (19), 2795 (22), 2823 (6), 4152 (7). - 6056 (2: 1). 6541 (23); Killip 1555 (12). ‚ 19770 (13), 34321 (19); Klein 3916 (19): Koninck 157 (25 ›); Kral 69245 (19), 72050 (1), 75939 (11); Kra- povickas 14972 (24), 25202 (24): Kuhlmann 1667 (17), 1668 (17). 1671 (23). 1674 (15). s.n. (22); Kummrow 2204 (22). Lazor 3350 (19); Leavenworth 1649 (1 2: : eehmann 979 19); Leonard 7526 (24). 7537a (24). 3 (12), 895: „ 8957 (12); Liebmann 225 (9); p 19350 (15): Ama 58-3056 (24); Linares 3663 (9), 3904 (9); Linden = = x ы 5); Llano 47 (6); Llatas Quiroz 2322 (13); Loefgren gos 351 Pg" 1121 (12), 1543 (8): hado de Campos 111 Machuca Nuñez 4040 (1: 101 (22), 908 d 19226 (24); Maguire — (15); Mal- me 965 (22). 1562b (12), 2226 (22). s.n. (22): 1254 (13); Marin 851 (13); Martius 267 (22): Martínez F- 1953B (13); MeVaugh 19098 (13); Mendoga 2293 (22): — 2943 (24); — = (24); Меуег 2090 (22), ; Miranda 3883 : Molina 8702 (25). 12041 “tz? (24): Montes — ). 1953 (24). 9) | 1920 ( 24), 15353 (24), 16250 (24), n (24). 27042 (24); Montoro s.n. (22); Moraes 1713 ; Moran 5505 : Mori 13303 (22), 17298 (6); ا‎ T (25); Mor- rone 205 (24); Mostacedo 314 1783 (23), 1862 (1): Müller 2036 (13); Mutis 5453 (19), 5537 (13); Myers 3338 E (19), 3404 (24). Nee 8166 (24), 26848 (3), 40372 (7), 46904 (19). 19412 (22); Nelson 3424 (24); Nicora 4700 (22), 4708 (24). 6352 (24); Niederlein 570 (22); Norrmann 159 (24): Nunez 7175 (13). Oldenburger 5 (15); Oliveira 305 (7). 598 (7). O15 (2). 623 (2). 630 (2), 638 (16), 700 (16). 719 (12), 758* (2). 763 (19): Ollgaard 90721 (7); Orozco 182 (19): Orth 1059 22). 4664 (24), 7070 (24): Partch 09- 91 3): Pedersen 1032 (22), 2611 (24). 3783 (24). (22), 13371 (22), 15713 (24); q nnell 8174 (19); net 496 (22); Peterson 9478 (7), 11887 (22). 1 1801 (2 t); Pinheiro 142 (7), 517 (7), € A (7); Pinto d (19), 894 (19). 1066 (19). 1141 (19), 1467 (19). 1550 (19); Pire 14649 (24); Pires 2290 (24). 16684 (24): — r 214 (25). 1788 (25) 4351 (19), 5020 (24), 5902 (25). 0033 (25), 7487 (24). 24 ‚ 11326 (24), 11998 (2 D 19132 (15); Pohl 714 (19), . 2833 (24). (24). 12274 (6), 12361 (19), 13531 (9). 13761 (25). 13850 (24). 13922 (19); Porter 4300 (19); Prance 13823 (1): Pur- pus 2002 (13); Purseglove 6514 (6); Pursell 8580 (25). Quarín 2235 (24), 3340 (24), 3716 (24). 3852 (24). 1043 (21); Quintana 165 (24); Quiroga 38 (24). Rambo 40511 (22), 40690 (22), 40804. (22), 53900 (22), 54017 (22), 54991 (22); Ramia 1009 (24); Ramirez 961 (15). 1789 (24): Ramos 182a (6); Ratter 520 (22). 820 (21). 1915 (24). 1969 (19), 2883 (24), 2884. (25). 3335 (24); Reeder 3313 (13); Reeves 194 (25); Regnell 111-1345 (25). HI-1344 (24), 111-1335 (19), 1-466 (22); Re- ineck 431 (22); Reitz 10652 (22), 10659 о 10081 (22), 10906 (22), 10933 (22), 12477 (22); 1038 (13). ibi — 5409 (19); Riedel 966 (12), (22), 1048 (1), 2143 (22), s.n. (1), s.n. (12), s.n. (24); — ra L-L41 (19), 3 (19); Rizzo 4285 (22); Robinson 3189 (22); Rodrigues 662 (6); Rodriguez 206 (24); Rojas 2757 Qu. 9569 (22), 13404 (22), 13933 (22): p Rombouts 308 (19); Rondon 2135 24); Rosengurtt B-3197 (22), B-5749 22), B-5883 (24), B-7397 (22); Busia 1200 (25): Rzedowski 17636 (13). 2332 (7), 2341 (13); — M (24): (22) 4215 (22), 5 (22); t (13); Schenck 3155 m M 1323 (24). 1403 (24). 1436 (24), 6207 (24); Schmidt 53 (22), 136 (13); Schreiter 9251 (13), 9256 WA ч 300 (24). 7220 (24). 17275 (22); Schwacke 8444 (2 2214 (22), 5539 (24). 5703 (24). 10003 (24), тл (24); Sellow 1232 (19), 1239 (19), 1240 (22), 4416 (22). 5686 (24). s.n. (19), s.n. (22). s.n. (24): Semir 380 (19); Semple Romanezuk ' 12 (9) Rusby 1014 Sánchez Vega Saravia 14006 (13), Schaffner 120 (13), 15 Annals of the Missouri Botanical Garden 150 (19); Sendulsky 1239 (10); Sermeño s.n. (13); Shaw s.n. (19), s.n. (6); Silva 114 (24); Smith 142 (24), 143 (12). 2170 (25), 5907 (22), 6316 (22), 6912 (24), 8445 (19). 11529 (24), 12013 (22), 12106 (22), 14462 (6). 14579 (8). 14858 (8), 15569 (19), 16123 (22); Soderstrom 1124 (6); Sohns 427 (13); Soria 4442 (24); Soriano 986 (13); — ~ ; Souza 5548 (7); Spegazzini s.n. (22); Standley 13185 (12), 56275 4); Steinbach 3015 (3), 3712 (22), 5162 (22), 5385 6618 (13), 6808 (22), 7103 (24); Stevens 24761 (13), 25701 (13); Steyermark 29550 (9), 32312 (24), 42586 (25), 55 (25), 59150 (19), 75259 (19), 121984 (25), 127710 (24). 127714 (25); Stockwell 893 (13); Stuckert 14159 (13), 18694. (22); Swalle n 3074. (19), 3643 (6), 3687 (6). 3772 4078 (6), 4762 (6), 8344 (19), 8361 (22), 3418 30 EA 8518 (22), 8584 (22), 8613 (22). 8639 (22), 8698 (22), 8772 (22), 8784 (22), 8975 1/2 (22), 8994 (22), 9033 (22), 9083 (22), 9087 — 9132 (22), 9150 (22), 9159 (22), 9181 (22), 9540 (2 9590 (24), 10747 (9), 10809 (9). 10810 (12), 10993 (25 R ‘ ). ; Tamayo 1084 (25), 1867 (22), 2046* (25), 2281 (13), 2881 (2 4), : 3198 (15), s.n. (15); Tate 157 * (12); Triana 761 (25); (24 Ule 7919 (24), 80 32 (24). Valeur 498 (24); от 8340 (6); van der Werff 7863 (22): Vargas 2459 (7), 19047 (7); Venturi 8320 (7), 9905 (7); Viei v 54 (22); ee Carenzo 1818 (13); Vincelli 1146 (19). Warming s.n. (25); Webster 29732 (19); Weddell 1699* (8), 2556 (10); Widgre n s.n. (22), s.n. (19); Williams 10533 (24), 13088 (15), 14704 (24), 17344 (24), 42957 (25); Wood 4273 (22), 4306 (25), 7852 (13); Wurdack 27 (15). Zardini 26848 (24), 51117 (19); Zanoni 31952 (24); Zu- loaga 888 (22), 910 (24), 2383 (22), 3835 (11), 3839 (19), 4379 (6), 4413 (15), 4587 (22), 4606 (22), 4654 (22), 4697 (22), 4874 (22), 4904 (22), 4930 (22), 5371 (24), 5671 (24). 5788 (24), 5791 (22), 5885 (13). 6585 (22), 6817 (24). 6944. (22) T APPENDIX 3 List of specimens for outgroup taxa in the cladistic anal- ysi є. Anthae "nantiopsis | fie brigii L. Parodi ARGENTINA. Jujuy: Dpto. Dr. M. Belgrano, Zapla, Mina 9 de Octubre, Cabrera et al. 32666 (SI). Panicum I — laxum Sw. ARGEN PIN A. Corrientes: Monte Caseros, monte del ar- royo El Seibo, Nicora 4516 (SI). Misiones: Dep. Iguazú. Cataratas tar Cabrera et al 28942 (SI). Santa Fe: Río Paraná, Canal Viejo, Job 710 (SI). — ит — Kunth J.S.A. Arizona: Box Canyon Road, Gould 2889 (SI). Paspalum € NS Dissecta Paspalum acuminatum Raddi ARGENTINA. Tucumán: Dpto. ida Loc. Monteros, Lillo 38377 (SI); Dpto. Leales, Loc. Leales, Venturi 699 (SI). Entre Ríos: Dpto. Feliciano, río Guayquiraró, Ruta 12, Burkart et al. 23250 (SI). Paspalum — tum (L.) 1 : N side a Gulf Shores, Kral 51314 (MO). Louisiana: "Cher eniere Lake, Burkunas 605 (MO). P uns Group Eriantha — erianthum Nees ex ‘Trin. LIVIA. Santa Cruz: Ñuflo de Chávez, Est. Las Madres. Killeen 1369 (SI). BRAZIL. Distrito Federal: Area da Proflora, Filgueiras & Zuloaga 2207 (SI); Brasília, jen Sul, Zuloaga 3836 (SI) Paspalum guttatum Trin. BRAZIL. Distrito Federal: — en Brasília & Sobradin- ho, Irwin et P 9163 (MO). Goiás: s.l., Glaziou 22442 (G). Paspalum — ntum Trin. BRAZIL. ‚ Go of Cristalina, /rwin et al. 9858 (MO). i Santuario, Zuloaga & = as Gerais: — da Piedade, Morrone 4676 (SI). Paspalum Group Notata : Tavarory, 3 km S from administra- E of río Paraguay, Zardini & Tillería 29180 (S1): lta — — 6277 (SI). Pte. Hayes: Estero Pa- tiño, ruta Trans Chaco, km 164, Quarín 4049 (SI). — нати Trin. ex Doll ARGENTINA. Corrientes: "el del Timbof, о de Monte Caseros, Nicora 4629 (SI). BRAZIL. Río Grande o Sul: Centro Agronómico de Guaiba, Burkart pes (Sh. URUGUAY. Durazno: La Paloma, Schroeder s.n. (SI). jos notatum Flüggé ARGENT pa Córdoba: Santa María, Alta Grac la, E zker 1178 orrientes: San — 19 km S de ( alí, Ruta 9 Maas 1837 (SI); Monte Caseros, ai el Urug , Nicora 4587 (SI). Misiones: Dpto. Apóstoles, aida Burkart — (SI). Paspalum Group Rac — — How. & Bonpl. ex Flüggé) Kunth NTINA. Jujuy: Dpto. Dr. Manuel Belgrano, faldeo del rh Se nde то pos de Lozano, Deginani & Cialdella 135 (SI); Dpto. Capital, Lagunas de Yala, Cabrera et al. 306706 (SI); Termas de Reyes, Burkart et al. 11218 (SI). Paspalum — Scribn. & Merr. MEXICO. Hidalgo: near — Station, Pringle 8891 (SI). VENEZU A. Lara: Distrito Jiménez, 4—5 km of Sanare, Davidse & González 21366 (SI). Paspalum racemosum Lam. PERU. Lima: Chancay, Lomas of — S of Chancay, Stork & Vargas 9344 (SI). Trujillo: La Libertad, Cerro Cab- ezón, Sagástegui & Mostacero ze (SI); La Libertad, Bar- raza, Mostacero & Ramírez 690 (S1). = APPENDIX 4 Index to scientific names. Synonyms appear in italics. Biaristata, sect. 338, 352 Jeresia 337, 352 Ceresia membranacea 363 Ceresia, sect. 338, 339, 352 337, 338, 339, 340, 344, 346, 348, 349, 350, Ceresia, subg. 351, 352 Ceresia, subsect. 352 Panicum ceresia z: P. obtectun Paspalum aspidiotes | у. 339, 340, 352, 354, 381 * 351, P. aureolatum P. biaristatum 377 338, 339, q 346, 348, 35 i 356, 357, 379 P. bicilium 386, 38€ P. blepharophorum 337, xs 373, 388 P. blepharophorum var. 386 tenue P. buchtienii 338, чү 340, 344, 347, 351, 352, 357, 375 Volume 89, Number 3 2002 Denham et al. Paspalum subg. Ceresia 399 P. burmanii P. cachimboense P. carinatum P. ceresia P. ceresioides P. ciliatum P. contractum P. cordatum P. cujabense P. cymbiforme P. distichophyllum P. echinotrichum P. elegans P. erectifolium P. eucomum P. eucomum var. pilosior P. govasense P. gracile P. heterotrichon P. heterotrichon var. pau- cispicatum P. humboldtianum P. humboldtianum var. ele- rantissima P. humboldtianum var. stuckertii P. imbricatum P. kappleri P. lanciflorum 339, 346, 348, 351. 352. 360, 380 359, 360, 38( 339, 340, 344, 346, 348 352, 3 » 3 L 37i 362, : 70 338, 340, че 348, 351, 352, 363. 391 7l НЕ 338, 351. 377 338, er 346, Mp m 354, 365. 378, : 389 338, 339, 1T 346, 347, 351. 352 373, 388 377 352. 363 337. 338, 346, мы чш 352. 307, 366 т 9] m. 2 * 351. 352. 3 K 338, 339, 346, 351, 352 371, 383. 392 311 338, aye eo 344, 340. 347, 360, ш = 375, 376, 380, 388 375 338, 375 339, е awe 340, 348, 351 361 338, 339, ЗЛО. 346. 352. 301. 377 P. longiaristatum & P. macroblepharum I P. malmeanum P. membranaceum P. membranaceum var. aequiglume P. membranaceum var. inaequiglume P. niquelandiae P. pectinatum P. petrense P. phyllorhachis P. piligerum P. polyphyllum P. reticulinerve P. setiglume P. soboliferum ns P. splendens var. sphacela- tum P. stellatum P. stellatum f. паша monostachyum P. stuckertii P. trachycoleon P. wagenerianum Pectinata, sect. Tricholaena obtecta 378, : 19 386, 384 3: " 3 7 351, 379, 391 352. 362. 90 389 389 338, 339, 340, ж 348, 351, 35: 303. 389, 391 ч 33. 339, 346, 352, — 386. : 301, 389 338, 339, 340, 354. 37 | 377, 340. 18. 351. 383, Xx 352. 369, 331. E. 363 363 363 339. 346, 348, 351. 352. 2 79. 380 337, 338, 340, 346. 351. 352. 365, 378, 380 339. 348, 351. 352. 371. 383, < 3: ES gt 346, 348. 351, 352, 371. 383. 386, 392 377 337. 338, 339, 346, 317, 4 Е = 367. 375 "5 gr ae 346, 351. 352, 354. 388. 389 338. 354 338, 375. 376 361. 362 338. 367. 370 389 337. 338. 340, 346, 351. 350. 352 380, 39 PHYLOGENY OF THE TRIBE Alan W. Meerow,? Charles L. Guy,’ HYMENOCALLIDEAE Qin-Bao Li,* and Jason R. Clayton’ (AMARYLLIDACEAE) BASED ON MORPHOLOGY AND MOLECULAR CHARACTERS! ABSTRACT The gene ric limits of Hymenocallis have been variously proposed by different taxonomic workers, often without discussion or data. The genera Leptochiton, Ismene, Elisena, and Pseudostenomesson have been included with Hymen- ocallis, lumped together as the genus /smene, or maintained as distinct genera. Recent cladistic analysis of plastid and nrDNA for Aar рари — a distinct tribe Hymenocallideae. Cladistic analyses of morphology, and plastid (trnL-F region) and nuclear sl DNA (ITS) are prese е alone and in combination for the tribe. Leptochiton is sister to the rest of the genera in the tribe in all analyses. While Hymenocallis is always resolved as monophyletic, Ismene is variably paraphyle lic or monophyletic. The enda sequence data produce the most resolved and best- — jd phyloge ny, wherein Hymenocallis and Ismene are monophyletic sister genera. These data support an origin for the tribe in the Andes, with vicariant distribution of the largely Mesoamerican Hyme — Formal recognition of Ismene subg. Elisena and Pseudostenomesson is established. Key words: Amaryllidaceae, cladistics, molecular systematics, phylogeny. Systematics of the genus Hymenocallis Salisb. — menocallis, subg. Ismene (Salisb.) Baker ex Traub (Amaryllidaceae) and its allies have defied precise (including Leptochiton), subg. Elisena (Herb.) systematic understanding at both the specific and Traub, and subg. Pseudostenomesson (Velarde) generic levels (Flory, 1976: Meerow € Dehgan, Traub. Traub (1980) later reduced these subgenera 1985). The genera Hymenocallis and Ismene Salisb. to the rank of section without explanation. Ravenna were established by Salisbury (1812) for the Neo- — (1980) in his description of H. heliantha (= Lep- tropical species with fleshy seeds originally as- tochiton heliantha (Ravenna) Gereau € Meerow) signed to the Old World genus Pancratium L. The suggested that subgenera /smene (including Lepi- zygomorphic-flowered Elisena was described by dochiton). Elisena, and Pseudostenomesson should Herbert (1837), who recognized Hymenocallis and probably be all recognized as the genus /smene, Ismene as distinct genera. Baker (1888) subsumed distinct from Hymenocallis. Meerow and Dehgan Ismene within Hymenocallis but retained Elisena as (1985) suggested that Pseudostenomesson might distinct, as did Pax (1890). While Stapf (1933) warrant recognition at the rank of genus due to its treated A. quitoensis Herb. as a species of Pamian- extreme phenetic divergence. (funnelform-tubular the Stapf, Sealy (1937) considered the species to perianth) versus the “pancratioid” flower of Lepto- exhibit sufficient’ morphological divergence to be — chiton, Ismene subg. Ismene, and Hymenocallis. recognized as a monotypic genus, Leptochiton Sealy. € © — ’ancratoid” floral morphology refers to a large. Hutchinson (1934, 1959) retained both Elisena and white, fragrant, crateriform flower with a conspic- Ismene (presumably including Leptochiton) as dis- uous staminal cup (ef. Pancratium L.). This type of tinct. Velarde (1949) established the Peruvian ge- flower appears to be adapted for sphingid moth pol- nus Pseudostenomesson for a fleshy-seeded species — lination (Bauml, 1979; Grant, 1983; Morton, 1965). originally described as Stenomesson morissonii Var- Meerow (1990) treated Leptochiton as a distinct ge- gas as well as one new species. Traub (1962) rec- nus and recognized Hymenocallis and Ismene (in- ognized all four erstwhile genera as subgenera of cluding Elisena and Pseudostenomesson) as distinct, Hymenocallis in his synoptic treatment: subg. Hy- a treatment followed by Gereau et al. (1993) and е work was supported in part by NSF Grant DEB 9628787 to gi M and ( SDA-ARS-SHRS, 13601 Old C de < Miami, Florida 33158, U.S.A., and — hild ‘Tropical Garden, 11931 ом г tle — Miami, Florida 3315 Iniversity of Florida-IFAS, De iube nt " Environmental Horticulture, 1545 Fifield Hall. Gainesville, Florida — USDA. ARS- SHRS, 13001 Old Cutler Rd., Miami, Florida 33158. U.S.A. ANN. Missouri Bor. GARD. 89: 400-413. 2002. Volume 89, Number 3 2002 Meerow et al. Phylogeny of Tribe Hymenocallideae Meerow and Snijman (1998). No cladistic analysis has focused exclusively on testing the validity of this treatment, although at least one representative of each subgenus was included in overall molecular studies of Amaryllidaceae (Meerow et al., 1999, 2000a). Hymenocallis and its allied segregate genera are entirely Neotropical in distribution [a single West African taxon, Н. senegambica, was treated by Sea- ly (1954) as an early adventive introduction of H. caribaea]. Hymenocallis sensu stricto, with 50 to 60 species, is chiefly Mesoamerican and extends into the West Indies and the southeastern United States. It is sparingly represented in northern South Amer- ica. Leptochiton Sealy (2 spp.). Ismene (ca. 7 spp.). Elisena Herb. Velarde (2 spp.) are all endemic to the Central An- — 2 to 4 spp.), and Pseudostenomesson dean region of South America. Hymenocallis, Is- mene, and Leptochiton are contrasted in Table 1 Hymenocallis and allies have usually been allied with Eucharis Planch. in the tribe Eucharideae (Hutchinson, 1934, 1959; Traub, 1963; Dahlgren et al. 1985; Miiller-Doblies € Miiller-Doblies. 1996). Meerow (1989, 1995) argued that the link- age of these genera, largely through the perception that both lineages shared a fleshy seed, was mis- construed, and proposed that either subtribal or tribal recognition of Hymenocallis and allies was Miiller-Doblies Miiller-Doblies (1996) placed them in Eucharideae subtribe Hy- warranted. and menocallidinae, while Meerow and Snijman (1998) recognized a distinct tribe, Hymenocallideae. Fam- ily-wide analysis of plastid sequences (Meerow et al., 1999) and nrDNA analyses of the monophyletic 2000a) support a distinct Hymenocallideae as sister to the American clade of the family (Meerow et al.. newly recognized tribe Clinantheae (a segregate of the former Stenomesseae), but complete resolution of the intratribal relationships is not apparent in these large analyses. Both tribes are subclades of a well-supported, Andean, tetraploid clade of gen- era. In this paper, we present phylogenetic analyses of morphological and molecular data for the tribe Hymenocallideae, and seek to clarify the relation- ships within the tribe. MATERIALS AND METHODS SAMPLING Sequences for the plastid trnL-F region were newly obtained for H. eucharidifolia, which, along with H. latifolia, was used as an exemplar taxon of Hymenocallis (Table 2). Previously cited sequences were used for one species each of the three sub- genera of Ismene, one species of Leptochiton, and the outgroup Pamianthe peruviana (Table 2, Mee- 999). For ITS and the morphological data matrix, we increased our sampling with an ad- row et al., ditional four species of Hymenocallis and two ad- ditional species of Ismene subg. Ismene (Table 2). The aligned sequence matrices are available from the first author (miaamOars-grin.gov). MORPHOLOGICAL DATA Morphological and cytological character. state data were derived from the following sources: Traub (1962, 1980), Sealy (1954), Flory (1976), Velarde (1949), Bauml (1979), Meerow and Dehgan (1985); from examination of living material in research col- lections at the USDA, Miami, vations of Hymenocallis, Ismene, and Leptochiton Florida; field obser- species; and examination of herbarium material. The morphological matrix consists of 12 species representing 4 genera and 23 characters (Tables 3, t) SEQUENCE DATA The trnL-F (trnL intron and spacer region be- tween trn and trnF) matrix consisted of 6 taxa and 906 base positions. The nrDNA ITS sequence ma- trix (IISI. 5.85 intron, ITS2) consisted of 12 taxa and 636 bp. DNA EXTRACTION, AMPLIFICATION, PROTOCOLS AND SEQUENCING silica gel Genomic DNA was extracted. from dried leaf tissue as described by Meerow et al. (2000a). The trnL-trnF region was amplified using the primers of Taberlet et al. (1991) as described by Meerow et al. (1999). Amplification of the ri- bosomal DNA ITSI/5.8S/ITS2 region was accom- plished using flanking primers (185, 265) ABIOI and AB102 (Douzery et al., 1999), and the original White et al. (1990) internal primers ITS2 and 3 to amplify the spacers along with the intervening 5.95 sequence, as described by Meerow et al. (2000a). Amplified products were purified using QIAquick (Qiagen, Valencia, California) columns, following manufacturer's protocols. All polymerase chain re- actions (PCR) were performed on an ABI 9700 (Ap- plied Biosystems, Foster City, California). Cycle sequencing reactions were performed di- rectly on purified PCR products on the ABI 9700, using standard dideoxy cycle protocols for sequenc- ing with dye terminators on either an ABI 377 or ABI 310 automated. sequencer (according to the manufacturer's protocols; Applied Biosystems, Fos- ter City, California). 402 Annals of the Missouri Botanical Garden ‘Ob ‘OB-€EZ = uc 3[qegrre^ mq пошшо) jsou Ob ‘OF = UZ juasqy 9p — uc juasqy Ju9sq y OLIPPOL Ju9sq y juasa1g 01-с тс 0791 1y3regs pue 3u0| juour -e[y әәлу шим pedis jou *odeus ш әүдеивл pue jews 10 93181 dno [PULLEYS ‘Suoj oqn] 12212 “9|ISSIS Ápsow ‘ueu 0] [ “JURIBRLY ‘aM ‘эле ‘orydsowounoe *prongioueg "ugreqjs juaue -[ Әәл) *feouipur&oqns dno [eururejs ‘Buoy eqn) 'sno[npuod *oje[[oorpod *snodouinu Зиви8елу Jou “иәэ18 *[[euis + corydsowounoe чеүпчпу-шлодәпип y ‘areur[92p pue 2по| sjuaure[y әәлу *aqni Əy} шолу paxop -әр ‘adie, = dno jeurweys ioys eqn) :әјешүәәр “3|ISSISGNS “IZ “JURIBRIY JOU ‘1M ‘a31| *ongdaoui02 47 ‘PaAINOUL pue uoys juour -P[tj әәлр шим Uso peduijs *93.P] dno ¡eurure]s "Juo| x eqn) ‘spu -I[99P 10 [PJUOZLIOY *ajej[22rpad ој 3[rssosqns *([-z “JURIBRIJ *wo[[a4 10 3]« “эвлер omdaourourjoe "prongaoueq 'pexinour pue poys јиәше[у 331] TUTYIIM пәәл# рэа pue 93.18] dno [eururejs ‘Suoj əqn} дәәлә(дп) “Q[ISSOS “JUPIBRIY “MOT[OA 10 3]I[« 9318] *orgdaourourjoe *prorgaoueq juasqy juasa4q juaso4q ]u3s214 juasqy EOLI3UIPOS AJA `вәгрир ISAM “SA AS ш 000€ әлоде ‘nəd ѕиоцеләјә цац о] pru je Jopenoy pue плә,] ѕиоцелә[ә ущ 0] MOJ I? səpuy penu suone4o[9 MOT те ‘nd AN pu? Iopenoq MS 0с “Bo c S1]]DIOUGULA H uossoul -0uajsopnosq “BQNS auauls d `q I Dua] "oqns suaus] əuəus] "qns ousus] u0114201d7] Jequinu эшоѕошоц?у роо paas uo uP[auio]4uqd әүпәо] iad səma (Zojoydiow [P10] 4 wa]sopnasd 3p]e3uo[ uornqugst(] satdads Јо 19QUINN] snuazqns 10 snug ‘areapı[[P2ouauAH aqua 3eaoPpr[[Areury јо eıauadqns pue рләпәЗ ay) jo uosureduio") E "MEL 403 — — Phylogeny of Tribe Hymenocallideae Meerow et al. "|Р 19 моләәр ‘|B 19 MOLI в ә моләәрү "JE 19 MOII9[A үе 19 MOLI] qe Ja MOLI], 0801 1PAV ye 19 MOLI9 N Je 19 моләә[\| ye 19 MOLI ÍA “pe 19 MO.139 [A JE jo мого (6661) [e 19 моләәрү (6661) [e 19 моләәң (6661) e 19 моләәу (6661) [E 19 *o129[N — (6661) "T€ 19 MOI99]A OLOLITAY (6661) 18 19 Mo132]|N (6661) TF 19 моләәрү (6661) 1° 19 моләәр (6661) `I? 19 моләәрұ — (6661) 18 19 Mo12929[N (6661) ЧЕ 19 Mo1229[N S20LLEAV (6661) 18 19 M0129]|A 7. jdeyg pumantad ouquniumg ATRAS (qlo H) saisusopmb uo poda] 1 MOIDI Y NLA (әрїеүәд) 118D24Da M0122 N (әрлејәд) иохғәшоиәјѕорпәѕ "апе ouauis] моләәү C[pur]) руруәйї8ио] `] MOLVA СӘҢ) очокту 7845 ouaus] ‘boe f vaopfissioanu `] моләәр Y пвәлә=) (SPBIRA) NSƏYNDY *] «Моң (ARG y ZINY) sonoupum suaus] “SUBS эмо ^H "Neg, vyofipuvyonə ^H ‘moo (TTA) 2170/70 y "WOY “JA Pond)? Y 199MS (Чөң) pipofimop si]]roouaui&g Joords 4-qu4] 3u93 “ju ПОР ѕпотләла J0 ‘OU UOTSSIIIR que gus) похве, Volume 89, Number 3 2002 `рәїкәїриї ISIMISUYJO ss9[un HLA ye pausodop 318 SIIYONOA ПУ тгәвәрццүвәоцәшАн jo savduanbas VNC 19] sI9qunu UO0ISSIIIB que gus) Mou pue SIIYINOA y 9149], 404 Annals of the Missouri Botanical Garden Table 3. Characters and character states used in the cladistic analyses of Hymenocallideae based on morphology. Character States and coding 1. Elongate pseudostem . Flower number 3. Flowers sessile/pe dicellate 7. Ponant — pancra erianth symme 10. 11. Staminal cup shape 12. Staminal с es striping 13. Free filame 14. Free — 15. Pollen grain size 16. Pollen grain 17. Exine reticulum 18. Ovules per locule 19. Seed per locule 20. Phytomelan on testa n Ф 21. Seed cc 22. Saad shape 23. Most common diploid chromosome number erect absent 2-10 (0); solitary (1); >10 (2) sessile (0); pedicellate (1) 0); declinate/horizontal (1); pendent (2) shorter than tepals (0); longer than or equal to tepals (1) straight (0); curved (1) — 20 ( )); 1 numerous (0); 2-5 (1); 1 (2) esent (0); absent (1) oat not fleshy (0); fleshy (1) P € (0); globose (1) 46 (0); : p 0); present (1) atioid (0); funnelform-tubular (1); + funnelform (2 — actinomorphic (0); zygomorphic (1) white (0); vellow (1); green (2) present (0); absent (1) rotate or funnelform (0); cylindrical (1) present (0); absent (1) incurved (0); straight (1); declinate (2) longer than cup (0); shorter than cup (1) very large (0); large (1); auriculate 9 not т medium (2) е (0); n — edium (1 6:00. " 2-10 (2); 2—4 (3) 34 (1); 46, 40 (2) SEQUENCE ALIGNMENTS Both sequence matrices were readily aligned manually using the program Sequencher (Gene- Codes, Inc., Ann Arbor, needed to be inserted. Michigan) as few gaps CLADISTIC ANALYSES Pamianthe (tribe Clinantheae) was used as out- group for all analyses. In larger sequence analyses Table 4. morphisms: + = (0,1); * = (0,1,2). (Meerow et al., 1999, 2000a), this genus resolves as most closely related to the Hymenocallideae. Pa- mianthe and Leptochiton (the latter putatively the least derived genus in the Hymenocallideae; see discussion below) share two four-base sequence el- ements in the trnL-F region (bp325-328, 821-824) that are absent from the rest of the Hymenocalli- deae. Phylogenetic analyses were run using PAUP* version 4.0b8 beta (Swofford, 1998). An exhaustive search of all possible tree topologies was conducted Character state matrix for cladistic analysis of 23 morphological characters in Hymenocallideae. Poly- Taxon Matrix Character Hymenocallis acutifolia Hymenoc allis eucharidifolia Hymenoc allis tubiflora smene amancaes Ismene hawkesii Ismene longipetala Ismene narc — Ismene vargasii Leptochiton — Pamianthe peruviana 1 2 12345678901234567890123 00001000000110000221112 0*001000000130000221112 0*001000000130000221112 0*001000000130000221112 0*001000000130000221112 10111+00100001000321110 10111+00000001000321110 10110021011120211321110 10111+00000001000321110 10120110211110211321110 01001100+00001000110111 10111000000001110000000 Volume 89, Number 3 2002 Meerow et al. Phylogeny of Tribe Hymenocallideae for trnL-F. For ITS, the morphological, and all com- bined analyses, branch and bound searches were conducted. Support for internal nodes of the trees was determined with 5000 replicates of branch and 1985) and by calculation of Bremer (1988) decay indices (DI) us- 1999). A branch and bound search was implemented for bound bootstrapping (Felsenstein, ing the program TreeRot (Sorenson, each constraint statement postulated by TreeRot. A bootstrap value of 50-64% was considered weak, 65-74% moderate, and 75-100% strong support. Combining independent character matrices, whether both molecular or molecular and morpho- logical, very often increases the resolution of the ingroup and the bootstrap support of the internal nodes of the phylogenetic trees (Olmstead & Sweere, 1994; Chase et al., 1995; Yukawa et al.. 1996; Rudall et al.. 1998; Soltis et al.. 1998: Mee- row et al., 1999), Nonetheless, there is controversy about whether different data sets should be ana- lyzed separately or together (De Queiroz et al., 1995 1996). Congruence of the independent matrices has generally been demon- ; Huelsenbeck et al., strated before they are combined, but it has also been argued that incongruence should not be a pre- determined factor against doing so (Dubuisson et al. 1998; Seelanan et al.. 1997). Miyamoto and Fitch (1995) argued that data sets should always be analyzed independently, as underlying assump- tions, constraints, or weighting strategies will vary from data set to data set. Kluge (1989) and Nixon and Carpenter (1996) argued that simultaneous analysis of multiple data sets better maximizes par- simony and allows secondary signals to appear from the combined data. Bull et al. (1993), Rodrigo et al. (1993), and De Queiroz (1993) advocated com- bining data only after a statistical test of congru- ence, what Huelsenbeck et al. (1996) called “ ditional combination." con- Before combining the data sels, we performed a partition homogeneity test (Farris et al.. 1994, 1995) on the variously com- bined matrices, using a branch and bound search. RESULTS MORPHOLOGICAL MATRIX With all characters unordered, two most parsi- monious trees (Fig. LA, one shown) were found of length = 37, consistency index (CI) = 0.86, and retention index (RI) = 0.88. Sixteen of the 23 char- acters used were parsimony informative. In both ees, Hymenocallis is monophyletic (bootstrap = ш. DI = 1). while /smene is paraphyletic. /smene longipetala (subg. Elisena) and I vargast (subg. Pseudostenomesson) are sisters in both trees. Lep- tochiton is sister to both Hymenocallis and Ismene in one tree (Fig. ТА). The 6 apomorphies at the ancestral node are an increase in pollen grain size, auriculate pollen grains, reduction in ovule number from more than 20 to 16 to 20; reduction in number of seeds per locule; and evolution of globose, fleshy seeds. Apomorphies for Hymenocallis (Vig. ТА) are the absence of an elongate pseudostem, predomi- i erect flowers, and 2n = 46, 40 Other than Hymenocallis, the only clade with strong bootstrap support is the sister re- nantly sessile an chromosomes. lationship of /smene subg. Elisena and subgenus Pseudostenomesson (100%, DI = 6). apomorphies: perigone tube length reduction, non- based on 7 fra- grance, cylindrical staminal cup, and smaller non- pancratioid floral morphology. loss of floral auriculate pollen grains with less coarse exine reticulum. If all of the characters are ordered as irreversible, a single tree is found of length = 48, CI = 0.67 and RI = 0.88 (Fig. 1B). There is moderate bootstrap support for a monophyletic Zs- mene (65%: DI = 2; apomorphies: elongate pseu- dostem, pedicellate and declinate/horizontal flow- ers, and 2—4 ovules per with C locule). There is weak support for the sister relationship of Hymenocallis and /smene (56%. DI in ovule and seed number, respectively: and the 1; apomorphies: reduction loss of phytomelan from the testa). Leptochiton is moderately supported as sister to both (65%, DI = 1: apomorphies: reduction in ovule and seed num- ber and the evolution of a fleshy seed). /smene subg. Ismene has a 91% bootstrap and DI = 2. /smene subg. Elisena (1. longipetala) and subgenus Pseu- aa (I. vargasit) are again sister groups 1 100% bootstrap and a DI = 9. A monophyletic ola receives 87% bootstrap support with DI = 4. Hymenocallis latifolia, H. glauca, and Н. eucharidifolia form a monophyletic group with ^ > bootstrap support of 60 and DI = 1. This same tree topology (Fig. 1B) is 40 steps long with CI = 0.80 and RI with the topology as a constraint with all characters = 0.80 if a branch and bound search is run unordered. PLASTID trnL-F SEQUENCES Using trnL-F sequences, which provide 7 parsi- mony-informative base substitutions, three equally most parsimonious trees are found of length = 82, CI = 0.99, and RI = 0.88 (Fig. 2. one tree shown). All three trees resolve a monophyletic /smene with 81% bootstrap support (DI = 2), and Leptochiton as sister to the rest of the tribe but without support. A monophyletic Hymenocallis is resolved as sister to /smene in one tree (Fig. 2). but Hymenocallis and Annals of the 406 Missouri Botanical Garden (€ 314BL 398) qougaq ey} BUO[E sarqdıouode әле SIYIURIG MO| 94 s1oquainu pue soul] [E21LI9A `ѕәѕәцјиәлва ut JIP (ор) зәори! Арәәр pue sa3ejuaniad de.nsi00g "sy13ua]| Youelg 318 SIYIUPIG gaoge SI9QUINN "9[QUS19A9.LIL se p2312p4o 31E SLIJIBIBYO Пе J! punoj 331] snoruouirsaed вош a[3uIŞ “> `рәләрдоип ѕләЈәвЈецо ПЕ Чим punoj S331} snoruoursıed 150Шш OM] jo 910 'V— “SIIPOBIBYO ¡eor3o]oyd.1ow uo pəseq әвәрү[еәоиәшАң 10] ѕшеі8ореүэ eueiAnjed ayjueweg sisueojinb uojiu2ojde1 nseBsen eueuis| еүезәа!биој eueuis| EJO|JISSIDJ8U Guauis| пѕәумец eueuis| S9PIUBUJE Guauis| EIJOJI]8| SIJJEÉ2OuOUIÁAH eone/6 sijesouawAyY eyopjipueyona ѕ=цүвэоиәшАн EJ10/J1GN] sijje20uauiAH еојцпәе sije2oueuiAH L1 ‘91 'SL'EL sdas ё Sppe £z JapeJeyo 19 ebueyo lr '/8)$ 02 ‘61 ‘81 U `95)Е eueiAniad ayjuejwed sisusoynb uojiu2ojde BIOYISSIIIEU ӘШӘШ$®] пѕәумец әиәшѕ] ѕәвэиеше әиәшѕ] ııseBıen әиәш8] ејејәаібиој eueuis| еуојцеј ѕүеэоиәшАн ganejD =вэоиәшАн EJJOJIPUEYINA sijjesouawAY &JojJIqn) sij[e2ouauiAH Boyne SIJ|B2O UOUIÁH L1 '9V'SIE ‘LL 'OL'Z'S d əmi Meerow 407 Volume 89, Number 3 al. Nu of Tribe Hymenocallideae Ismene narcissiflora Ismene longipetala Ismene vargasii Hymenocallis eucharidifolia 2 20 " og. ge — — Hymenocallis latifolia 7 А „ А Leptochiton quitoensis Pamianthe peruviana Figure 2. One of the three most parsimonious trees found by cladistic analysis of plastid trnL-F DNA sequences for the тЫ ‘allideae. Numbers above branches are branch lengths; numbers below branches are bootstrap per- centages, follow all three trees. Ismene form a clade in all three (73% bootstrap, DI = 1). Ismene subg. Ismene (1. narcissiflora) and Eli- sena (1. longipetala) are resolved as sister groups in all three trees with a bootstrap of 70% (DI = 1). ITS SEQUENCES ITS provides 50 parsimony. informative charac- ters, and 9 trees of length = 209, C 0.73. and RI = 0.77 were found (Fig. : 3). In all o of the trees, Leptochiton is resolved as sister to both Hymeno- callis and Ismene (Fig. 3A), but without significant support. jede allis is monophyletic (bootstrap — 97%. DI = 5). but /smene is monophyletic in only 2 of the 9 trees (Fig. 3B, one shown). However, Ismene subg. Ismene (1. amancaes, 1. hawkesit, 1. narcissiflora) is monophyletic with weak bootstrap support (59%) and DI = 1 (Fig. 3B). ed by decay indices (italic). The large arrow indicates a node that collapses in the strict consensus of COMBINED £rnrl.-F AND ITS SEQUENCES The P value from the partition homogeneity test = 0.93. indicating that the trnL-F and ITS se- quence matrices were highly congruent. Six most р arsimonious trees were found of length = 292, СІ = 0.92, and RI = 0.77 (Fig. 4). In all trees, Hy- menocallis and Ismene are monophyletic sister gen- era with bootstrap support of 94% and a DI = 3. Leptochiton is sister to both, but without significant support. Bootstrap support for a monophyletic Hy- menocallis is 98% (DI = 5), but only 68% (DI = 1) for a monophyletic /smene. The only other inter- nal resolution within /smene that receives bootstrap support is a sister relationship between /. narcis- siflora and I. hawkesii (both within subg. Ismene) at 84% with DI = Annals of the 408 Missouri Botanical Garden "(9i[e11) вәә1їрш 4вәәр Aq рәмо[үо} sedejuao1ad 21151004 318 ѕәцәиві ^o[oq ѕләдшпи :sy¡3ua] Y9UBIG әле SIYIUP.IG әлофе sroquinv 'sioouawukg 0] dno13 191818 one[4udouour e SI 2uauis| YOTYM UL S221 OM] JO эц) “g— “SIAM snoruourrsed. A([enbo әши jo snsuasuoo LYS “y — ‘seouenbes GJ] YN Qu uo paseq arapı[[e2ouauAH 10} sweisopey) сє 310314 euegejaniad әціиешед sisueojinb uojiu2ojde] eyojipueyonea $=јеэоиәш/Ан eanejD si[[R2ouauiÁH &JOJJIqn] $1јеэоиәш/Ан EIJOJI1NIE sij[RB20uGUIÁH EI/OJ1J8] ѕ=1івэоиәш/Ан пѕебіел ououus / ejejedibuo| eueuis| saeouewe ououus / lisaymey eueuis| BlOJJISsI2.eu eueuis| V BuelAniad aujuejuegd sisuaojinb uojiu2oojde] IISEDJ8A ououus / eٳfoJpıeyone‎ si[j[eR20uauJAH ganejD si[jR2ouauiAH H EIOMOMD si[RB2ououlAH 8IJ0JI]n2P $1јеэоиәш/н вое Sij|EB20uGulÁH ejejadiBuo] eueuis| seeouewe әиәш] пѕәҳмец әиәшѕ] EJO|JISSIDIBU ououus / Volume 89, Number 3 2002 Meerow et al. Phylogeny of Tribe Hymenocallideae 409 Figure 4. ITS sequences. Numbers above branches are branch lengths: numbers below branches are bootstrap percentages fol- lowed by decay indices (italic). T COMBINED SEQUENCE AND MORPHOLOGICAL MATRICES The P value of the partition homogeneity test was 0.0003, tween the morphological and DNA sequence data indicating significant incongruence be- matrices. Much of the apparent incongruence can be attributed to the weak resolution of the morpho- logically based topologies, and we felt that it would still be informative to combine the two matrices in a single analysis. Of the 1565 characters included, 76 were parsimony informative. A single tree was found of length = 332, СІ = 0.92, and RI = 0.79 (Fig. 5A). Hymenocallis is monophyletic with 100% bootstrap support (DI = 8), but /smene is paraphy- Ismene narcissiflora Ismene hawkesii Ismene amancaes Ismene vargasii Hymenocallis glauca Hymenocallis eucharidifolia Pamianthe peruviana PERU PERU PERU Ismene longipetala ecuanor, PERU PERU Hymenocallis latifolia FLORIDA. UPPER ANTILLES Hymenocallis acutifolia mexico Hymenocallis tubiflora NORTHERN souTH AMERICA MEXICO MEXICO Leptochiton quitoensis SW ECUADOR, NW PERU BOLIVIA, PERU One of six most parsimonious trees found by cladistic analysis of combined plastid trnL-F and nrDNA ). The larger arrow indicates a node that collapses in the strict consensus of all six trees. letic. Bootstrap support for the monophyly of Ismene (DI = 2), subg. Elisena (1. longipetala) and Pseudostenomes- subg. /smene rises to 81% but /smene son (d. vargastt) are sister groups — = 97%, DI = = 1) as sister to Hymenocallis. Leptoc — Is again I) weakly supported (bootstrap = 57%, DI sister to the other members of Hymenocallideae but without support. If trees one step longer were also retained in the search, in addition to the single shortest tree (Fig. 5A). a single, fully resolved tree 333. CI = 0.90, and RI = 0.77 was found (Fig. 5B). In this tree (Fig. 5B), both Hymen- ocallis and Ismene are monophyletic sister genera, of length = as are Ismene subg. Elisena and Pseudostenomes- son. Annals of the 410 Missouri Botanical Garden "suua| qoueaq 318 SIYIURIG JAOQB SIIQUIN y `рэч!в191 318 Vc ut painjord 331] uey} 19200] days 3UO Saa] jl punoj 3341] [tuo ippe әш E ° (оен) 839)IpUI \вдәр íq рэмо[[0ј sogequa21ad 421181004 ӘЛР SIYIURIG MO[3q s19qunu "sy 1Suo] youe1q ӘЛЕ SIYIUPIG 9AOqE SI9QUNN *331] snoruouits ed jsou әш "W— "SJOLI]EUI әәиәпһәѕ YNA pue [eotsojoydiouw p2urquioo uo peseq әрәреәоиәшАң 10] ѕшрізореүэ 6 3131 g V euelAniad ayjueiwed euejansad ayjuejwed sisueogjinb uodu20jde | sisuaoynb uo}y901de7 еџоџриецэпә sjjje2ouaw AH EIJOJIPIIEYINO sijjesouawAyY гопејб sijje2ouaw AY eonejb sjjje2ouaw AY = 6S елордап1 sijje20ouauiAH ds BIOJJIqn] SJjje2oua ui AH 95 EIJOJI jn 2B siJj[B20ua ui AH eojnnoe Się20uə WÁH Boye =еэоиәш/Ан BIJOJI18] SiJ&e2ouaurAH ¡sebiea auaws] ysebien әйәш$| ejejedibuo| eueuis| ejejadibuo] euauis| ѕәвэиеше әиәш] saesguewe auauis[ пѕәҳмец әиәшѕ] пѕәҳмец әиәш] 2 210115812Јви BUGIS] EJO|JISSIDJBU ӘШӘШ$| zz LL 6 oL ez $ 8 y zu lz ц 6z LL L 6 t zu Volume 89, Number 3 Meerow et al. 411 2002 Phylogeny of Tribe Hymenocallideae DISCUSSION the non-fleshy seeded Andean endemic Clinan- Both plastid (Meerow et al., 1999) and ITS (Mee- row et al., 2000a) sequences strongly support the position of the tribe Hymenocallideae as a mono- phyletic group within the Andean tetraploid clade of the endemic American Amaryllidaceae that i sister to the newly recognized tribe Clinantheae Meerow (Meerow et al., 2000a). Clinantheae are. uniformly The seeds of the flat, winged, and with phytomelanous testas. There are links between Leptochiton and Pamianthe that Stapf implicitly recognized, most notably the plesiomorphic pres- ence of phytomelan in the testa of Leptochiton’s seed [of which Meerow & Dehgan (1985) were un- aware], but also the numerous ovules of this genus (plesiomorphic as well). In the ITS phylogeny pre- sented by Meerow et al. (2000a), support for Pa- mianthe as sister to the rest of Clinantheae (vs. a dry, sister group relationship to Hymenocallideae or an unresolved position) was considerably weaker when the aligned matrix was not successively weighted. This is not surprising given that both genera occupy a basal phylogenetic position in their respective clades herein. The difficulty of relying on morphological char- acters alone to generate phylogenies in Amarylli- daceae has been discussed (Meerow, 1995; Meerow 20005). for many morphological characters in the family. et al., given a high degree of homoplasy Our analysis (Fig. 1) generates trees with relatively high CI and RI, but parsimony is still not able to resolve /smene nor consistently place Leptochiton in the basal position within the tribe with unordered morphological characters alone, in contrast to se- quence data (Figs. 2-4), which also provide (in the combined trnL-F and ITS matrix). over three times the number of phylogenetically informative char- acters of morphology alone. The combined plastid and nuclear sequence matrix produces the most ful- ly resolved shortest trees. To “force” this topology upon any of the other conflicting data matrices re- quires either ordering characters or accepting lon- ger trees (albeit only one step longer in the com- bined sequence and morphological analysis). When biogeographic information is optimized upon the combined plastid and nrDNA tree (Fig. 4), the gene phylogeny supports an origin for the tribe in the central Andes, inarguably a locus of diversity for the Andean tetraploid clade of the 2000a). with a vi- cariance event that gave rise to the largely North American Hymenocallis. Leptochiton, with 16 to 20 Amaryllidaceae (Meerow et al., ovules per locule and a phytomelanous testa, oc- cupies a relict position in the tribe with links to theae. However, it is the genus /smene that reflects the patterns of floral morphological diversity that occur in the Eustephieae, Clinantheae, and Steno- messeae (sensu Meerow et al., 2000a). /smene subg. Ismene retains the plesiomorphic pancratioid floral morphology of Leptochiton, Pamianthe, and Hymen- ocallis, while the smaller /smene subg. Elisena and subg. Pseudostenomesson express floral novelties. Ismene subg. Pseudostenomesson, occurring at the highest elevations of any member of the tribe, might be the youngest element of the polymorphic /smene. since the Andes likely did not extend above 1000 m elevation before the Pliocene (10 MY BP: Van der Hammen, 1974, 1979). Analogous patterns of floral diversity are found throughout the tetraploid An- dean clade of the American Amaryllidaceae. In the Clinantheae, the low- to mid-elevation genera Pa- mianthe and Paramongaia Velarde have pancra- tioid floral morphology, while the mostly high-ele- Clinanthus Herb. putatively ornithophilous flowers. In the more distantly related vation has colorful, petiolate-leafed Stenomesseae, Eucharis has the pancratioid flower; Plagiolirion resembles a mini- ature /smene subg. Elisena; and Stenomesson and Urceolina exhibit colorful, putatively ornithophilous flowers. Finally, in the Eustephieae, which is sister to rest of the Andean clade (Meerow et al.. 2000a). the full range of variation is evident in a single genus. Hieronymiella Pax (Hunziker, 1969). recurrent pattern suggests a scenario of rapid mo- 1984) within this 1987). The relatively low number of phylogenetically in- This sale evolution (sensu Stebbins, monophyletic, tetraploid group (Meerow, formative base substitutions in our sequence anal- 50 for ITS) supports a hypothesis of a relatively recent radia- tion within the Hymenocallideae tied to the rise of the Andes. This seems most significant relative to yses of non-coding regions (7 for trnL-F; Ismene, the most polymorphic of the three hymen- ocallid genera, and the only one that has adapted to high elevation. Hymenocallis is most speciose in Mexico (Bauml, 1979). with a secondary area of diversity in the southeastern United States (Smith & Flory, 1990, 2001: Smith et al.. 2001). Only species have been reported from South America: the broadly and coastally distributed H. littoralis, H. pedalis, and H. tubiflora. The genus does not occur at all in the Andes, and H. tubiflora is the three described only species of the three that is restricted to north- South America (including Trinidad-Tobago). The known distribution. of егп the Hymenocallideae suggests two possible hypotheses, either a long-dis- — ance dispersal event from the Andean center of 412 Annals of th Missouri ы Garden origin, or extinction of intervening populations of a proto-Hymenocallis ancestor. The fleshy seed of Hy- menocallis is the largest of all the endemic Amer- ican Amaryllidaceae, exhibits no dormancy, and germinates within 3—4 weeks after release, whether substrate (Whitehead & Brown, 1940; obs.). The relatively heavy seed does not im- or not in pers. mediately seem amenable to long-distance dispers- al, and no dispersal agent other than water has even been suggested for the genus. Thus ancestral ex- tinction is a more convincing hypothesis, but with- out a better understanding of the historical bioge- and a well-resolved ography of Hymenocallis - phylogeny of the genus a likely explanation for its distribution cannot be determined. In summary, combined trnL-F and ITS sequenc- es support the Meerow and Snijman (1998) treat- ment of Hymenocallideae with three genera: Hy- menocallis, Ismene, and Leptochiton. Leptochiton is sister to the Hymenocallis/Ismene clade and retains two plesiomorphic characters of the Andean tetra- ploid clade: 16 to 20 ovules per locule and a phy- tomelanous seed coat. The central Andean ende- mism of /smene and Leptochiton and the absence of Hymenocallis from this region further suggest a vi- cariance event at some point subsequent to the or- igin of the tribe. It is thus appropriate to formalize the recognition of the two new subgeneric combi- nations within /smene. Ismene subg. Elisena (Herbert) Meerow, comb. nov. Basionym: Elisena Herb., ceae, 75, 201. 1837. TYPE: /smene ringens (Ruiz & Pav.) Gereau & Meerow, Novon 3: 29. 1993. Ismene subg. Amaryllida- Pseudostenomesson (Velarde) Meerow, comb nov. Basionym: Pseudosteno- messon Velarde. Rev. Cienc. (Lima) 51: 47—51 1949. TYPE: /smene vargasii (Velarde) Gereau & Meerow, in L. Brako & J. — Monogr. Syst. Bot. Missouri Bot. Gard. 45: 1253. 1993. Literature Cited Baker, J. G. 1888. Handbook of the Amaryllideae. ( Bell and Sons, London. Bauml, J. A. 1979. A Study of — ed in Mexic versity, Ithaca, New Yor Bremer, К. 1988. The limi of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 198-213. Bull, J. J., J. P. Huelsenbeck, Swofford & P. J. Waddell. T 'urtitioning — com- ining — in phylogenetic analysis. Syst. Biol. 42: 384-397 C — M. We D. W. Stevenson, P. Wilkin € P. J. Rudall. 995. Monocot systematics: A combined analysis. Pp. »eorge V AN Genus Hymenocallis . Thesis, Cornell Uni- . Cunningham, D. L. ex 685-730 in P. J. Rudall, P. J. Cribb, D. F. Cutler & C. J. Humphries (editors), Monocotyledons: — and Evolution, Vol. 2. Roya ‚ Kew кзы n, R. M. T., H. T. Clifford & P. E ‚ 1985. Тһе amilies of the Monocotyledons. ее Вег- — ЖО, L^ d 1993. For consensus (sometimes). Syst. Biol. + 68-372. moms & J. Kim. 1995, Separate versus combined analysis of phylogenetic evidence. Ann. Rev. Ecol. Syst. 26: € э ›81. Douzery, J. P., А. ни оп, ДЕ Когеѕ, Н. Kurzweil, Р. Linder & M Chase: Э. Molecular ibe ties of Diseae ос ra contribution es nuc ae ribosomal ITS sequences. Amer. J. Bo E Dubuisson, ebant-Mauri & y ps ier. 1998. Molecules m alan: Conflicts a congruence ia the fern genus Tric — ا‎ eae) Molec. nel. Evol. 9: 390—X "arris, J. S., A luge & E Bult 1994, nificance of incongruence. Cladistics 10: : — ———, M. Kiillersjó, A. G. Kluge & C. Bult. 1995. Con- struc ang a significance test for incongruence. Syst. Biol. 44: 570—572. Felsenstein, J. 1985. Confidence limits on а таран An approach using the bootstrap. Evolution 39: 783— 791. = Testing sig- 5-319. Flory, W. S. 1976. Distribution, chromosome number and types of e species of Hymenocallis. The Nucleus 19: 204—2 Gereau, R. E. A. W. Meerow & L. Brako. 1993. New Combinations in Hippeastrum, Ismene, and Leptoc hiton (Amaryllidac eae) for the Flora of Peru. Novon 3: : Grant, V. 1983. The systematic and — cal distri- — жр in the temperate North n flora. Bot. 144: 439-449, Hammen, E Van der on The Pleistocene changes of — and climate in tropical South America. J. Biogeogr. 1: 2-26. 179. History of the flora, vegetation and climate in the C Vienn Cordillera Oriental during the million years. Pp. 25-32 in H. Larsen & L. B. Tropical Botany. Academic ast five Holm- Nielsen (editors), Press, 1837. ê andi Huelsenbeck, | P. J. J. Bull & €. W. Cunningham. 1996, ues. data in кин tic analysis. Trends Ecol. Evol. | 52-15 Hunziker, A. un Estudios sobre Amaryllidaceae. III. Sinopsis Dilanê) de Hieronymiella, y — Ar- gentinas юше о Kurtziana 5: 343-3 Hutchinson, J. 4. Families of — Plants. n 2. Monoc — Ist ed. MacMillan, London. . 1959. Families of Flowering Plants, Vol. 2. Mono- cotyle dex 2nd ed. Clarendon Press, Oxford, U.K. Kluge, A. G. 1989 phylo- ic hypot hod of а among FEpicrates (Boidae, Se ш! Syst. Zool. -25. Meerow, A. W. 1987. The — ne systematic rela- tionships of о — h ye e Baker ано ў :eae). Taxon 36: 566-5 Amaryllidaceae. James Ridgeway and . A concern for evidence and a 8 = 9. Systematics of the — lilies, Eucharis and culpan ннн Ann. Missouri Bot. Gard. 76: 136-22 = Volume 89, Number 3 2002 Meerow et al. 413 Phylogeny of Tribe Hymenocallideae ————. 1990. 202, ише In G. Harling (edi- tor), Flora of Ecuador No. 41. University of Göteborg. ————. 1995. Towards a — ny of the — 169-179 in P. J. Rudall, P. J. Cribb, . J. Humphries (editors), Monoc — Systematic 5 үе Evolution, Vol. 2. Royal Botanic Gar- de n5, — ^ é Jehgan. 1985. The auriculate pollen grain of Hence кл quitoensis Herb. (Amaryllidaceae) and its systematic in eee mer. J. Bot. 72: 540—5: Sehr: D. A. Snijman. 1998. alii Pp. 83— ы їп К. Kubitzki (editor), Vascular Plants, Vol. 3. Springer- — zerlin. ——, М. К. Fay, €. L. Guy, Q-B. Li, К. Q. Zaman & hase. 1999. —— er maryllidaceae based on cladistic analysis of plastid rhel. and trnL-F sequence data. Amer. J. hey 86: 1325-1345 ————, C. L. Guy, Q. Li & S-L. Yang. 2000a. Реза of the потем Pese eae i based on nrDNA ITS sequences. Syst. Bot. 25: 70 ———, M. Кау, M. W. Chase, С. E Q. Li, D. Snij- man & S-L. Yang. 2000b. Bis f ‘the Amarylli- daceae: Molecules and morphology. Pp. 308-382 in K. Ф ~ Families and Genera of е» 5 Wilson & D. Wallace (editors), Мох cols: Systematics ind Evolution. CSIRO Press, Sydne —— M. M. &W.M. h. 1995. Testing species phy logenies and phylogenetic methods with congru- 14: 61—76. 1965. The experimental taxonomy of the est African — Рапат L. (Amaryllida- eae). Kew Bull. 19: : 4 Müller- Doblies, D. & U. мл. Doblies. 1996. Tribes and subtribes 2 some species combinations in Amarylli- daceae J. St.-Hil. emend. R. Dahlgren et al. 1985. Fed- des Rep. 107070 S.c. Caen de sis. р — s 12: —24 . А. — in — netic systematic using scalps molecular е sets in the Solanaceae, Syst. Biol. 43: 467—481. Pax, F. id In A. Engler Pflanzenfamilien Il. Teil 5: mann, Leipzig. р) = On simultaneous — Combining data — Ап empirical approach lasted. С. & Amary llidoide 'ae-Narcisseae-Eucharidinae. Prantl (editors). Die natürlichen 110-11. Wilhelm Engel- ~ Ravenna. P. 1980. A new yellow-flowered — (Amaryllid ) from north Peru. Bot. Not. 133: 98. Rodrigo, A. ( . Kelly-Borges, P. R. Bergquist & P. L. Bergquist. ed A randomization test of the null hy- pothesis that two cladograms are sample estimates of a — phylogenetic tree. New Zealand J Bot. 31: a 257-26 к. а PJ, Chase, D. F. Cutler, J. Rusby & . De Кы 1998, ——— M de * sys- tematics of Asteliac eae and Hypoxidaceae. Bot. J. Linn. Soc ‚ 127: 1— Salisbury, R. A. . On the a of rare plants. Trans. Hort. з sh lon 1: 261—3€ Sealy, J. R. 1937. Lepidochiton — nsis. Bot. Mag. t. 949] Review of the genus Hymenocallis. Kew — Bull. pa 201-240, Seelanan, T., A. Schnabel & J. Wendel. 1997. Congruence and consensus in the cotton tribe (Malvaceae). Syst. Bot. 22: 259-2 Smith, G. L. & W. 5, Flory. 1990. Studies on — allis henryae (A 3rittonia 42: — ‚ 2001. — — exits eae) in Texas, with a new varietal combination. Novon 11: 229-232. —— ———, L. С. Anderson — W. S, Flory. 2001. A New Species P Pines callis (Am aceae) in Lower Central — P anhandle exon | |: 233-2 Soltis, D. E., H Soltis E. Mort, M. W. ar . M. Morton. 1998. Inferring complex —— s using parsimony: An empirical ap- pu lh using t Pa DNA data sets for angio- sperms. Syst. Biol. 47: 3242. — M. D. 1999. “Tree Hot, version 2. University of Mic — Boston University, Boston, Massachusetts. Stapf, O. 1933. Pamianthe peruviana. Bot. Mag. t. 9315. ие G. L. 1984. Mosai mosaic se — and angiosperm phylogeny. Bot. J. Linn. Soc. 68: 1 164. Swofford, D. L. ENS v. 4.0 с ev oluti tion, Analysis Using 1998. PAUP: Phylogenetic ) beta. Sinauer, Sutherland, Massac Жы Taberlet. P.. E. Gielly, С. Pautou & J. 1991. Uni- versal primers for amplification of three non-coding, re- gions of chloroplast DNA. РІ. Molec. Biol. 17: 1105- 1110 Traub, H. P. 1962. Ke species ш Нутепос allis. PL 3 Bouvet. y to ES aro nera, alliances and 59-72 2. ——— : iM eae. Plant 3 . Society, La Jolla, California. —— ———., 1980. Sec tions and alliances. genus Hymenocal- lis ded ti : 46-49. ;ene А of the American oO oN aan Life : Velarde, 1949. Pseudostenomesson un nuevo genero de iie eee Rev. Cienc. ma) ) 51: 47-5 White, T. J.. T. Bruns, s. Lee & J. Taylor. 1990. T fication and direct sequenc ом” of fungal ribosomal RNA genes for phylogenetics. Pp. М. Innis. D. Gelfand. | Sninsky & T. e ite — PCI cols: A Guide to Methods and Applications. Academic Press, 1 F Whitehead, — OF Brown. 1940. The seed of * spider ih СМИРАМ Amer. J. Bot. 129- was ex Н. Ohba, К. M. Cameron & M. Chase. ‚ Chloroplast E phylogeny of Ebr Bendio- bid (Orchidaceae): Insights from a combined analysis based on леі, sequences and restriction site variation. J. PL Res. 109: 169-176 PHYLOGENETIC AND Jun Wen,? Porter P. Lowry IL? BIOGEOGRAPHIC Jeffrey L. Walck,* and Ki-Oug Yoo’ DIVERSIFICATION IN OSMORHIZA (APIACEAE)! ABSTRACT Osmorhiza Raf. (Apiace ae) consists of 10 species — ‘tly distributed in temperate Asia (1 sp.) and the Americas (9 s pp.) Osmorhiza berteroi and О. depauperata show an American antitropical disjunction. ДИ North America, use two species are also disjunctly distributed in eastern pis western North America and the Great Lakes regions. A phylogene tic analysis was conduc * toc — inter- and — ific relationships based on sequences of the ITS and 5.85 regions of nrDNA. With Anthriscus, Geocaryum, and Myrrhis as outgroups, Hs monophyly of Osmorhiza is strongly — d. The ITS she — the basal position of the Asiatic O. aristata and the ‘monophyly of the nine orld species. The ITS sequence of Osmorhiza aristata is relatively dive ur from those of all other species even йи it is morphologically similar to the eastern North American О, claytonii and O. longistylis (which form a clade), suggesting early divergence followed by morphological stasis. Osmorhiza berteroi, O. brac 'hypoda, О. depauperata, О. mexicana, О, occidentalis, and О, purpurea constitute a monophyletic group (= western North American clade), The morphologically distinct O. glabrata from the central Andes forms a trichotomy with the eastern North American clade laytonii and O. longistylis) and the western North American clade in — and maximum likelihood analyses. The 11 populations studied of the widespread O. berteroi form a clade, and showed little sequence divergence, sug- gesting recent establishment of the widely disjunct populations following long-distance dispersal. Disjunct populations of O. depauperata from the Rocky Mountains and eastern North America have an identical ITS profile. Osmorhiza occidentalis, however, shows a high level of infraspecific sequence dive rgen nce. The ITS phylogeny and ihe low sequence divergence values suggest rapid рае Яе ‘ation of Osmorhiza in western North America. Key words: Apiaceae, biogeography, disjunction, morphological stasis, Osmorhiza. ze Osmorhiza Raf. (Apiaceae) comprises ten spe- genus (Lowry & Jones, 1984; see also Kartesz & cies, with one species in Asia, eight species in Ghandi, 1993), the sole Asian species, O. aristata, North America (three of which also occur in South is closely related to two species found in eastern America), and one species restricted to the central North America (O. claytonii and O. longistylis). The Andes. The distribution of species within the genus second distributional pattern is seen in O. berteroi provides an ideal model for studying the evolution and O. depauperata, both of which exhibit an an- of both intra- and intercontinental disjunctions, — titropical (often inappropriately referred to as “am- with three distinct patterns represented among its — phitropical,” see Cox, 1990) disjunction between temperate North and South America. Also, the more — members (Fig. 1; cf. also Lowry & Jones, 1984). According to the most recent classification of the distantly related O. mexicana subsp. mexicana has ' thank Mark McBroom and Kyoung-Ja Lee for laboratory assistance, and Dan Crawford, Fernando Chiang, Ric dnd R. Halse, Lawrence Janeway, Jan Jorgensen, Youngdong Kim, Sangtae Lee, Clodomiro Marticorena, Tod Stuess sy, Kouzi Yonekura, and Shiliang Zhou for help in obtaining le saf material. We are grateful to Bill Burger, Michael Dillon, and Tod Stuessy for helpful discussions on antitropical disjunctions, Carol Baskin for information on the evolution of seed dormancy, Steg ге 'n Downie for permission to use sequences of Anthriscus caucalis and Geocaryum macroc arpum, and three pecado reviewers, David Baum, and editor Victoria C. Hollowell for constructive comments. Curators of the following herbaria are acknowledged for pe — the examination of their specimens: A, CS, F, ‚ MSC, and RM. Jeff Walck thanks The Ohio Sti ч University for a postdoctoral fellowship. This study was supported in part by the Pritzker Laboratory for Molecular Systematics and Evolution of The Field Museum and by grants from the National Science Foundation (DEB-019605 1 de DEB-0108530) to J. Wen. * Department of Botany, Field Museum of Natural History, 1400 S Lake Shore Drive, Chicago, Illinois 60605-2496, U.S.A. wenOfieldmuseum.org. ' Missouri Botanical Garden, P.O. Box 299, St. Louis, sea — 0299, U.S.A.; Laboratoire de Phanérogamie, Muséum National d'Histoire Naturelle, 16 rue ns 75005 Paris, France. lowry@mobot.org. ' Department of Biology, P.O. Box 60, Middle Tennessee State ‘Unive 'rsit y, Murfreesboro, Tennessee 37132, U.S.A. jwale kGOmtsu.edu. ' Department of Botany, Field Muse ‘um of Natural History, 1400 S Lake Shore Drive, Chicago, Illinois 60605-2496, U.S.A. Present address: Division of Life Sciences, College of Natural Sciences, Kangwon National University, Chun- eheon, Kangwondo 200—701, Korea. а" ANN. Missouri Bor. GARD. 89: 414—428. 2002. > Volume 89, Number 3 2002 Wen et al. 415 Osmorhiza (Apiaceae) Y 18 >. Figure 1.—A. Distribution of Osmorhiza (: — eae). —В. and eastern North American disjunction. western North American disjunctions. a range that spans the American cordillera, with isolated populations extending from northern Mex- ico to northern Argentina. Finally, a third disjunc- tion is found in O. berteroi and O. depauperata, in which populations occur in western North America. the Great Lakes region, and northeastern North America. The disjunct distribution of plants between east- ern Asia and eastern North America is a classic topic in biogeography (Gray, 1840, 1878; Li, 1952: Graham, 1972; Thorne, 1972: Boufford & Spong- berg. 1983; Wu, 1983: Wen, 1998, 1999, 2001). and most genera exhibiting this pattern have now been well documented. Recent phylogenetic. stud- ies have confirmed the close affinity of the inter- continental disjuncts in many groups (reviewed in Wen. 1999), but most presumed species pairs stud- ied so far between the two continents do not show a direct sister-species relationship. Morphological stasis (i.e... the lack of significant morphological change in a species over long periods of time) has been suggested as a common phenomenon among these disjuncts (Parks & Wendel, 1990; Qiu et al.. 1995a, 1995b) and may at least partially explain distribution of Arrows indicate isolated areas of distebütiong Distribution of Osmorhiza sect. Osmorhiza, showing Asian f O. berteroi showing American antitropical, and eastern— To test the reported. findings (Wen, 1999, 2001). this hypothesis, Wen (1998) suggested the exami- nation of eastern. Asian—eastern North American disjunct taxa within the context of biogeographic studies covering the Northern Hemisphere or even at a global scale. In-depth comparative studies of disjunct groups that also have apparent close rel- atives in western North America may provide in- sight into the relative morphological, molecular, physiological, and cladogenetic rates of evolution of taxa in all three areas. \ntitropical disjunctions between western North America and southern South America were recog- nized as early as 1880 by Gray and Hooker, and have been discussed by many authors since (e.g.. Engler, 1882; DuReitz, 1940; Campbell, 1944: Constance, 1963; Raven, 1963; Cruden, 1966; Moore, 1972; Thorne, 1972, 1993; Cox, 1990; Pe- terson & Morrone, 1997; Peterson & Ortiz-Diaz, 998). Constance (1963) pointed out that the spe- cies exhibiting this pattern comprise a highly un- — representative sample of the floras in the two dis- that self-compatible and often autogamous. Based on junet areas, and many of them are 416 Annals of the Missouri Botanical Garden these and other considerations, Raven (1963) con- cluded that the most plausible explanation for the observed antitropical disjunctions is relatively re- cent long-distance dispersal, although Cruden (1966) presented arguments in favor of “mountain hopping” that involves shorter distance dispersal along the western cordillera. Fernald (1924, 1925, attention to disjunctions between western North 1935) was the first to draw America, the Great Lakes region, and northeastern North America, which he thought involved primar- ily arctic and western taxa that were able to survive during the Pleistocene in unglaciated areas (“nun- taks”) around Lake Superior, on the Gaspé Pen- but Й and in Newfoundland and Labrador, 3 could not uy their range following the glacia- tions because of their antiquity. However, all of Fer- nald's nunataks were in fact glaciated, and the no- tion of senescent species is now widely rejected (Wood, 1972 was proposed by Schofield (1969), who regarded ). A modified version of this hypothesis the eastern populations as remnants of a previously widespread flora whose members were able to sur- vive only in some areas (probably south of the gla- cial boundaries, but perhaps also in nunataks) and then move to their present sites following the Pleis- tocene, while being eliminated from the refugia. Lowry and Jones (1984) viewed this as a plausible explanation for the western North American—east- ern North American disjunctions in О, berteroi and O. depauperata, and further suggested that taxa oc- curring in the Great Lakes area and the northeast are most likely now restricted to sites with less competition from eastern boreal taxa and where cli- matic conditions (especially snowfall and moisture availability in spring) are similar to those in west- ern parts of the continent. Another fascinating aspect of Osmorhiza is that the type of seed dormancy differs among the spe- cies, and in particular between the putative close relatives in eastern Asia and eastern North Amer- ica. Very little comparative research has been done on physiological traits of species that exhibit this classical disjunct pattern (Terui & Okagami, 1993; Wen et al., 1996), ducted within a phylogenetic framework (Wen et with only one investigation con- al., 1996). Studies on seeds of five Osmorhiza spe- cies found that they have morphophysiological dor- mancy (MPD) (Baskin & Baskin, 1984, 1991: Bas- 1995; Walck et al., 2002). Seeds with MPD have embryos that are very small relative to kin et al., the size of the seed, and the embryo must grow to the full length of the seed to germinate. In addition to being small, the embryos also have physiological dormancy that must be broken before seeds can 1998). Seeds of О. claytonii and О. longistylis have nondeep complex germinate (Baskin & Baskin, MPD, i.e., they require warm followed by cold strat- ification to germinate, and gibberellic acid (GA,) substitutes for warm stratification (Baskin & Bas- kin, 1984, 1991). In contrast, seeds of O. aristata and those from western North American popula- tions of О, berteroi and O. occidentalis have deep complex MPD, requiring only cold stratification to overcome dormancy, and GA, does not substitute for stratification (Baskin et al., 1995; Walck et al., 2002). Thus, mapping the types of MPD on a phy- logenetic tree can provide an opportunity to eval- uate the evolution and adaptive significance of this physiological trait. ‘onstance and Shan (1948) proposed an infra- generic classification for Osmorhiza that was largely adopted by Lowry and Jones (1984) with a few modifications (Table 1). Several hypotheses of re- lationships are implied in this classification. Two subgenera are recognized, one of which (subg. Gly- cosma) comprises a single western North American species, O. occidentalis, which would thus repre- sent the sister species to a clade that includes the remaining nine species. Within the typical subge- nus, three sections are defined (each with three species), which are likewise assumed to comprise monophyletic groups. This classification thus offers 1 framework against which to test alternative hy- potheses of relationships using molecular sequence data that were not available previously. The objectives of this study are to: (1) recon- struct the phylogeny of Osmorhiza using sequence: of the internal transcribed spacer (ITS) and 5.85 regions of nuclear ribosomal DNA; (2) examine the biogeographic diversification within the group: (3) re-evaluate the earlier hypotheses regarding the possible origins of the observed disjunct distribu- tional patterns; and (4) examine the evolution of seed dormancy in a phylogenetic framework for dis- junct and type(s) of MPD are plesiomorphic versus derived. taxa of Osmorhiza determine which Sequences of ITS and 5.85 regions were employed because they have been shown to be appropriate to assess evolutionary relationships within other groups exhibiting disjunct distributions between Asia and North America (Fritsch, 1996; Lee et al., 1996; Wen & Zimmer, 1996; Wen et al., 1998) as well as interspecific relationships in Apiaceae 1998) other north temperate. plants (e.g.. Xiang et a 1998; Lee & Wen, 2001). (Downie et al., and evolution within many MATERIALS AND METHODS Forty-eight populations representing all ten spe- cies of Osmorhiza and three outgroup taxa were Volume 89, Number 3 2002 Wen et al. Osmorhiza (Apiaceae) Table (1984), as modified from Constance and Shan ( Taxa of Osmorhiza Raf. and their distributions following the classification scheme of Lowry and Jones 8) Classification Taxon Distribution Subgenus Glycosma таш Drude Subgenus Osmorhiz O. brachypoda Torr. O. glabrata Phil. О mexicana Griseb. Section Mexicanae п ИНВ & Shan ex Lowry & A. С. Jones subsp. bipatriata (Constance & Shan) wry € A. G. Jones subsp. mexicana Section Nudae Constance & Shan 0. berteroi DC. & G. Jones ex Lowry ¢ O. depauperata Phil. O. purpurea (J. M Coult. & Rose) Suksd. Section Osmorhiza ). claytonii (Michx.) € o longistylis (Vorr.) DC. O. occidentalis (Nutt.) Torr. s aristata (Thunb.) Rydb. W North America California, Nevada, and Arizona central Andes SW Texas and N Mexico N Mexico to N Argentina W North America, Great Lakes area, NE North America, and South America W North America, Great Lakes area, NE North America, and South America NW North America te — Asia . B. Clarke EN E Lu America rth America sampled in this study (Table 2). Populations of widespread disjunct taxa such as O. berteroi and O. depauperata were examined from throughout much of their distributional ranges. Anthriscus caucalis M. Bieb., Engstrand. chosen as the outgroups because of their close re- 1984: Geocaryum macrocarpum (Boiss. & Spruner) and Myrrhis odorata (L.) Scop. were lationship to — za (Lowry & Jones, Downie et al., 200 Total DNA was — with the CTAB method of Doyle and Doyle (1987) and purified over EsCl/ DNA amplifications were performed in 100-j.L reactions following Wen and Zimmer (1996) using the primers С26А and Nncel8810 (see Wen «€ Zimmer for sequences of The entire ITS and 5.85 regions were se- ethidium bromide gradients. primers). quenced manually from both directions following Wen et al. (1998) using four primers: (5.85. C264. ITS4, and NI8LI8 (see Wen € Zimmer for se- quences of primers). The DNA sequences obtained were assembled. and the boundaries between the coding and spacer regions were determined by comparing them with the sequences of Daucus carota L. (Yokota et al., 1989). The sequences were then exported to PAUP* (vers. 4.0, Swofford, 1999). substitutions, allowing Most mutations were base thus manual alignment. All the sequences have been deposited at GenBank (see r Table 2 for accession numbers). Phylogenetic analyses were performed with PAUP* using maximum parsimony (Swofford et al.. maximum likelihood (Felsenstein, 1981). 1987) methods. Parsimony analysis was performed using a branch- and-bound search with MULPARS and furthest ad- dition sequence options. The amount of support for 1996), and neighbor-joining (Saitou & Nei, monophyletic groups revealed in the maximally parsimonious tree(s) (MPTs) was examined with 1000 bootstrap replicates (Felsenstein, 1985) with random addition and heuristic search options. The maximum likelihood analysis was performed with the input order of sequences randomized and the transition/transversion ratio set at 1.42 based on the observed frequencies in the MPTs of the parsimony А neighbor-joining tree was constructed distance. (Kimura. analvsis. using Kimura. two-parameter 1980) Relative rate tests were performed with the meth- od of Wu and Li (1985) to detect any rate asym- metries of the ITS and 5.85 regions among taxa in Osmorhiza. The proportions of site differences were estimated using the Kimura two-parameter distance (Kimura, 1980). The optimal area cladogram was constructed from taxon cladograms using the optimality method in COMPONENT (vers. 2.0, Page. 1993). The fol- lowing options were used: nearest-neighbor inter- changes and minimizing the number of leaves add- endemism were defined for ed. Four areas of Osmorhiza based on the distribution of taxa and previous biogeographic studies of the North Tem- 1970: 1981: perate zone (e.g.. Wood, Patterson. 418 Annals of the Missouri Botanical Garden V86ESVAV C86ESPAV C86CSPAV 8LOESPAV I86€£€T.4 V 086€S€T AV 6L6€ST AV LLOESPAV 9LOESVAV €L6ESVAV CSOESPAV UPA] оләпұ “ODIXIJA TA ONW *oorxo]y ору “ODIXIA `о7) 3pga&e 4 *&xyonjuaN "sn `07) e3eieg “URBINA ус °09) PANPIEA, "еце" цроу "V 5 : °0) SATIN eming "Vig OpeloW [4 Nd *o2enueq Р 31U21 4 әр; : вое, әр “4014 ‘OYO &a][PA ешүеәт “opto 07) snowy *oaqon() peue `о7) 1әшилР^] *opea1opo?) “VS 03 SPJUOYRIOG “PIUIBATA 1524 “VS °07) INAL J "AXonquaN "VS '07) лоше *pue[&re|y “YSA o7 e3e1eg “URSIYOI “VSN 707) шер "erueA[4suuag “y Sr] 707) ээри “AYonjusy “VS оу зип wuuepnm y) VST ‘09 uone *uo224() ус °03 SPUR YORI) *uo3a4() “y SN *07) 9dse«) *ooqon() “epeue”) '07) snowy *23qon() *epeue) o9 om] *ue3rjory “W'S ‘o9 mng ‘вішоуео "vis n ‘ory — "PIUdOJI[e7) “WIS ‘Jug Uouov dar ‘AOIg Buetfey7 ‘euyy (4) 2816 poumy f W X 1 тот а d (SO) 9/21 Funryy 4 (SD) (86-1A-01) "ws SUDIQU Hs (4) £99 UDADPIH "NES Y PPM TT (4) с/р vay T (V) 198 vay f (ү) 228 vay f (OW) 61 FZ 1” 19 12021 "S (SO) 6131 210421 “Wx (SO) 8111 24042 "IW (4) crzr voy f (4) orcg uM Tx (A) 29€ охор ^N "S Y PPM 71 f (4) cog UDÁDPIY N 7S X PPA 7T f (SD) ESE РИО А (OW) 0667 7” 12 1] мот d d (OW) 6f6F JP 1? [] тот d d (SO) 8t&6 PIPNSO ‘H A (S9) 014€ ә] `q X «omoupf Y 7] (4) 2211 чәзппоу 7») Y vzavg ^w (4) GSES] `P 12 Assanig [ (4) 929 Y9S1NOY 7) ӯ гәр "y (A) €29 asoy ©) Y эгэр "Ws (4) us ung M ^g Y 977 HT (4) Yes Dang2uoq ^N (4) "ws NOYZ “Sy Dnjmuuqpdiq ‘dsqns рирлхәш cf, : | i 0 DUDIIXIUL 'dsqns Dunpoixout `0 siasiduo] °0 070190]8 °0 pımıadnndap ‘0 1101012 `0 vpodíy. 2010) ( ) 10191199 '() DIDISUID `0 U0ISSIJIB xueg Ud‘) JIINOS 19YINOA ПОХР], “(€ 3[qep 99S) S]S9] Ә]Р1 PAN POI əy} ur posn әләм whe чим po»xreur suornoo[[o?) 'pe[duies dnoi3jno 34] pue sardads 021410ш5() jo SUOISSIIOY A 3[4B.L Volume 89, Number 3 002 Wen et al. 419 Osmorhiza (Apiaceae) Continued. Table 2. GenBank accession Source Voucher Taxon AF453985 AF453986 orado, Garfield Co. U.S.A., Co *B. Painter et al. 18 (CS) J. Wen 3860 (F) O. occidentalis — (MN mM b- oO © е ¢ DD DQO uU» min ш E ы ш a 333 a з „оо C . 2255 ы Itc o > =e Vo AO. € x Laon. # oos eod dd ue т” ЖЫ = ue s BEES bef a © зы d = с "a “a 0 осоо 4j uiui d unt [db Mi = = ™ N QN Mm +s DO € دہ‎ 2 ES > © = NOUS AS ‘Sy ой + bip ые Ше os S Rk 8 sa 6 = = 5 о S S & س‎ о A б — AA A д = ITA Seen U S.A., Nevada, Lander Co. AF453989 К. К. Halse 5561 (CS) U.S.A.. Oregon, Benton Co. U.S.A R. R. Halse 5565 (CS) AF453994 AF453995 Oregon, Tillamook Co. U.S.A., Oregon, Linn Co. + P. P. Lowry 11 4963 (MO) P. P. Lowry 11 5069 (MO) O. purpurea AF453996 U.S.A., Oregon, Linn Co. Downie et al. (1998) 070 (MO) 5 P. P. Lowry Il Anthriscus caucalis 3605. AF073606 901 AFO AFO (2000) Downie et al. Geocaryum macrocarpum ad Valiejo-Roman et al. (1998) Myrrhis odorata Hoey & Parks, 1991), work on Osmorhiza (Lowry & Jones, analysis. These areas are western North America, 1984), and results of this phylogenetic eastern North America, central Andes, and Asia. South America was not chosen as one of the areas of endemism because it is linked to western North America via disjunction and shares the same spe- cies, suggesting that it was recently colonized (also see Discussion). RESULTS CHARACTERISTICS OF ITS SEQUENCES The combined length of the ITS], 5.85, and 1782 regions in Osmorhiza species 15 005 bases, with an ITS] of 217-220 bases, a 5.85 of 164 bases, and an ITS2 of 224-225 bases. Two insertions in ITS] were required to align the sequences within the ge- nus. One occurs in the three accessions of Asian О. aristata (1 bp), and the other is a 3-bp insertion 1 O. mexicana subsp. bipatriata. Two additional insertions are in ITS2, each 1-bp in length. The alignment of sequences of Osmorhiza with those of the outgroup species required five additional 1-bp and one 2-bp indels, and sequences of Osmorhiza and the outgroups could be aligned manually with- out difficulty. The sequence divergence of Osmor- hiza taxa and the outgroups ranges from 4.679 (be- tween Myrrhis odorata and О. berteroi) to 14.05% (between Anthriscus caucalis and O. mexicana subsp. mexicana). PHYLOGENETIC ANALYSES Of the 693 aligned positions, 116 sites were var- iable: of these, 62 were phylogenetically informa- tive. Treating gaps as new characters, the parsi- mony analysis generated a single most parsimonious tree (MPT) with a total length of 157 steps, a consistency index (CI) of 0.860 (0.763 ex- cluding uninformative characters), a retention in- dex (RI) of 0.912, and a rescaled:consistency index (RC) of 0.785 (Fig. 2). Treating gaps as missing data, the parsimony analysis produced a single MPT an identical topology to Figure 2 with a 141 steps, a Cl 0.851 (0.756 when se abel. characters are excluded), an. RI 0.974, and an RC = 0.916. The tree from the weighted parsimony analysis (weighting transver- sions over transitions 1.42 times) also had an iden- tical topology to the MPT shown in Figure 2. Sev- length « eral relationships are suggested by the parsimony analyses: (1) the three populations of O. aristata from Asia form a monophyletic group: (2) O. clay- Annals of the Missouri Botanical Garden 10 100 18 — 1 «50 C2 of of of of o eo © — — <50 50 Nondeep Complex MPD Clade alis eocaryum macrocarpum С Myrrhis odorata aristata China aristata Japan aristata Korea rteroi675 5 [South (c berteroi15556| berteroi4726 fe berteroi4732 Quebec berteroi4949 Oregon berteroi4950 Oregon [ l l [ E Asian Clade oo об 22 9 o N berteroi5710 California berteroi9248 California berteroi4707 Michigan brachypoda9253 California depauperata1118 S America W N American Clade d s9493 California dentalis4633 Nevad occidentalis5155 Nevada dentalis5561 — dentalis5565 Ore occidentalis9532 California purpurea4963 Oregon purpurea5069 Oregon purpurea5070 Oregon past plea — mexic mex кө " bipatria ta ytonii565 Kentucky [567 West Virginia laryland Mexico i812 V 11069 Kentucky ¡1089 Pennsylvania ¡4716 Michigan longistylis5 64 P Kentucky ngistylis 827 Virginia longistylis861 N Carolina longistylis4715 Michigan 1219 E М American Clade Central a2419 Andean Clade Outgroups Eurasia Figure 2. The single most i tree of Osmorhiza with a — length of 157 steps, a СЇ of 0.860 (0.763 excluding uninformative characters), an RI of 0.912, above lines names are accession numbers as in and ar 1 RC of 0.785, treating gaps as new characters. Numbers are br ranch lengths and those below are bootstrap v: ues 's in 1000 replicates. Numbers following the taxon Table 2 to help identify the source of study material. The solid black bar at the node of the О. — O. longistylis clade indicates the most parsimonious explanation of the evolution of the nondeep complex morphophysiologic: al dormaney (MPI tonii and O, longistylis from eastern North America American clade); (3) the multiple populations examined of the widespread form a clade (the O. depauperata form a monophyletic group; (4) O. depauperata and O. occidentalis form a clade: (5 O. brachypoda and O. purpurea are allied with the O. depauperata—O. occidentalis clade; (6) popula- tions of the widespread O. berteroi form a mono- phyletic group; (7) several species (O. berteroi, О. brachypoda, O. depauperata, O. mexicana, O. oe- cidentalis, and O. purpurea) form a largely western North pauperata extending to eastern North America and American clade with O. berteroi and O. de- South America, and O. mexicana ranging from Tex- as to South America (the W N American clade); and (8) О, glabrata from the central Andes forms Volume 89, Number 3 2002 Wen et al. Osmorhiza (Apiaceae) Table 3. at the 5% level.) Relative rate tests to detect rate asymmetry. Myrrhis odorata was used as the reference taxon. (* Significant Species 1 Species 2 K; K K,-K,, + SE O. aristata O. berteroi 0.06582 0.05165 0.01417 + 0.00611* О. aristata O. claytonii 0.06582 0.06015 0.00567 + 0.00619 O. aristata O. depauperata 0.00582 0.06759 —0.00177 + 0.00701 O. aristata O. glabrata 0.00582 0.05347 0.01235 + 0.00545* O. aristata О. mexicana 0.060582 0.0644 0.00140 + 0.00669 O. aristata O. occidentalis 0.06582 0.06993 —0.00411 + 0.00752 O. berteroi O. claytonii 0.05165 0.06015 — 0.00850 + 0.00499 O. berteroi O. depauperata 0.05165 0.06759 —0.01594. + 0.00450* O. berteroi O. glabrata 0.05165 0.05347 —0.00182 + 0.00397 O. berteroi O. mexicana 0.05165 0.06442 —0.01277 + 0.00398* O. berteroi O. occidentalis 0.05165 0.06993 —0.01828 + 0.00493* O. claytonti O. depauperata 0.06015 0.06759 —0.00744. + 0.00619 O. claytonii O. glabrata 0.06015 0.05347 0.00668 + 0.00415 O. claytonii O. mexicana 0.06015 0.06442 —0.00427 + 0.00574 O. claytonii O. occidentalis 0.06015 0.06993 —0.00978 + 0.00641 O. depauperata O. glabrata 0.06759 0.05347 0.01412 + 0.00520* O. depauperata O. mexicana 0.00759 0.00442 0.00317 + 0.00481 О. depauperata O. occidentalis 0.06759 0.06993 —0.00234. + 0.00347 O. glabrata O. mexicana 0.05347 0.06442 —0.01095 + 0.00477* O. glabrata O. occidentalis 0.052347 0.06995 —0.01646 + 0.00550* O. mexicana O. occidentalis 0.06442 0.06993 —0.00551 + 0.00568 a trichotomy with the E N American clade and the W N American clade. The maximum likelihood tree (MLET, not shown, with a log likelihood of = 1512.26) has an identica topology to the MPT. The neighbor-joining tree (NJT) is similar to the MPT and the МЇЛ but differs in that (1 ) O. glabrata is placed sister to the O, claytoniti-O. longistylis clade in the NJT: and (2) the O. claytonii-O. longistylis—O. glabrata clade is sister to the Asian О. aristata. The new clades are weakly supported with bootstrap values less than 50% SEQUENCE DIVERGENCE Treating gaps as missing data, the Kimura two- parameter distance among species of Osmorhiza was estimated to be 0.209-3.784%. gence values are comparable to those in These diver- lralia sect. Aralia, which shows a similar disjunct pattern of distribution, with representatives in eastern Asia and in eastern and western North America (Wen et c al., 1998). O. occidentalis from western North America and € The highest divergence occurs between aristata from eastern China, and the lowest — O. claytonii and O. longistylis from eastern North America, which are sympatric throughout much of their ranges. Overall, divergence values between O. aristata and other members of the genus are com- paratively high, ranging from 1.627% (with O. clay- tonii) to 3.784% (with O. occidentalis). Among the New World species, sequence divergence of O. gla- brata from the central Andes ranges from 0.601% (with O. longistylis) to 2.552% (with O. occidental- is). The western North American O. occidentalis has a relatively high level of sequence divergence, ranging from 0.319% (with O. 2.552% (with O. glabrata). Infraspecific variation — depauperata) to was detected in six species for which multiple sam- ples were available, O. aristata (O-0.302%), O. ber- terol (O-O.473%), O. claytonit (O-0.517%), O. mex- icana (0-1.415%). O. occidentalis (0-1.013%), and O. purpurea (0-0.148%)., lations sampled of O. depauperata and O. longis- whereas multiple popu- tylis showed no variation in their ITS profiles, de- fact thal distributions. The absence or relatively low level of spite the these species have wide sequence variation within these taxa supports the species level cireumscriptions proposed by Lowry and Jones (1984). On the other hand, infraspecific occidentalis was 1.013% be- tween populations from the Rocky Mountains and variation within O, those from California. RELATIVE RATE TESTS Twenty-one relative rate tests (Wu & Li, 1985) were conducted to detect rate asymmetry (Table 3). Myrrhis odorata was used as the reference taxon. Rate differences of most pairs of species were not Annals of the Missouri Botanical Garden “7? 0 — YY МА 3B 75'E са 5 $ Figure 3. Biogeography of Osmorhiza.—A. Area cladogram of Osmorhiza based on the nuclear ribosomal ITS phylogeny. —B. Model of diversification in Osmorhiza. statistically significant at the 5% level, whereas eight species pairs showed significant differences. These tests suggest that O. berteroi had a slower rate of nucleotide substitutions in the ITS regions than O. aristata, O. depauperata, O. mexicana, and O. occidentalis. They also show that O. glabrata had a slower rate of nucleotide substitution than O, aristata, O. depauperata, O. mexicana, and O. oc- cidentalis. The molecular clock hypothesis for the ITS sequences in Osmorhiza was therefore rejected. AREA CLADOGRAM CONSTRUCTION The area cladogram (Fig. 3A) was constructed with the MPT and MLT using the optimality method in COMPONENT (vers. 2.0, Page, 1993). Among the four areas of endemism, Asia is basal and east- ern North America, western North America, and central Andes form a trichotomy. DISCUSSION PHYLOGENY AND PATTERNS OF DIFFERENTIATION 'arsimony and maximum likelihood analyses support the basal position of the Asian O. aristata within the genus and the monophyly of the New World taxa; these relationships were previously suggested in Downie et al. (2000) in a broader anal- ysis of the tribe Scandiceae of Apiaceae. Within the New World species, three major subclades are suggested: (1) the eastern North American O. clay- tonii and O. longistylis; (2) a largely western North American clade comprising O. berteroi, O. brachy- poda, O. depauperata, O. mexicana, O. occidentalis, and O. purpurea; and (3) the central Andean en- demic O. glabrata. The neighbor-joining tree, how- ever, groups the central Andean endemic O. gla- brata with the eastern North American O. claytonii and O. longistylis. The O. claytonii-O. longistylis— O. glabrata group is then sister to O. aristata from Asia. As in the MPT and the MLT, the NJT also suggests a largely western North American group comprising O. berteroi, O. brachypoda, O. depau- perata, O. mexicana, O. occidentalis, and O. pur- purea. The ITS phylogeny conflicts in several aspects with the phylogenetic hypothesis implied in the in- frageneric classification of Lowry and Jones (1984) based on morphology (cf. Table 1). Osmorhiza oc- cidentalis is a morphologically distinct species and was thus originally described in the monotypic ge- nus Glycosma by Nuttall (in Torrey & Gray, 1840). Recent treatments have included this species with- in Osmorhiza, but placed it in its own section (Con- stance & Shan, 1948) or subgenus (Lowry & Jones, 1984) because of the numerous features that dis- tinguish it from other members of the genus, in- cluding glabrous fruits lacking a caudate append- age. numerous staminate umbellules and flowers, bipinnate leaves, and yellow to greenish yellow flowers. The inclusion of O. occidentalis within a well-supported western North. American clade the ITS phylogeny suggests, however, that its dis- Volume 89, Number 3 2002 Wen et al. 423 Osmorhiza (Apiaceae) tinctive morphological characters are most likely attributable to rapid evolution within the species, resulting in many autapomorphies. Another discrepancy between the current infra- generic classification and the ITS phylogeny in- volves the three species currently included in Os- morhiza sect. Osmorhiza: the Asian O. aristata and the eastern North American O. claytonii and O. longistylis. These taxa form a morphologically co- herent group and were even treated as a single spe- cies by some authors (e.g.. Gray, 1859; Clarke. 1879; Kuntze, 1891; Boivin, 1968). They share sev- eral diagnostic features, including ап involucre composed of (12)2-3(-5) conspicuous, foliaceous bractlets, and styles (including the high-conic sty- lopodium) that are 1-3.6 mm long. However. based on ITS data, O. aristata is not only basal within the genus, but also shows a high level of sequence di- vergence from its congeners (1.627—3.784%), sug- gesting its early divergence within the group. The relative antiquity of O. aristata and the compara- tively high level of morphological similarity. be- tween it and the eastern North American species are consistent with the hypothesis of morphological stasis among eastern Asian—eastern North Ameri- can disjuncts, as proposed by Parks and Wendel (1990). are sister to the western North American clade and Taxa of the eastern North American clade not to O. aristata. The morphological similarity be- tween 0. aristata, O. claytonit, and O. longistylis could, however, be explained by their having ex- perienced morphological stasis while concurrently the largely western North American clade diversi- fied more rapidly. Morphological stasis was first proposed by Parks and Wendel (1990) as a possible explanation for the presence of morphological similarities among eastern Asian—eastern North American disjunct taxa that are not closely related or do not represent sister groups. They found that two morphologically similar disjunct species of Liriodendron (L. chinense (Hemsley) Sargent and L. tulipifera L.) show a high level of allozyme and cpDNA divergence. Molecu- lar and fossil data suggest the divergence time of mid- — the two species to be 10-16 million years ago dle Miocene). Morphological stasis has also been suggested to explain similarities between the Asian and North American species of Aralia sect. Dimor- phanthus (Wen, 2000). Liquidambar (Hoey Parks, 1991; Shi et al., 1998), and Magnolia sect. Rytidospermum (Qiu et al., 1995a, 1995b) A third conflict between the current infrageneric 2 classification of Osmorhiza and the ITS data in- volves the central Andean endemic 0. glabrata. The relationships and taxonomic placement of this species have been somewhat ambiguous. Constance and Shan (1948) included О. glabrata in their sec- tion “Glycosmae” along with O. occidentalis, largely based on apparent similarities in their fruits. By contrast, Lowry and Jones (1984) placed the An- dean endemic in section Mexicanae, which also comprises O. brachypoda and O. mexicana. Shared characters among these species include short styles (0.5—1.2 mm long), a low-conic stylopodium, a con- spicuous involucre of 1—6 bractlets, and an absent or poorly developed involucel subtending the um- bellules. By contrast, the maximum parsimony and the maximum likelihood analyses suggest that О. glabrata forms a trichotomy with the eastern North American clade (О. claytonii and O. longistylis) and The neighbor-joining tree, however, places O. glabrata the largely western North American. clade. eastern North American O. claytonti— The ITS phylogeny (Fig. 2) thus suggests the relative antiquity of O. glabrata sister to the O. longistylis group. among the New World species. It seems most plau- sible that O. glabrata was derived from a North American ancestor, perhaps shared with O. clay- tonii and O. longistylis (as in the NJT) or with taxa of the eastern North American clade and the west- ern North American clade (Fig. 2). The present nar- row distribution of O. glabrata in the central Andes may be due to local survival following the Pleisto- cene glaciations, which have been considered im- portant in the evolution of the Andean flora (Vuil- 1971; 1975, 1983; 1982). Additional data are, however, leumier, Simpson, Prance, needed to as- certain the exact phylogenetic position of O. gla- rata. INFRASPECIFIC VARIATION Differentiation within O. occidentalis. Osmorhiza occidentalis showed a relatively high level of infra- specifie sequence divergence, with a Kimura two- parameter distance of 1.013% between populations from the Rocky Mountains and California. Morpho- logically, O. occidentalis is somewhat variable, and several segregate species have been recognized in the past based primarily on differences in inflores- cence structure, leaf pubescence, and fruit size (see 1984). were regarded as minor by Lowry and Jones (1984), These differences, however, owry & Jones, who provided data showing that a broadly defined O. occidentalis comprises a morphologically coher- ent group despite the presence of a few rather atyp- ical collections (some of which served as the basis for the segregate species). Because O. occidentalis is so distinct from its congeners, it may be that infraspecific morphological variation has been over- 424 Annals of the Missouri Botanical Garden looked by previous workers. Additional studies should be undertaken to re-assess variability within O. occidentalis and to evaluate whether the molec- ular divergence is correlated with morphology and/ or geographic distribution. Differentiation between O. mexicana subsp. mex- icana and subsp. bipatriata. The two subspecies of O. mexicana show a relatively high level of ITS sequence divergence (1.415%). Moreover, they did not form a monophyle tic group in the phylogenetic analysis (Fig. 2). Initially, Constance and Shan 1948) recognized these taxa as distinct species, but Lowry and Jones (1984) treated them as sub- species because of the presence of occasional mor- n „ч. phological intermediates at several localities northern Mexico, including one area (Cerro Poposí, Nuevo León) where they co-occur with typical ma- terial of both subspecies. The biphyly of O. mexi- cana, as currently circumscribed, suggests the need to reexamine the status of subspecies bipatriata and the possible causes of the morphological interme- diates. Although no interspecific hybridization has been reported for Osmorhiza, occasional morpho- logical intermediates have been observed in sym- patric populations of O. berteroi and O. occidentalis in western Oregon (R. Halse, pers. comm.). BIOGEOGRAPHY Diversification in major geographic areas. The ITS phylogeny (Fig. 2) shows that rates of clado- genesis vary among the four major areas occupied by Osmorhiza: eastern Asia, eastern North America, western North America, and the central Andes. Cladogenesis clearly appears to have been more rapid in the New World than in Asia, and the west- ern North American clade shows a particularly high level of species diversity (Fig. 3B). It is unusual that Asia is species-depauperate considering that most other Asian—North American disjunct genera show a higher species diversity there (Tiffney, 1985; Wen, 1999). The number of Osmorhiza spe- cies occurring in each area may in part be the re- sult of differential rates of extinction, especially during the Quaternary glaciations. However, it is generally accepted that North America was more severely affected by glaciation than eastern Asia due to its more complex topography and the north— south rather than primarily east-west orientation of its mountain ranges (Axelrod et al., 1998; Wen, ¢ . Western North America is the center of species diversity in Osmorhiza, with six of the ten species occurring in this region. The monophyly of the western North American clade is well supported in all three analyses presented here. Morphologically, however, the clade is rather diverse, with members from both subgenera and two of the three sections recognized by Lowry and Jones (1984). The ITS phylogeny suggests two successive diversifications within Osmorhiza in western North America. First, the common ancestor of the western North Ameri- can clade may have diversified into O. berteroi, O. mexicana, and the ancestor of the O. brachypoda— O. depauperata—O. occidentalis-O. purpurea sub- clade. The latter subclade then appears to have dif- ferentiated further. This diversification among western North American Osmorhiza may have been highly influenced by the availability of a range of habitats associated with the uplifting of the Rocky Mountains and the western cordillera during the Tertiary (Barbour & Christensen, 1993; Graham, 1993). The Andean endemic O. glabrata most likely represents an isolated relict. Its phylogenetic po- sition is not well-resolved in the ITS phylogeny. It may have diverged rapidly early in the evolutionary history of the North American Osmorhiza, persist- ing in a relatively small portion of the central An- des. The fruits of O. glabrata are glabrous to sparsely hispid, which may afford them limited op- portunities for long-distance dispersal by animals. It should be noted, that O. occidentalis almost always has completely glabrous fruits but however, nevertheless extends throughout a much larger area of western North America (Lowry & Jones, 1984). Eastern Asian—eastern North American disjunc- — tion. The relative antiquity of the eastern Asian— eastern North American disjunction in Osmorhiza is suggested by the phylogenetically basal position of the Asiatic O. aristata, the monophyly of the diverse New World species (Fig. 2), and the rela- tively high ITS sequence divergence between О, aristata and its congeners. The most commonly ac- cepted interpretation of the origin of eastern Asian— eastern. North American disjunctions is that this pattern reflects an initial widespread distribution of temperate forest elements in the Northern Hemi- sphere during the mid Tertiary followed by subse- quent extirpations in. western North America and western Europe as a result of late Tertiary and Qua- ternary climatic cooling (Graham, 1993; Manches- ter, 1999; Wen, 1999). No fossils of Osmorhiza have been reported that could help to date the biogeo- graphic disjunction. Also, the rejection of the mo- lecular clock hypothesis for the ITS sequences in Osmorhiza (see Results) makes it questionable to attempt an indirect estimate of the times of diver- gence between the eastern. Asian and the North American members of the genus. Volume 89, Number 3 2002 Wen et al. 425 Osmorhiza (Apiaceae) The apparent close relatives of Osmorhiza, the genera Myrrhis and Geocaryum (Downie et al., 2000), are restricted to the Old World, which is consistent with a hypothesized Old World origin of Osmorhiza (Downie et al., 2000). The basal position of the Asian O. aristata and the Old World distri- bution of Myrrhis and Geocaryum are consistent with the idea that the ancestor of the North Amer- ican Osmorhiza migrated from Asia. American antitropical disjunction. Osmorhiza berteroi and O. depauperata show a similar pattern of antitropical disjunction between western North America and temperate South America. Constance (1963) suggested that these species may have mi- grated south in a step-wise manner along the west- ern American cordillera throughout the Tertiary. with subsequent elimination of populations from in- tervening tropical areas. The present-day distribu- tion of O. mexicana subsp. mexicana, which com- prises isolated populations that bridge the areas currently occupied by O. berteroi and O. depauper- ata, appears to support the idea that the now-dis- junct species could also have had a more contin- uous distribution in the past. However. the absence of ITS sequence divergence between the western North American and South American populations of both species suggests a recent origin of the ob- served M disjunctions. Species of Osmor- hiza are facultatively autogamous, which would en- able establishment of a new population from a single propagule, and their armed fruits with cau- date appendages and numerous retrorse bristles ap- pear to be well adapted for epizoochorous dispersal 1984). (1963), the pattern of disjunction of temperate taxa North South America corresponds closely to the migration (Lowry & Jones, s indicated by Raven between western America and southern routes of many bird species, and this probably ac- counts for the disjunctions seen in О, berteroi and O. depauperata (Lowry & Jones, 1984). Thus, the lack of ITS sequence divergence, the monophyly of both species, and their fruit morphology are all con- sistent with an origin of the antitropical disjunction due to relatively recent long-distance dispersal via birds from western North America to South Amer- ica. A similar explanation has been used for several disjunct taxa of grasses (Peterson € Morrone, 1997; Peterson & Ortíz-Diaz, 1998). North The phytogeographic disjunction between Eastern—western American — disjunc- tion. America, the Great Lakes region. North cussed repeatedly since it was first reported by Fer- nald in 1925 (Fernald, 1925, 1935: see also Steb- 1935; 1953; 1969; western North and northeastern America has been dis- bins, Rousseau, Drury, Schofield, 1969; Morisset, 1971: Voss, 1972: Wood, 1972; Miller & Thompson, 1979). erally regarded that the disjunct populations of spe- It is now gen- cies showing this pattern are remnants of a previ- ously widespread flora that probably survived south of the glacial boundaries, but perhaps also in nun- ataks (cf. Fernald, 1925), and then migrated to their present sites following the Pleistocene, while being eliminated from the refugia in the south (Schofield, 1969; Miller & Thompson, 1979). In explaining the western North American—eastern North American disjunctions in O. berteroi and 0. depauperata, Lowry and Jones (1984) also emphasized that taxa occurring in the Great Lakes area and the northeast are most likely now restricted to sites with less competition from eastern boreal taxa and where cli- matic conditions are similar to those in western parts of the continent. In the present study, the western cordilleran populations of O. berteroi showed an identical ITS profile to those of the Great Lakes region, while the eastern populations had The fact that there is little or no sequence divergence within just a single nucleotide substitution. each of these species is consistent with a relatively recent origin of the disjunct pattern during the gla- cial or postglacial periods. PHYSIOLOGICAL ADAPTATION——SEED DORMANCY Disjunct an ies of herbaceous plants with relict distributions in. both. eastern. North America and Asia appear to exhibit stasis in ecological traits (Ricklefs € Latham, 1992). Similarly, closely re- ated species groups with members occurring in one or the other area may also show stasis with regard lo various traits, including the type of morphophys- iological dormancy (MPD) they exhibit. This is the case for seeds of the eastern North American—Asian disjunct species pairs — diphylla (L.) Pers.—J. dubia (Maxim.) Ben f. (Ber- beridaceae) and Panax на L.—P. ginseng C. A. Meyer (Araliaceae), all of which have deep simple MPD (cf. Baskin € Baskin, 1998). In Os- seeds of O. aristata from Asia morhiza, however, and 0. berteroi and O. occidentalis from western North America share the same type of MPD (deep complex MPD), whereas those of the eastern North American O, claytonit and O. longistylis exhibit a different type of dormancy (nondeep complex MPD). Wake et al. (1983) suggested that changes in some traits, including physiological ones, might help an organism to compensate for maintaining a stable morphology over long periods of time. This would appear to be the case in members of Os- morhiza. 426 Annals of the Missouri Botanical Garden Although much additional work must be com- pleted before we have a robust understanding of the evolution of various types of MPD, some insights into their evolution may be drawn from the phylog- eny of Osmorhiza. The plesiomorphic condition in the genus appears to be deep complex MPD, where- as nondeep complex MPD is derived. This is fur- ther supported by studies that show seeds of the outgroup Myrrhis odorata also have deep complex MPD (Lhotska, 1977; Deno, 1994). The suggestion of Baskin et al. (1995) that deep complex MPD may have been derived from nondeep complex MPD is not upheld according to the ITS phylogeny of Os- morhiza. Seeds of most (and probably all) Osmorhiza spe- cies are dispersed during summer/autumn, with germination occurring in spring (Baskin & Baskin, 1984, 1991; Baskin et al., 1995; Walek et al., 2002). A cold stratification period would thus en- sure that germination does not occur until the fol- lowing spring, since dormancy is broken with cold (winter) stratification. The selective forces respon- sible for the evolution of the requirement for a warm stratification pretreatment prior to cold strat- ification observed in the eastern North American O. claytonii and O. longistylis are unclear. The most parsimonious explanation based on the ITS phylog- eny is that the evolution of this trait occurred in the common ancestor of these closely related sister species. It seems reasonable, however, to assume that a common environmental selection pressure (or a set of pressures) was responsible for the acqui- sition of the warm + cold stratification require- ment. Interestingly, species of Erythronium have a similar geographic pattern of MPD types, with seeds from the western North American E. gran- diflorum Pursh exhibiting deep complex MPD and those from the eastern North American E. albidum Nutt. and E. americanum Ker “ж i non- deep complex MPD (Baskin et al., )5). Literature Cited 1998. His- 55 in A.-L Axelrod, D. I., I. Al-Shehbaz € P. H. Raven. tory of the men flora of China. Pp. 43— Zhang & S.-G. 1 (editors), Floristic Characteristics and Diversity of Pd Asian Plants. China Higher Ed- ucation — Beijing. Barbour, M. G. N. L. Christensen. 1993. Vegetation. Pp. 97-1: 3l in F NA Editorial Committee, Flora of North America North of Mexico, Vol. 1. Oxford Univ. Ne ^w 'or * Baskin, С. C. & J. M. Baskin. 1998. Seeds: Ecology, Bio- ER and pu of Dormancy and Germina- Aca ademi E Mant Press, tion. © Press, San Diego. E. al dormancy in seeds of two genera & J. M. Baskin. 1995. Two types of (Osmorhiza and e а ап Arcto-Tertiary distribution pattern. 82: 293-298 Baskin, J. М. & С. С. 84. Cemuinaliun eco- physiology of the woodland he à —— longistylis (Umbelliferae). — ч Bot. 37—692. ——— QA ——— к, p come morphophys- Ismorhiza claytonüi T r in — J. Bot. 3. Flora of the prairie provinces. Phytologia & iological ласеае к, (Ар 78: 588-593. Boivin, ig 5 a eae D. E. & S. A. Spongberg. 1983. Eastern Asian— astern North American phytogeographic val relation- ships—A history from the time of Linnae ^i to the twen- tieth € gr Missouri Bot. Gard. 70: 423—439. — ll, D. 944. Relations of the te е rate — с orth me PR America. Proc. Calif. Acad. Sci. 25: 3 — — 1879. Umbelliferae. Pp. 665-720 in J. D. eer (editor), Flora of British India, Vol. 2. ewe Ker 1963 а val relations in the her- baceous Hora of the Coast of North and South America: Á symposium. ра — = historical re- view. zn rt. Rev. n 09-1 —— . H. Sha — n genus Osmorhiza (Um- be — = © 2 study in pue affinities. Univ. Calif. deu : 111-15 НА e. L. — * Cox, C. re . New — 'al theories and old biogeo- phi al pre ms. J. Biogeogr. 17: 117-130. Cruden, Birds as agents of long-distance dis- persal or — plant groups of RE tempe rate West- Evolution 20: 517 Deno, N. C. 1994. Seed Germination: Theory and Prac- tice, nda a oth ра Published by the author, State C өн, Pennsylva Dow 5. RS. panier Katz-Downie & К. l — 1998. Molecular systematics of Apiaceae sub- family — ae: Phylogenetic analyses of nuclear ri- bosomal DNA i Чегин. transcribed spacer and plastid rpoC | intron sequences. Amer. J. Bot. 85: 563-591 —— ——-, D. S. Katz- — & K. Spalik. 2000. A phylog- eny of Evidence from nu- clear — DNA epe — ribed spacer se- quences. Amer. J. Bot. 87: Doyle, у 1 & 1 L. Doyle. 1 E is Ai ‘rapid DNA isolation proc — for — — s of fresh leaf tissue. Phy- toche m. e -15 Drury, Bi 1969. Plant persistence in the Gulf of St. Lawrence. Pp. 105-148 in H. Greenidge (ed- itor), Essays in — deeem and Ecology. Nova Sco- tia Museum, Halifa DuReitz, G. E. 1940. а of bipolar plant distribu- tion. Ac т — орт. Suec. 13: —282. Engler, ; 382. Versuch einer Entwicklungsgeschichte dir — ‘пме, insbesondere der Florengebiete seit der Tertiárperiod. Engelmann, Leipzig. Felsenstein, J. 1981. Evolutionary trees DNA se- A maximum likelihood approach. J. Molec. ern He misphe re. Apiaceae tribe Scandiceae: 5— from quences Evol. 17: 308-370 1 85 Confide nce limits on phylogenies: ү ар- proac ‘h — k the bootstrap. Evolution 39: 783- Fernald, M. L. 1924. Isolation and endemism in — eastern en rica and their relation to the age-and-area hypothesis. Amer. J. Bot. 11: 558—573. —.. 1925. Persistence of plants in unglac iated areas Arts 15: 239— T ә. of — America. Mem. Amer. Acad. Volume 89, Number 3 2002 Osmorhiza (Apiaceae) 1935. Critical plants of the Upper Great Lakes region of Ontario and Michigan. Rhodora 35: 1-16, 47— 120-140. 161-185, 203— — 230-247. 265—283, 298-315, 327-346, 364-386. : )5—403. Fritsch, P. 1996. Isozyme analysis of interc — dis- juncts within Styrax (Styracaceae): da rure for the Madrean- Tethyan hypothesis. Amer. J. Bot. 83: 342 355. Graham. A. 1972. Floristics and ind deir of Asia and Eastern North America. Else . Amsterdar 1993. History of the Хаара: Cretaceous MEI Pp. 57-70 in FNA Editorial — Flora of North America North of Mexico, Vo P 7s d Univ. Press, New York. . Dr. Siebold, Flora Japonica: sectio prima. — ornatui m usui inservientes; digessit Dr. J. G. Zuccarini: fasc. ). fol. (A review). Amer. J. Sci. 39: 175-176. 1859. Diagnostic characters of phanerogamous да ted in Japan by Charles Wright, Pacific Exploring Expedition, with observations upon the relationship of the Japanese flora to that of North America and of other parts of the north- ern temperate zone. Mem. Amer. Acad. Arts 6: 377- = botanist of the U.S. North I 53. 878. Forest geography = archaeology, a lec- the val rd University Natural r. J. Se far ts, ser. 3, 16: 85-94, ture delivered be " € History I iety. 183-19 E 1 D. Hooker. 1880. The vegetation. of the Rocky Mountain region and a Um with that of other parts " vu world. U.S. Geol. Geogr. Survey Ter- ritory Bull. 1—77. oey, M. T. Parka, 199]. [sozyme divergence between eastern Asian, rth American and Turkish species ds куы (Hamamelidaceae). Amer. J. Bot. 947. Kartesz, 1 T. №. Ghandi. 1993. Osmorhiza berterot (Api- correct name for mountain sweet-cice y. 1-182. A simple — for estimating evolu- aceae): The Brittonia 15: 18 itution through comparative . J. Molec. Evol. 16: tionary rates of base subs studies of nucleotide sequence 2 1891. Revisio Generum Plantarum. Felix. Lee, N. S., T. Sang. D. J. Crawford, S. H. Yeau & S.-C. Ў 1990. Molecular divergence between disjunc t taxa in eastern Asia and eastern North America. Amer. Bot. 83: 1373-1378. Lee. S. € J. Wen. 2001. A phylogenetic analysis of Prunus and the Amygdaloideae (Rosaceae) based on ITS se- quences of nuclear ribosomal DNA. Amer. n Bot. 88: 1: — Lhotska. M — — Li, H.-L. 1952. Floristic Asia and eastern North America. Soc. 42: 371—429. Lowry, P. P, H & A. G. Jones. 1984 [1985]. Systematics of Osmorhiza En (Apiaceae). Ann. Missouri Bot. Gard. 71: 1128-117 Manchester, 5. ^ 1999. Biogeographical relationships of North American Tertiary oras. Ann. Missouri. Bot. Gard. 86: 472—5: Miller, N. G. & G. G. Thompson. 1979. Boreal and west- . Notes on the ecology of ge rmination in Phytotax. 12: 209-213. © relationships between Eastern Philos. Folia Geobot. Trans. Amer. ern North American plants in the late Pleistocene of Vermont. J. Pr Arbor. 60: 167—218. Moore, D. M. 1972. Connections between cool temperate — тк ‘шаг reference to southern America. Pp. 115-138 in D. Н. Valentine (editor), Tax- onomy, мония d and Evolution. Academic Press. London. — P. 1971. Endemism in the vascular plants of the Gulf of St. Naturaliste Canad. 98: 139-149. Page, Н. D. M. 1993. COMPONENT, Muset “um ae History), Londor Parks, 1. & J. F. Wendel. 1990. Molecule divergence between we and North American species of Liriod- endron (Magnoliac eae) with — ations of fossil floras. mer. J. Bot. 77: 1243-12 Patterson, C. 1981. The d ИИ of the North Amer- ican fish fauna—4A problem of historical biogeography. Pp. 265-281 in P. E Forey (editor), Chance, Change and Challenge, Vol. 2. The Evolving Biosphere. British Museum (Natural — and Cambridge Univ. Press, London. Lawre ence region. Vers. 2.0. British Peterson, 1997. Allelic the кык eks al ane t Lycurus setosus (Poaceae: 1-34 Muhle — ae). Madrofio 44: 33 6. . Morrone. variation in — ‚ Ortiz-Diaz. 1098. Allelic variation in the — rop iad disjunct Muhle — ~ yi (Poaceae: Muhle nbergiinac) Brittonia 50: 381-3 . 1982. Biological ту ‘ation in the Trop- = Prance, ics. Columbia Univ. Press, New York. Qiu, Y.-L., E — & C. К. Parks. roplast DNA perinde study of the « eastern North America disjunct section — rmum Amer. J. 1 1582- 1995a. A chlo- astern Аха e ا‎ (Magnoliaceae). . R. Parks & M. W. Chase. 1995b. Molecular divergence in the eastern Asia—eastern North America disjunct section Rytidospermum of — (Magnoli- aceae). Amer. J. Bot. 82: 1589-15 Raven, Р. H. 1963. Amphitropical — in the lo- ras of North and South America. Quart. Rev. Biol. 3 15l- Ricklefs E. & R. E. Latham. 1992, corre — of geographic ranges suggests stasis in eco- logical traits of relict genera of Aero ИЙ perennial E iue d 139: 1305-1321. The value of iny as indicator of unglac dole a areas. Mém. Jard. Bot. Montréal 40; 1-8. 87. The neighbor-joining method: Intercontinental = 2 2 = = Pp L ^ ш = = Saitou, N. & M. Nei. 19 A new method v FAR ting phylogenetic trees. Molec. jus Evol. 25 Schofield, B. ен EL mds of northwestern North Veces Bryophytes and vascular plants. Madro- ño 2( : 5-207 Shi, 5.. H-T. E Cine Y. Chen, L.-H. Qu & J. Wen. 1998. — пу of the Hamamelidaceae based on the ITS se d nees of nuclear ribosomal DNA. Biochem. Syst. Ecol. 25: 55-6 — n . B. 1975. —— changes in $ vun of the high — al Andes. Paleobiology 1: 273-29 1983. An Жи. алб ШОТ of de hh Alain Ani Revista Chilena 50: 122. Stebbins. = L.. Jr. 1935. Some — ations on ni flora of the Bruce Peninsula, Ontario. Rhodora 37: € 3-7 1. e . L. 1999. PAUP*: Phylogenetic Analysis Us- ing Sunderland. Mass Vers. 4.0. — r, с is, tts. 428 Annals o Missouri Botanical Garden —— Olsen, P. J. bien & D. M. Hillis. 1996. — aba inference. Pp. 407-514 in D. M. Hillis Moritz € B. K. Mable (e — Molecular Syste m- atics, 2nd ed. Sinauer, Sunderland, Massachuse Terui, K. & N. Okagami. 1993. Temperature кенен оп seed к germination of east Asian and Tertiary relict spe- Amer. J. Bot. 80: oO ^ч T doscorea (Dioscoreaceae). 9 Thorne, К. 19 ranges of a TIRA — 1% Major и іп E oe Juart. Rev. Biol. 4 SEE Pp. | * М in pum A Ed- of North America North of xico, Vol. 1. O ыы Univ. Press, New York. Tiffney, B. H. T Perspectives on the origin of the flo- ristic similarity between eastern Asia and eastern North America. J. Arnold Arbor. 66: 73-94, Torrey, J. & A. Gray. 1840. Flora of North America, Vol. Part 4. Wiley € Putnam, New Yor Valiejo- — К. M., M. G. Pimenov, E. 1. Terentieva, S. R. De , D. S. Pa ees & A. V. Troitsky. 1998. Molecular. ae matics of e Using — rDNA internal transcribed spa quences to reso issues of evolutionary rel: dine Tol Zhurn. 83: 22. torial € EU e, к е е К. G. 1972. Michigan E ar 1. u onoc ols. — Inst. . Bull. Vuilleumie г, В. 5. 1971. Pleistocene c — in же айпа and flora Y South е Science 173: 771—780. £ ا = pe aad rms —188 Wake, D. B., С. Roth & M Wake. 1983. On the prob- lem = — in — evolution. J. Theor. Biol. 101: —224. Walck, : L., 5. N. Hidayati < N. Okagami. 2002. Seed germination ecophysiology of the Asian species Osmor- hiza aristata (Apiaceae): Comparison with its North American congeners and implic pde dor oo of types of dormancy. Amer. J. Bot. 998 eastern Aa * eastern North American disjunct pattern: onis bom phylo- genetic — Korean J. РІ. Taxon. 28: 63-81. ————. 1999. Evolution of eastern. Asian and eastern North b. 'an disjunet parte rn in flowering plants. Ann. Rev. Ecol. Syst. 30: 421—455. ~ Internal transcribed spacer phylogeny of e astern North American disjunct Aralia sect. —— (Araliaceae) and its biogeographic im- plications. Int. J. Pl. Sci. 161: 959-966 u “2001. Evolution of eastern Asian—eastern North American ЕТ disjunctions: A few additional issues. Int t. J. РІ. Sci. 162: $117-S122. & ы; 1996. Phylogeny and biogeog- raphy of Panne | „ (the — genus, Araliaceae): In- ferences from I Is sequences ¢ ue л "ar ribosomal DNA. Molec. Phylogenet. Evol. € - 166- . К. Jansen & К. — 1996. Evolution of E — rn Asian and eastern North American disjunct genus Symplocarpus (Araceae): Insights from chloro- plast DNA restriction site data. Biochem. Syst. Ecol. 24: 735-747. ———, 5. Shi, K. Jansen & E. A. Zimmer, 1998. Phyloge ny т ы of — sect. Aralia (Ar- aliac eae). . Bot. 85: 866-87 Wood, € "1970. Some floristic relationships between the onthe тп Appa 'hians and western North America. Pp. 331404 in P. C. Holt (editor), The Distributional History of the Biota i the Southern Appalachians, Part Il: Flora. Research Division Monograph 2. Virginia Polytechnic Institute and State University, Blacksburg, Virginia. . 1972. deal aan d ie phytogeography: The clas- sical — :h to the study of disjunctions. Ann. Mis- " 4. ue Bot. Gard. E 107-12: Wu, C.-I. € W.-H. 1985. Evidence for higher rates at nucleotide ышы in rodents than in man. Proc Natl. Acad. Sei. U.S.A. 82: 1741-1745. Wu, 7. Y. 1983. On the significance of Pacific intercon- tinental discontinuity. Ann. Missouri Bot. Gard. 70: 977-56 B Q.-Y Crawford, A. D. Wolfe, Y.-C. Tang & J. . De E 1998. Mi d ч н а of i har L. (Hippocastanaceae): ju cular phyloge- netic pe rspec live. d 52: ou Yokota aw . lida, | — 8. Tanifuji. 1989. m de otide s — nees of A: 5.85 rRNA gene and internal transcribed spacer regions in carrot and broad bean ribosomal DNA. J. Molec. Evol. 29; 294—301 A SYNOPSIS OF OCOTEA (LAURACEAE) IN CENTRAL AMERICA AND SOUTHERN MEXICO! Henk van der Werff ABSTRACT Ocotea is the largest genus of Lauraceae in Mesoamerica (Central America and Southern Mexico, i.e., the area between the Isthmus of Tehuantepec in Mexico and the Panamanian—Colombian border) with 102 species — jd. The most recent treatment of Mesoamerican Ocotea was published by Carolyn Allen in 1945. It included 3 it is curre ntly known some brief notes on characters useful in their identification, and is now se — outdated. A synopsis of the genus as consisting of a lo species, synonymy, type Табынай, and distribution rid country and altitu ide. Key words: Central Americ a, Lauraceae, Mesoamerica, Ocotea, spec ies from Mesoamerica is present d here, southern Mexico. Lauraceae are a large and ecologically important family of trees and shrubs in wet tropical forests from sea level to the tree line at about 3000 m. The family is rare or lacking in areas with a pronounced dry season. Lauraceae have the reputation of being a difficult family, partly because of problems with generic delimitation and partly because many spe- cies are trees with inconspicuous flowers and there- fore difficult to find and are infrequently collected with flowers. A key to the genera of Lauraceae in the Neotropics was published by van der Werff (1991); this characters but did not include fruit. characters. key makes it almost impossible to use because herbar- key was based on floral and vegetative Combining fruit and floral characters in : ium specimens do not as a rule have both flowers and fruits. By far the largest of the genera of Laur- aceae in the New World is Ocotea Aubl., with an estimated 300+ species in the Neotropics (van der Werff, 1991) or 350 including the African and Mad- agascan species (Rohwer, 1993). Ocotea is char- acterized by its stamens with 4 locelli; these ar- ranged in 2 pairs, unisexual or hermaphrodite flowers, a lack of papillae on the stamens or tepals in most species (papillae are present in a few spe- cies, but these have the locelli clearly arranged in 2 pairs), and the tepals free, not basally united, and falling off in old flowers or rarely persisting on the cupule. In this way Ocotea is not very well defined. and both van der Werff (1991) and Rohwer (1993) suggested that Ocotea serves as the catchall genus for species with 4-celled stamens that do not fit in the other, better-defined genera of Lauraceae. The Central American and Southern Mexican (from the area between the Isthmus of Tehuantepec in Mexico and the Panamanian-Colombian border, hereafter referred to as Mesoamerica) species of Ocotea were (1945), who included 33 spe- cies in the genus. Since then, additional collections ast revised by Allen from Chiapas and Guatemala have led to the de- scription of several new species by Lundell (1965. 1969, 1970, 1971, 1974a, 1974b, 1977, 1978). In the more recent treatment of Lauraceae for the Flo- ra Costaricensis (Burger & van der Werff, 1990) 42 species of Ocotea were recognized, including 7 new species. Rohwer (1991) reviewed the marginal spe- cies of the O. helicterifolia group, including the O. heydeana group, and described four new species. Van der Werff (1999) treated the core species of the O. helicterifolia group and found four additional new species. The study of recent collections of Me- soamerican Lauraceae has resulted in publications in which new Ocotea species have been or will be described (Gomez-Laurito, 1993, 1997; Hammel, 1986: Lorea-Hernandez & van der Werff, 2002 in Nelson, 1984; van der Werff, 1987, 1988a, press: 1988b. 1996, 2001: Wendt, 1998; Wendt & van der Werff. 1987). SPECIES GROUPS IN OCOTEA In a genus as diverse as Ocotea, one can expect to find several distinct. species Rohwer (1986) published an overview of the species of Nec- groups. ! I thank n curators of B, BR, CAS, F, K, NY, US, and TEX for loans and Gordon McPherson for critically reading the manuscr ? Missouri ‘Botanic al Garden, ANN. MISSOURI Bor. P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. henk.vanderwerff@mobot.org. GARD. 89: 429—451. 2002. Annals of the Missouri Botanical Garden Table 1. Ocotea species with unisexual flowers in the Mesoamerican region. Table 2. Species of Ocotea subg. Dendrodaphne in the Mesoamerican region, Ocotea adela Ocotea puberula Ocotea rubrinervis ex Rottb., Ocotea, Nees, and Rhodostemonodaphne Rohwer € Kubitz- tandra Roland. Pleurothyrium ki, largely based on a study of type specimens. He recognized a number of species groups in Ocotea based on characters of flowers and fruits. Among the Mesoamerican species the following groups can be recognized, starting with the best-defined groups: l. Species with unisexual flowers. This group comprises almost half the species of Ocotea and is best represented in South America. In Mesoamerica one finds 12 species of this group (Table 1). Most of those are also present in South America, and only two are endemic to Mesoamerica (O. adela van der Werff and O. atlantica van der Werff). Rohwer (1986) divided this group into numerous smaller groups, based largely on the South American spe- cies. The Mesoamerican species can all be confi- dently assigned to these smaller groups, but listing these would be too detailed for this synopsis. . Ocotea subg. Dendrodaphne (Beurl.) Mez. This group is characterized by a combination of the following characters: inflorescences in the axils of bracts near the tips of the twigs; cupules with a double-rimmed margin; flowers with spreading te- pals; and nearly sessile, tongue-shaped stamens with a sterile tip and a papillose indument. This is a small, but well-defined group consisting of about 10 species, with 4 species in Mesoamerica (Table 2 The Ocotea helicterifolia group. This group shares some characters with the species of Ocotea subg. Dendrodaphne and can be recognized by the following combination of characters: inflorescences in the axils of normal leaves; cupules with a single margin; flowers with spreading tepals and stamens subsessile or infrequently with a distinct filament, sometimes tongue-shaped, densely papillose and with a sterile tip or more commonly rectangular, without a sterile tip and very sparsely papillose. Ocotea dendrodaphne Ocotea klepperae Ocotea morae Ocotea veraguensis This group has been recently revised by Rohwer (1991), who studied O. heydeana (Mez & J. D. Smith) Bernardi and its allies, which have glabrous or very sparsely pubescent leaves, and O. sinuata (Mez) Rohwer and its allies, which have densely pubesc ent leaves and densely papillose stamens (1999), who studied O. helicterifolia (Meissner) Hemsley and its allies, characterized by pubescent leaves and stamens with few papillae, without a sterile tip. This group is strongly centered in Me- soamerica, with only one species in northern South America, and includes about 32 species (Table 3). The Ocotea insularis group. This group is characterized by two seemingly insignificant char- acters. One is the pubescence of the inner 3 sta- Table 3. Species of the Ocotea helicterifolia group in the Mesoamerican region. l. Ocotea acuminatissima 2. Ocotea arcuata 3. Ocotea — 4. Ocotea betazer E Ocotea botrar . Ocotea bonrgeauaiana . Ocotea brenesi " ( s congregata 2. =: 9. Ocotea corrugata 10. Ocotea gordonii 11. Ocotea helicterifolia 12. Ocotea heydeana 13. Ocotea — I4. Ocoted lenti 15. Ocotea agni 16. Ocotea тойс 17. Ocotea n 18. Ocotea pa шаса ] 9. Ocotea — * hrosorum 25. Ocoted rubriflora 26. Ocotea sinuata 27. Ocotea tonii 28. Ocotea valeriana 29. valerioides 30. Ocotea verticillata 31. Ocote — ^ad > Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea Table 4. Mesoamerican region. Species of the Ocotea insularis group in the 1. Ocotea atirrensis 2. Ocotea austinii 3. Ocotea barbatula Ocotea chi | 5. Ocotea contrerastt 6. Ocotea endresiana ea whitei 16. Ocotea viridiflora mens; these stamens have a patch of white hairs on the side facing the pistil at the junction of the fil- ament and the anther. The second character 15 found in the inflorescence: the secondary, tertiary. and further branches of the inflorescence are flat- tened, with the axis of flattening of the tertiary branches forming a 90° angle with those of the sec- ondary branches, those of the quaternary branches forming a 90° angle with the tertiary branches, and so on. In other characters this group conforms to the general Ocotea pattern: inflorescences in the axils of leaves; cupules with a single margin: flow- ers with erect to half-erect tepals and stamens with a distinct filament, not papillose, and without a sterile tip. This group is best represented in Me- soamerica, with one or a few species in northern South America. I would place 16 Mesoamerican species in this group (Table 4), including O. me- ziana С. К. Allen and its allies, which differ from the O. insularis group mainly in the light green drying leaves with raised reticulation and often yel- lowish major veins. My concept of this group considerably wider than that of Rohwer (1986). larger un eyond these reasonably well-defined, groups one can establish a larger number of poorly defined, small groups, but doing so is beyond the scope of this synopsis. Too many species would be transitional between such groups, and they would contribute little to our understanding. If Ocotea is the genus to which all species that cannot be placed in other, better-defined genera of Lauraceae are assigned, one may ask if Ocotea, as currently accepted, is monophyletic. The answer to this question can be drawn from a recent DNA- based phylogeny of Lauraceae (Chanderbali et al., 2001). In this study 22 species of Ocotea, repre- senting several different groups, were included. Those results show clearly that Ocotea is not monophyletie genus, but that the following clades can be recognized: 1. The 9 species with unisexual flowers formed a clade together with Endlicheria Nees and Rho- dostemonodaphne, both genera with unisexual flowers. 2. The 3 species of the O. helicterifolia group formed a clade. The 2 species of subgenus Dendrodaphne formed a clade. 4. The 4 Old World species formed a clade. 5. Three species with perfect flowers (including O. insularis (Meissner) Mez) formed a clade with 3 Aniba species and 1 Aiouea species. 6. A recently described species from Cameroon, O. ikonyopke van der Werff, was part of a clade with Cinnamomum species and was probably incor- rectly placed in Ocotea. These clades are well separated from one anoth- er, and maintaining Ocotea as a monophyletic ge- nus would require a transfer of several quite dis- tinct genera (such as Nectandra, Pleurothyrium, Aniba Aubl., Licaria Aubl.) into Ocotea, a solution that Edo not find acceptable. Hence Ocotea should be considered a polyphyletic genus. Mez (1889) recognized 198 species in Ocotea and divided the genus into 4 subgenera: subg. Hemiocotea Mez, with 1 species, now placed in the genus Urbano- dendron Mez; subg. Dendrodaphne, 8 species, with the same circumscription as it currently has; subg. Mespilodaphne (Nees) Mez, consisting of all re- maining species (51) with perfect flowers: and subg. Oreodaphne (Nees) Mez, including all species (138) The DNA study supports the monophyly of subgenera Dendrodaphne and Or- eodaphne, but not of subgenus Mespilodaphne. Mor- with unisexual flowers. phological characters for a division of subgenus Mespilodaphne in monophyletic groups have not yet been found This synopsis is based on a treatment of Ocotea for the Flora Mesoamericana and includes the 102 species found from the Isthmus of Tehuantepec in Mexico south to the border between Panama and Colombia. Mexican species only known from the area north of the Isthmus of Tehuantepec are not included. Altitudinal distribution for species also occurring outside of Mesoamerica is based on Me- soamerican specimens and does not reflect their al- titudinal distribution outside this area. Annals of the Missouri Botanical Garden KEY TO THE SPECIES OF OCOTEA IN MESOAMERICA 1: 3(12). 13 14(13). ee 2221). 22". 23(20). 23' Stamens tongue-shaped, with a distinct, sterile tip, moderately to densely papillose; filaments lacking E VEE lii ec P а. ns not tongue-shaped, usually т rectangular, without a ste rile tip, not inni or rarely with papillae on the margin; filaments usually Present ы A — l Lowermost pair of lateral veins ihi conspicuous Lcd domatia — 41.0. holdridgeiana Leaves without conspicuous poc * ‘t domatia ‘Twigs and usually the leaves densely pubescer ent ‘Twigs and leaves glabrous or ance m — { pee clustered— a - 99. O. verticillata Leaves evenly distributed along the twig - 5 Lower leaf surface completely covere d pe the indument, the hairs short and cure е — 66. О. pharomac hrosorum Lower leaf surface not completely covered by the indument, but partly or — completely visible ) Leaves elliptic, to 6 cm wide; petioles 5—9 mm long Tm 46. O. klepperae Leaves obovate or obovate-elliptic, 5—17 cm wide; petioles 1. 54 cm long 7 Inflorescences paniculate-cymose; pistils с pubescent п 84. nuata Inforescences racemose; pistils glabrous m 13. 0. urina ‘Twigs hollow | — Twigs solid ү Shrubs or small trees to 8 m; leaf bases plane, not revolute; inflorescences to 15 ст long; fruits ca. 2 X l em; сирше ca. | em diam., cup- to bowl-shaped, very weakly канон e е 1. 0.5 mm wide; widespread |... 1. 0. deboli Troes to 20 m; leaf bases slightly revolute; infloresc 'ences to 5 em long; fruits ca 5 E m cm, cupule ca. 5 em diam., bowl-shaped, with a conspicuous double margin, this ca. 1 em de Costa Rica E TAANI E A инал O A 0. morae — ‘ences about as long as the leaves, пената pubescent; Teal bakea — not гак pus 2 em, cupule to 1.2 em diam., with a double margin, the outer margin ca. 2 mn wide | widespr reac a — . 0.1 — Inflorescence ‘es much shorter than the le saves, de nsely pubescent; leaf bases slightly — fruits 5 X 3.5 em, cupule ca. 5 em diam., with a conspicuous double margin, this ca. 1 em wide: — Rica ‚ O. morae Lower leaf шк е complete sly hidden by the indument — а. 1: Lower leaf surface glabrous or partly covered by the indument 5 17 Flowers ш айди leaf base inrolled and decurrent along the entire length of the petiole; — lines visible on the lower leaf surface — 16. О. calophylla Flowers perfect; leaf base plane or if inrolled and decurrent, most of the petiole distinct; vernation lines not present — — — Leaf M inrolled and shortly decurrent on the petiole | 83. O. salvinii se plane and not decurrent 1 indias nt of the lower leaf surface pale brown or whitish у. 15 Indument of the lower leaf surface reddish brown zm 16 Domatia absent; flowers 6 mm diam., tepals 2-2.5 mm long — 0000 82. O. salvadorensis Domatia present as axillary tufts of hairs; flowers 4 mm — tepals 1.5 Ер long . 43. O. iridescens Receptacle glabrous inside; tepals papillose on the inner surface; leaves 9-16 cm long 81. O. rufescens Receptacle appressed pubescent inside; tepals pubescent on the inner Ре 5 leaves 7-10 ст long see discussion iul 82. O. salvadorensis Twigs hollow, often inhabited by ants me 18 Twigs solid 0 00000000 n 19 Inflorescences and flowers de ansely brown-tomentellous; leaves s broadly obovate, coriaceous, with raised reticulation 44. O. jefensis Inflorescences and flowers glabrous or sparsely puberulous; leaves (narrowly) elliptic or obov: ate, ы chartaceous, reticulation immersed or raised —.— ). atirrensis Flowers unisexua : — ИРЕЙ 20 Flowers perfect o — 000000 - 30 Lower leaf surface with erect hairs, disce sable Bin.) Ka ce I RR RR 21 ower leaf surface glabrous or with appressed hairs, hairs not discernable to the touch — 23 Indument completely covering the surface of the twigs 0000000000000 22 Surface of the twigs partially visible between the hairs 51. О. mac — e slightly decurrent on the petiole; laminae and pe ticles. not sharply offset; tertiary Еи or e lower leaf surface scalariform and clearly raised; indument on young twigs s agey O. fendleri В. ase of laminae obtuse, clearly differentiated from the petiole; tertiary venation on lower * — carcely raised and not scalariform; indument on young twigs very short, « pe lica Twigs sharply angular; upper leaf surface with raised, minute reticulation 7 a — ** ‘Twigs rounded or slightly angular; upper r leaf surface without raised reticulation . or vilis coarse retic lation — - Volume 89, Number 3 van der Werff 433 2002 Synopsis of Ocotea 24(23). F 25(24). ES 28(27). 28". 29(28). 29'. 30(19). 30' 31(30). 31 32(31). 36(: 35). 36 37 (ЗО). A. 38(37). 39(38). 39'. 40(35). 43(42). 43'. Hi 10). A x 14). 45 э 15). 25. 26(25). Pistils pubescent; cupules plate- like, with a double margin |... = . О. — Pistils glabrous; cupules various, but not plate-like and with a double r margin .- Pit domatia present in the axils of the lateral veins of some leaves, the domatia not pubese e d PO TD FRE . О. oblonga ри domua absent, but axillary tufts of hairs sometimes present 26 Axillary tufts of hairs present on the lower leaf surface; tepals persistent on the cupule; oil glan ds readily visible on the upper leaf surface as black dots 0 80. O. rubrinervis Axillary tufts of hairs lacking: te pals deciduous or persistent in fruit; oil glands not or scarcely visible Twigs densely and minutely gray appressed pubescent, the hairs small and scare ely ااا‎ visible; indument on the inflorescences and pedicels notably denser than on the flowers ). — Twigs sparsely pubescent, individual hairs visible, or glabrous: inflorescences and — UE 0 density of indument on flowers and inflorescences about the same PARA AREA ee ER 7. О. cernua 29 Piia of Ihe. outer 6 stamens free; inflorescences ети idus ‘rs sparse els ae ent ¢ Leaves 7-25 cm long, the tertiary venation slightly raised on the upper surface; lateral veins — on the upper ине же Р ОЕТИРНОРИ ЕН 7: ). рибе rula Leaves 4-11 cm long, the tertiary venation imme sed on the upper surface; lateral veins гань im- pressed on * upper surface |... ). ). adela Lower leaf surfaces and young twigs 5 with pre 'edominantly erect hairs, these discernable to the touc ^s 31 Lower leaf surfaces and young twigs glabrous or with predominantly appressed hairs, rarely with sparse erect hairs, but these not discernable to the peek Leaf bases decurrent on the petiole and usually inrolled 0 Leaf bases not decurrent on the petiole, usually flat 35 Young twigs appressed pubescent, the indument moderate or sparse and part of the surface alway visible; inner surface of tepals glabrous or nearly so 00 25. О. — Young twigs with erect or ascending hairs, the surface (nearly) completely covered; inner — 'e of tepals pubescent ишсен —— З. Leaf bases al recurved: lateral veins 9 to 12 pairs 0 86. O. stenoneura Leaf bases scarcely if at all recurved or inrolled: lateral veins 4 to 8 pair 7 34. Indument on lower leaf surface predominantly erect; leaves 10-20 cm E А 37. O. hartshorniana Indument on lower leaf surface predominantly appressed; le saves 5-12 em long —_ 57. O. montei erde nsis Outer surface of tepals densely pubescent, the surface completely — by the indument if inne tepals with less indument on the upper half, then at least outer 3 tepals densely pubescent 30 Outer surface of all tepals glabrous or variously pubescent; if variously pubescent, de surface of the Flowers 10—14 mm diam и 33. О. gomezil Flowers less than 8 mm diam 37 Inflorescences few-flowered, racemose or with some lateral Grane hes ending in a cyme; le aves сопасе‹ cC AI AAA . pse opal Inflorescences many-flowered, the lateral branches several times divided: le saves coriaceous or chart: . — — 38 — Е 7 — 3. O. amplifolia Petioles up to 15 mm Іор 39 Petioles ca. 30 mm long g — Leaves coriaceous, densely ferruginous pubescent on the lower surfac е; ‘cupule with entire margin: leaf apex obtuse or very shortly Een Tot LIS A RE A oe Pine EE 23. О. darcyt Leaves chartaceous, moderately to sparsely golden brown abuso ent on the lower surface; tepals | 56. О. mollijalia sisting on the cupule; leaf apex acuminate Inflorescences racemose or rarely with a few lateral cymes os Inflorescences panic "ulate-cymose 05 — 44 Midvein, lateral veins, and tertiary venation clearly impressed (leaves rugose- bullate) — 22. О. corrugata Venation immersed or raised, not impressed (leaves not rugose-bullate) —— — 42 Outer surface of tepals glabrous: anthers sessile 0000000000 sl 34. 0. gordonii Outer surface of tepals (sparsely) pubescent: stamens with filaments Y or more of the bli of the STI AA ыы RE EDT RI PRACT VU — Hairs on the lower surface of leaves asce nding. covering most of the lamina; indument PLAY eee авуз eee 55. O. mollicella isis on iile bus: en] suis e erect, most of the lamina eines indui brown or ү дитин ge арн A ھی ت ت ی‎ A o ТАК O DUPDUTGO Leaves clustere d. me OM cat ETE O 45 Leaves alternate, evenly distributed along Ше twigs — ee AN ee a n — 7 Outer surface of the tepals glabrot — — = СТО ТЕНЧ 14. 0. bourgeauviana Outer surface of the tepals генен TER NIE NER AEN _.. 46 Petioles to 6 mm long: leaves 15-25 X 5-6 c m; indument el twigs ye -llow-brown __ А 91. O. tonii Petioles at least 10 mm long; leaves 10-15 X 6-7 em; indument of twigs brown 20. О. congregata Annals — — Garden 47(44). 7! a. 4847). " 18). 50 49). 50 5 5 51 (48). 51”. 5530). 55'. 56(55) ә 60(59). 60”. 61(59). ol’. 62(61). 62’. 63(62). 03". 64(63). Flowers at anthesis 2-3 mm diam.; tepals erect to half-erect; inner surface of tepals and stame glabrous f ›2. ( ШАТ Flowers at anthesis 5-10 mm diam.; tepals spreading or half-erect; inner surfac e of tepals and/ stamens partially papillose or vani nt side Receptacle pubescent in 49 Receptacle glabrous inside — < - Ё 51 Leaves elliptic, to 15 em long : e 12. O. betazensis Leaves obovate, 20-40 cm long . a = 2 90 Indument completely covering young twigs and inflorescences 95. O. valerioides Surface of twigs and inflorescences visible between the indument _ — Е . lentü Outer surface of the tepals pubescent E : — | " 64, 0. = Outer surface of the tepals glabrous — сос. = Inner surface of the tepals pubescent; filaments distinct, c ca. Y the length of the anthers — E | Inner surface of the tepals glabrous; filaments not evident = Е m 54 Leaves to LO cm long, the tips obtuse, acute or shortly acuminate O. praetermissa Leaves 10-20 cm long. acuminate or gradually narrowed into a slender tip . 4 bourgeauviana Surface of young twigs completely covered by the indument; cupule cup- shaped 94. O. valeriana Surface of young twigs partially visible between the indument; cupule shallowly howl-shaped or plate- like 38. helicterifolia Leaf bases inrolled and/or decurrent on the petiole or reflexe d, or leaves subsessile with the — weakly recurved; free petiole usually absent or very short — m - 56 Leaf bases not inrolled or decurrent on the petiole; petioles usually distinct 65 Flowers densely reddish pubescent; terminal buds glabrous or fine ly appresse d pubesc ent distally — 90. O. tonduzti Flowers gray pubescent or Кини te mina buds uniformly pubescent or "nd glabrous — 57 Leaves usually 30-50 em long, the apex usually rounded; domatia lacki E rivularis qi usually less than 25 em long, apices obtuse or rounded; nee ie ‘king or — eaf bases reflexed, the lobes frequently covering the midrib; flowers glabrous ... — 27. О. н жа ын Leaf зеби decurrent or inrolled, but not reflexed; flowers usually puberulous 59 Flowers 4-5 mm diam E - . - — 60 Flowers 2.5-3 mm diam 61 Outer surface of the tepals densely whitish pubescent, the surface complete ly cove — vernation lines visible on the lower leaf surface; receptacle (deeply) — Costa Rica, Pan v : al aucosericea Outer surface of the te de moderately pubescent, the — е pab visible; vernation — not visible 8. ( rec — ‘le bowl-shaped; Mexico, Guatemala О. chiapensis Leaves 5-10 em wide, obovate, glabrous or with a few appressed hi hairs on the lower leaf surfac e; tertia venation not or scarcely raised on upper leaf surface ). insularis Leaves to 5 em wide, elliptic or oblong, appressed pubesc ent or ‘glabrous | on the lower leaf surface tertiary venation imme sed or raised Leaf base cuneate or acute, moderate Чу or weakly reflexed, not decurrent on the petiole; leaves sub- sessile — 42. O. insularis s.l. Leaf base decurrent on the petiole, usually inrolled; leaves petiolat ite 63 Tertiary venation raised on the upper surface: leaf tips obtuse; leaf bases inrolled and decurrent vie austinit Tertiary venation not or rare ly raise d on the upper — if raised, leaf tips acute or acuminate and/ or leaf bases decurrent, but not inrollec Apparent petiole due to decurrent leaf bases, 4—5 cm long; flowers sparsely — 70. 0. produci Apparent petiole to 3 em long, usually shorter; indument of flowers various 10 whitet Tepals papillose or partially papillose on the inner surface; te pals glabrous. or sparsely — nt outside; tepals spreading at anthesis; tepals usually longer than З mn 0 Tepals glabrous or pubescent on the inner surface; outer surface — or variously pubescent: tepals usually erect or half-erect at anthesis; — usually less than 3 mm long А 9 Hairs on the lower leaf surface erect (check along midrib and lateral ve ins) Е Е 67 Hairs on lower leaf surface appressed or lacking Inflorescences racemose; young twigs densely pubescent with minute, erect hairs 67. О. pittieri Inflorescences panic ulate-c 'ymose; young twigs moderately to densely appressed pubese ent ›. О. rhytidoirie ч Stamens with lum nis at least as long as the ене outer stamens с used inward 5. O. pausiaca Stamens with filaments less than half as long as the anthers; outer stamens straight Loss. 6 Pit domatia present in the axils of the lowermost late ice veins — 4. O. arcuata Pit domatia lacking, axillary tufts of hairs sometimes present — 70 Inflorescences densely and minute dy brownish pubescent, the surface е cove етей « or nearly so .. с. di ose ‘ences sparsely or moderate ‘ly pubescent, the surface largely visib » 72 Flowers 6-7 mm diam.; widespread, from southern Mexico to Panama | 79. O. rubriflora Flowers 10-12 mm diam.; Osa Peninsula in Costa Rica EA е | 50. О. macrantha Leaf margin at the base (narrowly) reflexed or inrolled - . А 3 — 68. O. platyphylla Volume 89, Number 3 van der Werff 435 2 002 Synopsis of Ocotea -l 2 N — гь а І لح لد لد لد لد لد لد له 1 b de ЕЧ‏ 3 79. 80(79). 80”. аш in 82 А 2" , — — А p 84”. 8584). 85' 86(83). 80". 91(86). 87'. 88(87). 88' 8919). 89”. 4 89). (90). 92(89). 92’. 93(92). 93'. Leaf margin at the base flat or nearly so ES w Inflorescences paniculate, with at least a few e flowered lateral cymes, rarely racemose but then with flowers 6.5 mm or more in diameter and leaves drying greenish 00 0 0 — 1 Inflorescences racemose; leaves drying dark green or flowers * io 5 mm X — 77 lama nt on twigs brown and + istent; indument on peduncle moderately dense 15 Indument on twigs pale and becoming quickly sparse with age; indument on pe * le rather sparse ateral veins 6 to 9; Guatemala and С iuis — í — 52. 0. — Lateral veins 4 to 5; Costa AA 102. Ocotea sp. A Flowers 4.5—5.5 mm diam.; lowlands — — 9. О. жоло Flowers 6.5-9 mm diam.; montane forests, — above 900 m o 40. O. heydeana Indume ‘nt on young twigs dense and consisting of short, erect hairs 00 _ 67. O. pittieri — on young twigs of variable density and consisting of appressed. peu somewhat ascending ha Rec ae le pee pubes ent inside: Costa Rica Rc PS 15. O. brenesú Receptacle sparsely or very sparsely pubescent inside; Mexico, Cuán ala _ 1. O. acuminatissima Lower leaf surface with pit domatia, cavities with a small slit-like or rounded же е, in the axils of the basal lateral veins or along the lower lateral veins; orifice — glabrous, rarely with a fringe of hairs A A sss 80 ere leaf surface without pit denda: sulla tufts of PONE in a , small depression sometimes present A E A E E E aes 89 Domatia at least s once, |, often — times, their diameter away е from. the midrib; three or more р domatia present in each leaf; domatia with a fringe of hairs cc 10. Vh Domatia adjac ent to or less than their diameter away from the midrib; domatia present in the pea several pairs of lateral veins or only in the axil of the basal pair; domatia usually glabrous 8l Flowers sparsely appressed pubescent: inflorescences moderately appressed pubescent |. 36. О. педаг Flowers glabrous; inflorescences glabrous or nearly so s 2 Leaves tripliveined; domatia only present in the b 'rmost pair of lateral veins 96. О. — Leaves pinnately veined; domatia often present in the axils of more than one pair rof lateral veins or along the lateral VES a AAA ا‎ v UE нд» 83 Terminal buds glabrous — ——— 84 Terminal buds pubescent ici. retratos | - 86 Leaves broadly elliptic, 17-27 X 8-15 BS GI. ee НИЯНЫН — ‚ 29. O. euvenosa Leaves elliptic, A MM 85 Tepals е at anthesis; leaves. n n des, vellowish green, the validi often ома in aes than the laminae ne 100. O. viridiflora Tepals erect or half. erect al anthesis: [en s, when dry, dark $ green to o Bhi dus the midrib cone ue Is with the laminae o — 89. O. tenera Leaf apices obtuse to нн leaves coriaceous: doehas conspicuous, restricted to the — of the — O. pullifolia apices acute; — ee teous to coriaceous: T, coriaceous, domatia A imd found e the lowermost pairs of lateral vei af à lowermost lateral veins |... Tertiary venation on мун le af "en e immersed or nearly so; hi ‘aves coriaceous to — каа eous, dark green wher = we о O RA A e : 5. Jorge-esc barii Tertiary venation on vd leer — e raise d: Je aves chartaceous, bled k or velldiial green when dry DONDE CON RR IM = u 88 Inflorescences to 20 cm long; mm "s hac k when dry, the ите! very vais E. 11. O. bernoulliana Inflorescences to 10 em long: leaves yellowish green when dry, the main veins often lighter in color than thé amina sees unt caen adea EE . 54. О. meziana Terminal buds deci тере times a few hairs present along the margins of the bracts das the s) TS ER E 90 Terminal buds pubescent лыдан с чайнады села л E рене асс оди 92 Flowers and inflorescences alata »us — 98. О. verapazensis Flowers and distal parts of inflorescences sparsely or moderately, minutely ole scent 91 Leaves 15-35 X 7-15 em, leaf base rounded, rarely obtuse or subcordate: tepals erect or ne ee E anthesis 22000000 85. 0. — Leaves 10-15 X 4—5 cm, leaf ‘hase acute; tepals spreading at anthesi 53. O. matudat Twigs densely pom ent, the surface — tely covered by the erect or ascending hairs el و‎ cle ee cs ا ا چ‎ 21. 0). contrerasti Twigs g — pen ми ent or, if densely pubescent, the hairs appressed 93 Leaves, when dry, yellow-green, rarely darker green: tertiary venation forming a raised retic iln a on vba lower leaf surface: midrib and lateral veins usually lighter in color than the —— leaf tis 94 Leaves, when dry, dark green: tertiary venation not forming a raised reticulum on the lower leaf seg major veins concolorous with or darker than the leaf tissue B ER . 95 436 Annals la — Garden 94(93). 94 95(93). 95'. id 5). — 97'. 98(97). 98". 99(98). 99”. 100(96). 100". 101(100). 101”. 102(101). 102' 103(102). 103’. 104(103). 105100) — 104). 106’. l. Ocotea acuminatissima (Lundell) Rohwer, Bot. Jahrb. Syst. 112: 379. 1991. Phoebe ac- uminatissima Lundell, Contr. Univ. Michigan Distribution. Herb. 6: 19. 1941. Cinnamomum acutatum from 900 to 2700 m. Twigs and leaves glabrous or with a few appressed hairs 47. O. laetevirens Twigs and major veins on the lower leaf surface ж er pubescent with very short, erect or asc alk y hairs; indument on the lower leaf surface similar, but sparser |. ). rovirosae Tepals to 1 mm long and inflorescences moderately to sparsely pubesc ent, the hairs asc scending e -] zl л - ). mu dom Tepals at least 1.5 mm long; if rarely shorter (1.3-1.5 mm), then inflorescences very sparse ly pubescent with appressed hairs Inflorescences densely pubescent, the: surfac e large ly: or completely covere ad by the indúument - 97 Inflorescences sparsely or moderately pubescent, the surface largely or entirely visi 100 Leaf apices rounded or obtuse; axillary tufts of hairs lac кш on the lower leaf surface = Leaf apices acute or acuminate; axillary tufts of айа present on the lower ЗГА ас eaves obovate, to 4 em wide, drying green; twigs, when — alate or — ne Jover leaf surface sparsely pubescent, the hairs short, erect and inconspicuous .€ Leaves elliptic or oblanceolate; when widest above the n lower leaf surface sparsely appressed pubescent or glabre eaves to 2.5 em wide, oblanceolate or narrowly obovate- valine. drying blac К; — ca. 15, mm long 9. O. chry — 98 eucuneata — ane blac k; twigs — or angular; О. кш — 2.5-8 cm wide, elliptic c or narrowly у elliptic --obovate, drying green to dark green; ра! a. 2 nm lon 93. О. — Young SO densely pubescent, the hairs ere ct; infloresce nces 15-20 cm long; domatia lacking a ). subalata Young twigs appressed pubescent or — ‘glabrous: if with some erect Tes беле — short and their orientation scarcely visible; inflorescences usually shorter than 15 em; domatia, as axilla ary tufts of hairs or shallow pits, often present 101 Leaves lanceolate; domatia consisting of shallow pits, these with a pubesc ent or sometimes glabrous margin ). effi pes Leaves elliptic, ovate or elliptic-obovate: domatia, if | present, consisting of axillary tufts of hairs 02 Leaves 13-25 X 4—8 cm, the tertiary venation raised on the lower surface 35. КОТ — generally smaller than 12 X 5 em; if occ asionally to 15 cm long, the tertiary ve nation immerse n the lower surface 103 — elliptic to oblong, the apices rounded or blunt; domatia consisting of tufts of hairs, : — the own diameter away from the midrib herbert Leaves ovate, obovate or e lliptic, the apices acute or acuminate; domatia, when present, in [iod Wr. the lateral veins close to the midrib - — Leaves ovate, distally tapering into a long and slender apex — re eaves broadly or narrowly elliptic, without a long, slender apex = Domatia present as axillary tufts of hairs; outer surface of tepals glabrous Domatia absent; outer — е — tepals sparsely pubescent Flowers glabrous or nearly without magnification 63. 0. nu ‚ O. strigosa ; leaves › drying dark green; domatia (as axillary tufts sol hairs visible 75. O. truncata Flowers pubescent, the indument covering the tepals almost completely; leaves drying green to green; domatia, when present, not visible without magnification ^^ 8 dar ‚ O. racemiflora pressed indument on the lower leaf surface (or leaves almost glabrous). Mexico (Chiapas), Guatemala, Kostermans, Reinwardtia 6: 20. 1961. nom. nov. for Phoebe acuminatissima Lundell. . — Өр IRA: 2. Ocotea adela van der Werff. Novon 11: 501. PYPE: Mexico. Chiapas: Matuda 2107 (holo- mo 44 к 5 a . . )О1. TYPE: Panama. Prov. Panamá: Cerro type, MICH not seen; isotype, NY!). Jefe, Croat 13049 (holotype, MO!). Phoebe чиси лр, Lundell, Contr. Univ. Michigan : : 14 2. TY er 31 Phoebe — Lundell, Wrightia 1: 151. 194. E namomun — — — ll) a Rein- wart A variable species characterized by its racemose ). YPE: Mexico. Chiapas: Matuda An inconspicuous species best recognized by its sed p not seen; isotype, | : unisexual flowers, leaves with slightly impressed — zu lateral veins, and shallow cupules with persistent tepals. It can be confused with the widespread Oco- dtia 6: 23. 1961. TYPE: Mexico. Chiapas: Ma- hos (News): Mes. Ми iut Sheela haran tuda 5140 tee, i not seen). бе сетш VANEGE ез, DML tnar эрестев паз orten nodding flowers, immersed (not impressed) lateral veins, and does not have persistent tepals on the inflorescences, papillose inner surface of the tepals, — cupule. spreading tepals at anthesis, and the sparse ap- Distribution. Panama, from 500 to 900 m. Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea 3. Ocotea amplifolia (Mez & Donnell Smith) van der Werff, Novon 11: 510. 2001. Phoebe am- plifolia Mez € Donnell Smith, Bot. Gaz. (Crawfordsville) 19: 261. 1894. Cinnamomum amplifolium (Mez & Donnell Smith) Koster- Reinwardtia 6: 20. 1961. TYPE: temala. Depto. Quiché: Heyde & Lux (syntype, GH!). Gua- 3035 mars, Only known from the type collection. The long petioles (3 em long) and the dense indument on the twigs, inflorescences, and flowers are distinctive. Distribution. Guatemala, 3000 m. 4. Ocotea arcuata Rohwer, Bot. Jahrb. Syst. 112: 380. 1991. TYPE: Panama. El Road, Mori et al. 6883 (holotype, Llano-Carti MO! A rarely collected species, best recognized by its flowers with spreading tepals with papillose inner surfaces, almost glabrous leaves with pit domatia in the axils of the lowermost lateral veins, and the loop-connected lateral veins. Distribution. | Panama, from 300 to 550 m. э. Ocotea atirrensis Mez & Donnell Smith, Bot. Jahrb. Syst. 30, Beibl. 67: 18. 1901. TYPE: Costa Rica. Donnell-Smith 4930 (syntype. US”. Ocotea — C. К. Allen, J. Arnold Arbor. 26: 345 19 TYPE: Pam. Coe lé: Allen 1211 (holotype, F not seen; isotype Ocotea ке, Mes [o — Syst. 30, Beibl. 67: 19. 190 : Costa Rica. Talamanca, Pittier & Tonduz, in * inst. ge езйн Costarica 9172 9179 (syntypes, !). Ocotea pentagona Mes, o — Syst. 30, Beibl. 67: 17. 1901. TYP a Rica. Rio — Tonduz in herb inst. * -geogr. Passus 7613 (syntype. BR!); prope San Mateo, Biolley 7106 шо BR!): without locality, Tonduz 8 = Pittier & Durand 3362 (syntype, BR!). Ocotea wedeliana €. K. Allen, J. Arnold Arbor. 26: 339. 1945. TYPE: Panama. Boc 'as del Toro: Cooper 539 (holotype, F not seen; isotype, GH!). A broad concept of this species is accepted here: distinctive are the hollow twigs, stamens without a sterile tip. and small cupules. For a different view- point, see van der Werff (1988a). Only a few Ocotea species in Mesoamerica have hollow twigs: Ocotea Jefensis van der Werff differs in its tomentellous flowers and leaves with decurrent, while O. dendrodaphne Mez has tongue-shaped sta- inrolled bases, mens with a sterile tip and a double-margined cu- pule. Distribution. ama, from 50 to 1200 m. Nicaragua, Costa Rica, and Pan- Novon 11: Za- 6. Ocotea atlantica van der Werff. 504. 2001. TYPE: Honduras. Atlántida: mora 1744 (holotype, MO!). Readily recognized by its small, unisexual flow- ers and the erect pubescence on the lower leaf sur- face and twigs; the indument obscures the surface of young twigs. Distribution. Rica in forest on the Atlantic slopes from 50 to 100 m. Honduras, Nicaragua, and Costa 7. Ocotea aurantiodora (Ruiz € Pavón) Mez. 295. 1889. Laurus aurantiodora Ruiz € Pavón, Fl. Peruv. 1. 1. 349. 1804. TYPE: Ruiz & Pavon n. (B not seen). — Jahrb. Kónigl. Bot. Gart. Berlin 5: Peru. Distinctive are the clearly angled twigs, the small, raised reticulation formed by the tertiary ve- nation on the upper leaf surface, and the unisexual Ocotea longifolia HBK, used for South American specimens, is probably a flowers. a name widely synonym. Distribution. much of South America, from 0 to 600 m. Nicaragua, Costa Rica, Panama; 8. Ocotea austinii C. K. Allen, J. Arnold Arbor. 26: 350. 1945. TYPE: Costa Rica. lustin Smith A 125 (holotype, F not seen; iso- type, MO)). Zarcero, Ocotea irazuensis Lundell, Wrightia 5: 339. 1977. TYP Costa Rica. Volcán Irazu, Proctor 32355 озы LL not seen; isotype, MO!). Best recognized by the raised reticulation on the upper leaf surface, the inrolled leaf bases, which are decurrent on the petioles, and the coriaceous, frequently oblong leaves. Distribution. Costa Rica and Panama, from 2000 to 3000 m. 9. Oc otea bajapazensis Lundell, Wrightia 6: 8. 1978. TYPE: Guatemala. Lundell & Contreras — (holotype, LL MO). — not seen; isotypes, F!, This species can be recognized by its sparsely appressed pubescent twigs, paniculate-cymose in- florescences, and small flowers (ca. 5 mm diam.). It is similar to O. heydeana, which has larger flow- ers (6.5-9 mm diam.) and occurs at higher alti- tudes. Distribution. Guatemala, from 100 to 300 m. Annals of the Missouri Botanical Garden 10. Ocotea barbatula Lundell, Wrightia 5: 336. 1977. TYPE: Guatemala. Baja Verapaz: Lun- dell & Contreras 19444 (holotype, LL not seen: isotype, MO!). This species is vegetatively similar to O. me- ziana, but differs in having pit domatia with a pu- bescent margin and well away from the midrib. More collections may well show that these differ- ences do not hold and that O. barbatula is better treated as a synonym of O. meziana. Distribution. Guatemala. 11. Ocotea bernoulliana Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 275. TYPE: Gua- temala. Mujulia, Bernoulli & Cario 2590 (syn- type, U!). Best recognized by the following combination of characters: twigs, leaves, terminal buds, and flowers glabrous, leaves with conspicuous pit domatia in the axils of lateral veins and smaller pit domatia along the lateral veins, and cupules with 6 promi- nent longitudinal ridges. This species is rarely col- lected. Distribution. 1000 to 1600 m. Chiapas and Guatemala, from 12. Ocotea betazensis (Mez) van der Werff, No- von 9: 572 39. Phoebe betazensis Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 192. 1889. TYPE: Mexico. Oaxaca: Liebmann 2, 3, 22, 23 (syntypes, C!), Galeotti 2885 (syntype, not seen), Jürgensen 575 (syntype, not seen). Ore — mexicana Meissner var. diminuta Meissner, Prodr. 15(1): 118. 1864. Ocotea mexicana — r) Hemsley var. diminuta к Не m- sley, Biol. Centr. Amer., Bot 1 Mexico. Jürgensen 575 (not see ш. g Characteristic are the rather broad, elliptic leaves, relatively long (12-23 mm) ments half as long as the anthers, and the densely petioles, fila- tomentellous or tomentose twigs. This species had been included in O. helicterifolia, but that species has frequently obovate leaves, shorter (4—15 mm) petioles, sparse indument on the twigs, and nearly sessile anthers. Distribution. Oaxaca, to be expected in Chia- pas, 2000—2600 m. 13. Ocotea botrantha Rohwer, Bot. Jahrb. Syst. 112(3): 375. 1991, non Ocotea matudai Lun- de 2 Persea matudai Lundell, Lloydia 4: 49. 1941. TYPE: Mexico. Matuda 1880 (holotype, E not seen; — MO». Easily recognized by its racemose inflorescences, arge flowers with tongue-shaped stamens, and densely pubescent twigs. It is closely related to O. sinuata, which differs in its paniculate-cymose in- florescences and densely pubescent pistils. Distribution. Chiapas, Guatemala, El Salvador, from 800 to 2400 m. 14. Ocotea bourgeauviana (Mez) van der Werff, Novon 9: 574. 1999. Phoebe bourgeauviana Mez, ps Kónigl. Bot. Gart. Berlin 5: 194. 1889. TYPE: Mexico. Bourgeau 2234 Manos MO!). Veracruz: — Wrightia 5: 34. Nectandra — Lundell, )74. Contreras 11186 (holo- » PE: Guatemala. Izabal: е, LL? 101 seen; isotype, MO!). Phot aan Schultes, Bot. Mus. Leafl. 9: 170. лппатотит chinantecorum (Se :hultes) Kos- — Reinwardtia 6: 20. 1961. TYPE: Mexico. Oaxaca: Schultes & Reko 827 (holotype, GH!). The combination of clustered leaves and tepals that are glabrous on the outside and pubescent on the inside is diagnostic. It can resemble O. helic- erifolia, but that species has the inner surface of the tepals glabrous. The two collections from Gua- temala and Honduras have the interior of the re- ceptacle pubescent, while the Mexican specimens have a glabrous interior of the receptacle. More col- lections are needed to determine if the specimens from Guatemala and Honduras are properly placed in О, bourgeauviana. Distribution. Mexico (Veracruz, Oaxaca, Chia- pas). Belize, Guatemala, Honduras, from 100 to 1200 m. 15. — brenesii Standley, Publ. Field Mus. at. . Bot. Ser. 18: 454. 1 . Nectandra dedi "Gtandley) € C. K. Allen, 1. Arnold Ar- bor. 26: 370. 1945. TYPE: Costa Rica. Brenes 13653 (holotype, F!). The racemose inflorescences and appressed in- dument on the young twigs characterize O. brenesit. It is similar to O. pittieri (Mez) van der Werff, but the latter species has a short, erect indument on the twigs and domatia in the form of axillary tufts of hairs. Distribution, 2000 m. Costa Rica, Panama, from 700 to 16. Ocotea calophylla Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 298. 1889. Pleurothyrium ve- lutinum Meissner, DC. Prodr. 15(1): 170. 1864. TYPE: Colombia. Jervise s.n. (holotype, K not seen). Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea 439 ж — Standley & Steyermark, Ceiba 1: 237 TYPE: Costa Rica. Cartago: Leon 2166 (ho- 10!). F!; isotype, MC A striking species easily recognized by its dense- ly pubescent leaves with a revolute base, the ver- nation lines on the lower surface, and its upper montane habitat. Distribution. from 2600 to 3000 m. ica, 17. Ocotea cernua (Nees) Mez, Jahrb. Königl. 1889. ( „о пе 36. TYPE Bot. Gart. Berlin 5: 377. cernua Nees, Syst. Laurin, 424. Martinique. Sieber 106 (syntype, MO!). A common and widely distributed species with nearly glabrous twigs, leaves, inflorescences, and flowers; nodding or reflexed. The cupules are deeply cup- shaped: veined. Distribution. Brazil, from O to 700 m. 18. Ocotea chiapensis (Lundell) Standley & Stevermark. Publ. ser. 23: dell. Contr. Michigan Herb. 6: 12. 1941. TYPE: Mexico. Chiapas: Matuda 22 (holo- type, MICH not seen; isotype, MO!). Field Mus. Nat. Hist.. Bot. Univ. Distinctive are the leaves with decurrent, in- rolled bases, domatia as axillary tufts of hairs, and the relatively large (5-6 mm diam.) flowers. It is similar to O. glaucosericea Rohwer from Costa Rica and Panama, but the latter species has a denser indument on leaves and flowers and usually lacks domatia. Distribution. Mexico (Guerrero, Oaxaca, Chia- pas). Guatemala, from 1000 to 2800 m. 19. Ocotea chrysobalanoides (Lundell) Lun- dell. Wrightia 5: 35. Persea. chrysoba- lanoides Lundell, Wrightia 1: 151. 1946. TYPE: Mexico. Chiapas: Matuda 5582 (holo- type. LL not seen: isotype, МО!). Only known from the type collection. This spe- cies can be recognized by its leaves with obtuse to rounded apices, the pubescent inner surface of the receptacle, large (0.8 mm long) staminodia, and the dense, gray indument on twigs and inflorescences. Distribution. Mexico (Chiapas). at 2000 m. 20. Ocotea congregata van der Werff, Novon 9: 514. 1999, TYPE: Mexico. Chiapas: Shilom Ton 8930 (holotype. MO!). Costa Rica. northern South Amer- the unisexual flowers are often somewhat leaves are frequently somewhat tripli- From Mexico south to Bolivia and . 1944. Nectandra chiapensis Lun- Can be recognized by its clustered leaves, the brown to dark brown indument on the twigs, and the rather long (10 mm or more) petioles. It can be confused with O. tonii (Lundell) van der Werff, which differs in its shorter (to 6 mm) petioles and yellowish brown indument on the young twigs. Distribution. Mexico (Chiapas). 900 to 1400 loose ly from 21. Ocotea contrerasii Lundell, Wrightia 5: 337. 1977. TYPE: Guatemala. Dept. Baja Ver- apaz: Lundell & Contreras 19588 (holotype, LL not seen: isotype, MO!). A rarely collected species related to O. meziana. from which it differs in the dense indument on the twigs, with erect or ascending hairs and spreading hairs along the major veins on the lower leaf sur- face. Distribution, Guatemala. 22. Ocotea eds i van der Werff. Novon 9: TY 574. 199 PE: Mexico. Oaxaca: Wendt et al. 6765 Prin] MEXU not seen; isotype, MO). Only known from the type collection and easily recognized by its rugose-bullate. pubescent leaves. lts racemose inflorescences are also a useful char- acter. Distribution. 1300 m Mexico (Oaxaca), from 1100 to 23. Ocotea darcyi van der Werff, Novon 11: 505. 2001. TYPE: Panama. Correa & Montenegro 10176 (holotype, PMA not seen; isotype, MO!). This species is closely related to O. stenoneura Mez & Pittier, which differs in its inrolled and de- current leaf bases, acute leaf apices, and elliptic leaf shape. It can also be confused with O. pseu- dopalmana Burger, but O. darcyi differs in having many-flowered inflorescences, a ferruginous indu- ment (not brown), and smaller flowers (4—4.5 vs. 6—8 mm diam.). mm Distribution. Panama, from 700 to 1000 m. 24. Ocotea eh es ras Mez, Jahrb. Kónigl. Bot. Gart Berlin 5: 238. macrophylla Beurl., ife 1854: 145. 1856, non Ocotea macro- phylla HBK. TYPE: Panama. Billberg s.n. (ho- lotype, 5 not seen). 389. Dendrodaphne Vetensk. Acad. Handl. Ocotea ovandensis Lundell, Contr. Univ. Michigan Herb. 6: 16. 1941. TYPE: Mexico. Chiapas: Matuda 444 (holotype, MICH not seen). 440 Annals of the Missouri Botanical Garden Ocotea quisara Mez & Donnell Smith, Bot. Gaz. (Craw- fordsville) 33: 259, 1902. ТҮРЕ: Costa Rica. J. Don- nell Smith 6756 —— BM not seen), J. Donnell Smith 6751 (syntypes, В, K not seen), J. Donnell Smith 6753 (syntype, K not seen) A widespread species, easily recognized by its hollow twigs and tongue-shaped stamens with a sterile tip. The poorly known O. morae Gomez- Laurito shares these characters, but has leaves with an inrolled base (flat in O. еш and much larger fruits (5 X 3.5 em vs. 2.5 X ] cm in O. dendrodaphne). distribution. From Mexico to Panama, 50- 1300 m 25. Ocotea — van der Werff, Fieldiana, Bot., n.s. 23: 79. 1990. TYPE: Costa Rica. Go- mez & Herrera erie (holotype, MO”). Distinctive are the obovate-elliptic leaves with inrolled, decurrent bases and with a rather sparse, erect indument on the lower surface, while the twigs are appressed pubescent. Distribution. Atlantic slopes of Nicaragua, Cos- ta Rica, and Panama, from 100 to 1000 m. 26. Ocotea effusa (Meissner) Hemsley, Biol. Centr. Amer, Bot. 3: 73. 1882. Oreodaphne effusa Meissner, DC. Prodr. 15(1): 120. 1804. TYPE: Mexico. Prope S. Pedro Nolasco, Talea etc., Jürgensen 906 (syntype, BM!). Best recognized by its lanceolate leaves with shallow, often pubescent pit domatia. The large (0.8 mm) staminodia are also a useful character. Several collections from higher altitude (1700-2000 m) are provisionally placed here. Distribution. Southern Mexico, Belize, Guate- mala, from 200 to 2000 m. 27. Ocotea endresiana — Jahrb. Kónigl. Bot. Gart. Berlin 5: 257. 1889. TYPE: Costa Rica. Endres 223 — E, Characterized by its glabrous leaves, inflores- cences, and flowers, reflexed leaf bases, and cu- pules without persistent tepals. Collections from Costa Rica tend to have well-developed domatia, but these are less conspicuous or even lacking in specimens from Panama. Large-leafed specimens & Williams, but that species has puberulous flowers. resemble O. rivularis Standl. Distribution. Costa Rica, Panama, from 200 to 1500 m 28. Ocotea eucuneata Lundell, Contr. Univ. Michigan Herb. 6: 16. 1941. TYPE: Belize. Stann Creek Distr., Gentle 3068 (holotype, MICH not seen; isotype, M Only known from the type collection, Distin- guishing features are the obovate leaves with a sparse, erect indument on the lower surface, which is best seen along the major veins, and the sharply angled young twigs. Distribution. Belize, at 200 m. 29. Ocotea euvenosa Lundell, Wrightia 4: 157. 1971. Ocotea venosa Lundell, Phytologia 12: 245. 1965, 1931. TYPE: Gua- temala. Alta Verapaz: Contreras 4678 (holo- non Gleason, type, LL). Known to me only from the type collection. Di- agnostic are the large (to 27 X 15 cm) leaves, dry- ing nearly black, with pinnate venation and pit domatia in the axils of the basal lateral veins. Closely related are O. bernouilliana, which has nar- rower leaves and pit domatia not only in the axils of the basal lateral veins but also along the lateral veins, and O. vanderwerffii (Kostermans) van der Werff, which has smaller, tripliveined leaves. Distribution. | Guatemala. 30. Ocotea fendleri (Meissner) Rohwer, Mitt. Inst. Allg. Bot. Hamburg 20: 152. 1986. Gym- Prodr. 15(1): Fendler 2395 nobalanus fendleri Meissner, DC. 142. 1864. TYPE: Venezuela. (holotype. G-DC not seen). Known in Mesoamerica from only two collec- tions, a pistillate and a sterile one; among the spe- cies with unisexual flowers best recognized by the dense, shaggy pubescence on the twigs, the decur- rent leaf bases, and the erect indument on the lower leaf surface. Identifications of the two Mesoameri- can collections are provisional. French Guy- Distribution. Panama, Venezuela, ana, at 1000—1100 m. 31. Ocotea floribunda (Swartz) Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 325. 1889. Laurus floribunda Sw., Prodr. Veg. Ind. Occ., 65. 1788. TYPE: Jamaica. Swartz s.n. (holotype, S not seen). A widespread species, best recognized by its pu- bescent pistil or pistillode, the rather large (6—7 mm diam.) flowers, and the platelike cupule with a double margin. Distribution. From Nicaragua south to Brazil and Peru, 100-1400 m. Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea 32. Ocotea glaucosericea Rohwer, Mitt. Inst. Allg. Bot. Hamburg 20: 144. 1986. Nectandra hypoglauca Standl. ex С. К. Allen, J. Arnold Arbor. 26: 399. 1945, ees) Mez, 1889. TYPE: Panama. Chiriquí: Davidson 531 (holotype, F not seen: isotype. O!). non O. hypoglauca Similar to O. chiapensis: see that species for a discussion of the differences. Distribution. Costa Rica, Panama. from 1500 to 2000 m. 33. Ocotea ign Burger, Fieldiana, Bot., n.s.. 23: 81. . TYPE: Costa Rica. Gomez- Jed ito et al. pus (holotype, CR not seen: iso- type, MO). A very distinctive species with large (10—14 mm diam.) flowers, densely pubescent twigs. leaves, and flowers, and persistent tepals on the cupule. Distribution. Costa Rica, from 800 to 1400 m. possibly Panama. 34. Ocotea gordonii van der Werff. Novon 9: 575. 1999, TYPE: Panama. Chiriquí: McPher- son 10421 (holotype, MO”). This species belongs to the O. helicterifolia group and is characterized by its racemose inflorescences, long (ca. 1 em) pedicels, and its pubescent inner surface of the receptacle. Most other species in this group have paniculate-cymose inflorescences. In Werff shares the racemose inflorescences, but that spe- Panama only O. purpurea (Mez) van der cies has smaller (5 mm vs. 8—10 mm diam.) flowers. Distribution. Panama, from 1000 to 1150 m. 35. Ocotea guatemalensis Lundell, Wrightia 5: 339. 1977. TYPE: Guatemala. Baja Verapaz: Lundell & Contreras 19754 (holotype. LL not seen; isotype, MO! — A rarely collected species with rather large (to 25 X 8 cm) leaves. raised reticulation on the lower leaf surface. The conspicuous domatia, and density of the pubescence on the inflorescences in- creases markedly from the base toward the flowers. A few collections from Costa Rica and Panama are provisionally included here; they differ in having smaller (ca. 3 mm vs. 4-5 mm diam.), less pubes- cent flowers. Distribution. Guatemala, possibly Costa Rica and Panama. 36. Ocotea haberi van der Werff, Novon 11: 505. 2001. TYPE: Costa Rica. Guanacaste: 4— 5 km NW of Monteverde, Haber, Guindon & Brenes 11093 (holotype, MO!). An inconspicuous species best recognized by its pubescent flowers, pit or pocket domatia, often with a narrow opening, and a pubescent inner surface of the receptacle. Most other species with pit dom- atia have glabrous flowers and a glabrous inner sur- face of the receptacle. from 800 to Distribution. Costa Rica, Panama. 1400 m 37. Ocotea hartshorniana Hammel, J. Arnold Arbor. 67: 128. 1986. TYPE: Costa Rica. He- redia: Hammel 11932 (holotype, DUKE not seen; isotype, MO!). Best recognized by the erect or ascending hairs on the lower leaf surface, the dense, reddish brown indument on twigs and inflorescences, and the de- current leaf bases. It is very similar to О. montev- which has 2 erdensis Burger. 2 predominantly ap- pressed indument on the lower leaf surface апа occurs at higher altitude. It can also be confused with O. stenoneura, but that species has inrolled, not flat leaf bases. Distribution. Costa Rica, Panama, from 100 to 300 m. 38. Ocotea helicterifolia (Meissner) Hemsley, Biol. Centr. Bot. 3: 73. 1882. Oreoda- phne helicterifolia Meissner, DC. Prodr. 15(1): 123. 1864. Phoebe helicterifolia (Meissner) Mez. Jahrb. Kónigl. Bot. Gart. Berlin 5: 193. 889. Cinnamomum helicterifolium (Meissner) Kostermans, Reinwardtia 6: 21. 1961. TYPE Mexico. Chiapas: Linden 1641 (syntype, K!). Amer., Nectandra corzoana Lundell, Wrightia 4: 102. 1969. Phoe- zoana (Lundell) Lundell, Wrightia 5: 342. 1977. ME corzoanum — Koster- mans, Reinwardtia 10: 422. TYPE: Mexico. Chiapas: Shilom Ton 3560 dine LL! Ocotea tenejapensis Lundell, ightia 4: 108. 1969. TYPE: Mexico. Chiapas: à Ton 779 (holotype. d x x I LED. Oreodaphne mexicana Meissner, DC. 1864 Prodr. 15( Oreodaphne mexicana var. sner, DC. Prodr. 15(1): 118. 1864. nom. supe var. mexicana. Ocotea mexicana — Hemsl, Biol. Centr. Amer., Bot. 3: 73. 1882. Ocotea mexi- cana var. cy (Me занае) E ux as C entr. Amer., Bot. 3: 73. 1882, nom. те rfl. - сапа. — Rd euet Mez, Jahrb. Кош. ‘Be Gart. Berlin 5: 194. 1889, non Phoebe mexicana Meissner. TYPE: Mexico. Galeotti 7004 (lect totype, — here, BR!), Jiirgensen 537 (syntype, not seen). le LES. subsessilis Meis- 3. Annals of the Missouri Botanical Garden m mexicana Meissner и longipes Meissner, DC. odr. 15(1): 118. 1864. Ocotea mexicana коне r) Hemsley var. — ome — er) em- sley, Biol. Centr. Amer., Bot. 3: 73. 1882. TYPE: Mexico. Cerca Orizaba, Botteri 1018 | — K not — м A dins umbrosa Nees bullata Meissner, DC. rodr. 1): 122. 1864. as umbrosa var. — (Nees) Hemsley, Biol. Centr. Amer., Bot. 3: 74. 1882 me E: Central America. Oersted 21 ы В de- royed). Phoe "s — Lundell, Contr. Univ. Michigan Herb. 6: 21. 1941. Cinnamomum obtusatum 1 m Kos- termans, Reinwardtia 6: 22. 1961. : Mexico. Chiapas: Matuda 1887 (holotype. s not seen; isotype, CAS!). A widespread and variable species, character- ized by the hirsute indument on twigs and leaves, paniculate-cymose inflorescences, glabrous flowers, alternate leaves, and a glabrous inner surface of the receptacle. Collections placed here from Costa Rica have obovate leaves and occur at lower altitudes; they might represent a different species. Ocotea te- nejapensis, only known from the type, 15 provision- ally included here. Distribution. From Mexico to Panama, at 1000— 1900 m (in Costa Rica at 50-600 m). Lundellia 1: 40. Wendt et al. isotype, 39. Ocotea heribertoi Wendt, 1998. TYPE: Mexico. Oaxaca: 6871 (holotype, MEXU not MO!) seen, Tall trees, known from only two collections and best recognized by their smooth leaves with an ob- tuse or rounded apex, presence of domatia at some distance from the midrib, and its large fruits (to 4.5 cm diam.). Distribution. Mexico (Oaxaca, to be expected in Chiapas), at 200—300 m. 40. Ocotea heydeana (Mez & Donnell Smith) Bernardi. Candollea 22: 93. 1967. Nectandra heydeana Mez & Donnell Smith, Bot. Gaz. (Crawfordsville) 19: 262. 1894. TYPE: Gua- temala. Heyde & Lux 4260 (syntype, MO!). Can be recognized by its rather large (6.5-9 mm diam.), glabrous flowers, paniculate-cymose inflo- rescences, and sparsely pubescent twigs. Similar to О. bajapazensis, which has smaller (4.5-5.5 mm diam.) flowers and occurs at lower altitudes, and to O. magnifolia (Lundell) Lundell, which has a dens- er indument on the twigs. Hondu- Distribution. Guatemala, El Salvador, ras, from 600 to 1900 m. 41. Ocotea holdridgeiana Burger, Fieldiana, Bot, n.s. 23: 83. 1990. TYPE: Costa Rica. Lent 1677 (holotype, F not seen; isotype, MO!). Easily recognized by the conspicuous pocket domatia in the axils of the basal lateral veins, the arge (10 mm diam.) flowers, and the stamens with a sterile tip. Leaves can be somewhat tripliveined. Distribution. Costa Rica, Panama, from 1600 to 2400 m. 42. Ocotea insularis (Meissner) Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 271. 1889. Phoebe insularis Meissner, DC. Prodr. 15(1): 33. 1864. TYPE: Costa Rica. Cocos Island, Menzies s.n. (holotype, K not seen: isotype, MO!). llen, J. pola Arbor. 26: 419. TYPE: ia White 225 (holotype, МО). п. ira Me & Pittier, Bull. Herb. 8: 232. . TYPE: Costa Rica. Tonduz in Herb. Inst. Phys. i Costar. 10415 (syntype, US!), idem 13399 (syntype, US!). Aiouea d C. K. 19 dér Boiss., As accepted here, a variable species with rather large leaves with decurrent bases, small (2.5-3.5 mm diam.) flowers with erect tepals, and tufts of hairs on the back of the inner stamens. Montane populations have often smaller, more coriaceous leaves. Domatia are often, but certainly not always. present. Distribution. Costa Rica, Panama, Colombia, and Ecuador, from 100 to 2000 m. 43. Ocotea iridescens Lorea-Hernandez & van der Werff, Brittonia 54: 2002 in press. TYPE: Mexico. Oaxaca: Salas. Among the Mexican species of Ocotea, this spe- cies is distinctive because of its dense, minute, and light-colored indument on the lower leaf surface, its acuminate leaves, and slender inflorescences. It is known from a few collections made near the bor- der of Oaxaca with Chiapas at 1400-1600 m. 44. Ocotea jefensis van der Werff, Novon 11: 506. 2001. TYPE: Panama. Prov. Panamá: Cerro Jefe, Carrasquilla 2123 (holotype, MO!). Diagnostic are the hollow twigs, obovate leaves, and small (4 mm diam.), tomentellous flowers. Its petioles are indistinct due to the decurrent leaf ba- ses. Distribution. Panama, from 200 to 900 m. 45. Ocotea jorge-escobarii Nelson, Ceiba 25: 173. 1984. TYPE: Honduras. Olancho: Nelson & Soto 8188 (holotype. TEFH not seen; iso- type, MO). Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea A montane species with dark green, stiff leaves, pit domatia along the basal lateral veins, and large. ribbed cupules. Closely related is O. pullifolia van der Werff. which differs in leaves with obtuse or rounded apices, and in having pit domatia restrict- ed to the axils of the basal lateral veins. Distribution. Honduras, Nicaragua, from 1000 to 1500 m. 46. Ocotea klepperae van der Werff. Novon 11: 508. 2001. TYPE: Costa Rica. Puntarenas: Hammel 22068 (holotype, INB not seen; iso- type, MO). The large flowers, tongue-shaped stamens with a sterile tip, and the position of the inflorescences in axils of bracts near the tips of the branches place this species in the subgenus Dendrodaphne. It dif- fers from the other Mesoamerican species of this subgenus in the dense indument on the twigs, in- florescences, and flowers. Distribution. Costa Rica (2 collections), from 10 to 300 m. 47. Ocotea laetevirens Standley & Steyermark, Field Mus. Publ. Bot. 23: 114. 1944. TYPE: Guatemala. Huehuetenango: Cerro Chiblac, Steyermark 49378 (holotype, F!). nov. DS». dias — Lundell, Wrightia 4: 133. 1970. Syn. : Mexico. Chiapas: Clarke 38 (holotype. As accepted here, a variable species that in- cludes all specimens of the O. meziana group with elabrous or sparsely appressed pubescent leaves, pubescent terminal buds, and without domatia or with domatia as axillary tufts of hairs. It is close to О. meziana, but that species has pit domatia on the lower leaf surface. Some collections from Costa Rica and Panama that are vegetatively indistin- guishable from O. laetevirens have 2-celled anthers: in spite of this, they are here included in O. lae- tevirens and not described as new species of Aiouea Aubl. Distribution. Mexico (Chiapas). Costa Rica, and Panama, from 100 to Guatemala, Honduras, 3000 m. 48. Ocotea lentii Burger, Fieldiana, Bot., n.s. 23: 86. 1990. TYPE: Costa Rica. Cartago: Lent /94 (holotype, F not seen; isotype, MO!). Distinctive are the large (18—40 cm), obovate leaves, sparsely pubescent leaves, receptacles pu- bescent on the inner surface and the large (5 X 2.5 valerioides ст) fruits. It can be confused with O. Burger, but that species has the twigs and inflores- cences completely covered by the indument, while in O. lentii part of the surface remains visible. The hairs of O. lentii are also coarser. Distribution. Costa Rica, from 700 to 1400 m. 49. Ocotea leucoxylon (Swartz) Lanessan, Pl. Util. Col. Franc. 156. 1886. Laurus leucoxylon Sw.. Prodr. Veg. Ind. Occ. 65. 1788. TYPE: Jamaica. Swartz s.n. (holotype, S not seen). Ocotea — Lundell, Wrightia 5: 54. 1974. TYPE: Guatemala. Izabal: Contreras 9924 (holotype. LL not is — MO!). — е — Publ. Field Mus. Nat. Hist Bot. Ser. 18: TYPE: Costa Rica. fient 6789 — Ма Diagnostic аге the small, unisexual flowers. leaves with a smooth upper surface, and the mi- nule, appressed, gray indument on the young twigs. The flowers are less densely pubescent than the pedicels. Cupules are shallow to platelike and often lenticellate. Distribution. From Mexico to Brazil and Peru, West Indies, from 200 to 1600 m. Ocotea macrantha van der Werff, Novon 11: 508. "РЕ: Costa Rica. Puntarenas: Canton de Osa, Aguilar 4688 (holotype. MO! 50. — Vegetatively very similar to O. rubriflora Mez: both have a short and dense indument on the twigs and However, O. arger flowers (10-12 mm vs. 6-7 mm diam.). Costa Rica, from 100 to 200 m. inflorescences. macrantha has Distribution. 51. Ocotea macropoda (HBK) Mez. Jahrb. Kón- igl. Bot. Gart. Berlin 5: 348. 1889. Persea ma- cropoda HBK, Nov. Gen. Sp. 2: 127. 1817. TYPE: Colombia. Humboldt & Bonpland s.n. holas P not seen). Mem. New York Bot. Gard. TYPE: Venezuela. Bernardi Ocotea babosa С. Allen, 9: 82. 1966. pis nov. 86 (holotype, NY! Best recognized by the unisexual flowers and erect indument on twigs, leaves, and flowers. In these characters O. macropoda resembles O. atlan- tica, but in the latter species the indument covers the surface of the twigs completely, while it remains partly visible on O. macropoda. Ocotea macropoda also has larger flowers than O. atlantica (5-6 mm vs. 3.5 mm diam.). I am not quite certain that the name O. macropoda is correctly applied here: the South American specimens (including the types of O. macropoda and О. babosa) come from much higher altitude (above 2000 m). 444 Annals of the Missouri Botanical Garden Distribution. Costa Rica, Panama, Colombia, pubescence, and the densely pubescent twigs. It is Venezuela, Ecuador, from 100 to 800 m. infrequently collected. Distribution. Costa Rica, from 1400 to 2300 m. 52. Ocotea magnifolia (Lundell) Lundell, Wrightia 5: 341. 1977. Nectandra magnifolia E: Gua- Lundell, Wrightia 4: 103. 1969. TYP temala. Alta Verapaz: Contreras 7865 (holo- type, LL not seen; isotype, MO"). Nectandra thornei Lundell, Wrightia 5: 335. 1977. Syn. TYPE: Mexico. Chiapas: Thorne & Lathrop 40526 (holotype, LL not seen; isotype, DS!). A poorly known and poorly defined species, best recognized by its indument, which is denser than in O. heydeana and sparser than in O. rubriflora Mez. Distribution. Mexico (Chiapas), Guatemala, from 100 to 200 m. 53. Ocotea matudai Lundell, Bull. Torrey Bot. e 69: 388. 1942. TYPE: Mexico. Chiapas: . Ovando, Matuda 4221 (holotype, MICH not seen; isotype, F!). A rarely collected species best recognized by its glabrous terminal buds and leaves; flowers have spreading tepals at anthesis. The distal parts of the inflorescences have a minute, predominantly erect indument, also a useful character. Distribution. Mexico (Chiapas), 1300-2700 m. 54. Ocotea УРИ К. Allen, J. Arnold Ar- bor. 26: 1945. TYPE: Costa Rica. Zar- cero, pines Smith H359 (holotype, F not seen; isotype, MO!). A rather common species, characterized by its greenish drying leaves with small pit domatia along the lateral veins, pubescent terminal buds, and gla- brous flowers. It is similar to O. viridiflora Lundell, which differs in its glabrous terminal buds, finely acute leaf apices, and smaller inflorescences (to 5 em in O, viridiflora, to 10 em in O. meziana). Also similar to O. laetevirens, but this species lacks pit domatia Distribution. Honduras, Nicaragua, Costa Rica, and Panama, from 100 to 1800 m. 55. Ocotea mollicella (Blake) van der Werff, Fieldiana, Bot., n.s. 23: 88. 1990. Phoebe mol- licella Blake, Contr. Gray Herb. 52: 64. 1917 Cinnamomum mollicellum (Blake) Koster- mans, Reinwardtia 6: 22. 1961. TYPE: Costa Rica. Copey, Tonduz 11676 (holotype, GH!). Readily recognized by its racemose inflorescenc- es, elliptic to narrowly elliptic leaves with soft, gray 56. Ocotea — Mez & а Bull. Herb. Boissi ser. 2, 3: 233. 19( TYPE: Costa Rica. om A Er " not seen). Best recognized by its densely pubescent twigs, pubescent, chartaceous leaves, and many-flowered inflorescences. See also under O. darcyi and O. pseudopalmana Burger. Costa Rica, Panama, 50 to istribution. from 1000 m. 57. Ocotea ЕТЕ Burger, Fieldiana, ot., n.s. 23: 89. 1990. TYPE: Costa Rica. Puntarenas: Pietas dd (holotype, CR not seen; isotype, MO!) Similar to O. hartshorniana; see that species for differences. Distribution. Costa Rica, from 800 to 1500 m. Novon 7: Costa Rica. Alajuela: Go- USJ not 58. Ocotea morae Gomez-Laurito, 145. 1997. TYPE: mez-Laurito & Mora 12817 (holotype, MO!). seen; Isotype, The tongue-shaped stamens and double-mar- gined cupule place this species in subgenus Den- drodaphne. lt can be confused with the widespread О. dendrodaphne, but differs Е its inrolled leaf base and large (5 X 3.5 cm) fru Distribution. Costa Rica, a 100 to 800 m. 59. Ocotea multiflora van der Werff, Novon 6: 481. 1996. TYPE: Costa Rica. Puntarenas: Reserva Forestal Golfo Dulce, Aguilar 791 (holotype, MO”). Tall trees, best recognized by its leaves with many (12 to 17) pairs of lateral veins, the many- flowered (2-2.5 mm diam.) flowers. It was included in Burger and van der Werff (1990) as Ocotea sp. B. Distribution. Only known from the Osa Penin- sula in Costa Rica, from 10 to 200 m. inflorescences, and small 60. Ocotea nigrita (Lundell) Lundell, Wrightia 5: 341. 1977. Nectandra nigrita Lundell, Wrightia 4: 132. 1970. TYPE: Guatemala. El Petén: Contreras 9465 (holotype, LL not seen; isotype, CAS!) An infrequently collected species, characterized by its oblanceolate, black-drying leaves with tufts Volume 89, Number 3 2 van der Werff Synopsis of Ocotea of hairs in the axils of the basal lateral veins and the small (3—4 mm diam.), densely puberulous flow- ers. This apana Wendt & van der Werff, which differs in its 5-2.5 em in O. s species can be confused with O. uxpan- longer and wider leaves (6-15 X 1. nigrita, 10-25 X 2.58 ст in 0. uxpanapana) and larger (3—4 mm vs. 4-6 mm diam.) flowers. Distribution. Guatemala, from 200 to 300 m. 61. Ocotea oblonga (Meissner) Mez, Jahrb. Kón- igl. Bot. Gart. Berlin 5: 367. 1889. Mespilo- daphne oblonga Meissner, DC. Prodr. 15(1): 107. 1864. TYPE: French Guyana, Sagot 491 (holotype, G-DC not seen). Phoebe mayana Lundell, Amer. Midl. Naturalist 20: 473. 1943. Ocotea mayana (Lundell) Pond П, Wrightia 2: 52. 1960. Reinwardtia 10: 446. 1988. — по! Cinnamomum der igi (Lundell) Kos- TYPE: Belize. seen: isotype, termans, Gentle 3187 (holotype, 10!). This species can be easily recognized by the combination of unisexual flowers and leaves with pit or slit domatia. The slender, appressed pubes- cent vegetative buds are also a good character. Distribution. From Mexico to Bolivia and Bra- zil. from O to 1000 m. 62. Ocotea oblongifolia van der Werff, Novon 11: 509. 2001. TYPE: Guatemala. Quezalte- nango: Finca St. John, ca. 5 km S of Sta. Maria de Jesus. Walker 442 dela US! — Diagnostic are the oblong, densely pubescent leaves with an obtuse to subcordate base and the small (2-3 mm diam.) flowers. The indument is similar to. what in the O. helicterifolia group, but the small flowers with erect to half-erect is found tepals are entirely different. Only known from the type collec- 1400-1800 m. Distribution. tion, Guatemala, 63. Ocotea parvula (Lundell) van der Werff. No- von 11: 510. 2001. Phoebe parvula Lundell, Wrightia 5: 343. 1977. Cinnamomum parvul- um (Lundell) Kostermans. Reinwardtia 10: 448. 1988. TYPE: Mexico. Chiapas: Ton 605 (holotype, LL). An infrequently collected species with ovate leaves, gradually tapering into the tip, domatia as axillary tufts of hairs, glabrous tepals, and small (to 5 cm long) inflorescences with frequently persisting bracts. See also O. strigosa van der Werff. Mexico (Chiapas, Oaxaca), Distribution. from 1800 to 2800 m. 64. — patula van der Werff, Novon 9: 577. : Costa Rica. Puntarenas: Aguilar et e. 2715 Delos: MO! Part of the O. helicterifolia group and similar to ). valeriana (Standl.) Burger, from which it differs in its densely tomentellous inflorescences (with the surface almost entirely or entirely covered), its shorter (2-3 mm vs. 6-8 mm) pedicels, and its pu- bescent flowers. Distribution. Known from two collections, Cos- ta Rica, 1000-1400 m. 65. Ocotea pausiaca Rohwer, Bot. Jahrb. Syst. 112: 387. 1991. TYPE: Panama. Colón: Knapp 5782 (holotype, MO! — Part of the O. heydeana group, characterized by rather large, glabrous or nearly glabrous flowers with spreading tepals, these with some papillae on the inner surface, and often dark-drying leaves. It is distinct within this group by its stamens with the filaments twice as long as the anthers and the outer six stamens curved inward. like O. rubriflora, which differs in the very dense, mi- nute indument on the young twigs. Distribution. Panama, from 900 to 1500 m. Vegetatively it is 66. Ocotea сог »rum Gomez-Lauri- . Novon 3: 31 1993. TYPE: Costa Rica. San — E et al. 12160 (holotype, CR not seen: isotype, МО!). A very distinctive species characterized by its stamens with a sterile tip and the densely tomen- tellous whitish lower leaf surfaces. It was included in Burger and van der Werff (1990) as “A species f uncertain. generic position” and in Rohwer (1991) as Ocotea sp. / Distribution. Costa Rica, Panama, from 1800 to 2200 m > 67. Ocotea pittieri Me z) van der Werff, Fieldi- Bot.. n.s. 23: 92, 1990. Phoebe pittieri . Bot. Jahrb. Syst. 7 Beibl. 67: 16. 1901. bum pittieri (Mez) Kostermans, Rein- wardtia 6: 23. 1961. TYPE: Costa Rica. Ala- juela: Tonduz 11893 (holotype, B not seen). Best recognized by its racemose inflorescences and the dense, short, erect indument on the young twigs. The indument on the lower leaf surface can e very sparse and difficult to see. Ocotea pittieri — is very similar to O. brenesii, which has appressed hairs on the twigs. The name 0. pittieri was applied in Burger and van der Werff (1990) to the species here treated as O. praetermissa van der Werff. 446 Annals of the Missouri Botanical Garden Distribution. Costa Rica. from 1800 to 2200 m. Ocotea ra — ех Brandegee, Univ. Calif. Publ. 68. Ocotea platyphylla (Lundell) Rohwer, Bot. Jahrb. Syst. 112: 390. 1991. Phoebe platy- P Lundell, Contr. Univ. Michigan Herb. ii 23. 1941. ) Nectandra dU (L pies . K. Allen, J. Arnold Arbor. )2. 1945. m Mexico. Chiapas: Villes 1930 ciii type, MICH not seen; isotype, MO!). Very similar to O. heydeana, from which it differs in its narrowly inrolled or revolute leaf bases. Distribution. Mexico (Chiapas), Guatemala, El Salvador, from 900 to 2600 m. 69. Ocotea praetermissa van der Werff, Novon 6: 482. 1996. TYPE: Costa Rica. Cartago: Burger et al. 12065 (holotype, MO!). Similar to O. purpurea, but O. praetermissa can be recognized by its paniculate-cymose (not race- mose) inflorescences and its glabrous (not pubes- cent) flowers. Distribution. 3200 m Costa Rica, Panama, from 2000 to 70. Ocotea producta (C. K. Allen) Rohwer, Mitt. Inst. Allg. Bot. Vie 20: 143. 1986. Nec- tandra — ta С. К. Allen, J. Arnold Arbor. 26: 352. 1945. TYPE: dosis Rica. Prov. San José: Skutch 3906 (holotype, GH not seen; iso- type. Only known from the type collection and char- acterized by the long (4—5 ст) apparent petioles with inrolled, decurrent leaf bases. The filaments of the inner 3 stamens are fused. Distribution, Costa Rica, 700 m. 71. Ocotea с Burger, Fieldiana, Bot., n.s. 23: 92. . TYPE: Costa Rica. San José: Lent 1679 — ED. Diagnostic are the coriaceous leaves with erect indument on the lower surface, the few-flowered in- florescences, the large (6-8 mm diam.) flowers. and confused with the montane habitat. It can be mollifolia, which has chartaceous leaves, smaller (3-6 mm diam.) flowers, and occurs usually below 1000 m. Distribution. Costa Rica, Panama, from 2200 to m. 72. Ocotea puberula (Richard) Nees, Laurin. 472. 1836. Laurus puberula Richard, Actes Soc. Hist. Nat. Paris 1: 108. 1792. TYPE: French Guyana. Le Blond. s.n. (holo- ) Syst. — type, P not seen). Bot. 7: 326. 1920. TYPE: Mexico. Veracruz: Purpus 8456 Ae i S not seen; isotype, MO!). A widespread species, best recognized by its unisexual flowers, the slightly raised reticulation on the upper leaf surface, and the rather large, (thinly) chartaceous leaves. Ocotea leucoxylon has the up- per surface of the leaves generally smooth and smaller (3—4 vs. 4—5 mm diam. Distribution. From Mexico to Argentina (in Me- soamerica known from Chiapas, Costa Rica, and flowers. — Panama), from 0 to 1300 m. 73. Mentes pullifolia van der Werff, Novon 11: 509. 2001. TYPE: Costa Rica. Puntarenas: Canto de Golfito, Herrera 4119 (holotype, INB). Closely related to O. jorge-escobarii from Hon- duras and Nicaragua. Ocotea pullifolia is best rec- ognized by its obtuse or rounded leaf apices and in having pit domatia restricted to the axils of the low- ermost lateral veins. In some Panamanian collec- tions the pit domatia are lacking. Distribution. Costa Rica, Panama, from 200 to 1100 m 74. Ocotea purpurea (Mez) van der Werff, No- von 9: 579. 1999. Phoebe purpurea Mez, Jahrb. Königl. Bot. Gart. Berlin 5: 196. 1889. TYPE: Guatemala. Alta Verapaz: von Tuerckheim 371 (syntypes, B, K, LE not seen, photo MO!). Nectandra о Lundell, Wrightia 5: 33. 1974. TYPE: Guatemala. Contreras 11235 (holotype, LL not seen; (gem ). Readily identified by its usually racemose inflo- rescences, rather small (to 11 em long) leaves, and sparsely pubescent flowers. Specimens from Pana- ma differ in having fewer lateral veins (mostly 4 pairs Distribution. Mexico (Oaxaca, Chiapas), Gua- temala, Honduras, Panama, from 1400 to 2600 т. 75. Ocotea racemiflora Lundell, Wrightia 4: 107. 1969. TYPE: Guatemala. Alta Verapaz: Contreras 7904 (holotype, LL!). Only known from the type collection, which is in fruit with a few detached, old flowers. The flat, thin cupules with persistent tepals and roundish fruits are not known from other Mesoamerican species. Distribution. Guatemala. 76. Ocotea rhytidotricha Rohwer, Bot. Jahrb. Syst. 112: 391. 1991. TYPE: Nicaragua. Ma- tagalpa: Hall & Bockus 7919 (holotype, MO!). Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea Part of the O. heydeana group and characterized by the short, erect indument on the lower leaf sur- face. The hairs are inconspicuous, not discernable to the touch, and are best seen along the major veins. Distribution. El Salvador, Honduras, Nicara- gua, from 1200 to 1600 m. 77. Ocotea rivularis Standley & L. O. Williams. Ceiba 1: 238. 1951. TYPE: Costa Rica. Pun- tarenas: Allen 5590 (holotype, EAP not seen: isotype, MO!). Readily recognized by its large (30—50 cm long). obovate leaves and pubescent flowers. It can be confused with O. endresiana, which has smaller (10-30 em long) leaves and glabrous flowers. Distribution. Costa Rica, from 50 to 300 m. 78. Ocotea rovirosae Lorea-Hernandez & van der Werff, Brittonia 54: 2002 in press. TYPE: Mexico. Veracruz. This species is best recognized by the combi- nation of its short, erect indument on twigs and leaves, the paniculate-cymose, many-flowered inflo- 4—35 Nectandra lundellii rescences, and the rather large ( em long), re green drying leaves. . К. Allen. n large. yellow-green drying leaves, but has typ- known from the same area, has simi- ical Nectandra flowers with the inner surface of the tepals and the stamens papillose: its tepals are also spreading (not erect or half-erect) at anthesis. Vera- Distribution. Mexico (Oaxaca, Tabasco. cruz), from 10 to 200 m. 79. Ocotea rubriflora Mez, Jahrb. Kónigl. Bot. Gart. Pre 5: 279. 1889. Nectandra rubriflora . K. Allen, J. Arnold Arbor. 26: 372 — (syntype, G!). Mexico. Tabasco: — s.Hn. Ocotea perseifolia Mez & ge Smith, Bot. Gaz. (Craw- fordsville) 20: 1¢ ҮРЕ: Guatemala. Izabal: J. Donnell Smith 1807 piedad B not seen, US not seen). This belongs to the O. heydeana group in which it stands apart by the dense and minute indument on the twigs and inflorescences, which covers the close to O. crantha, but that species has larger flowers (10—12 surface completely. Vegetatively ma- — vs. 6—7 mm diam. Distribution. From southern Mexico to Panama, at 100-350 m 80. Ocotea rubrinervis Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 351. 1889. TYPE: Peru. Spruce 4580 (syntype, B!); Panama. Duchassaing s.n. (syntype, not seen). known from the Pacific — n Mesoamerica only coast of Panama and best recognized by its unisex- ual flowers, the readily visible gland dots on the upper leaf surface, and the shallow cupules with persistent tepals. Distribution. Panama, from 0 to 50 m. Ocotea rufescens van der Werff, Novon 6: 479. 1996. TYPE: Costa Rica. Limón: Aguilar & Schmidt 1077 (holotype, MO!). 81. The only Ocotea species in Costa Rica and Pan- ama (where it is expected to grow as well) with a dense, reddish brown or reddish indument on twigs and lower leaf surface. Ocotea salvinii Mez, from Mexico and Guatemala, has a similar indument but has inrolled leaf bases. Distribution. Costa Rica, from 450 to 1500 m. 82. Ocotea salvadorensis (Lundell) van der Werff, Novon 6: 481. 1996. dorensis Lundell, Wrightia 4: Vectandra salva- 105. 1969. Phoe- be salvadorensis (Lundell) Lundell, Wrightia 5: 344. TYPE: El Allen 7173 (holotype, LL not seen: isotype, GH!). Salvador. Р. Only known from the type collection. The dense, tomentellous, gray-brown indument on the lower leaf surface is diagnostic. Such an indument also occurs in O. pharomachosorum, which has tongue- shaped stamens with a sterile tip. and O. iridescens, which has domatia. A few collections from montane cloud forest in Honduras will key to O. salrador- ensis, but differ from that species in their reddish brown indument on twigs and leaves, and the rather dense indument on the inner surface of the tepals. More collections are needed to determine if the specimens from Honduras represent an unde- scribed specie Distribution. El Salvador, at 2500 m. 83. Ocotea salvinii Mez. Jahrb. Kónigl. Bot. Gart. Berlin 5: 264. 1889. Phoebe. salvinii (Mez) Lundell, Contr. Univ. Michigan Herb. 6: 23. 1941. Cinnamomum salvinii (Mez) Koster- Reinwardtia 6: 23. 1961. TYPE: temala. Salvin s.n. (syntvpes, K not seen, W mans, Gua- [1 / not seen). Diagnostic are the ferruginous tomentellous leaves with an inrolled base and the perfect flowers. See also under O. rufescens. 448 Annals of the Missouri Botanical Garden Distribution, Chiapas and Guatemala, from 2400 to 3000 m. 84. Ocotea sinuata (Mez) Rohwer, Bot. Jahrb. Syst. 112(3): 373. 1991. Nectandra sinuata Mez, Jahrb. Kónigl. Bot. Gart. Berlin 5: 402 (1889). TYPE: Guatemala. Bernoulli & Cario 2581 (syntypes, B, GOET, K, LE all not seen). Best recognized by its large (13-16 mm diam.) flowers, anthers with a sterile tip and densely brown-tomentose twigs. It can be confused with O. botrantha; see that species for differences. distribution. From southern Mexico to Panama, but not reported from Honduras and Nicaragua, from 200 to 1500 m. 85. Ocotea standleyi C. K. Allen, J. Arnold Ar- bor. 26: 343. 1945. Phoebe — nan dley & Steyermark, Field M "ubl. Bot. 116. 1944, non Phoebe mac dE (Blume) Blume. 1851. TYPE: Guatemala. Alta Vera- paz: Standley 70009 (holotype, F not seen). Readily recognized by its large (15-35 cm long М leaves with a rounded to subcordate base, апа the glabrous terminal buds and leaves. A Licaria spe- cies occurring in Chiapas is vegetatively similar, but its flowers (with only 3 stamens) and cupule (deeply cup-shaped with a weakly developed dou- ble margin) set it apart. The Licaria species gen- erally occurs below 1000 m. Distribution. Mexico (Chiapas), Guatemala, from 800 to 2500 m, mostly above 1500 m 86. Ocotea stenoneura Mez & Pittier, Bull. Herb. 2, 3: 233. 1903. TYPE: Costa Rica. San Jose de Dota, Tonduz 13377 (lectotype, designated by Allen (1945: 334), GH not seen; isolectotype, US!). Boissier, ser. A rarely collected species with decurrent, revo- lute leaf bases, an erect, rather dense indument on the lower leaf surface, and prominently raised ve- nation on the lower leaf surface. Specimens placed here vary in cupule shape, and more collections are needed to determine if two species are involved. Distribution. Costa Rica, Panama, possibly Co- lombia and Ecuador, from 700 to 1700 m. 87. Ocotea strigosa van der Werff, Ann. Missou- ri Bot. Gard. 75: 723. 1988. TYPE: Nicaragua. Matagalpa: Stevens 22181 (holotype, MO! — An inconspicuous species with ovate leaves and few-flowered inflorescences. It is similar to O. par- vula, but that species has domatia and glabrous (not sparsely pubescent) flowers. Ocotea iridescens is also a close relative, but differs in its dense indu- ment on the lower leaf surface Distribution. Nicaragua, from 1000 to 1600 m. 88. Ocotea subalata Lundell, Lloydia 4: 48. 1941. TYPE: Mexico. Chiapas: Matuda 2957 (holotype, MICH not seen; isotype, F!). Useful char- acters for identification are the long (to 20 cm) in- florescences and the sharply angled or winged, Only known from two collections. densely pubescent young twigs. Distribution. Mexico (Chiapas), from 2100 to 2500 89. Ocotea tenera Mez & Donnell Smith, Bull. Herb. Boissier, ser. 2, 3: 234. 1903. TYPE: Costa Rica. Pittier 13396 (syntype, US!). Oc ba — Lundell, Wrightia 6: 9. 1978. Syn. nov. Guatemala. Lundell & Contreras 20948 (ho- ee LL not seen; isotype, ! Ocotea eucymosa Lundell, Wrightia 5: : 35. 1974. Syn. nov. TYPE: Guatemala. Contreras 11215 (holotype, LL not seen; isotype, MO! This species is best recognized by its small size (to 12 m tall), dark-drying leaves with gland dots on the upper leaf surface, glabrous terminal buds, and small, glabrous flowers with erect or half-erect tepals. Distribution. Guatemala and Costa Rica, from 100 to 1600 m. 90. Ocotea tonduzii Standley, Field Mus. Publ. Bot. 18: 456. 1937. Ocotea cuneata Mez, Bot. Jahrb. 30, Beiblatt 67: 17. 1901, non Ocotea cuneata (Grisebach) Gomez. 1889. TYPE: Costa Rica. Tonduz 2142 (syntype, BR!). reddish indument on the flowers, the large, glabrous (or finely ap- Characteristic are the dense, pressed pubescent) terminal buds, the raised ter- tiary venation on the lower leaf surface, and the glabrous, nearly sessile leaves with an inrolled base. Ocotea endresiana has a similar appearance but lacks the reddish indument of the flowers. Distribution. Costa Rica, from 1500 to 2400 m. 91. Ocotea tonii (Lundell) van der Werff, Novon 9: 579. 1999. Nectandra tonii Lundell, Wrigh- tia 4: 106. 1969. TYPE: Mexico. Chiapas: Shi- lom Ton 2014 (holotype, LL not seen; isotype, NY!). Distinctive are the clustered leaves with short (to 6 mm long) petioles, yellow-brown indument, pa- Volume 89, Number 3 2002 van der Werff Synopsis of Ocotea niculate-cymose inflorescences, and pubescent flowers. See also under O. congregata. Distribution. Mexico (Chiapas). from 1000 to 1500 m. 92. Ocotea truncata Lundell, Phytologia 12: 244. 1965. TYPE: Guatemala. Alta Verapaz: bu "731 (holotype, LL not seen; iso- type, US! м This species сап be recognized by its thin, dark green drying leaves with rather conspicuous dom- atia and its few-flowered inflorescences with gla- brous flowers. The truncate apex of the young fruits is less pronounced in mature fruits and scarcely helps in the identification of this species. Distribution. Mexico (Chiapas), from 700 to 1200 m. Guatemala, 93. Ocotea uxpanapana Wendt & van der Werff, nn. Bot. Gard. 74: 413. 1987. TYPE: Mexico. Veracruz: Mpio. Minatitlan, Wendt et al. 2869 (holotype, MEXU not seen: isotype, MO!). Missouri Currently only known from the Uxpanapa region in Veracruz. It resembles O. eucuneata and O. ni- grita but differs in its rather large (to 25 em long). elliptic larger. flowers. The strongly lobed cupules appear to be a unique fea- ture of this species, but cupules of O. eucuneata and O, nigrita are not yet known. Distribution. Mexico (Veracruz), to be expected in Chiapas, from 100 to 300 m. leaves and slightly 94. Ocotea valeriana (Standley) Burger, Fieldi- ana, Bot., n.s. 23: 96. 1990. Phoebe valeriana Standley, Publ. Field Mus. Nat. Hist., Bot. Ser. 18: 460. 1937. Cinnamomum valerianum (Standley) Kostermans, Reinwardtia 6: 24. 1961. TYPE: Costa Rica. Tonduz 11746 (ho- lotype, F not seen м — austinit €. K. Allen, J. Arnold Arbor. 26: 374. n TYPE: Costa Rica. Alajuela: Austin Smith P 2 — A not seen). — smithii C. Allen Arnold Arbor. 26: 317. f an niidina Kostermans, Rein- — 6: 2; . 1961, non Cinnamomum smithii Lu manoff. : Costa Rica. Alajuela: Austin Smith P.C. 307 Чаки F not seen). Similar to O. helicterifolia, from which it differs in its indument of the twigs (short matted hairs and longer, erect ones in O. valeriana vs. only long, erect ones in O. helicterifolia), its deeper cupules, and its presence at higher altitudes. Distribution. to 2200 m. Costa Rica and Panama, from 800 95. Ocotea valerioides Burger, Fieldiana, Bot.. n.s. 23: 97. 1990. TYPE: Costa Rica. Harts- horn 1530 (holotype, СК not seen; isotype, MO! Distinctive are the large (to 40 em long) leaves and the dense, tomentellous indument on the twigs and inflorescences. Differences with O. lentii are discussed under that species. Distribution. 500 m. Costa Rica, Panama, from 50 to 96. Ocotea vanderwerffii (Kostermans) van der Werff, Novon 11: 511. 2001. Phoebe — van der Werff. Ann. Missouri Bot. Gard. 406. 1987, non Ocotea glabra van der on Cinnamomum vanderwerffit Kostermans, Rein- wardtia 10: 454. 1988, non Cinnamomum gla- brum Ettinghausen. TYPE: Wendt et al. 4813 (holotype, N exico. Oaxaca: MO!). Ocotea vanderwerffii is best recognized by its dark-drying, tripliveined leaves with large pit dom- atia in the axils of the basal lateral veins and the glabrous flowers with erect tepals. It resembles 0. tenera, but that species has pinnately veined leaves and small pit domatia, which are not visible on the upper surface Mexico Distribution. Southern 1200 m. from 150 to 97. Ocotea br iar (Meissner) Mez, Jahrb. Kónigl. Bot. Berlin 5: 240. 1889. Sas- safridium veraguensis Meissner, DC. 15(1): 171. 1864. TYPE: Costa Rica. Bridges s.n. (syntype, K not seen), Oersted, Laur. 10 (syntype, В not seen), Warszewicz 1 (syntype, G not seen). Gart. Prodr. 65. 1917. Ocotea bakeri Blake, Contr. Gray РЕ GH not TYPE: Nicaragua. ` Boker 2493 eh een). Ocotea a rines L unde Il, Contr. Univ. Michigan Herb. 6: 941. TYPE: Mex Chiapas: Matuda 654 (h — MICH not see Ocotea paradoxa Mez, Bot. Jahrh. 30. Beibl. 67: 16. 1901. TYPE: Tonduz in Herb. Phys.- Geogr. Costar. 7648 (holotype, BR). xico. Costa Rica. A widespread and rather frequently collected species, best recognized by its papillose, tongue- shaped stamens with a sterile tip, solid stems, gla- brous (or nearly so) leaves, and its cupules with a double margin. Can be confused with O. dendro- 450 Annals of the Missouri Botanical Garden daphne, which has hollow twigs and a pubescent inside of the receptacle. Distribution. From southern Mexico to Panama, from 10 to 2000 m. 98. Ocotea verapazensis Standley & Steyer- Field Mus. Publ. Bot. 23: 114. 1944. ҮРК: Guatemala. Standley 71421 (holotype, En A poorly known species with glabrous terminal buds, twigs, and leaves, and slightly winged peti- Vegetatively similar to О, subalata, but that species has densely pubescent terminal buds and oles. pubescent inflorescences and flowers. Distribution. | Guatemala, Eon 1500 to 2000 m. 99. Ocotea verticillata Rohwer, Bot. Jahrb. Syst. 112(3): 369. 1991. TYPE: Mexico. Veracruz: Ibarra 2328 (holotype. MO!). Characteristic are the whorled leaves abruptly rounded at the base and its tongue-shaped stamens with a sterile tip. Related to O. botrantha and O. sinuata, both of which have alternate leaves. Distribution, Mexico (Veracruz, Oaxaca), from 100 to 1000 m. 100. Ocotea viridiflora Lundell, Wrightia 5: 36. 1974. TYPE: Panama. Chiriquí: Proctor 31916 (holotype, LL not seen; isotype, MO). Similar to O. meziana, but differing in its gla- brous terminal buds and sharply acute leaf apices. Collections from Panama have persistent tepals on the cupule; those from Costa Rica do not. Distribution. Costa Rica, Panama, from 1300 to 1800 m. Ocotea whitei Woodson, Ann. Missouri Bot. Gard. 24: 188. 1937. Nectandra whitei (Woodson) €. K. Allen, J. Arnold Arbor. 26: 398. 1945. TYPE: Panama. Chiriquí: Seibert 307 (holotype, MO!). 101. Wrightia 5: 338. 1977. TYPE: Ocotea eusericea Lundell, Panama. Chiriquí: Proctor 31858 (holotype, LL not seen: webs О! йш? — |. К. ‚ J. Arnold Arbor. 26: 352. DE — a. Skutch US GH de seen; isotype, MO As accepted here, a variable species best rec- ognized by the dorsally pubescent filaments of the inner 3 stamens, the rather narrow (2—5 cm wide), frequently oblanceolate or obovate-elliptic leaves with a decurrent and sometimes slightly inrolled base. Distribution. Nicaragua, Costa Rica, and Pan- ama, from LOO to 2500 m. IMPERFECTLY KNOWN SPECIES 102. Ocotea sp. A. Resembles O. brenesii, from which it differs in its cymose-paniculate (not racemose) inflorescence, a denser indument of the the twigs, and a — (not pubescent) inner surface of the receptacle. 1 is only known from one collection, Carvajal ж (MO), from Costa Rica. It is likely an undescribed species, but more material is needed for a descrip- tion. In Rohwer (1991) it was included as Ocotea sp. В P2 Literature Cited Allen, C. K. 1945. Studies in the Lauraceae VI. A prelim- inary survey of the Mexican ы Central American spe- cies. J. Amo — 26: 280-434. Burger, W. & der Werff. Costaricensis. qs Idiana, Bof, Chanderbali, A. S., Н. van der fT & S. Renner. 2001. Phylogeny and historical — hy of Lauraceae: Evidence from c — ist and nuclear genomes. Ann. Missouri Bot. Gard. 88: 104—134. Gomez-Laurito, J. 199: i new Ocotea (Lauraceae) from the high ı mountains S Costa Rica and Panama. Novon 313 1990. Lauraceae. In: Fl. n.s. 23: 1-129. . 1997. Ocotea morae (Lauraceae): A new species from Cos a Novon 7: 145-146. Hammel, В. E. 1986. New species and notes on Lauraceae from the ( J. Arnold :aribbean lowlands of Costa Rica. Arbor. 67: 123-136. Lorea-Hernandez, F. G. & Н. van der Werff. 2002. new species of Ocotea * southern Mexico. Brittonia Three In press. Lundell, C. L. 1965. Additions to the Lauraceae of Gua- temala. — 12: 243—246 ————. 1909. Studies of ond Plants I. Wrightia 4: 97-128. . 1970. Studies of American Plants П. Wrightia 4: 129-1 52. . 1971. Studies of American Plants Il. Wrightia 4: 153-170. . 1974a. Studies of American Plants VI. Wrightia : 23-44. . 1974b. Studies of American Plants VII. Wrightia 5: 51-7: — 1977. Studies of American Plants XIV. Wrightia 5: 331-359 А ————. 1978. Studies of American Plants XVI. Wrightia 6: 4—20. Mez, C. 1889. Lauraceae Americanae. Jahrb. Kónigl. Bot. art. Berlin 5: 1-550. Nelson, C. 1984. Una Ocotea (Lauraceae), una Salvia (La- biatae) y un peara i :ompositae) nuevos de Hon- duras. Ceiba 25: 173-17€ 1986. — einer Monographie der Gattung Ocotea Aubl. (I — sensu lato. Mitt. Inst. Allg. puo oe 20: 1-27 Borderline cases 9 ‘tween Ocotea, < Nectan- Volume 89, Number 3 van der Werff 451 2002 Synopsis of Ocotea dra and Phoebe (Lauraceae): The marginal species of | ———. 1991. A key to the genera of тн in the the O. helicterifolia group, including the O. heydeana New World. Ann. Missouri Bot, Gard. 78: 377-387. group. — Jahrb. Syst. 112(3): 365—397. —— ———. |996. Notes on Costa Rican amabas with the — —. 1993. Lauraceae. Pp. 366—391 in K. Kubitzki, J. dese "s gi se — new species. Novon 6: 476—483 — г & V. Bittrich (editors), The F bem апа. 1999. New taxa and combinations in the Ocotea E nera of ls ular Plants IL Springer- Verlag. Berlin. helicieiola i auraceae) species group. Novon 9: 571— van der Wer 1987. Six new — of Neotropical 283. ; аас Im — Bot. Gard. 74: 401—412 —— 2001. New = and combinations in а ios eae aie — sal сеае) from Central America. Novon 11: on 511. [988a. Eight new species sand ue DUM combi- Wendt, T. 1998. Ocotea wa pne (Lauraceae), г pe- nation of шры Lauraceae. Ann. Missouri Bot. 1 cies of rain forest canopy tree — he md p Te - Gard. 402— | | — Mexico. Lundellia 1: 40— 1988 88b. A new species of Ocotea (Lauraceae) from є H. van der Werff. 1 = А new spec iés of Ocos айтарга and a — on n Ocotea jorge-escobarii. Aun. tea (Lauraceae) D Vu Mexico. Ann. Mis- 727. :413— Missouri Bot. Gard. 7 3-7 souri Bot. Gard. Volume 89, Number 3, pp. 305—451 of the ANNALS OF m — IRI BOTANICAL GARDEN is published on October 1, 2002. — ANNALS -OF THE MISSOURI BOTANICAL GARDEN AND MISSOURI BOTANICAL GARDEN ANNUAL REPORT REE 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 Conservation Biology Diversity and Distributions Ecological Applications Ecological Monographs Ecology 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 Botanical Society Limnology and Oceanography Missouri Botanical Garden Annual Report New Phytologist Paleobiology Quarterly Review of Biology Systematic Biology Systematic Botany Information regarding JSTOR is available at http://www.jstor.org 120 Fifth Avenue, New York, NY 10011 . 3 1753 00294 3147 | 21 | | || ! CONTENTS Phylogenetic Relationships of the Genus Piptochaetium (Poaceae, Pooideae, Stipeae): Evidence from Morphological Data Ana Maria Cialdella & Liliana Ménica Giussani Systematic Revision and Phylogeny of Paspalum Subgenus Ceresia (Poaceae: Panicoideae: Paniceae) ............ Silvia 5. Denham, Fernando O. Zuloaga & Osvaldo Morrone Phylogeny of the Tribe Hymenocallideae (Amaryllidaceae) Based on Morphology and Molecular Characters lan W. Meerow, Charles L. Guy, Qin-Bao Li & Jason R. — Phylogenetic and Biogeographic Diversification in Osmorhiza (Apiaceae) —— Jun Wen, Porter P. Lowry II, Jeffrey L. Walck & Ki-Oug Yoo A Synopsis of Ocotea (Lauraceae) in Central America and Southern Mexico — Henk van der Werff Cover illustration. | Breonia richardsonii Razafim., drawn by Barbara Alongi. Annals of the Missouri Botanical Garden 2002 G Volume 89, Number 4 Fall 2002 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 Committee Victoria C. Hollowell ditor, Missouri Botanical Garden Amy Scheuler McPherson Managing Editor, Missouri Botanical paden 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, 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 © Missouri Botanical Carden 2002 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 mission of the Missouri Botanical Garden is to discover and share knowledge about plants and RD thei , in order to puan and enrich life. La k P © This paper 239.48-1992 (Permanence of Paper). olume 89 Annals umber 4: of the 002 Missouri Botanical Garden Ld PHYLOGENETIC Victor W. Steinmann? RELATIONSHIPS IN J. Mark Porter? EUPHORBIEAE (EUPHORBIACEAE) BASED ON ITS AND ndhF SEQUENCE DATA! and ABSTRACT The monophyly and phylogenetic relationships of the tribe Euphorbieae (Euphorbiaceae) were evaluated using sep- juences of the nrDNA internal transc ribed spacer (ITS) sec s representing nearly all of the previously arate — maximum parsimony analyses of nucleotide region and cpDNA coding region ndAF. The study included 223 ingroup speci recognized genera, subgenera, and sections within the К uphorbieae and 4 outgroup taxa from the tribe Hippomanose (E прогини eae е). Te th the ITS and ndhF analyses support the monophyly of Euphorbieae in addition to the monophy subtribes. Anthosteminae, Neoguillauminiinae, and Euphorbiinae. Within Euphorbiinae, there are four major vhich copos ds w with a previously recognized taxon: the three remaining clades d of a paraphyletic ety All other ind Endadenium of its three clades, — one of y of various subge nera and sections. The majority of the subtribe is compose genera currently recognized in the subtribe are nested within Euphorbia. In addition, Synadenium ai are nested within — nium. Within Euphorbia, the majority of the currently recognized subgenera are either par- aphyletic or polyphyle vlet iogeographical patterns examined in light of the also idence suggest that the tribe Euphorbieae arose in Africa. possibly before the breakup of Gondwanaland, at which time aha major — s of subtribe ^A | are extremely grateful to all those who have assisted with this research, In partic ‘ular we thank Susan Carter, J. Travis Columbus, Elisabeth Friar, Victoria Hollowell, Gordon McPherson, and Grady Webster for critically reviewing the — ap For — to e — their collections and to sample he еа material, we thank the curators . DAV. ENCB. E GB. K. MO. NY. RSA. SGO. SP, and TEX. The following institutions and and s ARIZ, C. iid extended assistance we pert o access a their living collections: Dylan Hannon; Chuck Hanson of Arid Lands Greer — ‘sı Mark Mayfield; Meena Singh: mour Linden; John Trager and Joe Clements of the Huntington Botanic Garden: Urs Egeli o of с, anten- — Zürich: and Herman Schwartz. Caden McPherson and Sebastian Teillier collecte d material of gemini cleopatra and Euphorbia germainii, respectively. Fieldwork, herbarium visits, and laboratory use and mate ‘rials were made possible by the generous support of the Andrew W. Mellon Foun- dation, the American Society of Plant | жикке, m Howard and Phoebe Brown Sc holarship, and the е Native Plant Research Award. Lauren Raz provided ITS sequences for C иаа degeneri, C. hypericifolia, a rata. Kenneth Wurdack gave useful information about outgroup selection. We are inde bted to the staff and —— Garden for encouragement and support, especially Eric Roalson for his = prost students at the Rancho Santa Ana Botanic Е help апа advi inc үш; — Ana Botanic Garden, " 1.Mx 1500 N. College Ave.. Claremont, California 91711, U.S.A. steinmann@ ‘Tine — Instituto de Ecología. A.C., Centro Regional del Bajío, A.P. 386, 61600 Pátzcuaro, Michoacán, Mexico. ANN. Missouri Bor. GARD. 89: 453—490. 2002. 454 Annals of the Missouri Botanical Garden Euphorbiinae were alre scription of Euphorbia that contains all of Key words: the about 20( sady present. It is argued that the best solution for Euphorbia classification is a broad circum- JO species of the subtribe Euphorbiinae. classification, Euphorbia, Euphorbieae, ITS, ndhF. It was only a year after Linnaeus (1753) first cir- cumscribed the genus Euphorbia L. (Euphorbi- aceae) to include all of the then-known members of the tribe Euphorbieae that other botanists began to divide the genus into several smaller genera (Miller, 1754; continued since, and almost 250 years later the is- sue as to whether Euphorbia should be recognized Trew, 1754). The controversy has in its initial broad sense or be separated into many smaller genera still has not been resolved. Euphor- bia and Euphorbieae are generally considered tax- onomically difficult, and a considerable degree of uncertainty has always existed about the relation- ships of the groups within them. This problem is due in great part to extreme morphological diver- sity, a large number of species, and a subcosmo- politan distribution. Few workers have been able to gain a complete understanding of the tribe through- out ils immense range, and there has never been a This lack of consistency has hindered and to some extent dis- couraged research within the genus Euphorbia as well as the universally accepted classification. Euphorbieae as a whole. Thus, a clear understanding of relationships within the tribe is of great importance in order to provide subsequent workers with a phylogenetic framework on which to base their studies. The tribe Euphorbieae is characterized by its synapomorphic pseudanthial inflorescence (termed a cyathium) composed of a gland-bearing involucre of several united bracts and their associated flowers and bracteoles. Each bract subtends a monochasial staminate inflorescence, and these monochasia sur- round a single pistillate flower. The individual flow- ers in Euphorbieae are highly reduced and repre- sented by a single stamen or ovary, with or without a perianth. The staminate flowers terminate slender pedicels, and the pistillate flowers can be long-ped- icellate or subsessile. This structure is highly com- plex, and there are still doubts as to its exact nature and evolution (Gilbert, 1994). According to the most recent synopsis of the Eu- phorbiaceae (Webster, 1994), the Euphorbieae con- tain 11 genera placed into three subtribes: Anthos- teminae (Baill.) G. L. Webster, Croizat, and Euphorbiinae. Neoguillauminiinae r ab- sence of a perianth (presumably a calyx) on the flowers distinguishes these taxa. In Anthosteminae both the staminate and. pistillate flowers possess a The presence « perianth; in Neoguillauminiinae only the pistillate flowers possess a perianth; and in Euphorbiinae neither the pistillate nor the staminate flowers pos- sess a well-developed perianth, although a rudi- mentary calyx-like structure is present below the pistillate flowers in a few species. Anthosteminae consist of two genera of tropical forest trees: Anthostema А. Juss. (4 spp. disjunct in west tropical Africa and Madagascar) and Dichos- temma Pierre (1 sp. in west tropical Africa). This subtribe is considered the least specialized because of the perianth on both the staminate and pistillate Further, each individual involucral bract closely envelops a cluster of many bracteoles and staminate flowers. flowers. The cyathia are bisexual or sta- minate. In Anthostema the cyathia are arranged in condensed axillary cymes. The involucre is made up of four united bracts and is slightly zygomorphic because it is spread in an open half circle. There are five large glands along the involucral bract mar- gins and between the clusters of staminate flowers. The 3-locular pistillate flower is not contained in the involucre but instead lies at the base of the involucre’s open side. Based on its open involucral morphology and the possible lateral position of the pistillate flower, this genus probably most closely resembles the ancestral inflorescence morphology of the Euphorbieae. In Dichostemma the cyathia are arranged in loose, open, axillary or terminal cymes. In contrast to the four monochasia- containing involucral bracts are united in a Anthostema, ring, and the involucre is completely closed and acti- nomorphic. The four involucral glands are con- tained. within this structure and attached to the in- ner walls of the and the base of the gynophore. In the center of the inflorescence is ei- ther a minute pistillode or a 4-locular pistillate flower. bracts Like Anthosteminae, the subtribe Neoguillau- miniinae also contains two genera: Calycopeplus (5 spp. in Australia) and Neoguillauminia Croizat (1 sp. in New Caledonia). In contrast to Anthostemi- nae, the staminate flowers lack a perianth. How- ever, a perianth is present on the pistillate flower, and this combination of features defines the sub- tribe. In addition, the bracts of the involucre do not tightly envelop the staminate monochasia, although the latter are enclosed within large bracteoles. Ca- lycopeplus are xerophytic shrubs with small, oppo- Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae site leaves and cyathia arranged in axillary clusters. The involucre is distinctly cupular and is made up of four bracts. The glands are located between the lobes and attached to the wall of the cupular in- voluere. There is a central pistillate flower sur- rounded by four staminate monochasia that are op- posite the involucral lobes. The sole representative of Neoguillauminia is a mesic forest tree or shrub with large, spirally arranged leaves. The cyathia are long-pedunculate and arranged in few-cyathiate ax- illary or subterminal groups. The involucre is com- posed of four to six bracts that extend into large, petaloid appendages. There are eight to twelve glands arranged in pairs between the four to six staminate monochasia and attached to the base of the involucre and sometimes also to the base of the gynophore. There is a single central pistillate flow- er. Neoguillauminia is noteworthy because the in- volucres are generally composed of five bracts, and a 5-merous involucre also characterizes Euphorbi- inae, discussed below. Calycopeplus, as well as An- thostema and Dichostemma, possess 4-merous in- volucres With about 2000 species and a subcosmopolitan distribution, the largest and most complex subtribe is Euphorbiinae. It is characterized by the lack of a perianth on both the staminate and pistillate flow- ers, although a rudimentary calyx-like structure is present below the pistillate flowers of a few species. In addition, the involucre is made up of five united bracts, not four as generally are found in the other subtribes of Euphorbieae, and the bracteoles that surround the staminate monochasia are generally reduced. The glands are mostly located along the rim of a cupular involucre. Again following the cir- cumscription of Webster (1994), the subtribe con- tains seven genera: Chamaesyce Gray, Cubanthus (Boiss.) Millsp., Endadenium Leach, Euphorbia. Monadenium Pax, Pedilanthus Necker ex Poit., and Synadenium Boiss. Poinsettia Graham and Elaeo- phorbia Stapf are also sometimes recognized as dis- tinct from Euphorbia, but their status is not as widely accepted, and most authors treat these as infrageneric taxa of Euphorbia. Extreme emphasis is placed on variation in the configuration of the eyathium, and with the exception of Chamaesyce, the genera segregated from Euphorbia are distin- guished on the basis of involucral features. Euphorbiinae are dominated by Euphorbia, which accounts for ca. 80% of the species and oc- curs throughout the geographic range of the sub- tribe. The genus is best known for the common Christmas poinsettia (Euphorbia pulcherrima Willd. ex Klotzsch) and is popular with horticulturists be- cause of the prevalence of succulents. One of the most fascinating features of the subtribe is its great diversity of growth forms. Large forest trees, shrubs, perennial herbs, geophytes, annuals, and a great diversity of succulents are all well represented. In comparison to most other genera of Euphorbiinae. the cyathial morphology of Euphorbia is relatively unspecialized. The cyathia are actinomorphic and generally possess one to five separate glands situ- ated on the rim of the involucre, and this plesiom- orphic feature unites the genus. With about 300 species, Chamaesyce is the larg- esl segregate genus from Euphorbia. lt occurs wide- ly, but most species are confined to the New World. Distinguished on the basis of vegetative morphol- ogy, its cyathia are nearly identical to those of many species of Euphorbia subg. Agaloma (Raf.) House. Chamaesyce is characterized by many unusual syn- apomorphies: apical abortion of the main shoot and subsequent. sympodial growth; interpetiolar stip- ules: opposite, frequently asymmetrical leaves; and C, photosynthesis (Koutnik, 1984. 1987). Numer- ous Euphorbieae specialists (e.g.. Carter, 1988a, 1992b: 1987) retain it as a subgenus of Euphorbia. Gilbert, The other five segregate genera of Euphorbia are also easily identifiable, but their differences, as mentioned above, involve involucral features. Three genera possess zygomorphic cyathia: in Pedilanthus primarily Mexico) there are two to six glands enclosed within an adaxial, spurlike exten- sion of the involucre; in Cubanthus (3 spp., Cuba and Hispaniola) there are two glands united into a (15 spp.. shieldlike structure on the outside of the involucre: and in Monadenium (ca. 70 spp., Africa) the glands are united into a single horseshoe-shaped structure. The two remaining segregate genera possess acti- nomorphic eyathia: in Synadenium (20 spp., Africa) the eyathia possess five united glands that form a complete ring around the top, and in Endadenium (1 sp.. Angola) the closed rim of the cyathium is not a gland but instead an apparently eglandular extension of the involucral wall with a ring of nec- tar-bearing depressions on the inside of the invo- lucre TAXONOMIC HISTORY OF THE TRIBE EUPHORBIEAE The taxonomic works treating Euphorbieae are numerous, and only a brief overview, focusing on those with broad and significant implications to modern Euphorbieae taxonomy, will be provided here. A more detailed account is found in Stein- mann (2001 One of the first prominent taxonomic treatments — of the Euphorbieae was that of Klotzsch and Garcke 456 Annals of the Missouri Botanical Garden (1859, 1860). Although these authors segregated that of Pax and Hoffmann (1931) in Engler’s Die both Anthostema and Pedilanthus into separate tribes, the Anthostemeae and Pedilantheae, they did concede their close relationship with Euphor- bieae. Their Euphorbieae contained 408 species, all members of present-day Euphorbiinae. Euphor- bia was recognized in a restricted sense, with only 27 species. The remainder were placed in 17 seg- regate genera, organized in two subtribes: Aniso- phyllae Klotzsch & Garcke Klotzsch & Garcke. The first and last monograph of the entire Eu- and Tithymalae phorbieae, and the basic framework for the classi- fication that is still in use today, was provided by (supplement 1866) for de Can- Boissier followed Klotzsch and Boissier in 1862 dolle's Prodromus. Garcke by recognizing Anthostemeae as distinct, but differed from them by submerging Pedilantheae into Euphorbieae. Euphorbieae contained 740 spe- cies. Also in contrast to Klotzsch and Garcke, Bois- sier recognized Euphorbia in a broad sense, and his Euphorbieae contained only three genera, all currently placed in Euphorbiinae: Pedilanthus (15 spp.). Synadenium (2 spp.), and Euphorbia (723 spp.). The taxa of Euphorbia were positioned into 27 sections and two “series”: Appendiculatae (corresponding to Klotzsch and Garcke's Anisophyllae) and (corresponding to Klotzsch and Garcke's Boiss. subtribe Exappendiculatae Joiss. subtribe Tithymalae). Calycopeplus and Neoguillau- minia were treated in Euphorbia, and Cubanthus was erected as a section of Pedilanthus. Bentham (1878) expounded greatly on the Eu- phorbiaceae as a prelude to his treatment of the family in Genera Plantarum (Bentham, 1880). He agreed that Euphorbia should be recognized in the broad sense and followed Boissier’s treatment with only minor modifications. Anthostemae was com- bined with Euphorbieae. Calycopeplus, first de- seribed by Planchon in 1861 but subsequently treated as a synonym of Euphorbia by Boissier (1862), was resurrected at the rank of genus. Ben- tham’s greatest change concerned the rank of Bois- sier's sections of Euphorbia, and he proposed a sys- tem containing only six sections, under which the majority of Boissier’s sections were reduced to sub- sections. Otherwise, Bentham did little to modify Boissier's grouping of species. Around the turn of the century, Pax (1894a) de- scribed Monadenium and Pierre (1896) described Dichostemma, the second known genus in the mod- Nearly two decades later, Millspaugh (1913) first recognized Cubanthus ern subtribe Anthosteminae. as a distinct genus. The next major treatment of the Euphorbieae was natürlichen Pflanzenfamilien. They essentially fol- Bentham's classification with only minor newly taxa. Again, a single tribe, the Euphorbieae, was Owe o accommodate described e modifications recognized without any further divisions. Anthoste- ma, Dichostemma, and Calycopeplus were all treat- ed as distinct. Euphorbia continued to be recog- nized in the broad sense, although the number of segregate genera had increased. In addition to the earlier recognized Pedilanthus and Synadenium, these authors segregated the genera Monadenium, Stenadenium Pax, Elaeophorbia, and Diplocy- athium H. Schmidt. now treated within Euphorbia (Webster, 1994), and Stenadenium is currently — as a synonym of Monadenium (Bally, 195° relegated back to a section of Pedilanthus. The latter two segregates are . 1961). Cubanthus was In 1937, Croizat described the genus Neoguil- lauminia and the subtribe Neoguillauminiinae to accommodate it. The type, N. cleopatra, was first described by Baillon (1861) as a species of Eu- phorbia and placed in his monotypic E. sect. De- cadenia Baill. Boissier (1862) treated this species within E. sect. Tithymalus. Wheeler (1943) published a broad classification of the entire Shortly after. Croizat's contribution, Euphorbieae. This was the first major conspectus to advocate the use of subgenus as the primary di- vision of Euphorbia, and it is this rank that pre- dominates in current. Euphorbiinae classification. Wheeler recognized a single Euphorbieae but abandoned Croizat's subtribes. The genera that he included were the same as those of Pax and Hoff- mann (1931) except that Diplocyathium was rele- gated to synonymy within Euphorbia, and Croizat's newly described Neoguillauminia was recognized. Euphorbia consisted of eight subgenera, and in general, these corresponded to the sections recog- (1931). Dressler (1957) provided a monograph of Pedi- nized by Pax and Hoffmann lanthus in which he convincingly demonstrated that the species of Cubanthus do not belong within Ped- ilanthus and should be treated separately. Dressler is also noteworthy because he advocated that Eu- phorbia should be recognized in the narrow sense and restricted to Old World succulents. In the years after Dressler's treatment, Webster made important contributions to Euphorbieae tax- onomy. He suggested the recognition of Chamae- syce as a genus and at the same time strongly sug- gested that Elaeophorbia should be treated within Euphorbia (Webster, 1967). Eight years later, Web- er (1975) circumscribed the tribe to its present, generally accepted. configuration. He created. the Volume 89, Number 4 2002 Steinmann & Porter 457 Phylogenetic Relationships in Euphorbieae subtribe Anthosteminae to accommodate Anthoste- ma and Dichostemma and resurrected Neoguillau- miniinae to include Calycopeplus and Neoguillau- minia. In his third subtribe Euphorbiinae, he recognized seven genera: Euphorbia, Chamaesyce, Cubanthus, Endadenium, Monadenium, Pedilan- thus, and Synadentum. Gilbert (1987) and Carter (1985, 1988a) have made the most recent modifications in Euphorbi- inae classification. Their changes primarily involve African members and the elevation of various pre- viously recognized sections of Euphorbia to sub- generic rank In summary, at the time of this writing most au- thors agree that Euphorbieae is a single tribe with three subtribes: the Anthosteminae, the Neoguillau- miniinae, and the Euphorbiinae. Within Anthos- teminae there are two genera, Anthostema and Di- chostemma. Within Neoguillauminiinae there are also two genera, Calycopeplus and Neoguillaumi- nia. The subtribe Euphorbiinae contains about sev- en genera. Those that are universally accepted are Endadenium, Monadenium, Euphorbia, Synaden- ium, Pedilanthus, and Cubanthus. Genera that are less frequently accepted are Chamaescye, Elaeo- Other have not even gained minor acceptance. Within Fu- phorbia, and Poinsettia. segregale genera phorbia, the common primary division employed is the rank of subgenus, and there are 9 to 11 gen- erally recognized subgenera: Agaloma, Chamaesyce Raf. (when not treated as a genus), Poinsettia (Gra- ham) House (when not treated as a genus), Esula Pers., Eremophyton (Boiss.) 7 C. Wheeler, Fu- phorbia, Lacanthis (Raf.) M. G. Gilbert, Tirucalli (Boiss.) S. Carter, Trichadenia pee S. Carter, Rhi- zanthium (Boiss.) L. C. and Lyctopsis (Boiss.) L. С. coherent classification of the genus is lacking. and heeler, Wheeler. However, a modern, global, some proposed sections of Euphorbia have not been adequately accommodated within a currently rec- ognized subgenus (e.g., E. sects. Arthrothamnus (Klotzsch & Garcke) Boiss., dri) Croizat, and Deuterocalli Croizat). Detailed ac- Denisophorbia (Lean- counts of the currently recognized infrageneric taxa of Euphorbia and their taxonomic histories are pre- sented in the discussion section. MATERIALS AND METHODS In total, 227 species were sampled (Appendix 1), including 4 outgroup and 223 ingroup species. Out- groups were chosen from tribe Hippomaneae be- cause it is traditionally considered to be closely related to the Euphorbieae (Webster, 1994 broad molecular phylogenetic reconstruction of the . and a м entire Euphorbiaceae also supports their affinity (Kenneth Wurdack, pers. comm.). An attempt was made to include as broad a sample as possible from Euphorbieae. Four species of Anthosteminae, 4 species of Neoguillauminiinae, and 215 species of Euphorbiinae were sequenced. These included all genera (except Cubanthus) recognized by Webster 1994). eight species of Pedilanthus, one species of Syna- The sole representative of Endadenium, —_ denium, eight species of Chamaesyce, and five spe- cies of Monadenium were included. One hundred ninety-two species of Euphorbia were represented, including all sections (except sects. Bongium Boiss. and Caulanthium Boiss.) treated by Boissier (1862) and most other important taxonomic groups recog- nized by subsequent workers (e.g., Webster, 1967; Gilbert, 1987; Carter, 1985, 1988a). The ITS 3 Anthosteminae, 2 Neoguillauminiinae, analysis included 216 species (4 out- groups. and 207 Euphorbiinae). The sample for the ndhF analysis was smaller and included 114 species (4 oulgroups, 4 Anthosteminae, 3 Neoguillauminiinae, and 103 Euphorbiinae). Following a preliminary analysis of the ITS sequence data, a subset of the sampled taxa representing the major clades and well-supported lineages was sequenced for the ndhV analysis. In addition, 15 species were includ- ed in the ndhF analysis that were not included in the ITS analysis because of problems obtaining “clean” ITS sequences for these taxa (see Appen- dix 1). Total DNA was isolated from either fresh, silica gel-dried, or herbarium material using a modified CTAB method (Doyle & Doyle, 1987). Two genic regions were employed in the phyloge- genomic netic. reconstructions: the cpDNA coding region ndh V region (ITS). ITS amplification using the polymer- and the nrDNA internal transcribed spacer ase chain reaction (PCR) followed the procedures described by Baldwin (1992) and Baldwin et al. 1995). Amplification of the ndAF region generally — followed the protocols described by Olmstead and Sweere (1994) and Kim and Jansen (1995). The 5’ quarter of the ndhF region was excluded due to problems amplifying it. Also, a primer ca. 50 bp internal to the 3^ end “2110Ri” (5'-TCA ATT ATT CGT TTA TCA А-3”) was designed because many taxa would not amplify using primer “2110R.” Four additional primers were specifically designed for this study: (1) (5'-FTA TTC AAT АТС TYT ATG GGG TAA-3’), (2) (5'-TAA CCC CAT АҢА GAT ATT GAA TAA-35, (3) (5'-TAG GAA TTC CYT TYA ATC AA-3'), and (4) (5'-TTG ATT RAA RGG AAT TCC TA-3’). The PCR products were electrophoresed using a 458 Annals of the Missouri Botanical Garden Tabl Table 1 Characteristics of the ITS and portion of the ndhF regions included in this study. ITS ITS] 5.85 ITS2 ndhF Raw length (bp) 591-660 210-267 164-167 202-241 1467-1506 Aligned length 739 303 169 267 587 Variable sites (proportion) 527 (0.71) 272 (0.90) 35 (0.21) 220 (0.82) 740 (0.47) Parsimony-informative sites = 70 (0.64) 244 (proportior GC content, mean (range) ( (( 0.58 (0.50-0.70) 0.58 (0.46-0.71) 0.56 (0.51—0.58) 0.61 (0.48-0.76) 22 (0.13) 204 (0.76) 519 (0.33) 0.32 (0.30-0.33) 1.5% agrose gel in a 0.5x TBE (pH 8.3) buffer, stained with ethidium bromide, and then cleaned using the PEG precipitation. protocol (Nickrent, 1996). Cycle-sequencing adhered to the manufac- turers specification using the PRISM® Deoxy™ Terminator Kit (Perkin-Elmer, cle-sequencing )ye- Inc.). Cy- was followed by ethanol purification, and sequencing used an Applied Bio- systems Model 373A Automated DNA Sequencing System. Sequences were assembled from automated DNA 3.0 TS sequences sequence chromatograms using Sequencher Inc.). were initially aligned with ClustalW v. 1.4 (Thomp- (Gene Codes Corporation, son et al., 1994), using a gap costgap extension cost ratio of 10:5, followed by visual modifications; ndhF sequences were aligned visually. Because of high divergence and the large number of taxa in- cluded in the study, alignment was problematic for certain highly variable regions of ITS sequences. However, the difficulties mostly occurred aligning the major lineages of Euphorbieae to each other, and alignment within major lineages was less prob- and ndhF duced gaps into some sequences, and these sites lematic. Alignment of both ITS intro- Missing data were coded with a question mark in the matrix. Align- were included in the analyses. ment matrices have been deposited at the library of Rancho Santa Ana Botanic Garden and submit- o TreeBASE treebase.index.html). e ted (http://herbaria.harvard.edu/ The aligned ITS and ndhF sequence matrices were analyzed separately using PAUP* 4.0b4a for Macintosh™ (Swofford, 2000) on a Macintosh™ G3. Due to the large data sets, maximum parsimony using heuristic searches (Acctran, 10 random ad- dition cycles, TBR branch swapping, steepest de- scent option not in effect) was employed. Maximum likelihood estimates of transition/transversion (TI/ TV) biases were measured for both the ITS and ndhk data sets individually as implemented in PAUP* 4.0b4a under the HKY model of nucleotide substitution and using the equal-weighted parsi- mony trees. These estimates were used to produce a TI/TV step matrix, employed in further maximum parsimony analyses. This weighting scheme was employed in order to model more closely the max- imum parsimony analyses to the given data set. In- dels were coded as missing data. Multiple most par- trees combined in a strict simonious were consensus tree. Tree robustness was estimated us- ing 10,000 “fast addition” heuristic bootstrap rep- licates. RESULTS Sequence variation for both ITS and ndhF shown in Table 1. For ITS sequences, pairwise lev- ” values) for the 34.4% (between Anthostema sp. nov. and Euphorbia insulana) to 1.1% (between E. alta and E. spathulata). For sub- els of divergence (uncorrected *p entire Euphorbieae ranged from tribe Anthosteminae, levels of divergence varied from 9.4% (between Dichostemma glaucescens and Anthostema madagascariense) to 2.2% (between A. madagascariense and A. sp. nov.); for the two in- cluded species of subtribe Neoguillauminiinae (Neoguillauminia cleopatra and Calycopeplus cas- uarinoides), the level of divergence was 7.696; for subtribe Euphorbiinae levels of divergence varied from 30.9% (between E. trichotoma and E. panch- ganiensis) to 1.1% (between E. alta and E. spathu- ata). As expected, levels of divergence for ndhF sequences were much lower than levels observed in. ITS these ranged from 11.796 (between Anthostema sp. sequences. For the entire Euphorbieae, nov. and E. oaxacana) to 0.1% (between E. bilobata and E. levels of divergence varied from 2.696 (between An- exstipulata). For subtribe Anthosteminae, thostema sp. nov. and Dichostemma glaucescens) to 0.496 (between Anthostema sp. nov. and A. mada- gascariense); for subtribe Neoguillauminiinae di- vergence levels varied from 296 (between Calyco- peplus collinus and C. paucifolius) to 1.496 (between Neoguillauminia cleopatra and Calycopeplus pau- cifolius); for subtribe Euphorbiinae levels of diver- gence varied from 9.6% (between E. aphylla and Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae 459 Table 2 weighted characters of the same topology. For e Statistics from weighted maximum ee analyses of ITS and ndhF datasets and statistics for un- analysis, number and length of most parsimonious trees, consistency index (CI). retention index (RI), and rescaled consistency index (RC) are reported. Number Analysis of trees Tree length Cl RI RC ITS (weighted) 2160 8838.2 0.1813 0.7219 0.1309 ITS (unweighted) 2160 6887 0.1802 0.7126 0.1284 ndhF (weighted) 19012 1818.2 0.5221 0.7926 0.4138 ndhF (unweighted) 19012 1985 0.5441 0.7992 0.4349 E. oaxacana) to 0.1% (between E. bilobata and E. exstipulata). Transition/transversion biases were calculated to be 1.8 for the ITS data set and 0.85 for the ndhF data set. These biases were used in the weighted maximum parsimony analyses. The ITS analysis resulted in the recovery of 2160 trees of 6887 steps (equal most parsimonious weighted). The ndhF analysis resulted in the re- covery of 19,012 most parsimonious trees of 1985 steps (equal weighted). Statistics for both analyses are detailed in Table 2. The strict consensus of the 2160 trees obtained in the ITS analysis is depicted in Figures 1, 5. 7, 9, 10, 12, and 13. The strict consensus of the 19012 trees obtained in the ndhF analysis is depicted in Figures 2, 6, 8, 11. and 14. With the exception of Figures 3 and 4. phylograms are presented in Steinmann (2001). For the ITS analysis, the bootstrap 50% majority rule tree (values depicted on strict consensus tree) is structurally identical to the strict consensus in- sofar as the majority rule tree is resolved. Many of the terminal clades are supported with high boot- strap (bs) percentage values. However, there is no support for the majority of the basal internal nodes. Notable exceptions are the ancestral nodes that de- fine the tribe Euphorbieae and its subtribes. In the ndhV analysis, the bootstrap 50% majority rule tree (values depicted on strict consensus tree in Fig. 2) is structurally identical to the strict consensus with one significant difference. In the bootstrap analysis, clade A is not placed as the sister taxon of clade B but instead is located at the earliest diverging clade in a group that contains clades C and D. and this is supported by a bootstrap value of 77%. This incongruency suggests that the topology of the ma- jor clades within the subtribe Euphorbiinae should be viewed with caution. The majority of clades found in the ndhF strict consensus tree are sup- ported in the bootstrap analysis, and in contrast to ITS (see Fig. 1). many of the internal, basal nodes also possess significant bootstrap support. It is worth noting that a combined analysis was conducted for the 99 species in common between the ITS and ndhF data sets. presented here because results from the indepen- dent ITS and ndhF analyses agree strongly with each other, and the combined analysis does not pro- vide novel insights. For the instances in which tree structure resulting from the ITS and ndhF analyses differed, the combined analysis generally provided The results are not the same structure as the ndhF analysis. DISCUSSION UTILITY OF THE ndhF AND ITS REGIONS IN EUPHORBIACEAE AND A COMPARISON OF THE RESULTING PHYLOGENIES Despite being one of the largest angiosperm fam- ilies and an important floristic Component. espe- cially in tropical regions, the Euphorbiaceae have not previously been the subject of a comprehensive molecular systematics investigation. If the results obtained here are any indication, both ndhF and the ITS data appear useful in inferring relationships within the family. In general, the two reconstruc- tions of the Euphorbieae are similar, and thus cor- roborate evidence of their phylogenetic signal. The — structure of both analyses is the same (Figs. 1, 2). and when differences do occur, these gener- ally involve clades that lack bootstrap support in one or both of the There are very few instances in which well-supported clades are po- sitioned differently in the separate analyses, e.g., the arrangement of Euphorbia meenae Blatt. & McCann, E. abdelkuri Balf.f., E. drupifera Thonn.. and E. poissonii Pax (Figs. 10, 11) analyses. MONOPHYLY OF THE EUPHORBIEAE AND ITS SUBTRIBES AND PARAPHYLY WITHIN SUBTRIBE EUPHORBIINAE In both analyses, Euphorbieae form a monophy- letic assemblage with respect to the four outgroups. However, there is not high bootstrap support for this grouping in either analysis. The unusual synapo- 460 Annals of the Missouri Botanical Garden Stillingia spinulosa Figure 1. Bootstrap — greater than 50% are shown above the Webster (1994) are indicated. Individual strict consensuses for c Jade 's A-D are found in Figures 5, 7, 9, 10, 12, ar morphic inflorescence structure of the tribe there- ore supports this relationship, and it is highly probable that the tribe is indeed monophyletic. All three subtribes of Euphorbieae also form monophy- letic groups (Figs. 1, 2), and these clades possess high bootstrap support, 100% for all three in the ndh¥ analysis. Further, in both analyses subtribe Anthosteminae is the earliest diverging subtribe in Euphorbieae and subtribes Neoguillauminiinae and Euphorbiinae are sister to each other. Anthostema and Calycopeplus are strongly supported to monophyletic (bs 100% and 90% in ndhF, respec- tively), and presumably so too are the monotypic Dichostemma and — w ^ Neoguillauminia. Therefore, clade D (Widespread) = clade C ie 82 (Widespread) 30 ЕЁ clade В zm (Primarily Laurasia) " 95 92 — clade A tr (Africa and Madagascar) С ‚ 0 , 5 T Calycopeplus casuarinoides ai 210 100 (Australia) 220 55 La | W 67_ Neoguillauminia cleopatra 7 >т A (New Caledonia) z т d madagascariense | > 100 (Madagascar) — Ic gc Anthostema sp. nov. Ax 100 mA (Madagascar) ez Z 1] > Dichostemma glaucescens т (Tropical West Africa) Sapium sebiferum 100 — © — — Omalanthus populifolius С — (2 * Е © Sebastiania cornuta © 98 GF #2] Strict consensus of 2160 trees based on a weighted maximum pamimony analysis of the ITS region. ; of the classification of nd 13. branches. The tribes and subtribes Webster's overall classification of the tribe Euphor- bieae (1975) is consistent with the molecular re- sults. Within Euphorbiinae, both analyses demonstrate that Chamaesyce, Pedilanthus, Monadenium, Syn- adenium, and Endadenium are all nested within a paraphyletic Euphorbia (see clades C and D, Figs. 9-14). Also, both Synadenium and Endadenium are nested within a paraphyletic Monadenium. RELATIONSHIPS AND MAJOR LINEAGES WITHIN THE EUPHORBIINAE Both analyses support that there are four major lineages within the subtribe Euphorbiinae. here Volume 89, Number 4 Steinmann & Porter 2002 Phylogenetic Relationships in Euphorbieae 74 clade D 91 (Widespread) m 75. elade С 32 100 (Widespread) © E 86 F clade B ЫШ (Primarily Laurasia) > 100 100 clade A — (Africa and Madagascar) 2 Calycopeplus collinus © т 90 (Australia) а s c Е S az 100 Co alye е paucifolius S = (Australia en} © zz | x 62 Neoguillauminia cleopatra = = (New Caledonia) > > "Е Anthostema madagascariense 78 (Madagascar) > 2, du 100 Anthostema 2 nov. Е E (Madagascar ny =E 100 Anthostema senegalense Sm (Tropical West Africa) > Dichostemma — escens (Tropical West Afri Sapium sebiferum Omalanthus populifolius Sebastiania cornuta SdNOaADLNO Stillingia spinulosa Figure 2. Strict consensus of 19,012 trees based on a weighted maximum рагайдопу analysis of the ndhF region. Bootstrap values greater than 50% are shown above the branches. The tribes and subtribes of the — ‘ation of Webster (1994) are indicated. Individual strict consensuses for dades A-D are found in Figures 6, 8, 11, and 14. Due to the lack of a comprehensive classifica- designated clades A-D. With the exception of clade tion, many sections of Euphorbia have not been ac- B (primarily composed of the temperate herbaceous group of Euphorbia subg. Esula), none of these lin- commodated within a currently recognized subge- eages closely correspond to any previously recog- nus. Also, some sections that have been suggested nized taxon. Instead, they are conglomerates of var- to belong to a particular subgenus are demonstrated ious sections and subgenera. All of these major to not be closely related. Therefore, the following lineages possess significant bootstrap support in the discussion about the major groups of Euphorbiinae ndhV analysis (bs 74-100%), while only clade А is is organized by taxa of various ranks. supported in the ITS analysis (bs 92%). There still is some question, however, as to the exact relation- ship among these lineages. Although both the ITS were included. In both the ITS and ndhF analyses and ndhF strict consensus trees show that clades (Figs. 12, 14), these species form a well-supported, A and B are sister to each other and together these monophyletic group (bs 91% and 99%, respective- are sister to clades C and D, there is no bootstrap ly) in clade D. Species previously thought to inter- support to this grouping in either analysis. grade with Chamaesyce (see Webster, 1967), such Chamaesyce. Eight species from Chamaesyce 462 Annals of the Missouri Botanical Garden Е. gradyi E. guerichiana | clade D E. й Е. tanguahuete саде С E. trichotoma E. calyptrata | clade B E. acalyphoides ] clade A copeplus casuarinoides eoguillauminia cleopatra lense v. Dichostemma glaucescens Sapium sebife Oma tieni — Sebastiania cornuta Stillingia spinulosa — 10 changes Figure 3. Phylogram of one of the equally most parsimonious trees resulting from — ITS analysis, showing branch lengths within Euphorbieae. Only the species with the least and most changes are of subtribe Euphorbiinae. Individual а of clades A-D are given in чы жиы (2001). £ 5 Wheeler and E. peper- › share their vegetative as Euphorbia innocua L. appear similarity due to convergence, as they do not group 3 omioides Boiss.. close to the Chamaesyce clade (Figs. 9, 1: A. Mey., 1949) to be possibly inter- Similarly, E. cheirolepis Fisch. & C. sug- gested by Prokhanof — mediate between Chamaesyce and other species of — is not supported to be closely related (Figs. 12, uA these putative relatives are clearly outside 14). At least based on the molecular re- Chamaesyce, and the characters presented in the introduction can distinguish all true species of Chamaesyce. As previously suggested by Dressler (1957) and Webster (1967), Chamaesyce is derived from within £. subg. Agaloma (Figs. 12, 14). Sampling of Chamaesyce was not broad enough to determine the monophyly of the subsections pro- posed by Boissier (1862). However, it is noteworthy that С. acuta and С. angusta, both in Chamaesyce subsect. Acutae Boiss., lie sister to the remainder of the species sampled from Chamaesyce subsects. Gymnadeniae Boiss., Cheloneae Boiss., Chamaesy- ce, and Hypericifoliae Boiss. Chamaesyce subsect. Acutae is the only group within Chamaesyce not possessing the derived C, photosynthetic pathway otherwise characteristic of the genus (Webster et al., 1975; Mayfield, 1991). Al- though their composition differs between the ITS Synadenium/Monadenium/Endadenium. and ndhF analyses, a total of six species from these genera, here referred to as the Synadenium alli- ance, were included. They form a well-supported, monophyletic group (bs 10096 in both analyses, see 10, 11). Both ITS and ndhF also suggest that Monadenium is paraphyletic with Synadenium and Figs. wn for eac һ of the major clades = Euphorbia. Endadenium nested inside it. Morphologically, the genera are united in the possession of fused invo- lucral glands. They belong to a well-supported C (bs 78% in ITS, 100% in ndhF) that also contains Kuphorbia subg. Euphorbia and Lacanthis. clade Pedilanthus. As their unusual zygomorphic cy- athia suggest, the eight species of Pedilanthus in- cluded here form a well-supported (bs 100% in both analyses). monophyletic group (Figs. 9. 11). Based primarily on its involucral appendages and predominantly Mexican distribution, Dressler (1957) and Webster (1967) suggested that Pedilan- thus arose from Euphorbia subg. Agaloma, with E. fulgens Karw. ex Klotzsch as the closest potential living intermediate. However, this analysis does not support any relationship between these two groups. In fact, they do not even belong to the same major clade of Euphorbiinae, with E. fulgens (Fig. 13) and other members of subgenus Agaloma (Figs. 11—13) belonging to clade D. Beyond being a member of clade C, the exact position of Pedilanthus is not strongly supported. In the ITS analysis (Figs. 9, 10), it lies sister to the clade of Old World taxa that includes the Synadenium alliance and E. subg. Eu- phorbia and Lacanthis. However, there is no boot- strap support for the relationship. The ndhF anal- ysis (Fig. 11) suggests a relationship with E. elata in a weakly supported clade (bs 72%) of various Neotropical species. At least based on biogeogra- phy. this association is more plausible. Involucral appendages have arisen on various occasions with- in Euphorbiinae, and their presence in Pedilanthus apparently represents an independent derivation of this feature. Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae E. guerichiana E. gradyi е " |elade D E. pervilleana Ped. tithymaloides | clade C —— cleo, Anthostema sp. nov. Anthostema senegalense Dichostemma glaucescens Sapium sebiferum Omalanthus populifolius Sebastiania cornuta — 5 changes E. p E. — 5 Е.а Са * ee collinus i — clade B | clade A s opeplus paucifolius opatra Anthostema madagascariense tillingia spinulosa Figure 4. Phylogram of one of the equally most -— trees resulting А from E ndhk analysis, showing branch lengths within Euphorbieae. Only the species with the c des of subtribe и Individual phylograms of c nsi A—D are given in Ste inmann 1 (2001). E. ^ed, = Pedilanthus Cubanthus. as a section of Pedilanthus, but as noted by Mills- paugh (1913) and confirmed by Dressler (1957), these two groups do not appear closely related. Al- Cubanthus was initially described though Cubanthus was not available for this study, Euphorbia gymnonota Urb. and E. punicea Sw. were included (clade C, Figs. 9, 11). Like members of Cubanthus, these two species possess involucral glands that are partially situated on the outside wall of the involucre in contrast to the rim of the invo- lucre, as is generally the situation in Euphorbia. Based on this shared feature, an overall similarity in habit, and an allopatric distribution, Cubanthus is presumably closely related to these taxa and probably would associate with them in a molecular analysis. Pax (1921) first proposed this taxon as a section of Euphorbia, but Euphorbia subg. Trichadenia. a decade later he synonymized it with Euphorbia sect. Tithymalus Roeper, which group within section Tithymalus he believed without elaborating on that it belonged. Carter (1985) resurrected E. sect. Trichadenia Pax as a subgenus and proposed two Somalica S. Carter (inflorescences sections: sect. umbellate and unbranched. with small deciduous bracts) and sect. Trichadenia (inflorescences branching. with large persistent bracts). Later she removed her section Somalica from Euphorbia subg. Trichadenia and placed it in subgenus Ly- 1988a). Although E. subg. Trichad- Carter is now generally accepted, its clopsts (Carter, ета (Pax) S exact rank and placement are disputed. Based pre- the major = Euphorbia, ast and most changes are үп for each of sumably on the fact that many species possess a pseudoumbellate inflorescence, Gilbert (1987, 1990). for example, preferred to treat it within £. subg. Esula. Various species referred to Euphorbia subg. Tri- chadenia by Pax, Carter, and Gilbert were included in this study: Euphorbia trichadenia Pax, E. goetzet Pax, E. platycephala Pax, E. grantii Oliv., and F. omariana M. G. Gilbert. With the exception of E. goetzei, which falls out in clade D as related to E. pirottae N. (Figs. 12, 14). subgenus Trichadenia are supported to be closely related and belong to clade A (Figs. 5, 6 Terrac. the members of ). However, they do not form a monophyletic group. Instead, they belong to a group together with Euphorbia Dactylanthes (Haw.) A. Medusea Meleuphorbia A. Berger, and Anthacantha (Lem. sects. (Haw.) Baill., (Haw.) Baill., ese sections, Berger. Treista Мм A. Berger. which lack a current. subgeneric placement, all contain tuberculate-stemmed South African succulents, many of which also possess glandular involucral processes. Their various char- acteristics are as follows: E. sect. Anthacantha (leaf-reduced stems, the inflorescences axillary with the cyathia borne at the tips of persistent. spiny peduncles; representative: E. atrispina N. E. Br.): Dactylanthes highly the involucral glands with . sect. (leaf-reduced, branched, short stems, long, fingerlike divisions and curved lip at the base: representative: E. globosa (Haw.) Sims): Ё sect Treisia (leafy stems and three conspicuous bracts just below the cyathium; representative: E. clava Annals of the Missouri Botanical Garden E.o platycephala He. sg. Trichadenia НЫНЫН Е monteiri — Pseudeuphorbium : ME sg. Trichadenia E. — E. obesa E. s. Meleuphorbia E. tubiglans E. — s. Anmacantha E. clava — E. s. Tre E. — оза — —Е. sg Rhizanthi. ium E. balsamifera ssp. сл — E. sg. Esula E. meul E hellii E. caputmedusae © L E esculenta RE s. Medusea ”— E. acalyphoides — = киш ыры Le sg. Eremophyton E. — les E.a Figure 5. of the ITS region. Bootstr rap values greater than 50% are shown above the branches. E. = Euphorbia. s. = sectio 8 I 8 I — subgenus, ssp. — subspecies. Jacq.); E. sect. Medusea (main stem normally partly sunken in the ground and with few to many sec- ondary branches crowning and radiating from the apex; representatives: E. caputmedusae L. and E. esculenta Marloth); E. Meleuphorbia (little- branched to unbranched aboveground stems that are frequently subglobose and with distinct angles; representatives: E. tubiglans Marloth ex R. A. Dyer, f., and E. meloformis Aiton). The close relationship between the species of subgenus Trichadenia and these taxa is not surprising and Gilbert (1987), who noted that there does not appear to be any important discon- sect. E. obesa Hook. was predicted by tinuity between these groups. Also closely allied are E. tuberosa L., the type of E. subg. Rhizan- thium; E. monteiri Hook. f., the type of E. sect. Pseudeuphorbium (Pax) A. Berger; and E. lignosa Marloth, the type of E. subg. Lyciopsis (Boiss.) L. C. Wheeler sect. Lignosae Pax & K. Hoffm Euphorbia subg. Rhizanthium. This taxon was initially proposed by Boissier (1862) as a section of Euphorbia to accommodate various geophytes 64 — E. s. Denisophorbia Euphorbiinae clade A, strict consensus of 2160 trees based on a we ighted 1 maximum parsimony analysis } yn, sg. from India and South Africa. Boissier (1862) treat- ed another geophyte, Euphorbia sessiliflora Roxb. from Burma (not included here), in his F. Caulanthium Wheeler (1943) combined these two sections under the name Rhizanthium and elevated its rank to subgenus within Euphor- bia. Gilbert (1987) provided a detailed overview of the group, concluding that subgenus Rhizanthium, sect. Boiss. as currently circumscribed, is a heterogeneous as- semblage of many unrelated species. This study vindicates his contention, and members of this tax- on belong to various unrelated groups in clades A and C (see Figs. 5, 6, 10, 11). A number of species in Euphorbia subg. Rhizan- thium were included here. Although Euphorbia ses- siliflora, the sole member of Boissiers Euphorbia sect. Caulanthium, was not available for inclusion, the later-described and apparently closely related E. panchganiensis Blatt. & McCann and E. meenae 5. Carter were represented. Both of these Indian geophytes are nested within the spine-shield suc- culents of E. subg. Euphorbia. Euphorbia tuberosa, trichadenia — E. sg. Trichadenia E lobosa — E. s. Dactylanthes гЗ LM E. sg. Lyciopsis s. Lignosae s da el hala * nt * E. sg. Trichadenia E — E. obesa — E. s. I PN 89 E. clava — E. s. Treisia 99 :. tuberosa Lg sg. Rhizanthium 62 — — Myr E. sg. Lyciopsis s. Somalica 7 E. — ssp. adenensis — E. sg. Esula p—— E.matabelensis — E. sg. Lyciopsis s. yciopsis 100 bb E. crotonoides 100 [——- E. acalyphoides E. sg. Eremophyton — — — 4 E. s. Denisophorbia Figure 6. of the ndh¥ region. Bootstrap values greater than 50% are shown above the branches. E. subgenus, ssp. = subspecies E uphorbiinae clade A, strict consensus of 19,012 trees based on a weighted maximum parsimony — = Euphorbia, s. = section, Volume 89, Number 4 2002 Steinmann & Porter Phytogenetic Relationships in Euphorbieae the type of subgenus Rhizanthium, is not closely related to the Indian species, but instead is related to a group of South African taxa currently recog- nized as E. subg. Trichadenia (clade A, Figs. 5, 6: see discussion above). Despite not truly being geo- phytes, E. longituberculosa Boiss. and E. pirottae The — were placed here by Pax and Hoffmann (1931 first species is not closely related to any of the above members, but instead its position is near E. acalyphoides Hochst. ex Boiss. in E. subg. Eremo- phyton sect. Pseudacalypha Boiss. in clade A (Figs. 5. 6). where it was placed by Carter (1988a). The second species also shows no affinities with any of the above members and comes out related to Eu- phorbia goetzei within clade D (Figs. 12, 14). Fu- phorbia primulifolia Baker, a Madagascan species treated in subgenus Rhizanthium by Denis (1921) and in E. subg. Lacanthis by Gilbert (1987), be- longs with other Madagascan species currently placed in subgenus Lacanthis (clade С, Fig. 10). When 1862), contained only Euphorbia cuneata Vahl, a spines- Euphorbia subg. Lyciopsis. first de- seribed as a section (Boissier, this taxon cent shrub from Africa with fasciculate leaves. The section was compared with E. sect. Tirucalli Boiss. gland characters but distinguished by the fas- ciculate leaves and non-succulent stems. Pax and Hoffmann (1931) expanded section Lyciopsis Boiss. to include three subsections: subsect. Lyciopsis (as Eulyciopsis). subsect. Espinosae (Pax & K. Hoffm.) Pax € K. Hoffm., and subsect. Lignosae (Pax & K. Hoffm.) Pax & K. Hoffm. The latter two were based on sections proposed a decade earlier (Pax & Hoff- 1921). Wheeler (1943) followed Pax's in rank mann in Pax, circumscription and made changes only Carter — this taxon t ~ when elevating » subgenus. (1988a) expanded it to include an additional sec- tion, E. sect. Somalica, a group erected by her to accommodate about 10 east African species with woody semisucculent branches, small scarious de- ciduous bracts, crenulate or pectinate involucral — and large, often ornamented capsules (Cart- . 1988b); E. sect. Somalica was previously placed in E. subg. Trichadenia (Carter, 1985). Members of all four currently recognized sections of Euphorbia subg. Lyciopsis were included in this analysis. The results suggest that the subgenus is polyphyletic, and its component sections represent four distinct groups. In both molecular analyses, E. espinosa Pax and E. guerichiana Pax form a well- supported lineage (bs 97% in ITS, 100% in ndhF) in the early diverging portion of clade D (Figs. 12, 14). They are far removed from the remainder of the sections of subgenus Lyciopsis, all of which be- long to clade A (Fig. 6). Due to problems obtaining TS sequences, the other species of this subgenus were included only in the ndhF analysis. Both А. lignosa and E. hamata (Haw.) Sweet, members of subgenus Lyciopsis sect. Lignosae, are placed in a well-supported clade (bs 87%) together with E. tri- chadenia and E. globosa. Their relationship here is not unexpected; Pax and Hoffmann (in Pax, 1921), when first describing this section, noted that the appendages of E. lignosa were similar to those of species in E. sect. Dactylanthes, of which E. glo- bosa is a memb The only species of Euphorbia subg. Lyctopsis sect. Lyciopsis that was included in this study is Æ. matabelensis. It comes out as an early diverging member of clade A, sister to E. crotonoides Boiss. However, this relationship should be viewed with reservation because there is little morphological similarity between these two species and no boot- strap support for this grouping. The final section, Euphorbia subg. Lyciopsis sect. — was represented in this study by two ra cies, E. scheffleri Pax and E. — two form a well-supported group (bs 99% in ndhk, Fig. 6 ifera Aiton and various other Euphorbia, Euphorbia socotrana Balf. М in a subclade containing E. balsam- balsamifera is vegetatively similar to species of subgenus Lyciopsis sect. Somalica, so the sugges- tion that they have arisen from common ancestry is reasonable. Euphorbia subg. Esula. This subgenus largely corresponds to Boissiers Euphorbia sect. Tithyma- lus. Wheeler (1943), when implementing his system of subgenera, resurrected the application of E. subg. Esula for this assemblage. Its current use cor- responds to the circumscriptions of these authors with some subsequent modifications. For example, E. ipecacuanhae L., the type of Boissier's section Tithymalus subsect. Ipecacuanhae Boiss., has been transferred to E. subg. Agaloma, and Tithymalus subsect. Inundatae has been proposed to accom- modate the remainder of the species treated in this 1967). In addition, E. sect. Ti- Crotonopsideae Boiss. was re- subsection (Webster, thymalus subsect. moved by Radcliffe-Smith (1974) and placed subgenus Cystidospermum (Prokh.) Prokh. (see dis- cussion under Euphorbia subg. Eremophyton). As employed in the discussion below, many of Bois- sier's subsections are elevated to the rank of section E. subg. Esu- when E. sect. Tithymalus is treated as Si a ~ With as many as 500 species, this is the largest subgenus currently recognized within Euphorbia. It probably is also the most taxonomically difficult. Annals of the Missouri Botanical Garden The greatest diversity is in northern temperate re- gions. Plants are mostly perennial herbs, but a va- riety of growth forms exist from diminutive annuals to shrubs or rarely small trees. The majority of E. subg. Esula possess alternate lower leaves and a well-developed pseudoumbellate inflorescence in which the stem terminates in a whorl of leaves and a fascicle of three to many branches with opposite leaves and dichotomous branching; cyathia are sit- uated in the axils of these upper leaves. The in- volucral glands lack appendages and are often trun- cate or bicornute. Exstipulate leaves also characterize this group, but some members cur- rently placed here possess stipules, e.g., the spe- cies of subgenus Esula sect. ia (Raf.) G L. Webster and section Tithymalus subsect. Inun- datae. About 45 species of Euphorbia subg. Esula, from the majority of Boissier’s subsections, were includ- ed in this study. Although most of Boissier’s sub- sections were sampled, not enough members from each were included to make any definitive infer- ences concerning their monophyly. Still, a number of conclusions can be reached. Euphorbia ipecacuanhae, the basis of Boissier's section Tithymalus subsect. Ipecacuanhae, was in- deed supported as being unrelated to other mem- bers of this subsection (currently treated as E. subg. Esula sect. Tithymalus subsect. Inundatae) as well as unrelated to other members of subgenus Esula: thus, its removal by Webster (1967) is justified. Eu- phorbia ipecacuanhae belongs with members of subgenus Agaloma in clade D (see Fig. 13). Also justified is Radcliffe-Smith’s placement of section Tithymalus subsect. Crotonopsideae within Euphor- bia subg. Cystidospermum (see discussion under £. subg. Eremophyton). Euphorbia subg. Esula sect. Tithymalus subsect. Inundatae was represented by three South Ameri- can species: E. papillosa A. St.-Hil., E. stenophylla (Klotzsch & Garcke) Boiss., and E. thinophila Phil. These do not demonstrate any close relationship to other members of subgenus Esula. Instead, the mo- lecular evidence supports that they are related to E. peperomioides Boiss. of section Nummulariopsis and Е. Phil. castrum Boiss. Together these form a monophyletic Boiss. germainit of section Portula- group in clade C (see Fig. 9). Euphorbia subg. Esula sect. Balsamis Webb. & Berthel., a taxon corresponding to Boissier's section Tithymalus subsect. Pachycladae, is clearly poly- phyletic (Figs. 5-9, 11). This is not surprising, con- sidering that the only features uniting its members are their shrubby habit together with branches that аге leafy only toward the tips but leafless with promine nt leaf scars proximally. Euphorbia balsam- ifera subsp. adenensis (see Figs. 5, 6) does not ap- pear closely related to any other species of subge- nus Esula sect. Balsamis sampled here. Instead, it belongs to clade A and based on the ITS results (Fig. 5), Schwartz from Yemen, a vegetatively similar spe- it has its affinity with E. meuleniana О. cies. Euphorbia plumerioides Teijsm. ех Hassk., also referred to this group by Boissier, shows no close relationship to E. balsamifera but instead be- longs to clade С (Fig. 9). Euphorbia plumerioides is a member of a group of about eight species occur- ring in Australia, Malesia, and Melanesia (Foster, 1994). In this analysis, the Hawaiian species, E. haeleeleana D. R. Herbst, is suggested to belong to this group. Also related is E. boóphthona C. A. Gardner, an herbaceous plant from Australia pre- viously treated in subgenus Eremophyton. Both E. balsamifera and members of the E. plumerioides group differ from typical members of subgenus Esu- la by lacking the characteristic pseudoumbellate inflorescence described above. A third group of species sampled from subgenus Esula sect. Bal- samis, E. dendroides L., E. longifolia Lam., E. re- el., and Ё. are demonstrated by the molec- gis-jubae Webb. atropurpurea Brouss. ex Willd., ular evidence to be related to typical members of the north temperate group of subgenus Esula dis- cussed below (see Figs. 7, 8), although collectively these do not form a monophyletic group. Another group that does not appear related to the remainder of Euphorbia subg. Esula is its section Adenorima. This taxon corresponds to Boissier's section Tithymalus subsect. Laurifoliae; another synonym is the genus Euphorbiodendron Millsp. As the last name suggests, the ca. 20 species compos- ing E. subg. Esula sect. Adenorima are usually trees. They occur primarily in tropical forests from Mexico to northern South America and the Carib- bean and possess the pseudoumbellate inflores- cence structure typical of subgenus Esula. It is for this reason that they traditionally have been placed here. A number of species presently referable to subgenus Esula sect. Adenorima were included in E. cestrifolia HBK, E. laurifolia Juss., E. elata Brandegee, E. tanquahuete Sessé & Mociño, E. calyculata HBK, and E. gym- nonota Urb. (Figs. 9, 11). These taxa do not appear this study: E. punicea Sw., closely related to the core Esula group discussed below, and the two groups are in separate major C and B, tively. In addition, the species of subgenus Esula sect. Adenorima do not form a monophyletic group, clades of Euphorbiinae, clades respec- but instead appear to represent an ancestral New World grade from which various other groups of Eu- Volume 89, Number 4 Steinmann & Porter 2002 Phylogenetic Relationships in Euphorbieae 80 92 = Schimper Г Е. sg. Tirucalli E. usambarica 91 E. aphylla E. regis-jubae E. atropurpurea 100 E. peplus 100 E. е Е. — E. m E. megalatlantica E. ro Е. turczanino 72 = E. amygdaloides : E. soongari m 7 — — E. esula E. sg. Esula 99 E. disco lL. E krausiana E. calyptrata 54 r— Е. lathyris LL E. myrsinites 100 E. spathulat — E.al E. acanthothamnus E. oblongata E. depauperata 100 ro Е. pilosa | L— E. longifolia E. stricta Figure 7. Euphorbiinae clade B, strict consensus of 2160 trees based on а weighted maximum parsimony analysis of the ITS region. Bootstrap values greater than 50% are shown above the branches. E. = Euphorbia, sg. = subgenus. phorbia have evolved. Dressler (1957) suggested that this taxon includes the most primitive members of the strongly supported here, the species do represent genus. Although that assumption is not some of the earliest diverging taxa in clade C and possess many traits that appear primitive for the genus. In both the ITS and ndhF analyses, the remain- ing species of Euphorbia subg. Esula sampled come out together (see Figs. 7, 8). This group c orresponds to the following subsections of Boissiers section Ti- thymalus: subsect. Decussatae Boiss., subsect. Op- Carunculares Boiss., positifoliae Boiss., subsect. subsect. Galarrhaei Boiss., subsect. Esulae Boiss., and subsect. Myrsiniteae Boiss. They form a well- supported, monophyletic assemblage (bs 86% in ndhV) only after the inclusion of the third group of species discussed above in subgenus Esula sect. Balsamis and the inclusion of E. schimperi Presl and A. been placed in E. subg. Tirucalli, but it should be mauritanica L. The latter two species have emphasized that their association with Æ. tirucalli „ and relatives is based primarily on gross. mor- phology; the possession of pseudoumbellate inflo- rescences and the lack of stipules supports their placement in subgenus sula, as the molecular data here strongly suggest. Collectively the taxa mentioned in the previous paragraph compose clade B and correspond broadly Г Euphorbia subg. Esula. The presence of Euphorbia to the temperate, Northern Hemisphere group o in temperate regions is almost entirely the result of the radiation of this lineage, and very few other Al- though the great majority of this group are restrict- species of Euphorbia are found in such areas. ed to temperate, Northern Hemisphere environ- ments or high-elevation montane tropical regions, this mauritanica and relatives are found in arid tropical is not a strict characterization. Euphorbia E E. dendroides, E. longifol- and subtropical regions; ia, E. regis-jubae, and E. atropurpurea occur in subtropical vegetation on the Canary Islands; and E. tropics. trichotoma inhabits beaches in the New World . aphylla atropurpurea ue 74 Al ats intica ped op esula HE. sg. Esula myrs unite anthothamnus | : Spath ule . Le meifolia gure 8. of us ndhY region. Bootstrap values greater than 50% are shown above the branches. E. Euphorbiinae clade B, strict consensus of 19,012 trees based on a weighted а — analysis bg pilose = Euphor = subgenus. 468 Annals of the Missouri Botanical Garden ‚10 96 d tehuacanus 96 Ped. bracteatus Р, carpus 100 Ped. connatus 7. thinophila — Esula е qu — БЕ 5. Nummulariopsis кенер, phyla g. sg. Esula \ po — E. s. Portulacastrum ete ‹ ollmeriana E. s. Stachydium agunillarum © 51 — (3) E. s. Pteroneurae г. pteroneura (2 аня (1) es. — — E. s. Euphorbiastrum ueri —R. s. оаа E oia |- E. sg. Esula unicea . calyculata E. sg. Esula Jesi iris — E. sg. Eremophyton Figure 9. E cp clade С (in part), strict consensus of 2160 trees based on a weighted maximum parsimony analysis of the ITS ау, values greater than 50% are shown above the branches. E. = Euphorbia, Ped. = Pedilanthus, s. = s i Em sg. — subgenus, ssp. — subspecies E. panchganiensis Le. sg. Rhizanthium . meenae 7. teke Г. — E. sg. Euphorbia 1 1 [ E E E -qpiphloides E. abdelkur. E і i г primulifolia E. г. perr .mi H і . ankarensis I— E. sg. Lacanthis ii . capmanambatoensis . iharanae E . hedyot oides | А E. mahabobokensis [—E. s. Denisophorbia E - elliotii г. alluaudii —E. s. Deuterocalli Syn. grantii fon. elegans Endadenium/ nd. gossweileri Monadenium fon. magnificum | | Synadenium — brunellii He. sg. Lacanthis Figure 10. о clade С (in part), strict consensus of 2160 trees based оп а we id d maximum parsimony analysis of the ITS region. Bootstrap values greater than 50% are — above the branches. E. = Euphorbia, End. = Endadenium, Mon. = Monadenium, Syn. = Synadenium, s. = section, sg. = subgenus. Volume 89, Number 4 Steinmann & Port 2002 Phylogenetic ае in Euphorbieae . meenae — E. sg. Rhizanthium . abdelkuri . drupifera Le sg. Euphorbia . poissonii . brunellii ` rubella I- E. sg. Lacanthis n. grantii | Endadenium/ тїит/ рей * vides sg. Lacanthis Дате. — ^s eye pen — — E. s. Deuterocalli — E. s. Portulacastrum {эе hila — E. sg. Esula :. lactiflua . calyculata 7. tanguahuete | — E. sg. Esula г. punicea * pteroneura (1) — E. s. Pteroneurae do niana — E. s. Euphorbiastrum eberbaueri — E. s. Pteroneur [ | 100 eH sip — E. sg. Esula "ed. tehuacanus Ред tithymaloides * Pedilanthus 100 ata — E. sg. Esula m.» 100 PL 6 ; :lagunillarum Le s. Stachydium ا‎ поа — Е. s. Pteroneurae 100 г eleele === оета | sg. Eremophyton 100 [—— E. perville — E. s. Denisophorbia ڪڪ‎ | E pisadas EE s. Tirucalli Figure 11. Euphorbiinae clade С, strict consensus of 19,012 trees based on a weighted maximum — analysis of the ndhF region. Bootstrap values | greater than 50% are shown above the branches. E. = Euphorbia, End. = Endadentum, Mon. ian Ped. = Pedilanthus. Syn. = Synadenium, s. = section, sg. = subge At least from the limited sample examined in this site, asymmetrical leaves and elongate, glandular study, the primary division within the restricted cir- stipules. These features together with axillary, 4- cumscription of Euphorbia subg. Esula, discussed glanded involucres are reminiscent of Chamaesyce, in the previous two paragraphs, is between those to which Wheeler (1943) believed that this section species that possess tuberculate ovaries and those belonged. Webster (1967) noted that E. peperom- species whose ovaries are smooth. It remains to be ¿oides bordered Chamaesyce. In this analysis, E. pe- | crease in sampling. If so, this would have important is instead related to E. stenophylla, a South Amer- seen whether this distinction will withstand an in- peromioides shows no affinity with Chamaesyce. taxonomic implications because many of Boissiers ican member of subgenus Esula sect. Tithymalus subsections contain both tuberculate- and smooth- subsect. /nundatae (Fig. 9). Although in habit Æ. ovaried species. peperomioides and E. paranensis differ greatly from It is worth emphasizing that the occurrence of a оћег members of this subsection, a close relation- pseudoumbellate inflorescence in all of the major ship is suggested by the common possession of su- lineages of Euphorbia suggests this is а symple- bulate, sepal-like lobes below the pistillate flowers, siomorphic feature that was present before the dif- a feature otherwise very rare in subtribe Euphor- ferentiation of subtribe Euphorbiinae. The struc ture biinae. bears some resemblance to the inflorescence of : | : . ; Euphorbia sect. Portulacastrum. Two species, Neoguillauminia and may be homologous. There- . | Dn | 5. . : PI — Euphorbia germainii and E. pentlandii Boiss. (the fore, the possession of this type of inflorescence : | : . . à . — mE latter not included), are contained in this section. should not be viewed as the defining characteristic ‘a ban Ы — n . Both are South American annuals with cleft invo- of E. subg. Esula. Instead, it is the combination of ; lucral appendages. Based on the presence of these a pseudumbellate inflorescence together with ex- involucral — посаде, section Portulacastrum has stipulate leaves that better characterizes the core been placed in E. subg. Agaloma, e.g., Wheeler subgenus Esula group, as represented by clade B. (1943). However, according to the molecular data, Euphorbia sect. Nummulariopsis. Euphorbia at least E. germainii does not fall out as related to peperomioides and the very similar E. paranensis the core Agaloma group. Instead, it nests within Dusén (the latter not included) are the only two South American members of subgenus Esula sect. species belonging to E. sect. Nummulariopsis. Both Tithymalus subsect. /nundatae in clade С (Fig. 9). are prostrate Brazilian perennial herbs with oppo- This placement is anomalous, and E. germainii and Annals of the Missouri Botanical Garden E. pentlandii are morphologically incongruous there because they possess well-developed involu- cral appendages and lack the characteristic sepal- like lobes below the pistillate flowers of the sub- section /nundatae. Euphorbia sect. Denisophorbia. This is a small group of approximately 20 species of leafy trees and shrubs, mostly confined to Madagascar. It was first proposed as a subsection of Euphorbia sect. Eu- phorbia by Leandri (1957). Croizat (1972) elevated the group to the rank of section. As mentioned by Leandri (1957), section Denisophorbia is difficult to define. The leaves are entire and alternate to spi- rally arranged. The cyathia are relatively large, lack appendages, and are solitary or in terminal pseu- doumbellate inflorescences. Seven species belong- ing to this group were included here: Euphorbia antso Denis, E. denisti гаа E. elliotii Lean- dri, E. hedyotoides N. . E. mahabobokensis Rauh, E. pervilleana Baill., and E. tetraptera Baker. These species do not form a monophyletic clade but instead come out as representing three separate groups. Euphorbia antso is the earliest diverging species in clade A (Figs. 5, 6) and shows no affinity with the other species of E. sect. pie, edge sampled here, all of which belong to clade C (Fi 9-11). Euphorbia denisii, E. pervilleana, and E. te- traptera form a monophyletic group that is sister to E. tirucalli and relatives, currently treated in sub- genus Tirucalli. Not closely related to this group are E. elliotii, E. hedyotoides, and E. mahabobok- ensis. These form a monophyletic clade sister to subg. Lacanthis proper. Euphorbia sect. Denisophorbia was proposed to be the most primitive group in Euphorbia (Webster et al., 1982). In part, this may be correct because Euphorbia antso is the earliest diverging species of clade A and possesses many of the primitive fea- tures for the subtribe (see discussion below under origin and biogeography of Euphorbieae). Also, with regard to ITS and ndhF molecular a this species possesses the least amount of genetic divergence in relation to the outgroup taxa (see Figs. 3, 4). Euphorbia subg. Tirucalli. The section Tirucal- li Boiss. was proposed in Euphorbia to accommo- date arid-adapted shrubs with long, slender, semi- succulent branches (Boissier, 1862). The leaves are reduced and the stem is green and photosynthetic. The section was elevated to subgenus by Carter (1985), and she later noted that it contains two well-defined groups (Carter, 1992a). The first group ы corresponds to E. tirucalli and relatives and is characterized by small scarious bracts, tightly con- gested inflorescences, and glandular stipules. The second group corresponds to species such as F. mauritanica and is characterized by leafy bracts, pseudoumbellate inflorescences, and a lack of stip- ules. Various members of Euphorbia subg. Tirucalli were included in this analysis, and the two groups recognized by Carter do not appear closely related. Instead, their similarities in growth form appear to have resulted from convergent evolution. Euphorbia irucalli and relatives (represented here by E. ar- buscula Balf. f., E. gregaria Marloth, and E. xylo- phylloides Brongn. ex Lem.) come out in clade C as sister to a group of leafy shrubs from Madagascar that are currently treated in section Denisophorbia (Figs. 10, 11). The remainder of the species of sub- genus Tirucalli (represented here by E. mauritanica and E. schimperi) are found nested within the north- temperate group of subgenus Esula in clade B (Fig. 7). The leafy bracts, pseudumbellate inflorescenc- es, and lack of stipules of these latter plants sub- stantiate this placement. Euphorbia lactiflua Phil. ex Boiss., a shrub from the deserts of Chile and the only New World spe- cies referred to this group (Boissier, 1862), is not related to any other species of the subgenus (Figs. 9, 11). In fact, its initial placement by Boissier is in itself very peculiar because this species is a leafy, scarcely succulent shrub. Euphorbia lactiflua is taxonomically isolated and shows no close rela- tionship with any other species of Euphorbia. In this analysis, its affinities are not determined with precision, but belongs to clade C, in a group of various Neotropical Euphorbia. Euphorbia sect. Euphorbiastrum. This taxon was first erected as a genus by Klotzsch and Garcke (1860) to accommodate their new species, Euphor- biastrum hoffmanniana Klotzsch & Garcke. Bois- sier (1862) then reduced Euphorbiastrum Klotzsch & Garcke to a section of Euphorbia. Its most dis- tinctive feature is that the involucres are solitary in the leaf axils and subtended by a condensed spiral of small, imbricate bracts. In this analysis, Eu- phorbia hoffmanniana (Klotzsch € Garcke) Boiss. occurs in clade C where it is related to E. weber- baueri Mansf. and E. pteroneura A. Berger in a well-supported group (bs 91% in ITS, 9896 in ndhF) together with E. cestrifolia (Figs. 9, 11). Ac- cording to the ITS evidence (Fig. 9), E. laurifolia is the basal member of this assemblage, but there is no bootstrap support for its placement. Members of this group are morphologically quite different. Euphorbia cestrifolia and E. hoffmanni- ana are leafy shrubs; E. weberbaueri is a leaf-re- Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae 471 duced, stem-succulent shrub; and E. pteroneura is a leafy, drought-deciduous herbaceous succulent. Despite their gross differences in habit, some fea- tures suggest relationships among these plants. For both £. complex, one, cestrifolia and E. weberbaueri possess well-developed glandular stipules. As well, in all taxa the cyathia are closely subtended and sometimes enclosed in a pair of bracts, the styles are entire or emarginate, and the seeds are similar in size and shape and possess a minute car- uncle. The involucres of E. weberbaueri and E. pter- oneura possess four glands while those of E. cestri- folia and E. hoffmanniana possess five glands. Euphorbia pteroneura was the basis for E. sect. Pteroneurae A. Berger (1906). Other taxa have been placed in this section, e.g., E. sipolisii and E. phos- phorea, but a close relationship between these and E. pteroneura is not supported here (Figs. 9. 11). Given the close affinity of E. hoffmanniana and E. pteroneura, section Pteroneurae is best reduced to synonymy under E. sect. Euphorbiastrum. Euphorbia sect. Stachydium. This section is characterized by a congested, primarily monocha- sial inflorescence on which the pairs of bracts are closely imbricate and fold over to obscure the in- volucres within. There are about five species in South America and one species, Euphorbia phyl- (1862). sect. Stachydium Boiss.. loclada Boiss.. in Namibia. Boissier who first described £. recog- nized two subsections: subsect. Americanae Boiss. (for the American species) and subsect. Capensis Boiss. (for E. phylloclada). Five species were in- cluded here: E. comosa Vell. E. gollmeriana Klotzsch ex Boiss., E. heterodoxa Müll. Arg.. E. lagunillarum Croizat, and E. phylloclada. ln this analysis, the two subsections are placed in different major clades of Euphorbiinae (clades С and D. spectively), and a close relationship between them is not supported. The similarity between the two subsections is due mostly to the unusual architec- ture of the inflorescence. Pax (1921) treated Eu- phorbia phylloclada as a member of E. sect. Pseu- dacalypha, based presumably on the axillary cyathia, but there is no affinity between this species and other members of Pseudacalypha either. Based on the possession of involucral append- ages, Euphorbia sect. Stachydium has been sug- gested to belong to subgenus Agaloma (Wheeler, 1943). However, only E. phylloclada occurs within this group (see discussion under subg. Agaloma). The South American species are members of clade C (Figs. 9, 11) and form a well-supported group (bs 93% in ITS, 100% in ndhF) that is sister to the £. phosphorea complex, a small assemblage of leafless. stem-succulents from eastern Brazil. Beyond its as- sociation with the E. phosphorea complex, the re- — of section Stachydium to other members ‘the genus is obscure, although the ndhF results E that it belongs to a weakly supported clade of various Neotropical taxa, including Pedilanthus and subgenus Esula sect. Adenorima. the most recent modifications in Euphorbia classifica- tion (Gilbert, 1987; Carter, 1988a). the subgenus Euphorbia is restricted to about 250 Old World stem succulents in which the base of each leaf is Euphorbia subg. Euphorbia. Following surrounded by a callous pad, termed a spine-shield. bearing a pair of spiny outgrowths and stipules modified into spines. It corresponds to Haworth’s (1812) and Klotzsch and Garcke's (1859, 1860) re- stricted genus Euphorbia. Boissier treated these milii Des Moul. (= E. splendens Boj. ex Hook.) under his E. sect. Dia- Although Bentham's (1880) and Pax and Hoffmann's (1931) classifications recog- ~ species together with E. canthium Boiss. nized section Diacanthium in the same sense as Boissier, it was reduced to a subsection of section Euphorbium Boiss., a taxon these authors used to accommodate essentially all of the succulent Ku- Wheeler (1943) followed Ben- tham’s and Pax and Hoffmann’s circumscription but Tithymalus: the name Tithymalus has otherwise usually been phorbia species. under the misapplied name E. subg. applied to the north temperate members of E. subg. Esula. Webster (1967) essentially followed Wheel- ers classification but correctly applied the name £. subg. Euphorbia to this assemblage. In both the ndhF and ITS analyses, Euphorbia subg. Euphorbia belongs to a well-supported clade (bs 78% in ITS, 100% in ndhF) together with sub- the 11). Within this clade, all of subgenus genus Lacanthis and Synadenium alliance (Figs. 10, Euphorbia sampled form a monophyletic group, but only after the inclusion of a few additional taxa. Elaeophorbia, an African group of four species with drupaceous fruits, has been recognized as a genus by some (e.g.. Carter, Webster, E. drupifera Thonn. and well supported 19884) and a section by oth- ers (e.g.. 1967). It was represented in this study by (bs 100% in both analyses) to be nested within sub- 11). meenae are also strongly supported (bs 100% in ITS) to be nested within subgenus Euphorbia (Figs. 10, 11). Both of these genus Euphorbia (Figs. 10, In addition, £. panchganiensis and Е. species are dwarf geophytes from India that are very different in appearance from the typical mem- bers of the subgenus. At least in the mature phase. thev lack the characteristic spine-shield structure. 472 Annals of the Missouri Botanical Garden Euphorbia panchganiensis is apparently closely al- lied to E. fusiformis Hamilton ex D. Don, which was placed by Boissier in E. sect. Rhizanthium (= E. subg. Rhizanthium sensu Wheeler, 1943). Based on almost identical capsules and seeds, Gilbert (1987) first suggested that E. fusiformis might have evolved from Asian members of subgenus Euphorbia. The molecular evidence supports his suspicions. The reduction to geophytic herbs has been documented in at least one African lineage of subgenus Eu- phorbia (Carter, 1994), and it appears that the In- dian geophytes represent a parallel derivation of this growth form from spiny shrubs. The Indian geo- phytic species should be examined at early onto- genetic stages to determine if indeed spine-shields are present in the seedlings. Carter (1994) recognized two sections and many subsections in her classification of Euphorbia subg. Euphorbia, but not enough taxa were included in this study to test the validity of these groups. Euphorbia subg. Lacanthis. Lacanthis Raf. (Rafinesque, 1837) originally contained only La- canthis splendens Raf. (= Euphorbia тий). The name slipped into obscurity until Gilbert (1987 resurrected it at the rank of subgenus, applying it to a group of species from Madagascar previously — treated together with the spine-shield taxa of E. subg. Euphorbia, sensu stricto, discussed above. Gilbert also included within subgenus Lacanthis (Raf.) Baill. ex Boiss. and the subg. Rhizanthium. As discussed in detail (Gilbert, 1987), there are numerous differences that suggest these species should be treated separately from the Gilbert the species of E. sect. Goniostema Madagascan members of E. narrowly defined subgenus Euphorbia. For exam- ple, in subgenus Lacanthis the inflorescences are much branched (vs. little branched), the bracts are well developed (vs. greatly reduced), and the seeds are oblong-cylindrical (vs. ovoid to subglobose). Additionally, in subgenus Euphorbia the spines are borne on a differentiated spine-shield and the stip- ules are represented by prickles just above the leaf, but in subgenus Lacanthis the spines are actually the stipules and a spine-shield is absent. Besides the subgenus Lacanthis a few geophytes from tropical Madagascan species, Gilbert also included in east Africa that are morphologically very similar to some of those from Madagascar. A broad array of species from Euphorbia subg. Lacanthis were sampled. included E. milit, E. pedilanthoides Denis, E. got- tlebei Rauh, Е. rossii Rauh & Buchloh, E. primu- lifolia, E. thouarsiana Baill., E. perrieri Drake, Е. ankarensis Boiteau, E. millotii Ursch € Leandri, Those from Madagascar geroldii Rauh, E. шии Rauh, and E. tharanae Rauh. From tropical rica, Ё. ru- ella Pax and E. brunellii Choiv. ex Chiarugi were represented. All taxa belong to clade C in a sub- clade containing subgenus Euphorbia and the Syn- adenium alliance (Figs. 10, 11). The tropical east African and the Madagascan taxa belong to differ- ent lineages within this subclade. The Madagascan members of subgenus Lacanthis are supported as a monophyletic assemblage belonging to a clade also containing species not previously associated with the group. Euphorbia alluaudii Drake, a Madagas- can endemic very similar in habit to E. tirucalli but treated by Croizat (1972) in his E. sect. Deuterocalli Croizat, represents the earliest diverging species within the group. Another group not previously as- sociated with subgenus Lacanthis is the E. hedyo- toides complex of E. sect. Denisophorbia. In this analysis, the complex forms the sister clade to sub- genus Lacanthis proper. It is suggested that the origin of Euphorbia subg. Lacanthis, sensu Gilbert, is separate from that of subgenus Euphorbia and spinescence and succu- lence in these two groups have resulted from in- dependent derivations. Euphorbia sect. Arthrothamnus. Klotzsch and Garcke (1860) first proposed this group as a genus to accommodate Euphorbia tirucalli and seven spe- cies from the Cape Region of South Africa. Boissier 1862) later treated it as Euphorbia sect. Arthro- thamnus. He removed F. tirucalli but expanded the group to include two species from the West Indies. —. The latter two were treated within his Ё s throthamnus subsect. Americanae Boiss., while the remainder of the Old World taxa were placed in section Arthrothamnus subsect. Capenses Boiss. As discussed further under Euphorbia subg. Agaloma, the two subsections of Arthrothamnus do not appear closely related, and this section should be restricted to about 20 species in South Africa and Namibia. These are dioecious, dichotomously branching shrubs with photosynthetic, articulate branches and small, opposite leaves. Two species of the group were included here, E. juttae Dinter and E. rhombifolia Boiss., and the close relation- ship of both of these is well supported (bs 100% in both analyses, see Figs. 12, 14). They belong to clade D and represent an early diverging lineage of this clade. Further relationships of these species to other Euphorbia are unclear. Euphorbia subg. Eremophyton. This group was first erected by Boissier (1862) as a section of Eu- phorbia to include Euphorbia eremophila A. Cunn., and E. E. agowensis Hochst. ex Boiss., gueinzii Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae 100 82 91 Figure 12. analysis of the ITS region. Bo = Chamaesyce, s. = on, sg. = subgenus A — 13. simony ees of the I section, sg. — subgenus. Fig. 13 — tyla E. sg. Poinsettia E. sg. Agaloma lobata s. Zygophyllidium E. sg. Poinsettia sg. A E.s s. ا‎ УТЯИ Chamaesyce . polycnemoides angusta a E > — "Odes gymnc clada чапа Ез е ا‎ . Sp. nov. сеа КЕ. sg. Cystidospermum E. sg. Eremophyton go ns ee homi lada Е. s. Stachydium rombolta E. s. Arthrothamnus E. sg. — — is spin s. Es, * d . scatorhiza г. platyclada `. goetzei — E. sg. Trichaden sis | sg. — vton 7. Od. » FOS SUC 7. inn а es 7. aff. Euphorbiinae clade D (in part), strict consensus of 2 ITS region. Bootstrap values greater than 50% are shown above the branches. E. = siana ssiana var. nov . ceroderma . antis pue a aron-rossi 7. ipe Ж ре ‘olor 7. cassythoides `. equi Sora . misella . macropus BA di za . pirottae— E. sg. n Euphorbiinae clade D (in part), strict consensus of 2160 trees based on a weighted maximum parsimony strap values greater than 50% are shown above the branches. £. = Euphorbia, Cham. xacana ше — 5 ‘ola E. sg. Agaloma s. Alectoroctonum/ Cyttarospermum/ — Petalom E: — pei ula ithymalopsis/ Сой Trichosterigma j. SEQOVIENSIS г. leucocephala a a осна ас — — E. s. Arthrothamnus unensis T n — E. sg. Agaloma s. Cyttarospermum/ ichili Tithymalopsis 2160 trees based on a weighted maximum par- Euphorbia, Annals of the Missouri Botanical Garden E. gradyi Е. delicatula A 3 ое иа E. sg. Agaloma E. e - s. Alectoroctonum/ 67 100 E. sinaloensis vitarospermum/ 94 90 E ms — Dichilium/ E. mis Tithymalopsis/ Е. apes ua Trichosterigma E. sp. no E. sphaerorhiza E. oerstedia E. SÍFIgOsc 99 E. — — E. pumicola 7 E А H— E. sg. Poinsettia E. chersonesa 2. Sg. : E. hormorhiza E. heterophylla E. pulcherrima گے‎ E. Perd dod E. bilobata E. sg. Agaloma E. bifurcata А eriantha s. ZY apko ldu 93 E. jali 99 — Cham. “articulate [hago Р L— Cham. ¢ sg. Agalor , appariciana — e a s. К оа id 100 r— E. glanduligera lt E —— laa — E. s. Stachydium — E. sg. Trichadenia m E. pi — wie a — тугоп 66 29 E. petiolata — E. se. € 'vstidospermum E. tannensis —E. sg. Eremophyton 35 E Veo ine 2: - 100 "c E. scatorhiz E. sg. —— . polya — | 100 l E. guerichiana E. sg. Lyciopsis E. espinosa s. Espinosae 100 — а n i й olia — [ge — ibifoli Fe s. Arthrothamnus Figure 14. Euphorbiinae clade D, strict consensus of 19,012 trees based on a weighted maximum parsimony analysis of the ndhF region. Bootstrap values greater than 50% are shown above the branches. E. = Euphorbia, Cham. = Chamaesyce. Boiss. It was originally circumscribed to contain herbs or subshrubs with alternate lower and oppo- site upper stem leaves; glandular or subulate stip- ules; axillary or terminal, solitary involucres with four to five glands lacking appendages; and carun- culate to ecarunculate seeds. Carter (1985) empha- sized also petiolate leaves and exserted capsules as characteristics of subgenus Eremophyton. As these traits suggest, there is no single synapomorphy that unites this group, and all of the features used to delineate the subgenus can be found in various combinations in other taxa of Euphorbia. In 1880, Bentham synonymized Boissier's E. sects. Bongium Boiss., Cheirolepidium Boiss., and Pseudacalypha under his expanded section Eremophyton, and this system was followed by Wheeler (1943) when he elevated the section to subgenus. Euphorbia sect. м Cheirolepidium was removed by Prokhanov (1933 and formed the basis of his genus Cystidopermum Prokh. Cystidospermum was later reduced to a sub- genus of Euphorbia (Prokhanov, 1949), and Rad- cliffe-Smith (1974) agreed that subgenus Cystidos- sufficiently distinct from the permum was remainder of E. subg. Eremophyton to warrant its separate recognition. He additionally referred to it Boissiers E. sect. Tithymalus subsect. Crotonopsi- deae. Two species of Euphorbia subg. Cystidospermum (sensu Radcliffe-Smith, 1974) were included in this study: E. cheirolepis and E. petiolata Banks & Sol. Various species of subgenus Eremophyton proper were also included: E. boophthona, E. pirottae, E. agowensis, E. scatorhiza S. Carter, E. cheirolepis, E. eremophila, E. tannensis Spreng., E. polyantha Pax, phoides. In the analyses these taxa do not group crotonoides, E. longituberculosa, and E. acaly- together and are widely scattered throughout the 6, 9, 11, 14). Therefore, subgenus Eremophyton, as currently rec- subtribe Euphorbiinae (Figs. 5, ognized, is highly polyphyletic. three species of Euphorbia subg. Eremophy- ton sect. Pseudacalypha included in this study (£. acalyphoides, E. crotonoides, and E. longitubercu- losa) all grouped as closely related members of clade monophyletic group but instead a grade of taxa A. However, they did not form a strictly (Figs. 5, 6). Interestingly, E. matabelensis (a mem- ber of subg. Lyciopsis sect. Lyciopsis) also occurred together with the species of subgenus Eremophyton sect. Pseudacalypha as sister to Euphorbia croto- Volume 89, Number 4 2002 Steinmann & Porter 475 Phylogenetic Relationships in Euphorbieae noides in the ndhF analysis, although there is no bootstrap support for this relationship and the two are morphologically very different. One of the more unusual results of this study is the placement of Euphorbia boóphthona, an Aus- tralian member of subgenus Eremophyton, in a 94% in ITS) together with E. plumerioides and E. haeleeleana, two ar- strongly supported clade (bs borescent Pacific Island taxa presently treated in subgenus Esula sect. Balsamis. In most respects, E. bodphthona closely resembles other Australian members of subgenus Eremophyton, e.g.. E. tan- nensis and E. eremophila, and these Australian taxa have been thought to represent a natural group (Hassall, 1977). It is worth noting that according & Hassall (1977), E. two species as well as the > boóphthona differs from these other Australian mem- bers of the subgenus in being an octoploid (n = 28), but no tetraploid has yet been found in this subgenus. Radcliffe-Smith's (1974) decision to unite Bois- siers Euphorbia sect. Tithymalus subsect. Croton- opsideae (represented here by E. petiolata) with subgenus Cystidospermum (represented here by А. cheirolepis) was supported, and the two species are sister to each other (bs 92% in ITS, see Fig. 12) in this analysis. They belong to a clade also contain- ing two members of E. subg. Eremophyton proper, E. tannensis and E. eremophila. Thus, subgenus Cystidospermum probably is best treated as a syn- onym of subgenus Eremophyton. The Madagascan endemic E. platyclada Rauh, whose relationship with other Euphorbia was previously unknown (Rauh, 1998), also groups with these taxa in the ndhV analysis, but this relationship is not well sup- ported (Fig. 14). In the ITS analysis (Fig. 12). F. platyclada comes out with E. As suggested by Carter (1992b), these latter two species are sister taxa in so far as this scatorhiza and Е. polyantha. sample is concerned. Many species of Euphorbia subg. Eremophyton possess a great similarity to those of subgenus Aga- loma, especially taxa of its section Zygophyllidium Boiss. For example, contrary to reports in the lit- erature, the southwest Asian E. petiolata and E. cheirolepis actually possess involucral appendages. Undoubtedly, if these species occurred in the New World, they would be treated within subgenus Aga- loma without question. Therefore, the molecular ev- idence that E. subg. Agaloma has either evolved from a portion of subgenus Eremophyton or that the two are sister taxa and share common ancestry is plausible. Euphorbia subg. Agaloma. This taxon is based on another one of Rafinesque’s genera, Agaloma af. It was first erected (Rafinesque, 1838) to ac- commodate Euphorbia corollata L. and two related species. Interestingly, Rafinesque also published six other genera that correspond to the current cir- cumscription of this subgenus (Aklema Raf.. Lepa- dena Raf., Peccana Raf., Petaloma Raf., Raf., and Zalitea Raf.). Euphorbia subg. Agaloma 4). At was first treated at this rank by House (192 Vallaris that time it only accommodated E. corollata L., and other members of modern subgenus Agaloma were placed in subgenus Lepadena (Raf.) House. Wheel- er (1943) was the first to adopt subgenus Agaloma in its current circumscription, a concept corre- sponding to E. sect. Adenopetalum Boiss., sensu Bentham (1880) and Pax and Hoffmann (1931). The group is broadly defined to contain New World members of Euphorbiinae with petaloid involucral appendages but excludes species of Chamaesyce and Pedilanthus. Beside the presence of involucral appendages, little else unites all members of the group. Trees, shrubs, perennial herbs, geophytes, annuals, and stem-succulents are represented. The subgenus comprises about 150 species and is con- difficult 1975; ). In addition, it has frequently sidered taxonomically (Johnston, Buck & Huft, 1977 been suspected of being paraphyletic, Chamaesyce and E. subg. Poinsettia nested within with both = Euphorbia subg. Ава ота was the best-sampled group in this study. Fifty-six species from all of its recognized sections were included. With the excep- tion of two of these sections, Euphorbia sects. Stachydium and Portulacastrum (discussed previ- ously), all species of subgenus Agaloma belong to 12-14). Indeed, previous suspicions were supported, and clade D and form a single subclade (Figs. both Chamaesyce and E. subg. Poinsettia are shown to have evolved from within subgenus Agaloma. Therefore, a monophyletic subgenus Agaloma must Poin- also include both Chamaesyce and E. subg. settia. This entire subclade is hereafter referred to s the Agaloma alliance. The r that Euphorbia phylloclada and E. glanduligera Рах, two annual African species from the Namibian analysis strongly supports (bs 93%) desert, also belong to the Agaloma alliance (Fig. 14). Both species possess well-developed involucral appendages and morphologically are easily accom- modated within subgenus Agaloma, where they cer- tainly would have been placed if it were not for their African distribution. Boissier treated Кирһог- bia phylloclada as the sole representative of E. sect Stachydium subsect. Capensis. However, the molec- ular data do not support a close relationship be- Annals of the Missouri Botanical Garden tween this species and other members of the section Stachydium). Pax (1894b) placed E. glanduligera in Chamaesyce, (see discussion under E. sect. and the similarity is indeed strong. In the ITS anal- ysis, these two species are placed in a clade sister to some members of subgenus Eremophyton (Fig. 12). However, this relationship lacks any bootstrap support, and the ndhF analysis probably provides a more accurate reflection of relationships. Within the New World, group in the Agaloma alliance consists of Euphor- the earliest. diverging bia subg. Agaloma sects. Ephedropeplus Müll. Arg. and Crossadenia Boiss. The former section is rep- pep here by E. appariciana Rizzini, E. sp. nov. . and E. gymnoclada Boiss., while the latter sec- tion is represented here by E. crossadenia, E. ses- silifolia, and E. sarcodes. In the ITS analysis, these two sections form a single well-supported clade (bs 99%) sister to the remainder of the Agaloma alli- 12, 13). The sample in the ndhF study was not sufficient to investigate their monophyly, ance (Figs. but here too the single species included, E. appar- iciana, is sister to the remainder of the Agaloma alliance. Both sections are restricted to eastern Bra- zil and represent a morphologically diverse assem- blage of about a dozen arid-adapted perennial The in- volucral appendages are deeply cleft to subentire, herbs, small shrubs, or stem succulents. and in one species they are lacking. Traditionally subgenus Agaloma sects. Crossadenia and Ephed- ropeplus have been separated by the possession of opposite or whorled, highly reduced scalelike leaves in section Ephedropeplus and the possession of alternate, well-developed leaves in section Cros- sadenia. However, this distinction does not appear to hold, and with respect to each other, the sections are not monophyletic. Therefore, E. subg. Agaloma sect. Ephedropeplus is best reduced to synonymy under section Crossadenia. In both the ITS and ndhF analyses, Chamaesyce is the next diverging lineage in the Agaloma alli- ance, and it is sister to the remainder of the spe- cies, excluding the earlier diverging Namibian taxa and members of Euphorbia subg. Agaloma sect. Crossadenia (Figs. 12-14). However, statistical support for this topology is lacking. What is sup- ported is that Chamaesyce is an ancient lineage that diverged earlier in the evolution of the Agaloma alliance. Beyond this, however, the exact relation- ship of its members to other Euphorbiinae is not evident. After the divergence of Chamaesyce, there are two major groups within the Agaloma alliance. These are sister to each other and present in both the ITS and ndhF analyses, with considerable sup- port (bs 99%) in the latter (Figs. 12-14). The first clade contains species currently placed in Euphor- bia subg. Agaloma sect. Zygophyllidium together with members of subgenus Poinsettia; the second clade contains the remainder of sections treated in subgenus Agaloma. Euphorbia subg. Agaloma sect. Zygophyllidium corresponds to a North American and Mexican as- semblage of about a dozen species. The group is poorly defined, but united by their annual or rarely short-lived perennial habit. In addition, many spe- cies possess serrate leaves. The section was rep- resented in this study by £. idoli Engelm., E. bilobata Engelm., Е. eriantha Benth., Е. exstipulata Engelm., E. jaliscensis B. L. Rob. & (жеши, and E. lacera Boiss. It is not monophyletic because sub- genus Poinsettia is nested within it (see discussion under E. subg. Poinsettia). Sister to the Zygophyllidium/Poinsettia group is the core of Euphorbia subg. Agaloma, an assem- blage comprised of section Alectoroctonum (Schltdl.) Baill. Benth., section Cyttarospermum Boiss. (representatives: E. E. aff. ariensis HBK, E. calcicola delicatula Boiss., E. eglandulosa V. W. E. gradyi V. W. Steinm. & A. Ram.-Roa, E. graminea Jacq., Steyerm., E. lagunensis Huft, E. misella S. Watson, Rob. & Greenm., (representatives: E. colletioides E. leucocephala Lotsy, and E. sp. nov. 2), adiantoides Lam., Fern., E. Steinm., E. guatemalensis Standl. & E. oaxacana B. L. E. ocymoidea L., Е. rzedowskii Mc Vaugh, E. segoviensis (Klotzsch & Garcke) Boiss., E. orae E E E. subpeltata S. — E. succedanea sinaloensis Brandegee, E. son- L. С. Wheeler, and E. whitei L. C. Wheeler), sec- tion A Boiss. үжен эч E. insulana Vell. and E. oerstediana (Klotzsch € Garcke) Boiss.), section Petaloma (Raf.) Boiss. (represen- tative: E. bicolor Engelm. & A. Gray), section Ti- thymalopsis (Klotzsch & Garcke) Boiss. (represen- tatives: E. A. Н. Holmgren € N. H. Holmgren, E. innocua L. C. Wheeler, E. L., E. macropus (Klotzsch € Garcke) Boiss., and E. sphaerorhiza Benth.), and section Tri- aaron-rossil cuanhae chosterigma (Klotzsch & Garcke) Boiss. (represen- tatives: E. antisyphilitica Zucc., E. californica Benth., E. ceroderma 1. M. Johnst., E. fulgens Karw. ex Klotzsch, E. gentryi V. W. Steinm. & T. F. Daniel, E. macvaughii Carvajal & Lomelí, E. misera Benth., E. rossiana Pax, and E. rossiana var. nov.). In comparison to members of subgenus Agaloma sect. Zygophyllidium, species of this group are highly variable in habit, and the leaves are strictly entire. This group accounts for nearly 90% of the species that have been treated in subgenus Aga- loma, and it ranges from Argentina and Chile to Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae the northern United States and the Caribbean. Col- lectively, these sections form a monophyletic group in both analyses. However, the clade only has boot- strap support (94%) in the ndhF analysis (Fig. 14). Here too belong E. cassythoides Boiss., a Caribbean taxon previously placed in Euphorbia sect. Arthro- 1862) and E. equisetiformis A. yalapagos Islands thamnus (Boissier, Stewart, an endemic to t whose relationship to other Euphorbia was uncer- 1971). Although the sample of species is not broad tain (Burch, enough to make a definitive conclusion, this study supports Parks (1998) narrow circumscription « Euphorbia subg. Agaloma sect. Tithymalopsis. Ac- cording to the molecular data, the section, as de- fined by Huft (1979) and Boissier (1862), is poly- phyletic and composed of at least three separate neither Huft nor Boissier can lineages. However, be criticized. In fact, the placement of similar-ap- pearing, geophytic taxa such as E. macropus and E. sphaerorhiza in separate clades with nongeophy- tic taxa is odd and represents one instance where Of the re- maining five sections of subgenus Agaloma, only molecular results are counterintuitive. sections Petaloma and Dichilium are likely to be monophyletic, but not enough taxa from these were included to test this hypothesis. What is clearly shown is that subgenus Agaloma sects. Alectoroc- tonum, Cyttarospermum, and Trichosterigma are polvphvletic. Euphorbia subg. Poinsettia. According to the most recent treatment of Euphorbia subg. Poinsettia (Mayfield, 1997), this taxon is a strictly New World assemblage of 24 species. Dressler (1962) recog- nized 11 species. The group is characterized by cup-shaped involucral glands that are generally re- duced to one (rarely more) per involucre. Pigmen- tation of the subeyathial leaves, as exemplified by E. pulcherrima. occurs in many but not all of the Growth form varies from annuals, species, peren- nial herbs, geophytes, to shrubs. Nine species of subgenus Poinsettia were included in this study: E. chersonesa Hult, E. heterophylla L.. E. hormorhiza Radcl.-5m., ma, E. pumicola Huft, E. radians Benth.. Е. E. pentadactyla Griseb., Е. зүн! stri- and E. zonosperma ar Arg. gosa Hook. & Arn., In both analyses, these species are nested wills a clade that corresponds to Euphorbia subg. Aga- loma sect. Zygophyllidium (Figs. 12, 14). This is not surprising because some species of this section (e.g.. E. bifurcata) possess involucres with a single gland and nearly identical seeds to those of mem- bers of subgenus Poinsettia. In addition, other taxa (e.g.. E. jaliscensis) possess panduriform leaves that í are otherwise known only from a few species of subgenus Poinsettia. The subgenus is noteworthy because of the extreme amount of genetic diver- gence among its members. It forms a monophyletic group only in the ndhF analysis (Fig. 14). In the ITS analysis, subgenus Poinsettia comes out in two groups nested within subgenus Agaloma sect. Zyg- ophyllidium (Fig. 12). A close affinity between sub- genus Poinsettia and species of subgenus Agaloma 1962; Webster, 1967), but the two groups do not appear closely related (Figs. 12-14). Dressler's (1962) re- moval of E. eriantha from subgenus Poinsettia sect. Dichilium has been suggested (Dressler, ^ justified by molecular evidence. Euphorbia cherso- E. heterophylla L. was placed in subgenus Poinsettia by Millspaugh nesa (= var. eriocarpa Millsp.) (1889) but placed in subgenus Agaloma by Hufi (1984). The molecular data strongly support that it is sister to E. pumicola in subgenus Poinsettia. Morphologically, E. chersonesa is noteworthy be- cause it is intermediate between subgenus Agaloma sect. Zygophyllidium and other members of sub- genus Poinsettia in involucral gland characteristics; the glands are not deeply cupped, but they do ap- parently lack appendages. The herbaceous habit of the species of Euphorbia subg. Agaloma sect. Zygophyllidium and many of the early diverging members of subgenus Poinsettia suggests that woodiness in species such as £F. pul- cherrima is secondarily derived. In addition, the lack of involucral appendages appears to represent a reversal to unappendaged involucres. ORIGIN AND BIOGEOGRAPHY OF EUPHORBIEAE The tribe Euphorbieae demonstrates a complex The earliest diverging African The next diverging biogeographical pattern. subtribe Anthosteminae, is strictly 1, 2). clade. subtribe Neoguillauminiinae, occurs in Aus- tralia and New Caledonia (Figs. 1, 2). Of the four major clades of Euphorbiinae, one, clade A (Figs. clade. and Madagascan (Figs. э. б). is stric ‘tly African and Madagascan, and one, clade B (Figs. 7. 8). is primarily distributed in tem- perate regions of the Northern Hemisphere. The (Figs. 9-11) and D (Figs. 12-14), are widespread, but the earliest diverging Africa and Madagascar. If present-day distributions are in- two remaining clades, € lineages within these two clades occur in dicative of historical ranges, then the molecular ev- idence is consistent with a hypothesized African origin of the tribe before the breakup of Gondwan- aland. Although clade B is almost entirely Laurasian in distribution, there are a few members in Africa. The 478 Annals of the Missouri Botanical Garden clade is absent from Australia and essentially ab- sent from South America, where only two species occur, و‎ spathulata and E. philippiana Boiss. The first of these has an amphitropical dis- tribution also occurs in North America. The second is a Chilean endemic similar to North American species. Considering that approximately 85% of the temperate South American species of angiosperms have an origin in the Northern Hemi- sphere (Raven, 1963), the presence of these species in South America is likely the result of dispersal Within clade B there are two major subclades, and both possess species events from North America. in North America and Eurasia. It is possible that the current distribution of clade B represents either an ancient dispersal event to Laurasia or evidence that the diversification of Euphorbiinae predates the split up of Pangaea. Clade C has a wide distribution that is best de- scribed as pantropical. According to the ndhF re- sults herein, the earliest diverging lineages of this clade occur in Africa, Madagascar, and Australasia. The Hawaiian endemic species Euphorbia haelee- leana belongs here and appears to have arrived at the islands by long-distance dispersal from related taxa, e.g., Euphorbia plumerioides, in the Pacific Islands. In the ndhF analysis, all of the Neotropical members of clade C form a single lineage that is sister to a lineage containing the subgenera Eu- phorbia and Lacanthis together with the Synaden- ium alliance, again suggesting that these two groups arose before the breakup of Gondwanaland. Simi- larly, the sister clade relationship of subgenus La- canthis and subgenus Euphorbia together with the Synadenium alliance suggest that the common an- cestor of both these groups was present before Mad- agascar began to separate from Africa. Clade D has a distribution similar to clade С. Both ITS and ndhF analyses indicate that the ear- liest diverging lineages are in Africa, Madagascar, Also like in clade C, all of the New World species belong to a single lineage. In- and southwest Asia. terestingly, according to the ndhF evidence, of the two earliest-diverging groups in the lineage con- taining the New World taxa of clade D, one occurs in Namibia and the other occurs in arid eastern Brazil. Thus, there is a clear, well-supported link between New and Old World species in this group. The near parallel distribution in the sister clades C and D provides further evidence that the distri- bution of New and Old World taxa is the result of vicariance caused by the breakup of Gondwana- land. Therefore, the molecular evidence corrobo- rates that despite being a very specialized group, the Euphorbiinae are also a very ancient group. Based on biogeographic patterns, Croizat (1940) postulated that Euphorbia was already differenti- ated into modern subgenera by the mid Cretaceous (ca. 100 million years ago), and this appears quite possible The fruits in Euphorbieae are generally dry, ex- plosively dehiscent, and initially dispersed only a few meters from their source. For this reason, there are few instances of long-distance dispersal within the tribe. One notable exception involves species of Chamaesyce, a primarily New World taxon whose range parallels that of the entire Euphorbieae. Web- ster (1967) proposed that this group probably orig- inated in the New World and molecular evidence supports his hypothesis. Because many Chamae- syce possess mucilaginous seeds, much of their great success in distribution is likely because seeds are able to adhere to animals and thus achieve long-distance dispersal (see discussion in Jordan & Hayden, 1992). Euphorbia and helps explain why Chamaesyce is This feature is otherwise rare in one of the few relatively derived taxa within Eu- phorbiinae that exhibits a transoceanic distribution. Webster et al. (1982) stated that the most-likely primitive Euphorbiinae were Euphorbia sect. Denisophorbia. Dressler (1957) believed that they Al- though the molecular evidence does not support were in sections Balsamis and Adenorima. their hypotheses, it does not provide solid insight as to what actually is the most primitive Euphorbia either, and it appears difficult to identify a single most-ancestral group. Dresslers and Webster's choices are sound because all three sections rep- resent early diverging lineages within the genus. The molecular evidence does provide some in- sight as to the features that ancestral Euphorbia likely possessed. First. they were probably a trop- ical tree or shrub, because early diverging lineages within the various clades of Euphorbia are mostly woody tropical plants. Phyllotaxy was presumably spiral, and it is likely that a pseudoumbellate in- florescence structure was well developed: these characteristics are found in all the major clades of Euphorbia, and their widespread presence is par- simoniously explained by their presence in the common ancestor of all four major clades. Stipules were probably absent or minute and glanduliform. There were likely five glands on the involucre. Al- though gland reduction is common, in many species the first involucre formed in an inflorescence often pos- with reduced glands, e.g., E. graminea Jacq., sesses five glands and only subsequent involucres possess fewer glands. Presumably involucral ap- pendages were absent, because they are present only in derived groups of Euphorbiinae. It is note- Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae worthy that the Madagascan endemic Euphorbia antso is the least derived species with regard to molecular evolution (see Figs. 3, 4) and possesses most of the features discussed above as ancestral. Although the ancestral condition is probably woody, in terms of species number, the two most successful lineages within the subtribe, Chamaesyce and the temperate E. subg. Esula clade, are primarily com- prised of herbaceous species. The large amount of diversification within Eu- phorbiinae compared to the other subtribes of Eu- phorbieae and the sister tribe Hippomaneae sug- gests that one or more key innovations may have promoted rapid evolution and diversification within this lineage. If such innovations actually exist, they likely involve the eyathium. Although cyathiate in- florescences also occur in Anthosteminae and Neo- guillauminiinae, both these subtribes contain very few species and have narrow distributions. The par- ticular feature that occurs in. Euphorbiinae but is absent in the other subtribes is that, with few ex- ceptions, the nectar-producing involucral glands are situated on the rim of the involucre and not partially enclosed within this structure. Therefore, this feature may have greatly enhanced successful insect attraction and pollination, thus giving mem- bers of Euphorbiinae a selective advantage, which in turn has driven the patterns of speciation and diversification presently observed in extant mem- bers of the subtribe. COMPARISON WITH MORPHOLOGICAL DATA There are two previous phylogenetic studies of the Euphorbieae based on morphological characters (Park, 1996; Park & Elisens, 2000). The first. of these treated only the New World taxa of subtribe Euphorbiinae, while the second treated the entire tribe. The molecular results differ drastically from the results obtained in either analysis. Even the two separate morphological analyses yielded different topologies, and in neither study was there signifi- cant statistical support for the majority of the clades. Part of the problem with these morpholog- ical analyses may be that too few characters were included to resolve the taxa. For example, in the first study of the New World species only 37 char- acters were used to resolve relationships among 49 terminal taxa, in this case species groups. In ad- dition, many characters employed in both analyses are highly variable even among closely related spe- cies, and their use to discern relationships within such a large and diverse tribe as Euphorbieae is unlikely to provide accurate results. FUTURE OF EUPHORBIINAE CLASSIFICATION The current classification of Euphorbieae sub- s tribes Anthosteminae and Neoguillauminiinae well supported herein and recognizes only mono- phyletic groups. No taxonomic adjustments are nec- essary for these subtribes, but as mentioned above, problems remain for the classification of subtribe К uphorbiinge. ‘he current taxonomic trend is going in the di- rection of splitting Euphorbia. This has already been observed with Websters (1967) removal of More Carter. (1994: 378) stated that subgenus Euphorbia “could be separat- Gilbert 235) also leaned toward an eventual dis- Chamaesyce. recently, ed as a genus in its own right." Likewise, (1987: memberment of Euphorbia, but well aware of the "profound nomenclatural consequences? associated with such a change, he justified only changes in — within Euphorbia. Ve disagree that Euphorbia should be divided. * opinion is that the best long-term solution to the problem of Euphorbiinae classification is to ex- pand Euphorbia to encompass all members of the subtribe. Some might contend that this is an un- desirable step backward 250 years to Linnaeus's broad concept of Euphorbia. However, we believe that this solution is more favorable than leaving the genus in its current paraphyletic circumscription or restricting Euphorbia to only the subgenus Kuphor- bia. The first taxonomic problem with limiting Fu- phorbia to subgenus Euphorbia is that ca. 90% of the species currently in the genus would need to be accommodated in other genera. Thus, Euphor- bia, a well-known and easily identifiable taxon known throughout the world, would no longer exist in most parts of the globe. Instead, there would be a multitude of genera completely unknown to most, and the boundaries and circumscription of these would be vague and certainly debated for quite some time. On the contrary, a broad Euphorbia would require changing the names of ca. 100 (vs ^ а. 1700!) species and would only affect groups with relatively limited distributions. These changes would only mildly broaden the current concept of Euphorbia to encompass taxa that possess an un- usual involucral morphology, a feature we believe has received undue taxonomic weight. Another reason that splitting Euphorbia is un- satisfactory concerns the unusual nature of evolu- The situation here is that — ion within this group. basic cyathial morphology in the genus is highly but morphology is highly conserved, vegetative à plastic. This has led to much parallel evolution in Annals cour — Garden growth form with little change in floral form. There- fore, if the relationships suggested by the molecular evidence do indeed accurately depict the phylog- eny of the group, then there are a number of well- defined monophyletic lineages nested within a par- aphyletic background of relatively undifferentiated groups or groups that have undergone a high degree of parallel vegetative evolution. To propose various genera whose members are superficially nearly identical fails to serve one of the primary purposes of a system of classification, e., to provide a predictable system that allows for the separation of taxa and for the ability to make assumptions about relationship based on morpho- logical features. If Euphorbia classification is to ac- tually reflect relationship, then there will be excep- tions and inconsistencies in defining new genera. We fear that any system that attempts to dismember Euphorbia will continue to have only limited suc- cess and acceptance. Croizat (1965: 574) emphasized the problem as- sociated with subgeneric groups within Euphorbia and stated "the infraspecific combinations of char- acters are so intricate as to make it really difficult to identify a truly natural subgeneric taxon." The same problem occurs with an elevation in rank to genus, but the broad implications are greatly more severe. Recognizing highly similar subgeneric taxa is much less of a problem because such rank is usually of primary interest to specialists in the group. Webster (1967: 398) stated, "If the various microgenera of Euphorbieae cannot be easily dis- tinguished, there seems little reason to adopt them simply because they represent evolutionary units; it is quite easy to discuss the evolution of these taxa if they are referred to as infrageneric compo- nents of Euphorbia.” We could not agree more. As mentioned in the introduction, the majority of genera currently segregated from Euphorbia are distinguished by cyathial traits. In. some regards, cyathial morphology has received undue weight, and this may be because there has been great di- versification in vegetative morphology throughout Euphorbiinae, but the overall structure of the cy- athium has remained relatively conserved. Thus, changes in cyathial morphology are viewed as tax- onomically significant occurrences, and little sys- tematic emphasis has been placed on other fea- tures, including relationships. Leach (1973: 32), when describing Endadenium, justified his new ge- nus on the basis of its distinctive arrangement of glands within the involucre, further saying that En- "combines so many of the characteristies African members of the tribe that its dadenium of the other assignment to any one of the genera involved would so effectively blur, if not nullify, the diagnostic cri- teria as to make it virtually impossible to retain these as separate genera on any but purely arbitrary grounds." However, the current recognition of gen- era separated solely on the basis of cyathial char- acters while ignoring, as in the case of Endaden- ium, all of the characters that do indeed blend the segregate genera into Euphorbia, is in itself some- what arbitrary. One of the arguments frequently used to justify the division of Euphorbia is that the genus is so large and heterogeneous that it is unwieldy. The incredible speciation and diversification that have occurred within Euphorbia are largely biological facts. Dismembering the genus would make it smaller and less heterogeneous but collectively as unwieldy an assemblage as ever. Another argument advocated by splitters is that the generic concept used for these assemblages is too broad in com- parison with other groups; for the sake of taxonomic consistency, Euphorbia must be divided. In re- sponse to this, it is worth pointing out that no two of plants are completely comparable Each has an independent evolutionary the argument that Euphorbia lineages each other. history. Therefore, should be divided simply because its evolutionary past has involved amazing speciation, dispersal, and diversification is unfounded. In our opinion, recognizing the genus in its broad sense has some benefit because it conveys the incredible evolution- ary history of this group. Can a workable system of classification be de- veloped for the Euphorbiinae that is based on the concept of monophyly? We believe that the answer is yes, but the units of such a classification should be infrageneric. Certainly many modifications will be necessary and time required before a stable sys- tem of classification is developed. However, it is better that such a process be undertaken at the infrageneric level, where the far-reaching nomen- clatural implications associated with generic changes can be avoided. Literature Cited pee Euphorbiaceae Neo-Caledonicae. Ad- Baillon, H. D = 5 М 1992, Ее netic utility of the internal spacers of nuclear ribosomal DNA in plants: b exile the Compositae. Molec. Phy- logenet. Evol. xu — J. M. Porter, M. К. Wojciechows- Domne 1995. The ITS A valuable source of Missouri Bot. . S. Campbell & M. J. region of nuclear ribosomal DNA: evidence 'е оп — phylogeny. Ann. Gard. 82: 247-277 P. | O. varieties and Bally, 1959. Some new species, Volume 89, Number 4 2002 Steinmann & Porter 481 Phylogenetic Relationships in Euphorbieae forms of Monadenium (Euphorbiaceae). Candollea 17: 25- . "1961. The Genus Monadenium. Benteli Pulish- ‚ Ber Be ania. G. 1878. Notes on Euphorbiaceae. J. Linn. Soc., Bot. 17: 185-267. m Euphorbiaceae. Pp. 239-340 in G. Ben- . Hooker (editors), Genera Plantarum 3. Lovell — London Berger, А. 1907 [1906]. Sukkulente Euphorbien. Eugen Ulmer, — Boissier, E. 1862. Euphorbieae. Pp. 3-188 in A. de € Prodromus Systematis Naturalis, Pan 15(2). Vie d Masson & Fils, Paris. — ——. 18 Peru ae (addenda et corrigenda). Pp. р in А. de Candolle (editor), Prodromus Sys- E matis Naturalis: Part 15(2). Victor Masson & Fils, Paris. Buck, W. & M. Hufi. 1977. Two new species of — subge nus Agaloma from Mexico. J. Arnold Arbor. 343-348. Bute һ, р 1971. d о | 509 in I. L. Wiggins . M. F Torá of the Calg Is бш. Stanford Univ. Press, Stanford. Carter, S. 1985. New species and taxonomic changes in Euphorbia from east and northeast tropical Africa and Oman. Kew A0: 809-825. 988a. Euphorbieae. Pp. hill (editor), das of Tropic al * Africa, оонай: aceae. А. / - 587 4 er (editors). a new с from . Balkema, Rotterdam. — ———,. 1988}. do qo section Somalica in Somalia. PE ^ 6—38. ——— e species of Euphorbia subgenus Ti- ruc alli ^ (Euphozbige eae) from Somalia and Oman. Nor- dic J. Bot. 12: 675-679. 1992b. New herbaceous and woody species of E oes ^ — eae) from Somalia. Nordic J. Bot. 12 — \ preliminary classification of е subgenus Euphorbia. Ann. Missouri Bot. Gard. 81 366—319. Croizat, E C. 1937. Notes of Euphorbiaceae, with a new genus and a new subtribe of the Euphorbieae. pine J. Sci. 64: 397 412. ———. |940. A significant new spec ies from New Guin- ea: Euphorbia mmo Croiz., D nov. Bull. Jard. Bot. Buitenzorg, ser. HI, 16: 351-35 1965. An моне НЫ ‘tion to ihe sification of Euphorbia L., with rican ar s M: ilagasy species. 1. Webbia 20: 573-706. in introduction to the subgeneric clas- sification s Euphorbia L; with stress on the South Af- rican and pi agasy species. Ш. Webbia 27: 1-221. Denis, M. 1921. Les Panhorst des ie Tie. Pp. 66-69 їп Philip- —— тепе el аз- Australes Imprimerie Nemourienne, rs J. > 8 J. L. Doyle. 1987. A rapid DNA isolation р > for ET — of fresh leaf tissue. Phy- Su] Bull. -15 Dressler, R. L. ve ne res iE (Euphorbi- aceae). Contr. Gray Heeb. — — 62. A synopsis a тола (Е uphorbiaceae). Ann. Missouri Bot. Gard. 48: 329—341. Foster, P. I. 1994 Revision ү Euphorbia — Teijsm. ex. Hassk. (Euphorbiaceae) and allies. Austro- baileya 4: 227-24 Gilbert. M. G. 1987. co new geophytic species of Eu- е — к оп the кк grouping of African members. Kew Bull. 42: 231—244. ——— 199% . Six new species — A ege (subgenus Esula) Dx "ED ges Kew Bull. 45: 265-276. === The relationships of ie Euphorbieae (Eu- Missouri Bot. X ~ — eae). Ann. Hassall, D. С. 1977. The genus RT in Australia. Austral. г Bot. 25: 429-453. Haworth, 1612. Synopsis Plantarum Succulentar- um. Ric ha — Londor House, H. 192 — of Me sit State. Knfstated list of ferns and flowering Bull. New York State Mus. Huft E Ji 1497 9. A Monograph ч Euphorbia section Ti- thymalopss Unpublished Ph.I | of — igan, Ann Arbor. 84. . Dissertation, Univer- ——— A review of Euphorbia (E — eae) in Buc ш: Ann. Missouri Bot. Gard. 7 )21— 1027. М. С. 1975. of the Chihuahuan desert region and adjacent areas. Wrightia : 120-1 " Jordan, M. S. & W. Lo testa in —— Collect. 21: 79-89. Johnston, Studies of the Euphorbia species 1992. A survey of mu Bot. (Barce dona) 2) iden. kim, K.-J. € R. K. Jansen. 1995. ndhF lution and the major clades in А» e family. Proc. Natl. Aci id. Sci., a : 10379- Klotzsch, J. E. & nar 1860 Bos Hr. Klotzsch las über Li inné's ens he Pflanzenklasse Tri- rliner Herbarium's im Allge: die natürliche Ordnung Euphorbiaceae insbesondere а m Konig. Akad. Wiss. Berlin 1859: 36— sequence evo- coccae des Be nen und Preuss. Tricoc- 4 00. Linné's кү he Pflanzenklasse ae. — . Wiss. Berlin 1859 [Phys. Ab- hal Koutnik. pm Chamaesyce (Eupl newly onion genus in southern Addis 'a. S, Afric ап J. Bot. 3: 262— 1987. : taxonomic revision of the Hawaiian species of the genus Chamaesyce (Euphorbiaceae). Al- 331-388. — 1: C. d interesting taxa of the tribe Africa Leach, L. 1973. New an Ы (Euphorbiac eae) from Portuguese Garcia de Orta, Ser. Bot. 1: 31—42. 1957. Euphorbia mandravioky, nom. nov., el un nom nouveau por une sous-section du genre Eu- phorbe. Bull. Soc. Bot. France 104: 499—501. Linnaeus, С. 1753. Species Plantarum. Impensis Lauren- tii Salvii, dera m. Mayfield, M. 1991. aceae). a new — johnstonii (E (Euphorbi- aulipas. Mexico — notes on Dahana н Асшае. Sida 11 9 species from Таш 57 1997. A Systematic Treatment of Euphorbia sub- genus ‚Ренан (Е uphorbiaceae). Unpublished Ph.D Disse — University of Texas, Austin. Miller, P 17 The Gardeners Dictionary, Abridged 4th ed. ne r At published), 1 Millspaugh, C. F. 1889. Euphorbiaceae. Proc. 217-230. 1913. The ew other = rican E uphorbiac eae. ed 37 . Bot. 353- London C айин та to North Ame m California Acad. Sci., 2nd ser., genera Pedilanthus = Cubanthus, ield Mus. Nat 482 Annals of the Missouri Botanical Garden sh nt, D. 1996. Molecular Methods in Plant Biol- y, 2nd ES Nic krent (self published), Carbondale. — R. G. € J. A. Sweere. Combining data in eileen systematics: An empirical approach us- ing three vd cular data sets in Solanaceae. Syst. Biol. 43: 467-48 Park, K.-R. (7 6. Phylogeny of New World subtribe Eu- ri (Euphorbiaceae). Korean J. Pl. Taxon. 26: 35-250. 1998. Monograph of Euphorbia n Tithymal- opsis с eae). Edinbu — Bot. 55: 161-208. —— & V. sens. 2000. tribe Euphorbieae (Euphorbiaceae). Int. J. Pl. Sei. 161: 425434. Pax, F 1894a. Euphorbiaceae africanae. Il. Bot. Jahrb. 19: 76-127. 1894b. —— s A Mes Plantae Gürichianae. Bot. Jahrb. 19: 142-143. 5: . Euphorbiaceae. Рр. 1-168 in A. Engler e (editors), Die Vegetation der Erde. IX. Die n welt Afrikas. HI. 2. Wilhelm Engelmann, Leip- Zg . Hoffmann. 1931. Euphorbiaceae. Pp. 11- 233 11 ln sler & K. Prantl (editors), Die natürlichen меана 2nd ed. 19C. Wilhelm Engelmann, Leipzi Pierre, ‘a 1896. Plantes du Gabon: 0 Bull. Mens. Soc. Linn. Paris 1(159); 1259— Planchon, P. Е, E а vrai nature Ps — des eu- ~ — phorbes expliquée par nouveau genre d'Euphorbiacées. Bull. Soc. Bot. j e 8: : -3: Prokhanov, J. 1. 1933. Conspectus systematicus Tithy malorum Asiae Medios. Trans. Rubber and a ha nst. Moscow. 1949, Euphorbia. Pp. 304—495, 134—144 in Ko- maros et al. (« editors), dos U.R.S.S Rade Jiffe-Smith, A. The taxonomic ‘posited of Eu- phorbia petiolata Sen p — n | of E. posti. (Eu- phorbiaceae). Kew Bull. 29: 503-5 Vili study of Rafinesque, C. S. 1837. Flora Telluriana 2. Rafinesque self — as Iphia. 3. Flora Telluriana 4. Rafinesque (self pub- lis hed), Pie lohia: ~ Rauh, W. ‚ Succulent and ae Plants of Mad- agasc e p 2. Strawberry Press, Mill Valley. Raven, P. H. 1963. мет ins ш relationships in the floras s Py and South America. Quart. Rev. Biol. 38: 151-17 Steinmann, V. ү. 2001. The Evolution of Succulence in the New World Species of Euphorbia (Euphorbiaceae). Unpublished Ph. se e Claremont Graduate Iniversity, Claremo Swofford, D. L. 2000. PAUP, Phylogenetic Analysis Us- ing Parsimony (*and Other Methods). Version 4. Sin- and, Т husetts. D. G. Higgins & T. J. Gibson. 1994. ISTAL W: — ving the sensitivity of progressive AU i sequence alignment through sequence weight- gap penalties and weight matrix Acids Res. 22: 4673-4680. Trew, €. $ 1754. Herbarium Blackwellianum, Vol. pen у, Webster, С. | 967. The genera of Euphorbiaceae in the southeastern United States. J. Arnold Arbor. 48: 303— 361 — — 5 = 2° —s = جس‎ - У. en o = = — p — c onspectus o a new classification of the Е eae. Taxon 24: 593—601 ——. 1994. Synopsis of the genera and раа taxa of l Euphorbia eae. Ann. Missouri Bot. Gard. 33-14 t wn & B. N. Smith. 1975. Systematics of synt — fixation pathways in Euphor- bia. Tuo n 24: 27-33. . Rupert & D. Koutnik. 1982. Systematic sig- niña ‘ance of pollen nuclear number in En ласеае, tribe epee Amer. J. Bot. 69: 407— Wheeler, 2 1943. The genera of ne — X50 Amer. B de 30: 45€ Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae ppe "i cultivated * Get ed taxa, the native origin of the plant is given in parentheses ndix |. icher information for the included species of Euphorbieae and ойон sequenced in this study. ITS ndhk GenBank GenBank Taxon Origin and voucher accession accession OUTGROUPS Omalanthus populifolius Graham Cultivated (nat. Australia), Steinmann AF537585 AF538202 23 (RSA) Sapium sebiferum (L.) Roxb. Cultivated (nat. China). Steinmann 1424 AF537580 AF538261 Sebastiania cornuta McVaugh Mexico, Sonora, Steinmann 589 (RSA) AF537587 AF538263 Stillingia spinulosa Torr. Mexico, Sonora, Felger 92-381 (RSA) AF537588 AF538264 EUPHORBIEAE SUBTRIBE ANTHOSTEMINAE Anthostema madagascariense Baill. Madagascar, Pascal 586 (MO) AF537582 AF538257 Anthostema senegalense A. Juss. Senegal. Bamps 7759 (MO) — AF538259 Anthostema sp. nov. — ar, Miller et al. 8840 (MO) AF537583 AF538258 Dichostemma glaucescens L. Pierre Gabon, McPherson 15531 (DAV) AF537584 AF538200 EUPHORBIEAE SUBTRIBE NEOGUILLAUMINIINAE Calycopeplus casuarinoides L.S. Sm. Cultivated (nat. Australia). Steinmann AF537580 -— 7 (RSA) Calycopeplus collinus РЛ. Foster Australia, гап der Werff 11848 (DAV) — AF538254 Calycopeplus paucifolius (Klotzsch) Australia, Craven 7139 (RS — AF538255 Baill. Neoguillauminia cleopatra (Baill.) Cro- New Caledonia. WePherson 17882 (MO) AF537581 AF538250 izat EUPHORBIEAE SUBTRIBE EUPHORBIINAE Chamaesyce acuta (Engelm.) Millsp. U.S.A.. Texas. Mayfield 1989 (RSA) AF537450 AF538176 Chamaesyce angusta (Engelm.) Small Chamaesyce articulata (Burm.) Britton Chamaesyce carunculata (Waterf.) Shin- ners Chamaesyce degeneri (Sherff) Croizat & Degener Chamaesyce hypericifolia (L.) Millsp. Small Chamaesyce polycnemoides (Boiss.) J. м Chamaesyce prostrata (Aiton Soják Endadenium gossweileri (N.E. Br.) L.C. Leach — aaron-rossú A.H. Holmgren ¿ N.H. Holmgren Euphorbia abdelkuri Balf.f. Euphorbia acalyphoides Hochst. ex 0188. Euphorbia acanthothamnus Heldr. & Sart. ex Boiss. Euphorbia adiantoides Lam. Euphorbia agowensis Hochst. ex Boiss. Euphorbia alluaudii Drake Euphorbia alta Norton Euphorbia amygdaloides L. Mexico, акак U.S.A., i Mayfield 1328 (RSA) irgin Isalnds. Steinmann 94-10 ) Mexico, Chihuahua. Bowers et al. 2939 (ARIZ) U.S.A., U.S.A., U.S.A.. Hawaii. Hawaii. Hawaii. Motley 338 (NY) Motley 1804 (NY) Motley 1802 (NY) Botswana. Snow «€ Chatakuta 6905 (MO) Cultivated (nat. SA) U.S.A.. Arizona. Ross s.n. (RSA) Cultivated (nat. SA) Kenya. Luke et al. TPR177 (MO) Angola). Steinmann Abd-el-Kuri). Steinmann Cultivated — Greece/Turke у), Stein- mann 142. ) Ecuador. pita & Anderson 22548 (G Ethiopia, Gilbert & Thulin 132 (MO) Cultivated (nat. Madagascar). Steinmann 1481 (RSA U.S.A. ) . Arizona. Sanders 5905 (RSA) C — d * Europe). Steinmann 1428 (RS. AF537449 AF537446 AF537447 AF537444 AF537443 AF537445 AF537448 AF537471 AF537396 AF537458 AF537395 AF537419 AF537468 AF537553 AF538175 AF538198 AF538194 AF538251 AF538232 AF538197 Annals of the Missouri Botanical Garden Appendix |. Continued. ITS ndhk GenBank GenBank Taxon Origin and voucher accession accession Euphorbia ankarensis Boiteau Cultivated (nat. Madagascar), Steinmann AF537462 — 2 (RSA) Euphorbia antisyphilitica Zucc. Cultivated (nat. U.S.A., Texas), Stein- AF537398 — mann 1451 (RSA) Euphorbia antso Denis Cultivated — ии ar), Steinmann AF537579 AF538253 1473-B (RS Euphorbia aphylla Brouss. ex Willd. Cultivated va. A anary Islands), Stein- AF537540 AF538225 mann 1466 (RSA) Euphorbia appariciana Колт Cultivated ina Brazil), Steinmann 1442 AF537455 AF538177 (RSA) Euphorbia arbuscula Balf.f. Cultivated (nat. Socotra), Steinmann AF537496 — 1435 (RSA) Euphorbia aff. ariensis HBK Mexico, Nayarit, Steinmann 1148 (RSA) AF537409 — Euphorbia atrispina N.E. Br. Cultivated (nat. Republic of S. Africa), AF537568 — Steinmann 1478 (RSA) Euphorbia atropurpurea Brouss. ex Cultivated (nat. Canary Islands), Stein- AF537542 AF538230 Willd. mann 1489 (RSA) Euphorbia attastoma Rizzini Cultivated (nat. Brazil), Steinmann 1487 AF537511 — (RSA) Euphorbia balsamifera Aiton ssp. ade- Cultivated (northeast tropical Africa), AF537571 AF538250 nensis (Deflers) Bally а 1480 (RSA) Euphorbia bicolor Engelm. & A. Gray . Texas, Van Devender 96-290 AF537380 — pes Euphorbia bifurcata Engelm. Mexico, Nuevo León, Nesom 7703 AF537434 AF538173 SA) Euphorbia bilobata Engelm. U.S.A.. Arizona, Steinmann 938 (RSA) AF537435 AF538172 Euphorbia boöphthona C.A. Gardner Australia, Coveny 3054 (RS AF537515 AF538207 Euphorbia brunellii Chiov. Cultivated (nat. east аре 'al Africa), AF537480 AF538203 Steinmann 1495 (RSA) Euphorbia calcicola Fern. Mexico, Morelos, Steinmann 801 (RSA) AF537385 — Euphorbia californica Benth. Cultivated (nat. Mexico), Steinmann AF537377 — 92 (RSA) Euphorbia calyculata HBK Mexico, Puebla, d L. 7261c (NY) AF537524 AF538221 Euphorbia calyptrata Coss. € Kralik Morocco, Podlech - 8 (RSA) AF537549 Euphorbia capmanambatoensis Rauh Cultivated (nat. oe ar), Steinmann AF537470 — 1468 (RSA) Euphorbia caputmedusae L. Cultivated (nat. Republic of S. Africa), AF537574 — Steinmann 1463 (RSA) Euphorbia cassythoides Boiss. Cayman Islands, Proctor 47858 (NY) AF537387 — Euphorbia ceroderma He Johnst. Mexico, Sonora, Steinmann 1393 (RSA) AF537389 AF538153 Euphorbia cestrifolia | Ecuador, Harling 27200 (GB) AF537521 AF538213 пако cheirolepis * h. & С.А. Central Asia, Vasák s.n. Ms AF537424 — Mey ооб chersonesa Hult Mexico, Baja California Sur, Steinmann АЕ537436 AF538174 1252 (RS Euphorbia clava Jacq. Cultivated (nat. Republic of S. Africa). AF537509 AF538245 Steinmann 1472-B (RSA) Euphorbia colletiodes Benth. Mexico, Sonora, Steinmann 93-387 AF537405 — (ARIZ) Euphorbia comosa Vell. Brazil, Webster 25425 (DAV) AF537503 AF538222 Euphorbia crossadenia Pax & K. Hoffm. Brazil, Graças et ds 886 (SP) AF537451 — Euphorbia crotonoides Boiss. Tanzania, M.R. 23765 (NY) AF537578 AF538238 Mexico, — Steinmann 1199 AF537393 AF538152 Euphorbia delicatula Boiss. (RSA) Volume 89, Number 4 2002 Steinmann & Port 485 Phylogenetic э Н in Euphorbieae Appendix |. Continued. ҺЕ GenBank GenBank ‘Taxon Origin and voucher accession accession Greece, Strid 25582 (RSA) AF537539 — Euphorbia dendroides L. Euphorbia denisii Oudejans Euphorbia depauperata Hochst. ex A. Rich. Euphorbia discolor Ledeb. Euphorbia drupifera Thonn. Euphorbia eanophylla Croizat Euphorbia eglandulosa V.W. Steinm. Euphorbia elata Brandegee Euphorbia elliotii Leandri Euphorbia epiphylloides Kurz Euphorbia equisetiformis A. Stewart Euphorbia eremophila A. Cunn. Euphorbia eriantha Benth. Euphorbia esculenta Marloth Euphorbia espinosa Pax Euphorbia esula L. Euphorbia exstipulata Engelm. Euphorbia fulgens Karw. ex Klotzsch Euphorbia gentryi V.W. Steinm. & T.F. Daniel Euphorbia germainii Phil. Euphorbia geroldii Rauh Euphorbia glanduligera Pax Euphorbia globosa (Haw.) Sims Euphorbia goetzei Pax Euphorbia gollmeriana Klotzsch ex Boiss. Euphorbia gottlebei Rauh Euphorbia gradyi V.W. Steinm. & A. Ram.-Roa Euphorbia graminea Jacq. Euphorbia grantii Oliv. Euphorbia gregaria Marloth Euphorbia guatemalensis Standl. & Steyerm. E uphorbia guerichiana Pax Euphorbia gymnoclada Boiss. Euphorbia gymnonota Urb. Cultivated (nat. Madagascar), Steinmann 1434 (RSA) Malawi. Kaunda & Tawakali 906 (NY) Former U.S.S.R., McNeal 710 (RSA) Cultivated (nat. Africa), Steinmann 1488 (RSA) Bolivia, Beck 11399 (DAV) хїсо, ‚© hiapas, Breedlove 70137 (С ы m 6694 (NY) Madagascar, Dorr et al. 3985 (MO) Cultivated (nat. Andaman Islands), Steinmann 1459 (RSA) Ecuador, Galapagos Islands, Eliason & Eliason 1573 (К) Australia, Vasek 680914-51 (RSA) U. Arizona. Steinmann 925 (RSA) Cultivated (nat. Re — of S. Africa). Steinmann 1474 (RSA) Cultivated (nat. ne a), Steinmann 1494 (RSA) U.S.A.. New Jersey (nat. Eurasia), Stein- mann 1427 (RSA) U.S.A.. Arizona, Steinmann 934 (RSA) Mexico, Oaxaca. Campos 813 (RSA) Mexico, Sonora, Steinmann 94-357 (ARIZ) Chile. Teillier 4267 (SGO) Cultivated (nat. Madagascar). Steinmann 1467 (RSA) Namibia, Koutnik 2015 (DAV) Cultivated (nat. Re pacc of S. Africa). Steinmann 1454-A (RS Tanzania, Taylor et if 6490 (MO) Venezuela, Ramírez 2696 (DAV) Cultivated (nat. Madagascar), Steinmann 71 (RSA) Mexico, Oaxaca, Steinmann 784 (RSA) Mexico, Sonora, Steinmann 94-107 (RSA) Tanzania, Bidgood et al. 1186 (MO) Cultivated (nat. Republic of S. Africa), Steinmann 1445-B (RSA) Mexi hiapas, Steinmann 1170 HE. — of South Africa, Balkwill et al. 6022 (MO) Brazil, Webster 25853 (DAV) Bahama Islands, Correll & Wasshausen 46769 (NY) AF537497 AF537547 AF537480 AF537498 AF537394 AF537495 AF537478 AF537484 AF537388 AF537423 AF537440 AF537416 AF537540 AF537433 AF537404. AF537400 AF537499 AF537475 AF537420 AF537413 AF537501 AF537459 AF537407 AF537410 AF537527 AF537408 AF537415 AF537450 AF537507 AF538191 AF538217 AF538107 AF538190 AF538229 AF538171 AF538154 AF538205 AF538178 AF538239 AF538185 AF538220 AF538151 AF538155 AF538242 AF538182 Annals of the Missouri Botanical Garden Appendix l. Continued. ITS idhk GenBank GenBank Taxon Origin and voucher accession accession Euphorbia haeleeleana К R. Herbst U.S.A., Hawaii, Fernstemacher s.n. (NY) AF537514 AF538206 Euphorbia hallii R.A. Cultivated (nat. Re public of S. Africa), AF537573 Steinmann 1475 (RS! Euphorbia hamata (Haw.) Sweet Cultivated (nat. Republic of S. Africa), — AF538237 Steinmann 1454-B (RSA) Euphorbia hedyotoides N.E. Br. Cultivated (nat. Madagascar), Steinmann AF537460 AF538196 2-A (RSA) Euphorbia heterodoxa Müll. Arg. Brazil, Webster 25810 (DAV) AF537500 — Euphorbia heterophylla L. Mexico, Sonora, Van Devender 98-1157 AF537429 AF538170 (ARIZ) Euphorbia — (Klotzsch & Costa Rica, Haber 10501 (F) AF537508 AF538211 Garcke) B Euphorbia — Radel.-Sm. Mexico, Tamaulipas, Mayfield & Patter- AF537431 AF538165 son 1843 (TE) Euphorbia iharanae Rauh Cultivated (nat. Madagascar), Steinmann AF537477 — 1458 — Euphorbia innocua LC. Wheeler U.S.A., Texas, — ld 2168 (RSA) AF537380 AF538161 Euphorbia insulana Vell. Brazil, ns 03 (N AF537411 Euphorbia ipecacuanhae L.. U.S.A., South ( cali Spongberg & AF537397 — Boufford 1718 (MO) Euphorbia jaliscensis B.L. Rob. & Mexico, Jalisco, Steinmann 754 (RSA) AF537442 AF538160 Greenm. Euphorbia juttae Dinter Cultivated (nat. Namibia), Steinmann AF537418 AF538188 Euphorbia kraussiana Bernh. و‎ of S. Africa, Stalmans 372 AF537548 — О) Euphorbia lacera Boiss. Mexico. К. de Mexico, Castilla & Tejero AF537441 — 958 (ENCB Euphorbia lactiflua Phil. Chile, Dillon & Teillier 5105 (F) AF537528 AF538219 Euphorbia lagunensis Hufi Mexico, Baja California Sur, Steinmann AF537379 — 272 (RSA) Euphorbia lagunillarum Croiz. Cultivated (Venezuela), Steinmann 1621 AF537502 — (RSA) Euphorbia lathyris L. U.S.A., California (nat. Eurasia), Stein- AF537550 — mann 1426 (RSA) Euphorbia laurifolia Juss. Ecuador, Mena C61 (NY) AF537509 — Euphorbia leucocephala Lotsy Cultive ate d (nat. Guatemala), Steinmann AF537381 — 94-17 (RSA Euphorbia lignosa Marloth C abes d (nat. Namibia), Steinmann — AF538240 455 (RSA) Euphorbia longifolia Lam. Canary Islands, Legaard 9905 (DAV) AF537558 AF5382: Euphorbia longituberculosa Boiss. Cultivated (nat. east tropical Africa), AF537577 АЕ ien Steinmann 1479 (RS! Euphorbia macropus (Klotzsch & Gar- Mexico, Nuevo León, Mayfield 1294 AF537378 — ‘ke 0188. ( T EX) Euphorbia macvaughii Carvajal & Lo- Cultivated (nat. Mexico, Michoacán), AF537382 — melí Steinmann 1486 (RSA) Euphorbia mahabobokensis Rauh Cultivated (nat. Madagascar), Steinmann AF537522 — 1456 (RSA) Euphorbia matabelensis Pax Botswana, Smith 4229 (MO) — AF538247 Euphorbia mauritanica L.. Cultivated (nat. southern Africa), Stein- AF537531 mann 1432 (RSA) Euphorbia medicaginea Boiss. Morocco, Podlech 41883 (RSA) AF537535 — Euphorbia meenae S. Carte India, Singh s.n. (RSA) AF537483 AF538202 Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae Appendix l. Continued. ITS ndhl GenBank GenBank Taxon Origin and voucher accession accession Euphorbia megalatlantica Ball Morocco, Podlech 41177 (RSA) АЕ537536 AF538226 Euphorbia meloformis Aiton Cultivated (nat. Republic of S. Africa), AF537505 Steinmann 1490 (RSA) Euphorbia meuleniana O. Schwartz Cultivated (nat. Yemen), Steinmann AF537572 — 1448 (RSA) Euphorbia milii Des Moul. Cultivated (nat. Madagascar), Steinmann AF537461 AF538195 1476 (RSA) Euphorbia millotii Ursch & Leandri Cultivated _ Madagascar), Steinmann AF537403 — 1477 (R Euphorbia misella S. Watson Mexico, C — Steinmann 1032 AF537384 AF538160 (RSA) Euphorbia misera Benth. Mexico, Baja California, Steinmann AF537383 1285 (RSA) Euphorbia monteiri Hook.f. RS Long & Rae 290 (K) AF537503 — Euphorbia myrsinites L. Colorado (nat. Eurasia), Stein- AF537551 AF538231 mann 1430 (RSA) Euphorbia namuskluftensis L.C. Leach Cultivated (nat. Bebo: Steinmann AF537502 — 191 (RSA) Euphorbia oaxacana B.L. Rob. & Mexico, Colima, Steinmann 1094 (RSA) AF537373 — Greenm. Euphorbia obesa Hook.f. Cultivated (nat. Re public of S. Africa), AF537566 AF538244 ips inmann 1465 ( Euphorbia oblongata Griseb. U.S.A., California (nat. Europe), Halse AF537555 4. : 34 (К; Euphorbia ocymoidea L. Mexico, Jalisco, Steinmann 1139 (RSA) AF537392 Euphorbia oerstediana (Klotzsch € Gar- U.S.A., Puerto Rico, Axelrod & Sastre AF538159 0185. 6278 (NY) Euphorbia omariana M.G. Gilbert Ethiopia, Friss et al. 3674 (К) AF537500 AF538243 nd panchganiensis Blatt. & India, Singh s.n. (RSA) AF537375 — Са Euphorbia papillosa A. St.- Argentina, Renvoize 3068 (RSA) AF537510 — Euphorbia pedilanthoides a Cultivated (nat. Madagascar), Steinmann — AF538192 38 (RSA) Euphorbia pentadactyla Griseb. Argentina, Cantino 700 (. han, AF537428 — Euphorbia peperomioides Boiss. Brazil, Nakjima et al. 1596 (DAV) AF537523 — Euphorbia peplus U.S.A., California (nat. E. Stein- AF537532 AF538228 mann 1433 (RS Euphorbia perrieri Drake us d (nat. Madagascar), Steinmann AF537463 = 3 (RSA) Euphorbia pervilleana Baill. — (nat. Madagascar), Steinmann AF537518 AF538209 1444 (RSA) Euphorbia petiolata Banks & Sol. Middle East, Liston 7-85-388/3 (RSA) AF537422 AF538180 Euphorbia phosphorea Mart. AF537512 AF538223 Euphorbia phylloclada Boiss. Euphorbia pilosa L. Euphorbia pirottae N. Terrac. Euphorbia platycephala Pax Euphorbia platyclada Rauh Euphorbia plumerioides Teijsm. ex Hassk Cultivated (nat. Brazil), Steinmann 1446 RSA) Republic of S. Africa, Oliver et al. 6611 K Former U.S.S.R., Elias et al. 7182 (RSA) Cultivated (nat. ' 1440 (RSA) Tanzania, Bidgood et al. 2249 (К) Cultivated (nat. Madagascar), Steinmann 147 (RSA) Australia, Fryxell & Craven 4022 (RSA) Tanzania), Steinmann AF537427 AF537557 AF537417 AF537501 AF537421 AF537513 AF538179 AF538234 AF538180 AF538241 AF538187 488 Annals of the Missouri Botanical Garden Appendix l. Continued. GenBank GenBank Taxon Origin and voucher accession accession Euphorbia poissonii Pax Cultivated (west Piu ral Africa), Stein- AF537482 AF538193 mann 1498 Euphorbia polyantha Pax Kenya, Faden & — 74/477 (К) — AF538189 Euphorbia primulifolia Baker Cultivated (nat. Madagascar), Steinmann AF537466 1 (RSA Euphorbia pteroneura A. Berger (1) Mexico, Chiapas, Mayfield 980 (TEX) AF537506 AF538210 Euphorbia pteroneura A. Berger (2) Cultivated (nat. Mexico/Cent. America), AF537505 Steinmann 1622 (RSA) Euphorbia pteroneura A. Berger (3) Cultivated (nat. Mexico/Cent. America), AF537504 — Steinmann 1623 (RS. Euphorbia pulcherrima Willd. ex Mexico, Nayarit, Steinmann 1070 (RSA) AF537432 AF538168 Klotzsch Euphorbia pumicola Hufi ee xico, Baja California Sur, Steinmann AF537437 AF538164 1 (ARIZ) Euphorbia punicea Sw C — d (nat. Jamaica), Raz 193 (NY) AF537516 AF538208 Euphorbia radians Be nth. exico, Sonora, Steinmann 944 (RSA) AF537438 AF538169 Euphorbia regis-jubae Webb & Berthel. C шый (nat. Canary Islands), Stein- AF537541 mann 1431 (RSA) Euphorbia rhombifolia Boiss. Cultivated (nat. Re "public of S. Africa), AF537414 AF538183 Steinmann 1439 (RSA) Euphorbia robusta (Engelm.) Small .5.А., Colorado, — 1429 AF537533 AF538227 (RSA) Euphorbia rossiana Pax Mexico, Puebla, Steinmann 1195 (RSA) AF537374 — Euphorbia rossiana var. nov. Mexico, Guerrero, Steinmann 1199 AF537390 — (RSA) Euphorbia rossii Rauh & Buchloh Cultivated — Madagascar), Steinmann AF537465 — 1449 (RS Euphorbia rubella Pax Cultivated (ninh east tropical Africa), AF537487 AF538204 Steinmann 1464 (RSA Euphorbia rzedowskii Mc Vaugh Mexico, Michoacán, Steinmann 1110 AF537399 — (RSA) Euphorbia sarcodes Boiss. Brazil, Cordeiro et al. 2233 (SP) AF537454 — Euphorbia scatorhiza S. Carter Cultivated (nat. Somalia), Steinmann AF537420 AF538181 1441 (RSA) Euphorbia scheffleri Рах Cultivated (nat. Kenya), Steinmann 1452 — AF538249 RSA) Euphorbia schimperi Presl Cultivated (nat. Arabian Peninsula/NE 11537537 — tropical Africa), Steinmann 1499 (RS/ Euphorbia sessilifolia Klotzsch ex Boiss. Brazil, Arbo 5418 (DAV) AE537453 — ) Euphorbia sessilifolia Klotzsch ex. Boiss. Brazil, Cordeiro et al. 2218 (SP) AF537452 — 2 Euphorbia segoviensis (Klotzsch & Gar- Mexico, Chiapas, Steinmann 1174 AF537400 — cke) Boiss. RS/ Euphorbia sinaloensis Brandegee Mexico, Sonora, Steinmann 94-130 AF537401 AF538156 Euphorbia sipolisii N.E. Br Cultivated (nat. Brazil), Steinmann 1443 AF537517 — (RSA) Euphorbia socotrana Balf.f. Cultivated A Socotra), Steinmann AF538248 1436 (RS Euphorbia sonorae Rose Mexico, id Steinmann 964. (RSA) AF537402 — Euphorbia soongarica Boiss. Former U.S.S.R., Elias RSA) AF537545 — Euphorbia spathulata Lam. U.S.A., California, Banks 1601 (RSA) AF537552 AF538233 Euphorbia sphaerorhiza Benth. Mexico, Sonora, Steinmann 1020 (RSA) AF537412 AF538158 Volume 89, Number 4 2002 Steinmann & Porter Phylogenetic Relationships in Euphorbieae Appendix l. Continued. ITS GenBank GenBank Taxon Origin and voucher accession accession AF537529 — Euphorbia stenophylla (Klotzsch & Gar- cke) Boiss. Euphorbia stricta L. Euphorbia strigosa Hook. & Arn. Euphorbia subpeltata 5. Watson Euphorbia succedanea L.C. Wheeler Euphorbia tannensis Spreng. Euphorbia tanquahuete Sessé & Mociño Euphorbia teke Schweinf. ex Pax Euphorbia tetraptera Baker Euphorbia thinophila Phil. Euphorbia thouarsiana Baill. Euphorbia tirucalli L. Euphorbia trichadenia Pax Euphorbia trichotoma HBK Euphorbia tuberosa L. Euphorbia tubiglans Marloth ex R.A. Euphorbia turczaninowii Kar. & Kir. Euphorbia usambarica Pax Euphorbia weberbaueri Mansf. Euphorbia whitei L.C. Wheeler Euphorbia zonosperma Múll.A Euphorbia xylophylloides Brongn. ех Lem. Euphorbia sp. Euphorbia sp. nov. | Euphorbia sp. nov. 2 Monadenium elegans S. Carter Monadenium ellenbeckti N.F. Br. Monadenium lindenii S. Carter Monadenium magnificum E.A. Bruce Pedilanthus bracteatus Jacq. Pedilanthus calcaratus Schltdl. Pedilanthus connatus Dressler & Saca- mano Pedilanthus cymbiferus Schltdl. Brazil, Aparecida da Silva & dos Santos 3267 (DAV Austria, Wallnofer 8531 (NY) Mexico, Nayarit, Steinmann 1079 (RSA) Mexico, Morelos, Steinmann 794 (RSA) Mexico, Aguascalientes, Steinmann 745 (RSA) Australia, Fryxell et al. 4475 (RSA) Cultivated (Mexico), Steinmann 1620 RSA Cultivated (nat. east tropical Africa), Steinmann 1470 (RSA — ar, Rapes ohita — (DAV) Chile, Dillon & Teillie 7 (Е — аг, pda & Ranaivoja- ona 14565 (K) Cultivated (nat. Africa/Madagascar), Steinmann 1445-A (RSA) Cultivated (nat. Zimbabwe/Angola), Steinmann 1461 (RS! Belize. Hill 20357 (MO) C — d (nat. dye of S. Africa), Steinmann 1472 USA) Cultivated (nat. Re publi of S. Africa), Steinmann 1462 (RS China, Liston 827-4 (RS, А) Tanzania, Balslev 291 (NY Cultivated (nat. Peru), Steinmann 1347 (RSA Mexico, Oaxaca, Torres 10833 (DAV) Brazil. Plowman et al. 8579 (F) Cultivated (nat. Madagascar), Steinmann 1450 (RSA) Cultivated (nat. probably Africa), Stein- mann 1469 (RSA) Brazil. Coredeiro et al. 2203 (SP) Tamaulipas, Mayfield 1851 | lexico. Cultivated (nat. Tanzania), Steinmann 1473-A (RSA Cultivated (nat. east ida “al Africa). Steinmann 1453 (RS Cultivated (nat. Somalia), Steinmann 1485 (RSA) Cultivated 1496 (RS — d (nat. Mexico). Steinmann 1460 (RSA) — о, a raeruz, Cházaro B. & de Cházaro 7294 (NY) Mexico. Jalisco, Sacamano s.n. (MO) Tanzania), Steinmann Mexico, Puebla, Steinmann 1624 (RSA) AF537559 AF537439 AF537376 AF537403 AF537425 AF537525 AF537485 AF537526 AF537530 AF537474 AF537504 AF537507 AF537543 AF537391 AF537430 AF537407 AF537481 AF537457 AF537470 AF537472 AF537489 AF537492 AF537493 AF537491 AF538163 AF538162 AF538184 AF538224 AF538218 AF538230 AF538246 AF538212 AF538214 AF538199 AF538200 490 Annals of the Missouri Botanical Garden Appendix l. Continued. ITS idhF GenBank GenBank Taxon Origin and voucher accession accession Pedilanthus finkii Boiss. Mexico, Oaxaca, Meave del Castillo AF537520 — 1551 (MO) Pedilanthus macrocarpus Benth. Mexico, Baja California, Steinmann AF537490 — 235 (RS Pedilanthus tehuacanus Brandegee Mexico, Puebla, Steinmann 1400 (RSA) AF537488 AF538215 Pedilanthus tithymaloides (L.) Poit. Guatemala, Castillo 2713 (NY) AF537494 АЕ538216 Synadenium grantii Hook.f. Cultivated (east tropical Africa), Stein- AF537469 AF538201 mann 1497 ( — ~~ SA) MOLECULAR Jason C. Bradford? PHYLOGENETICS AND MORPHOLOGICAL EVOLUTION IN CUNONIEAE (CUNONIACEA E)! ABSTRACT The Cunonieae are the largest tribe in the flowering plant family Cunoniaceae and include the widespread genus Weinmannia. This study aims to understand phylogene lic re lationships within Cunonieae e by using DNA sequences in a parsimony-cladistic analysis. Sequence ph loci included the internal transcribed s (ITS-1 ind ITS- 2) of nuclear е дерине | DNA. and the tral. intron and trnl-F spacer of chloroplast DNA. жайы p» aaa ific amplification of non-orthologous ITS-2 copies made it necessary to exclude the IT 5-2 data, but otherwise the nuclear and ¢ : 0 ala sels were congruent. The results place Vesselowskya as the sister genus to other Cunonieae and support the нат of Panc 5 Cunonia, and all five sections of Weinmannia, but do not ae how these а аге related. The os of Weinmannia sect. Weinmannia is upheld, with W. trichosperma from temperate forests of South America and W. tinctoria from the Mascarene Islands placed basal to a large clade of tropical American species. Although e maintain the monophyly of Weinmannia, this is neither verified nor statistically refuted by the molecular data. Likewise, Cunonia, with one isolated South African species, has only weak molecular support but ¢ = morphological ee | Lack of support for relationships among major clades within Cunonieae makes it difficult to suggest patterns of ораев, ‘al evolution. However. a well-qupportid phylogenetic hypothesis within инна 5 sect. Leiospermum is used to discuss hete тоору in inflorescence architecture. Uniquely derived features of the inflorescence are found in the New Caledonian species Weinmannia dichotoma and in the New Zealand species W. silvicola and W. racemosa. These heterotopic changes involve alternate patterns in the fate of terminal meristems and s arrangement of metamers bearing racemes. In an appendix the correct orthographies and original publications of all five sections of Weinmannia are жейде: types are also designated for Weinmannia sections /nspersae. and сае in order to — al em. Key words: cladistics, Cunonia, Cunoniaceae, Cunonieae, evolution, Fasciculatae, heterotopy. гон Ак e archi- lecture, — sae, ITS. J a molecular systematics, Pancheria, paralogous loci, Spicatae, trnk. ‚ Vesse- lowskya, Weinmannia. The flowering plant family Cunoniaceae R. Br. 1998a). Weinmannia is divided into five sections, (Oxalidales) (Angiosperm Phylogeny Group, 1998) with each one largely restricted to a particular geo- comprises about 300 species in 26 genera. Plants graphic region. The Cunonieae are also composed f the two other largest genera in Cunoniaceae, of the family are trees and shrubs in wet tropical and cool temperate forests, with most genera oc- Pancheria, with about 30 species endemic to New curring in eastern Australia, Melanesia, and New Caledonia (Guillaumin, 1940; Morat, 1993), and Cunonia, with about 25 species in New Caledonia Guinea. About 210 Cunoniaceae species are in ¿ monophyletic group of four genera called the tribe and 1 species in the South African Cape region Cunonieae (Bradford & Barnes, 2001). Weinmannia (Hoogland et al.. 1997). Vesselowskya. the remain- is by far the largest and most widely distributed ing genus in the tribe, has only two species endem- genus, with over 150 species found in the Ameri- ic to eastern Australia (Rozefelds et al., 2001). cas, islands of the eastern Indian Ocean, Malesia. A few recent publications have provided new in- and the South Pacific (Bradford, 1998: Hopkins. sights on the taxonomy and phylogeny of Cunon- ' This research has been supported by a U.S. National Science Foundation — grani (BIR-9250779) to Wash- ington University; a Mellon Foundation * к the Missouri Botanical Garden; a U.S. National Science Foundation Dissertation Improvement Grant (#5747 Э) to В. Schaal and J. Bradford: and а travel grant from the American Society of Plant Taxonomists to J. Bradford. " с A1 or the ueri and ms of many colleagues from institutions ا‎ the world, including: AKU, ОБУР, COL. FRIM, LOJA, MA OU, NSW, P, PORT, QCNE, SAN, SAR, 5 , TAN, TEF, UNIMAS, USZ, V И i. and the ^ etii p iety (W. E to B. Schaal for work in her laboratory al edes University; and to C. Feuillet, V. Hollowell, H. Fortune Hopkins, J. Rauscher, and A. Rozefelds for help on the manuscript. ? Research Scientist, Missouri Botanic z Garden. Visiting Sc ‘holar, мү Science and Policy, 2132 Wickson Hall. University of California at Davis, Davis, California 95616, U.S.A. bradford@ice.ucdavis.edu. ANN. MISSOURI Bor. GARD. 89: 491-503. 2002. 492 Annals of the Missouri Botanical Garden ieae. In a study of the relationship between Wein- mannia and its putative close relative, Cunonia, using a cladistic analysis of morphological features (Bradford, monophyly of Weinmannia, four of its sections, and 1998) there was weak support for the of Cunonia. A family-level analysis by Bradford and Barnes (2001) based on morphology and chlo- roplast DNA sequences (rbcL and trnL-trnF) estab- lished a new tribal classification, and clarified ge- neric. circumscriptions by proposing apomorphic morphological characters for each genus. Several recent generic revisions have provided valuable details on taxonomic distribution and mor- phology for many Cunonieae species. Rozefelds et al. (2001) gave a table summarizing similarities and differences among genera of Cunonieae and de- scribed a new species of Vesselowskya. New species descriptions and keys have also been produced for Cunonia (Hoogland et al., 1997) and Malagasy Weinmannia (Bradford, 2001; Bradford & Miller, )01). Revisions have been completed for Wein- mannia of Malesia and the South Pacific (Hopkins, 1998a, b, c; Hopkins & Florence, 1998). Weinman- nia of the Americas are poorly studied in their en- tirety, although some national and regional treat- ments have been done (Harling, 1999; Bradford & Berry, 1998). Despite these efforts, relationships among Cu- nonieae genera are unclear, the monophyly of Wein- mannia is in doubt, and the monophyly of some sections within Weinmannia is poorly established. In this study, I use DNA sequences from chloro- plast and nuclear loci to clarify phylogenetic rela- tionships within Cunonieae. In addition, I show that the phylogenetic hypotheses generated by this mo- lecular data can help reevaluate character evolu- tion within the tribe, especially with respect to in- florescence architecture. METHODS Based on the family-level analysis of Bradford and Barnes (2001 phyletic, and taxon sampling was designed to max- ) Cunonieae are clearly mono- imize the geographic, phylogenetic, and morpholog- ical diversity within this clade that has been 1998). One distinctive species endemic to Sulawesi, Wein- elucidated by previous studies (Bradford, mannia descombesiana, is missing; otherwise, sam- pling is broad, including exemplars from 45 species of Cunonieae (Table 1). Codieae may be the most closely related tribe to Cunonieae, but I was unable to obtain internal transcribed spacer (ITS) nrDNA sequences of Codieae to use as outgroups. Instead, TS trees were rooted using two species from an- other closely related tribe, Caldcluvieae (Bradford & Barnes, 2001). The trnL-trnF cpDNA (Taberlet et al., 1991) data set uses several outgroup taxa from closely related tribes. Collections were made from native populations and botanical gardens between 1994 and 1998. Fresh leaves were dried in silica gel for DNA pres- ervation. All DNA samples are vouchered with her- barium specimens and were deposited at MO and in the country of origin. Table 1 lists source and voucher information of each herbarium specimen and GenBank accession numbers for all DNA se- quences. Detailed information is available for Brad- the TROPICOS ((http://www.mobot.org)). ford collections on database I sequenced both the nuclear ITS region (Bald- win, 1992) and two adjacent chloroplast loci, the tral. intron and the intergenic spacer between the tral 1991). The ITS region was sequenced first, which helped es- 3’ exon and trnF (Taberlet et al., tablish likely monophyletic groups. A smaller set of trnL-trnF sequences was obtained from a sub- sample of each distinct lineage that was discerned from ITS data. Standard methods were used to ex- tract, amplify, and sequence DNA loci, and these are described in Bradford and Barnes (2001). BLAST ((http://www.ncbi.nlm.nih.gov/BLAST/)) comparisons were done to confirm that sequences were of angiosperm origin and not from possible fungal or other contaminants, and indeed similar nucleotide sequences were of appropriate loci and within the eudicot clade. Standard ITS primers (Baldwin, 1992) did not strongly amplify or produce clear sequences of ITS- 2 for many species in Weinmannia sections Leios- permum and Inspersae, or in Pancheria, Geissois and Caldeluvia. Hypothesizing that high G-C con- tent was interfering with PCR amplification, I de- signed alternative primers with higher annealing temperatures based on published sequences of 26S rDNA (Kuzoff et al., 1998) and 5.8 rDNA sequenc- es from my own work. The new primers did yield clear sequences as hoped, but preliminary cladistic analysis using ITS-2 data alone resolved clades with a mixture of ingroups and outgroups, a result incongruent with ITS-1, chloroplast, and morpho- logical data (see Bradford, 2000, chapter 1, for a figure of these results). This suggested that the ITS- 2 region amplified using the new primers was not orthologous to ITS-1 sequences obtained using the standard (e.g., ITS4) primers. To test this, a Parti- tion Homogeneity Test (PAUP*4.0b6a; Swofford, 2001) was used to compare the ITS-1 and ITS-2 data sets, and they were found to be significantly 0.002). Because incongruent (500 replicates, P = Volume 89, Number 4 2002 Bradford 493 Molecular Phylogenetics in Cunonieae ITS-1 and ITS-2 are adjacent loci they should have the same evolutionary history; significant incongru- ence between them strongly suggests that non-or- thologous loci were amplified. The ITS region is known to contain multiple copies of the ribosomal genes, as well as pseudogenes, and different PCR conditions can preferentially amplify different par- alogous loci (Buckler et al., 1997). Until the issue of paralogy can be resolved, the decision was made to exclude the ITS-2 data from analyses of organ- ismal phylogeny; these sequences, however, are available in GenBank (phylogenetic data set with the range AF521255-AF521298) Sequences were aligned by eye in Se-Al (Ram- baut, 1995) and exported in a NEXUS format. In- sertions and deletions were scored as binary char- acters (e.g.. present or absent). Any regions with ambiguous sequence or uncertain alignment were ignored during analysis. Parsimony cladistic anal- yses were implemented in PAUP*4.0b6a (Swofford. 2001). For all parsimony analyses, the following op- tions were used: characters unweighted and unor- dered, searches heuristic, starting trees obtained via random stepwise addition, TBR branch swap- ping. COLLAPSE option on, STEEPEST DE- SCENT option off, MULTREES on. Support for clades was estimated with bootstrap values (using the “Fast” bootstrap option with 10,000 replicates in PAUP) and decay values (using Autodecay: Er- iksson, 1999). The nuclear and chloroplast data sets were com- bined after checking for compatibility using the Partition Homogeneity PAUP. T could not reject the null hypothesis that the data sets represent the same evolutionary history (500 = 0.06 Test in ns test replicates, RESULTS ITs-] ANALYSIS The final ITS-1 data set included 48 sequences from 47 species (45 in Cunonieae) and a matrix of 260 characters. The number of equally parsimoni- ous trees found during heuristic searches made it impossible to complete branch swapping. Several searches were done using random taxon addition, and each analysis found the same large tree island with 268 steps = 0.70, RI = 0.86). The strict consensus of this tree island shown in Figure | is based on over 30, ООО equally parsimonious trees. The ITS-1 data strongly support the sister-group relationship between Vesselowskya and the rest of Cunonieae (Fig. 1). Pancheria is monophyletic and weakly placed as the sister to clades of Weinmannia and Cunonia. Five traditionally recognized groups are resolved as clades, including Cunonia, and four sections of Weinmannia: Weinmannia, In- spersae, Spicatae, and Fasciculatae. Species from Weinmannia sect. Leiospermum form part of a large polytomy. The two sections from Madagascar, sec- are placed as sects. tion /nspersae and section Spicatae, sister taxa to each other. Not all of these clades have high “Fast” bootstrap values, however, and the data give no support for relationships between these arger clades. trnL-F ANALYSIS The trnL-F data set included 996 characters from sequences representing 38 species (29 in Cu- nonieae). Ingroup sampling was less intensive than with the ITS data set, but included sufficient sam- pling from all major ITS clades. The parsimony analysis found 420 trees of 244 steps (CI = 0.73, RI = 0.82) (Fig. 2). On the strict monophyletic, and Vesselowskya is placed as the sister group to other Cunonieae. In contrast to ITS results. Weinmannia sect. Weinmannia is the sister taxon to a large clade containing Cunonia, Panch- eria, and all other species of Weinmannia. Cunonia capensis, from South Africa, groups with Pancheria while other Cunonia are monophyletic and sister to the remaining Weinmannia. Malagasy Weinmannia (sects. Inspersae and Spicatae) form a clade, as do sections Fasciculatae and Leiospermum. The inter- nal topology of the cladogram has low bootstrap val- consensus tree, Cunonieae are ues. COMBINED ANALYSIS For the combined analysis, 29 taxa, including 27 from the Cunonieae, were sequenced for both ITS and trn L-F. The final data set included 1254 char- acters. The parsimony analysis found a single most parsimonious tree of 388 steps (СІ = 0.74, RI 0.81) (Fig. 3). The base of the tree is structured similar to the ITS trees: Vesselowskya is basal to all Cunoniae, and Pancheria is a sister taxon to the clade con- taining Weinmannia and Cunonia. Although there is strong support for the position of Vesselowskya, no other internal branches have high “Fast” boot- strap or decay values. High “Fast” ES and decay values do support most of the commonly rec- including Pancheria and all five Cunonia does ognized taxa, Weinmannia sections. By contrast, not form a clade in the “Fast” bootstrap consensus tree and has a decay value of one. 86SS8PAV Þ6L66ZAV LPI66ZAV spue[sp Á191908 (ауа 'OW) 226 paofpvag 9400] 7AV "f. SISUIIDIDA pruupumiog LOSS8PAV CSA) puez мәм (OW) 9£6 paofpvag Y 7] »souia2na тииршиә 96SS8PAV spue[s] Á191908 (d¥d ‘OW) £16 piofpvag “18104 *£) Dsopfrasvd тииошиә C6CC9t. AV C6166Z4AV CP 166c4V вшорәүе”) MIN (QON 'OW) £S0S surgdog SU) x "usuo1g DUOJOYIP piuupDumnog (ZZ) voq *qq umuaedsoiz] иоцәәѕ (IST) 7] отииошитәц 6£9S8PAV €1Z667AV 09166c.1V eensny (MSN ON) 628 paofpoag MOT[9d Y вәше ^M "Y “Splajezoy visnuaa DAYSMO]ASSAA (c) dueg néysnojossa| $ SE9SBTAV ешорәег) MIN (NON ON) 819 pzofpvag шшперте) ортол DILIYIUD] © 1£9€9t AV PiuOpa[e7) MON (QON ON) 029 p10fpmıg шшперте) xo su) y "usuoag sapioopai]iud DILIYIUD б 9€9CQt.4 V ZIZ66ZAV 6€ 166c.1V вшорә[ег) мәк (NON ON) 019 peofpoig ‘uedweg хә "әт отету D1421|2ung © СЄОСӨРАҮУ LIZ66ZAV 8S L66ZAV vruopaſp MƏN (NON ON) 209 pu0/pv1g чә рирмәрЗиә DILIYIUD Y 2 PEOCRPAV bluopoa,ſp MIN (NON ON) 919 ptofpoig "uedureq mssayunsg DILIYIUD| 9 = (08) SUN x "usuoig nuauyoungq 5 el €€£9c9r. JV eruopo[e7) MƏN (QON ON) S£9 paofppag suc) Ņ "uxuoig pjjayoynd DIUOUN) os ZEOCBHAV BIuOpayey) MON (QON ON) 609 puo/pv1g [YOG рирјиош DIUOUN) g 2 [€9S8PAV OIZ66ZAV ZST66ZAV етиорәүегу MƏN (QON 70И) 209 p4ofpvig sut) y `иЗиолң рур\уЧолээш Dıuoun?) £z 089684V 60Z66ZAV 9€ L66ZAV CSA) euyy quinos (OW) S£2 p4ofpvig "| szsuadp? pruoun-) 6Z9SBPAV 8076674 V CC [66c.1V eruopo[e?) MON (QON ОИ) 219 psofpoig SHS y “UBUOIG IDSUD]DG piuoum) 8c9€9t AV LOZ66ZAV PSL66ZAV вшорәүг) MIN (QON ОИ) £19 pl0fpvıg “AYO suaqnaoaqm piuoun:) (QZ) "| Pruoun;) (O12) "8 ^H AVAINON AD [POS8PAV L1Z66ZAV POL66ZAV spue[s| uowojog (JISH ОИ) Org pzofpvag JUNY 22192122 sisdoavardg SZZO6TAV ZLI66ZAV ity (VANS ON) cge psofpoig "pos 214078 “yo v9]MM4 ШЕ (VAS `ОИ) 610¢ surydoy a1dsa][1) pquadns sI0SS124) 81Z66ZAV COL66ZAV ецедѕпу (HA ON) 658 paofpvag эпи 4 Dupnunuaq S108812) PZZO6TAV [L166ZAV eruopo[e7) MƏN (ПОМ ON) 009 psofposg urune[[1n<) 40]0951p mpo) EZZ66ZAV 02 L66ZAV ei[Ensny (MSN ON) 258 paofpvag swoipuy prjofia4os DUI021]]D7) Ot9cor. AV 9126624Ұ £9166c.1V (erpeagsn y ) op?) (16191€ OH) "ws saung uoq ‘q n;vjoiupd pianpop]p-) C1Z66Z4V c9166c.1V рчрүвәў Moy (A ОМ) 606 p:ojpvag чипу) “y Dijofisos DUDIIY ÞIZ66ZAV I9166c4V Pre. sn y (THA ON) Erg peofpoig MUA cp 7D) Y Avzneg (C[[onjy 74) оғорлотира DUDY Y SAANOHILNO ISLI As quu р—ә qua (рәї®лцүпә) uonoa[o^) (uOXP] ш saroads Jo 13quinu) snug‘) UIZLIO JANEN sardadg пәм ӘР exeyqns 5и pue әвәшоипт) и! satoeds jo 19qunu ay1 jo sS3Je wusa Jūuə lmno ]s9q əy} *ssoisoid ul HOM pue SUOISIAQI ]u32294 uo posed “DIUUDUU1A4 jo suor]oos 3 pue *p1ouos *aearuoun?) “sdno133no Aq poziuez10 oie sie[dusxy “USAS aie a9uanbas yore Jo Joquinu yuegua‘) әці pue sutouanbas VN( ш pasn sie[duioxa 10} SISYINOA °[ AQEL 495 Bradford Volume 89, Number 4 2002 Molecular Phylogenetics in Cunonieae [I9C914 V CS aU) (OW) 8&2 p0/pv1g “ARTY DULLIASOYIALL] рииюшиләд\ OL9OC8PAV erquiopor) (100 ON) 752 pio[pvag 7] 2s0ju2u10] DIUUDUUIIAf 609€8T.4 V 66 L66EAV 9V 166cAV Sn me] (06282 VIA) "ws “waaay q MIO оир POOC8HAV р|әпләпәд (LMOd ON) S&P pyofpvag € лои ‘ds рииюэшизәд\ 809<8 AV eIquio[o7) (ТОЭ (OW) Sz pio[ppag DNY DIJOf IÓ A DIUUDUUI9Mf 109€9t4V PIAI[OQI (ZSA ON) gee puofppvag spel, PYIUDJUD]]2 DIUUDUIUIAA 909€8T.4V 86166c4V CV LOOTA V ida hara! (ZSA OW) EZS pofpvag Aqsny пЎирф AAA и соосу eog (781 ON) ZFS pao/poag '"uolorp] DID]NIUND юииэшипәу\ (11) "] эютииршщшәдү uonoos CZOCEPAV JROSPSEPR IA (NVI OW) €€9 paofppag z ou “ds DIUUDULUIA LEOS8TAV 90c66c.1V ES LOOGA V леоѕевереру (МҮЛ, ON) 0882 42q102]0 y quu nzioquoma pruupDunen 9C9S8TAV с0с66с4Ү сс166с4У лвәвр#ере|ү (NVL OW) 4Е©9 рго/рргя чә$ хә "D SISUILIDISDID DU AUDA (6) p1ojJpeag 77) "f хә їрїриләң apsiadsuy U00YI3S PZOSBPAV IROSRSEPR N (NVL OW) 059 p0/pv1g Jaye DAYIDISOUI]S ртииюшиләд\ EZOCOPAV POZ66ZAV LSL66ZAV JIPOSRSE PE N (NVI ON) SIZ pl0fpmıg ipreudag umapznsimmzups тииршиәд CCOSBTAV COC66cAV 0€166c41V 1еоѕеӣвреуі (МУЛ, OW) 282 199 шоуу 1oxeg wopfimunu инип 12968 A V JPISPBRPR N (NVI ON) 269 P0/pD1g piojpeag ^S "f TNA `$ Y sisuatofoapui piuunuuiam 0c9c€9v.1 V JPISPBRPR IA (NVI ON) €69 рор ipreudagp DUDILIQUINY DIUUDUUIIA, 619€9T AV IRISPSEPP N (NVI OIN) 689 pio[ppag "ny, Punitelod DIUUDULUIDAf 8196844V IRISPBR PRIA (NVL OW) cro p0/pv1g piojpeag °) f (1ріешәя) руп DIUUDUUIIA (9g) p1ojpeaq ^7» "f ‘Xe Iparuleg 2D7D21dg uonoas LI9S8TAV ifr] (VAAS d ON) £co€ sudo Хеле) "y mon DIUUDUAHOR 9[9C€9T.4 V ваши р лере (4AN d OI) 62€ piofppag piojpeag `7 ‘f у surydoy 7 ^H 2pun]3004 DIUUDUUIAA, CI9cort AV O2UIO£ (NVS ON) 100€ surgdog uo(] (| Xo “US (Z) nouixpaf nniununiog €I9€8T4V c0c66c4V ША АЙ epsutuoq PISAP[P N (JAM d ON) 826 paofppag uoq (Y хә "ws (1) naumpaf DIU DULL Ta TI9S8VAV L0c66c4V St 1000d V SHOES (dISH ОИ) r18 Pl0/p»ıg4 "us *) “Y pnzixo LUDA — GLOS8SVAV 00c66c4V LV I66cAV e1sAeJe W (d ‘NIY ON) [TOS suzydoy SUNG PEU piuupunnam (61) surydoy *) ‘н Y pue[goog хә rpieurog 2D7p]n212$D,[ uonoos €£09€8T AV frg (VANS d ON) [ros surydoy MOOS ыы DIUUDUU124f c09€8T.1V spue[s] Á191908 (ауа ON) [26 p2ofppag эңе 209590 piuununiog 166684V SUOUIO[OG («ІЯ ON) ZE8 paofpvag (29661 “sunydoy jo) [ слой ‘ds DIUUDWUIUIAA 009€9T.4 V PIUOP3]P) MIN (NON ON) 229 p0/pv1g SU x "идиота 0701295 DIUUDUUIIA, LOOS8VAV 61664V PV LO6CAV рче[вэи MON (AMV COW) 216 P20/p01g ey Se pay Oe DIUUDUUTA A 66SE8VAV 9616671 V tr LOOZAV BOWES (OIN) 008 Plofpmıg ARI "V. SISUaOWIDS шири ISLI Aa que pro ud (pajean¡no) UOT] JO") (uoxe] ur saroads Jo ләдшпи) snud‘) UIBLIO JANBN saınad¢ ‘ponunuoy “р әде, 496 Annals of the Missouri Botanical Garden ununodsolo7 p— Weinmannia raiatensis Weinmannia sp.nov. 1 (—— — Weinmannia dichotoma (0 —— ——— Weinmannia serrata p— —— Weinmannia parviflora lL Weinmannia samoensis p— — Weinmannia racemosa (000 0————— MWeinmannia silvicola ا Weinmannia‏ Weinmannia —‏ Weinmannia sp. nov. 3 Weinmannia auriculata Weinmannia tinctoria Un N S833 3333 AAAA zc 2 3 8383383 gs БЕ Sis 3SES ЖЕЕ за > LEES €IUUEUIUI9 AA Weinmannia trichosperma S 33 = 3 3 Е & & & з = = 32222 & RR y SRA $ & 2% F = = % 333*32 & & ASS 99]9[DOIOS? y Weinmannia arguta Weinmannia bojeriana E Weinmanni Weinmannia stenostachya Weinmannia sp. nov. 2 Weinmannia madagascariensis Weinmannia rutenbergii oes1odsu] 73 Pancheria brunhessii [— —— Caldcluvia paniculata | — L—— Spiraeopsis celebica Figure 1. Outgroups Strict consensus of >30,000 most parsimonious trees of ITS-1 ше псев. “Fast” bootstrap values аге above branches. Each section of Weinmannia is labeled to the right of the tree DISCUSSION In several respects, the phylogeny of the Cunon- iae resulting from the combined analysis of ITS and trnL-F (Fig. 3) is highly congruent with previously published studies using morphological characters (Bradford, 1998). It differs primarily by showing that Weinmannia sect. within section Fasciculatae. Most significantly, the Weinmannia is not nested combined data support the monophyly of Panch- eria, Cunonia (albeit weakly), and all five sections of Weinmannia. This analysis also suggests that Malagasy taxa form a clade. Although the combined ‚ “Fast” bootstrap and decay analyses do not give much data set does produce a highly resolved tree support for internal nodes, indicating that relation- ships between major clades within Cunonieae are still poorly understood. In contrast to the results of Bradford (1998), the molecular data sets do not uphold the monophyly Volume 89, Number 4 2002 Bradford 497 Molecular Phylogenetics in Cunonieae 98 63 59 Weinmannia bojeriana = Weinmannia sp. nov. 2 Weinmannia minutiflora Weinmannia dichotoma Weinmannia racemosa Weinmannia silvicola Weinmannia raiatensis шпшщләд$отәт Weinmannia samoensis Weinmannia clemensiae Weinmannia — Weinmannia richii oeje[norose,] Weinmannia sanguisugarum Weinmannia louveliana Weinmannia madagascariensis Weinmannia rutenbergii PRES [. Cunonia capensis р, h Н J, Ў, | о | » oes1odsur W oejeoidg Weinmannia marojejyensis h. Cunonia balansae Cunonia macrophylla Pancheria hirsuta Pancheria phylliraeoides Weinmannia bangii Weinmannia trichosperma Weinmannia tinctoria Weinmannia tomentosa ттийешшәү\ Vesselowskya venusta Callicoma serratifolia Codia discolor Pullea cf. glabra Figure 2. Strict consensus of trnl-trnF trees. “Fast” mannia is labeled. Cunonieae taxa are shown in bold type of Weinmannia. Weinmannia sect. Weinmannia has a very long branch and is placed as a sister group to Cunonia and other Weinmannia (Fig. 3). long branch and the short internal nodes make it This possible that these results are not dependable. To test whether this data set could statistically reject the hypothesis of a monophyletic Weinmannia, | built a constraint tree in MacClade (Maddison & Maddison, 1992) making Weinmannia monophylet- ic and enforced this topology in PAUP while re- analyzing the combined data set. A single tree of 390 steps was found, only one step more than the tree found in the unconstrained analysis. A Ackama paniculosa 84 Ackama rosifolia Caldcluvia paniculata loh ZU PF 100 Geissois benthamii Geissois superba Á— values are above branches. Each section of Wein- COXON the most parsimonious tree with the monophyletic Weinmannia tree (Templeton, 1983; Mason-Gamer & Kellogg. 1996), and the null hypothesis of a monophyle tic tac could not be rejected N ‚Т = 2. P = 0.56). It would therefore be pre- mature to — Bosse paraphyletic based signed-rank test was then used to compare on this data. Morphologically, Weinmannia is recognized eas- ily by the presence of hairs on the seeds, which are lacking in other Cunonieae. Also, and Vesselowskya have winged seeds, but Pancheria, Cu- попа, i To account for wings are lacking in Weinmannia. 498 Annals of the Missouri Botanical Garden Weinmannia bojeriana d Woi ia marpiei Weinmannia dichotoma New Caledonia Weinmannia raiatensis Society Islands Weinmannia samoensis Samoa Weinmannia racemosa New Zealand ogroed qnos Weinmannia silvicola New Zealand Weinmannia sanguisugarum Weinmannia sp. nov. 2 IeoseaepejA Weinmannia madagascariensis Weinmannia rutenbergii Weinmannia clemensiae Malesia Weinmannia fraxinea Cunonia macrophylla ^ Melanesia Weinmannia richii Cunonia atrorubens nnnin New Caledonia Cunonia capensis — South Africa Ў W, nn han gii e 60 g 1 Tropical America NS 81 Weinmannia tomentosa 96 $ A 1 4100 Weinmannia tinctoria = Mascarenes 19 100. 79 12 < 100 |? 11 Caldeliuyi Pancheria phylliraeoides m Weinmannia trichosperma — Temperate South America Pancheria engleriana New Caledonia Pancheria hirsuta Vesselowskya venusta — Eastern Australia paniculata = South America —— 5 changes Figure 3. given above branches, de the geographic occurrenc of monoecious, usually with асе flowers. diu uM be En single most parsimonious T below. The values these character states with a paraphyletic Wein- mannia, two additional morphological steps are re- quired: either seed hairs were gained twice and seed wings were lost twice (once on each of the two Weinmannia lineages), or a reversal of both char- acters occurred in Cunonia (Fig. 4). Other char- acters supporting the monophyly of Weinmannia == celebica = Malesia spec le 9. Spec ies plac ‘ed in bold type Outgroups tree from the combined analysis. “Fast” bootstrap values are > clades corresponding to each section of Weinmannia are labeled, as is have a dioecious breeding system; others are have been found in micromorphological studies by R. Barnes (see Bradford & Barnes, 2001) in which multicellular hair bases were found in all sections of Weinmannia, but not in every species. These kinds of hairs have never been found in other gen- era of Cunoniaceae. Although combined molecular data provide little Volume 89, Number 4 2002 Bradford 499 Molecular Phylogenetics in Cunonieae Hairs on seeds Wings absent Weinmannia Cunonia Wings absent Hairs on seeds Weinmannia Pancheria Vesselowskya Weinmannia Wings present No hairs on seeds Hairs on seeds Cunonia Wings absent Weinmannia Pancheria Vesselowskya Hairs on seeds Wings absent Weinmannia Cunonia Pancheria Vesselowskya Figure 4. Alternative hypotheses for relationships — Cunonieae. —a & b. Weinmannia is pue with respect to Cunonia. This involves 38 o cs ular data set, ones. —4 'inmannia is mot 390 — in * с ambined molecular data set, and 2 more morphological one support for the cladistic relationships among Pancheria, Cunonia, and Weinmannia sections, new insights into relationships within genera and sections have emerged. The South African species Cunonia capensis was recognized by Bradford (1998) as being morpholog- ically similar to two very distinctive New Caledon- ian species, C. macrophylla and C. schinziana. All three species have larger flowers and fruits than other Cunonia, and similar inflorescence architec- tures composed of axillary pairs of stout racemes at the ends of stems. This inflorescence architecture is more reduced than the compound racemes typi- found in other cally Cunonia (Hoogland et al., 1997). Morphological cladistic analyses. showed this group of species to be a highly derived clade within Cunonia. In contrast, the combined molec- ular data make Cunonia capensis basal within Cu- nonia, and C. macrophylla basal within New Ca- ledonian Cunonia (Fig. 3). ost Cunoniaceae, including Cunonieae, are pollinated by small flying insects, especially bees. The distinctive features of Cunonia capensis and C. macrophylla may be due to their unusual pollina- tion biology. Observations show their visitors in- clude nectar-feeding birds that perch at the base of the raceme (Hopkins, pers. obs.; Coates Palgrave, 1983), which makes sense considering their rela- tively large flowers and simple. rigid inflorescence structure. The combined cladogram (Fig. 3) sug- gests that similarities between Cunonia capensis and Cunonia macrophylla may be plesiomorphic. Given that bird pollination is only known in one other Cunoniaceae genus (Geissois), it seems un- likely that bird-pollination was plesiomorphic dur- ing the origin of Cunonia, although it may be for the extant clade. Alternatively, bird pollination may be convergent in Cunonia capensis and C. macro- phylla, but testing these hypotheses for ancient lin- eages is nearly impossible. Although molecular systematics only gives weak support to the monophyly of Cunonia, the genus is morphologically distinct. At least two characters are shared only by species of Cunonia: fruits that have a circumbasal-acropetal dehiscence, a char- acter unique in the family, and floral disks that are adnate to the base of the ovary, unlike any other Cunonieae (Bradford, ds Bradford & Barnes, 2001; Rozefelds et al., 2001) The most — clade in the analyses is This is the that of Weinmannia sect. Weinmannia. argest section in the genus and is disjunct between the Americas, where over 70 species occur, and the Mascarene Islands, where 2 species are endemic. The Mascarene species are distinguished by being but otherwise are similar to American species (Bradford, 1998). Although the topology is not strongly supported, Mascarene species, repre- dioecious. sented here by Weinmannia tinctoria, are nested between W. trichosperma, from temperate forests of South America, and a clade of species from Neo- tropical. montane forests. Weinmannia trichosperma is apparently a rem- nant of a more ancient, temperate lineage within Weinmannia sect. Weinmannia—a lineage that macrofossils show may have once occurred in Tas- mania as well (Carpenter & Buchanan, 1993). That a derived tropical clade of the section is disjunct between the neotropics and the Mascarenes sug- gests that interchange across the Atlantic and In- 500 Annals of the Missouri Botanical Garden Y €————— Racemes ——— al d 4 XÊ . Y - Р * М, m" N N. A T Y y Y e "мү NTP * К TT $ N Y Б = ЖФ. > y 7 NL a М ` als 4 y Li ES "d t H A for " at ar > Ж ALI [LPS 5 = LZ i - : Inflorescence Modules (IM) Figur Two common inflorescence architectures in. Weinmannia are shown; left, section Leiospermum; right, section Tec latae. Racemes, internodes, and meristems are organized in particular patterns to form higher order units called Inflorescence Modules (IMs), and IMs are organized in particular patterns along the main stem to comprise the Total Inflorescence. dian Oceans has occurred more recently than in- terchange between temperate and tropical America. However, few molecular characters support this unexpected relationship, additional data are required to test it. In Bradford (1998), Weinmannia sect. Fascicu- latae was paraphyletic with respect to a highly de- since only a rived Weinmannia sect. Weinmannia. This view is overturned by the molecular evidence. Weinmannia sect. Fasciculatae has high bootstrap and decay values (Figs. 2—4). Missing from this study, how- ever, is sequence data for Weinmannia descombes- iana, an unusual species placed in section Fasci- culatae, but with some traits suggestive of section Leiospermum (Hopkins, 1998b). Weinmannia sect. Leiospermum is widely distrib- uted in the South Pacific, from the Bismarcks to the Marquesas, and from Rapa to New Zealand (Hopkins, 1998a, 1998c: Hopkins & Florence, 1998). Bradford (1998) recognized three species groups within the section: New Caledonian species, New Zealand species, and other South Pacific spe- cies. The two species from New Zealand, Weinman- nia racemosa and W. silvicola, each have distinctive and unique inflorescence architectures. South Pa- cific species outside of New Caledonia are dioe- cious. As a whole, New Caledonian species have no obviously derived features; however, Weinman- nia dichotoma aborts its terminal meristem at every node. Although each individual data set is ambig- uous, the combined analysis strongly supports the monophyly of section Leiospermum and places the ew Caledonian clade sister to clades from the South Pacific and New Zealand (Fig. 3). A well-supported phylogeny presents an oppor- tunity to reevaluate some of characters 1 discussed in previous studies (Brad- ~ the morphological ford, 1998). This earlier work emphasized inflores- cence architecture and heterotopy, and some back- ground on this is warranted here. The inflorescence in Cunonieae comprises nested sets of structures. The most familiar of these are the flower-bearing ost genera, but ball-shaped 2001). These flower- bearing axes, along with internodes and meristems, axes, raceme-like in m in Pancheria (Rozefelds et al.. are typically arranged in repeated units | term In- florescence Modules (IMs) (Fig. 5). The structure of IMs and their arrangement is highly variable among Cunonieae genera and Weinmannia sections (Fig. 6). This observation led to a system of coding in- florescence architecture based on principles of po- q . Molecular systematics has confirmed the general sitional homology (see Bradford, 1998, for details perception from comparative morphology that inflo- rescence evolution represents heterotopy, and that these characters can be effectively coded for mor- phological cladistic analyses. For example, phylo- genetic support for some clades in the morpholog- ical cladistic analysis was based mainly on these characters. The monophyly of Weinmannia sect. Leiospermum was supported by two characters of the inflorescence: having a sequential arrangement of metamers within an IM, and having the largest IMs in the terminal 5, 6, 7). Furthermore, improved support for clades within — i.e., acrotonic) position (Figs. section Leiospermum can be used to make specific statements about the pattern of heterotopic chang- es. From the generalized inflorescence form of sec- tion Leiospermum (see top and bottom left diagrams Volume 89, Number 4 2002 radford 501 Molecular Phylogenetics in Cunonieae LA Leiospermum Inspersae Spicatae Fasciculatae Weinmannia Cunonia Figure 6. phylogeny. The phylogeny is based upon Figure 3. but though ities to both sections cl ide 9 havi were collapsed. Cunonia and Pancheria — fences are t C. macrophylla and C. capensis are similar to section Боан Fasciculatae and Leiospermum. For detailed illustrations of inflorescence divers Pancheria Vesselowskya The general breadth of inflorescence diversity in Weinmannia is illustrated and linked to taxonomy and ng no bootstrap suppor rt and a decay value of only є Ч ) ч r see REM (1998, 2001), Hoogland et al. (1997), the publication series of Hoskins (1998), and ren felds et al. (2001). of Fig. 7). it is clear that species from New Zealand (Weinmannia racemosa and W. silvicola) have de- rived heterotopic changes in their inflorescences (see right diagrams of Fig. 7). Weinmannia race- mosa has regained vegetative growth beyond the inflorescence with a terminal vegetative bud, a re- versal to the plesiomorphic state found in other sec- tions of Weinmannia (Fig. 6). This species has also lost the development of lateral inflorescence mod- ules. The other species from New Zealand. Wein- тапта silvicola, has developmental asymmetry: al- though it does not always produce lateral IMs, when present, they develop only from one axillary bud. Furthermore, Weinmannia silvicola is the only spe- cies to produce sequential metamers within an IM and then abort the apical meristem. Other species, such as Weinmannia dichotoma from New Caledon- ia (see left diagram in Fig. 7), abort the apical mer- istem at the first metamer within an IM. Although abortion of the apical meristem occurs as part of normal variation among IMs within many Weinman- nia plants in the South Pacific, the fixation of this trait is apparently derived within Weinmannia di- chotoma. (See also Hopkins, 1998a, 1998c; Hop- 1998.) Without a more resolved phylogeny it is difficult kins & Florence, to re-evaluate other characters discussed previously Bradford, 1998), Although the — such as dioecy. strong phylogenetic hypothesis for Weinmannia sect. Leiospermum indicates that dioecy evolved once within this clade (Fig. 3). it is difficult to dis- cern the general pattern of breeding system evo- lution within Cunonieae (see also Bradford, 1998: 2001). Rozefelds et al., CONCLUSIONS Molecular systematies has enabled us to confi- dently delineate some major lineages within Cu- nonieae and provided sufficient resolution in some Un- fortunately, little is known still about how Cunon- clades to re-examine inflorescence evolution. ieae genera and Weinmannia sections relate, except Until a better phylogenetic hypothesis is available, it is for the basal position of Vesselowskya (Fig. 6). м best to retain the genera as currently circum- seribed. despite questions about the monophyly of Weinmannia. Remaining to be addressed is char- acter evolution for many traits that vary among gen- era of Cunonieae and sections of Weinmannia, al- though evolution within some genera and sections has been clarified. This work is the first in-depth айели to under- stand the phylogeny of Cunonieae. While taxon sampling was broad, the number of sequenced loci was likely insufficient to resolve all nodes. Ques- 502 Annals of the Missouri Botanical Garden Bismarcks, Solomons, Vanuatu, Fiji, Samoa, Cooks, Societies, Rapa, Marquesas New Caledonia Weinmannia dichotoma e.g. Weinmannia serrata Figure 7. The number of species and location of each clade is li tions about character evolution may best be studied by comparing variation within species, sections, and genera rather than at the tribal level. For ex- ample, Malagasy Weinmannia have the richest va- riety of inflorescence architecture, sympatric Cu- nonia species have a diversity of floral coloration and scents, and in many Weinmannia species from Malesia and the South Pacific is “leaky” with morphologically male, female, and bisexual dioecy flowers found within a single population or plant. Most importantly, studies on the spatial ecology of species are almost totally lacking and would be useful for effective conservation measures. Literature Cited Angiosperm Phylogeny Group. 1998. An ordinal classifi- cation for the families of flowe "ring plants. Ann. Missouri ot. Gard. 85: 531-553 ,. 1992. Phylogenetic utility of the internal transcribed spacers of ribosomal DNA in plants: An e.g. Weinmannia samoensis sp abeled. Some common forms of inflorescence archite shown, and the most unusual species, Weinmannia silvicola, W. racemosa, and W. dichotoma, are highlighted. New Zealand Weinmannia racemosa Relationships among clades within Weinmannia sect. Leiospermum based on the analyses presented here. } | cture are — from the Compositae. Molec, Phylogenet. Evol. 1: Bradford. j. C. 1998. A analysis of species- groups in Weinmannia (Cunoniaceae) based on mor- К апа inflorescence architecture. Апп. Missouri Bot. Gard. 85: 565—593. "2000 . Phylogenetic Systematics of Cunoniaceae (Oxalidales). with an Emphasis on Spec nflorescence E volution in Genera. $a cladistic . Thesis, Washington nd St. Louis. . 200 The applic ation of a cladis the КҮ e ‘ation and identification of lada (C u- noniaceae) in o ar and the Comoro Islands. Ad- ansonia, sér, З, : 3: 237-246. & R. В 2001. Phylogenetics and clas- sification of Cunoniaceae (Oxalidales) using chloroplast DNA sequences and morphology. Syst. Bot. 26: 354— 385. &I . Berry. 1998. Cunoniaceae. Pp. 462—469 in P. E Be чту, В. К. Holst & К. Yatski evych (vol. edi- tors), Pos of me Venezuelan Guayana, Vol. 4. Missouri Botanical — n — St. & г. 2001. notes on he йога of the M ouis. New taxa and nomenclatural arojejy massif, Madagascar— Volume 89, Number 4 Bradford 503 Molecular Phylogenetics in Cunonieae V. Cunoniaceae: Weinmannia. Adansonia, sér. 3, 23: о Buckler, — banal DN — netic implications. с netics 145: 812—832. Carpenter, R. J. & A. M. Buchanan. 1993. Oligocene leaves, p and flowers of the Cunoniaceae from Ceth- ana, Tasmania. — Syst. Bot. 6: 91-109. Coates Palgrave, Trees of Southern Struik Poblishen. Ca ape m wn. Eriksson, T. 1999. AutoDecay Ver. 4.0. Computer program distributed by the author. Bergius Foundation, Royal Swedish Ac — of r iences. Stockholm. A. Ippolito yi T. P. Holtsford. 1997. The : Divergent paralogues and Africa. Greuter, W., J. McNeill. Barrie, H. M. Burdet, De — TS. iuc j^ T Nicolson, P. C. Si J. E. ‚Р. Trehar urland & D. L. Hawk- sent (editors) 2000. Inte d Code of e al menclature (Saint Louis Code). Regnum Veg. 138. Guillaumin, A. 1940. Matériaux pour la Flore de — LVL Revision des Cunoniacées. s . France 87: 242-256. тена 3 1999. Flora of Ecuador; 78. Cunoniaceae. De- partment of Systematic Botany—Institute of Biological > iences, University of Aarhus, Nordlandsvej 68. dk- 3240 Risskov, Denmark. Ha R. D., J. Jérémie & H. C. F. Hopkins. 1997. enre Cunonia (Cunoniaceae) en Nouvelle-Calédon- De SC de cinq espéces nouvelles. Adansonia. la Nou- Bull. 3, Hopkins, x C F. 1 998 a. noniaceae) in Malesia and the Pac and an account of the species of Western Malesia. the Adansonia, sér. 3, 20 A revision of d (Cu- fic 1. Introduction Lesser Sundas and the Moluccas. 541. 1998b. A revision of Weinmannia (Cunoniaceae) in Maleaia and the Pacific. 2. Sulawesi and the Phil- — sér. 3. 20: 43-66. 1908 revision of E einmannia (С unoniaceae) in — and the Pacific. Guinea, Solon Islands, Vanuatu and Fiji, ge notes on the species « Samoa, Rarotonga bog Caledonia and New Zealand. Adansonia, sér. б 11-106. & J. p A revison of Weinmannia (Cunoniaceae) in Malesia and the Pacific. 4. The So- ciety, Marquesas and Austral Islands. Adansonia, sér. 3, 20: 107—130. ippine S. Kuzoff, I J . D. E. Soltis, P. S. Soltis & E. A. ибо — lo potential ы entire S rDNA sequences in plants. Molec. Biol. Evol. 15: 251-263. Maddison, W. . R. Maddison. 1992. MacC * Analysis of — and Character Evolution, V 3 Sine r, Sunderland, Massachusetts. Mason-Gamer, R. J. & E. A. Kellogg. 1990. Testing for р ylogenetic conflict among molecular data sets in the ribe Triticeae ги пеае). Syst. Biol. 45: 524—545. Mane Р. 1993. Our knowledge of the flora of New Cale- donia: — and diversity in relation to vegetation types and substrates. Biodiversity Lett. 1: 72-81. Rambaut, A. 1995. Se-Al, — program distribute by the Кетем Oxford. Rozefelds, A. C., Н. W. nes & B. Pellow. 2001. A new species "and shi cal morphology of ا‎ lowskya 175-192. equence Alignment Program. (Cunoniaceae). Austral. Syst. Bot. 14: Swofford, D. L. 2001. PAUP*: Phylogenetic — Us- ing Painon, Ver. 4.0b6a. Sinauer, Sunderland, Mas- Uni- Gielly, G. Pautou & J. Bouvet. 1991. versal | s for amplification of three non- — re- gions of eorla DNA. PI. Molec. Biol. 1105- 1109. 1983. Phylogenetic inference from re- Temple ton, A. R eavage sile maps with partic- striction endonuclease c ular reference to the evolution of humans and the apes. 244. Evolution 37: 221— APPENDIX 1 The p are correct orthographies (following Greu- ter et al., JO, ICBN Art. 21.2) and original publications of all five pep see of Weinmannia. Types are also desig- naled here for two sections. . Nat., T 10, 2: 997, 1005, 1367. < Weir nmannia pinnata Сайан. 18: Welomannis Le; 1759, nom. cons Simplic TM Bernat l., sine typo. тла sect. 289. 1963, nom. D Weinmannia sect. ospermum (D. Don) Engl., Nat. flanzenfam, З (2а). 101. 1891. Leiospermum D. Don, Edinburgh New Philos. м 9: E 1830. TYPE: Weinmannia racemosa L A Suppl. 227. 1781. (lec- totype, designated by H Uu 19982: 21) Weinmannia sect. Racdmasae Berman, Bot. Jahrb. Syst. 83: 132, 185. 1964. — Weinmannia sect. Fasciculatae Be mardi ex — & H. C. Hopkins, Adansonia, sér. 3: 998. с Пав fraxinea (D. Don) a Sides sect. Inspersae Bernardi ex J. C. sect. nov. Weinmannia Bot. Jahrb. Syst. 83: 132, sine typo. TYPE: — d DC This section was originally published without a ons Bradford, spe n 1€ es Weinmannia - sect. Spi sect. nov. Weinmannia sect. Jahsh Syst. 83: 132. 1964, nom. T : Weinmannia bojeriana Tul. шш published without a type species catae Bernardi ex J. C. Bradford, nicatae Bernall, Bot. invali d.. sine typo This section was — CONTRIBUCION A LA Sandra S. Aliscioni? FILOGENIA DEL GÉNERO PASPALUM (POACEAE: PANICOIDEAE: PANICEAE)! RESUMEN El género Paspalum L. presenta alrededor de 330 especies y una amplia distribución geográfica, — ipalmente en las regiones tropicales y subtropicales de América. Debido a la marcada variabilidad mo rfológica entre sus especies, — autores intentaron dividir al género en subgéneros, secciones о grupos informales de especies. En el presente rabajo se realizó un análisis cladístico del género Paspalum utilizand s anatómicos foliares y exomorlológic os con la — de poner a prue ‘ba su origen monofilético y кегин ‘er relaciones filogenéticas entre las especies mas representativas. Asimismo se analizó el grado de homoplasia de los caracteres con e D to de estimar su valor diagnóstico y calidad informativa en la caracterización de los distintos grupos. Para el anális 47 especies pertenecientes a 27 grupos establecidos según Chase con algunas modific 'aciones propuestas | por Ае lla et al. y Morrone et al. Se identificaron 36 caracteres exomorfológicos y anatómicos foliares. La matriz de datos fue analizada bajo el criterio de parsimonia con pesos iguales y con — implicito de caracteres utilizando los programas Nona y Pee-Wee respectivamente. Sobre la base de los caracteres utilizados en este análisis no pudo establecerse un origen común para el género confirmándose que el mismo representa una asociación parafilética de especies, estrec mente relacionadas con Axonopus P. Beauv. y Thrasya Kunth, géneros también pertenecientes a la tribu Paniceae. 3 ч ABSTRACT Paspalum L. has — ‘ly 330 species and a large geographic : distribution, mainly in tropical and subtropical regions of America. Due to the morphological variability among the species, different authors have proposed dividing the genus into ш sections, or informal groups. A cladistic ТИР of Paspalum was conducted using anatomical foliar and оре characters to test its monophyly and its phylogenetic relationships among representative species. The degree of character homoplasy was have kinks in order to estimate the diagnostic value of anatomical cha racia rs — groups. For the cladistic analys ed species of 27 groups established by Chase, with modification ble by Cialdella et al. and Morrone et к nd 36 exomorphologi ‘al and. anatomical foliar characters were considered. The Yo ata matrix was analyzed with Nona oa Pee-Wee, parsimony programs using equal weights and implied weighting. The present phylogenetic analysis confirms that Paspalum is a paraphyletic group and shows close relation- ships with Axonopus and Thrasya in Paniceae. Key words: Axonopus, Paniceae, Panicoideae, Paspalum, Poaceae, Thrasya. El género Paspalum L. es uno de los más im- La importancia económica del género radica en portantes dentro de la tribu Paniceae debido al que muchas de sus especies son consideradas ex- elevado número de especies que presenta y a su celentes forrajeras naturales, como por ejemplo Р. amplia distribución geográfica, habitando princi- notatum Fliiggé, “pasto horqueta” y P. dilatatum palmente regiones tropicales y subtropicales de Poir., “pasto miel." América, con pocas especies en África y Asia. La El género incluye hierbas perennes, rara vez falta de un estudio taxonómico global para este gé- anuales, cespitosas o rastreras, estoloníferas o ri- nero hace difícil estimar en la actualidad el número zomatosas. Se distingue de otros miembros de la total de especies. Según Chase (1929) el mismo tribu por presentar inflorescencias con racimos es- incluye aproximadamente 400 taxones; Clayton y piciformes unilaterales; articulación de la espigui- Renvoize (1986) consideran que comprende alre- Ila en la base junto al pedicelo; espiguillas abaxia- dedor de 330 especies. les, solitarias o apareadas, dispuestas en 2 a 4 ! El presente trabajo conatituye parte de la Tesis Doctoral prese ада en la Fac 'ultad Es Ciencias ач Че 1а Universidad Nacional de La Plata. Expreso mi — "imiento a mis directores, el Dr. F. Zuloaga y la . M. Arriaga por todo el apoyo — y a mis compañeros y lectura crítica del manus ? Instituto de Botánica Darwinion. Labardén 200-C.C. 22-B 1642 HYD San lsidro, Buenos Aires, Argentina. saliscioni@darwin.edu.ar. ANN. Missouni Bor. Garb. 89: 504—523. 2002. \agesen, L. Giussani y O. Morrone por las sugerencias aportadas Volume 89, Number 4 2002 Aliscioni : 505 Filogenia de Paspalum hileras, dorsiventralmente comprimidas, plano-con- vexas o cóncavo-convexas; gluma inferior usual- mente ausente y antecio superior endurecido obo- voide a elipsoide (Chase, 1929; Burkart, 1969; 1990). Debido al elevado nümero de especies y a la Judziewicz, marcada variabilidad morfológica entre las mismas, diversos autores han intentado dividir a Paspalum en subgéneros, secciones o grupos informales, de- tectándose dificultades para dicha delimitación. Estas categorías fueron establecidas sobre la base de caracteres exomorfológicos, principalmente a ni- vel de la espiguilla (pilosidad de la espiguilla, mor- fología de la gluma superior, textura y coloración del antecio, etc.) sin considerarse en la mayoría de los casos los caracteres anatómicos foliares. Los es- tudios taxonómicos llevados a cabo hasta el mo- mento son fragmentarios: el género Paspalum ha sido tratado en revisiones parciales y floras regio- nales (Chase, 1927; Parodi, 1928, 1932, 1937: Hitchcock, 1951; 1954, 1957, 1966, 1967: Burkart, 1969; Rosengurtt et al., 1970; Quarín. 1975; Sendulsky & Burman, 1978, 1980a. b: da Silva et al.. 1979: Pohl, 1980; Smith et al., 1982: Renvoize, 1984, 1987a: Crins, 1991; Judziewicz. 1990; Filgueiras, 1993; Cialdella et al.. 1995: Mo- 1995, 1996, 2000; Rodriguez, 1998; 2000). Cabe destacar el valioso Barreto, rrone et al., Rodriguez et al., aporte realizado por Chase (1929) al estudiar las especies de Paspalum que crecen en América del Norte, estableciendo 27 grupos de especies con un total de 140 taxones distribuidos en los mismos. La anatomía foliar de Paspalum fue estudiada por Tiirpe (1967), quien describió detalladamente las características histofoliares de las especies ar- gentinas sin aportar datos que permitan inferir re- Arriaga (1998) lle- varon a cabo un estudio anatómico comparado en laciones entre ellas. Aliscioni y los grupos Quadrifaria y Virgata del género Pas- palum, destacando la semejanza anatómica entre ambos. Posteriormente, Aliscioni (1999) realizó un estudio histofoliar de las especies más representa- tivas del género, observando que Paspalum es ana- tómicamente heterogéneo presentando caracteres diferenciales a nivel de especie o grupo de espe- cies, algunos de los cuales mantienen cierta rela- ción con los agrupamientos establecidos sobre la base de caracteres exomorfológicos. Por otro lado la autora muestra que los únicos caracteres anató- micos comunes a todas las especies estudiadas son aquellos asociados al tipo de vía fotosintética, los cuales permiten corroborar que el género presenta anatomía Kranz del subtipo MS, fisiológicamente C, NADP-me. Tiirpe (1967) define a Paspalum como un género anatómicamente homogéneo y considera que los ca- racteres diferenciales que se pueden observar en las hojas, reflejan las condiciones ecológicas del medio en el cual habitan las especies. Sin embargo, Aliscioni (2000) estudió la anatomía ecológica de algunas especies de Paspalum y describió diferen- tes estrategias adaptativas presentes en el género destacando que especies de hábitats muy diferentes pueden manifestar características anatómicas si- milares. 2 Existen estudios filogenéticos realizados en 1 tribu Paniceae los cuales incluyen un escaso nú- mero de especies de Paspalum. Sobre la base de datos morfológicos, Zuloaga et al. (2000) no logra- ron resolver las relaciones dentro del género, mien- (2001) y Giussani et al. 2001) realizaron análisis moleculares concluyendo tras que Duvall et al. — en ambos casos que Paspalum sería parafilético. En el presente trabajo se realizó un análisis fi- logenético con el objeto de poner a prueba la mo- nofilia de Paspalum. Se plantea una hipótesis sobre a historia evolutiva del género estableciendo po- sibles relaciones filogenéticas entre las especies más representativas. Asimismo se analiza el valor diagnóstico de los caracteres exomorfológicos y anatómicos a la luz de las relaciones filogenéticas, comparando el grado de homoplasia presente en cada uno de ellos con el objeto de evaluar su im- portancia en la caracterización de los distintos gru- pos de especies. MATERIALES Y MÉTODOS SELECCIÓN DEL GRUPO INTERNO El análisis cladístico se realizó sobre la base de 47 especies correspondientes a 27 grupos infor- males de Paspalum establecidos según Chase (1929, inéd.) con modificaciones propuestas por Cialdella et al. (1995) para el grupo Bonplandiana y por Morrone et al. (1995, 1996) en los grupos Racemosa y Dissecta. Para seleccionar los taxones terminales se consideraron representantes de la mayoría de los grupos informales del género, tra- tando de abarcar la máxima variabilidad morfoló- gica como así también cubrir el área de distribu- ción geográfica que presenta el género en América. Se seleccionó una especie representativa de los grupos Conjugata, Dissecta, Gardneriana, Inaequi- valvia, Macrophylla, Orbiculata y Saccharoidea, y dos especies para cada uno de los restantes grupos. Cuando estuvieron disponibles, se seleccionaron más de seis ejemplares por especie. El Apéndice 1 muestra la lista de ejemplares examinados. 506 Annals of the Mou. Botanical Garden SELECCIÓN DEL GRUPO EXTERNO El grupo externo se seleccionó considerando ta- xones que presentan un gran nümero de sinapo- morfías potenciales compartidas con el grupo in- terno (Nixon & Carpenter, 1993). Para ello se seleccionaron representantes de géneros de la tribu Paniceae los cuales presentan similitudes morfo- lógicas y anatómicas con Paspalum (Burman, 1985; Nicora & Rügolo de Agrasar, 1987; Watson & Dall- witz, 1989; Crins, 1991; Morrone et al., 1993) y que están relacionados filogenéticamente en análi- sis realizados con datos moleculares (Gómez-Mar- tínez & Culham, 2000; Duvall et al., 2001; Gius- sani et al., 2001; Gómez-Martínez et al., inéd.). Dichos géneros son Thrasya, Axonopus y Anthae- nantiopsis Mez ex Pilger. El género Thrasya fue relacionado por varios au- tores con Paspalum, siendo ambos muy afines. Ni- cora y Rügolo de Agrasar (1987) describen a Thra- sya considerándola semejante al grupo Ceresia de Paspalum por compartir la presencia de inflores- cencias con racimos espiciformes solitarios y membránaceo o folioso. Por otro lado, Thrasya se diferencia de las especies de Ceresia por la apa- y raquis rente disposición uniseriada de las espiguillas y la presencia de la lemma estéril hendida. A diferencia de esto, Burman (1985) compara a Thrasya con los grupos Dissecta y Decumbentes del género Pas- palum refiriendose a ciertos caracteres de Thrasya, como por ejemplo la lemma inferior hendida, un 'arácter que ha sido observado en especies del gru- po Decumbentes, y la presencia de lemma inferior delgada la cual acerca a Thrasya al grupo Dissecta. Thrasya también comparte con Paspalum el mismo tipo anatómico y fisiológico de vía fotosintética. Giussani et al. (2001) y Duvall et al. (2001) reali- zaron análisis cladísticos con datos moleculares, mostrando la estrecha relación filogenética entre ambos géneros. En el presente análisis se seleccio- nó como — del género Thrasya a T. pas- paloides Kunt Watson y Dallwitz (1989) realizaron un análisis fenético en el cual muestran a Axonopus y Paspa- — lum como géneros próximos. También Crins (1991 cita la relación fenética estrecha entre ambos. Chase (1906) y Webster (1988) muestran que di- chos géneros sólo se diferencian en la orientación de la espiguilla en relación al raquis, siendo ad- axial en Axonopus y abaxial en Paspalum. En re- lación al tipo fotosintético, Axonopus es fisiológi- camente С, NADP-me y igual que Paspalum. Gómez-Martínez y Culham (2000) realizaron estudios filogenéticos con datos moleculares que confirman la cercana relación fi- anatomicamente MS, al logenética entre ambos géneros. En el presente análisis se seleccionó como representante del gé- nero Axonopus a А. compressus (Sw.) P. Beauv. Morrone et al. (1993) realizaron la revisión del género Anthaenantiopsis señalando que el mismo presenta algunos caracteres morfológicos comunes con Paspalum. Ambos géneros presentan especies cespitosas, con inflorescencias contraídas a ligera- mente abiertas, espiguillas pilosas solitarias o apa- readas, antecio superior endurecido, piloso o gla- bro, tipo fotosintético C,, subtipo anatómico MS y ntimero cromosómico básico x = 10. Por otro lado, Anthaenantiopsis se diferencia de Paspalum por presentar inflorescencia espiciforme no unilateral; espiguillas biconvexas, gluma, flor y pálea inferior siempre presentes. Sobre la base de los caracteres anatómicos foliares descriptos por Morrone et al. (1993) se seleccionó A. rojasiana L. Parodi como representante del género por presentar transcorte foliar expandido al igual que la mayoría de las es- pecies de Paspalum y haces vasculares con vaina parenquimática externa remanente, también se observó en P. inaequivalve (Aliscioni, carácter que El reconocimiento del tipo anatómico y fisioló- gico asociado a la vía fotosintética constituye una serie de caracteres valiosos para realizar clasifica- ciones (Hattersley & Watson, 1992), especialmente dentro de la tribu Paniceae la cual incluye géneros con diferentes tipos fotosintéticos. Con la finalidad de mejorar la resolución del grupo externo e im- pedir incorrectas topologías del grupo interno (Nix- on & Carpenter, 1993), se incluyeron como grupo externo representantes de algunos de los tipos fo- tosintéticos presentes en la tribu, diferentes al de Paspalum. De esta manera se seleccionaron Erio- chloa Kunth (tipo fotosintético C, РСК, subtipo anatómico PS) y Panicum laxum Sw. (tipo fotosin- tético C,). A pesar de que Eriochloa comparte un menor nümero de sinapomorfías con Paspalum, ambos gé- neros presentan inflorescencias constituidas por ra- cimos unilaterales, diferenciándose Eriochloa por la presencia de espiguillas con un callo basal dis- coide y número cromosómico básico x = 9 (Arriaga, 2000). Debido a que Eriochloa es un género mo- nofilético y anatómicamente homogéneo (Arriaga & Aliscioni, 2000) se seleccionó E. distachya Kunth como representante del género. Brown (1977) y Hattersley y Watson (1992) pro- ponen distintos esquemas evolutivos para la tribu Paniceae, pero en ambos casos sugieren a un gé- nero anatómicamente non-Kranz como precursor de la tribu entre los que incluyen representantes C, de Panicum. Por tal motivo se asignó Panicun la- Volume 89, Number 4 2002 Aliscioni Filogenia de Paspalum 507 xum Sw. como raíz del árbol filogenético, ya que dicha especie al igual que toda la sección Laxa del subgénero Phanopyrum es fotosintétic — C con anatomía non-Kranz (Zuloaga et al., 3 CARACTERES Se identificaron un total de 36 caracteres de los cuales 14 corresponden a caracteres exomorfológi- De los 16, 21, 23, 26,31 y 32) pennant a caracteres multiestado cos y 22 a caracteres anatómicos foliares. 36 caracteres utilizados, 7 de ellos (3, los cuales han sido tratados como no aditivos. La lista de caracteres se incluye en el Apéndice 2 y la matriz básica de datos en el Apéndice 3. Las observaciones se realizaron sobre más de 300 ejemplares de herbario pertenecientes a las si- guientes — iones: BA. — BAF, CAY, COL, CPUN . FCQ, G, HUT, LPB, MEXU, MO, NY, Р. P UNE SE, SL v USAM, USM, VEN, segtin siglas tomadas de Holmgren et al. (1990). Para la obtención de los caracteres anatómicos se realizaron transcortes foliares y preparados epi- dérmicos los cuales fueron observados mediante la utilización de microscopia óptica. Se seleccionaron caracteres histofoliares cualitativos que resultaron ser consistentes para cada una de las especies es- tudiadas, evitando incorporar aquellos que pudie- sen modificarse segtin el grado de desarrollo de la lámina. Los caracteres anatómicos fueron designa- dos siguiendo la terminología propuesta por Ellis (1976. 197‹ También se incluyeron todos los caracteres exo- morfológicos diagnósticos de los grupos informales de especies (Chase, 1929; Cialdella et al., 1995: Morrone et al., 1995, 1996; Zuloaga & Morrone, inéd.) correspondiendo principalmente a caracte- rísticas de las inflorescencias y J las espiguillas. ANÁLISIS FILOGENÉTICO El análisis cladístico se realizó sobre la base del 1983). Para ello se realizaron estrategias de búsqueda con principio de máxima parsimonia (Farris, pesos iguales y pesos implícitos de caracteres, uti- 1.8 (Goloboff. 8 (Goloboff, 1997b) Los caracteres fueron pesados de ands los programas ver. y Pee-Wee ver res- pectivamente. acuerdo a su grado de homoplasia, asignando me- nor valor a aquellos caracteres más homoplásicos (Goloboff, 1993). La relación entre el peso asignado a un carácter y su grado de homoplasia se expresa mediante una función cóncava que incluye una constante de concavidad (K) la cual puede variar segtin la intensidad con la que el programa pena- liza a los caracteres homoplásicos. Ramírez (1998) sugiere que el valor óptimo de concavidad en un análisis, sería matriz-dependiente y plantea que la utilización de funciones moderadas de pesado (K > 3) encuentra patrones jerárquicos más predicti- vos. Goloboff (1995) considera que funciones de pesado muy intensas (valores más bajos de K) no deberían ser usadas, ya que en la práctica actúan de manera similar a un análisis de clique. En el presente análisis se discuten los resultados obte- nidos con pesos iguales y con pesado implícito de caracteres bajo el valor medio de constante de con- cavidad (K — 3). Debido al elevado nimero de taxones terminales se optó por el método de búsqueda heurístico em- pleando la opción mult* con 100 replicaciones la cual genera árboles de Wagner con secuencias de adición al azar en cada réplica, seguido de reaco- modamientos del tipo TBR, reteniéndose 20 árboles por cada réplica (h/20). ron bajo la opción amb- la cual colapsa las ramas Las búsquedas se realiza- que están apoyadas solamente por optimizaciones ambiguas. Las ramas de los árboles obtenidos fue- ron posteriormente permutadas con el comando max*. Con el objeto de sortear posibles islas de árboles (Maddison, 1991) se utilizaron las opciones Jump*1, jump*2 y jump*3 las cuales realizan rea- comodamientos del tipo TBR en árboles con Та 3 pasos más largos respectivamente. Por último se obtuvo el árbol de consenso estricto mediante la opción nelsen (nel). Como medida de apoyo para establecer el so- porte de los diferentes clados se calculó el índice de Bremer (BS) (Bremer, 1994) utilizando Nona ver. 1.8 (Goloboff, 1997a) para el análisis con pesos iguales y Pee-Wee ver. 2.8 (Goloboff, 1997b) para el análisis con pesos implícitos. Asimismo, se cal- culó la frecuencia de grupo de Jackknife (Farris et al., 1996) utilizando el archivo de instrucciones de Nona/Pee-Wee jak.run, (con 1000 repeticiones, eli- minando al azar el 36% de los caracteres de la matriz y utilizando la opción de búsqueda mult*5) y el programa fq.exe para calcular el árbol de con- senso de mayoría (Goloboff, 1997a, b). Para graficar los árboles y optimizar los carac- teres se utilizó el programa Winclada Beta version 0.9.9 (Nixon, 1999). La optimización de caracteres fue posteriormente verificada con el comando apo/ de los programas Nona (Goloboff, 1997a) y Pee- Wee (Goloboff, 1997b). RESULTADOS El análisis cladístico realizado con pesos iguales de caracteres (matriz analizada con Nona) encontró 2903 árboles, todos ellos igualmente parsimoniosos 508 Annals of the Missouri Botanical Garden = 0.29, Ri = 0.09), Utilizando pesado implícito de caracteres, Pee- Wee y de 143 pasos de longitud (Ci encontró 28 árboles de 151 pasos de longitud (fit 1.2 y fit re-escalado = 43%). A continuación se describen los resultados obtenidos del análisis con Pee-Wee, mencionando las diferencias presen- tes en el análisis de Nona. Ambos análisis muestran que Paspalum se com- porta como un grupo parafilético constituyendo un clado de alto soporte (bs — 2) junto con Thrasya y Axonopus (Fig. 1), sustentado por tres sinapomor- fías: ausencia de gluma inferior (carácter que re- vierte secundariamente en Thrasya); antecio coriá- ceo (carácter que revierte secundariamente en P. stellatum y P. polyphyllum del grupo Ceresia [ Pers. | chb.; P. bertonii y P. lilloi del grupo Bertoniana [sensu Chase, inéd.]; у P. saccharoides del grupo Saccharoidea [sensu Chase, inéd.]); y ausencia de vaina parenquimática rodeando los haces vascula- res (Fig. 2A, tituye el grupo hermano de dicho clado (Fi B). Asimismo, Anthaenontiopsis cons- En el análisis realizado con Nona, el clado con- formado por Paspalum + Thrasya + Axonopus se observa con un alto grado de colapsamiento; a di- ferencia de ello el árbol de consenso obtenido a partir del análisis realizado con Pee-Wee muestra una topología más resuelta. Sin embargo ambos análisis coinciden en que los tinicos grupos del gé- nero Paspalum que constituyen clados monofiléti- cos son Bertoniana (sensu Chase, inéd.) con bs — 1995) con bs — 0.5, Anachyris Chase, con bs — 5 y Filiformia 1.3. Racemosa (sensu Morrone et al., — (sensu Chase, 1929) con bs = 0.2; y que el grupo 1995) e rafilético constituyendo un clado monofilético como Bonplandiana (sensu Cialdella et al., s pa- grupo hermano de Racemosa (bs = 0.8). Por otro lado, los grupos Caespitosa (sensu Chase, 1929), 1929), Eriantha (sensu Chase, inéd.), Fasciculata (sensu Chase, 1929), Linearia (sensu Chase, 1929), Livida (sensu Chase, 1929), Notata (sensu Chase, 1929), Paniculata (sensu Chase, 1929), Parviflora (sensu Chase, 1929), Plicatula (sensu Chase, 1929), Qua- drifaria (sensu Barreto, Ceresia, Corcovadensia (sensu Chase, 1966) y Virgata (sensu Chase, 1929) son claramente polifiléticos o cons- tituyen politomías no resueltas. El grupo Disticha (sensu Chase, 1929) forma un clado monofilético sólo en el análisis realizado con Nona, mientras que el grupo Dilatata (sensu Chase, 1929) es monofi- lético sólo en el análisis de Pee-Wee; aunque en ambos casos sustentados por un bajo soporte. A continuación se describen los grupos monofi- léticos resultantes del análisis realizado con pesos implícitos (Fig. 1) y las sinapomorfías que los sus- tentan (Fig. 2A, Paspalum lineare del grupo Linearia, junto con las especies del grupo Filiformia se unen en un 2.5) por compartir la presencia de porciones laterales del transcorte fo- clado de alto soporte (bs = liar inconspicuas o ausentes. Paspalum lineare es el taxón hermano del grupo Filiformia, siendo este último un grupo monofilético aunque de menor so- porte (bs = 0.2), caracterizado por la presencia de transcorte foliar de contorno semicircular (clado 1, Fig. 1). Paspalum proximun del grupo Linearia, es el taxón hermano del grupo Bertoniana uniéndose en un clado de bajo soporte (bs = 0.2) por compartir la presencia de parénquima incoloro en las alas del transcorte. Las especies del grupo Bertoniana for- man un grupo monofilético de alto soporte (bs = 1.3) sustentado por la presencia de transcorte foliar convoluto o subconvoluto (clado 2, Fig. 1). Paspalum orbiculatum del grupo Orbiculata, se comporta como taxón hermano del clado constituí- do por las especies de los grupos Plicatula, Gard- neriana y Virgata sustentado con alto soporte (bs = 1.3) por compartir la presencia de antecio obovoi- de. Las especies de los grupos Plicatula, Gardne- riana y Virgata se reúnen en un clado de alto so- porte (bs = 1.8) por compartir la presencia de antecio de coloración castaño (clado 3, Fig. 1). Todas las especies de Paspalum consideradas de los grupos Anachyris, Bonplandiana, Ceresia, Con- jugata, Corcovadensia, Dissecta, Livida, Parviflora, Racemosa y Saccharoidea, junto con una especie de los grupos Eriantha (P. ammodes), Caespitosa (P. y Quadrifaria (P. coryphaeum), forman un clado en el cual se inclu- — indecorum), Notata (P. pumilum) yen Thrasya paspaloides y Axonopus compressus (bs — 0.2). Dentro de este gran clado puede destacarse el nodo que retine a las especies de Racemosa y Bonplandiana (bs — 0.8), sustentado por la presen- cia de células distintivas Kranz, unido a P. repens = del grupo Dissecta como taxón hermano (bs = (clado 4, Fig. 1). Las especies P. malacophyllum y P. simplex del grupo Anachyris se reunen por l: ausencia de la gluma superior en un clado de alto — М soporte (bs = 5) (clado 5, Fig. Los restantes nodos del árbol de consenso, a pe- sar que algunos de ellos se muestran resueltos, no poseen sinapomorfías que los sustenten como cla- dos monofiléticos. n relación al peso que recibieron los caracteres 1 función de la homoplasia, los caracteres exo- Е #11 (nervios de la lemma superior), 13 (forma del antecio), 14 (coloración del antecio), y anatómicos: #15 (porciones laterales del trans- corte foliar), 24 (cloroplastos en la vaina mesto- mática), 26 (vaina parenquimática), 27 (células dis- Volume 89, Number 4 Alisc 2002 ioni 509 Filogenia de Paspalum PANICUM LAXUM RIOCHLOA ANTHAENANTIOPSIS Paspalum erianthum (Eriantha) P. quadrifarium (Quadrifaria) P. caespitosum (Caespitosa) Р. inaequivalve (Inaequivalvia — P. urvillei (Dilatata) P. fasciculatum (Fasciculata) . notatum (Notata) P. lineare (Linearia) 6 P. filiforme (Filiformia) P. lindenianum (Filiformia P. distichum (Disticha) P. vaginatum (Disticha) P. equitans (Fasciculata) P. prox imum (Linearia) P. bertonii (Bertoniana) P. lilloi (Bertoniana) P. orbiculatum (Orbiculata) P. modestum (Plicatula) P. virgatum (Virgata) P. plicatulum (Plicatula) P. gardnerianum (Gardneriana) P. regnellii (Virgata) P sa haroides (Sa ha id P. repens (Dissecta) P. bonplandianum (Bonplandiana) P. pallidum (Bonplandiana) P. penicillatum (Racemosa) P. pygmaeum (Racemosa) P. stellatum (Ceresia) THRASYA P. lividum (Livida) P. indecorum (Caespitosa) . mandiocanum (Corcovadensia) P. alcalinum (Livida м P. coryphaeum (Quadrifaria) . malacophyllum (Anachyris) P. simplex (Anachyris) P. crispatum (Parviflora) P. polyphyllum (Ceresia) x © US P. conjugatum (Conjugata) P. parviflorum (Parviflora) Figura 1. Arbol de consenso estricto obtenido a partir del análisis realizado con Pee-Wee (con pesado implícito de caracteres). El número sobre las ramas indica el valor de frecuencia de grupo de jackknife y el número ubic * debajo de las ramas indica el índice de Bremer. Los clados señalados (1-5) se discuten en el texto. Clado 1: P lineare (Linearia) + grupo Filiformia. Clado 2: P. proximum (Linearia) + grupo Dorn Clado 3: P. orbiculatum — ulata) + grupo Plicatula + grupo Gardneriana + grupo Virgata. Clado 4: P. repens (Dissecta) + grupo Racemosa + grupo Bonplandiana. Clado 5: grupo Anachyri Missouri Botanical Garden Annals of the 510 “SOx JuP[q sojn: ӘЛЛӘ UOD UE. р as soorspg[douiou S3139818 SO] А so1dou sojn: 919 uO2 UB: әри! 35 serpourodeurs se] enay el ua BIJSINW VS anb [9918 [9 ua ap UILINVO sisajugaed 31]u2 SOpE.LI2. IUD $3J19]2EI8» A s9[oq.re so] sopo} Ud UILINVO stsojugied әлә SOPE.113: ӘПӘ ug]s2 OU anb SIIAJIPIBO :Ope]s9 [9 Р. әри! SEUIEI se] әр ofeqap оләшпи [9 A ДӘ]ӘР.ГР [? B. ори sewel sel 91405 оләшпи IH "99M -994 uoo sopru3]qo sosoruouirsaed seul So[Oqie 8с SO] әр oun әр ose] oj В S319]2EIE2 so] әр uot. »eziuimd() “dv Fa engi J V V1VOMIA !әибәл y E 0 g - z ‘Bly ua епициод VNVIY3NOYVO WnueveUpseb ¿E? ай ын ЛЕ 2 VNVINOLY38 VINWANIT unulxoJd “y v1vinolosv4 sueynbo y L 0 0 0 VLVLON unjejou A — (607 6 . c VLVINOIOSW4 unje[noroseJ “y ЧОДА Boy V1VIVIIO Jenn d &jejejp ‘а VTIAHdOWOVN unj|Kqdo428ul y VLVWINIINVA 02 92 CL VIA1VAIQO3VNI 8۸/eA!nbeeu! y , ET | VSOLIAS3VWI unsojidsoeo - | 9c vC CC VINWSINOWND unuejupenb y LE LE VHLNVIY3 шпуиеиә uinjedseg be i sisdonueuseyju id гоо ооо coRPOUT ZE Bl Л 129 t шпхеј шпдшед 511 Aliscioni Volume 89, Number 4 2002 Filogenia de Paspalum 'uor)enumnuo^) "c 210814 3 q VeOTSIANWd штошмеа y "2 Un 0 v1vonrNoo unjebníuoo ‘а F О, ‘eine eu 2 sndouox s vIS3339 winijAyddjod y 21 г YYOTHIAYYd uinjedsuo y 0 cé xejduns y 001) 0 SIHAHOVNV TOI O uinjjkgdooejeui y m 8 VIYV3IYOVNO uneeud/oo y TS L L? VQIAI шпиеэје y 1 EF Z WOON чедо nini d O ce 0 0 — gy шпигәориеш y 0 VISN3QVAO2HOO [e 8SUOp0A02100 ч F Lt eÁses ae ZG = BE x gg Z VATION N Peete VIS3339 шп]}е/ә}$ ` S = 18/1978 “y AMARA! uneewbÁd y | VSOW3OVMH [3 uinjejioiued y v uinpi|jed VNVIONY IdNOS8 UN wnuelpuejduog y VLOASSIO suede! y jd LLL 0 VLVLON unjund y “(| i VHLNVINS ѕәрошше y VAGIONVHOOVS sepiojeuooes uinjedseg V -z ‘B14 ap uoroenunuoo wv 512 Annals of the Missouri Botanical Garden tintivas Kranz), 30 (parénquima incoloro dispuesto en las alas), 33 (macropelos adaxiales) y 34 (ma- . — cropelos abaxiales) presentan peso = 10 (fit = 1 correspondiendo todos ellos a caracteres libres de homoplasia. Por otro lado, los caracteres exomor- fológicos: #1 (ciclo de la planta), 6 (espiguilla ab- axial), 9 (gluma inferior), 10 (gluma superior), y anatómicos: 420 (cavidades aeríferas) y P (c 'élulas 75). Por último, el carácter anatómico: #18 wipes foliar buliformes) presentan peso — 7.5 (fit — abaxial con costillas y surcos diferenciados) pre- 6.0 (fit tantes caracteres presentan peso = 5.0 (fit = 0.6), mientras que los res- = 0.5). senta peso = DISCUSIÓN Y CONCLUSIONES Sobre la base de los caracteres utilizados en el presente análisis, no pudo establecerse un origen monofilético para el género Paspalum. Siguiendo el método práctico propuesto por Farris (1974, 199] para identificar clados polifiléticos o parafiléticos sobre un árbol determinado, Paspalum se comporta en este análisis como grupo parafilético. Los resultados obtenidos concuerdan en parte con otros estudios llevados a cabo por diversos au- tores, realizados sobre la base de caracteres mor- fológicos y moleculares, coincidiendo todos ellos en interpretar a Paspalum como género no monofilé- tico (Zuloaga et al., 2000; Duvall et al., 2001 2001; Gómez-Martínez et al., inéd.). En un estudio amplio sobre la filogenia de la ; Giu- ssani et al., tribu Paniceae realizado con datos morfológicos, Zuloaga et al. (2000) incluyeron representantes de diferentes grupos de Paspalum, los cuales no for- maron un grupo monofilético por reunirse en una politomía no resuelta junto con géneros afines. Por Gómez-Martínez y Culham (2000) anali- zaron filogenéticamente la tribu Paniceae sobre la otro lado, base de datos moleculares obtenidos a partir de la secuenciación del gen de cloroplasto trnL-trnFk, donde Paspalum es monofilético formando un clado cuyo grupo hermano es Panicum obtusum; sin em- bargo no se consideró a Thrasya en este estudio y a relación con Axonopus no queda resuelta. Pos- teriormente, Gómez-Martínez et al. (inéd.) sugieren a condición parafilética de Paspalum por incluirse en un clado sin resolución interna junto con An- thaenantiopsis y Panicum obtusum. Recientes estudios filogenéticos sobre la subfa- milia Panicoideae, aportan mayor información so- bre la historia evolutiva de Paspalum. Giussani et al. (2001) realizaron un estudio sobre una matriz molecular obtenida a partir de la secuenciación de gen de cloroplasto ndhF. Estos autores incluyeron varias especies de Paspalum concluyendo que el mismo representa un grupo parafilético, con la in- corporación dentro del mismo de Thrasya formando entre ambos un clado monofilético de alto soporte. Asimismo, Duvall et al. (2001) comprueban lo mis- mo a partir de análisis realizados con el gen de cloroplasto rpoC2. A diferencia de los estudios mencionados ante- riormente, Renvoize (1972) sugirió que Paspalum sería polifilético por presentar caracteres exomor- fológicos semejantes en especies de grupos que es- tán sistemáticamente distanciados dentro del gé- nero, aunque no probó su hipótesis realizando estudios de filogenia. Los resultados obtenidos en el presente estudio muestran que solamente los grupos Anachyris, Ber- toniana, Filiformia y Racemosa se resolvieron en clados con relativo soporte indicando que los mis- mos representarían asociaciones monofiléticas de especies. En relación a los datos utilizados, se detectaron caracteres exomórfológicos y anatómicos libres de homoplasia los cuales proveen evidencia para la caracterización de los grupos. Dentro de los carac- teres anatómicos, el desarrollo de las porciones la- terales de la lámina, la presencia de células distin- tivas Kranz, la presencia de parénquima incoloro en las alas de la lámina y la presencia de macro- pelos en ambas epidermis, podrían ser valiosos en futuros estudios sistemáticos del género; algunos de ellos junto con otros caracteres exomorfológicos, definen grupos monofiléticos como el caso del clado que reune a los grupos Bonplandiana y Racemosa, en relación a las células distintivas Kranz. Dentro de los caracteres exomorfológicos utilizados tradi- cionalmente en la caracterización de los grupos de especies, solamente la presencia de nervios cons- picuos en la lemma superior, la forma y coloración del antecio se manifestaron libres de homoplasia. Contrariamente, otros caracteres exomorfológicos como por ejemplo el hábito de la planta, el número de racimos por inflorescencia, el contorno del ra- quis de los racimos, la pilosidad de la espiguilla entre otros, manifestaron un alto grado de homo- plasia lo cual disminuye su valor informativo de- biendo ser utilizados con cierta cautela en la de- limitación taxonómica de los grupos. Según Farris (1991) solamente los grupos mo- nofiléticos tienen una existencia real por poseer una historia evolutiva independiente; por tal motivo los resultados obtenidos no sustentan la existencia de Paspalum como grupo natural. Debido a la am- plitud y a la diversidad observada dentro de este género, el presente estudio muestra una aproxi- mación hacia su historia filogenética y suma una nueva evidencia a que el mismo debe ser reestruc- Volume 89, Number 4 2002 Aliscioni Filogenia de Paspalum turado taxonómicamente, incluyendo a Thrasya y Axonopus, aunque la inclusión de éste ültimo no es consensuada por estudios de otra naturaleza. Literatura Citada Aliscioni, S. S. 1999, Estudio inr Comparado de ro Paspalum L. Panicoideae: Paniceae). Trabajo ав en Fac. Сі. Nat. UNLP: ‚20 Especies Americanas del Gé (Poa- ceae: de Tesis Doctoral 275. . Anatomía ecológica пуат D — ae: 13 8: 1874 & a en especies del género Panicoideae: Paniceae). Darviniii- na rriaga. 1998. Estudio histofoliar compa- rado de lok grupos Virgata y Quadrifaria del género Pas- palum L. (Poaceae: Panicoideae: Paniceae). Candollea 93: 333-348. Arriaga, M. 2000. Austral South American species of Eriochloa. Pp. 141-148 en S. W. L. Jacobs & J. Everett (editores). Gr : Systematics and Evolution. Vol. 2 CSIRO, Collingwood, Australia & 5 Aliscioni. 2000. Estudio biogeográfico a Eriochloa Kunth Aj ais eaea). Gayana Bot. 57 (supl.): : Barreto, 1. L. 1954. Las especies afines a Paspalum vir- gatum en s ‘eres ‘a del Sur. Revista Argent. Agron. 21: 125-14 . 1957, E s especies de Paspalum con dos racimos HE wow en Rio Grande do Sul (Brasil). Revista Ar- gent. 1. 24: 89-11 . Las especies afines a dide di quadrifa- rium — ae) en la América del S re y templado. Darwiniana 14: 13 ——. 1907. As espécies de Paspalum (Gramineae) — a Paspalum corcovadense do Rio G. Sul. — — ultura do Rio Grande do Sul, Bol. -13 — Kk. 1994. — h support and tree stability. Cla- distics 10: 295—304. Brow n, W. V. 1977. The — syndrome and its subtypes ass systematics. Me ‚ Torrey Bot. Club 23: 1—97. йш, 1969. P is 369—411 en A. Burkart (editor), У llustrada En — (Argentina). Colecc 1. Inst. i^ э pee | The genus Tray H. B. K. (Grami- Миё, 4: | Notes on genera wt Paniceae. Biol. Soc. Wah 19: 183— — 1927 . Paspalum. En E der clima sub- ээ. cretaria ~ А P roc. Hitchcock (editor), The grasses of Ecuador, Ган. pee Bolivia. Contr. U.S. Natl. He — 434—455 229. he — Meri 'an species of Paspalum. Contr. U. S. Natl. Herb. 28: 1-310. rédito. парат pl South America. Hitche and C я Library, Bota stitution, Washington, Cialdella, A. ы О. Могод & F. О. Zuloaga. visión de las especies del género Pa: ak \ ly Department, анна E 1995. Re- um (Poaceae: Panic via Paniceae) Grupo Bonplandiana. Darwinia- na 33: 5 . Vega. 1996. Estudios sobre la variación амак d las espiguillas en т = la Tribu Paniceae (Poaceae , и 34: Clayton, W. D. € S. A. Renvoize. on nium: — 's of ilie — Kew Bull., H.M.S.O., London Crins, W. i р The genera of Paniceae (Gramineae: enera Grami- Mns Ser. XIII. Pa- nicoideae) i in n ои United States. J. Arnold Supp pl. Ser. 1: 171-312 Davic >. Ў — Japan in the Poaceae. Pp. 143—155 en T. R. Soderstrom, К. EN Hilu, C. 5. Camp- bell & worth (editores). Grass — matics — Evolution, Smithsonian N Ны Press, Washing- Dengler. N. G., R. E. Dengler & P. W. Hatte rsley. 1985. Differing ontogenetic origins of PCR (“Kranz”) she — in leaf blade >s of C, grasses (Poaceae). Amer. J. Bot. 204—302. Duvall, M. R., J. D. Noll & A. H. Minn. 2001. Phyloge- netics oí Paniceae (Poaceae). Amer. J. Bot. 88: 1988- 1992. Ellis, R. 1976. A procedure for standardizi leaf anatomy in the Poaceae, I. The leaf blade as viewed in — section. Bothalia 12: бй 9. — —— 1979. A procedure for stan leaf antomy in the Poaceae, Il. he — rmis as seen in surface Bothalia 12: 641-671. Farris, J. S. 1974. dics он of paraphyly and و‎ — Zool. 23: 548—554. . The — * of pityingenal tic analyses. Pp. 2 en EN Platnick & V. Funk (editores), Advances in C а 's, Vol. 2. Columbia Univ. Ne w York . Hennig defined paraphyly. Cladistics 7: 1t1 у iew. _ =) Press, — ——., V. A. Albert, M. Kallersjé, D. Lipscomb & A. ( Kluge. 1996. а. P SYM outperforms — bor- joining, ^ ladistics 12: 99-124, Filgueiras, 1993. аео and critical notes um (Po ceae: Pa- on some Brazilian species of Paspalu niceae). Acta Amazon. 23: 147 E v Giussani, L., J. Н. Cota-Sánchez, ). Zuloaga & E. A. ellogg. 2001 A molecular — of the grass sub- es Panicoideae cr aceae) E — origins of C, photosynthesis. Amer. J. Bot 1993. Estimating — ler we ira dorini Cladistics 9: 83—91. )5. — and we дЕп: А reply to Turn- er ба fani e. Cladistics 11: 91-104. Та. Nona ver. id New Yor 1997b. Pee-Wee ver. 2: 8. Computer program and documentation, New York. Gomez-Martinez, R. & А. Culham. 2000. Phylogeny of the d — with boa on the tribe Pa- ce from the trnL-F cpl L. Jacol E Goloboff, P. A tree searc Уе 1.8. — program and uceae: E iden 136-1 39 . en S. W. obs & J. beet — Gras Syste matics 2 Evolution, Vol. 2. CSIRO Collingwood, с Hattersley “Matson, 1992. Diversification of pho- 1 te is p 38-116 en G. P. € артап (editor). Grass Evolution and Domestication. Cambridge Univ. Press, London. Hite lico rock, A. S. 1951. Manual of the Grasses of the Unit- ed States. Se er iss “ón. U.S.D.A. en Publ. pos Туан n, P. K., N. rnett. 1990. Holmgren & I Index He — Part 1 I. The urbi x of the World. Regnum Veg. 120. Judziewiez, E. J. 1990. Flora of the Guianas. Series A Phanerogams. Fasc. 8. Koeltz Scientific Books, Kónigs- tein, Kellogg. E. A. 2000. Molecular ~ Кз ergo ws evo- — in the Andropogoneae. Pp W. L. Jacobs & J. Everett (editors). Gras — 8 wd Evolution, Vol. 2. CSIRO, Collingwood. Australia 514 Annals of the Missouri Botanical Garden Maddison, D. R. 1991. multiple islands of most-parsimonious trees. Syst. 40: 315— 328. е O., T. Filgueiras, Е. Zuloaga & J. Dubcovsky. 1993. Revision of Anthaenantiopsis (Poaceae: Panicoi- deae: 53. a eae). Syst. Bo ‚ F. Zuloaga & E. — 1995. Revisién del REN Racemosa del género Paspalum (Poaceae: Pani- coideae: Paniceae). Ann. Missouri Bot. Gard. 82: 82— 116 za & F. Zuloaga. 1996. Revisión de las énero Paspalum 1. (Poaceae: Panicoi- deae: — grupo Dissecta (s. str.). Candollea 51: 103- espec ie 5 ^ 1 E De nham, S. Aliscioni & F. Zuloaga. 2000. Re- даба de las especies de Paspalum (Paniceae), subgé- nero ; — Candollea 55: 105-13. 5. Nicora, E. & Z. Rúgolo de Agrasar. 1987. Los géneros de ташы as de América Austral. Не misferio Sur, Buenos Nixon. K. 1999. Winclada d ta) ver. 0.9.9. Published by the author, ae a, New Yor —— ё arpenter, 1993. On outgroups. Cladistics 9: 413—426. Parodi, L. Serie. H. 1928. Gramíneas ся P n Argentina. П di ig (Buenos Aires) 9 . 1932. Algunas gramíneas nuevas o críticas. Phy- sis (Buenos Aires) 9: 12-24. . 1937. Contribución al estudio de las Gramíneas del género Paspalum de la Flora Uruguaya. Revista us. La Plata, Secc. Bot. 1: 21 1-29 ). Pohl, К. 1980. Gramineae. En W. Burger (editor), Flora Costaricensis. Fieldiana, Bot. 4: 350—392. Quarín, С. 1975. Notas E el género Paspalum (Gra- mineae). eps 3: 10. Ramírez, M . How жа should a weighting func- tion be? Pp 2 en Hennig XVII. 17th Meeting of the Willi Hennig Society. Sáo Paulo, Brasil. T S. A. 1972. Studies in the Gramineae: кут n 451-45 55 The — of Bahia. ХХХ. Royal Botanical байам. p» 1987a. New grasses from Paraná, Brazil. Kew Bull. a E -925. 37b. A survey of — As anatomy in grasses XI. Pa niceae. Kew Bull. < 9-708. ш z, Н. R. 1998. El — ro Ceresia (Pers.) Rei- del género — L. (Gramineae) en Vene- zue Kj Emstia 8: 7-5 , E. amac o de & L. Guevara. 20( а sección Pani iflora Rodriguez del género Paspalum a (G pe en Venezuela. Ernstia 10: —143. — ngurtt, R. Arrillaga de Maffei & Ehe un re de Artucio. pon Gramíneas Uruguayas. e rsidad de la Re E a, Montevidec Rua, G. H Weberling. 1995. Growth form and inflo- rescence structure of Paspalum lL. (Poaceae, Paniceae): Torres =) ‘omparative morphological approach. Beitr. Biol. Pune n 69: 363—431. Sánchez, Arriaga. 1989. El síndrome Kranz en Arundincllae (Panicoideae-Poaceae) de la flora Suda- mericana, Parodiana 5: 2 Sendulsky, T. & A. G. Burman. 1978. Paspalum species of the Serra do Cipó (1): A epa to the study of the Brazilian Poaceae. Rev. Brasil. Bot. 1: 1-15. . 19804. Paspalum species a the Serra do Cipó 61-277. The discovery and importance ч Zo — ID: A не to the — of the Brazilian Poa- ceae. . Brasil. p 23- l 980b. A new species r Paspalum (Gramineae) from Brazil. А 32: 487-489. Silva, T. S. da G. Burman & T. Sendulsky. 1979. Es- pecies de Paspalum н da wy do Cardoso, Estado de Sâo — he ie 11-2€ Smith, L. C. i ali & R. M. Klein. 1982. — . 1, Paspalum—Zea, Pp. 910-1407 en ‚ Flora Ilustrada Catarinense. Itajaf, Stebbins, G. L. 1982. Major trends of evolution in the Poaceae and their possible significance. Pp. 3-36 en J. R. Estes, R. J. Tyrl & J. N. Brunken (editores), Grasses and Grasslands: a matics and Ecology. Univ. Oklahoma Press, Norma Türpe, A. M. 1967 [1966]. Histoasonomia o a especies argentinas del км ro Paspalum. Lillo 72. 19€ Vegetti, A. C. . Análisis tipológico d la ilr 'scen- cia en — (P oaceae). Kurtziana 19 —160. Watson, L. & М. Dallwitz. 1982. — Grass Ge nera: Anatomy, Miranda and Key Australian. Na- tional Unive rsity, Research Sc hool of Biological Scienc- anberra. . 1989. Grass Genera of the World: Il- vesc of € — ‘ters, Descriptions, Classification, Interactive Identification, Information Retrieval. The Australian National — rsity, Research School of Bi- ological Sciences, Ca Webster, R. 1988. ae “of He North American Paniceae (Poaceae: Panicoideae). Sy: 6-609 Zuloaga, F. O., R. Ellis € revision of Panicum subgenus Phanopyrum section Laxa (Poa- ceae: Panicoideae: Paniceae). Ann. Missouri Bot. Gard. 79: 770-818. O. Morrone & L. Giussani. 2000. A cladistic analysis of the Paniceae: A p — th. Pp. —135 en : W. L. Jacobs & J. arcc Grasses! Systematics and Pur "Vol. 2 CS Collingwood, peers ÍNDICE DE NOMBRES CIENTÍFICOS GRUPO INTERNO — Paspalum alcalinum Mez (LIVIDA) Р. ammodes Trin. (ERIANTHA) P. bertonii Hack. (BERTONIANA P. bonplandianum Humb. & Bonpl. ex Flüggé (BON- PLANDIANA) P. caespitosum Flüggé (C o ү" P. conjugatum Berg. (CONJU P. corcovadense Raddi (CORC — P. coryphaeum Trin, (QUADRIFARIA) P. crispatum Hack. (PARVIFLORA) P. dilatatum Poir. (DILATATA) distichum L. (D om P. equitans Mez (FASCICUI P. fasciculatum Willd. ex Е (FASCICULATA) P. filiforme Sw. (FILIFORMIA Р. gardnerianum Nees (GARDNERIANA) P inaequivalee Raddi (INAEQU IVALVIA) j 17084) P. lindenianum A. Rich. (FILIFORMIA) : lineare Trin. (LINEARIA) Volume 89, Number 4 2002 Aliscioni 515 Filogenia de Paspalum P. lividum Trin. ex Schldl. (LIVIDA) P. macrophyllum Kunth м P. malacophyllum Trin. (ANA S P. mandiocanum Trin. (CORC on ADENSIA) P. modestum Mez (PLICATULA) P. notatum Flüggé (NOTATA) P. orbiculatum Poir. (ORBICULATA) P. pallidum Kunth (BONPLANDIANA) P. paniculatum L. (PANICULATA) P. parviflorum Rhodé ex Ms — IFLORA) F. penicillatum Hook. f. (R SA) P. plicatulum Michx. (PL IC a М F. polyphyllum Nees ex Trin. (CERESIA) P. pygmaeum Hack. (R AC ЕМ P. quadrifarium Lam. ZR Su ae P. regnellii Mez (VIRGATA) Р. repens Berg. (DISSECTA) P. saccharoides Nees (SACC pia P. simplex Morong ex Britton (ANACHY P. stellatum Humb. & — ex Fliiggé * RESIA) P. urvillei Steud. (DILATATA) P. vaginatum Sw. (DIST Tera P. virgatum L. (VIRGATA) GRUPO EXTERNO Anthaenantiopsis rojasiana L. Parodi Axonopus compressus (Sw.) eauv. Eriochloa distac E Kunth Panicum laxum Thrasya — Kunth APENDICE 1. MATERIALES EXAMINADOS. ESPECIES DEL GRUPO INTERNO Paspalum subg. Anachyris Chase ۴ — ои Trin. NTIN atamarca: Altivalle de las Gran- р p rentz en (US). PARAGUAY. Central: Asunción, Rojas 6732 (US). 2. Paspalum simplex. — ex Britton PARAGUAY. Central: Asunción, Hartley & Rojas 54185 (US): Trinidad. Pate: & Rojas 7895 (US). a 10811 (US). Cordillera: Piribebuy, Rojas 13445 (US Chaco: Bahía Negra, Rojas 13813 (US). Neembucú щй: Estancia "Yacaré Pilar,” Rosengurtt B-5527 (US): Río Pilcomayo, Morong 1583 (US). Grupo Bertoniana (sensu Chase, inéd. ) 1. Paspalum bertonii Hack. (GENTINA. Misiones: Candelaria, en isla del Río Paraná, — 988 (51): Iguazú, Salto Uruguay. Hunziker 9941 (SI); —€— Porta 55 (SI); San Ig- nacio, Cabrera et al. 28787 . Boelcke et al. 5393 (SI); Jardín América, Quarín D a BRASIL. Pa- гапа: Guairá, Reitz & Klein 12139 (US 2. к lilloi Hac ASIL, Prodi Salto Iguassá, Rambo 53619 Grupo Bonplandiana (sensu Cialdella et al., 1995) l. Paspalum bonplandianum Humb. & Bonpl. ex Flüggé ER P |. Cajamarca: Cajamarca, Smith & Sánchez Vega 7508 (MO, SI). Lambayeque: Ferreñafe, Sagás- tegui et al. 12831 (HUT, SI). ECU = Carchi: Tul- cán, Werff & Gudifio 11042 (MO, SI). 2. — к Kunth CUAD mbabura: S hi, Peñafiel & Ta- mayo 892 MEL QCNE. SI). I Cajamarca: Ca- — а, Se p Cabanillas — & Guevara 643 (CPUN, SI); Cayayup, Ochoa 1580 сүл oe Ferre 2s Llatas uiis 1920 (HUT, SI). Grupo Caespitosa (sensu Chase, 1929) 1. Paspalum caespitosum Flüggé BELICE. El Cayo: — Ridge, Lundell 6426 (SI, US). MEXICO. Las Ruinas de Coba, Tellez & Ca- brera 2353a (MEXU, SI). 2. Paspalum — 'orum Mez ARGENTINA. Corrientes: Santo Tomé, Krapovic- kas et al. domi I). Misiones: Cainguás, Zuloaga & Deginani 506 (SI); Candelaria, номда 's 1357 (SD): Santa Ana. Montes 2334 (SI); Villa Venecia. Renvoize 3024 (М). PARAGUAY. Guairá: Iturbe. Montes 12739 (CTES). — 'aspalum subg. Ceresia (Pers.) Rchb. 1. сү polyphyllum Nees ex Trin OLIVIA. Santa Cruz: Sara, — 5385 (US). a Mato Grosso: Tres Chase 10740 Gerais: Serra de Cipó, Archer et al. 4995 (US). Parana: Ponta Grossa, Swallen 8772 (US). Rio anta Ca- agoas, Rojas 10209 (US). Hassler 11561 (US). ^ Papam stellatum Humb. & Bonpl. ex Flügg tGENTINA. Misiones: Posadas. — 594 — COLOMBIA. Magdalena: Santa Marta. Smith (US). PARAGUAY. Amambay: Hassler 11058 E A Caaguazú: Estancia “Plate San José.” Rosen- gurtt B 5883 (US). Central: Asunción, Rojas 11005 (US): Tarumandy, Schinini 6207 (US). Cordillera: Cordillera de Altos, Fiebrig 664 (US). = Central: Ypacaraí. Lago Ypacaraí. Grupo Conjugata (sensu Chase, 1929) l. Paspalum conjugatum Berg. RGENTINA. Jujuy: ong Grande, Taylor et al. 11377 (MO, SI). Misiones: Apóstoles, Cabrera & Saénz 29087 (S1): ‘Candelaria Montes 2455 (SD; El Do- ral. Belgrano, Hunziker et 6 (SD: — Mont 6950 (SI); Posadas, Bur- (SD: 5 vier, Cabrera et al. 28659 (Sl). BOLIVIA. Santa оа Sara, Steinbach 1910 (SI): Warnes. Zabala 627 (SI) Grupo Corcovadensia (sensu Chase, 1929) 1. Paspalum corcovadense Raddi BRASIL. Rio de Janeiro: Leblón, ни 3377 (US). Rio Grande do Sul: Arth 1/02 (US). Santa Catarina: ltajaí. Smith & Reitz 6076 (US). Sao Pau- lo: Capital Chávaca dos Morrinhos, Pickel 5803 (US). 2. Paspalum mandiocanum Trin. ASIL. Rio de Janeiro: Dist. Froes 51-11331 (US). PARAGU AY. AI araná: Íra- а, Montes 11029 (US). Central: A dde "n, Rojas l. 3346 6 (US): Lago Ypacaraf, Hassler 12481 (US). Guai- : Villarrica, Rosengurtt B-5929 (US). Itapúa: En- — Black & Annals of the Missouri Botanical Garden carnación, Rojas 11191 (US). Paraguarí: Caapucú, Anderson 1151 (US); s.l., Rojas 1334611 (US Grupo Dilatata (sensu Chase, 1929) 1. Paspalum уц Poir. ARGENTINA. Buenos Aires: Bañados entre San Fernando y Pac — Correa s.n (SD; La Plata, Cabrera 3420 (SD). San Luis: Pedernera, Rosa 48 (SI). Santa Fe: — — 31 (SD. BRASIL. Rio Grande s. Barreto 1601 (SI); São Gabriel, Barreto (SI); Sao Simao, del Mato 12382 (SI). Santa Ca- tarina: Agua Doce, Smith & Klein 13529 (SI). CUBA. Pinar del Río: near Mariel, Ekman s.n. (US). URU- GUAY. Montevideo: Atahualpa, Herter et al. 334 (SI). 2, de pres urvillei Steud. BRASIL. Mato Grosso: vicinity of Dourados, Chase 10951 US. Paraná: Coloniba, Dombrowski 4391 ^ a — (US). Santa Catarina: ltajaí, Klein 2749 (US). Sao Paulo: Alto da Serra, Lemos 31504 (US). COLOMBIA. ти Urrao, Pohl & Betancur 15504 (US). PAR- GUAY. Guaira: Villarrica, Jorgensen 3535 (US), Jor- gensen 3635 (US). е Dissecta (sensu Morrone et al., 1996) . Paspalum л ns Berg GENTINA. Huson Aires: Delta del Paraná, Burkart rhe (SD, Burkart 4343 — Chaco: Rio de Oro, Holmbe rg s.n. (SI). Entre Ríos: La Paz, Pedersen ; Puerto La Paz, Duos et al. 26830 (SI). Santa Fe: Garay, León & Fossati 922 (S1); San Jeró- nimo, Morello 16873 (S1): Islas del Río Paraná. More llo з (SI). PERU. Loreto: Maynas, Gentry . VENEZL UE es A. Guárico: Camaguán. Gin el s 4353 (SI, VEN). 33 Grupo Disticha (sensu Chase, 1929) I. Paspalum distichum L. BOLIVIA. Coe — Sl, tchcock thichas, Hite ei 22882 (US 0, ADOR. vie hine hat Н, | betw. Terrier Rouge and Fort Liberté, Bartlett 17 183 (US). P iir AGUAY. Central: in regione r 12452 (US) lacus Ypacaraí, Hass Paraguay, Pavetti & Rojas 9448 (US); in regione cursus inferioris fluminis Pilcomayo, Mani 322 (US). PERÜ. Piura: Hacienda Buenos Ап PUE RTO RICO. — e: s.l.. 2. ii ever vaginatum Sw. BRA = Bz hia: — i: Marapanim, Davidse et al. 17785 B (US) Santa bs Smith Anderson 668 (US). Chase 6483 (US). alenat- Montabo, Hoff 5168 (C AY, Р, US). ragı bertad: Trujillo, Angulo T 0279 (US Chorillos, Asplund 10906 (US). Grupo Gardneriana (sensu Chase, 1929) 1 ш Mon rianum Nee BRASIL. Goié ig ei Glaziou 22583 ( Glaziou z 586 (US). Maranhão: Grajahú, Swallen 3652 (US). Piauí: s.l., 22793 "Hite hcock 19908 (US). EC- - Asplund 6571 (US). Ji playas del Río Belém & Peirheiro 2985 179 агі: Caacupú, Anderson 1125 (U S). Р E RÚ. La Li- S). Lima: Р); s.l., Barra do Corda to Gardner 3507 (P). GUAYANA. Ebini Experimental Station, Harrison 1786 (US). COLOMBIA. Arauca: 10 km SW of Cravo Norte, Blydnstein 1522 (US). Grupo Eriantha (sensu Chase, inéd.) 1. Paspalum ammodes Trin. BRASIL. Goiás: Serra do Caiapó, Irwin & Soder. trom 7040 (SI). Paraná: Jaguarihyva, Dusén 16025 (G). São Paulo: San Jose dos Campos, Morello s.n. ( 2. Paspalum — Nees ex Tri BRASIL. Goiás: s.l., Glaziou "22491 (С). Glaziou 22489 (G), Glaziou 22. 589 (G). Paraná: Turma 22, Du- sén 15669 (G). Rio de Janeiro: Ы (С). PARAGUAY. Amambay: in re gione — cur- sus superioris fluminis а Hassler 11924 (6); Sierra de Amambay, Lomadas Esperanza, — lores (G). Grupo Fasciculata (sensu Chase, 1929) l. Paspalum equitans Mez . Corrientes: Santo Tomé, Quarín et al. 2753 (CT ES, SI, US). PARAGUAY. Paraguarí: Este de la Cordillera de Villa Rica-Marais, Balansa 87 Р). 2. Paspalum. jesse а Willd. ex Fliiggé 3RASIL. Pará: Boa Vista, Swallen 31651 (US). CO- Pennell 3933 (NY, US); S). Magdalena: Cos- US). PARAGUAY. LOMBIA. Bolívar: Magangue, Island of Mompos, Curran 366 (U! ta del Caribe, Dugand 5259 t OL, Central: s.l., Morong 535 (U Grupo Filiformia (sensu Chase, 1929) 1. Paspalum. — Sw. CL r del Río: — 1. Ekman 946 (SI, US). 2. парат аа А. Кісі BA. Pinar del Rio: Mariel. Ekman 947 (SI, US). Grupo Inaequivalvia (sensu Chase, inéd.) и — e Raddi BOLIVIA. La Paz: Milluguaya, Caripata, Hitchcock 22672 (US). B RASIL . Mato Grosso do Sul: Corumbá on Rio Paraguay. Chase 11133 (US). PARAGUAY. Caacu s.l.. Archer et al. 4799 (US). Central: in regione lacus Ypacaray, Hassler 12401 (US): s.l., Mo- ron ). Cordillera Caacupe Ramírez 406 (US): in regione cursus inferioris — Pilcomayo. Rojas 96 (US); s.l., Balansa 106 ( Grupo Linearia (sensu Chase, 1929) 1. Paspalum — Trin. BRASIL. Ceara: s.l.. — — 22473 5 (G). Minas Gerais: s.l., M, Piauí: s.l.. Gardner 2975 (G, P). Sao — 351 (P); s.l., Riedel 949 (G); s.l., Gardner 2979 (P). Goiás: Ca- Wed- 2. Paspalum @ Mez VASIL. Santa Catarina: — Luz, Smith & Klein 13319 (P) Agua Doce, Smith & Klein 13653 (US); Fachinal dos Guedes, cami e Klein 14014 (US). Paraná: dores Jonsson 307a (P). Sao Paulo: e — 28535 (US); Estacás — bacs Pickel 5520 (US); ж Campo Alegre, Black 51- 10994 M PARAGUAY. Guairá: Villa P. а, "Balan- sa 69 (G). Volume 89, Number 4 2002 Aliscioni 517 Filogenia de Paspalum Grupo Livida (sensu Chase, 1929) l. Paspalum alcalinum Mez COLOMBIA. Cundinamar La Esperanza, García Barriga 10009 (US). PARAG UAY. Alto Par- aguay: Puerto Casado, Rojas aw (US); Cerro Galván, Rojas ds (US). Chaco: . Casado, Rosengurtt B- 5663 (l s.l., oe В- — (US); Estancia Gus- tafson, Sa 190 (US). 2. Paspalum lividum Trin, ex Schldl. {GENTINA. Entre Ríos: Paraná, аа А 661 (Sl); San José, Colón, Zucol 655 (SI). Santa ) de Julio, Pensiero & Vegetti 2740 (SF); San Justo. poen 2736 (SF). — c Grupo Macrophylla (sensu Chase, inéd.) . Paspalum mac pne Kunth COLOMBIA oquia: Fredania, Archer 505 (US); Coc arina, Lindi 14 (P). Cauca: La Manuelita, Pittier 897 (US). Grupo Notata (sensu Chase, 1929) l. Paspalum notatum Fliiggé )MBIA. Antioquia: Medellín, Archer 785 (US). Cauca: La Manuelita, Pittier 845 (US). Cundi- namarca: Fusagasuga, Juzepczuk 5317 (US). El Valle: Hacienda “El Trejo,” Hernando García 6412 (US). — — iras, Fosbe rg 19240 (US); s.l., Dawe 854 (US) ‚емга: Jardín Botánico, Archer 1721 a ES Bernardino Rojas 1006 (US). Ñeem- bucú: Estancia del Sr. Medina Pilar, Ramirez 131 (US). TT betw ios buy and Paraguarí, Jor- gensen 3537 (US): pa 110 (US) 2, т сл 5 ASIL. Rio de Janeiro: s.l., Sellow s.n. (P); Co- С hana leen 4343 (P); Jardín — Alta Boa Cha: 3 (US). Santa Catarina: Blumenau, bs 10. 366 em Florianopolis, Canasvieiras, Klein al. 6478 (US). ноя Orbiculata (sensu Chase, 1929) . Paspalum orbiculatum Voir. BOLIVIA. Pando: Madre de Dios, Nee 31353 (US). COLOMBIA. Amazonas: Trapecio amazónico, Schul- tes & Black 8508 (US); Boca Loreto—Yaco, Schultes & Black 46-171 (US). Antioquia: alrededores del Rio сс Gutiérrez & Barkley 17C178 (US). El Valle: Río alema. Cuatrecasas 21118 (US). PAR. A- GUAY. € al: Carapeg guá, Rojas 3350 (US); Arroyo Tobaty, yee "4839 (US). Grupo Paniculata (sensu Chase, 1929) 1. Paspalum juergensii Hack. BOLIVIA. La Paz: Nor Yungas, Beck 9152 (US): Sur PN Hitchcock 22627 (US). BRASIL. Serra do Mar, Dusén 3624 (SI). Rio Grande do Sul: I т 41248 (US): Inst. Agr. Sul-Pelotas- R. - Sul. | € dp Botánico), da Costa Sacco 119 dei COLC utumayo: Paianza el Fabaño, Gamta bise MS. ECUADOR. Tungurahua: Valley of Pastaza River, Hitchcock 21842 (US). Hitchcock 21749 (US) Parana: 7 2. Paspalum paniculatum BRASIL. Bahia: as Pedane ‘Go, Chase 7868 (US). COLOMBIA. Boya Valley of Rio Pomera, (US 2200414). Cauca: "n Tambo, Godfrey s.n. (US). Cun- dinamarca: Estación Santana, Dugand & Jaramillo 3891 (US). El Valle: Cordillera Occidental, Cuatre- casas 19685 (US). ECUADOR. Imbabura: Lita, Acos- ta-Solts 12567 (Us 5). PARAG UAY. Amambay: tiplanitie et declivibus “Sierra de Amambay,” Rojas 10134 (US), Rojas 10738 (US). Caazapá: Abai region San Juan Nepomuceno, Rojas 5835 (US). TRINIDAD TOBAGO. Port of Spain: s.l., Hitchcock 10036 (US). in al- Grupo Parviflora (sensu Chase, 1929) l. Paspalum crispatum Hac Е BRASIL. Mato Grosso: Santa Rita do Araguaya, Chase 11844 (US), Chase 12051 (US). PARA( GUAY. Amambay: in altiplanitie Sierra de Amambay, Hassler 11975 (US). 2. Paspalum parviflorum d ex Pligg BRASIL. Amapá: Rio Pedreira, Pires & Cavalcante 52142 (US); Macapá, м ۳ & Fróes 51- 12300 (US); betw. Rios Cuyubin and Flechal, Pires & — 5237 ) (US). Mato Grosso: entre Bards de Pimienta Buena, Kuhlmann 1682 (US zal. — 6489 b Pará: Missis Island. Goeldi 95 (Us ; Pombal, ná do Tapará, Black 52-15565 Geom 5116 (US). GUYANA FRANCE- л Че тогу Leándre, Hoock 1136 (US); — Le Prieur 5 (US). LS Grupo Plicatula (sensu Chase, 1929) l. deco modestum Mez ¿NTINA. Corrientes Capital, — 608 men : SD: Itatí, Arrocera Rzepecki, Ahum 920 (CTES, SD; Solari, Curup Cuatiá, Vander shail 1393 (SI). Entre Rios: La Paz, Burkart & Bacigalupo ; Puerto Soto, Burkart & Bacigalupo 21093 (SI). PARAG UAY. Central: orillas de lago Ypacaraí, Hicken 111 (SI); San Bernardino, Rojas 1111 (US). 2. Paspalum go wee Michx BRASIL. Paraná: — Dusén 13491 (G). Rio Grande do Sul: Bagé, Estancia Taruma, rem to 1600 (SI); Faz. Experimental de Criagao Ci Cruzes, Ba- rreto 1377 (S1); Potrero № 24, Tora 1475 (SD; 12 km de Vacaría, — 1720 (CTES, SD; Passo Fundo, Quarín 1718 (CTES, SI). Santa Catarina: Agua Doce, Smith & Klein es (SI). T Grupo Quadrifaria (sensu Barreto, 1966) l. Paspalum coryphaeum Trin. BRASIL. Bahia: subida al pico das Almas, Zuloaga et al. 4837 (SI); Br-242, Zuloaga et al. 4774 (SI). Mi- nas Gerais: Morro da Gloria, Glaziou 16684 (P). Rio de 00 iro: Nova Friburgo, "ue; 13328 (P); Pào de sucar, Chase 8393 Ш (US). 2. — quadrifarium Lam. RGENTINA. — entes: Curuzü- — d ü, Ro- 1644 (SI). URUGUA tigas: Arroyo Gu sengurtt 11328 dem Тош Río Yí, a 641 (US). Grupo Racemosa (sensu Morrone et al., 1995) l. Paspalum penicillatum Hook. f. BOLIVIA. La Paz: Nor Yungas, Beck 9158 (SI, A): 0.7 km W of Chuspipata, Solomon 13692 (MO, SI). PERU. La Libertad: Otuzco, Sagdstegui et al. 11569 (HUT, SI). Lima: Chancay, Trovar 375 (SI, 518 Annals of the Missouri Botanical Garden USM); Lima, Lomas de Atocongo, Cerrate 832 (SI. loaga et al. 3319 (SI). Corrientes: 42 km E de Itu- USM). zaingó, — et uL 2298 (SI). 2. Paspalum pygmaeum Hack. ARGENTINA, Jujuy: Santa = Burkart & Tron- coso 11670 (SD; s.l., Parodi 16452 (SI). Tucumán: Taff, Burkart 5280 (SI). BO LIVIA. "bdo ni — ollo, Beck et al. 18079 (LPB, SI). La Paz: In- , Villavicencio 288 (LPB, SI); Murillo, 17 kn al este de La Cumbre por el camino Unduavi, Solomon 18263 (LPB, MO, SI); Omasuyos, Tiquina, Beck 2923 (SI). PERU. Aneash: Bolognesi, joc 1375 ( USM). EUM Huancavelica, Patacancha, To- var "du 55 (SI, USM). Junín: Huancayo, Tovar 2164 (US US — м N Grupo Saccharoidea (sensu Chase, inéd.) l. — fe ee чаны Nees 3 a Paz: Quebrada de Zongo, к a — Avila 1831 (S1). COLOMBIA. Meta: 10 de la carretera Villavicencio a Bogotá, Zuloaga 3955 (SI); Quindio, Monte negro, Zuloaga & Londoño 4186 (SD. PANAMÁ. € facer savannas near El Valle, Duke & к. 6612 (51, US). Chiriqui: Fotuna Dam area, Churchill & mete 6141 FONS SI). VENEZUE — entre Piedra = > Federal: « entre Caracas y | a — Burkart 16994 a a (SI sl). Grupo Virgata (sensu Chase, 1929) 1. Paspalum — Mez ARGENTIN . Misiones: Iguazú, Guaglianone et al. 995 (SI). BR ASIL. Minas Gerais: Viçosa, Chase 942 SI). Rio de Janeiro: Petrópolis, Glaziou 9049 (P). PARAG UAY. Central: Asunción, Rojas 5800 (SI). 2. eco virgatum L. GUYAN/ ree — Lamaha € — Hitchcock 1655 52 — JAMA . Savana entre Ew ton y Linstead, Hitchcock өз 1 TRINIDAD Y ro. BAGO. Te — ‘Prune hfield, Eggers 5553 (Р). XICO. Tabasco: Tenosique, Matuda 3519 MER — = ESPECIES DEL GRUPO EXTERNO Anthaenantiopsis rojasiana L. Pare ARG ENTINA ;orrientes: Trans Tomé, Burkart 19628 (BAA, SD); Estancia Garruchos, Cabrera 11871 Misiones: Posadas, Parodi 4552 (BAA). Axonopus ам (Sw.) P. Beauv. ARGENTINA. Santa Fe: Arroyo Natulipiahué, Cas- tellanos 19443 (BA). BOLIVIA. Santa Cruz: Sara, Steinbach 1865 (BA), Steinbach 6 PARA- ;UAY. Alto Paraná: Puerto Bertoni, Bertoni 11313 (BA). URUGUAY. e Pau 11212 (BA). Thrasya paspaloides K BRASIL. Distrito Federal: Brasilia, Zuloaga s.n. (SI). Erioc bn distac hya Kunth OL. anta Cruz: Velasco, & Carrión Rua calcarea c ursus supe rioris fluminis Apa, Hasslerl 103. (BAA, BAF, G) Panicum — Sw. ARGENTINA. Buenos Aires: Isla Martín García, — 46626 (BAA). Chaco: Puerto Antequera, Zu- APÉNDICE 2. LISTA DE CARACTERES. CARACTERES 1. Ciclo de vida de la ‘sient (O) perenne; (1) anual. La mayoría de las especies del género son perennes; a s. Ejemplo de g son las espe- cies del grupo Rac 'emosa D rone et al., 1995), como asf también — especies de los grupos Parrilla y Pli- catula (Chase, 1929). Según ; Rua y Weberling (1995) las especies — ocupan habitats específicos como altas montañas, márgenes de bosques montañosos o sabanas con una marcada estación seca. Por otro lado, Morrone et al. (1995) interpretan al cielo de vida anual como un ca- rácter evolucionado dentro del género, basándose en los criterios evolutivos propuestos por MM (1982), Wat- son y Dallwitz (1982) y Davidse (1987). 2. Habito de la planta: (0) cespitosa; (1) no cespitosa. Rua y Weberling (1995) desarrollaron distintos modelos para Hus ribir las formas de crecimiento presentes en Pas- palum; considerando que el hábito cespitoso es el más miis en el género. Por otro lado, las eden presentar tallos decumbentes, rastreros, especies no ces- pitosas pu largamente rizomatosos o estoloníferos; pertenecen a gru- pos restringidos dentro del género los cuales siete qe bajo condiciones ambientales particulares. 3. Número — racimos en las — encias: (0) 1 racimo; (1) 2 racimos; (2) más de 2 racimo La inflorescencia de Paspalum ha sido descripta tra- dic و‎ 'nie como una panoja constituida por racimos espiciformes insertos a lo largo de un eje. Vegetti (1987) describe a la inflorescencia de Paspalum como de tipo politélico debido a que los racimos o paro ‘ladios largos te rminan en una espiguilla estéril o en una porción estéril del raquis. Segtin Vegetti (1987) y Rua y Weberling (1995) as difer ‘rencias estructurales de las inflorescencias dentro in desarrollo variable de la zona de Paspalum se deben a ento de la zona internodal del eje principal ‹ Basado en estos conceptos, Vegetti (1987) considera que las inflorescencias con 1 ó 2 racimos serían más evolucionadas dentro del género 4. — del raquis de los racimos de las inflorescen- s: (0) subtrigono; (1) expandido. El raquis de los racimos prese nta un transcorte subtrí- gone en la mayorfa de las los racimos se presenta anchamente — con pro- yecciones laterales hialinas o foliosas. Esto se observa en as especies de los grupos Ceresia, o Racemosa y Bonplandiana, como así también en algunas especies de los grupos Lachnea, Gardneriana y Plicatula. Thrasya bi т comparte con las « espec les anteriormente citadas la presencia de inflorescencias con raquis expandido. Espiguilla adaxial: (0) ausente; (1) presente. аз espiguillas que se dispone n solitarias sobre el r quis se definen con posición adaxial cuando el dorso de Volume 89, Number 4 2002 Aliscioni 519 Filogenia de Paspalum la gluma inferior se orienta contra el raquis del racimo. ribu la Zuloaga el al. (2000) c onaiderati que la espiguilla adaxial es homóloga a la espiguilla de posición terminal que co- rresponde a una ramificación secundaria del raquis del racimo Espiguilla abaxial: (0) ausente; (1) presente. Las espiguillas que se disponen solitarias sobre el ra- quis se definen con posición abaxial cuando el dorso de la gluma inferior se — opuesto al raquis del racimo. Segün Zuloaga et al. en las especies de la tribu Paniceae que presentan perde pareadas, la Ed abaxial es homóloga a la espiguilla de posic ión axial, cual se genera a partir de una ramificación terciaria de 1 raquis del racimo. 7. Número de hileras de espiguillas por racimo: (0) dos: (1) cuatro Los racimos presentan dos hileras de espiguillas cuan- do las espiguillas son solitarias y cortamente pediceladas. En este caso, es simple determinar la orientación de la misma (Clayton & Renvoize, 1986), siendo adaxial en Axonopus y abaxial en las especies de Paspalum con es- piguillas solitarias. Sin embargo, dentro del género Pas- palum algunas especies poseen espiguillas — ob- servá nio cuatro hileras de espiguillas por racimo nümero de hileras de espiguillas en cada racimo o para- cladio largo es una consecuencia del grado de ramifica- ción del eje de los paracladios cortos (Rua & Weberling, 1995). La presencia de espiguillas dispuestas en pares es un carácter frecuente dentro de las Paniceae y se consi- dera que es una condición primitiva siendo interpretada como una sinapomorfía potencial de la tribu (Kellogg. 2000) 8. Pilosidad de la espiguilla: (0) glabra; (1) pilosa. sencia de espiguillas provistas de pelos es pro- Tiv pis una es — ion para la dispersion por ad- hesión o a través del viento de la diáspora. Ejemplos de esto son Paspalum conjugan el cual presenta espigui- llas con pelos distrib un delicado borde par altamente efectivo para la dis- persion por adhesión: y P. saccharoides que crece en zonas esenta espiguillas con Mos pelos se- sc vales facili- 87). E algo elevadas v pr dons intercalados tan la dipersion — el vie grupo Ceresia se caracteriza por P presencia ү, es smod llas — uamente pilosas, con pelos sedosos distribui- samente sobre los márgenes; un patrón similar = © — & = = к palum, como Dilatata y Dissecta, la presencia de espigui- llas pilosas es un carácter polimórfo dentro del grupo. 9. Gluma inferior: (0) ausente; (1) presente. Uni de ae carac terística as cae a miel — pigu — bargo, muc * especies de Pu E n phos: alguna espiguilla del racimo con ls gluma inferior par- с ialmente desarrollada, en algunos ejempla- s de P. distichum y р. аш (C hase, 1929). omo ocur 10. Gluma superior: (0) ausente: (1) presente. La ausencia de gluma superior es un carácter poco fre- cuente dentro de Paspalum, siendo una característica dis ntiva del subgénero Anachyris. A pesar de ello, — especies de dicho subgénero poseen gluma jd con diferentes grados de desarrollo (Morrone et al., ). Asi- mismo, algunas especies pertenecientes a ш, grupos Ra- cemosa, Gardneñana. Parviflora y Reimaria, también ca- recen de gluma superior — & Morrone, inéd.). Renvoize (1972) considera que 1 de la gluma superior se pudo haber originado — ndientementé va- rias veces dentro ^i género, posiblemente como una len- dencia adaptativ 11. Nervios de la lemma superior: (0) conspicuos; (1) no conspicuos. 1 presencia de lemma superior con nervios conspi- cuamente marcados es característica. de Anachyris; la misma presenta la cara abaxial sur cinco nervios prominentes, sie к ésta una sinapomorfia del subgénero (Morrone et al., 2000). subgénero 12. Consistencia del antecio: (0) membrandceo; (1) cortá- ceo. Según Cialdella y Vega (1996) el ane cimiento y es- pecialización del antecio superior en las Paniceae podría = — = ш. V Ф — — c 22 ее -l — led que much lum, junte imente con otros géneros de la misma tribu pre- sentan antecios de consistencia más dura debido a que estarían adaptados a mecanismos de dispersión de la di- ora a través de animales herbívoros. La dureza del an- * io permitirſa que el mismo se mantenga viable atin des- pués de atravesar el tracto digestivo del animal. B z Forma del antecio: (0) elipsoide; (1) obovoide. . Coloración del antecio: (0) pajizo; (1) castaño. CARACTERES ANATÓMICOS FOLIARES . Porciones laterales del transcorte: (0) ausentes o in- conspicuas; (1) desarrolladas. La mayoría de las especies de Paspalum presentan ho- jas planas o expandidas: sin embargo en alguno F laterales de la lamina no desarrollan, originando ho- as de aspecto filiforme las cuale foliares reducidos a : zona de la costilla central de la do: lámina (Aliscioni, 200( S Casos las s presentan transcortes — T 16. Contorno del transcorte foliar: (0) expandido o en for- nvoluto a ma de “V” abiert JNA ct — ado; (2) co subc onvoluto; (3) jo ; (4) semicircular (ver Alis- cioni, 2000, fig. 3). 17. Superficie foliar — con costillas y surcos diferen- ciados: (0) ausente: (1) presente. 18. Superfici ie н — con — y surcos diferen- ciados: (0) ausente; (1) p 19. Costilla — (0) adas (1) — 20. Cavidades aeríferas: (0) ausentes: (1) p Las cavidades aeríferas observadas en distintas be cies de Paspalum se presentan en la costilla central de la lámina; las mismas pueden presentar distinto grado de desarrollo pero en todos los casos se originan a partir de la lisis de células рае as i incoloras. La presencia de dicho carácte s que habitan lugares con exceso de — (Aliseioni, 2000, fig. 2). er se aso 2). ri y alternancia de haces vasculares: (0) un haz de 3° orden entre un haz de l° y uno de 2° 520 Annals of the Missouri Botanical Garden orden; (1) tres haces de 3° orden entre un haz de 1° y uno de 2° orden; (2) sin distinción entre haces de 2° y orden. 22. Pe osición de los haces vasculares: (0) todos los haces dispuestos en el centro del transcorte: (1) haces de 1° 2° orden dispuestos en el centro del transc orte y haces 5 3° orden desplazados hacia la cara aba: — La posición de los haces vasculares en el transcorte y su patrón de — ‘ión es un importante carácter diag- nóstico (E llis, 1976); sin embargo en algunas especies la proporción а eed de los diferentes haces varía desde la jm media de la lámina hacia el margen. Con la finalidad de « um observaciones fuesen comparables, los carac- ie y 22 fueron — en la porcion media de cada una de las semilámin: 23. Relación entre los grupos > células — y los haces озен — 30 : (0) células buliformes — soie buliformes sobre un en; 2 ts buliformes sobre tres o más haces de 3° orden La ubicación de las células buliformes con respecto a los haces vasculares es un carácter que se asoc la, en par- te, al фа fotosintético de la especie. Según Renvoize 9679/9 1 los géneros de la tribu — fotosintéti- camente С, М den subtipo MS, « hana Thrasya у disi, tei las células bulifor- mes se ubican siempre sobre haces vasculares menores о Axonopus, A diferencia de esto, las células buliformes se ubican en- tre haces vasculares en Panicum laxum, especie fotosin- téticamente C, y en Клос hloa, — que presenta tipo anatómico- -fisiológic o C, PCK, subtipo PS. 24. sn mestomdtica: (0) sin — (1) con clo- 25. Vaina ООС (0) sin cloroplastos; (1) con cloroplas оз caracteres 24 y 25 hacen referencia a caracterís- tic as anatómicas que permiten inferir la vía fotosintética species. Para establec тег las — sintéticos se siguió el criterio de Brow 971) y lar et al. (1985). Dichos autores inde la vaina Kranz de las ыо. C, con subtipo anatómico MS, a la vaina interna de las especies С, con subtipo anatómico PS y a la vaina interna de las especies C, non- Kranz. s son mestomáticas por su origen ya que se ge ^ ran del mismo meristema que A diferencia de ello, las « En los tres casos estas vainas origina al tejido vascular. spe- cies C, anatómicamente PS, presentan la vaina Kranz en »sición externa la cual se homologa a la vaina paren- quimática de sea "e les Cy, pues en ambos casos estas el meristema fundamental juntamente con lis células инш as del mesofilo. 26. Vaina parenquimática: (0) ausente; (1) remanente; (2) presente. Morrone et al. (1993) ооа en el género Anthae- nantiopsis la presencia de algunas células por fuera de la vaina mestomática, de aspe * poen y con — gruesas, con o sin cloroplastos especializados; dichas cé- lulas son interpretadas por los autores como restos de una vaina a a en proceso de reduc Los caracteres 24, 25 y 26 son características anató- nicas M MM a distintos procesos fotosintéticos y fue- ron ооа con el objeto de resolver relaciones еп los taxones exter — = os, con lo cual se puede prevenir algu- nas topologias erróneas del grupo interno causadas por incorrectas resoluciones del grupo externo (Nixon & Car- penter, 1993). 27. Células distintivas Kranz: (0) ausentes; (1) presentes. Dichas células presentan carac terísticas ultraestructu- por más de cuatro células clorenquimati as; las célula distintivas se conectan a los haces vasculares vecinos por — de venillas transversales (Sánchez & Arriaga, 1989). Segün Haneler y Watson (1992) podrían inter- pretarse como reducciones extremas de haces vasculares de menor orden. Su presencia dentro del género Paspalum fue corroborada para los grupos Bonplandiana y Race- mosa, A citadas por Cialdella et al. (1995) y Morrone et al. (1995) respectivamente. 28. Células buliformes: (0) cubriendo casi toda la super- ficie; (1) en grupos definidos. = Las células buliformes se presentan en la epidermis adaxial y pueden formar grupos regulares bien definidos con aspecto de abanico, dispuestos sobre uno a tres haces terciarios. En otros casos ibuyen en casi toda la excepto sobre los haces mayores, diferencián- dose de las restantes células epidérmicas solamente por ser de mayor tamaño se distr superficie, 29. Grupos de células buliformes asociadas a tejido paren- quimático incoloro: (0) ausentes; (1) presentes. 30. Parénquima incoloro dispuesto en las alas: (0) ausen- presente. — ~ Este carácter hace — ‘ia н — чойо asociado a el parénquima incoloro asociado as células les y la zona parenquimática de la costilla central. 31. Pared tangencial externa de las células ы аў adaxiales: (0) recta; (1) convexa; (2) papi 32. Pared Si d externa de las células ніта abaxiales: (0) recta; (1) convexa; (2) papilos Los caracteres 31 y 32 se refieren a las células epidér- micas observadas en corte transversal. Las mismas difie- ren en la forma que adquiere la pared tangencial externa, pudiendo ser totalmente recta шан 0) o queada o convexa dando a la « (estado Se — célula papilosa (estado 2 mente cuando la pared tangencial externa se evagina no- toriamente Bn una protuberancia digitiforme. 33. Macropelos adaxiales: (0) ausentes; (1) presentes. 34. Macropelos abaxiales: (0) ausentes; (1) presentes. La presencia de macropelos es un carácter muy fre- cuente en el género Paspalum; la mayoría de las especies poseen macrope plos en una o ambas epidermis, pudiendo ser largos y a engrosadas, con la base bulbosa hundida en la ке ине asociada en algunos casos a células epidérmicas sobreele- vadas formando un cojín basal. Si bien los mismos pre- sentan una marc A variación en cuanto a su longitud, frecuencia, espesor de la pared y ángulo de emergencia, sólo se nid ea aquellos aspectos más estables por Volume 89, Number 4 2002 Aliscioni 521 Filogenia de Paspalum lo cual tales características no fueron d incluidas en el aná- lisis debido su variación intraespecífic 35. Asperezas adaxiales: (0) ausentes; (1) presentes. Asperezas abaxiales: (0) ausentes; (1) presentes. El término aspereza se refiere a células epidérmicas con la pared tangenc ‘ial externa aguda; las mismas pueden di- el tamaño re- »yección exte de la misma (Ellis, 1976). Dentro (A género Paspalum és se observaron formas intermedias por al se reunió a am- bos tipos dentro de un mismo carácter. ferenciarse € N — < % ш Y — Annals of the Missouri Botanical Garden Apéndice 3. Matriz de datos. Referencias: | | indica carácter polimorfo; — indica carácter no comparable 1 2 3 4 5 6 1 8 9 10 11 12 13 14 15 16 17 18 Panicum laxum 0 1] 2 0 l 1 1 [01] l | 1 0 0 0 1 0 l 0 Eriochloa 0 0 0 0 1 0 0 1 l l 1 0 0 0 1 0 О 0 Anthaenantiopsis 0 0 2 0 l l l l l | 1 0 0 0 1 0 1 0 Axonopus 0 1] 2 0 l 0 0 0 0 l 1 1 0 0 l 1 0 0 Thrasya 0 0 0 l l l l l l l 1 — 0 0 1.0 0 0 P. malacophyllum 0 0 2 | l l l 0 0 0 0 1 0 0 | 0 0 0 2 simplex 0 0 2 l l 1 1 0 0 0 0 1 0 0 1 0 0 0 P. bertonii 0 0 1 l 0 | 0 l 0 | 1 0 0 0 1 2 l 0 P. lilloi 0 0 l 0 0 | 0 1 0 1 | 0 0 0 | 2 1 0 P. bonplandianum 0 | 2 | 0 | 0 0 0 l l 1 0 0 1 0 0 0 P. pallidum 0 1! 2 І 0 | 0 0 0 | 1 |] 0 0 1 0 0 0 P. caespitosum 0 0 2 0 l | 1 0 0 1 1 1 0 0 1 0 0 0 P. indecorum 0 0 2 0 l | |] 0 0 | 1 1 0 0 1 0 1 0 P. polyphyllum 0 0 [2] 0 1 1 1 1 0 1 1 о 0 ото 0 0 P. stellatum 0 0 [12] 1 0 | 0 l 0 | 1 0 0 0 1 0 1 0 р. conjugatum 0 | l 0 0 | 0 l 0 | 1 1| 0 0 1 0 0 0 P. corcovadense 0 0 2 0 l l l 0 0 1 1 1 0 0 1 0 0 0 P. mandiocanum 0 0 2 0 1 1 |] 0 0 1 1 1 0 0 1 0 0 0 P. dilatatum 0 0 2 0 1 | |] | 0 1 1 1 0 0 1 0 0 0 ? urvillei 0 0 2 0 1 l l l 0 1 1 1 0 0 1 0 0 0 P. repens 0 | 2 l 0 1 0 | 0 | 1 1 0 0 1 0 1 1 P. distichum 0 1] 1 0 0 1 0 [OI] [Ol] 1 1 1 0 0 1 0 l l P. vaginatum 0 1] 1 0 0 1 0 01] 0 111 0 0 1 0 l 0 P. gardnerianum 0 0 2 0 l | 1 [Ol] 0 0 1 1 | | 1 0 0 0 Р. ammodes 0 0 2 0 0 1 0 1 0 ] ] 1 0 0 |] 1 0 0 P. erianthum 0 0 2 0 l | | l 0 | 1 1 0 0 1 0 1 0 P. equitans 0 0 2 0 0 | 0 0 0 1 1 1 0 0 1.0 1 0 P. fasciculatum 0 | 2 0 0 | 0 0 [OI] 1 1 1 0 0 1 0 1 0 P. filiforme 0 0 0 0 0 1 0 0 0 ] 1 1 0 0 0 4 — — P. lindnerianum 0 0 [l] 0 0 1 0 0 0 | 1 |! 0 0 0 4 س‎ — P. inaequivalve 0 1] 2 0 1 1 1 0 0 1 1 1 0 0 1 0 0 0 P. lineare 0 0 [12] 0 0 1 0 [01] 0 1 1 1 0 0 0 3 — — P. proximum 0 0 [12] 0 0 1 0 [01] 0 1 1 1 0 0 1] l 0 0 P. alcalinum 0 0 2 0 1 1 1 [01] 0 1 1 |! 0 0 1 0 0 0 P. lividum 0 1 2 І 1 1 1 [01] 0 |o l 1| 0 0 1 0 0 0 Р. macrophyllum 0 0 2 0 l | ] 0 0 | 1 |] 0 0 | 0 0 0 P. notatum 0 | | 0 0 | 0 0 0 1 1 1 0 0 1 0 1 0 P. pumilum 0 0 l 0 0 1 0 0 [01] 1 l l 0 0 1 0 0 0 P. orbiculatum 0 | | 0 0 | 0 0 0 1 1 1 1 0 1 0 0 0 P. juergensii 0 0 2 0 l 1 1 [01] 0 | 1 1 0 0 1 0 0 0 P. paniculatum 0 0 2 0 1 l | [01] 0 l l l 0 0 1 0 0 0 P. crispatum 0 0 [12] 0 1 1 1 0 0 l l 1 0 0 1 0 0 0 P. parviflorum г O [I2] 0 0 | 0 0 0 l l 1 0 0 1 0 0 0 Р. modestum 0 | 2 0 0 1 0 [Ol] 0 l 1 |] 1 1 1 0 0 0 P. plicatulum 0 0 2 0 l 1 1 [Ol] 0 | l 1 l | 1 0 0 0 P. coryphaeum 0 |] 2 0 l l 1 [Ol] 0 | |] 1 [Ol] O 1 0 0 1 P. quadrifarium 0 0 2 0 l | 1 [0l] 0 1 1 1 [ol] O 1 0 l [01] P. penicillatum l l 2 l 0 | 0 0 0 1 | l 0 0 1 0 0 0 P. pygmaeum 1| 0 2 1 0 1] 0 0 0 ! ! ! 0 0 ! о O 0 P. saccharoides 0 0 2 0 0 | 0 1 0 l 1 0 0 0 1 0 | 0 Р. regnellii 0 0 2 0 l | 1 [0l] 0 1 1 I] 1 | | 0 0 0 P. virgatum 0 0 z 0 | 1 І [01] 0 | 11 1 1 1 0 [0l] 0 Volume 89, Number 4 Aliscioni 523 2002 Filogenia de Paspalum Apéndice 3. Extendido. 19 20 21 22 23 24 25 26 27 28 29 30 3132 33 34 35 36 Panicum laxum l 0 Z 0 о о [Ol] 2 0 1 0 0 00 0 0 0 0 Eriochloa 0 0 0 0 0 0 1 2 0 1 0 0 0 1 [I] 0 | l Anthaenantiopsis 1 0 0 1 2 1 [Ol] 1 0 I | 0 0 0 [Ol] [01] [OI] [Ol] Axonopus 1 0 [Ol] 0 2 1 — 000 0 о 1 о [Ol] [01] 0 0 Thrasya 1 0 [Ol] |! 2 1 — 00 ] 0 0 0 | l l l l P. malacophyllum l 0 l 0 2 |] — 000 0 0 0 ! l l l l P. simplex [01] 0 l 0 2 1 — 0 0 0 0 0 0 ! l l | 1 P. bertonit 0 0 1 l 2 1 — 00 ] 0 1 2 0 0 0 0 0 P. lilloi 0 0 l 1 2 1 — 0 0 1 0 1 2 0 0 0 0 0 P. bonplandianum |01] 0 0 0 l 1 — 0 1 1 0 0 1 1 l 1 [Ol]. [01] P. pallidum 01] 0 0 0 l 1 — 0 10 0 0 1I 1 | | 0 0 P. caespitosum l 0 0 1 l 1 — 0 0 1 0 0 0 0 [OI] [Ol] ! 1 P. indecorum 1 0 | 1 2 1 — 0 0 1 0 0 0 2 l l l 1 P. polyphyllum 1 0 l 0 2 1 — 00 0 0 0-2 2 ] 1 0 0 P. stellatum 1 0 l l 2 ] — 00 1 0 0 1 0 1 1 [01] [01] P. conjugatum 1 0 | 0 2 1 — 000 0 0 1 O [OI] [Ol] 0 0 P. corcovadense ] о [Ol] 1 [R] 1 0 0 0 I 0 0 0 1 [Ol] 1 [Ol] 1 P. mandiocamum 1 O [Ol] [Ol] [I2] 1 0.0 ] 0 0 0 1 [Ol] ! | | P. dilatatum 1 0 l l | 1 — 00 I 0 0 1 1 0 [о] ! | P. urvillei l 0 1 l l 1 — 0 0 1 0 0 ] 1 0 0 | | P. repens 1 l 0 0 | 1 — 0 0 1 0 0 2 2 1 l l l P. distichum 0 0 0 l | 1 — 0 0 1 0 0 2 2 0 0 0 0 P. vaginatum 0 0 0 1 | 1 — 0 0 1 о [0] 2 0 0 0 [OL] [01] P. gardnerianum [Ol] O [Ol] 0 [12] | — 00 1 0 0 1 1 [о [Ol] 1 [01] Р. ammodes 1 0 0 l | 1 — 00 I 0 0 1l l l l | | P. erianthum 1 0 [Ol] [01] [I2] 1 — 0 0 I о O 1 0 [OI] [Ol] 1I [OT] P. equitans 1 0 Jo] 1 [2] 1 — 001 0 0 20 0 0 0 [01] P. fasciculatum 1 о [Ol] 1 [R] 1 — 0 0 I 0 о 0 0 0 0 1 [Ol] P. filiforme 1 о — — — 1 — о 0 1 0 0 0 0 0 0 | 1 P. lindnerianum 1 о — — — | — о 0 I 0 о 00 0 0 | 0 P. inaequivalve 1 0 0 0 l 1 0 0 0 1I Q0 0 0 0 [OI] [OI] [01] [01] P. lineare l 0 — — 1 — 00— — — —0 — 0 — 0 P. proximum 1 0 0 l — 1 — 0 0 — — 1 2 0 0 0 [Ol] 0 P. alcalinum 1 [Ol] ! 1 | 1 — 0 0 1 0 O° I 2 l l l 1 ividum 1 [Ol] | 0 2 1 — 0 0 | 0 о 1 2 l l l l P. macrophyllum l 0 l 0 2 1 — 0 0 1 0 0 1I 1 0 O [Ol] [Ol] P. notatum [01] 0 0 1 | 1 — 00 1 0 0 0 0 [Ol] [Ol] 1 0 P. pumilum [01] O [01] 1 |2 1 — 00]! 0 0 0 1 l | [Ol] [01] P. orbiculatum 0 0 0 0 l 1 — 0 0 1 0 0 1 1 [Ol] [Ol] 0 0 P. juergensii 1 0 0 0 | 1 — 0 0 1 0 0 0 1 [OI] [Ol] 1 | P. paniculatum [01] 0 0 0 | 1 — 0 0 1 0 0 0 1 [Ol] 0 l l P. crispatum l 0 1 0 2 1 — 000 0 0 2 2 |] l 1 [01] P. parviflorum [01] 0 1 1 2 | — 0 0 0 0 0 1 ] 1 ] [Ol] 0 Р. modestum 1 1 0 | l 1 — 0 0 1 0 0 1 1 0 0 0 0 P. plicatulum 1 0 l l 2 1 — 00 ] O O 1 1 [01] [01] | 1 P. coryphaeum [01] 0 1 | 2 1 — о 0 0 0 0 1 1 l l | l P. quadrifarium 1 0 0 | | 1 — 0 0 I 0 0 1 | 0 0 l l P. penicillatum 1 0 0 0 | ї — 0 1 0 Q O 1 1 [01] [01] 0 0 P. pygmaeum [01] 0 0 0 l 1 — 010 0 0 1l | l l 0 0 P. saccharoides 0 о [Ol] 1 [l2] | — 0 0 I 0 0 2 0 1 0 l 0 P. regnellii mic 0 6 1 li — 001 9 9 lJT5 NR 1.3 P. virgatum І [Ol] 0 1 l 1 — 0 0 1 [Ol] 0 1 1 [Ol] 0 [I] 0 GOCHNATIA (ASTERACEAE, Susana E. Freire”, Liliana Katinas’, and MUTISIEAE) AND THE Gisela Sancho? GOCHNATIA COMPLEX: TAXONOMIC IMPLICATIONS FROM MORPHOLOGY! ABSTRACT Gochnatia is one of the largest genera of the tribe инен (Asteraceae) and has been traditionally charac ۰ by its homogamous capitula with isomorphic corollas. A morphological study of Gochnatia and associated genera, i.e., Actinoseris, Chucoa, Cnicothamnus, Cyclolepis, Hyalis, aa Nouelia, Pleiotaxis, and Wunderlichia, рн out to evaluate the circumscription of Gochnatia and its sections, and the a of this complex of genera, The бике studied involve habit, leat баи — ушевсепсе typ Dess тап, and venabionA a 9 5 ST sex, — of oe ee a о es, ache nial pu bee 'ence, end pay а MSN * these f d: (1) although Gochnatia is highly variable in most of the characters studied. it can be defined by this suite of a isomorphic to subdimorphic — ulate anther appendages, and smooth style branc hes; (2) ol = > ы & = = £z 5 = РА = = ^ D A £ = sections of Gochnatia needed to be re-evaluated. As result of this, two sections, i.e., sect. Discoseris and sect. Penta- phorus, are confirmed; two е иш аге opel i i.e., sect. Glomerata sect. nov. and sect. Rotundifolia sect. ; three sections are redefined, i.e., sect. Hedraiophyllum, sect. Leucomeris, and sect. Gochnatia, while sect. Anassnaphiotdes | is formally published; and (3) the combination of apic — anther appendages and smooth style brane — is unique to Actinoseris, Cnicothamnus, Cyclolepis, Gochnatia, Hyalis, lanthopappus, and Nouelia within the Mutisieae. This group of genera is recognized here as the Gochnatia complex, with Gochnatia as the basal genus of this нен Chucoa, Pleiotaxis, and Wunderlichia do not belong to the Gochnatia complex. Key words: Asteraceae, Compositae, Gochnatia, Gochnatia complex, infrageneric classification, morphology, Mutis- ieae. Gochnatia Kunth is one of the largest genera of 1838; Jervis, 1954; Cabrera, 1971) since the genus the tribe Mutisieae, subtribe Mutisiinae sensu lato was established by Kunth (1818). Cabrera (1971), (including Gochnatiinae; Robinson, 1991; Bremer, in his monograph of the genus, divided it into six 1994). It comprises 68 species, nearly all Neotrop- sections: sect. Discoseris, sect. Gochnatia, sect. ical and 2 endemic to the mountains of southeast- Hedraiophyllum, sect. Leucomeris, sect. Moquinias- ern Asia. All the species of Gochnatia have been trum, and sect. Pentaphorus. He suggested the ar- traditionally described as discoid with actinomor- — tificial delimitation of some sections, such as Hed- phic, deeply 5-lobed corollas, features that are ple- raiophyllum, and in other cases emphasized siomorphic within Mutisieae (Bremer, 1994). The geographical distribution in distinguishing groups only apomorphic character suggested for the genus of species, such as in section Gochnatia. is the acuminate to apiculate apical anther ap- Jeffrey (1967), when relating Gochnatia to other | I | h 8 »endage of the stamens (Bremer, 1994). This char- taxa, established the Stifftia-series of genera. which I 8 J 8 acter, however, is shared by other genera of Mutis- all have short, rounded, glabrous style arms, and іеае. commonly glabrous corollas. The Stifftia-series was The infrageneric taxonomy of Gochnatia has then divided into four subseries by slight differ- been much discussed (Lessing, 1832; de Candolle, ences in the shape of the style arms, although these ' We thank Vicki Funk, Per Ola Karis, Richard Keating. Harold Robinson, Peter Stevens, and Tod Stuessy for he Ipful Ma ин on an earlier draft of this manuscript, and one anonymous reviewer comments. We are espec ially grateful to John Pr who enthusiastically improved the manuse ript with his suggestions. We are ue — ful to the curators of G, K, LP. МО, NY, S, Sl, and US for the loan of specimens, and Nadia Roque for access to her prepublication manuscript on lanthopappus (now published i in Novon). Finally, we thank Victoria Hollowell * 'ause ы editorial w was extremely helpful, and Víctor H. Calvetti for inking our — pencil illustrations. Support for this study by the ;onsejo Nac ional de Investigaciones Científicas y Técnicas (CONICET), ), Argentina, the Smithsonian Institution, Wash- ington, D.C. (for G.S.), and the National Geographic Socie ty (grant 5776-96) is gratefully acknowledged. ? División Plantas Vasculares, Museo de La Plata, Paseo del Bosque, 1900 La Plata, Argentina. freire@museo. fenym.unlp.edu.ar (author for correspondence). ANN. Missouni Bor. GARD. 89: 524-550. 2002. Volume 89, Number 4 2002 Freire et al. 525 Gochnatia Morphology differences were not described for each subseries. Of these subseries, the Gochnatia-subseries includ- ed Achnopogon, Cnicothamnus, Gochnatia, Nouelia, and Oldenburgia. Cabrera (1971) considered Actinoseris, Cyclole- pis, and Pleiotaxis closely — to Gochnatia by their apiculate anthers , he also associated Chucoa with Gochnatia CREE 19 ansen (1991) considered only Actinoseris and Cyclolepis of Cabrera's group to be close to Gochnatia, excluding Pleiotaxis and Chucoa, and added Hyalis and Nouelia to what he called the “Gochnatia-group.” Hansen argued the presence of cone-like involu- cres, i.e., with light brownish bracts imbricately ar- ranged and resembling the cone of a spruce as a probable synapomorphy of the group. Recently, Roque and Hind (2001) created the new genus /an- thopappus for the species Actinoseris corymbosa, In this work, the authors grouped the genera Actinos- eris, Chucoa, Cnicothamnus, Cyclolepis, Gochnatia, Hyalis, lanthopappus, and Nouelia by their apicu- late to acuminate apical anther appendages. Fur- thermore, Roque (2001) and Roque and Pirani (2001) re-circumscribed the genus Richterago Kun- tze according to the genus concept first employed y Lessing (1830) to include the species of Acti- noseris and the species of Gochnatia sect. Discos- — eris. We consider, however, that the characters used to distinguish Richterago are widespread in Gochn- atia or are shared with other genera, such as her- baceous to subshrubby habit (shared among Gochn- atia sect. Discoseris, Actinoseris, and other such related genera as Hyalis and lanthopappus), leaves rosulate to alternate (alternate leaves are present in most species of Gochnatia), venation pinnate (pre- sent in most species of Gochnatia), capitula ho- mogamous and discoid (present in most species of Gochnatia) or heterogamous and radiate (present in some species of Actinoseris), as well as pappus fea- tures, i.e., pappus uniseriate, with 25 to 42 bristles, united into a fleshy ring (present in species of Gochnatia sect. Gochnatia). For these reasons, we affirm here the traditional concept of Gochnatia and Actinoseris (excluding A. corymbosa = lanthopap- pus corymbosus) as was established by Cabrera (1970, 1971). Some phylogenetic studies relate Gochnatia with other genera of Mutisieae; for example, Jansen and Palmer (1987) related Gochnatia with Ainsliaea, Onoseris, and Stifftia, and Karis et al. (2001) relat- ed it with Mutisia and Trixis. However, because of the few taxa sampled in the tribe, these analyses are not considered here to address relationships at the generic level. In other analyses (Karis et al., 1992; Jansen & Kim, 1996), Gochnatia was found to be an isolated taxon within Cichorioideae. The cladogram of Karis et al. (1992), based on morphological characters, shows Gochnatia as sister to most Asteraceae, ex- cluding Barnadesioideae and four genera of Gochn- atiinae. The ndhF tree of Jansen and Kim (1996) also shows Gochnatia as an independent lineage positioned basal to most Asteraceae, excluding Bar- nadesioideae and the core of Mutisieae examined. ere are other genera within Mutisieae such as Quelchia and Neblinaea from the Guayana High- land with apiculate anther appendages, but mor- phological features of these genera (Pruski, 1991; Bremer, 1994) mark a departure from Gochnatia and associated taxa. On the other hand, the plan- altive Brazilian genus Wunderlichia, with shortly apiculate anther appendages, was either placed in (Pruski, 1991) or it was considered isolated within Mutisieae (Hansen, 1991: 1992; Bremer, 1994). The group selected here for the analysis is main- the “Stenopadus group” Karis et al., ly represented by the genera with apiculate anther appendages apes, by Jeffrey (in part, 1967), Cabrera (1971, 1977), Hansen (1991), and Roque and Hind al It is comprised of Actinoseris sen- su Cabrera, excluding A. corymbosa (7 species), Chucoa (1 species), Cnicothamnus (2 species), Cy- clolepis (1 species), Gochnatia (08 species), Hyalis (2 species), lanthopappus (1 species), Nouelia (1 species), Pleiotaxis (ca. 25 species), and the con- troversial Wunderlichia (6 species). There are several potential diagnostic features in Gochnatia and relatives such as the habit, leaf mor- phology, types of capitulescence, involucre, florets trichomes, and pappus that have never been stud- ied comparatively or in detail. A detailed morpho- logical study of Gochnatia and its associated taxa is needed as a first step to provide a robust base for discussion. On the basis of morphological evidence, the goals of this study are: (1) to evaluate the circum- scription of Gochnatia and its sections, and (2) to eroup Gochnatia with other genera based on their similarities. MATERIAL AND METHODS Herbarium specimens of the studied taxa, i.e., Actinoseris, Chucoa, Cnicothamnus, Cyclolepis, Gochnatia (64 of its 68 species), Hyalis, lantho- pappus, Nouelia, Pleiotaxis, and Wunderlichia (Ap- pendix 1), were examined to assess characters used previously to distinguish taxa in this group of gen- era, and to search for additional characters. The data were augmented by literature studies (Fran- 526 Annals of the Missouri Botanical Garden chet, 1888; Cabrera, 1950, 1951, 1955, 1970, 1971; Jeffrey, 1967; Barroso & Maguire, 1973; Zar- dini, 1975; Sancho, 1997, 2000; 1997; Roque & Hind, 2001; Roque & Pirani, 1997, [ Roque, Vegetative and floral parts were dissected and observed after boiling in water and stained with 2% safranin. Freehand sections of leaves were per- formed and these sections examined to determine the presence of hypodermis, and then stained with safranin. Drawings were made by the authors using E stereomicroscope Wild M5 and ё microscope Leitz SM Lux with the camera lucida technique. Apical anther appendages were described and drawn to include the portion ranging from the the- cae apex to the apex of the stamen. Descriptive terminology for the trichomes follows Ramayya (1962), synonyms of Whenever possible, additional trichome terminology were added (Payne, 1978; Font Quer, 1979; Moreno, 1984; Harris & Harris, 1994; Metcalfe & Chalk, 1950). Some authors use the term “trichome,” whereas others apply the term “hair” in their pubescence classifications. Both terms are cited here following each author classification. RESULTS For Gochnatia with their corresponding species accord- ing to Cabrera (1971) are listed in Table 1. comparison with our results, sections of VEGETATIVE CHARACTERS Habit. Таха under study display four types of habit: small trees, shrubs, subshrubs, and perennial herbs. Several species of Gochnatia (e.g.. G. arbo- rescens, G. decora, G. elliptica, G. ilicifolia, G. mag- na, G. oligocephala, G. palosanto, G. polymorpha, G. spectabilis, G. Wunderlichia tortuensis), Cnicothamnus, Most and and are small trees. species of Nouelia are shrubs. Gochnatia sect. Discoseris, Hyalis, and lan- Gochnatia, Chucoa, Cyclolepis, thopappus are subshrubs. Actinoseris and Pleiotaxis are perennial herbs. Cyclolepis is the only spinose genus and is aphyllous at maturity. Leaves. All taxa analyzed have alternate, oc- casionally rosulate, subsessile to shortly petiolate leaves. Leaf consistency: The leaves are coriaceous or subcoriaceous, but only some species have leaves with an adaxial hypodermis. It is absent in some species of Gochnatia (e.g., G. amplexifolia, G. are- quipensis, G. cardenasii, G. discoidea, G. foliolosa, G. glutinosa, G. hypoleuca, G. intertexta, G. magna, G. palosanto, G. purpusii, G. rotundifolia, G. ver- nonioides), Actinoseris, Cnicothamnus, Cyclolepis, yalis, lanthopappus, and in some species of Pleio- taxis (e.g.. P. eximia, P. huillensis). Occasionally, the hypodermis is discontinuous (e.g., Gochnatia argentina, G. discolor, G. orbiculata, G. ramboi). Leaf shape: The leaves are always simple ir the taxa under study, but show great variation in shape, from linear to suborbicular. Most species of Gochnatia sect. Gochnatia and Nouelia have leaves ovate to ovate-elliptic, obtuse to subobtuse at the apex; occasionally ovate leaves are acute at the apex (e.g.. G. polymorpha, G. vernonioides), or cor- date at the base (e.g., G. cordata, G. haumaniana) (Fig. 1A). Most of the Caribbean species of Gochn- atia sect. Gochnatia, Actinoseris (A. hatschbachii, A. polymorpha, A. radiata), Chucoa, lanthopappus, and Wunderlichia have obovate leaves (Fig. 1B). Elliptic leaves (Fig. 1C) are rounded in the base and apex in G. discoidea, attenuate at the base and apex in G. magna, G. sect. Leucomeris, Actinoseris (A. polyphylla), Cnicothamnus, and some species of Pleiotaxis (e.g., P. huillensis, P. newtonii, P. selina, P. subscaposa). Gochnatia sect. Pentaphorus, G. ar- gyrea, Actinoseris (A. angustifolia, A. stenophylla), Cyclolepis, Hyalis, and most species of Pleiotaxis (e.g.. P. ambigua, P. decipiens, P. dewevrei, P. li- nearifolia, P. rogersii) have linear leaves (Fig. 1D). Leaf margin: Only the Caribbean species of Gochnatia sect. Gochnatia (except G. crassifolia and G. picardae) and Chucoa have spinose margins (Fig. 1E). In the remaining taxa the leaf margin is predominantly entire or denticulate (Fig. 1F). sal venation: The venation is predominantly pinnate, although in taxa such as Gochnatia are- quipensis, G. glutinosa, G. rotundifolia, Hyalis, and lanthopappus the leaves are three-veined (actinod- romous). CAPITULESCENCE Capitula are borne singly or 2 or 3 at the end of the branches or may be clustered in pseudoco- rymbs, pseudoracemes, or pseudopanicles, in open to condensed or glomerulose synflorescences de- pending upon the length of the peduncles. Solitary or few capitula (2 or 3) are short-pedun- culate to glomerulose in Gochnatia sect. Gochnatia, Cnicothamnus, Nouelia, and some species of Wun- derlichia (W. azulensis, W. crulsiana, W. mirabilis) (Fig. 2A). They are scapiform, long-pedunculate in Gochnatia suffrutescens, some species of Actinoseris e.g., A. angustifolia, A. hatschbachii, A. polymor- pha, A. radiata), — Chucoa, and some species of Pleiotaxis (e.g.. P. subscaposa) (Fig. 2B). 527 Freire et al. Volume 89, Number 4 2002 Gochnatia Morphology (q 2d4) y ada) sndded :sausyor ur sarey їрүприрүз ayeyideo ((әүетштәв]) spre; оош» “(әле -noide Apdniqe) apne? o3epuadde ләцир :рәдо| -c Ájdəəp Аләл о} Аүдәәр se[[o100 :(e 10 Z) e] -mudes Алгүов “шлеш (әлциә) asourds YIM soAg»] g edá1 sndded ‘рез 9jeruroe[ *91enuo] -e 93epuodde ләццир :a]« se[[o100 :surg eur au үр aje[orpro “SOLIR[[AYd snosqepsqns YIM әләпүол -u1 te[njrdeo Але ISNOIGR]S "рәшәл-әәлу SAARI] g 2d4 sndded :s[1e] ajeruroe[ “(aypenoide Арапе) ayepneo osepuedde toque (е 40 Z) e] -njdeo Are [Os :рәшәл (-9291g1qns) Ayeyeuuid SOAR’ y edaq sndded :s[ıe ature] *aje[noide Apdnaqe a3epuadde tsqui&i000pnosd asodeos ш sno1ouinu 10 *oje[nounpod-zuo| “Á1B]1]OS e¡mideo “SQnIysqng UB9q4 UB”) eqno) ә паә URJIUTWO iqndoy ueorunuoq eqn’) ueaqq.ie") eqn’) rag s nd N eunuoziy MN mjog S ernog enog рцорәлр “x т1тәшоЎ 74) Di]upoauua °“) 21111112 “) nunuyə °“) sisuaqno °“) D1 jofissp49 4) sapionjdpajspuy “1998 DIJDUYIO?) DIJOJIPUNIOL 2) y, Dijo fipunioy “1998 DIDUYIOS) sapio1uouA424 ©) пкр °<) puizpjnd °<) n4opfiaamo °<) NSDUIPADI °<) DUDIAO %9 sisuadimbaap “9 DIDUYIOY *1238 DIIDUYIOY sua2sajnaffns °<) DIPIOISIP °“) piofixojduim 1) S142980081(] 295 DDU yoo) “вә” 0170/1211 `2) — шегү Y 5їллә[ (uoo) 11220102. ‘< шегу (oxe[g Y `$) pyrunauua `t ure y (идә) noudijo °“ шерү Y swf qq) nuvwyə °< шерү Y Swf (erquae7)) sisuaqno °“) шүү y suf (чори) DIJOfISSDALI °<) UB] Y Y smf (uopug) 77720102. °<) шегу Y smf (UOWLIG) 270212703 °<) zougui[( (191) nyong 4 шерү Y smf (пори) DIDNUIND °<) _ :sə19əds ивәфивә ‘5597 pijofipunjoa 9 yyuny уәр1о1иоиләа x axe[g `{ `$ MSDUIPIDI °<) Seg A ND д "“mpues sisuadinbasp “9 :sər9əds ивоәшү YINOG Di]Dui204) “WIS DIJDUYIOS) 8.19.1Q8”) susyso mans 1) n ~ юләлЧЕТ) e1o1qe7) (1әирлее)) ю1уоїхә]4шъ °<) A ‘SSI J) DIPIOISTP “) x. STLISOIST(] 029% DIJDUYIOL) вләәвлецә 21]souser(] UOHNLYSI(] Apnis sių} 0} BUIPIOIIE suonoag (1261) &191qe7) nsuas SUOT}IIG 'uorpas YORA JO sotveds adA] = y ‘sosoyjuoird ш 3uneodde so1njeoj пошшоәип YIM “paptacid әле uonoos YORE 10] SISJORIBYO onsouzerp pue uonnquysip olydeldos93 әчү “(0007 “OYouRs) pjpidaomsuap “) pue (pr6[ Paqe) торур 4) jo uonrppe эці чим "([76[) &1o1q97) Aq рәишцәр suoroes əy) чим poreduioo *A3oJoudaour əy} Jo siseq əy} uo &pnis siy} ur pausi[qeiso. 927041902) шцим ѕиоцоә `| әре, Annals of the 528 Botanical Garden issouri M g 2d4 snd -ded :(y100us) sje} ayeruroey *(ayepnorde Apidniqp) ayenuayye a2epuadde Joyque :(морәќ) INYM (ZL 01 8)9 01 y s1219g :o[ounpad əy} our Surpua] -хә ѕәпе|үАҷа ш#теш әці 1? IPJO *snoaqe[3 Чим әлоп[олш :saporuedopnasd 10 вәшәәвлорпәва ‘squAıo20pnasd әѕо[пләшо[8 ut epnyideo sno1auny OOIX2[N PIAI[Oq S “eunuadly N OSA UWS S (y adá1) Я addy snd оотхәр -ded :S[18] щоошѕ (әјепиәце) a1e[noıde Apidniqe asepuadde sayjue :(0)с)0) o1 zT S1310 :squi&i -ooopnasd aso[nıau0[3 ut 1911920] мә} e[niide7) OOIX2]N ODIXIA ивэддив7 eqn) eqn) тш 7) отирѕотра `£) ponajodty °<) SILIWOINA] "29s DDU 1204) (u0noas Jayjouer 0] рәлош) SUIISILOQLD су, D]D42UI0]4) "1998 DIPDUYIOS) nuospn e) SISUINILO] °4) чәјоцѕ 9 DUDIDIGDS °<) DANIIL 74) әррірла °<) Dsojnosogfionvd ^) торага “4 DYJUDIAO °<) pijofisnigo “5 pupjuow `£) оуоцә2о1лш ғ) sisuanjupui р]хә]лә]и1 °<) 'шиәәлсу X UOSUIGOY nynus °<) e1a1qe7) озирѕотра `) ABI “Y (90) Donajodty “9 әәдәриғія "< `], usndind «) ?IoIqe7^) puspu 4) `55ә^] nmpa4o2 °“), әәЗәрирд `$ ||, suaosa40quD 4) штї]]ХцЧолюлрәр 71998 DIDUYI0y UIRTY Y smf (uogyug) nuospan е) zouguif (917) sisuanz40] ©) шегү Y smf (uonug) Lafpys “4 UIR[Y Y stAJ9(. DUDIDLIDS °<) uey Y smf (uonug) 011221 °<) ZaugUIf (q1) eppavoid “9 £131q*7) ('2u91H) Dsoynosopfionnd ©) ute[y Y smf (uonug) pijofiaund e) pıemoH (91) DYJUDFII0 °<) шү Y smf (uonug) vyofisniqo °4 шегу Y smf (uonug) DUDJUOU *) IN y swf ("qasuz)) рурудәәоләтш =) шү Y Swf ("qosue)) sisuenjupw °<) uey Y тллә[ (uoo) риртѕтош 7) UIRTY Y smf ("әѕце)) юухәмәйи “9 вләўәвлецә onsouzeT uonnquisi] Ápnis sip 0] Зитрлодәр suonoag (1261) &191qe7) nsuas suonoag 'penuguo) `I [qe] 529 Freire et al. Volume 89, Number 4 2002-- Gochnatia Morphology q 2d4 sndded [тв] әјвтшовү *Цзоошѕ “ae pnoide Apidniqe a3epuadde ıaqıue :әвү 10 IHYM “(OZ 01 QT 10 *) *€) € SIALOY SIWI -opnasd 4jeo[ “әѕојпләшо[8 ur e[njtdeo ѕполәшпи euquagiy M 9114, [Pus :әјеүприве[8 *paura4-221gj *Ap[9jeuutd *saAea[ 1eaurq [Ze1g S peg 5 PIAIT[Og ‘nag Jo sapuy pzeig S Aengeıeq ‘zeg © eunuesiy *Aenzni(] ‘Aenseieg “лея [Ze AS [zeig 4 [2219 Y [Zeg S |izeag "AenSe1eq [12219 [12819 7) od) sndded :s[rej оош» *9]RIULOR] °21] pzeig -noide храпе ozepuadde 1aque :(orqdaouiost) pzeig orgdaouipqns se[[o100 :(proosrp) unoprosrp ‘(snow pzeig 4l -рдошоц) snowue301919y e[njrdgo :so[oruedopnosd eunuoziy *Aen2ni(] ‘рар *^enzeieq 3500] 0} әѕо[пләшо[8 ut e[nirdeo вполәшпи ‘siey pzeig я (әтеәзеру) рәшле-с 01 -e “g YIM вәлвә{ :(snoto Aengeıeq ‘zeg S -әоцош) snowRsAjod *snoroaorpou&2 “піце 10 S221], [zeig S Purjuo21y N esy EISV Dsou11n]8 °<) psojonof `9 snioydnjuad “1998 DIIDUYIOS) ри1тјәа °<) pydiowsjod ‘4) Dip navund ©) D]p]n21Q40 °<) pnjoudoooZi]o 4) punssijjow °<) DUDIUDUNDY °<) B191GB") 1Y9DQYISIDY 74) 1L9UPADI °<) ppunquojf 1) 4OJOISIP E OYOURY 4) (8191427) pypydasisuap 19 5597] DIDPLO) 9F DUDIJIYIUDIG `0 DUNUITID °<) штп]]Хцдотюлрәц 152s DIDUYIO) sinqpi2ads 4) D4029p ©) шү Y XooH (uoq `@) nsouimj2 2 шү Y оон (чод O) 80701/0f 9x snioydpjuag "jes D11DuY4204) e1a1qe7) ("uog) nu1ngoa R.191QB (sso) Dp1pa4os * P191QB7) DUDAGSNL PIOIqP^) 10QUUD4 e1o1qe7) vayopnd e191q87) ('sso']) DydiowAjod e1a1qe-) ("ә") рурупәлита * RIOIGe’) (әшүрү) 220101410 ° &1a1qe7) (12upae4)) DIDYdav037]0 E191{E7) (эщ[ер) Dumssioui EI23JQE7) DUDIUDUTIDY 2 2.1918 (128g) тзәира›З РЛӘЛЧЕТ) эри nquoy * ЛӘХРЧ 10]02%р вләлв”) (9q) Pupljeyoun]q * BIOIGe’) (QUEJA) рәгс * еләге”) (E1a1qe7)) puijuazan ° p, UO So y SESS umajspiuimmbopy "2998 DI]DUYIO*) sso] (uoq `(]) 81729072905 2) y, e.191G8") (Zin M) D4022p s149u102Tio'] 90295 DIJDUYIOY) 9 вләәвтецә oysousei(y попи] Apnis stu 0} urp1oooe suonoag (1261) 191087) nsuas suonoag penunuo) "р [QPL 530 Annals of the Missouri Botanical Garden 0 Pro Pde Cge Gfo Gar Hla Aan i Gpo Gro С Figure 1. Leaf shape. —A. Ovate (le fti to rigai к» d rotundifolia (Handro 157, LP), G. ramboi — 50005 S), G. КОО (López & Sagdstegui 3354, LP), G. cordata (Serrano 6, LP), Nouelia insignis (l'Abbé D 2498, US). —B. Obovate (left to right): Wunderlic e crulsiana — et al. 2615, MO), Gochnatia intertexta A-1680, Ag G. pauciflosculosa (Wilson 7428, К), G. seg cae Ekman 4023, S), lanthopappu e s & Cuezzo 2304, LP), Actinoseris radiata (Hatsc hbach 690, LP), Chucoa ilicifolia (López Miran P). — “Elliptic (left to — Gochnatia discoidea (Blanchet 3345, LP), G. magna (Cronquist 11277, N ү), G. dites e е) 32, LP), Volume 89, Number 4 2002 Freire et al. Gochnatia Morphology Few to several capitula (more than 4) arranged in short-pedunculate to glomerulose pseudoco- rymbs are found in Gochnatia arborescens, G. mag- na, G. purpusii, Gochnatia sect. Leucomeris, Hyalis lancifolia, and Wunderlichia (W. bahiensis, W. insig- nis) (Fig. 2C). Numerous capitula arranged in scap- iform, long-pedunculate pseudocorymbs are found in Gochnatia sect. Discoseris (except G. suffrutes- cens), Actinoseris (А. stenophylla), Hyalis argentea, lanthopappus, and Wunderlichia senaeii (Fig. 2D). sochnatia sect. Pentaphorus and Cyclolepis have numerous capitula borne in short-pedunculate to glomerulose clusters, that are in turn arranged in leafy pseudoracemes (Fig. 2E). compact or glomerulose clusters, which lack leaves. at the tip of the branches, are characteristic. of Gochnatia palosanto (Fig. 2F). Loose pseudora- cemes are present only in some species of Pleio- laxis (e.g.. P. angusterugosa, P. gombensis, P. oxy- lepis, P. racemosa) (Fig. : Numerous Pseudoracemes in - capitula — in glomerulose pseudopanicles are present in most species of Moquiniastrum, in G. cordata, G. hypoleuca, and G. smithii (Fig. 2H). Short-pedun- culate (Fig. 21) or loose pseudopanicles (Fig. 2J) are present in species of Gochnatia sect. Moqui- Gochnatia sect. niastrum (G. argentina, G. argyrea, respectively). INVOLUCRE Shape. The involucre shape is either oblong to campanulate or turbinate. Oblong to campanulate of most species o Gochnatia (Fig. 3A), Chucoa (Fig. 3B), Cnicoth- amnus (Fig. 3C), Cyclolepis (Fig. 3D), Hyalis (Fig. involucres are characteristic = 3E), Nouelia (Fig. 3F), Pleiotaxis (Fig. 3G), and Vunderlichia (Fig. 3H). Turbinate involucres, on the other hand, occur in a few Caribbean species of Gochnatia sect. Gochnatia (e.g., G. cubensis, G. intertexta, G. pauciflosculosa), in some species of G. sect. cee died (G. hypoleuca, G. palosanto, G. smithii) ( 3D, G. suffrutescens, and /anthopappus (Fig. 3J). Fig. Actinoseris, Size. The involucre size ranges from 2 to 45 (2—)4—7(—8) mm high, with three main categories: mm, 10—18 mm, and 20-45 mm high. (A) The smallest involucres [(2-)4-7(-8) mm] are present in Gochnatia sect. Pentaphorus, Gochn- alia sect. Moquiniastrum (except G. argyrea with 9-10 mm), G. hypoleuca, G. microcephala, G. pa- losanto, G. smithii, Cyclolepis, and Hyalis. (B) Intermediate involucres (10-18 mm high are present in some species of Gochnatia (e.g.. G. arborescens, G. cordata, G. discoidea, G. ilicifolia. G. magna, G. patazina, G. purpusu, G. recurva, б. rotundifolia, G. vernonioides), М Actinoserts, Chucoa, and /anthopappus. (С) The biggest involucres (20—45 mm) are dis- played by a few Caribbean species of Gochnatia Gochnatia (e.g., G. G. G. ekmanii, G. picardae, G. sagraeana), Cnicothamnus, Nouelia, Pleiotaxis, and Wunderlichia. sect. cowellii, cubensis, Series of phyllaries. The phyllaries are arranged in several imbricate series. The highest number is (6)7 to 10 series, and this is found in a few Carib- bean species of Gochnatia sect. Gochnatia (e.g.. б. cowellii, G. cubensis, G. picardae, G. recurva), б. arborescens, Cnicothamnus (Fig. 3C), Nouelia (Fig. 3F). Pleiotaxis (Fig. 3G), and Wunderlichia (Fig. 3H). Three- to six-seriate involucres occur in most species of Gochnatia (Fig. ЗА), and Actinoserts, Chucoa (Fig. 3B), Cyclolepis (Fig. 3D), Hyalis (Fig. 3E) and /anthopappus (Fig. 3J). Gochnatia sect. Leucomeris and three species of G. sect. Hedrato- phyllum (G. hypoleuca, G. palosanto. G. smithii) show a peculiar feature in their involucres. The ca- pitula are 3- to 4-seriate, but the peduncle bracts extend into the involucre giving a 7- to 12-seriate condition (Fig. 31) Phyllary pubescence and shape. In almost all the taxa studied the phyllaries are dorsally pubes- cent or subglabrous. However, Gochnatia sect. Leu- comeris, G. hypoleuca, G. palosanto, G. smithii, and G. rotundifolia have dorsally glabrous phyllaries with ciliate margins. ost genera have linear to oblong or obovate phyllaries with entire margins. Cnicothamnus has phyllaries with an apical appendage rounded rhombic, with lacerate margins (Fig. 3C), and Wun- derlichia has phyllaries with scarious, colored, and occasionally fimbriate margins (Fig. 3 € Actinoseris polyphylla (Hatschbach 35304, LP), Cnicothamnus lorentzii А Жет " 10780, US). —D. Linear (left to right): Gochnatia foliolosa (Cabrera 34 Hyalis lancifolia (Sc hinini — LP), Actinoseris angustifolia ( (Hatsc m 1 — Б A Pleiotaxis rogersii (Re US (Rojas 7104 ;ochnatia montana (Ekman 18725, ), 5), P. dewevrei — S), Cyclolepis genistoides E Chucoa ilicifolia (López yr 1090, e S). G. cubensis (Ekman 9632, S), G. intertexta (Alain A- Pereira 8609 & Pabst 7984, 1, р), G. rotundifolia | (Handro 15 — Р ). Pletotaxis huillensis (Gossweiler argyrea (Hatschbach 9578, LP). Чуп 1568, P). E margin. —E. Spinose (left to Pon ‚ G. enneantha (Ekman Н-1549 ^ Dentice — to entire (left to right): G. — LP), G. foliolosa (Cabrera 3451, LP). LP ) 532 Annals of the Missouri Botanical Garden С D Figure 2. Types of capitulescence. Solitary or few capitula short-pedunculate. —B. Solitary or few capitula long- 8 YF у I 8 к culate and scapiform. —C. Р — short-pedunculate. —D. Pseudocorymbs — -pedunculate and scapiform. Pseudoracemes glomerulose and leafy. —F. Pseudoracemes apically glomerulose. —G. Loose pseudoracemes. —H. Ps "edd xinicles glomerulose. —I. Ps E anicles short-pedunculate. —J. Loose pseudopanicles. F B I | | I PALEAE (1991) included it in the “Stenopadus group.’ Cabrera (1971) described Gochnatia as typically Wunderlichia is the only genus in the group — epaleate, and only exceptionally with some pa- with receptacular paleae. For this reason, Pruski leae, but mentioned no species. However, we Volume 89, Number 4 2002 Freire et al. Gochnatia Morphology 10 mm. 40 mm. Figure 3. Involucre. A-H. Oblong to campanulate. —A. | Chucoa ilicifolia (López Miranda 1090, LP). —C. € (Tinto 2038, LP). —E. Hyalis lancifolia (Schinini 1 rogersii (Rolyns 1568, US palosanto (Schreiter in 1925, LP — — la found no evidence of paleae in any species of Gochnatia examined. FLORETS Corolla ment. morphology | and sex arrange- Corollas in Gochnatia and allies may be isomorphic, subdimorphic, or clearly dimorphic. 10 mm. inicothamnus lorentzii (Padaci 84, LP) 6098, LP). —F. Nouelia insignis (Maire 2516, NY). —G. Pleiotaxis Wunderlichia crulsiana (Ratter et al. 2615, MO). ‚ —J. lanthopappus corymbosus (Palacios & Cuezzo 2304, LP) 10 mm. Gochnatia floribunda (Hatschbach 43151, NY). —B. . —D. Cyclolepis genistoides І, J. Turbinate. —I. Gochnatia LP). All florets of the capitula may be hermaphroditic or female (homogamous capitula), or the capitula may have central hermaphroditic florets and mar- ginal female florets (heterogamous capitula). These features are combined as follows: (A) Corollas isomorphic. Plants with discoid and homogamous capitula made up of hermaphroditic florets with isomorphic tubular corollas are char- 533 534 Annals of the Missouri Botanical Garden acteristic of most species of Gochnatia (Fig. 4A, B), Chucoa (Fig. АС), Pleiotaxis (Fig. 4D), and Wun- derlichia (Fig. 4E). Only б. argyrea of section Mo- quiniastrum has isomorphic tubulose corollas and heterogamous capitula (Sancho, 2000). Very occa- sionally marginal female florets (with staminodes) are found in Chucoa (Cabrera, 1955, and our own observations). (B) Corollas subdimorphic. Such corollas are present only in Gochnatia sect. Moquiniastrum (ex- cept G. argyrea and G. gardneri) (Fig. 4F) and Cy- clolepis (Fig. 4G). Gochnatia sect. Moquiniastrum includes monoecious, polygamous, and gynodioe- cious plants (Sancho, 2000). Polygamous and gyn- odioecious species have disciform capitula made up of subdimorphic corollas, i.e., central hermaph- roditic florets with tubular actinomorphic corollas and coiled lobes (Fig. 4F,). and outer, female florets with tubular subzygomorphic corollas and erect or coiled lobes of different lengths (Fig. 4F,). In the gynodioecious genus Cyclolepis the female florets have tubular-filiform corollas (Fig. 4G,), and the hermaphroditic florets have tubular corollas (Fig. 4G,). (C) Corollas dimorphic. Plants with radiate, ho- mogamous capitula comprising dimorphic, her- maphroditic florets, with outer bilabiate or subbi- labiate and central tubular corollas, are present in some species of Actinoseris (e.g., A. polyphylla, A. revoluta, A. stenophylla) (Fig. 4H), Cnicothamnus (Fig. 41), Hyalis (Fig. 4J). Plants with heterogamous capitula and the type of — — and Nouelia (Fig. 41, dimorphic corollas previously mentioned are found in lanthopappus (Fig. АК) and some species of Ac- g., А. diata); the outer florets are female and the central tinoseris (e. arenaria, A. hatschbachii, A. ra- florets hermaphroditic. These truly dimorphic co- rollas are absent in Gochnatia. It may be noted that tubular, deeply lobed co- rollas with coiled lobes (reaching up to 2/3 of the corolla) are present in most species of Gochnatia (Fig. 4A), and in Actinoseris, Chucoa (Fig. 4C), Cni- cothamnus, Cyclolepis, Hyalis, lanthopappus, Noue- lia, and Pleiotaxis (Fig. 4D). Tubular, very deeply lobed corollas with straight lobes (more than % the length of the corolla, almost reaching the base) are present in most Caribbean species of Gochnatia sect. Gochnatia (e.g., б. calcicola, С. cubensis, С. gomezii, G. oligantha) (Fig. АВ). Bilabiate corollas have an external 3-dentate lip and an internal 2- Actinoseris revo- Nouelia) (Fig. 4J—L). and subbilabiate corollas have an external 3-den- 4, Acti- cleft lip (3+2 arrangement, e.g., luta, Hyalis, lanthopappus, tate and one entire, internal lip (3+1, e.g noseris hatschbachit, (Fig. 4H, А. polyphylla, Cnicothamnus) Number of florets per capitulum. The number of florets per capitulum ranges from 4 to 300. A species of Gochnatia (G. crassifolia, G. magna, G. recurva, G. rotundifolia, G. sagraeana) and most species of Actinoseris, Cnicothamnus, and Wunder- lichia have 50 to 300 florets per capitulum. Some species of Gochnatia (e.g., G. amplexifolia, G. bo- liviana, G. cordata, G. cowellii, G. patazina, G. ver- nonioides), some species of Actinoseris (e.g., A. hatschbachii, A. revoluta), lanthopappus, and Noue- lia have capitula with 30 to 50 florets. Some spe- cies of Gochnatia (e.g., G. arequipensis, G. carden- ilicifolia, G. purpusii), G. sect. Moquiniastrum, Chucoa, and Cyclolepis have 7 7 to asi, б. 30 florets per capitulum, whereas some species of Gochnatia (e.g., G. hypoleuca, G. smithii), G. sects. Leucomeris and Pentaphorus, a few Caribbean spe- cies of G. sect. Gochnatia (e.g., G. calcicola, G. cub- ensis, G. gomezii, G. oligantha), and Hyalis have capitula with 4 to 6 florets. Corolla color. The corollas in Gochnatia are predominantly yellow or cream. In some species (e.g.. G. rotundifolia, G. decora, G. spectabilis, G. foliolosa) they are white and/or lilac. In others (e.g., G. cowellii, G. ilicifolia, G. intertexta, С. mantuen- sis, G. sagraeana) they are orange. The corollas are white, pink, or purple in Actinoseris; in Cnicoth- amnus they are orange; in Chucoa and Cyclolepis, yellow; in Hyalis lilac; taxis deep red, pink, white, or cream; in Wunder- in Nouelia white; in Pleio- lichia white or yellow. lanthopappus shows white marginal corollas and purple disc corollas. ANTHERS As well as most species of the tribe Mutisieae, all the taxa in this study have an apical anther appendage and basal tails. The apical appendage can be viewed as an adaptation to protect the pollen in the anther tube from moisture and insect pred- ators until the stigma and style push upward through it for pollen presentation (Stuessy et al., 1996). With the exception of Chucoa, with acute anther appendages (Fig. 5A), anther appendages in Gochn- atia and related taxa have usually been described as apiculate. Pleiotaxis differs in its thickened and bulbous apical appendage (Fig. 5B). Analysis of this feature in all the genera under study revealed further variation in its shape that is particularly useful in grouping species within Gochnatia. Vhe appendage can be short (anthers Volume 89, Number 4 Freire et al. 535 2 Gochnatia Morphology Figure 4. A-E. Isomorphic corolla. —A. Gochnatia — (C — 11277, NY). —B. G. e — ulosa (Eggers 3866, К). —C. Chucoa ilicifolia — Miranda 1090, LP). —D. Pletotaxis diu evrei " 21 —E. Wun- derlichia mirabilis — еШ & Stutts 999, NY). F, e Subdimorphic corolla. —F. Gochnatia оу А (Hashimoto 624, LP): F,, disc corolla, F,, marginal — —G. Cyclolepis genistoides (Cabrera 3782, LP): G,, marginal corolla. G,, dise corolla. H-L. Dimorphic corolla. —H. ринин polyphylla (Hatschbach 35304, LP): H,, disc corolla, H,, marginal corolla. —I. Сп e lorentzii (Cabrera et al. 22576, LP): I, marginal corolla, L,, disc corolla. —J. Hyalis lancifolia (Schinini 16098, LP): J,, marginal corolla. J;. disc corolla. К. —— corymbosus — Cuezzo 2304, LP): K,, marginal m K,, disc corolla. —L. Nouelia insignis (Maire 2516, NY): L,. marginal coro L,, disc corolla. 536 Annals of the Missouri-Botanical-Garden - Gpar Gpat Wi A Sect. — Clo | ire Cge AN Gde Gsp Gro E F is ou} PA AAA NA Sect. Discoseris Sect. TT Nin Waz Gar ES Sgl Gco Ja Sect. Pentaphorus 0,3 mm. K L Figure 5. Apical anther appendage. —A. Not apiculate: — ilicifolia (López Miranda 1090, LP). —B. Apiculate bulbous: Pleiotaxis eximia (Rolyns 1836, US). C-F. Caudate anther appendage. —C. Gochnatia sect. Сос "o — Cabre ra left to right): Goc — fp nasii (Cordo & Ferrer 88-B-17, SI), G. cowellii (Britton & Cowell 1 urviflora (Jeréz et al. 49120, LP), G. ilicifolia (Small & — 8526, К), С. montana (Ekman 18725, * nant (Shafer 2938, NY), С. e s & Sagástegui 3409, LP), G. recurva (León 20946, LP). —D. Actinoseris T Volume 89, Number 4 2002 Freire et al. |... . . Gochnatia Morphology apiculate) to very long (anthers apically caudate). In addition, it can abruptly terminate in a sharp point, or gradually taper above into an attenuate point. The following three combinations were ob- served: (1) Apically caudate (Fig. 5C-E): in most spe- cies of Gochnatia sect. Gochnatia (except G. boliviana, G. buchii, G. microcephala, G. tortuensis, and б. vernonioides with abruptly apiculate apices) (Fig. 5C), Actinoseris (Fig. 5D), and Cnicothamnus (Fig. 5E). (2) Attenuate apiculate (Fig. 5G): in Gochnatia secl. Leucomeris, hypoleuca, G. palosanto, G. magna, and G. Е. (3) Abruptly apiculate (Fig. 5J—L): in Gochnatia sect. Discoseris, sect. Moquiniastrum (short append- age in G. argentina, G. densicephala, and G. flori- bunda), and sect. Pentaphorus, Gochnatia arbores- сепз, ©. cordata, С. purpusit (Fig. 5J), Nouelia oe 5K), and Wunderlichia (very short) (Fig. 5L Although useful in distinguishing taxa, overlap exists between these broad categories such м some as between apically caudate-attenuate (e.g.. lantho- pappus, Fig. 5F), attenuate-abruptly apiculate (e.g.. Cyclolepis, Fig. 5H; Hyalis, Fig. 51), and abruptly apiculate-caudate (e.g., G. velutina, Fig. 5J) Anther tails are free and can be smooth or la- They are smooth in the Caribbean species — ciniate. of Gochnatia sect. Gochnatia (except G. attenuata, G. ilicifolia, and G. microcephala), some species of Gochnatia sect. Moquiniastrum (e.g., G. barrosit, С. densicephala, G. floribunda, С. paniculata), some species of Gochnatia sect. Hedraiophyllum (e.g.. G. arborescens, G. magna, G. purpusii), Chucoa, Cni- cothamnus, and Wunderlichia. The tails are lacini- ate, at least in one side, in most species of Gochn- alia (e.g., G. sect. Discoseris, G. sect. Leucomeris, the South American species of G. sect. Gochnatia, and G. glutinosa), and in Actinoseris, Cyclolepis, Hyalis, lanthopappus, Nouelia, and Pleiotaxis. STYLE Most taxa have smooth styles except for Wunder- lichia (Fig. 6A) and Chucoa (Fig. 6B) with dorsally papillose styles, i.e., with the two branches (round- ed and acute at the apex, respectively) covered by short sweeping hairs, and Pleiotaxis (Fig. 6C) with a subapical crown of short hairs. All species of Gochnatia (Fig. 6D, E), Actinoseris (Fig. OF). Cni- cothamnus (Fig. 6G), Cyclolepis (Fig. 6H), Hyalis (Fig. 61), /anthopappus (Fig. 6J), and Nouelia (Fig. OK) have smooth, apically rounded styles. Most have the inner surface of the branches covered by stigmatic papillae prolonged into the outer surface of the style constituting a ridge of cells, which is less evident in Cnicothamnus. TRICHOMES Leaf pubescence. Excluding some species such as Gochnatia rotundifolia, Actinoseris hatschbachii, А. stenophylla, and Hyalis lancifolia, which have glabrous leaves, at least when mature, there are five different types of trichomes in Gochnatia and its relatives. (1) Oblique septate flagellate hairs: one or two foot cells, one- or more-celled stalks or stipes, and unicellular, very long, flagellate, tubular heads (Fig. 7A). This trichome type is present in most species of Gochnatia, Chucoa, Cnicothamnus, Pleiotaxis, and Wunderlichia. (2) Two-armed hairs (or T-shaped, malpighia- ceous, anvil, dolabriform hairs): one or two foot uniseriate, and unicellular heads. cells, stalks, which constitutes the head, shape of a hammer and later becomes T-shaped by further outgrowth of the two ends (Fig. 7B, С). This trichome type is present in Gochnatia sect. Moqui- niastrum (Fig. 7B) and G. cordata, Cyclolepis (Fig. generally two- to more-celled The apical cell, initially assumes the tE — оо 690, LP). — Cnicothamnus lorentzii (Ruiz Leal — Attenuate-apiculate anther appendage. .L Maung Mya 2 (Palacios & Cuezzo 2304, LP). —( Cabrera (from left i to pied Gochnatia n ora т — 3215, LP), G. palosanto (Schreiter in (Joly . LP). —H. ptly apiculate anthe : арре amplexifolia (Hatschbach 35312, LP), — — endage. J,, Cyclolepis genistoides — 3172 Fa m LP. — alia sect. G. discoidea. — 3345, LP). J Go Cabrera (from ie to right): ra — a (Hatschbach 9578, (Rojas, herb. Hassler 9752, maniana paniculata (Gardner 4810, US), ie polymorpha (Glaziou in 1876, 64, ee ;ochnatia sect. Pentaphorus sensu Cabrera (from left to right): С vernonioides (López Р). G. purpusit (Pur — NY). (Harley et al. 25209, MO) (very shortly apiculate). ‚б. oligocephala (Blanchet 328 . LP). J, (from left to right): С. ‚ Nouelia insignis (Maire 2 ‘wo — Е. ake ен corymbosus ;ochnatia sect. comeris sensu P), G. — ies 32, LP). G. G. hypoleuca LP). G. — (Fernández 3666, NY), С. ат im Hyalis lancifolia (Cabrera 4083, — Discoseris sensu Cabrera (from ve jm EL chnatia sect. Р), G. les Ge — 5771 et al., 8, US), б. orbiculata (Brade 5523, US), ( LP). G. velutina (Smith & Klein 14885, , 6. >. fe оса arborescens (Johnston 4023, LP), G cordata 516, NY). —L. Wunderlichia azulensis 538 Annals of the Missouri Botanical Garden A,B,D-F,H,1,K 0,3 mm 0,3 mm. Figure 6. Style branches. A-( —B. Chucoa ilicifolia (López Miranda 1090, s. —D. Gochnatia discoidea nchet Be LP (Hatschbach 28756, re —G. Cr Kiesling 114, LP). —l. Hyalis argentea (Ruiz Leal 3701, LP). —K. Nouelia — (Maire 2516, NY). ). —К. 7C), and Hyalis argentea. In Nouelia one end of the apical cell is very short. (3) Three- to 5-armed hairs (or 3- to 5-branched, stellate hairs sensu Cabrera, 1971): similar to the 2-armed hairs, but the apical cell has З to branches (Fig. 7D). The apical cell does not divide and thus the head remains one-celled. This tvpe is л found only in Gochnatia barrosii and G. rusbyana of section Moquiniastrum. (4) Multistoried T-shaped hairs: similar to the 2-armed hairs, but further periclinal divisions take place in the apical cell. The head is thus comprised of 3 or 4 one-celled layers, all oriented transver- sally and parallel, but of different lengths (Fig. 7E). This type of trichome, not very common in Aster- aceae, has been reported in the tribe Senecioneae (Robinson, 1989). Janthopappus is the only genus with this type of trichome. (5) Biseriate glandular hairs: comprised of 2 rows of cells in the body, with two to many cells in each row, enclosed by a persistent or collapsed cu- ticular vesicle (Fig. TF, €). Glandular hairs are G. cordata аг — zi ¿ni — lorentzii n: Leal 14162, e 2 branches. —4A. Wunderlic — azulensis (Harley et al. 25209, MO Р). —C. Pleiotaxis eximia (Rolyns 1836, US). D-K. Smooth style brane he ctinoseris angustifolia Р). — yck — Ы (Zardini & Р). —J. к р Жозе. (Palacios & Cuezzo 2304, widespread in all the taxa studied, and especially in species of Gochnatia sect. Pentaphorus (i.e., Gochnatia foliolosa, С. glutinosa) where they cover almost the entire surface of the leaf, with the fla- gellate hairs restricted to the margins. Achenial pubescence. Glabrous or slightly pa- pillose achenes occur in Chucoa, in a few species of Pleiotaxis (e.g., P. decipiens, P. huillensis, P. li- nearifolia, P. welwitschii), and Wunderlichia (e.g.. W. crulsiana, W. insignis, W. senaei). The remaining taxa have villose achenes. M W. bahiensis, Three types of achenial trichomes were found: Duplex or twin hairs, leaf-like hairs (2-armed, oblique-septate flagellate), and glandular hairs. (1) Duplex or twin hairs: Twin hairs are the common type in Asteraceae, comprised of two tri- angular or rectangular basal cells, one sometimes reduced, and two cylindrical or elliptical hair cells, equal or subequal in length, generally in contact up to their tips (Hess, 1938; Freire & Ka- tinas, 1995) (Fig. 7H, 1). All the taxa studied, in- Volume 89, Number 4 2002 Freire et al. 539 Gochnatia Morphology E $ B lo = E E wo © alo C =. € B,C,E 0,1 mm E Е 5 | : A D FG 7 005mm | \\ 0,1 mm. Y J K Trichomes. A-G. Leaf hairs. —A. Oblique septate flagellate hair: d tortuensis (Ekman Н-3553. Figure 7 S) E 3, C. 2- ~~ hair. —B. бос — — (Pedersen 8587, LP). —C. Goc chnatia barrosit үсе th 16945, e p n i) alacios & Cuezzo 2304, LP). К, LP). —G. С — жем (Fabris e ы. H-L. Nicora, LP). - to 5-armed hai (Blanchet 3345. lichia azulensts — el tal 25209, MO). H-I. I hair. —K. Dup Cowell 10183, NY). cluding most species of Gochnatia, have achenes with twin hairs, usually very long and filiform. In some cases twin hairs have one hair cell very short (e.g.. Gochnatia purpusii, G. recurva, G. tortuensis, Clk Wunderlichia azulensis, W. mirabilis) (Fig. 71), they are septate (e.g., Gochnatia hatsch- bachii, G. oligocephala, Hyalis, Pletotaxis eximia, Wunderlichia azulensis, W. mirabilis) (Fig. 7J), or have only one hair cell (Wunderlichia azulensis) (Fig. 7K). (2) Leaf-like hairs: achenes similar in morphology to the leaf hairs, as Two-armed hairs in the described above, were found in Gochnatia orbicu- lata. Only a few species of Gochnatia (e.g.. G. cub- ensis, G. magna, G. ramboi) have oblique-septate flagellate hairs (3) Ac heal biseriate glandular hairs: These are Juplex hairs yclolepis genistoides (Correa 3172 & Multistor ed hair: G. Biseriate каги уе н — К. Gochnatia o Achenial hairs. H-k. Duplex c Wunder the hair cells shorter tate dudes ried T-sha ith one of ex hair with only one hair cell. —L. Capitate каб biseriate hair: бос — — (Britton & similar to those on the leaves (Fig. 7F, G) and are very widespread in the group under study, occur- ring with the other types. A modification of the typ- ical biseriate glandular hair with a very enlarged head, i.e., capitate glandular hair (Metcalfe & Chalk, 1950) (Fig. 7L). was found in the Caribbean species of Gochnatia sect. Gochnatia. PAPPUS In all taxa analyzed the pappus is comprised of rigid and scabrid bristles. However, there is inter- esting variation in bristle length and width, modi- fications of the lateral cells at the apex of the bris- tles, and in the number of series of the bristles. Five pappus types were found. Type A. All the bristles have the same length 540 Annals of the Missouri Botanical Garden 0,1 mm Figure 8. ae General aspect (on the left) and detail of bristle apex (on the right). —A. Type A: Gochnatia recurva E 20946, LP). —B. Type B: Gochnatia decora (Maung Mya 5309, LP). —C. Type C: Ge rchnatia cordata (Serrano 6, LP). е D: Goc — и (Fabris 1343, LP). —E. Type E: Wunderlichia mirabilis (Irwin et al. 9913, NY) and width (Fig. 8A). This is present in the Ca- Туре B. All the bristles have the same width, ribbean species of Gochnatia sect. Gochnatia but about half are shorter than the others (Fig. 8B). (except G. cubensis, G. oligantha, G. sagraeana, This type is present in Gochnatia sect. Leucomeris, G. tortuensis), G. sect. Discoseris, and Actinoseris. the South American species of section Gochnatia, Volume 89, Number 4 2002 Freire et al. Gochnatia Morphology most species of section Hedraiophyllum, and Chu- coa, Hyalis, and Nouelia. Туре C. АП the bristles have the same width, half of them are shorter, and the longest are plu- mose at the apex (Fig. 8C). This pappus is present in Gochnatia cordata, G. sect. Moquiniastrum, and lanthopappus. Type D. tively wide and flat (somewhat paleaceous) and the Half of the bristles are long and rela- other half are short and thin (Fig. 8D). This type is present in Gochnatia sect. Pentaphorus and some species of Pleiotaxis (e.g., P. dewevrei, P. eximia, Р. pule — all with the long bristles acute at the apex, and in Cnicothamnus, which has the long bristles — at the apex. Type E. Half of the bristles are long and rela- tively wide and flat (somewhat paleaceous) and half but the — This kind « are short and thin; all are scabrid, are plumose at the apex (Fig. 8E). pappus is found in Cyclolepis, Pleiotaxis (e.g.. P. huillersii, P. rogersii), and Wun- derlichia. " The type A pappus. i.e. the same length, is 1-seriate with ЗО to 40 bristles. Types B-E, i are 2- to 3-seriate with more than 50 bristles. e.. with the bristles of different length. DISCUSSION CHARACTER VALUE Some characters such as habit, most leaf fea- tures, and pappus vary within Gochnatia. Other characters, although constant and common to all species of Gochnatia, are not exclusive to it, such as multiseriate involucres, tailed anthers with apic- ulate anther appendages, and smooth style branch- es. Table 2 shows this and the main morphological characters that allow the genera associated with Gochnatia to be distinguished. Some of the char- acters studied merit a brief discussion. The involucre of Gochnatia, resembling a spruce cone, suggested by Hansen (1991) as a distinguish- ing character for the group of Gochnatia and associated genera, also occurs in other Mutisieae such as Aphyllocladus, Dinoseris, Hyaloseris. and Stenopadus, which are not very closely related to Gochnatia. We do not therefore consider that this feature has diagnostic value. Corolla morphology has been used as a key di- agnostic character in Mutisieae. The tubular acti- nomorphic (Fig. 4A—E), tubular-filiform (Fig. 4G,). tubular subzygomorphic (Fig. 4F,), subbilabiate (Fig. 4H,. Т). and bilabiate corollas (Fig. 4J,. K,. some species d with all the bristles of L,) present in the taxa studied show the great var- Cabrera (1961, 1977) characterized Gochnatiinae and Barnadesi- iation of this character in Mutisieae. inae (the latter subtribe currently constitutes the subfamily Barnadesioideae; Bremer & Jansen, 1992) in the first comprehensive key to subtribes of Mutisieae by having more or less actinomorphic. deeply 5-lobed disc corollas, with equal or unequal segments, but never truly bilabiate, and bilabiate or subligulate ray corollas. The Mutisiinae and Nas- sauviinae, on the other hand, have clearly bilabiate (the disc florets exceptionally actinomorphic) or lig- ulate disc and ray florets. According to this key all the taxa studied, although variable in their corollas, would correspond to the subtribe Gochnatiinae (sensu Cabrera, 1977). Other authors (e.g.. Robin- son, 1991: Bremer, 1994) regard the distinction be- tween Gochnatiinae and Mutisiinae, based only on the actinomorphic versus bilabiate disc florets, to be artificial and recognize only Mutisiinae sensu lato. Apiculate anther appendages have been strongly considered to be an — character within Mu- tisieae (Cabrera, 1977; Hansen, 1991; Karis et al., 992: Bremer, 1994), The shape of the anther ap- and abruptly caudate, attenuate, 5C-L), Gochnatia and also among genera. Indeed, Gochn- pendage, i.e., apiculate (Fig. led to distinction. within atia can be associated with Actinoseris. Cnicoth- amnus, Cyclolepis, Hyalis, lanthopappus, Nouelia, and Wunderlichia by the common possession of flat (not bulbous), — anther appendages. Chucoa and Pleiotaxis (Fig. 5А, B), based on their acute and bulbous anther i f respectively, are very different. The current tribal position of Pleio- taxis is controversial. Some authors regard Pleio- taxis as forming part of a mutisiean. “Dicoma- group” (Bremer, 1994; Ortiz, 2000; Ortiz & Coutinho, 2001). According to Hansen (1991), this group of genera (including Pleiotaxis) should be ex- cluded from Mutisieae by features of style branch- es, anthers, and pollen type. Smooth styles are revealed as another informa- tive feature in this group of taxa, although the value Some authors of this character is controversial. (Bremer, 1987) consider the glabrous styles to be plesiomorphic within Asteraceae (although the Lob- eliaceae, characterized by hairy style branches, were used as an outgroup), while others (Stuessy et al.. 1996) postulate the smooth condition to be de- rived with Calyceraceae as the outgroup. At present Calyceraceae are the sister 2001; Ur- tubey & Stuessy, 2001) is widely accepted. Since the hypothesis that group of Asteraceae (e.g., Albach et al., the style branches in Calyceraceae are papillose Annals of the Missouri Botanical Garden qo 542 J 9d Я ‘q sedA] g ә44], 7) ad Ay g eday, 4 әчАү, q edA] g oda], y әд], ‘q “y sadA] snddeg ÁI A[[esıop -[eoıdeqns Аеѕлор пиәәѕәфпа пиэәѕәфпа yyoours yjoous q10ouis u1oouis yloours :u22saqnd yyoous yjoouls — soqouedq IJÁIS IPUI oous 3]eruroe| әүрїшәР] 3]eruroe| 3peiuroe| apeiuroe| оош» yjoous әүрїшәР] -вү "[joours sre dou y asoq ayenuaye -mq *pauo ayenuaye ЕТЕ ayenuaye ‘apne? храпе -Jr Apdnaqe -9]ppneo -Apidniqe -Andruqe ayepneo ayepneo ‘храпе osepuodde :9je[noirde :9je[noirde :9je[norde :9]e[norde :3]e¡norde :9]e ¡norde :aye¡norde əmər :9je[norde ap pnaido porde тәщиұ (әицм WERI *ogue “AM a[dınd a[dınd -10 °2P[1[) MOT[SÁ “OPIUM ‘yurd “pal IJYM pue әйцм әр MOT[OA IBUBIO моәл “уша #әйум — швәлә “морәл 10[02 P[[O107) (orqdaour -Ipqns) orgdaouiost orqd.ouiost ornqdaouip orqd.rouirp omjdiounp — orngdaouitpqns ongdaourp orydaowost orqd.rouip nydiowosi se[[0407) 3184229] agepuadde juasqe juasqe juasqe juasqe juasqe juasqe “SIQUIOYA juasqe juasqe juasqe ¡eorde sourep sug (snoui snoul (snow -e30.19194) -e3019194 -e3019194) snowersowoy snoureaoulou snowe3owoy —snourm2019]3Qq snourezourou snouresouloq snowegowoy snoure3ouoy ‘snowesowoy snowpsowoy epnuder, 91eo[ed әүвә[едә әвә[вЧә әвәрә 3jeo[eda ajeo[edo a1e2[eda 3jeo[edo 3jeo[pdo әүвә[еаә әүәруЧәдән pourre -¢ 01 зәшоцә padeys-J -g 'peuue = 1210) aje[ozep aje[[o2eg pouue-z — porrojsnnur рәшие-2 pauLie-z 31R[[93P y a1P[[938y aje|[22eg -Z wm[pSeg ^ 1e[pnpue[32-uoN sqniys snoyAyde (sqnuusqns) saad] sqiou sqnaus sqniysqns sqniysqns *asouids s331} sqnays sqıay *sqnuys ‘s22. I qeu Diyana puny sixDjo19]d р1]әпо\/ snddodoyrun] sig s1d210]2&7) snuuruJ]1021u7) poony’) suasouioy D11muwu2o-) “SNJN ur әле P19Uu93 13410 шол) DIJDUYIOL) ysin3unsip ey} S19]JB.IB Uy”) "sasagjuaaed € 1 - = I әле $31n]£9J UOUIUIODU N '99€jp[oq UL ӘЛЕ SI9]OETIEQO 3AIsn[oxa эц], "P19u238 pare 1dosse pue DIJDUYIOL) jo uoroutrjstp a41 мо[е ey} SIIJOPIBYO [e2130[oydıou шеу Volume 89, Number 4 Freire et al. 543 2002 Gochnatia Morphology (Hansen, 1992, and our own observations) the sect. Gochnatia, sect. Hedraiophyllum, and sect. smooth style branches in Asteraceae become an ad- vanced character for the family. The multistoried T-shaped hair (Fig. 7E). the 2-. and 3- to 5-armed hairs (Fig. 7B-D), the capitate glandular achenial hair (Fig. 7L), and the pappus types established here (Fig. 8) are revealed as new diagnostic characters. In fact: (1) the multistoried T-shaped hair is exclusive to lanthopappus and be- the 2-armed hairs are present in Gochnatia cordata, G. sect. —— Cyclolepis, Hyalis, and Noue- lia; (3) t to 5-armed hairs are exclusive to some species of Gochnatia sect. Moquiniastrum; (4) the capitate glandular achenial hairs, with a very enlarged head, distinguish most of the Caribbean comes another character to distinguish it: (2) the : species of Gochnatia sect. Gochnatia; and (5) pap- pus types A, B, C, D are present in Gochnatia and allow the distinction of sections within the genus (see below), while type E is present only in Cyclo- lepis, Pleiotaxis, and Wunderlichia. CIRCUMSCRIPTION OF GOCHNATIA AND ITS SECTIONS Our morphological analysis confirms that Gochn- atia has no single exclusive feature that distin- guishes it from related taxa, but it can be defined by a set of characters, i.e., homogamous (rarely het- erogamous) capitula, isomorphic (tubular) to sub- dimorphic (tubular and tubular subzygomorphic) corollas, apiculate anther appendages, and smooth style branches (Table 2). After analyzing the morphological features of the sections established by Cabrera (1971) were reviewed and some changes are pro- posed Gochnatia, The distinctiveness of two of the six Cabrera sec- tions, sect. Discoseris and sect. Pentaphorus, was confirmed and additional characters supporting them were found. For instance, Discoseris has pap- pus type A and Pentaphorus type D. Gochnatia cordata, placed by Cabrera in section Hedraiophyllum, has characters that link it to sec- tion Moquiniastrum such as the 2-armed foliar hairs (Fig. 7B), numerous capitula arranged in glomeru- lose pseudopanicles (Fig. 2H), and pappus type C (Fig. i However, it constitutes the type species of section 8C) and must be included in this section. Hedraiophyllum established by Lessing (as a sub- genus) in 1832 and has priority over the name of section Moquiniastrum established by Cabrera in 1971 (with G. polymorpha as the type species). Thus, the name Hedraiophyllum is retained in what was known until now as section Moquiniastrum. Phe remaining three sections sensu Cabrera, Leucomeris, are redefined resulting in the establish- ment of five sections: sect. Gochnatia, sect. Rotundifolia. Cabrera (1971) established two groups in his key to section Gochnatia, the “South American species” and the Anastraphioides, sect. Glomerata, sect. sect. Leucomeris, and “Caribbean species.” Species of this sec- tion have some characters in common, such as sol- itary or 2 to 3 capitula and caudate anther ap- pendages, suggest that it should be separated into two sections: sect. Anas- traphioides and sect. but other characters Gochnatia. The Caribbean species form the new section Anastraphioides clear- ly differentiated by spiny leaf margins (Fig. 1E). corollas very deeply 5-lobed (Fig. 4B), and pappus type A (Fig. 8A). The South American species, on the other hand, have predominantly ovate leaves with entire margins (Fig. 1A, F), corollas deeply 2-lobed, and pappus type B (Fig. 8B), and corre- spond to section Gochnatia. The species Gochnatia rotundifolia, included by Cabrera in section Gochnatia, has characters that show a departure from the other sections of the ge- nus such as glabrous and 3-veined leaves, white corollas, and anther appendages attenuate. It would approach Cabrera’s section Leucomeris by its phyl- laries glabrous with ciliate margins and conspicu- ous parallel veins, but G. rotundifolia lacks the in- volucre extending into the peduncle typical of this section. Consequently, this species is placed in the new, monotypic section Rotundifolia. Gochnatia sect. Hedraiophyllum sensu Cabrera, which was already recognized by Cabrera (1971) to be artificial, is split off. Three species of this sec- tion, G. arborescens, G. magna, and б. purpusti, are placed in the new section Glomerata characterized by the exclusive presence of numerous capitula ar- ranged in glomerulose pseudocorymbs (Fig. 2C). As mentioned above, the type species of section Hed- raiophyllum, G. cordata, was moved to Cabrera’s Gochnatia sect. Moquiniastrum. The remaining spe- cies of section Hedraiophyllum, G. hypoleuca, G. palosanto, and G. smithii, were placed in section Leucomeris largely based on their involucre with glabrous, conspicuously veined phyllaries, ciliolate al the margins, extending into the peduncle (Fig. 31). In summary, we propose the following eight sec- tions for the genus Gochnatia: sect. Anastraphioi- des, atta, sect. Hedraiophyllum, sect. Leucomeris, sect. Discoseris, sect. Glomerata, sect. Gochn- sect. Pentaphorus, and sect. Rotundifolia. This new in- frageneric classification, with descriptions and а key to the sections, is shown in Appendix 2. Annals of the Missouri Botanical Garden 544 DIJDUYIO*) UTYJIM suoroos osje pue “BXB] pa1e]91 ғи шолу DIJDUYIO) JO uonoutjstp əy} зшмо[е *xo[duioo DIJDUYIO*) 34] suluyap SIIJIBIB YO jeorsojoydiow шеу 'uJuojideos ejnydeo t '6 210314 'ejejnounped uoj squiÁsodopnasd =. asojnjawo¡B squiÁjooopnesd | ‘ец рәшше-ому x 'esouids uiDueui je a] pi | vile `2 adA} sndded Т ‘а edA sndded mm “y addy sndded ПП 'ejeuo' деб sauej¡Ayd f] 'eyeues 2}-/ Ajjuquedde “mh | 1 ѕпошебошоц A ѕпошебвојәјәц snddedoujue| sepioiudejseuy SUM JO 38]!| yes sejjoJo9 \ snsoydejuag | 19S BJEIBUO/O “pas (|! $ШӘ$025!0] Pas ецеицооо pas BIOJPUNJOY pas SYM SE||O0J02 ıeıpaH es suewoone? Pas xejduio2 ецеиц2оо Volume 89, Number 4 2002 Freire eta : 545 Gochnatia Morphology THE GOCHNATIA COMPLEX The combination of apiculate anther appendages and smooth style branches is exclusive to Actinos- eris, Cnicothamnus, Cyclolepis, Gochnatia, Hyalis, lanthopappus, and Nouelia within Mutisieae. Con- sequently, these genera are recognized here as the Gochnatia complex (Fig. 9). Chucoa, Pleiotaxis, and Wunderlichia, as already mentioned, differ in their styles and anthers. Within the Gochnatia complex, the genus most morphologically similar to Gochnatia is Cyclolepis. This genus shares trichomes 2-armed, subdimorph- ic corollas, and gynodioecy with section Hedraio- phyllum, which constitutes a pivotal group among the remaining species of Gochnatia and Cyclolepis, providing a “link” between both genera. Cyclolepis, however, can be distinguished from Gochnatia by its leafless, spiny branches, tubular, filiform female florets (Fig. 4G,), and pappus type E (Fig. 8E). The hypothesis that a shrubby habit, inde ho- mogamous, and solitary or few — with yellow, and actinomorphic corollas represent primitive con- ditions in Asteraceae has been widely discussed and recognized (Maguire & Wurdack, 1957; Ca- rlquist, 1976; Bremer, 1987, 1994; Pruski, 1991: Stuessy et al., 1996). From this point of view, most species of Gochnatia have a set of plesiomorphic characters when compared with the remaining gen- era of the complex. This hypothesis is consistent with Cabrera's (1977) idea that Gochnatia is the basal genus in a complex from which the other gen- era in the subtribe are derived. Literature Cited Albach, D. C., P. S. Soltis, D. E. Soltis & R. G. Olmstead. 2001. Phylogenetic analysis of Asterids based on se- к of four genes. Ann. Missouri Bot. Gard. 88: 3-212. © G. M. & В. Maguire. 1973. A review of the genus Wunderlichia (Mutisieae, Compositae). Rev. Brasil. Biol. 33(3): 379-406. Beltrán, п. — R. Ferreyra. 2001. Una especie nueva de e para ey y Boli id Gochnatia lanceolata. C — Newsl. : Bremer, K. 1987. — Pn of the Astera- ceae. Cladistics 3: 210-253. 994. Asteraceae. Cladistics and Classification. Timber Press, Portland, Oregon. & R. K. Jansen. 1992. A new шуу" of As- leraceae. Ann. Missouri Bot. Gard. Cabrera, А. L. 1950. Observaciones ma p géneros Gochnatia y Moquinia. Notas Mus. La Plata, Bot. 15(74): 37-48. . 1951. Notas sobre — de América Aus- tral. Darwiniana 9(3—4): : 386. 1955. Un nuevo э 66 nero de Mutisieas del Perú. — Bot. 6: 40-44. PL argentinas: Clave para la de- Bol. Soc. e de los géneros. Revista , Bernardino Rivadavia Inst. Nac. Bot. 2: 291-362. . 1970, Actinoseris, nuevo género de compuestas. Bol. Boc. Argent. Bot. 13: 45—52. 1971. Revisión del género Gochnatia E neon tae). Revista Mus. La Plata. Bot. 12(66): 1-16 974. Tres Compositae nuevas de Minas Gerais (Brasil) Mus. Bot. Munic. 15: 1—4. — . Mutisieae. Syateriatis review. Pp. 1039— 1066 i in v b Heywood, J. B. Harborne & B. L. Turner (editors), The Biology a SUME of the Compositae Academic Press, Loi Candolle, A. P. de. 1838. —— 7(1): 24. Par Carlquist, S. 1976. Tribal Ln and м УВ ny of the Asteraceae. Aliso 8: 465-49 Font — P. 1979. Diccionario de —* Ed. Labor, S. A., Barcelona, España. Franc he M. A. 1888. Bot. 5: 65-67. Freire, s E. & L. Katinas. 1995. Morphology and ontog- eny of the cypsela hairs of Nassauviinae (Asteraceae, Mutisieae). Pp. 107—143 in D. J. N. Hind, C. Jeffrey & G. Pope (editors), Advances in | Compositae Systematics. (ew. Mus. Argent. Ci. Invest. Ci. Nat. Les Mutisiacées du Yun-nan. J. Phe ореви oh in Compositae tribe Vui a Bot. 109: )92. Studies in the € aceae with a discus- sion a its relationship to C ae Nordic J. Bot. 12: 63-75. Harris, J. ( M. W. Harris. 1994. Plant Identification Terminology: An i. Glossary. Spring wake Pub- lishing, Spring Lake, Utah. Hess, R. 1938. Vergleic — Untersuchungen über die ТРЕ ИЕА der Kompositen. Bot. Jahrb. Syst. 68: 135—496. Jansen, K. K.-J. Kim. 1996. Implications of chlo- roplast DNA E for the classific: ve and е of the Asteraceae. Pp. 317-339 in D. J. N. Hind & H J. Beentje (editors), Compositae: — лас. Proceed- ings of the DE e Compositae Conference, Kew 1994, A . The Royal Botanic Gardens, Kew. d Palmer. 1987. A chloroplast DNA inver- sion bs an ancient pound split in En я . 84: ег — (Asteraceae). Proc. Natl. Acad. 5818-5822 Jeffrey, c — Notes on с, a The сне in East M Africa. Kew Bull. 1 Jervis, R. N. 1954. A Summary of E РЯ including а Revision of the West Indian — which Comprise the Sec tion е du Ph.D. Disserta- tion, e of Michigan, Ann Arbo Karis, O.. M. Kallersjö & K. Bremer. 1992. Phylogenetic — sis Е the Cichorioideae (Asteraceae), with empha- s on the Mutisieae. Ann. Missouri Bot. Gard. 79: 421— ===, Р, e & M. Killersjó. 2001. New evidence for the systematic position of Gundelia L. with notes on delimitation of Arctoteae (Asteraceae). Taxon 50: 105 114 Kunth, C. 5. 1818. In F. W. ips A. J. A. Bonpland & C. S. Kunth, Nov. gen. sp. 4 Реа С. Е. 1830. Synanthereis таа Regii Berli- nensis, Dissertatio Tertia. Linnaea 5: 237-263. . I6 e лана Syn. Gen. Compos.: 102-10 «de Mais, B. Wurdack. 1957. The botany of 546 Annals of the Missouri Botanical Garden Guayana — Part Il. Mem. New York Bot. Gard. 93): 235-3 Metcalfe, C. R. & L. Chalk. 1950. Anatomy of the Dicot- — Leaves, Stem, and Wood in Relation to Tax- onomy, with Notes on Economic Uses, Vol. П. Claren- don Press, Oxford. Moreno, N. P. 1984. Glosario Botánico "yen pañía xw Continental, S. A. de C. V. (C.E.C.S Xalapa, Mexic Ortiz, S. 2000. A phylogenetic analysis of Dicoma d and related genera (A ieae) based on morphological and anatomic characters. Ann. Missouri Bot. Gard. 87: 459-481. & A. P. Coutinho. 2001. Achyrothalamus reduced to гуен (Asteraceae, Mutisieae). Taxon 50: 389—40: Com- . A — Р steraceae: Cichorioideae: Mut Payne, W. W. 1978. A pon of plant hair terminology. Brittonia 30: 239-255 Pruski, J. К. 1991. — of the Guayana High- * I—V. The Mutisieae of the Lost World of Brazil, C olombia and Guyana. Bol. Mus. Paraense, ser. Bot. 7 335-392. 1962. Studies in the trichomes of some Com- India 4: Ramayya, N. e — M General structure. Bull. Bot. Surv. -18 — A 1989. A revision of the genus Dresslero- thamnus (Asteraceae: Senecioneae). Syst. Bot. 14: 380- 1991. Two new species of Stifftia with notes on lationships of genus (Asteraceae: Mutisieae). Sys ‚ 16: 685-69 Roque, N. 1 (Less.) Cabrera (C — eS Kew Bull. 52: 1 . Five new species of Richterago (Compos- itae, ы A genus endemic to Brazil. Novon 11: 341-34 a re 997. ie reassessment of Actinoseris polymorpha Mutisieae) и а пем 97-204. = & . М. Hind. 2001. /anthopappus, a new ge- nus of us tibe Mutisieae (Compositae). Novon 11: 97— 101. —— — & J. R. Pirani. 1997. Flora da Serra do — Minas Gerais: Co WO NT SET ae e Mutis 35. Bol. Bot., Univ. Sáo dapes x 151—1& & . 2001. Re — ment of the name Ri- chterago — жер recircumscription of the genus to include species formerly treated as Actinoseris (Endl.) Cabrera (Compositae, Mutisieae). 1155- 160. Sancho, G. 1997. Revisión Sistemática, Análisis Cladís- tico y Biogeográfic o de la Se Taxon 50: ción И del a (Asteraceae, Mutisieae). Ph.D. Dis- Universidad Nacional de la Plata, La Plata, Género Gochnati sertation, Argentina. ————, 2000. Revisión y filogenia de la sección Moqui- niastrum Cabrera del género Gochnatia Kunth (Aster- ‚ Mutisieae). Колин 54(5): 61-122. Stuessy, T. F., T. Sang & M. L. 1990. and locura of the Barnadesioideae, with impli- cations for early о м Compositae. Pp. 463—190 in D. J. N. Hind & H. J. Bee ntje (editors). e Systematics. Proceedings p: the e Compos- itae Conference, Kew 1994, Vol. 1. The Royal Botanic w Devore. Phylogeny — T €. Urtubey, E. & T. F. Stuessy. 2001. New hypotheses of \ I phylogenetic relationships in Barnadesioideae (Astera- 3 ceae) based on morphology. Taxon 50: 1043-1063. Zardini, К. М. 1975. A new species of | Actinoseris Com- positae). Bol. мы Bot. Munic. 23: 1— — APPENDIX 1. Index to specimens examined, with vouchers. Note: During preparation of this work, the new species Gochn- atia. jd eolata Beltrán & Ferreyra was published (Beltrán erreyra (C 'arolina 01. журе US; 2 "zj )1). Spec — rn pe п! species о & Peca 336, US) have \ асще enis appendages, so this species must be excluded from Gochnatia and does not belong to the Gochnatia complex Actinoseris А. ек BRAZIL. Minas Gerais: Jaboticatuba, schbach 28756 (LP); Serra do Cipó, Hatschbach 986 5 (LP). A. утара: BRAZIL. ade 7 (LP). А. — "BRAZIL . Minas Gerais: Santa Ana do Ria- cho, Hatsc on h 35304 (LP). A. radiata: a а a Paraná: Campo Largo, Hatschbach 690 Minas Gerais: Serra de Cipó, А. илину: Minas Gerais: Santa Ana do iacho, de Tm т; 35388 & Koczicki (LP). Cyclolepis C. La PARAGUAY. Chaco dea Rojas 7104 (LP). GENTINA. — bas | de las Con- chas, us 3782 (LP). — Gral. Cone- ssa, Correa 3172 & — (LP). Salta: La Cande- laria, Schreiter 6630 (LP). San Juan: La Laja, Tinto (LP). Neuquén: Plotier, Banda, Zardini & Kiesling 114 (LP). Chucoa C. ilic jm PERU. La T dia Santiago de Chuco, Ló- z Miranda 1090 (LP). Cnicothamnus C. azafran: ARGENTINA. Salta: Serranías del Crestón, Bortagaray 22 d Ab P). Jujuy: Santa Bárbara, Cabrera et al. 26287 C. lorentzii: ARG AT г INA. Salta: Caldera, Cabrera et al. Capital, Bonavía 61 (LP); camino de cornisa Salta- Jujuy, Zardini 1292 (LP). Gochnatia G. amplexifolia: BRAZIL. Minas Gerais: Riacho ao 35312 (LP). G. arborescens: MEXICO. Baja California: Isla Cerralvo, — 4023 (L P: entre Santonio y Puerto de Ba- » los Muertos, Wiggins 5632 (US); along the ^ac ей Coast, 14 mi. S of Pescadero, Spjut & Edson )85 (US). Santa Ana do `P ж ( С. gio PERU. Arequipa: Monte Chiwata, Eyer- dam & Beetle 22120 (LP). G. Кошто Tucumán: Trancas, Meyer — (LP). Entre Rios: Paracao, Paraná, Schulz 3 (LP): Leales, vanduti 703 (LP). G. argyrea: BRAZIL. Paraná: Vila Velha, Dusén (LP), Dusén 9115 (G), Hatschbach 9578 (LP) G. attenuata: CUBA. Oriente: Sierra de Nipe, Ekman 19174 (1 4035 Volume 89, Number 4 2002 Freire et al. Gochnatia Morphology ~ м 2 ~ a c a EN ` c a см $ ea, M м ч ~ ` "s ч ~ i c a м ` ы 0 BOLIVIA. Santa C М бшем: . calcio: М — nasil: € — re crassifolia: CUBA. Oriente: >. cubensis . CUTU — BOLIVIA. Tarija: . de nsicephala: BRAZIL. d divider r dise — ВК, ,. ше» . in . foliolosa: 345 М gardneri: BRAZIL. r. — LP p. gouni К i BA. .. hatsc nid hit: n» haumaniana: . — MEXIC( Hatschbach Macedo 1138 Paraná: Cianorte, serais: [tuitaba, or Tros E i: BRAZIL. 5 (LP). Min Minas ia a PARAGUAY. Amambay: Parque Nacional ¿erro Cord, Sancho 8 (LP). blanchetiana: BRAZIL. Ceará: Serra do лк ;ardner 1735 (K). Goiás: Glaziou 21663 (G ruz: Florida, — & Gutiérrez 33804 (LP). DOMINICAN REPUBLIC. (LP nte Cristo, Jiménez 1598 (LP). IB Falda del Morro, Oriente: Guantánamo, Ekman 64 (S). ARGENTINA. Jujuy: Tafna, Cordo & Fer- rer 88-B-17 (SD. ARGENTINA. Entre Ríos: Federación, Bur- 3169 & Crespo (LP): Santa Ana, Serrano 6 (L P. T AZIL. Rio Grande do Sul: Porto Alegre, Rambo 545 (LP). .. cowellii: CUBA. Santa Clara: Santa Clara, Britton «e Cowell 10183 (NY Baracoa, Ekman 4023 (S). : CUBA. Oriente: Sierra de Nipe. Ekman 4767 (S). 9632 (S). — Fiebrig 2838 (LP). ¿NTINA. Jujuy: Humahuaca, Meyer 21. id Д n p La Cande a Schreiter 9409 (LP): La ndelaria, Jeréz et al. 49120 (LP . 2 MYANMAR. — Lweji, Maung Mya 5309 (LP). Rio de Janeiro: Rio de Ja- neiro, Glaziou 11072 (K). Minas Gerais: Pocos de — Leoncini 88 (LP): Pico da Bandeira, Shep- 5771 et al. (UEC). BRAZIL. 5 (LP). Bahia: Igregia Velha, Blanchet AZIL. Minas Gerais: Claussen 1301 (NY ). J. A. Sin. loc., Wright 2875 (GH). —— AN REPUBLIC. Cordillera E en- tral, Samana, Los Haitises, Ekman H-15498 Minas Gerais: Morro d Frfo, : Diamantina, Hatschbach 30192 istrito Fede ral: Brasilia, Hatschbach 43151 ardner (US). D Y CHILE. Santiago: Cerro Renca, Cabrera Valparaíso: i 51 (LP). El Quisco, Маһи 10336 Goyaz: Capella da Passe. Gardner ). ARGENTINA. Tucumán: Tafí. Fabris 1343 ndoza: Cerro de La Gloria, King 183 (LP). Oriente: Región de Moa, Cerro Mir- . Marie-Victorin et al. 21591 (GH). BRAZIL. Minas Gerais: Jabotic atubas, Hatschbach 29951 (LP); Datas, pr hbach 301. LP). aflore BRAZIL. Mato Grosso do Porá, Meyer 18770 (LP). PARAGUAY. Sierra de — Rojas 6575 (LP) — 9752| (G) 1 (L Sul: Amambay: ). Rojas [herb. ^st. Los cinco hermanos, Sancho Ponta ). Hidalgo: С ;onzález Quintero 3215 5 (LP). Р). Cañada del Vaquero, Coahuila: Sierra Gav- а. Johnston 7223 А ilicifolia: BAHAMAS ISLANDS. Andros Island: Cop- pice, Small & Carter 8526 (К) G. G. G. G. G. ~ ` ~ Ss re кш . microcephala: CUBA. Oriente: NY : — BRA А тота па: . E. igocephala: . iris ata: pare ili CUI 38 (NY), . pauc for ulosa: А — BRAZIL. a — . recurva: rusbyana: BOLIVIA. Yungas: В . sagraeana: CUBA. ~ Shaferi: CUBA. Oriente: ñ — 4 us tabilis: intertexta: CUBA. Pinar del Río: Cajalbana, Alain A- 1680 (NY magna: MEXICO. San Luis Potosí: Cronquist 11277 (NY): Queretaro, 5 km SW Jalpan, Fernández 3666 NY). CUBA. Pinar del Río: 8 (LP). Guane, Shafer Boca Guantánamo a itua, Bro. — 4874 a Rio Grande dis Sul: Malme 646 5); pr. 3 Santa aah Malme 1261 (5). A. Pinar del Río: Guane. Ekman 5 (S). e е СОВА. juger 2:4 тп Baracoa region, NY). sa de Prada, /ебп 119 e ligar REPUBLICA eh AN, i Monte Cristi, of Villa Isabel, Jiménez 3614 (US BRAZIL. Bahia: — ra ae Jacobina, Blanchet 3288 (U: S); Río Branco, Curran 284 (NY), Salzmann s.n. = ВК, AZIL. Sao Paulo: Moóca. Brade 5523 S). Rio Grande do Sul: Guaiba. Sancho 48 (LP): e ше y poe — 998 (К). — С”) T y palosanto: ARG A. Jujuy: San Pedro, Cabrera F — ee үз E Tucumán: Vipos, Trancas, Venturi 1296 (LP), Schreiter in 1925 (LP). BRAZIL. Minas е rais: Р), Gardner 4810 JA. Oriente: Massa. Brade а 'n Savannas, Shafer . pain PERU. La Libertad: Pataz entre Huaylillas ‘A Pavabamba, López & Sagdstegut 3409 (LP). BAHAMAS NDS. Fortune Is- land: Eggers 3866 (К); Mariguana Island, 10 mi. W of Abraham Bay, Wilson 7. picardae: HAITI. , eroix-des Bau- quets, gorge of Grande-Riviére de cul-de sac. Ekman Н-5385 (К). — Massif de la Selle М — BRAZIL. Rio de Janeiro: Tijuca, Glaziou 6 (LP). Sao Paulo: Paranaiba do Sul, Hash- imoto 624 (LP). Rio Grande do Sul: Pereira 8609 & Pabst 7984 (LP). PARAGUAY. Amambay: San- es 26, Sancho 32 (LP). San Pedro: Lima. Pedersen (LP). Sao Paulo: Campos de Emas, Ca- brera 12311 (LP); Ityrapina, белг 8296 (LP). — . purpustt: MEXICO, Puebla: Tehuacán, Purpus 4248 (NY). BRAZIL. > (5). — e p) CUBA. Oriente: 946 (LP). Xanxeré, Rambo Palmeira. Rambo Santa Catarina: Rio Grande do Sul: entre Moa y Punta Andén, León 20 , rotundifolia: BRAZIL. Sao Paulo: Capital. a Handro 157 (LP); Vila Esperanga, Joly 59 Alto Urumbamba, Zamalloa 2015 (LP). Habana: León 7094 (NY); Vedado, Cusco: Main 2532 (NY). La Caridad, López Figueiras 1738 3 (NY). М шт MEXICO, Oaxaca: Cuesta de Coyula, Conzatti 5 (US). BRAZIL. Parana: Ponta por Hatschbach (LP); Senges, M hbach 2 K). INDIA. tar Pradesh: )un District, Calrola 32 (LP). 7422 ea Dehra 548 Annals of the cae Botanical Garden G. — HAITI. "aix, Ekman H-355. G. vargasii: PERU. — Abancay, Vargas 16317 (LP). C. — BRAZIL. Hatschbach 23447 (LP), lein 14885 (LP). G. iii hat Re RU. Amazonas: Chachapoyas, Tingo, — 97 (LP 2 e i Chachapoyas y Leimebam- ez еі al 1 (LP). La Libertad: Bolívar, — 1701 (L E ps & Sagástegui 3354 (LP). — du Nord-Ouest, Port-de т Parana: Рома Grossa, Vila Velha López 4364 (LP), Smith & Hyalis H. — еа (LP). Н. lane ifolia: ARGENTIN haco: San Fernando, Ca- brera 4083 (LP). Sc — 16098 (LP). ARGENTINA. Mendoza: Tupungato, Ruiz 701 (LP lanthopappus f, us aca ARGENTINA. Corrientes: Paso Troncón, 'alacios & Cuezzo 2304 (LP). Nouelia N. insignis: CHINA. Maire 2516 (NY). Yunnan: l’ Abbe Delavay 2498 (US), Rock 11714 (US). Pleiotaxis P. dewevrei: DEMOCRATIC kulu, de Hitte 288 (U$ P. eximia: DEMOCRATIC REPUBLIC OF CONGO. Tshinloingwe (Hout Katanga), Rolyns 1836 (US). ZIMBABWE. 20 km de Mangula, Lavranos 22745 REPUBLIC OF CONGO. Lu- S). (US). P. huillensis: ANGOLA. Gossweiler 10780 (US). P. pulcherrima: ANGOLA. Gossweiler — US Р, je ide ZAIRE—DE MOC RATIC REPUBLIC O CONGO. E Tem thville, Rolyns 1568 (US). Huila, Hampata, Distr. do Cuanza Sul, Seles, ` K Wunderlichia W. yes nsis: BRAZIL. l. 25209 (MO). W. vilia BRAZIL. Goiás: Chapada dos Veadeiros, ter et al. 2615 (MO). W. —— BR) AZIL. Goiás: Serra do Cristais, Irwin et al. 9913 (NY); about 52 km W of Alto Paraiso, Mar- tinelli & Stutts 999 (NY edra Azul, Harley Minas Gerais: | ~ ч APPENDIX 2. New infrageneric classification of Gochnatia. KEY TO THE SECTIONS OF GOCHNATIA l. Pappus type A (all bristles are thin and have the same length and width) and/or pappus type B (all bristles are thin and have the same width, but about half are shorter) C (all bristles are thin and have ~ and N the longer are plumose at the apex) or pappus type D (half of the bristles are long and wide, and the other half are short and thin) 2(1). Subshrubs; capitula solitary, very long-pedun- culate, or Wan in se apiform pse ges 'orymbs G. . Discoseris ~ 24. Tres or shrubs; capitula short-pedunculate or essile D 3(2). кубе phyllaries dorsally oo or subglabrous and ciliolate at the marg 3'. Involucral phyllaries dorsally tomentose and. not ciliolate at the ma 3). Capitula arranged in glomerulose pse eudoco- Seu еа les: TAINS ә ÉS RE rymbs, pseudoracemes, or | phyllarie s extending into the pedun G. ‚ Leucomeris 4. Solitary capitula; — not «ш ы into the peduncle G. sect. ni de. 5(3). Solitary capit tula o r2or3 5. € apitula i in glomerulose — 'orymbs G. sect. Glomerata ly 6(5). Leaves spiny; corollas deeply to very deeply lobe t. Anastraphioides O”. Leaves not spiny; corollas des Arn Tobe d G. sect. Gochnatia Pappus type ( 2: :; plants not glandulose, with : and 3- to 5-armed hairs; gynodioecious or po- lygamous бес ious: corolli as subdimorphic (isomorphic), 9-15(50) ; sect. ا‎ 7'. Pappus type D; plants | gl Caes monoecious corollas isomorophic, 5 (3, 7, or 10-20 7(1). ~ A ER sect. Pentaphorus Gochnatia sect. Anastraphioides Jervis ex S. E. Freire, „ Katinas & G. Sancho, sect. nov. TYPE: Gochnatia ilicifolia Less. — Arbores vel frutices, foliis alternis, spinoso-dentatis vel intege rrimis. Capitula apicibus ramulorum, solitaria vel 2- . sessilia. Inve — campanulatum vel turbinatum. Flores 4—150, lutei vel aurantiaci, isomorphi, herma- — corolla шы pentasecta vel profunde penta- Ча. Antherae. appendicibus connectivalibus caudatis w ec = * abruptis, appendicibus basalibus integerrimis vel la— ciniatis. Pappus uniseriatus vel biseriatus. Shrubs or small trees, monoecious. Leaves alternate, pet- iolate or shortly petiolate; obovate; oblong, or elliptic with margins spinose-dentate upper surface generally Hee lower tomentose (with flagellate hairs). — homagamous, scoid, sessile, solitary at the tip of the branches (rarely 2 or 3). Involucre campanulate or turbinate. Phyllaries in 1 to 10(15) series, dorsally tomentose. Florets 4 to 150, isomorphic, tubulose, yellow or orange, hermaphroditic, )pendages deeply to very deeply 5-lobed. Anthers with aj caudate (rarely abruptly apiculate), and — smooth, occasionally laciniate tails. Style bilobate or shortly bifid; style brane hes rounded, dors ally glabrous. Cypsclas with of the same length (rarely biseriate with a reduced number of outer short bristles) all thin. Twe nt y-six spec les: ©. attenuata, G. buchii, G. calcicola, pha G. par vifolia. G. — G. ріс shaferi, С. sagraeana, б. Lortuensis, G Distribution. Domini- Bahamas Islands, Cuba, Haiti, can Republic, i Puerto Rico Observations. As pointed out by Jervis (1954) and Ca- Volume 89, Number 4 2002 Freire et al. Gochnatia Morphology brera (1971), Bf рн D. Don (Trans. Linn. Soc. Lon- don 16: 295. onsidered an unknown : cause Its type ini ilicifolia D. Don Joannes Fraser" 'imen in the Lambert herbarium) has never been locate d. The original diagnosis of Anastraphia does not agree with the later interpretation of the genus 6). by de Candolle (1838: 26 0) is c Gochnatia sect. Discoseris O — Endlicher) € cabrera, Revista 6: 7 — NS " Ш) % A ^A = = ~ Z c c -— = Post & Kuntze, Lex. Gen. Cima 1904, nom. rfl. TYPE: Seris discoidea Less. Di Gochnatia discoidea (Less.) Cabrera) supe — monoecious. Leaves alternate, petiolate; wate-elliptic with margins entire or denticulate; — veined; glabrous or tomentose on both sides i — homogamous. discoid. solitary or arranged in scapose pseudocorymbs. — те коо to turbinate. Phyl- laries in 4 or long-pedunculate and orsally tomentose. Florets numer- ous (ca. 30), yellow, сие , isomorphic, tubulose, deeply 5-lobed. Anthers with appendages abruptly apic- ulate and laciniate tails. Style — bifid; style branches rounded, dorsally glabrous. Cypselas with duplex and glandular hairs. Pappus uniseriate with all bristles of the same length and thin 5 ser ies, Three species: G. amplexifolia, G. discoidea (type spe- cies), G. suffrutescens. Distribution. Southeastern Brazil. G n S. E. Freire, L. Katinas € TYPE: Goc hnatia arborescens — — ‘ted here). Goc — sect. S; Arbores vel frutices, foliis — Mis iiie Capit ula pauca, ibus nite od glomerata. Involucrum лл Flores 12-50, lu- tei, ison — bi maphroditi, corollis tubulosis, pentasec- is. Antherae appendicibus connectivalibu abruptis, sessilia subsessi la, s attenuatis vel appendicibus basalibus integerrimis. biseriatus vel uniseriatus. Small trees or shrubs, monoecious. shortly petiolate; ovate to elliptic with Pappus Leaves alternate, margins entire or dentic — pinnate Ny veined; upper surface glabrous or tomentulose, lower ace usually der AE Malla аў. Cda homogamous, discoid, ses- sile, few together in terminal glome — kG ory р. Inoue ‘re campanulate. Phyllaries in 5 6 series (rarely 8 to 10), dorsally tomentose or gla Florets 12 to 20 bu ca. 50), yellow, —— isomorphic, tubulose, deeply 5- lobo d. Anthers y asely tomentose ped with ap- pendages abruptly apiculate (rarely —— and smooth tails. Style shortly bifid or bilobate; style branches round- ed, dorsally glabrous. Cypselas with duplex and glandular hairs. Pappus biseriate of numerous scabrid bristles, with a reduced number of outer short bristles (rarely uniseriate with all bristles of the same length), all thin Three spec ies: G. arborescens (type species), G. magna. G. ригриѕи. Distribution. Northern Mexico. Gochnatia Kunth sect. Gochnatia, Nov. Gen. Sp. 4: 15. 1818. TYPE: Gochnatia vernonioides Kunth. Shrubs, monoecious. Leaves alternate, shortly petiolate; ovate or ovate-elli with margins entire (rarely dentic- ulate); pinnately, occa ia anal subtri-veined; upper surface generally glabrous or ы: клы ower surface densely tomentose (м flagellat Майа), Сари e omentose above. Florets pe in series, — lly tomentose or only to + ellow, ече a litic, — 5- iud ers with appendages caudate, more rare ly abruptly apic eae and iss tails. Style shortly bifid; style branches rounded, dorsally glabrous. Cypse las with la hairs m Mosis glandular hairs. Pappus bis- eriate of numerous scabrid bristles s, with a ber of outer, short bristles, all thir tubulose, reduced num- Seven species: G. arequipensis, G. boliviana, G. carden- asii, G. curv a ora, G. patazina, G. vargasti, G. vernonioides (type specie Distribution. Andes of Peru, Bolivia, and northwestern Argentina. Goc 'hnatia sect. е (Lessing) DC., Pre 7(1): 24. 1838. Gochnatia subg. He leia PA Syn. а Compos.: 103. 1832. TYPE: Gochn- atia cordata Less. Gochnatia sect. Moquiniastrum Cabrera, Revista Mus. L Plata 12, Secc. Bot. 66: 73. 1971. TYPE: А polymorpha Less. (= G. polymorpha (Less.) Cabrera). — Shrubs or trees, commonly By nodioeci ‘ous or polvgamous Leaves alternate, petiolate or shortly petiolate: ovate, elliptic (rarely linear or cordate) with margins en- tire, rarely denticulate; pinnately veined; upper surface glabrous (rare ly tomentose), lower surface densely tomen- with 2-, 3- to 5-armed hairs, occasionally flage llate) ири heterogamous (homogamous), dise ¡form — discoid), subsessi dioecious. tose ( ssile to pedunculate, numerous arranged in terminal and usually loose leafy pseudopanicles. Involucre oblong to campanulate. — in 3 or 4 series, dorsally tomentose. Florets 9 to 15(50), creamy or white, tubulose, deeply 5-lobed, subdindtiphie (occasionally isomorphic): functionally female, corollas slightly zygomorphic, straight lobes: — ac — Anthers кү ар and laciniate or sm iUis ails. Style s bilobate): style bra selas — duplex. epe nl and f biseri of numerous scabri with with resu- pinate lobes. ж d а метав of outer short bristles, all thin, and the longest are plumose at the ay ea, G. bar- cies), G. den- Twe 2n -one speci rosii, : G. argentina, G. argyr ‚ blanc hetiana. G. cordata (type spec on G. Дог ш. G. — iniana, G. mollissima, G. olig phala, G. orbiculata, G. Dan c. аиа G. "e chra, G. ramboi, G. rusbyana, G. sordida, G. velutina. OCE= Distribution. Andes of Peru and Bolivia, eastern Bra- ‚ Paraguay, Uruguay, and central-eastern Argentina N Сос hoatia sect. Leucomeris (D. Don) Cabrera, Revista Mu . La Plata 12, Secc. 66: 128. 1971. Leucomeris D. 3o Prodr. Fl. Nepal.: 169. 1825. Gochnatia ru Less., Syn. Gen. SURE . Don) sochnatia spectabilis (D Less. (= — —— spectabilis D. Don). 550 Annals of the Missouri Botanical Garden Trees or shrubs, monoecious. Leaves alternate, shortly pe tiolate; elliptic, margins entire or denticulate; pinnately — ined; Md r surface generally parous to tomentose, surface densely tomentose (with flagellate > hairs) o = تب tending to the peduncle. Florets 4 (rarely yellow), hermaphroditic, isomorphic, tubulose, deeply 5-lobed. Anthers with appendages attenuate (rarely abruptly apiculate) and lac iniate tails (rarely smooth). I ily with glandular hairs. Pappus Бегай of nu- merous scabrid bristles, with a reduced number of outer short bristles, all thin Five species: G. hypoleuca, б. RON. G. smithii, G. decora, C spectabilis (type species Distribution. Mexico, Brazil, + region of Bolivi: and Argentina, and southeastern Asi Gochnatia sect. Pentaphorus (D. Don) DC., Prodr. A 1): 24. 1838. e D. Don, Trans. Linn. S 16: 296. 1830. subg. P uei ). Don) Hook. & Arn., Cor Bot. Mag. 1: 108. Ta a PE: * — "foliolosa D. Don ex Hook. & Ап — = monoecious. Leaves alternate, sessile; linear- ate with margins entire or — por⸗ and low- Shrubs, ovate or linear-obov: tion denticulate; pinnately or three-veined: upp er surfaces glandulate (with or without ode hairs). Capitula homogamous, discoid, sessile or subsessile, nu- erous in leafy glomerulose pseudorac Involucr 1 cemes. re campanulate. Phyllaries i in 4 to 6 series, dorsally glabrous or — with ciliolate margin. Florets 5 (rarely 3 r 10 to 20), white or lilac, hermaphroditic, isomorphic. lee deeply 5-lobed. Anthers with appendages abruptly apiculate md smooth or laciniate a pare bi- lobate or shortly bifi d; style branches rounded, dorsally relatively vids and long. and the others short and thin. Two species: G. foliolosa, G. glutinosa (type species). Distribution. Western Argentina and central Chile. Gochnatia sect. Rotundifolia 5 . E. Freire, L. Kat Sancho, sect. nov. TYPE: Gochnatia rotundifolia Less. (selected here). Frutices, foliis alternis, integerrimis vel denticulatis. Capitula apicibus ramulorum solitaria, sessilia. Involu- crum campanulatum; bracteis involucralibus subglabris, margine ciliatis. Flores multi, E — herma- phroditi, corollis tubulosis, pentasectis. Antherae appen- dicibus connectivalibus attenuatis р appendici- bus basalibus laciniatis. Pappus biseriatus, setosus. Shrubs, monoec ious. Leaves alternate, shortly petiolate; broadly ellip ulate; three veined, ru see on both surface PUDE leaves with flagella te hairs). Cia homogamous, dis- Phyllaries c with margins entire to dent 2 ur z => 5 t3 lose, deeply 5- ва е and laciniate tails. Е pour style branches rounded, dorsally glabrous. Cypselas with duplex and glandular hairs. Pappus biseriate of numerous scabrid bristles, with a reduced number of outer, short bristles, all thin One species: G. rotundifolia. Distribution. Southeastern Brazil. EL GENERO QUERCUS (FAGACEAE) EN EL ESTADO DE MEXICO! Silvia Romero Rangel, Ezequiel Carlos Rojas Zenteno? y María de Lourdes Aguilar Enríquez RESUMEN Con base en trabajo de campo y E herbario, se re — en 23 especies de Querc us para el Estado de México. Diez pertenecen a la sección Quercus y a la sección — distribuidas ше Centro 8 ıérica y sólo Quercus rugosa se encuentra . El trabajo incluye una clave dicotómica y des os Estados Unidos Lobatae. Quince de las especies son endémicas de M éxico, siete en el norte de la frontera mexicana. en ‘ripciones morfológicas de cada uno de los taxa. re ней nte, para cada especie se proporcionan datos e nir ree os, fenológicos y ecológicos ABSTRACT Based on 1 field and herbarium work, in this paper 23 species of Quercus are rec ognized for the State of México. Ten xtend to Central America, and only Quere 1 Quercus = 13 to section Lobatae. Fifteen of the species are endemic to Mexico, seven is rugosa is found north of the Mexican border in the United States. The paper includes a dichotomous key and эйе ыг ‘al descriptions for all the taxa, as well as ethnobotanical, pheno- ogical, and ec al data "Kay words: Fagaceae, — Quercus. El género Quercus se distribuye mundialmente en zonas templadas y subtropicales del hemisferio norte. Se calcula que está conformado aproxima- damente por 500 especies distribuidas mundial- mente. México es el país que posee el mayor nú- 150 distribuidas en las zonas monta- mero de especies del mundo, entre 135 y (Nixon, 1993b), ñosas de todos los estados y territorios, a excepción del sur en Yucatán y Quintana Roo (Rzedowski. 1978 las especies del género Quercus se les en- cuentra formando bosques de encino, comunidades muy características de las zonas montafiosas de México. De hecho junto con los miembros del gé- nero Pinus constituyen la mayor parte de la cu- bierta vegetal de áreas de clima templado y semi- húmedo; sin embargo no se limitan a estas condiciones, pues penetran en regiones de clima caliente formando también bosques, no faltan en las francamente húmedas, siendo elementos del bosque tropical perennifolio y bosque mesófilo de montaña, y aún existen en las semiáridas siendo parte del matorral xerófilo, asumiendo con frecuen- cia forma arbustiva (Rzedowski, 1978). Existen dos trabajos que intentan registrar las especies de la entidad en estudio. Uno es de Mar- tínez (1954) titulado “Encinos del Estado de Méx- ico”: en él se refieren 45 especies, número que se ha modificado por la gran cantidad de sinónimos. El otro es de Espinosa (Rzedowski & Rzedowski. 1979), quien publicó, dentro de la “Flora fanero- eámica del Valle de México”, los encinos de esta zona. Una porción de esta región pertenece al Es- tado de México; en él se mencionan 13 especies para el área de estudio. Otros trabajos importantes que consideran especies de la entidad estudiada son los de Muller y MeVaugh (1972), MeVaugh (1974) y de Aguilar y Romero (1995). El Estado de México ocupa una superficie de 22.500 km? y está comprendido en dos provincias fisiograficas: la del Eje Neovolcánico que abarca la mayor parte de la superficie del norte, con altitudes ! Se agradece a los encargados de los herbarios CHAPA, CODAGEM, ENCB, INIF, IZTA y MEXU por el préstamo A los Doctores Jerzy Rzedowski R., de sus colecciones. Patricia Dávila A. y Rafael Lir a Š. por sus comentarios y valiosas sugerencias. A la Bióloga Adriana Bernal P. por la realización de los dibujos de ә ejemp = comunicación pertenece al proyecto encinos de México que se lleva a cabo en la UBIPRO de lat UNAM Campus Iztacala. 2? UBIPRO. _ NAM Campus lztacala. Av. р 4090. sromero@servidor.unam.t 5 Педа IZTA-UN. AM Cam México. C.P. 54090. zenteno@servidor.unam. mpus Iztacala. nh De los Barrios s.n De ы Barrios s.n. Los Reyes Iztacala Tlalnepantla, Estado de México, . Los Reyes Iztacala Tlalnepantla, Estado de México, ANN. Missourt Bor. GARD. 89: 551—593. 2002. Annals of the Missouri Botanical Garden que van de 2000 a 5452 m y la de la Sierra Madre del Sur que abarca el extremo SE, donde las alti- tudes son de 600 a 3000 m (INEGI, 1987) Los encinos son abundantes en las regiones mon- tafiosas de la entidad estudiada, formando parte de los bosques de coníferas, bosques de Quercus, bosques mixtos de Pinus—Quercus y Cupressus— Quercus; aunque también habitan en otras comu- nidades tales como bosque mesófilo de montaña, bosque tropical caducifolio, no faltando en matorral xerófilo, bosque de galería y pastizal. Quercus es considerado como un género taxo- nómicamente complicado, debido a su gran varia- bilidad morfológica, a la capacidad de formar hí- bridos y a que sus estructuras reproductivas no se han estudiado lo suficiente (Romero et al., 2000). Esto ha llevado a la existencia de una gran canti- dad de sinónimos. MATERIALES Y MÉTODOS El trabajo se basó fundamentalmente en la re- visión y estudio de los ejemplares existentes en los herbarios CHAPA, CODAGEM, ENCB, INIF, IZTA y MEXU. Se obtuvo la variación morfológica regional de los ejemplares citados y de las observaciones rea- lizadas en campo por los autores. Para determinar la aplicación de los nombres se revisaron las descripciones originales y los ejem- plares tipo provenientes de herbarios; cuando estos últimos no estuvieron disponibles para su análisis, se utilizaron en su lugar las fotografías de los tipos contenidas en Trelease (1924). Estos últimos y los herbarios donde están depositados se citan como lo hace dicho autor. En este trabajo se incluyen sólo los sinónimos encontrados para el Estado de Méx- ico. Las descripciones morfológicas se realizaron de acuerdo al formato del tratamiento del género Quer- cus para la Flora Novo-Galiciana (Mc Vaugh, 1974) El hábitat, las descripciones morfológicas, la dis- tribución geográfica y los dibujos de las especies están basados en los ejemplares revisados para cada especie. TRATAMIENTO TAXONÓMICO PARA EL GÉNERO QUERCUS EN EL ESTADO DE MEXICO Son 23 las especies reconocidas en este trabajo; 10 pertenecen a la sección Quercus y 13 a la sec- ción Lobatae, de acuerdo a la clasificación del gé- nero Quercus propuesta por Nixon (1993a). TYPE: Quercus L., Syst. Pl. ed. 2, II, 994. 1753. (fide ING) 1938).] Quercus robur L. . [Sinonimia com- pleta en Camus Árboles o arbustos; reunidas en las puntas de las ramas; estípulas su- monoicos; yemas foliares buladas o liguladas, generalmente decíduas, a ve- ces persistentes, más bien asociadas con las yemas que con las hojas; hojas alternas, generalmente pe- cioladas, nunca totalmente sésiles; amentos mas- culinos largos y colgantes, flores con el cáliz 5- lobulado, fusionado a un perianto que envuelve a los estambres; estambres de 5 a 10, libres, con an- teras cortas y filamentos delgados; flores femeninas en forma de racimo reducido con un raquis leñoso corto o largo y con una o varias flores, el cáliz con 6 lóbulos que se adhieren a la base de los estilos y se fusionan en un tubo; pistilo de 3 carpelos que forman un ovario trilocular, cada lóculo con 2 óvu- los; estilos 3, libres; fruto unilocular con una se- milla, los otros 5 óvulos son abortados; semilla en- vuelta en una cubierta rígida formando la bellota que está protegida parc ‘ialmente en su base por una cúpula cubierta de escamas; N = 12 I. Quercus sect. Quercus Base del perianto femenino adnado al ovario; es- tilos cortos y anchos; pared interna del endocarpo glabra; escamas de la cúpula aquilladas, engrosa- das en la base; dientes de las hojas mucronados; óvulos abortivos basales. 1. О. deserticola Trel. 2. Q. frutex Trel. 3. О. glabrescens Benth. 4. Q. glaucoides M. Mart. & Gal. 5. Q. laeta Liebm. 6. Q. magnolufolia Née 7. Q. obtusata Humb. & Bonpl. 8. Q. peduncularis Née 9. Q. rugosa Née 10. Q. splendens Née П. Quercus sect. Lobatae Loudon., Hort. Brit. 385. 1830. TYPE: Quercus aquatica Walt. (lectotipo, designado por Nixon (1993a)). Base del perianto femenino libre; estilos por lo general alargados, lineares; pared interna del pe- ricarpo pubescente; escamas de la cápula, no aqui- lladas, no engrosadas en la base; dientes de las hojas, por lo general aristados; óvulos abortivos en posición apical o lateral. Q. acutifolia Née 12. Q. candicans Née 13. Q. castanea Née Volume 89, Number 4 Romero et al. 553 2002 Quercus en el Estado de México 14. Q. conspersa Benth. rarlas es que Q. frutex posee pubescencia en el 15. Q. crassifolia Humb. & Bonpl. envés de las hojas que permite ver la epidermis, l6. Q. crassipes Humb. & Bonpl. formada por tricomas estrellado-estipitados con me- V7. Q. dysophylla Benth. nos de nueve rayos; mientras que Q. repanda posee 18. Q. elliptica Née pubescencia muy densa que no deja ver la epider- 19. Q. hintonii E. F. Warb. mis, formada por tricomas estrellados sésiles o cor- 20. Q. laurina Humb. & Bonpl. tamente estipitados con más de 15 rayos. 21. Q. mexicana Humb. & Bonpl. 22. Q. scytophylla Liebm. Il. QUERCUS SECT. LOBATAE 23. Q. urbanii Trel. Quercus affinis. Schiede. Citada por Martínez De los 45 nombres citados en la bibliografía para (1954), ips à similar a Quercus laurina, con la que el Estado de México, 20 resultaron ser sinónimos {Че confundida. La pU distingue por —— de las especies tratadas en este trabajo y las cinco Ye™mas foliares de forma conoidal, base de las hojas siguientes las hemos considerado no existentes en cuneada, nervaduras planas, y distribución en alti- la entidad estudiada; se indican las secciones a las tudes menores a los 2400 = mientras que Q. lau- que pertenecen (Nixon, 1993a). rina posee las yemas ovoides, la base de la hoja atenuada о redondeada, nervaduras realzadas y dis- 1. QUERCUS SECT. QUERCUS tribución en altitudes mayores a los 2200 m. Quercus мы Née fue confundida también Quercus sanchez-colinii Mart. Citada por Martí- por Martínez (1954) con Q. laurina. La primera nez (1954), seguramente corresponde a una varia- presenta de 15 a 25 а primarias еп las ción de Q. laeta debido a la similitud morfológica hojas y se distribuye en la vertiente del Pacífico, entre ellas. Sin embargo no se han vuelto a colectar en altitudes de 600-2000 m. La segunda posee un 4 a 12); se — ejemplares con hojas de borde ondulado como las número menor de nervaduras primarias observadas en el tipo de la primera especie. le encuentra en estados de la vertiente del Pacífico Quercus repanda Humb. € Bonpl. Citada por Es- y del centro del país, en altitudes de 2240-3150 m. pinosa (Rzedowski & Rzedowski, 1979), fue con- Quercus aristata Hook & Arn. Citada por Mar- fundida con Q. frutex debido a que ambas presen- — tínez (1954), no se encontró en los herbarios revi- tan formas arbustivas y similitud en la morfología sados, ni se logró colectar en los sitios referidos en de la hoja. Una diferencia que hace posible sepa- la descripción original. CLAVE DE LAS ESPECIES DE QUERCUS DEL ESTADO DE MEXICO, MEXICO la. rue de borde entero, algunas veces mucronadas o aristadas en el ápice, pero sin dientes laterales nvés de las hojas glabro o con tricomas concentrados en las axilas de las nervaduras o a lo Em de las mismas, frecuentemente glandular. Hojas comunmente glauca is, ápice nunca aristado; cúpula con escamas engrosadas en la base; pared interna del pericarpo glabra la. Superficies maduras labras | por — to o con algunos tricomas cerca de la Pu nerva- duras rojizo amarillentas 0000 ). glaucoides 4b. Superficies maduras con tricomas a lo largo de la nervadura central, NE. ve des | NI ORES AT 10. О. — 3b. Hojas no glaucas, ápice comunmente aristado; escamas de la cüpula no engrosadas en la base sa. Ramillas y pecíolos — pubescentes —— Be Ж ыны 18. Q. elliptica 5b. стт y pecíolos glabrescentes o glabros. jas de 11-17 em de largo: con el envés con abundantes tricomas glandulares ......... RR И КЕЛЕНКЕ IO EN. A AA — 14. Q. conspersa 6b. Hojas de 5-11 cm de largo; con escasos tricomas glandulares |... 20. Q. — 2b. Envés de la hoja pubescente, con los tricomas uniformemente distribuidos, — la lámina, no con- centrados en las axilas de las nervaduras o a lo largo de las mismas 7a. Apice de la hoja aristado; pared interna del pericarpo lanos: Ва. Envés de la hoja con tricomas contortos, epidermis — RE E XN 2]. Q. mexicana 8b. Envés de la hoja con tricomas no contortos; epidermis —— 9a. Pubescencia amarilla 17. Q. dysophylla 9b. Аган аду ө grisácea. 10 Narvadurda conspicuamente elevadas en el envés, de 5-12 en cada lado, bellota de 5-15 mm de largo . castanea 10b. Nervaduras ligeramente elevadas en el envés, de 10-19 en cada TM bello ta de 12-30 mm de largo 16. Q. crassipes + 554 Annals of the Missouri Botanical Garden Tb. Apice de la hoja con un mucrón; pared interna del pericarpo glabre 11а. Ar ‘bustos poe de 0.4 poa m de alto; n comunmente de 2—4.5 cm de — pula del fruto de 7-13 mm diár — ). frutex 11Ь. Шы 's — de 2— ij m de а. һојав comunmente de 4—7.5 ст de largo: cpu de fruto de 14-17 mm diár — 1. Q. deserticola Ib. Hojas con el borde dentado, ondulado o con n los bordes aristados o mucronados. 12a. Hojas con el borde ondulado o con dientes mucronados, nunca aristado. 13a. Envés de la hoja glabro, glabrescente o con tricomas concentrados en las axilas de las nervaduras o a lo largo de las mismas. 14a. Hojas con el envés glauco. 15a. Superficies — — por completo; árboles de 4-10 m de alto; My de 4— 4. > SIAM de dina ше oides I5b. Superficies maduras con tricomas que tienden a concentrarse en la nervadura с * мга árboles de 5 m de alto; estípulas de 5-8 mm de largo . Е Wu ). s andén 14b. Hojas con el envés no — е 3. — ). g 13b. Envés de la hoja pubescente, con los tricomas uniformemente distribuidos sobre la TEA no concentrados en las axilas de las ne — o a lo largo de las mismas. 16a. iue densamente pubescente Arbustos rizomatosos de 040-2. 50 m “ alto; hojas comunmente de 2—4.5 cm de argo; ctipula del fruto de 7-13 mm dia 2. Q. frutex Vib. н 's pequefios de 2-7 m de alto; hojas comunmente de 4—7.5 cm de largo; pis Q -] a — el fruto de 14—17 mm diám . deserticola 16b. Ramillas labras o glabrescentes. Tricomas del envés de las hojas estipitados 19a. Envés de las hojas con indumento sinclar frutos sésiles о en ip culos de nm de largo... s l. Q. dy — 19b. Envés de las hojas con indumento o blanquecir ino; . frutos en n pedúne a de 3.5- oni ае largo ers ырш e ыманы Мема ымы ШЫ 8. Q. — J rg 18b. P — 'encia del envés formada ү por tricomas sésiles o con un estípite Pi 20a. Tri 6. Q. magnoliifolia = 4 ricomas sésiles 0 20b. Tricomas con un A DB A corlo. 2la. Epidermis del envés de las hojas con algunos tricomas puse AD MA A e NE E ы а айлына E Ыссар ан аат нышын елаш 0. laeta 21b. — rmis del envés de las hojas con abundantes tricomas glandt is — de la hoja con mucrones de hasta 2 mm de largo; bellota vide o angostamente elíptica — 9. O. rugosa 22b. Marge n de la hoja con mucrones cortos, menores de 1 mm, que se 'urvan hacia el envés; bellota globosa o cilíndrico-ovoide |... 7. Q. obtusata 12b. Hojas con borde aristado. 23a. Envés de la hoja glabro, o con tricomas concentrados en las axilas de las nervaduras o a lo largo de las mismas. 24a. Epidermis — yemas de 1.5—4 mm de largo; árboles altos, de hasta 30 m, crecen en 10-315 20. altitudes de AAA AAA Q. laurina 24b. Epidermis no — sa; yemas ed mm de [ox árboles Шош; de hasta 12 m de alto, ‹ altitudes menores de 2 y | 11. Q. acutifolia 23b. Envés de la hoja con los tricomas тё memente distribuidos sobre la lámina, no concentrados 1 las axilas de las nervaduras o a lo largo de las mismas. 25a. Envés de las hojas blanquecino, — m она рог tricomas sésiles 26a. Hojas con el haz verde lustroso, de 1 nervaduras en cada lado; yemas de 3-5 mm de — estípulas de 10-15 mm гу ie frutos de 20 mm de largo y de 17 mm de AAA 12. Q. candicans 206b. Hojas con el haz ale grisdceo, no Шанбай, de 5-9 nervaduras en cada lado; pec "= de 14 cm de largo; — de hasta 5 mm de ا‎ frutos de 8-10 mm de larg de 10 mm de ancho — s 22. Q. : — 25b. Envés ids las hojas no blanqueci ino, indumento formado por tricomas estipitac ados 27a. Envés de las hojas con ose grisáceo; hojas elfpticas, elíptico- oblonga, oblan- ceoladas o lanceoladas |... s . castanea 27b. Envés de las a con indumento amarillento: hojas obovadas, oblongo- ola suborbiculares u orbiculares, ovado-elípticas, raramente elípticas. 28a. Ramillas de 5-11 mm diám., — panduriformes, suborbic — orbiculares, obovado-elípticas u ovadas, de 15-30 em de largo y de 17-34 em de e :ho . M e cias 3. Q. urbanii 28b. — de 2-5 mm diám., hojas planas, lane eoladas, ovado lanc coladas obo- adas, oblongo-obovadas o elípticas, de 4-20 em de largo y de 3-12 em de ancho. Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México 29a. Epidermis de la hoja ampulosa; cúpula del fruto hemisférica; — ovoide 15. Q. crassifolia 29b. Epidermis de la hoja lisa; cúpula del fruto poc -uliforme a | pate Ew. be- — — 1. Quercus e Trelease, Mem. Nal Acad. Sci. 20: 79, pl. 113. 1924. TIPO: Méx- ico. p. Uhde 309 (B). Quercus alveolata Trel., Mem. Natl. Acad. Sci. 20: 80, pl. 114. 1924. TIPO: Cerro del Gavilán, Puebla, Purpus 1 (NY). em. Natl. Acad. Sci. 20: 81, pl. Quercus texcocana Trel., 117. 1924. TIPO: e rro Texcotsingo, cerca de Te "x coco, Endlich 653 (B). Arbolillo de 2—7 m de alto, corteza gris; ramillas de 1-3 mm diám., densamente pubescentes, verde- amarillentas, el indumento se ennegrece con el tiempo, formado por tricomas estrellados: lenticelas de hasta 1 mm, blancas, visibles sólo en ramillas viejas; yemas de 2-3.5 mm de largo, ovoides, es- camas escariosas con los bordes ciliados; estípulas de 3-5 mm de largo, lineares o filiformes, pilosas, rojizas, persistentes en hojas Jóvenes y yemas; ho- jas jóvenes rojizas, haz verde, con el indumento formado por abundantes tricomas estrellados cortos. estipitados, envés más pálidos, con pubescencia más densa formada también por tricomas estrella- dos estipitados, pero con las ramas más largas: ho- jas maduras oblongas, elfpticas, elfptico-oblongas u obovadas. lámina (2.5-)4-7.5(-8.5) X (1.3—)2—3.5(—4.5) ст, ápice agudo u obtuso, mu- subcoriáceas, cronado, base cordada o subcordada, borde entero. revoluto, ondulado o dentado, 2 a 5 dientes mucro- nados de cada lado, frecuentemente asimétricos; nervaduras de 6 a 9 de cada lado, que se continúan en el diente; haz verde lustroso con tricomas estre- llados cortos, estipitados, uniformemente distribuí- dos, más abundantes en la nervadura central, ner- vaduras impresas; envés pálido, con abundantes tricomas estrellados, estipitados, con las ramas más largas que los del haz y tricomas glandulares rojizos sobre la epidermis ampulosa y papilosa, nervaduras elevadas; pecíolos de 2-5 ] mm diám., pubescentes al igual que las ramillas, mm de largo, de 0.5— base engrosada; amentos femeninos con 3 a 10 flo- res en pedtinculos de hasta 30 mm, densamente pubescentes: fruto anual, solitario o en grupos de 2 о З sobre pedünculos de 2-9 mm de largo: cúpula hemisférica, de 14—17(—20) mm diám., con pubescencia blanca, las de la base engrosadas, las escamas las superiores menos pubescentes, ápices obtusos, ligeramente elevados; bellota ovoide, pared interna del pericarpo glabra, de 11-13(219) mm de largo, de 11-15 mm diám., un tercio de su largo incluida en la cúpula. Figura ota globosa a comprimida J. Q. hintonii Reconocimiento. Quercus deserticola se reco- noce por ser un arbolito con las ramillas densa- mente pubescentes y hojas de hasta 7.5 em de largo con bordes revolutos y sin aristas. Distribución y hábitat. En México en los esta- os de Distrito Federal, Guanajuato, Hidalgo, Ja- lisco, Estado de México, Michoacán y Querétaro. En bosque de Quercus, pastizal y matorral xerófilo, asocia con Alnus se 2600-2800 m. Fenología. y Cupressus, en altitudes de Florece en abril y fructifica de julio a diciembre. Nombres populares y usos. Encino, encino to- cuz, encino chico. Su corteza se usa para curar encías, amacizar los dientes y en curtiduría; su madera como leña, para fabricar pulpa para papel, carbón, postes para cer- ca, arados, cabos para herramientas y horcones (González. 1986: Bello € Labat. 1987). MÉXICO. Estado de Méxi- co: Seule de San Luis Aculco, Capse 99 (ENCB "Асар. Ciudad Adolfo López Mateos y cerca- nías, Маны 3 28314 (MEXU); parte alta del Cerro Chi- — — de Ciudad Adolfo López Mateos, Rzedowski Ejemplares examinadas ¿spinos 32588 (ENCB); Coyotepec, Sierra de Alcaparrosa, Fer- n sa s.n. (IZTA); Sierra de Alcaparrosa, Callardo 1 B (ТИТА): Sierra de — Romero 153 (ИТА); Eca- tepec de Morelos, 6 km al E de San Cristóbal Ecatepec. Rzedowski 32154 (ENCB); Huixquilucan, cerca de la Pre- sa El Capulín, Fraccionamiento La Herradura, Rzedowski 25875 (ENCB); Dos Ríos, Rzedowski 35586 (ENCB); Ix- — a, parte alta del Cerro del Wen Rzedowski 29655 NCB); Teoloyucan, Sierra de rosa, Núñez 1718 (ZT, "s Te potzotlán, Teporzotlán, Mia 30 (ЕМСВ); Si- erra 06 Alcaparrosa, Núñez ZTA); parte baja de la ı de Alcaparrosa, 2 km al NN V A oo Rze- e эи shi 29912 (ENCB); parte alta de la Sierra de Alca- rosa, cerca de la estación de * mont: Re 59910 (ENCB): Texcoco, Cerro de ' E de Texcoco, Pulido 305 (C HAPA. bón. Cañón E de la Presa Taxhimay, Rojas & Romero 3315 (IZTA). 2. pos reus frutex Trelease, Mem. Natl. Acad. 20: 82, pl. 120. ee TIPO: México. ени Bourgeau 68 Arbusto rizomatoso de 0.40—2.50 m de alto, a veces es un árbol de hasta 7 m de alto; corteza conformada por escamas cuadrangulares, de color gris opaco; ramillas de 1—1.5 mm diám., densa- mente pubescentes, pubescencia persistente, for- 556 Annals of the Missouri Botanical Garden ` AY j ys Y Ay oue 0 ў 6 SEE SAN vie е : 0.6mm Figura 1. Quercus deserticola. —A. Rama. —B. Fruto. —C. Tricomas. —D. Rama. (A-C: Núñez 1663; D: Camacho 379.) mada por pelos estrellados, de color amarillento a grisáceo; lenticelas blancas, de hasta 1 mm de lar- go, a veces protuberantes y entonces muy evidentes a través de la pubescencia, yemas esféricas a ovoi- des, de 1-3 mm de largo, con escamas pilosas en los bordes; estípulas lineares, de 3—5 mm de largo, pilosas principalmente en la base y en el ápice, glabrescentes con la edad, con frecuencia en las hojas maduras; hojas maduras subcoriáceas, elíp- tico-oblongas, ovado-lanceoladas u oblanceoladas, lámina (1.5-)2 4.56) X (0.5-)1-2(-2.9) ст, ápi- ce redondeado o agudo, mucronado, base redon- deada o subcordada, borde entero, ondulado o con 2 a 4 dientes en las % partes superiores, revoluto; nervaduras primarias de 6 a 11 en cada lado; haz lustroso, con pelos estrellados dispersos, abundan- tes en la nervadura central cerca del pecfolo, ner- vaduras impresas; envés amarillento, pubescente, indumento formado por pelos estrellados estipita- dos, con rayos largos extendidos que dejan ver la epidermis ligeramente ampulosa y densamente pi- losa; nervaduras elevadas: pecíolos de 1—4 X 0.5- | mm, pubescentes; amentos masculinos de hasta 1.5 em de largo, raquis densamente pubescentes; Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México Figura 2. perianto sésil, de 2 mm diám., lóbulos ciliados; an- teras oblongas de 1 mm de largo, filamentos de 1.5 mm de largo: fruto solitario o en pares, sobre un pedúnculo de 3-10 mm de largo: cúpula hemisfé- rica de 7-13 mm diám., con escamas triangulares, pubescentes, excepto en el dorso superior, engro- sadas en la base; bellota ovoide, pared interna del pericarpo glabra, de 5-11 mm diám., incluída %4 de su largo en la cúpula. Figura 2. Reconocimiento. Quercus frutex se reconoce por ser un arbusto rizomatoso, con ramillas densamente pubescentes y hojas pequeñas con el ápice mucro- nado y bordes sin aristas. Puede confundirse con Q. microphylla; por ello es que otros autores la han considerado existente en el Estado de México. Se distinguen porque Q. microphylla posee el envés de las hojas con pelos estrellados sésiles y Q. frutex los presenta estipitados. Distribución y hábitat. En México en los esta- dos de Jalisco, Hidalgo, Estado de México, Mi- choacán y Tlaxcala. Se le encuentra en bosques de Quercus, pastizal y formando manchones densos en matorral xerófilo y pastizal. Se asocia con Pinus, Juniperus, Cupressus y Alnus; también se le en- cuentra en vegetación perturbada, en altitudes de 2360—3000 m. x MEAS ADT LD RIYA Y PES ELE KSSS Quercus frutex.—A. Rama. —B. Fruto. —C. Tricomas. (Román 622.) Fenología. Florece en junio y fructifica de julio a septiembre. Nombres populares y usos. Encino, encino com- pasillo, encino chaparro. Bello y Labat (1987) mencionan que se utiliza para postes de cercas y como leña. Ejemplares examinados. MÉXICO. Estado de Méxi- сө: Atizapán, 3 km al NW de Ciudad Adolfo López Ma- teos, Cruz 620 (ENCB); 4 km al N de Ciudad Adolfo Ló- ‚ López 19 (ENCB); Ciudad Adolfo López Rzedowski 30776 (ENCB): elos, Sierra de Guadalupe. al N de Coa- epec, Rzedowski 15715 (ENCB); Coyotepec, Sierra de Al- caparrosa, Gallardo 3, 4 (IZTA); Chalco, Tlachavote, A. Ventura 3016 (ENCB); El Oro de Hidalgo, Bassoco, Rojas & Romero 3327 (УТА); Huehuetoca, ladera Suroeste del cerro Sincoque, Rojas & Romero 479 (IZTA); ladera E del cerro Sincoque, Rojas & Romero 700, 701, 1623 (ПУТА): Huixquilucan. alrededores de Dos Ríos, Román 326 ENCB): Fraccionamiento La Herradura, Rzedowski 25143 (ENCB): Ixtapaluca, 2 km al E de la Colonia Agrícola Ávila Camacho, Rzedowski 37294 (ENCB); Naucalpan de Juárez, Villa Alpina, Rzedowski 32554 (ENCB); Otumba de Gómez Farías, Cerro Cuixi, Santa Bárbara, Rzedowski 16884 (ENCB); Teotihuacán, Cerro Patlachique, Chavelas ES-1796 (ENCB); San Martín de las Pirámides, ladera N de Cerro Gordo, Castilla & Tejero 388 (IZTA): Tenango de Tepopula, Rancho San Luis Aculco, Hinton s.n. (ENCB): Tepetlaoxtoc, La Venta, Ventura 678 (ENCB): Tepotzotlán. Pp vt m | a Ж D = S 558 Annals of the Missouri Botanical Garden ® Sierra de la Muerta (Sierra de Alcaparrosa), al N Tepotzotlán, Lot & Wendt 143 (CODAGEM): — bon del Cerro de la Cruz, dowski 33270 (ENCB); Cerro de la Cruz, 6 km al — Rzedowski 37055 (ENC B); Texcoco, 20 km al NE de Texcoco, Cruz 1845 ; ; 17 km al E de Texcoco, Charles & Janice e 3397 (С НАРА); 15 km al E de — :0, Muller CB); Cerro a — 7 km al E de Texcoco, p : 466 (ENCB, MEXU); San — Ixayoc, Ventura 845 (ENCB); Ville 1 on Carbón, Cañón E bs la Presa ' himay, Rojas & Romero 3312 (IZTA): mero, S de la Presa del Consuelo, La Coln 28 (IZ TA); 3 km al N de Magú, Rzedowski 16850 2 km al NW de Cahuacán, Azedowski 3: Zumpango, T km al N de Zumpango, Hernández s.n. (ENCB) 3. Quercus glabrescens Bentham, Pl. Hartw. 56: 348. 1840. TIPO: México. Hidalgo: Hartweg 428 (В). Arbol de 10—15 m de alto; ramillas de 1-2.5 mm diám. con lenticelas de color claro, de hasta 1 mm; yemas ovoides de 2.5-3 mm de largo, escamas ova- das con margen ciliado; estípulas oblanceoladas, de 7 mm de largo, glabrescentes, escariosas, persis- tentes en las yemas; hojas maduras subcoriáceas a coriáceas, oblanceoladas, elípticas u obovadas, lámina (3.5-)4-11 X (1.2-)1.7—4(—5.3) em, ápice agudo, base redondeada, borde revoluto, cartilagi- noso, ondulado o con 3 a 5 dientes mucronados a cada lado en la mitad superior de la hoja, mucrones de hasta 1 mm de largo; nervaduras primarias de 10 a 14 en cada lado, ascendentes, casi rectas, las superiores pasan directamente a formar los mucro- nes; haz de color verde obscuro contrastando con el envés verde más claro, glabrescente, con pelos glandulares simples sobre las nervaduras y pelos estrellados sésiles, nervaduras impresas; envés gla- bro o con pelos estrellados sésiles o con un estípite muy corto, ásperos, sobre la nervadura principal, epidermis lustrosa, lisa a papilosa, a veces ligera- mente ampulosa, nervaduras elevadas; pecíolos de 3-10 mm de largo, de 0.5-1 mm diám., dumento formado por pelos estrellados sésiles y ás- con in- peros; amentos masculinos de 2 cm de largo. pe- rianto de 2 mm diám., lóbulos ciliados; amentos femeninos de hasta 3 flores, de 1—1.8 em de largo, raquis glabrescente; frutos solitarios o en grupos de 2 a 3, sésiles o sobre un pedünculo 4—5 mm de largo; cúpula hemisférica, de 1.3-1.7 ст diám., con las escamas pubescentes, las superiores adpre- sas, ápice agudo o redondeado, bases engrosadas; bellota ovoide, pared interna del pericarpo glabra, de 1.5-1.8 cm de largo, de 1.2-1.3 cm diám., cluída en la cúpula un tercio de su largo. Figura 3. in- Reconocimiento. noce porque sus ramillas у el envés de las hojas Quercus glabrescens se reco- poseen escasa pubescencia, y porque el borde de ésta tiene dientes mucronados ubicados en la parte cal. Distribución y hábitat. dos de Distrito Federal, Hidalgo, Estado de México, Oaxaca, San Luis Potosí, Tlaxcala y Veracruz. En bosque de Pinus, Pinus-Quercus y en bosque me- sófilo de montaña, se asocia con Alnus, Abies, Fra- en altitudes de 2450— ap — =. En México en los esta- Clethra y Cornus, Florece de febrero a junio y fructi- Encino, encino blan- No se conocen usos de esta especie; sin embar- go, De la Paz (1985) la recomienda para la elabo- ración de mangos de herramienta, pisos industria- les, cercas, tarimas, construcciones pesadas y para construcciones en general donde se requiera resis- lencia. Ejemplares examinados. MÉXICO. Estado de Méxi- co: oo 414 (EN CB): юн de ‘la Joya A Ale lealica. | s.n. (CHAPA, ENCB, MEXU); Ozumba de Alzate, Techi- ies o, W del Popoc atepetl, Boyas 576 (MEXU); ' ехсосо, 1 al SE de San Pablo Ixayoc, Rzedowski 24171 (Е Pablo Ixayoc, Е: X l); Cañada al SE de San Rafael, Barrios 19 (CHAI ^. ENC B. AEXU); alrededores s, Rzedowski 31890 (CHAPA, ENCB, 1 km al E de San Rafael E. Rzedowski 32452 eid ); base del Iztaccfhuatl, 4 km al SW de San Rafael, Vera 114 (ENCB); Villa del Carbón. San Jerónimo, Martínez 39015 (MEXU — 4. Quercus glaucoides M. Martens & Galeotti, Bull. Acad. Roy. Sci. Bruxelles, Vol. 10, pt. 1: 209. 1843. TIPO: México. Oaxaca: Galeotti 103 (BR). Arbol de 4—10 т de alto, con el tronco de 25—40 cm diám., corteza gris; ramillas de 1-3 mm diám., de color castafio-rojizo, glabras o pilosas cerca de la base de las yemas, con lenticelas de 0.5-1] mm de largo, protuberantes, pálidas; yemas de 1-2.5 mm de largo, ovoides, obtusas; escamas ciliadas de color castafio, con los bordes papiráceas, ápices obtusos; estípulas de 4—5 mm de largo, lineares, pilosas en los márgenes; hojas jóvenes muy delgadas, no glau- cas, haz y envés con abundante pubescencia for- mada por tricomas estrellados largos, antes de la ma- durez las hojas son glabras; hojas maduras glabras o con algunos tricomas cerca de la base, glaucas, lámina (8—)10—13(—15) X (3.5—)4—7.5 em, coriáceas, oblanceoladas, elípticas u obovadas, ápice obtuso, Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México Figura 3. redondeado o retuso, a veces subagudo, base cor- dada, redondeada, a veces ligeramente oblicua, bor- de engrosado, cartilaginoso, plano o ligeramente re- voluto, entero, con ondulaciones o con 4 a 7 dientes anchos y obtusos de cada lado; nervaduras de (4)7 a 12 de cada lado, ascendentes, arqueadas o casi rectas, se ramifican cerca del borde de la hoja; haz glabro, de color verde grisáceo, algo lustroso, ner- vaduras rojizas o amarillentas, la central y primarias ligeramente elevadas; envés glabro o con algunos tri- comas cerca de la base, epidermis glauca-cerosa, papilosa, blanquecina, nervaduras elevadas; pecíolos Quercus glabrescens.—A. Rama. —B. Fruto. —C. Tricoma. —D. Hoja. (A-C: Robledo 3; D: Chen 2.) de 2-8 mm de largo, de 1-2 mm diám., de color castaño. rojizo o negro; flores desconocidas; fruto anual, en pares o en grupos de tres, sésiles, o en pedúnculos de hasta 6 cm de largo; cúpula hemis- férica, de 9-16 mm diám., escamas adpresas, en- grosadas en la base, con abundante pubescencia, ápices triangulares; bellota ovoide, pared interna del pericarpo glabra, de 8-12 mm de largo, de 8-9 mm diám., incluída un tercio de su largo en la cúpula. Figura 4. Reconocimiento. Quercus glaucoides se reco- Annals of the Missouri Botanical Garden Figura 4. noce por sus hojas de color verde-azuloso con el envés glabro y glauco-acerado. Кагаз veces, mues- tra similitud con Q. obtusata, debido a que esta tiltima puede presentar el envés de las hojas glauco y casi glabro, y los pecíolos cortos de color obscuro, características muy frecuentes de Q. glaucoides; se distinguen porque ésta presenta frutos ovoides, mientras que О. obtusata presenta por lo general, frutos globosos. Distribución y hábitat. En México en los esta- dos de Jalisco, Michoacán, Estado de México, Gua- najuato, Querétaro, Hidalgo, Morelos, Puebla. Gue- rrero, Oaxaca. En bosque de Pinus-Quercus, se le encuentra asociado con Pinus pringlei, Quercus conspersa, Q. castanea y Q. obtusata, en altitudes de 750-1800 m. Fenología. | Fructifica de junio a agosto. Quercus glaucoides.—A. Rama. —B. Fruto. (Romero 4031.) Nombres populares y usos. Tocuz, encino blan- co, encino negro, encino roble, encino y roble. = Se le usa como leña, para horcones, manufactura de algunos implementos agrícolas y posiblemente por el área de distribución se explote como material celulósico (González, 1986); también se emplea para elaborar carbón, cabos para herramientas, puertas de golpe y postes para cercas (Bello & La- yat, 1987). Las hojas y frutos se usan como forraje: las primeras se emplean como medicinales (Váz- quez, 1992). Ejemplares examinados. MÉXICO. Estado de Méxi- co: Ocuilan de Arteaga, 1 km al S de Chalma, Muller 9220 (MEXU): Santo Tomás de los Plátanos, La Junta, Martínez 29358 (CODAGEM, MEXU); Sultepec, km 18 carretera Sultepec-San Miguel Totolmoloya, Torres 384 (IZTA); km 12 carretera Sultepec-San Miguel Totolmolo- Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México ya, Torres 388 (17ГА); Tejupilco, Cuadrilla de López, Salto del / gua, — 6 е esa de Nanchititla, Mar- tinez Sont (MEXU); Nanc кещ Rodríguez 201 (INIF): Tena de González, Luvianos, Breedlove s.n. (MEXU); Cerro de la Culebra, Luvianos Progreso, Matuda 31472 (CODAGEM, MEXU); Tlatlaya, Llano Tlatlaya, Martínez 30151 (MEXU); Valle de d Boege s.n. (МЕХ); V tínez 2609 (INIF); Valle de Bravo, | . (MEXU): Tiloxtoc, San Bartolo NW de Valle de [A “Muller 9058 (MEXU); Zumpahuacan, cerca de San Gaspar, Tejero & Castilla 2862 (IZTA). Hincón Grande у Valle de Bravo, Valle ul apt Mar- Martínez Quercus laeta Liebmann, Overs. Kongel. Danske Vidensk. Selsk. Forh. Medlemmers Arbeider: 179. 1854. TIPO: México. Casa Grande, Hartweg 419 (B). Quercus — Trel., Mem. Natl. Acad. Sei. 20: 61, pl 60. . TIPO: México. Distrito Federal: Contreras, — poe 5c (B) Árbol de 6-10 m de alto, con el tronco de 25- 40 em diám.; ramillas de 1-2(-3) mm diám., roji- zas, glabrescentes, con indumento formado por tri- comas estrellados y simples largos, con el tiempo se ennegrece, lenticelas menores de 0.5-1 mm, pá- lidas, en general no muy abundantes; yemas de 2— 4(—5) mm de largo, ovoides, de color castaño, es- camas con el dorso pubescente y los bordes cilia- dos: estípulas 4—6(—7) mm de largo, lineares, pu- bescentes, persistentes en las yemas apicales; hojas jóvenes de color verde-grisáceo, haz con tricomas estrellados dispersos, más abundantes en la ner- vadura central, envés muy pubescente, pubescen- cia formada por tricomas estrellados con sus ramas extendidas; hojas maduras de color verde obscuro, coriáceas, decíduas, elípticas, oblonga-lanceoladas, oblanceoladas u obovadas, lámina 5.5-12(-14) (2-)3-5.5(-8.5) cm, ápice obtuso o agudo, base re- dondeada, cordada, a veces cuneada, borde revo- luto, engrosado, ondulado o dentado, a veces en- tero, con 5 a 7 dientes en cada lado, en las dos terceras partes superiores de las hojas, dientes ob- tusos, con mucrones engrosados y curvos; nerva- duras de 7-11(—12) en cada lado, ascendentes, cur- vas о casi rectas, se prolongan hasta los dientes. y se dividen en la mitad o cerca del borde de la hoja: haz lustroso, glabro o con tricomas estrellados uni- formemente distribuidos y tricomas simples en ner- vadura central, ésta principalmente pubescente en su base, nervaduras impresas, envés pubescente con tricomas estrellados largos, por lo general con sus ramas extendidas, a veces se enredan un poco en su base, estípite corto, tricomas glandulares es- casos; epidermis ligeramente ampulosa, papilosa, pecíolos de (2-)5-10(-13) mm de largo, de 1-2.5 mm diám., pubescentes, en- nervaduras elevadas; sanchándose en la base; flores desconocidas: fruto аши, solitario о en grupos de 3, еп pedtinculos de .5(—4—7) em de largo; cúpula hemisférica de 8— 11 cm diám., escamas pubescentes, principalmente en la base, ésta engrosada, ápices triangulares, ob- tusos; bellota ovoide, de 6-13 mm de largo, de 6— 7 mm diám., con una tercera parte o la mitad de su largo incluída en la cúpula. Figura 5. Reconocimiento. Quercus laeta se reconoce por sus hojas con el borde mucronado v envés con pu- bescencia formada por tricomas estrellados con las Esta muestra similitud con Q. obtusata v se distingue ramas extendidas y estípite corto. especie porque posee en el envés de las hojas abundantes tricomas glandulares y escasos tricomas estrellados con las ramas enredadas entre sí; además el tamafio de sus hojas es mayor. Distribución y hábitat. En México en los esta- dos de Aguascalientes, Coahuila, Distrito Federal. Durango, Guanajuato, Hidalgo, Jalisco. Estado de México, Michoacán, Nayarit, Nuevo León, San Luis Potosí, Sinaloa y Zacatecas. En bosque de Quercus, bosque de Pinus—Quercus, en matorral xerófilo y encinar secundario, se asocia con Q. obtusata, Q. castanea, Pinus, Abies, Cupressus, Arbutus, Fraxinus, en altitudes de 2350-2750 m. Fenología. Alnus y Florece en febrero y fructifica de julio a diciembre. Vombres populares y usos. Encino colorado. Su madera resulta difícil de trabajar, aunque se usa para la elaboración de algunas herramientas: mas no se considera buena para la elaboración de muebles. También se utiliza como leña, carbón pos- tería, horcones, cercas y como material celulósico (González. 1986). Ejemplares — ч ICO. Estado de Méxi- co: Atizapál 3. Zar LIA Rzedow ski 3258. (ZT А); Тар tepec, km 67 carretera ral ne ron dtp . Cisneros 5 (IZTA); Denhxi, Martínez 64 ( Tenango del Valle, Rancho Yeca, s.n т. Ton Sierra de 1574 (VZTA Nicolás Rome ro-T lazala de F Quintero 36 ПУТА): km 28 carretera a Tlalnepantla del Carbón, Rojas & = 5165 (IZ TA): Barrio IV, Rojas & Romero 3192 (Z Quercus magnoliifolia Née, Anales Ci. Nat 9: 208. TIPO: México. Guerrero: entre Chilpancingo y Tixtla, Née s.n. (MA). 1801 (como magnoliaefolia). Quercus macrophylla Née, Anales Ci. Nat E 274. 1801. TIPO: Entre € * ша y Tixtla, Née s. a ace ат МЕ t E. E Warb., Bull. Mise. : 85. IPO: México. К эе, de México: — 6360 562 Annals of the Missouri Botanical Garden Figura 5. Árbol de 3-15 m de alto, tronco de 20-50 ст diám.; ramillas de 2—4 mm diám., glabrescentes, con tricomas estrellados y glandulares cortos, éstos de color ámbar-rojizo; lenticelas muy notorias, pá- lidas, menores de 0.5-1.5 mm de largo; yemas de 1.5-5 mm de largo, ovoides, de color castafio, grue- sas, pubescentes, principalmente en los márgenes: estípulas filiformes, de 2-10 mm de largo, pubes- centes; hojas jóvenes rojizas por la abundancia de tricomas glandulares, con tricomas estrellados muy cortos, dispersos; envés más pálido; hojas maduras, por lo general obovadas o elípticas, lámina (7—)10— 27(-32) X 6-12.5(-21) em, ápice obtuso, redon- deado o agudo, base atenuada, cuneada, redondea- da o auriculada, borde engrosado, ligeramente re- dentado o sinuado-dentado, voluto, sinuado, dientes 8 a 14 en cada lado, con un mucrón corto Quercus laeta.—A. Rama. —B. Fruto. —C. Tricomas. (Urbina 4.) y doblado en el ápice de cada diente u ondulación; nervaduras primarias 11 a 20 en cada lado, ascen- dentes, rectas o ligeramente curvadas, pasan direc- tamente a formar los mucrones, a veces nervaduras secundarias los forman; haz verde lustroso, glabro, con tricomas estrellados y glandulares en la base de la nervadura central, nervaduras impresas, las secundarias y más pequefias poco notorias; envés verde-amarillento, más claro que el haz, con pu- bescencia abundante о escasa, formada de tricomas glandulares y estrellados sésiles muy cortos, adpre- sos a la epidermis, ésta ampulosa y papilosa, ner- vaduras elevadas; pecíolos de 5-10 Х 1—4 mm, más anchos de la base, pubescentes, principalmen- te por el envés de las hojas; amentos masculinos con numerosas flores, de 6—10 cm de largo, perian- to membranoso, largamente ciliado, de 3—4 mm Volume 89, Number 4 Romero et al. 563 2002 Quercus en el Estado de México у Figura 6. Quercus magnoliifolia.—A. Rama. —B. Fruto. —C. Tricoma. (Soto 283.) diám., anteras de 10 a 12, de 1-1.5 mm de largo; gadas cuando maduras, de una a dos terceras partes flores femeninas solitarias, en grupos de 2 o З ó de su largo incluída en la cúpula. Figura más, dispersas sobre un pedúnculo de 5-10 em de largo, fruto anual, la cúpula de 14-25 cm diám., Reconocimiento. Quercus magnoliifolia se re- escamas obtusas pubescentes, engrosadas en su conoce por sus hojas generalmente grandes, obo- dorso, ápice más delgado, angostamente triangular: vadas o casi elípticas, con mucrones, haz verde lus- bellota ovoide, pared interna del pericarpo glabra, troso y nervaduras casi rectas. de 8-22 mm de largo, de 7-15 mm diám., arru- Distribución y hábitat. En México en los esta- 564 Annals of the Missouri Botanical Garden dos de Colima, Guerrero, Jalisco, Nayarit, bosque de Pinus, Quercus y Pinus—Quercus, se aso- Estado de Méxi- co, Michoacan, Oaxaca y Sinaloa. En cia con Pinus oocarpa, Quercus hintonii y Q. ellip- lica; se le encuentra también en encinares pertur- bados, en altitudes de 1700—2600 m. Fenología. Florece en febrero y fructifica en junio y julio. Nombres populares y usos. Encino, encino ro- ble, roble, encino amarillo, encino napis, encino prieto, encino blanco, encino bermejo, encino ave- llano. Su madera se utiliza como leña, carbón, postería, horcones, para fabricar mangos para herramientas, bancos, vigas, postes para corral y para extraer ce- lulosa. Su follaje sirve para fabricar techos rústicos de viviendas del campo y sus frutos como forraje para cerdos (González, 1986; Bello € Labat, 1987; 'ázquez, 1992). MÉXICO. Estado de Méxi- co: Amatepec, Amatepe cercanías, Matuda 29822 (CODAGEM); Clachic hhilpan, 7 km al NE de Amatepec, 1049 (INIF); Las Trojas, Rodriguez 185 —— Ixtapan del Oro, cerro ubicado al SW de la presa Tiloxto d 850 (INIF); Cerro el € vin nco, al E de la — Valle de аа Estrada 852 (INIF); Ocuilan де Artea- ga Ex Arteaga, cerro la — Nah 1179 (INIF); Nuevo Santo Tomás de los — »8. La Junta, Matuda 29356 (CODAG EM); Sultepec, Las Tinajas, Кейш s.n. (СНА- РА, 12 carre s ra Sultepec-San Miguel Totol- moloya, Torres 90, 110, 194 (IZTA); km 18 carretera Sul- 2 (IZTA); Tejupilco, Nanchititla, JAG EM): Canada de Nanchititla, García 454 (CODAG En ENCB); Nanchititla, García s.n. (INIF); proximidade % de Tenería, Guízar 155 (CHAPA, ENCB); camino de Almoloya de la Granadas, Guízar 351 (ENCB); Cañada de Nanchititla, Soto е Moreno 183, 283 (IZTA); Popup a ra бе zález, 70 carretera federal 104, : km 65 de 1 Ejemplares — — Castilla & Tejero 1796 (VZT hs km 3-4 carretera Zacual- pan-Mamatla, Fragoso 339 (17ГА). 7. Quercus obtusata Humboldt € Bonpland, Pl. Aequinoct. 2: (26). Pl. 76. 1809. TIPO: Méx- ico. Michoacán: Ario de Rosales, Bonpland 4329 (isotipo, B!). Bull. Misc. Inform.: 88. 1939. Quercus atringlans Warb., TIPO: México. Estado de México: Temascaltepec de González, Hinton 6549 (K). Arbol de 6-20 m de alto, tronco de 40-60 em diám. o más, con la corteza gris, escamosa; ramillas —)2—3 mm indumento formado por tricomas glandulares glabrescentes, rojizas a gris o negras, de (1 diám., y estrellados, lenticelas pálidas, de hasta 2 mm de argo; yemas ovoides, de (1.5—)2—4(—5) mm de largo, de color castafio obscuro, escamas pubescentes en los márgenes; estípulas lineares, de 5-8 mm de lar- go, membranosas, pubescentes; hojas jóvenes con el haz rojizo por la abundancia de los tricomas glan- dulares, envés amarillento, densamente pubescente, indumento formado por tricomas estrellados entre- lazados; hojas maduras decíduas, gruesas y coriá- ceas, rugulosas, obovadas a — obovadas о elípticas, lámina (4—)6-17(-22) х (2—)3—8(—11) em, ápice obtuso o anchamente м а veces algo agudo, borde engrosado, revoluto, dentado, sinuado o dentado-sinuado, con З a 9 dientes u ondulaciones, que frecuentemente se distribuyen desde el ápice hasta la base de la hoja, cada diente u ondulación termina en un mucrón corto, que se dobla junto con el borde revoluto, nervaduras de 7 a 12 en cada lado, ascendentes, rectas o ligeramente curvas, cada nervadura secundaria pasa a formar un mucrón que coincide con el ápice del diente u ondulación, a veces el mucrón se forma entre ondulaciones; haz verde lustroso, glabrescente, con tricomas simples y estrellados dispersos, más abundantes en la base de la nervadura central, nervaduras primarias impresas o ligeramente protuberantes; envés verde-amarillen- to, opaco, con pubescencia dispersa formada por tri- comas estrellados con estípite muy corto, a veces muy escasos, de aproximadamente 8 rayos, enreda- dos entre sí o algo extendidos, de pocos a abundan- tes tricomas glandulares color ámbar o rojizos, a ve- ces se forman gotas de exudado, epidermis ligeramente ampulosa y papilosa; pecfolos glabres- )-11(-15) mm de largo, de 1-2 mm centes de (3— diám., a veces de color obscuro; amentos masculinos de 3 cm de largo, con muchas flores distribuidas a lo largo del raquis, perianto de 2 mm diám., larga- mente pubescente, 6 estambres, anteras de 1 mm de largo, filamento de 1 mm de largo; amentos feme- ninos de 3 a 6 o más flores distribuidas en la mitad distal de un pedúnculo de 2-3.5 cm de largo, pu- bescentes; fruto anual, solitario o en grupos de 2 o 3 ó más, pedünculos de 1.8—3.5 mm de largo; cú- pulas hemisféricas, de poco a muy profundas, de 12-18 mm diám.; escamas muy pubescentes, ápice agudo, bases engrosadas; bellota globosa, a veces cilíndrico-ovoide, con la pared interna del pericarpo glabra, de 6-20 mm de largo, de 11-19 mm diám., hasta un tercio de su largo incluída en la cúpula. Figura 7. Reconocimiento. | Quercus obtusata se reconoce por sus hojas obovadas, con mucrones robustos que se doblan hacia el envés, éste con abundantes tri- comas glandulares y escasos tricomas estrellados con las ramas enredadas entre sí; las bellotas son globosas. Quercus obtusata muestra similitud con Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México 565 Figura 7. Q. laeta la cual se reconoce por sus hojas con el borde mucronado y envés con pubescencia formada por escasos tricomas glandulares y abundantes tri- comas estrellados con las ramas extendidas y estí- pite corto. Distribución y hábitat. En México en los esta- dos de Guanajuato, Hidalgo, Jalisco, Estado de México, Michoacán, Morelos, Nayarit, Puebla, San Luis Potosí y Zacatecas. En bosques de Pinus. Quercus obtusata.—A. Rama. —B. Frutos. —C. Tricomas. —D. Hoja. (Torres 195.) Quercus y Pinus-Quercus, se asocia con Pinus leiophylla, P. montezumae, P. michoacana, Cupres- sus, Quercus candicans, Q. urbanii y Q. glaucoides; también se le encuentra en bosque mesófilo de montaña y pastizal con matorral xerófilo de Acacia y Opuntia. Es frecuente en encinares perturbados. en altitudes de 1430-2850 m. Fenología. | Florece de abril a mayo y fructifica de agosto a octubre. 566 Annals of the Missouri Botanical Garden Nombres populares y usos. Roble, encino, en- cino prieto, encino calicahuac, encino negro, en- cino cosahuicahuatl, encino blanco, encino roble, encino rojo, encino chino, tocuz, uricua, charari. Se utiliza para leña, carbón, postes para cerca, implementos agrícolas, horcones, cabos para herra- mienta, curtir pieles; la corteza tiene usos medici- nales, y como material de construcción (Bello & Labat, 1987). —— duri dus an MÉXICO. Estado de Méx Acambay, c ondingá, Estrada 1443 (IZTA); Almo- loya de Andrés, Rodríguez s.n. (INIF): Coatepec de Harinas, 4 km al E de Coatepec de Harinas, artínez 2697, 2840 (INIF); Donato Guerra, El Mirador de Donato Guerra, Toluca-Valle de Bravo, Rojas & Ro- mero 3242 (IZTA); Ocuilan de Arteaga de Arteaga, Loma de Fuego, Gutiérrez 403 dd cerro de La Llovizna, Nah 1180, 1181, 2340 (INIF); $ ded INIF); km 12 carretera Suhepee-San Miguel Totol- — — 195 (IZTA); Tejupilco, N pm 200 (CHAPA, INIF); Temascalte pec de ( Ype ricones, Hinton 6549 (INIF); Estancia Vieja, 10 S de Temascaltepec de González, Moreno 168 (INIF); Real г Arriba, Rocha s.n. 0 NIF); Tenancingo, Las Cumbres, Rodriguez 31 (INIF); Las Cumbres, Rodriguez s.n. (CHA PA, INIF); Te mange de n popula, Rancho San Luis Acul- co, Hinton 17989 ; Tepozotlán, Cerro de pie Ca- be маз, Madrigal & Vela s s.n. (INIF); Texcoco, 1 km al SE de San Pablo Ixayoc, zedowski 24193 (INIF); Valle de Bravo, 2 km al E de Pipioltepec, Pineda & Ochoa о (INIF); Godines Tehuaste — Rodriguez 205, 206 (IN Villa del Carbón, km 28 carretera Tlalnepantla—Villa de i — Rojas & Romero 3165 — Las Golondrinas, a 1000 m del límite con Hidalgo, Rojas & — 3306. 3310 (IZTA); Villa Nicolás Romero, 3 km al W de Ca- huacán, Rzedowski 32631 (INIF); Zacualpan, Subestación Coronas, Jiménez 13 (IZTA — 8. Quercus peduncularis Née, Anales Ci. Nat. 3: 270. 1801. TIPO: México. Acapulco, Née s.n. (MA). Entre México v Querc P — — E. E Warb., AD Misc. — : 87. . TIPO: México. Estado de México: Temascal- ài pec doi Conzale ‘z, Hinton — (holotipo, K no vis- to). Árbol de 4 m de alto; ramillas de 2-2.5 mm diám., glabrescentes, en un principio con un denso tomento formado por tricomas estrellados estipita- dos con rayos largos; lenticelas visibles sólo en ra- millas viejas, de color claro, de hasta 1 mm de largo; yemas de 3—4 mm de largo, ovoides, con las escamas obtusas, ciliadas en los bordes y dorso; estípulas de 5-7 mm de largo, subuladas o lineares, con tomento largo o glabrescente, persistentes aún en las hojas maduras; hojas jóvenes muy delgadas, de color verde obscuro, haz con fina pubescencia formada por tricomas estrellados cortos, algunos argos, envés densamente pubescente, de color cla- ro; hojas maduras subcoriáceas, oblanceoladas, elípticas u obovadas, lámina (4-)7-10 X (2.5—)3— 5 cm, ápice obtuso o agudo, base atenuada, redon- deada o subcordada, a veces asimétrica, borde car- tilaginoso, revoluto, casi siempre dentado, con 7 a 10 dientes mucronados a cada lado, desde la base de la hoja, nervaduras 9 a 14 en cada lado, ascen- dentes, casi rectas, pasan directamente al diente, se ramifican cerca del borde; haz lustroso, de color verde obscuro, glabro, excepto la base que presenta tricomas estrellados parecidos a los del envés, ner- vadura central algo convexa, las primeras impresas; envés con indumento blanquecino formado por tri- comas estrellados estipitados, los rayos largos y extendidos que se entrecruzan con otros, epidermis ampulosa y papilosa, nervaduras pálidas, convexas; pecíolos de 2 cidas; fruto anual, solitario o en pares sobre pe- mm diám.; —5 mm, pubescentes; flores descono- dúnculos de 3.5—4 em de largo, de cúpula hemisférica de 12-13 mm diám., pubescentes, sus ápices glabrescentes, agudos o escamas acuminados; bellota ovoide, pared interna del pe- ricarpo glabra, de 8-13 mm diám., de uno a dos tercios incluida en la cúpula. Figura 8. Reconocimiento. Quercus peduncularis es una especie muy escasa en la entidad; sin embargo se encuentran comunidades en donde esta especie es abundante. Se reconoce porque el envés de sus ho- jas es tomentoso y blanquecino y los pedúnculos de los frutos son largos. Distribución y hábitat. Еп México en los esta- dos de Colima, Chiapas, Guerrero, Jalisco, Estado de México, Michoacán, Oaxaca y Veracruz, también en América Central. Bosque de Pinus—Quercus, se asocia con Alnus, en altitudes de 2550 m. Fenología. | Fructifica en julio. Nombres populares y usos. Encino, roble, en- cino blanco, encino avellano, encino zopilote negro, mazcahuite. La madera se emplea como carbón, lefia y ma- terial celulósico para papel (González, 1986). Ejemplares examinados. MÉXICO. Estado de Méxi co: Te mascaltepec de González, San Lucas, Hinton 6376 Te mascaltepec de González, Hinton 67331 (MEXU): NE de San Francisco Oxtotilpan, Orozco 268 (IZTA). — — Quercus rugosa Née, Anales Ci. Nat. 3: 275. 1801. TIPO: México. Estado de México: Huix- quilucan y Ocuilan de Arteaga, Née s.n. (MA). Acad. Sci. 20: 77, Tlalpu- Mem. Natl. : México. Michoacán: Quercus conglomerata Tre ^ pl. 106. 1924. ТІР xahua, Hartweg 429 е ). Quercus decipiens Mart. & Gal., Bull. Acad. Roy. Sci. Bru- xelles, Vol. 10, pt. 1: 214. 1843. Syn. nov. TIPO Méx xico. Mirador, Galeotti 131 (BR). Volume 89, Number 4 Romero et al. Quercus en el Estado de México 567 Figura 8. Quercus reticulata Humb. & Bonpl., Pl. Aequinoct. 2: 40(35). pl. 86. 1809. TIPO: México. Guanajuato: Santa Rosa, Bonpland 4408 (B). ` Arbol de (5—)10—30 m de alto, con el tronco de hasta 1 m 6 más diám.; ramillas gruesas, de 1.5 mm diám., glabrescentes, de color gris-castaño, tri- comas del indumento estrellado-estipitados, que se ennegrecen con el tiempo; lenticelas hasta de 1 mm, en las ramas viejas más grandes y protube- rantes; yemas de 2-5 mm de largo, ovoides, de col- or café-rojizo, escamas coriáceas, pubescentes; es- típulas lineares u oblanceoladas, decíduas. de 5—7 Ç: К NES - "уо; КАЛД убу ЖОНИ oo К a) 0.4mm Quercus peduncularis.—A. Rama. —B. Fruto. С. Tricomas. (Hinton 6331.) mm de largo, piloso-sedosas; hojas jóvenes con el haz rojizo y abundantes tricomas glandulares, vena principal muy pubescente, envés con densa pubes- cencia pálida, epidermis totalmente cubierta de tri- comas glandulares amarillentos, nervadura con tri- comas glandulares rojizos; hojas maduras deciduas, coriáceas, rígidas, con frecuencia cóncavas, muy - rugosas, elípticas, elíptico-obovadas, obovadas « casi suborbiculares, lámina (3.2-)5-15(-19) X 2— 8-10) em, ápice obtuso, a veces agudo, base re- dondeada o cordada, borde sinuado o dentado, en- grosado, plano o revoluto, dientes de 5 a 12 en cada 568 Annals of the Missouri Botanical Garden lado distribuidos en las % partes superiores, agudos u obtusos, con mucrón corto hasta de 1-2 mm de largo; nervaduras de 7 a 12 en cada lado, rectas o curvadas, pueden pasar directamente al mucrón о ramificarse antes y las ramificaciones pasar a los mucrones adyacentes; haz algo lustroso, verde o grisáceo, glabro o con pocos tricomas estrellados y tricomas glandulares, principalmente en la nerva- dura media, nervaduras central, primarias y secun- darias impresas, las más finas elevadas formando un retículo pálido; envés densamente pubescente o con tricomas esparcidos; pubescencia formada por tricomas estrellados con un estípite corto, sus rayos enredados, tricomas glandulares vermiformes de color ámbar o rojizos; epidermis ligeramente o mar- cadamente ampulosa, papilosa; pecíolos pubescen- tes, de (35-10-13) mm de largo, de 1-3 mm diám., rojizos; amentos masculinos de 15-20 mm de largo, pubescentes, perianto campanulado, de 2 mm diám., bordes ondulados y largamente ciliados, anteras glabras; amentos femeninos con 15 a 20 flores en pedtinculos pubescentes; fruto anual, so- litario o en grupos de 2—5(-8), en pedúnculos de (0.8-)1.5-5(-8) cm de largo; cúpula hemisférica, de 8-15 mm diám., engrosadas en la base, pubescentes en el dorso, de de 6-10 mm de alto, escamas color café rojizo; bellota ovoide o angostamente elíptica, pared interna del pericarpo glabro, ápice agudo de 9-28 mm de largo, de 7-14 mm diám., incluída una tercera o hasta la mitad de su largo. Figura 9. Reconocimiento. Quercus rugosa se reconoce por sus hojas coriáceas y cóncavas, los mucrones largos y envés con tricomas glandulares y depósitos de mucílago. Quercus rugosa puede confundirse con Q. obtusata; ésta última se distingue por la presencia de mucrones callosos en los márgenes de las hojas, mientras que Q. rugosa los presenta pro- minentes de hasta 2 mm de largo. Distribución y hábitat. En México en los esta- dos de Aguascalientes, Coahuila, Chihuahua. Dis- trito Federal, Durango, Guanajuato, Hidalgo, Jalis- co, Estado de México, Michoacán, Puebla, Veracruz y Zacatecas. En bosques de Pinus—Quercus, de Pi- nus, de Quercus, de Abies, en matorral xerófilo, en encinares perturbados y cultivos agrícolas; se aso- cia con Pinus letophylla, P. oocarpa, P. teocote, P. pseudostrobus, Quercus castanea, Q. candicans, Q. laurina y Q. altitudes de 1700- 3500 m. Fenología. crassipes, en Florece en agosto y fructifica de septiembre a noviembre. Nombres populares y usos. Roble, doga, encino, encino roble, encino quebracho, encino hojarasco, encino negro, encino cuero, encino blanco, encino asta, encino avellano, encino miel, encino blanco liso, encino quiebra hacha. La corteza se utiliza en el tratamiento de la di- sentería, dolor de muelas, hemorragias y fortaleci- miento de los dientes; usada junto con las hojas sirve para preparar la infusión que se utiliza para el tratamiento contra el dolor muscular, la tos (Chi- no & Jacques, 1986). Se sabe también que las be- llotas sirven para elaborar café o pueden consu- mirse tostadas; también se usan como forraje (Vázquez, 1992). La madera se utiliza para la ob- tención de papel y como leña (González, 1986). Ejemplares examinados. MÉXICO. Estado de Méxi- co: Acambay, cerro Detiña, Estrada 1467 (IZTA); Aculco de Pyle km 117 carretera a Toluca, Rojas & Romero 2150 (IZTA); Axapusco, Cerro del Tepayo, Jaltepec, Ven- tura 675 (INIF); Amanalco de Becerra, San Jerónimo Amanalco, Ventura = (ENCB); alrededores de Amanalco le Becerra, Villa 163 (INIF); Amecameca, San Antonio Zoyatzingo, Avila s „ (МЕ); cerro 7 — Román 109 (IZTA); pee cerro Chiluc › km al SW de Ciudad Adolfo López Mateos, Райла: к (С HAPA, INIF); Coyotepec, Sierra de Alcaparrosa, Fernández 134 (ПУТА); parte alta de la Sierra de Alcaparrosa, 6 km al W de Coyotepec, Rzedowski 37251 (ENCB); Chapa de Mota, cercanías del — astronómico, Rojas & Romero 2796 (IZTA); Donato Guerra, Mirador de Donato Guerra, Rojas & ау 3243 (IZTA); Ecatepec de Morelos, Sierra de Guadalu wi Suárez 11 (VZTA); — — Matuda 29023, 29026, 29029 (CHAP: - e dalgo, La Cima, Gómez 1 (IZTA); cercanías p BE Hidalgo, Matuda 28621, 32613 (CODAGEM); ы Jilotepec, Matuda 26668 (CODAGEM); cerro de Jilotepec, Matuda 27821 (CODAGEM); Llano Grande, 6 km a de Canalejas, Rojas & Romero 3142 (IZTA); Santa Ana e Las Manzanas, Hernández 13 (IZTA); Jiquipil- 21 carretera Naucalpan—Jiquipilco, (ZT A): Ocuilan de Arteaga, Ocuilan de зправа: Martínez 31748 (IZTA); cerro de La Llovizna, Nah AM-5 (INIF); — de Gómez Farías, Rancher i Buenavista, 8 km al К de San Marcos, Fuentes АШ-32 y АШ-33 (MEXU); San n de las Pirám dr cerro Gordo, € wes E Tejero s.n. (ENCB): Sultepec, Real de Abajo, Машаа 29 м APA); Las Trojes, — 180 (INIF); енен erro Gordo, Castilla & Tejero 604 (IZTA); Pagos de popula, cerca de Amecameca, Matuda 18839 (CODA- E GEM): Teoloyucan, Sierra de Alcaparrosa, Núñez 43 (IZTA); эж к Sierra de Alcaparrosa, Núñez 2176 (IZTA); Texcoco, 28 km al E de Texcoco, sobre brecha maderera hacia Tlaloc, García s.n. (INIF); 10 km al E de Texcoco, al principio de la brecha maderera, García s.n. A); cerro Tetzc ‘ultzingo, 8 km al E de Texcoco, Pu- lido 428 (ENCB); 13 km al E de Texcoco, carretera a San Dieguito, Ricafo 28 (CHAI A): Santa María Tecuanulco, Vengura 877 (CHAPA, Ps В); Timilpan, Sierra de San André Cam wacho 288, 397 (IZTA); Tlalmanalco, x de Tialmanaleo Medellín n. (INIF);; 3 km al SW de San Rafael Tlalmanalco, — 1033 (ENCB); Tlazala de Fabela, 2 km al SE de Tlazala de Fabela, Román 289 (ENCB); Valle de ee entre Amanalco de Becerra Valle de Bravo, Matuda 28628, 28630 (MEXU); San EA San José Allende, Rojas & Ro- ~ T Alle ande, alre de dores de S Romero Volume 89, Number 4 et al. 2002 Quercus en el Estado de México 569 Figura 9. Quercus rugosa. —A. Rama. —B. Fruto. —C. Tricomas. —D. Rama. (A-C: Román 295; D: Camacho 60.) mero 2057 (ZTA); Villa del Carbón, Predio Piedra Azul. 215. 1801. TIPO: México. Guerrero: cerca de Ávila s.n. (INTE); Villa del Carbón. Matuda 693 (CHAPA); Tixtla, Née s.n. (MA) cercanías de Villa del Carbón, Matuda s.n. (CHAPA); Vil- — к id la Victoria, cercanías de Villa Victoria. Matuda 26620 (CODAGEM). pem Árbol de 10—15 m de alto. corteza gris: ramillas le 2 = mm diám., rojizas, con la edad cambian a 10. Quercus splendens Née, Anales Ci. Nat. 3: color café claro, con algunos tricomas estrellados I 8 570 Annals of the Missouri Botanical Garden POOR APP Re 3 B y EZT TS C Figura 10. Quercus splendens.—A. Rama. —B. cerca de los pecíolos y yemas, con numerosas len- ticelas pálidas, de hasta | mm de largo; yemas de 2 0 3 mm de largo, ovoides, de color castafio rojizo: estípulas de 5-8 mm de largo, subuladas, con in- dumento formado por tricomas largos, persistente en la yema terminal; hojas jóvenes con pubescencia fina, formada de tricomas estrellados dispersos en el haz, envés con tricomas simples y estrellados sobre la nervadura central; hojas maduras decí- duas, coriáceas, lanceoladas, elípticas, espatuladas, lámina 7-18 X 2-6 cm, borde ondulado, revoluto, engrosado, cartilaginoso, ápice obtuso o emargina- do, base cordada o redondeada; nervaduras de 11 a de cada lado, ascendentes, ligeramente ar- queadas, dividiéndose cerca del borde: haz glabro con las nervaduras impresas; envés con pocos tri- Fruto, —C. Tricomas. (Torres 529.) comas estrellados pequefios y abundantes tricomas simples adpresos a lo largo de la nervadura central, nervaduras prominentes; epidermis glauco-cerosa, papilosa; pecfolos de 5-9 mm de largo, de 1-2 mm diám. en su zona más ancha, rojizos o negros; flores desconocidas; fruto anual, solitario o en grupos de 2 a 3, sésiles; cúpula hemisférica, de 9-16 mm diám., escamas con los ápices triangulares, obtusas, canescentes, engrosadas en la base, sobre todo las basales: bellota oblata, pared interna del pericarpo glabra, de 7-9 mm de largo, de 11—15 mm diám., incluída en su totalidad en la cúpula. Figura 10. Reconocimiento. Se reconoce por sus hojas ver- de-azulosas, sin aristas y la pubescencia del envés que tiende a concentrarse en la nervadura central. Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México Distribución y hábitat. En México en los esta- Estado de Méx- ico, Nayarit, Michoacán, Morelos, Sinaloa y Oaxa- ca. En bosques de Pinus-Quercus y bosque tropical caducifolio, en altitudes de 1500—1900 m. Fenología. | Fructi dos de Durango, Guerrero, Jalisco, Fructifica en julio y agosto. Nombres populares y usos. Encino. No se tiene reportado algün uso. MÉXICO. Estado de Méxi- co: Ama у, cercanfas, Matuda 29824 (ENCB); ec km 18 carretera Sultepec—Salayatla, Torres 251a (IZTA); km 12 carretera Sultepec-San Miguel Totolmoloya, Torres 529 (IZTA); Tejupilco, cerro Nanchi- titla, Matuda 31527 (CODAGEM, ENCB); Mesa de Nan- chititla, Matuda 32817 (CODAGEM). E — — An 11. Quercus acutifolia Née, Anales Ci. Nat. 3: 267. 1801. TIPO: México. Guerrero: cerca de Tixtla, Née s.n. (holotipo, MA!). Arbol de 12 m de alto, tronco de 30 ст diám.: corteza obscura; ramillas de 1-2.5 mm diám., con costillas, color castaño rojizo, lustrosas, glabras, lenticelas de hasta 0.5 mm, de color claro abun- dantes; yemas ovoides, con el ápice agudo, de 4-6 mm de largo, escamas ciliadas en sus bordes: es- típulas de 5 mm, membranosas, linear-oblanceola- das, decíduas; hojas jóvenes con tricomas simples y estrellados dispersos, más abundantes en la ner- vadura central: hojas maduras subcoriáceas, rígi- das, verdes, envés más pálido, angostamente elíp- ticas, lanceoladas u oblanceoladas, lámina 6.5- 14.5 X 2.3-4.2 cm, ápice acuminado o agudo, base obtusa o cuneada, frecuentemente asimétrica, bor- de cartilaginoso, ligeramente revoluto, con 8 a 10 dientes aristados a cada lado, aristas de hasta 5 mm de largo; nervaduras de 8 a 14 de cada lado. casi rectas, dividiéndose a veces desde la mitad, las nervaduras primarias y secundarias forman aris- tas; haz glabro, lustroso, nervaduras impresas. la central algo elevada principalmente en la base: en- vés glabro. lustroso, epidermis lisa, no papilosa, nervaduras elevadas, principalmente la central: pe- cíolos de 9-22 X a y obscura; flores mm, glabros, su base más an- y frutos desconocidos. Figu- ra 11 Reconocimiento. (Quercus acutifolia es muy es- casa en el Estado de México; se reconoce por sus hojas dentadas y aristadas y envés casi glabro. Q. acutifolia muestra similitud morfológica con Q. conspersa, pero esta última se distingue porque el borde de sus hojas es entero y por su envés per- sistentemente glandular y marcadamente amari- llento por la presencia de las glándulas. Distribución y hábitat. En México en los esta- dos de Jalisco, México y Michoacán. Se le encuen- tra en bosque mesófilo de montaña, en altitudes menores a los 2200 m Fenología. Florece en enero; su fructificación es desconocida. Nombres populares y usos. Encino sencillo, en- cino blanco, encino laurelillo, encino tepezcohuite. La madera se utiliza como leña o para elaborar carbón, postería y pulpa para papel; se recomienda para chapa fina, muebles de alta calidad ebanística, pisos lambrines, cajas para empaque, mangos y ca- bos para herramienta (De la Paz, 1976). MEXICO. Estado de México: Valle de Bravo, en parte húmeda, Boege 1786 (MEXU); Rancho Rincón Grande Muller 9095 (MEXU). Ejemplares examinados. Piloxtoc, 12. Quercus candicans Née, Anales Ci. Nat. 3: 277. 1801. TIPO. México. cuaro, Pringle 3955 (MA). Michoacán: Pátz- Árbol de hasta 15 m de alto; tronco hasta de ramillas (1.5-)2-3(-3.5) mm diám., abundante pubescencia amarilla que disminuye m diám.: con con el tiempo, con lenticelas de 1-3 mm de largo, visibles en ramillas con la pubescencia disminuida; 5(—6) mm yemas ovoides de color castaño, de (2-)3— ` largo, con escamas pilosas; estípulas lineares de ~ 10-15 mm de largo, decíduas; hojas jóvenes algo lustrosas, haz con abundantes tricomas estrellados cortos y tricomas simples dispersos, envés con pu- bescencia densa blanca, semejante a la de las hojas maduras; hojas maduras coriáceas y gruesas, obov- adas, lámina (3.5-)5-19(-23.5) X 3-11(-14) cm, ápice obtuso o agudo, aristado, base subcordada a truncada o estrecha hacia el pecíolo, borde revo- luto, cartilaginoso, conspicuamente dentado o dien- tes mal definidos, hasta con 25 aristas de cada lado distribuidas en las dos terceras partes superiores de la hoja, aristas hasta de 5 mm de largo; nerva- duras 8 a 14 cada lado, rectas o ligeramente as- cendentes pasando directamente al diente; haz lus- verde tricomas estrellados y muy dispersos, pero abundantes en la troso, de color obscuro, con nervadura central cerca del pecíolo, las nervaduras más finas forman un retículo blanco, nervaduras central y primarias impresas o ligeramente eleva- das; envés con pubescencia densa blanca que con el tiempo sé amarillenta, formada con tricomas es- trellados con muchos rayos, sésiles y pueden pre- sentar un corto estípite; epidermis ampulosa, pap- Поѕа. nervaduras elevadas; pecíolo (3—)10—15(—-20) 0.5-1.5 illa a rojiza; amentos masculinos laxos, perianto de mm, con abundante pubescencia amar- 2.5-3 mm diám., pilosos en la parte externa y en 572 Annals of the Missouri Botanical Garden Figura 11. el lugar de inserción de los estambres, anteras ex- ertas de 1.5 mm de largo apendiculadas, filamentos de 2.5 mm de largo; fruto anual o bianual, solitario o en pares sobre pedünculos de 15 mm de largo, pubescente; cúpula hemisférica de 19-23 mm diám., borde recto, escamas gruesas, con pubes- cencia corta y muy abundante, ápices redondeados argo, de a agudos, glabros; bellota de 20 mm de 17 mm diám., anchamente ovoide, incluida un ter- cio de su largo en la cúpula. Figura 12. Reconocimiento, Se reconoce por sus hojas con Quercus acutifolia.—A. Rama. —B. Fruto. С. Tricomas. (Muller 9096.) dientes aristados, haz verde lustroso y envés con pubescencia blanca. Distribución y hábitat. dos de Chiapas, Chihuahua, Durango, Guanajuato, Hidalgo Jalisco, Estado de México, Nayarit, Oaxaca y Sinaloa, también en Guatemala. Se le encuentra de manera escasa o abundante en bosques de Quer- cus, Quercus—Pinus y bosque mesófilo de montaña, con Clethra, en altitudes de 2000- En México en los esta- y se asocia 2600 m. Fenología. Florece en mayo y fructifica en no- viembre. Volume 89, Number 4 Romero et al. 573 2002 Quercus en el Estado de México Figura 12. Quercus candicans.—A. Rama. —B. Fruto. —C. Tricoma. —D. Hoja. (Rodríguez s.n.) Encino de asta. en- Méxic Acambay, ladera S del cerro Hoyo de Lobo, Nombres populares y usos. cino cenizo, encino papatla, encino blanco, ahu- amextli. Su corteza se usa para dolor de muelas; su efecto dura hasta 15 días. De la Paz (1976) propone que se utilice para muebles у gabinetes de alta calidad ebanística, chapa fina, pisos para residencias, mar- cos para puertas y ventanas, cajas de empaque, cof- res, mangos y cabos de herramientas e implementos agrícolas. Ejemplares examinados. MÉXICO. Estado de Rojas & Romero 2059-A (IZTA); Coatepec de Harinas, 1 km al E de Coatepec de Harinas, Martinez 2845 (MEXU); 3 km al NE de Coatepec de Harinas sobre el camino a Agua Amarga, Rzedowski 30356 (ENCB): Huixquilucan, alrededores de Dos Ríos, Román 317 В); 1 km al NW de Santiago Yancuitlalpan, Rze- dowski 33489 (ENCB, MEXU); Jilote pee, Cerro de Jil- otepec, Matuda 29054 (CODAGEM, ENCB, MEXU); Santa Ana Jilotzingo, Sierra del s Alto, 3 km al / de Santa Ana Jilotzingo, Rzedowski 22412 (ENCB, MEXU — de la Sal, km 106 de la carretera a вео de la . Martinez 2847 (MEXU): en ladera des da, Matuda 28814 (CODA( tH T Malinalco, ¿EM Annals of the Missouri Botanical Garden MEXU); Ocuilan de Arteaga, km 18 de la terracería Ocuilan-Cuernavaca, Cortés 25 (IZTA); 1 km al S Ocuilan de Arteaga, Muller 9218 MENU js PCR km 16 carretera Sultepec—San Miguel Totolmoloya, Tor- res 302 (IZTA); Temascaltepec de González, Comunal de Tequisquiapan, Huerta € Ramos T-14 (ENCB, MEXU); El Guajero, rumbo a Real de Arriba, Rojas & Romero 2370 (VZTA); km 57 carretera Toluca—Temas- caltepec, Rojas & Romero 2426 (IZTA): ү 51 Toluca- ев & Romero — (IZTA); Tenan- i үз nanc ingo, Martinez 65 (MEXU); us de Yeca, Ávila s.n. (E у B, IZTA); Villa del Gonz zález s.n. (IZTA); San Jerónimo, (U); Cerro Bufa, Matuda 29753 2 Guerrero, 5 km al N de Villa — . (ENCB); Villa Nicolás Romero, km ¢ carretera V illa Ni ‘olds Romero—Tlazala, límite ш - ipal de Villa Nicolás Romero, — 33 (IZTA): Bar- rio IV, Rojas & Romero 3191 (IZTA); Cahuacán, Кой аў & Romero 3191 A (IZTA); 3 km al NW de Cahuacán, Rzedowski 32626 (ENCB). epec, Aojas me tho Y . Quercus castanea Née, Anales Ci. Nat. 276. 1801. TIPO: México. Hidalgo: Née s.n. (MA). Quercus axillaris Fourn. ex Trel., Mem. Natl. Acad. Sci. 20: 176, pl. 352. 1924. TIPO: México. San Nicolas, Valle de México, Bourgeau 1138 (B). Quercus pulchella Humb. * Bonpl., Pl. (38)44, pl. 88. 1809. TIPO: México. Santa Rosa, ie s.n. (P). Árbol de 5-15 m de alto, tronco de 40—80 cm ramillas de (0.5—)1—2 mm diám., café claro a obscuro, pubescentes, a veces glabras Aequinoct. 2: Guanajuato: diám.; de color con varias costillas, lenticelas pálidas de 0.5-1 mm de largo: yemas de (1—)2—3.5(—7) mm de largo, ovo- ides o agudas, de color café, con escamas ovadas, ciliadas en los márgenes y dorso superior, cori- áceas; estípulas de 5-6 mm de largo, lanceoladas, de color claro, membranosas, con tricomas largos muy principalmente en los márgenes, caedizas pronto; hojas jóvenes muy parecidas a las maduras, haz verde con pubescencia principalmente en la base y nervaduras, envés densamente pubescente; hojas maduras gruesas, elípticas, elfptico-oblongas, oblanceoladas o lanceoladas, láminas (1.8—)3—9(— 11.6) istado, base redondeada a cordada, borde plano o 14(-5.5) em, ápice agudo u obtuso, ar- ligeramente revoluto, entero o dentado, con 1-7 aristas de hasta 2 mm de largo, en cada lado del tercio o mitad superior; nervaduras primarias de (3235-12 en cada lado, ascendentes, se ramifican y reticulan cerca del borde, generalmente se con- tinúan en una arista; haz verde claro a obscuro, lustroso, finamente rugoso por las nervaduras, gla- bro excepto cerca del pecíolo, nervadura central elevada principalmente en la base, nervaduras pri- marias y secundarias impresas, las menores forman un retículo pálido; envés muy pubescente, la abun- dancia de tricomas disminuye con la edad, pubes- cencia grisácea, formada de tricomas estrellado-es- tipitados, sus rayos rugosos y enredados, epidermis con ámpulas prominentes, nervaduras conspicua- mente elevadas, pálidas; pecíolos de (2.5—)4—10(— se ensanchan en la base, pu- 5) X 0.5-1.5 mm, bescente o glabros; amentos masculinos de 6 cm de largo, pedtinculos pubescentes, perianto de 1.5— 2 mm diám., bordes ciliados, anteras de 1.5 mm de amentos fe- argo, filamentos de | mm de largo: meninos con | o 2 flores, pedúnculos de 3-5 mm de largo, con pubescencia muy corta formada de tricomas estrellados; fruto anual, 1 о 2 sésiles o sobre un pedúnculo de 1-7 mm de largo; cúpula hemisférica de 9-14 mm diám., con escamas algo engrosadas en la base, ápice obtuso y papiráceo, pubescentes a casi glabras, de color café rojizo; bellota anchamente ovoide, pared interna del per- icarpo lanosa, de 5-15 mm de largo y de 8-11 mm diám., incluida en la cúpula de un tercio a un me- dio de su largo. Figura 13. Reconocimiento. Esta especie se reconoce por sus hojas aristadas y envés con las nervaduras con- spicuamente elevadas y reticuladas. Distribución y hábitat. En México en los esta- dos de Colima, Distrito Federal, Durango, Guana- juato, Guerrero, Hidalgo, Jalisco, Estado de Méxi- Morelos, Nayarit, Oaxaca, Puebla, San Luis Potosí, Sonora, Sinaloa y Veracruz, tam- co, Michoacán, bién en Guatemala. Se encuentra en bosques de Pinus, Quercus y Pinus—Quercus, pastizal con ma- torral xerófilo y bosque mesófilo de montaña; es frecuente encontrarla en encinares perturbados. Se le ha visto asociada con Quercus obtusata, Q. glau- coides, Q. conspersa, Clethra, Arbutus y Pinus mi- choacana, en altitudes de 1900-3500 m. Fenología. Florece en junio-julio y fructifica de agosto a diciembre. Nombres populares y usos. Encino, encino blan- co, encino negro, encino amarillo, encino rojo, palo colorado, encino pipitillo, roble, encino prieto, te- pozcohuite chino, encino chaparro, encino colorado y aguacatillo. De la Paz (1982) recomienda su madera para pisos de residencias, vehículos (de motor y no mo- torizados), tarimas para carga y descarga, mangos y cabos de herramienta, implementos agrícolas, y diversos tipos de recipientes y armazones de con- strucción. — ке MEXICO. Estado de Méxi- co: Acamba ITO — a 1442 (IZTA); AI- moloya del ш. Can de Andrés, Rodriguez 147 (ENC — | km A NW W cae San Antonio Zoy- atzingo, Reyes 211 (ENCB, MEXU); Atizapán, vertiente SE del cerro Chiluca, 6 km al SW de C Pe Adolfo López Volume 89, Number 4 2002 Romero et al. 575 Quercus en el Estado de México Figura 13. dríguez s.n.) Mateos, Patiño 335 (ENCB); — nte SE del cerro a hil- uca, Polaco 214, 335 (ENCB); Cerro de Tigre, al NW de — Adolfo López Mateos, Ps dowski 32036 (E NE В, XU); parte alta del cerro Chiluca, cerca de Ciudad нея те Mateos, Rzedowski 32586 (ENCB, MEXU): El Oro de Hidalgo, entre el Oro y Villa Victoria, Matuda 28644 (CODAGEM, MEXU); cercanías del Oro, Машаа 28646 (ENCB): ds pa an, entre San Bartolo y Toluca. е 495 (ЕМ km al N de San Bartolito, García 189 (ENCB); Magdale na, Rzedowski EUN (E NC B = Jilotzingo, Sierra del Monte Alto, : Ana Jilotzingo. Rzedowski 22414 (ENCB): Malinalco. Mal- Quercus castanea.—A. Rama. —B. Fruto. —C . Tricoma. —D. Hojas. —E. Inflorescencia. (A-E: Ro- inalco, Matuda 28776 (ENCB); Naucalpan de Juárez, Juárez, COTECOCA 15072 (ENCB): al- San Francisco C himalpa, Rzedowski 20117 (ENCB. ME XU); 8 km al W de San Bartolo, Rzedowski 32243 (ENCB, МЕХ); Oc ile de Arteaga, El Ahuehue- te, Rzedowski 26922 (ENCB); Otzoloapan, Cerro de Pinal, Matuda s.n. (CODAGEM, MEXU); Polotitlán, km 147 car- retera a Toluca, Rojas & Romero 3099 (1; ); San Felipe del Progreso, entre el Oro y Villa Victoria, Matuda 2864 (ENCB) Nuevo Santo Tomás de los Pl: Nuevo Santo Tomás de los Plátan (ENCB): Sultepec, Real de Abajo, а 29300 (ENCB. МЕХ); 2 km al NE de Sultepec, Rodríguez 166 (ENCB): km 18 carretera Sultepec-San Miguel Totolmoloya, Torres 576 Annals of the Missouri Botanical Garden 317 (IZTA); Tejupilco, Nanchititla, Rodríguez 199 (ENCB, — |ustroso, de color olivo, glabro, excepto en la ner- MEXU); Temascaltepec de González, Tenería, Guizar 499 (MEXU); San Lucas del Pulque, Hinton 6738 (MEXU); Real de Arriba, Huerta RA-1 (ENCB); Temascaltepec de González, Mat d — (ENCB); La Labor, Mat uda 2763 (ENCB, MEXU); Estancia Vieja, 10 km al S de Te- mascaltepec de ( Ludis Moreno 147 (Е NC B); km 4 Val- le de Bravo-Temascaltepec, dirección a Pulque, Orozco & Rojas 520 E (IZTA); Barrio la Mesa, — Ке ны вни Orozco & Rojas 610 A (IZTA); km alle de B — mascaltepec, dirección a la Albarrada, * o 284 E 5 km al SW de Temascaltepec de González, EARS 208. 30 (ENCB); d — cingo, Martínez 698 (ENCB); Tenancingo, r 697 (MEXU); 10 km : N de Tenancingo, Muller 9194 (Е NC B); Comunal Real de Arriba, Ramos & Rocha s.n. (MEXU); Tenango, entre Te e y€ iai m, Martínez s s.n. Pp А арча 2 km al N de Magú, Cruz 493, 496 (ENCB, XU); Sierra de Alca aparrosa, Núñez 1710 ZA Cer- canfas de la Presa Cone “pe ión, Rzedowski 22908 (ENCB); i Alc — rrosa, 2 km al NNW de 8 (ENCB); Tlatlaya, Tlatlaya y cercanfas, Matuda 29829 (CODAGEM, ENCB, ME XU} Valle de Bravo, Valle de Bravo, Martínez 2612 (ENCB); — buea pec, Rodrigues 204 (ENCB); Villa del 5 'rogreso. Industrial, d is =! T Es = € arbón, 5 al W de I porta — (E NCB, MEXU); Villa Guerrero, Villa rero, COTECOCA 15077 (ENCB); Villa Nicolás — З km al E de Cahuacán, Espinosa 621 (ENCB); Cahuacán, Germán & Trejo 43 (MEXU); km 32 carretera — tla—Villa del Carbón, Rojas & Romero 2019 (IZTA); 1 km 2 S de — 'án, Rzedowski 16815 (ENCB); * ualpan, km 3-4 carretera Zacualpan-Mamatla, Fragoso 340 (IZTA). 14. Quercus conspersa Bentham, Pl. Hartw. 91. 1842. TIPO: Guatemala. Las Casillas, Hartweg 617 (isotipo, NY!). Árbol de 9 m de alto, tronco de 50 cm diám., ramillas de 2-3 mm diám., gla- bras, de color castafio rojizo; lenticelas numerosas, de hasta 0.5 mm de largo; corteza obscura; blancas, protuberantes, yemas de 2—3 mm de largo, ovoides, ápice agudo, con escamas ciliadas en sus bordes, de color cas- tafio-rojizo; estípulas de 7-8 mm de largo, lineares, pilosas; persistentes en las hojas jóvenes; hojas jóv- enes rojizas, haz con tricomas simples y estrellados, escasos y dispersos, envés con abundantes tricomas glandulares de color ámbar y estrellados con un largo estípite y rayos enredados entre sí, abundan- tes en las axilas de las nervaduras primarias; hojas maduras gruesas, coriáceas, decíduas, elíptico-ova- das, lanceoladas u oblanceoladas, lámina (8—)11— 17(-22) X (2-)3—6.5 ст, ápice acuminado o agudo aristado, base obtusa, subcordada o atenuada, a ve- ces asimétrica, borde grueso, cartilaginoso, revo- luto, entero; nervaduras primarias de 7 a 10 en cada lado, ascendentes, en hojas elíptico-ovadas son casi paralelas, ramificándose, a veces, desde la haz base de la nervadura hasta cerca del borde; vadura central en la que se observan abundantes tricomas glandulares obscuros y tricomas estrella- dos pequefios en la base de la hoja, nervaduras impresas, envés también lustroso, amarillento por la presencia de tricomas glandulares color ámbar, con mechones de tricomas estrellados con estípite largo y rayos enredados entre sí, concentrados en las axilas de las nervaduras primarias, epidermis papilosa, nervaduras prominentes; pecíolos de 8— 15 mm de largo, engrosados en su base, de color rojizo a negro, en un principio pubescentes, luego glabros-rugosos; flores desconocidas; fruto bianual, solitario, sobre un pedúnculo grueso, de 4-6 mm de largo; cúpula poco profunda, de 14 mm diám., borde enrollado; escamas con pubescencia blanca, adpresas, no engrosadas en la base, ápices delto- ides u obtusos; bellota comprimida, pared interna del pericarpo lanosa, de 5 mm de largo, de 10 mm diám., la mitad de su longitud incluida en la cu- pula. Figura 14. Reconocimiento. Quercus conspersa se reconoce por sus hojas de borde entero, haz lustroso de color verde-olivo y envés amarillento por la presencia de tricomas glandulares. Esta especie se confunde con Q. acutifolia (ver Q. acutifolia). Distribución y hábitat. En México en los estados de Chiapas, Guerrero, Jalisco, Estado de México, Michoacán, Oaxaca y Veracruz, también en Amér- ica Central. Se le encuentra en bosque de Quercus— Pinus, y se le encuentra asociado a Pinus oocarpa, en altitudes menores de 2000 m. Fenología. | Fructifica de diciembre a febrero. Nombres populares y usos. Encino, roble amar- illo, encino rojo, encino colorado, encino pipitillo, tepozcohuite, encino cáscara, encino blanco. Su madera es utilizada para la elaboración de chapa y pulpa para papel (González, 1986). Ejemplares examinados. MÉXICO. Estado de Méxi- co: Temasca ерес de González, Pungarancho, Hinton 7570 (MEXU); 0.5 km al SE de Albarrada, 7 km al W de Temascaltepec de González, Muller 9104 (ME XU); km 15 desviación San Pedro Tenayac, Orozco 144 (IZTA) 15. Quercus crassifolia Humboldt & сас РІ. Aequinoct., 2: 49, pl. 91. 1809. ТІР México. Guerrero: Chilpancingo, a п. (В, Arbolillo de 2-3 m de alto o árbol de hasta 15 m de alto, tronco hasta de 1 m diám., ramillas de 2-5 mm diám., con abundantes tricomas estrella- dos estipitados cubriendo totalmente la superficie, pubescencia amarilla a café que se obscurece y cae con el tiempo; lenticelas numerosas de hasta 3 mm, Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México 577 Figura 14. claras, visibles sólo en las ramas que han perdido la pubescencia; yemas de 3-7 mm, ovoides, pilo- sas; estípulas de 8-12 mm de largo, oblanceoladas, escariosas, sedoso pubescentes, con el tiempo gla- bras, persistentes en las yemas, a veces también en las hojas maduras; hojas jóvenes lanoso-tomento- Quercus conspersa.—A. Rama. —HB-D. Frutos. (Hinton 7570.) sas, haz rojizo y envés café, con aristas de hasta 3 mm de largo, indumento formado por tricomas es- trellados estipitados y glandulares rojizos sobre la epidermis obscura; hojas maduras por lo general obovadas, oblongo-obovadas, elípticas, coriáceas, 5-)7-17(-20) х 3.5-10(-11.5) cm, ápice ámina (4 578 Annals of the Missouri Botanical Garden agudo, acuminado u obtuso, aristado, base cordada o redondeada, borde ondulado o dentado, revoluto, cartilaginoso, de 3 a 8 dientes en cada lado, ubi- cados casi desde la base de la hoja. 1 a 10 aristas. dientes y aristas con frecuencia asimétricos; ner- vaduras 6 a 12 a cada lado, ascendentes, ramifi- cándose cerca del borde, pasan directamente a los dientes y forman las aristas, con frecuencia hacen — lo mismo las ramificaciones; haz lustroso, glabro, excepto la nervadura central que posee abundantes tricomas estrellados, sobre todo cerca del pecíolo, nervaduras impresas; envés amarillo anaranjado o café, lanoso-tomentoso, indumento formado por tri- comas estrellados con un largo estípite, sus ramas muy largas y entrelazadas, cubriendo la epidermis ampulosa y papilosa, nervaduras elevadas; pecíolo con pubescencia densa formada por tricomas es- trellados más cortos que los del envés, ensanchado en su base, de 6-15(-17) X (1-)2-3 mm: amentos masculinos de 8-9 mm de largo, raquis pubescente, — perianto amplio, de 2-2.5 mm diám., anteras cur- vadas de 2 mm de largo, amarillentas-cafés, apen- diculadas; fruto anual o bianual, solitario o en pares, sésiles o en pedúnculos de 3-5 mm de largo: cúpula hemisférica, de 11-13 mm diám., con las escamas delgadas adpresas, pubescentes, princi- palmente en sus ápices, éstos redondeados; bellota ovoide, pared interna del pericarpo lanosa, de 10— 17 mm de largo, de 8—11 mm diám., incluida en la cúpula % de su largo. Figura 15 Reconocimiento. Quercus crassifolia se recono- ce por sus hojas aristadas con el envés lanoso amarillo, anaranjado o café y ramillas de 1-5 mm Чат. Esta especie muestra similitud con Q. hin- tonii; ésta última se distingue porque el envés de sus hojas posee epidermis lisa y las ramas de los tricomas estrellados son más largas. En ocasiones también puede confundirse con Q. dysophylla, pero ésta última posee hojas de forma ovada, lanceolada o elíptica, por lo general de borde entero; mientras que Q. crassifolia casi siempre las presenta obov- adas o dentadas. Además los pecíolos son más cor- tos en Q. dysoph ylla. Distribución y hábitat. En México en los esta- dos de Chiapas, Guerrero, Hidalgo, Jalisco, Estado de México, Michoacán, Oaxaca, Puebla, Querétaro, San Luis Potosí, Tlaxcala y Veracruz, también en Guatemala. Se le encuentra en bosques de Quercus, у Quercus—Pinus, se asocia con Pinus leiophylla, О. laurina y Q. crassipes, y se le ha encontrado en zonas de regeneración después de incendios, en al- titudes de 2500-2800 m. Fenología. | Florece en abril y fructifica en oc- tubre. Nombres populares y usos. Encino chicharrón, encino colorado, encino prieto, encino roble, encino hojarasco, encino huaje, chi-ka-chi, hoja ancha. La madera de esta especie se utiliza para la man- ufactura de implementos agrícolas, para horcones de casas, mangos de herramientas y leña: el carbón obtenido de esta especie resulta durable y de buen peso (Camacho, 1985); los retoños son comidos una vez cocidos, molidos y mezclados con maíz; junto con carrizo de monte, sauco, toronjil y zarza se em- plea para hacer un agua caliente para mujeres des- pués del parto; la corteza se usa para aliviar dolores de encías, para curtir pieles y en la preparación de bebidas de Agave (Pennington, 1969). Ejemplares — MEXICO. Estado de Méxi- co: Acambay, Martinez s.n. (ENCB); Aculco de Espinosa, l farther da (МЕХ): Axapusco, Cerro del Tipayo, ЕМСВ); Coatepec de Harinas, sobre el cam- Amar E — ski 30351 (ENCB); Capul- huac, Rodríguez s.n. CB); El Oro de Hidalgo, Bassoco 0, Rojas & 3. оа — en (ИТА — Ne B): Peña Desc ‘ani, Min nez s.n. MT XI 2 Jil- otepec, Matuda 26673 (ENCB); Cerro de Jilotepec, Ma- "na 29058 (CODAGEM); Santa Ana Jilotzingo, Sierra de Monte Alto, al NE de S: Jilotzingo, Rzedowski 22411 (ENCB); Temascaltepec de González, San Francis- co Oxtotilpan, re gion Matlatzinca, Fragoso R. 253 (IZTA); AN Sierra de San — Camacho 153, 233, 293, ZTA); 4 km al V cabecera municipal, Gómez с vo 4 — Tlazala de Fabela, alrededores de Tlazala de Fabela, Román 275 (MEXU); 2 km al SE de Tlazala de F al, 290 (ENCB); Villa del — Cerro de la Cabra, Bringas s.n. (ENCB); 1 km a de Villa del C у Cruz 169 (ENCB); Cerro La An ar- tinez s.n. (ME XU): cercanías de Villa del Carbón, Maiuda — 28767 (E NCB); San Jerónimo, Matuda 29017 (С )DA- GEM); San Je rónimo, Mao 29160 (ENCB); km 15 Ji- quipilco-Villa del € n, Rojas & о 3175 (IZTA): El Plan, Rodríguez s. s.n. - (zi TA 0: Villa N cola А Котего, 5 SW « — = — wn 7 - > — 1831 0 (ENC В, МЕХ); 2 km al NW huacán, Rzedowski 33755. 16. Quercus crassipes Humboldt & Bonpland, PI. Aequinoct. 2: 37. Pl. 83. 1809. TIPO: Méx- ico. Guanajuato: Santa Rosa, Bonpland s.n. otipo, B!). — 18- Árbol de 4—17 m de alto o más, con el tronco de 0.40—1 m diám.; corteza de placas alargadas о de color obscuro; ramillas de (5—)1 2 mm diám., con pubescencia densa amarilla, formada por tri- comas estrellados con estípite muy pequeño: len- — ticelas hasta de 1 mm de largo, desde pálidas hasta del mismo color de las ramas; yemas de 1.54.5 mm de largo. ovoides, de color café-rojizo, escamas coriáceas, bordes ciliados; estípulas de 7-8 mm de Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México Figura 15. 3081; B-D: Camacho 15. Quercus а —A. Ката. —B. Fruto. largo. linear-lanceoladas, membranosas pubescen- tes en el dorso, decíduas; hojas jóvenes con abun- dante pubescencia amarilla en haz y envés, prin- cipalmente en la nervadura central; hojas maduras coriáceas, angostamente elípticas, lanceoladas и oblanceoladas, lámina 2-9(-10.8) X (0.6) 1-34) em, ápice mucronado o con arista de 3 mm de lar- go. base redondeada o subcordada, borde entero. —С. Tricoma. =). Hoja. —E (A: Romero . Inflorescencia. revoluto, engrosado: nervaduras de 10 a 19 en cada lado, rectas o algo curvadas, formando ángulos casi rectos, bifureados cerca del margen: haz algo lus- troso, color verde o grisáceo, glabro o con pequeños tricomas estrellados dispersos, muy abundantes en a base de la hoja. nervadura central elevada, las primeras impresas, las más finas forman un retículo pálido sobre un fondo verde; envés con pubescen- 579 580 Annals of the Missouri Botanical Garden AJ S SIN 7272» (LLL Figura 16. 20.) cia densa, grisáceo, tricomas estrellados estipita- dos, con 5 a 6 rayos extendidos, epidermis ampu- losa; nervaduras ligeramente elevadas: pecíolos amarillentos o rojizos, pubescentes o casi glabros, (1-)2-7(-10) mm de largo, de 0.5-1 mm diám.: 4—5.5 cm de largo; flores con el perianto escarioso de 4 X 3 mm, café-rojizo, pubescentes; estambres 5, de 3 mm de largo, an- teras apiculadas; flores femeninas de amentos masculinos de o 2 sobre pedúnculos de 5 mm de largo o menos, de 2—2.5 mm diám.; fruto bianual. solitario o por pares en pedünculos de 2-8 mm de largo: cúpula hemisfér- Quercus crassipes.—A. Rama. —B. Fruto. 2» + << Д ERES LLZ — 1ст у, ES ^ XS RE A Z KE (LEED SAN: Мах — SER NN 0.5mm — |. Tricomas. — Ноја. (A-C: Romero 2055; D: Acosta ica, de 11-17 mm diám., à veces invo- lutos, las escamas engrosadas en la base, pubes- centes, a veces glabrescentes; bellota ovoide, pared interna del pericarpo lanosa, de 12—17(—30) mm de largo, de 8-15 mm diám., cerca de la tercera parte de su largo incluida en la cúpula. Figura 16. márgenes Reconocimiento. Quercus crassipes se reconoce por sus hojas con el ápice aristado y las nervaduras que forman ángulos casi rectos: muestra similitud con Q. mexicana. Esta se diferencia porque el en- vés de la hoja posee tricomas estrellados con sus Volume 89, Number 4 Romero et al. Quercus en el Estado de México ramas enredadas entre sf, de manera que a simple vista se observan como puntuaciones. Distribución y hábitat. En México en los esta- dos de Jalisco, Guanajuato, Querétaro, Hidalgo, Colima, Michoacán, Estado de México, Distrito Federal. Morelos, Tlaxcala y Puebla. Se le encuen- tra en bosques de Quercus, Pinus-Quercus у Quer- cus—Cupressus, bosque mesófilo de montafia, mator- ral xerófilo, en sitios de transición de pastizal a bosque mixto; se asocia con Pinus pseudostrobus, P. leiophylla, P. montezumae y P. hartwegii, Quercus laurina, Q. crassifolia, Q. obtusata y Q. castanea, en altitudes de 1900-3500 m Fenología. | Florece en mayo y fructifica de sep- tiembre a enero. Nombres populares y usos. lorado, encino chilillo, encino oreja de ratón y en- Encino, encino co- cino laurel. De la Paz (1982) recomienda su madera para pisos de residencias, auditorios, museos, almace- nes, pistas de baile (en forma de parquet y ado- quín), para chapa fina, muebles y gabinetes de alta calidad ebanística, lambrín, decoración de estudios y corredores, cocinas integrales, baúles, canastos, macetas, cofres y diversos artículos decorativos, mangos para herramientas, lomos y mangos de ce- pillos, brochas y de utensilios de cocina, pasama- nos, huellas (escalones) y descansos de escaleras, hormas para zapatos y cajas para pianos. — ip orco MÉXICO. Estado de Méxi- ulco de Espinosa, Bosque Nado, 10 km antes de Ad Ds o de Espinosa, Matuda 28731 (ENCB); Almoloya de Juárez, km 11 de la carretera Toluca-Valle de Bravo, Ks- trada 346 (ENCB); Amecameca, San Antonio Zovatzingo, Ávila s.n. (ENCB); 1 km al NE de Santo Tomás Atzingo, Cervantes 181 (ENCB); 3 km al SW de San Antonio Zoy- atzingo, DM 528 (ENCB); Cerro Sacromonte, Rzedows- ki 26695a (ENCB); Capulhuac, Cerro т | km al W de San Minne Almaya, Rodriguez s.n. (EN Coacalco, 6 km al S de Coacalco, Sierra de Dono “Resins ski 30777 (ENCB); Chalco, Ladera E del Cerro La Tijera, km al SE de Santa Ana, Pineda Va ҮРҮ, Chicoloa- pan, 4 km al E de Coatepec, Román 311 (ENCB); El Oro de Hidalgo, El Oro de Hidalgo, ee 26420, 28647 (CODAGEM); cercanías del Oro, Matuda 28624 (ENCB); Morelos, 6 km al N de Santa Clara, Ramírez ENCB); а Fraccionamiento La Herrad- ura, аа s.n. (ENCB); Dos Ríos, Matuda 28405, 28406 (ENCB); alrededores de Dos ys I1: 327 (ENCB); Ixtapaluca, Río Frío, García 92 ( ied km 42 carretera México-Puebla, Hernández 64. B); Jilote- Denxi, Martínez 67 (IZTA); Jilotepec, Mud: 26674 Matuda 29052 29065 (ENCB); o de Jocotitlán, Matuda 30238 (ENCB); Otimba de ok Farías, 15 km al NE de Texcoco, Ochoa 40 (ENCB); Cerro Cuixi al E de Santa Bárbara, Rzedowski 16889 (ENCB); Otzoloapan, Cerro de Pinal, Matuda 31754, 31870 (CODAGEM); Ozumba de Alzate, 1 S de Ozumba de Alzate, COTECOCA 15036 (ENCB); — ~= No Js pec, (CODAGEM, ENCB); Cerro de qo nd — Tenango de Tepopula, San Luis Aculco de Espinosa, Hin- ton 115, 117 (ENCB); 3 km al W de Tenango de Tepopula, Pineda 731 (ENCB); San Martín de las Pirámides, Cum- bre del Cerro Gordo, Rzedowski 18801 (ENC В); Temas- caltepec de González, Cieneguillas, 14 km al W de Te- mascaltepec de González, Moreno 185 (CHAPA, ENCB): Temascalapa, Cerro Gordo, Castilla & Tejero 634 (ENCB, IZTA); Tepetlaoxtoc, 18 km al E de Texcoco, por la car- retera a Calpulalpan, De la Cruz 15 (ENCB); 14 km al E de Texcoco, Koch 75134 (ENCB); Tepotzotlán, 3 km al W de Magi, Mendiola. 188 (ENCB); parte alta de la Sierra 1 de la estación de microondas, Rze- dowski 29956 (CH. APA. E NCB); Texcoco, 20 km al NE de Texcoco, Cruz 1834 (ENCB); Cerro Tetzahuitl, Chávez 2233 (ENCB): 24 km al E de Texcoco, García s.n. (CHA- ›А); 19 km al E de Texcoco, Koch 75134 (ENCB); Tlal- Medellín s.n. (ENCB); Tlazala de Fabela, 4 km Santiago Tlazala, Jiménez s.n. (ENCB); 7 km al E CB); Cafiada de Onofre, T — le Alcaparrosa, cerc manalco, al E de ! de Santiago Tlazala, Luna s.n. Román 279 (ENCB); . Centro Ceremonial Toluca, Cuellar s.n. (ENCB); Valle di Bravo, ladera E del Cerro Gordo. — 9107 (ENC B); Villa del Carbón, Predio Pie- dri a Azul, Ávila s.n. n NC B); San Jerónimo, Matuda 29014 „Matuda 29735 (ENC г tahuacan, ; 5 km al SW de Cahuacán, > Santiago Tlazala, E de era Tisxala. (E NC B): 2 km de Cahuacán, Mortecinos 200 (E km al NW de Cahuacán, Reedow ski 33758 (ENCB). de ( N 17. Quercus dysophylla Bentham, PI. Hartw. 55. 1840. TIPO: México. Huasca: Hartweg 421 (holotipo, K!). 132, Esperanza, Natl. Acad. Sci. 20: TIPO: México. Mem. Quercus esperanzae Trel., pl. 248. 1924. Syn. nov. Purpus 5332 (NY!). Trel., Mem. Natl. Acad. Sci. 20: 177, pl. Quercus — ri 35 )24. Syn. nov. TIPO: México. Cofre de — — 255 (Р!). Quercus — Trel., Mem. Natl. Acad. Sei. 20: 131, pl. 247. |. TIPO: México. Cajalpa, cerca de Tolue a, к Fe (P). Árbol de 5-20 m de alto, diám., estriadas, glabrescentes, el indumento se ex- ramillas de 1-3 mm tiende a los pecíolos y nervadura central, formado de tricomas estrellados estipitados, de color cas- año-rojizo; lenticelas hasta de 1 mm de largo, más claras que el tallo; yemas de 2-5 mm de largo, ovoides, de color café claro, con el ápice redon- deado y los bordes largamente ciliados; estípulas caedizas, lineares, de 4-9 mm de largo, pubescen- tes en toda su superficie; hojas maduras gruesas y coriáceas, ovadas, lanceoladas o elípticas, lámina (3235-15-17) X (12-56) ст, ápice obtuso o agudo, mueronado o aristado, aristas de hasta 2 mm — de largo, base redondeada o cordada, a veces asi- métrica, borde revoluto, entero u ondulado, a veces con algunos dientes, con O a 13 aristas en cada lado, de hasta 2 mm de largo; nervaduras de 9 a 15 en cada lado, rectas o algo curvas, se bifurcan Annals of the Missouri Botanical Garden C Figura 17. en el tercio distal o cerca del margen; haz verde pálido, glabro, con tricomas estrellados, estipitados, dispersos sobre la lámina, abundantes sobre la ner- vadura central, nervaduras impresas; envés con pu- bescencia amarilla, formada por tricomas estrella- dos, estipitados, y por escasos o abundantes tricomas simples glandulares que dejan ver la epi- dermis ampulosa, nervaduras elevadas; pecíolos de (2—)4—7(—12) mm de largo. de 1-2 mm diám., den- samente pubescentes, se ennegrecen con la edad; flores desconocidas; fruto solitario o en grupos de dos, sésiles o en pedtinculos de 1—2 mm de largo, con lenticelas claras; cúpula turbinada, de 10—12 mm diám., de 8-12 mm de alto, borde recto, a veces involuto; escamas no engrosadas en la base. no adpresas, excepto las superiores, canescentes, ápice obtuso, márgenes más obscuros; bellota ovo- ide, pared interna de pericarpo lanosa, de 12-17 mm de largo, de 10—11 mm diám., de color castaño claro, incluída la mitad de su largo en la ctipula. Figura 17. Quercus dysophylla.—A. Rama. —B. Fruto. —C. Tricomas. (Romero 1.) Reconocimiento. Quercus dysophylla se recon- oce por sus hojas lanceoladas o elípticas, con el borde entero y aristado, envés con pubescencia amarilla y epidermis ampulosa; puede confundirse con Q. crassifolia y Q. hintonii. Quercus crassifolia se reconoce por sus hojas obovadas, dentadas, ar- istadas con el envés lanoso amarillo, anaranjado o café, y epidermis ampulosa. Quercus hintonii se distingue porque el envés de sus hojas posee epi- dermis lisa y las ramas de los tricomas estrellados son más largas que las especies anteriores. Distribución y hábitat. México en los estados de Distrito Federal, Guanajuato, Hidalgo, Estado de México, Michoacán, San Luis Potosí, Tlaxcala y Veracruz. Se le encuentra en bosques de Quercus, Pinus—Quercus, bosque mesófilo de montaña y bosque de Pinus perturbado, y se asocia con Quer- cus rugosa y Q. castanea, en altitudes de 2490— 2850 m. Fenología. Florece en marzo y abril y fructifica en septiembre y octubre. Volume 89, Number 4 2002 Romero et al. 583 Quercus en el Estado de México Nombres populares y usos. Encino, laurelillo. Bello y Labat (1987) sefialan que se utiliza como lefia y postes de cercas. MÉXIC O. Estado de Méxi- Ejemplares examinados. co: Amecameca, Alcansi, Amecameca, Vela “i de Chalma, al S de Huixquilucan, Геб 713 (ЕМСВ); alrededores de Huixquilucan, Román 332 (IZTA); Tepot- zollán, Sierra de Alcaparrosa, Núñez 1677 (ENCB): parte alta de la Sierra de Alcaparrosa, Rzedowski 31264 (ENCB); Timilpan, San Andrés — Camacho 357. 360, 241 (IZTA); Tlazala de Fabela, 4 km al E de Santiago Tlazala, Jiménez 141 (ENCB): Villa del Carbón. Villa del Carbón, Ramos 1 (IZTA); Villa Nicolás Romero, Villa Ni- colás Romero, Jiménez s.n. (ENCB); 2 km al NW de Ca- — ‘an, Rzedowski 33757 (ENCB); | km al NE de Ca- асап, Tirado s.n. (ENCB) . Quercus elliptica Née, Anales Ci. Nat. 3: 278. 1801. TIPO: México. Hidalgo: Ixmiquil- pan y Zimapán, Née s.n. (I Árbol de 8-15 m de alto, con el tronco de 15— 70 em diám., corteza gris obscura; ramillas den- samente pubescentes, de 2-3 mm didm., al prin- cipio de color amarillo, después se ennegrecen, el indumento formado por tricomas estrellados unos largos y otros cortos, erectos; lenticelas escasas cu- biertas por la pubescencia, de hasta 0.5 mm de largo: vemas de 2—4 mm de largo, ovoides, de color castaño obscuro; estípulas de 7-8 mm, oblanceo- ladas, ciliadas, principalmente en ápice v base. membranosas, escariosas, decíduas; hojas jóvenes con el haz v envés cubiertos de pequefios pelos estrellados, los de la nervadura central más largos. y pelos simples glandulares color ámbar: hojas maduras rígidas, coriáceas, elípticas, lanceoladas u oblanceoladas. lámina (2.1—)4.5-12 X (1.8—)3—5(— 6.5) em, ápice obtuso, agudo o emarginado, cuando es agudo a veces con arista de hasta 2 mm de largo. base redondeada, cordada o auriculada, borde en- tero, engrosado, cartilaginoso, ligeramente revoluto: nervaduras de 8 a 16 en cada lado, se ramifican en el borde, forman arcos interconectados: haz verde pálido, lustroso, casi glabro, con pequeños pelos muy dispersos, en la base de la nervadura central se agrupan pelos con rayos más largos, nervaduras impresas o ligeramente elevadas; envés glabro, a veces con pelos estrellados dispersos sobre la lám- ina, con aproximadamente 6 rayos cortos extendi- dos, en las axilas de las nervaduras son más largos. nervaduras elevadas, pálidas, gruesas, epidermis ligeramente ampulosa, papilosa; pecíolos de (3—)4— 7 mm de largo, de 1-1.5 mm diám., densamente pubescentes; amentos masculinos de 5-9 cm de largo, de muchas flores en posición alterna, flores sésiles, perianto de 2—4 mm diám., de 1.5-2 mm de largo, lóbulos ciliados, 4 a 7 estambres, anteras oblongo-elipsoidales, de 1-1.5 mm de largo, gla- bras, filamentos de 1-2 mm de largo, raquis muy pubescente; flores femeninas solitarias, en pares о en grupos de 4 a 5 sobre un pedúnculo pubescente ~ — 1-1.5 ст de largo: frutos de 1 a 3, sobre un pedúnculo de 2-3 mm de largo; cúpula pateliforme, de 3—4 X no engrosadas en la base, ápice obtuso, pubescen- 14-15 mm diám., escamas triangulares. te, bordes más obscuros: bellota ancha, pared interna del pericarpo lanosa, ovoide de 13-15 X 14-15 mm didm., incluída en la cúpula menos de una tercera parte. Figura 18. Discusión. Quercus elliptica se reconoce por sus hojas elípticas, ápice aristado y envés glabrescente. Distribución y hábitat. En México en los esta- dos de Guerrero, Jalisco, Estado de México, Mi- choacán, Nayarit, Oaxaca, Sinaloa y Veracruz, tam- bién en América Central. Habita en bosques de Pinus—Quercus y de galería, en altitudes de 1500— 1900 m. Fenología. Florece de febrero a marzo y fruc- tifica en octubre. Nombres populares y usos. Encino cucharita, encino tapahuite, encino nanche, encino colorado, encino laurel, encino cáscara. Su madera se usa como leña y para elaborar carbón (González, 1986), también para bancos, muebles rústicos, mangos, cabos de herramienta, vigas de construcción y pos- tes. Ejemplares examinados. MEXICO. Estado de Méxi- co: Tejupilco, Paso del Jate, Parque Cinegético Nanchi- titla, Abundiz 525. 536 (IZTA); Reserva Cinegética de Nanchititla, у=. 1966 (IZTA); en falda \ del cerro, Ma- tuda 31544 (CODAGE d km 2 3 ELC ийни t ae hititla. Rojas & Romero pes A); km 12 El Corupo-Nanchi- titla, Rojas & Romero iir (IZTA); 2 I al NW de Nan- chititla, Rzedowskt 22116 (ENCB); Nanchititla, alrede- dores de la población, Rzedowski 30303, 30309 (ENCB). 19. Quercus hintonii E. F. Warburg, Bull. Misc. Inform.: 91. 1939. TIPO: México. México: Temascaltepec de González, Nanchi- titla, Hinton 6359 (holotipo, K!). Estado de шс ig dg F. Warb.. Bull. Misc. Inform.: 95. ПРО: México. Estado de México: Temascal- — de González, Berros, Hinton 6568 (holotipo. K!). кн эга E. F. Warb., Bull. Misc. Inform.: 94. . TIPO: México. Estado de México: Р рес ae ;onzález, Mina de Agua, Hinton 6576 (hol- otip Temascal- T— Inform.: 90. 1939. Quercus р F. Warb., Bull. Misc. T Distrito de Te- IPO: México. Estado de México: Annals of the Missouri Botanical Garden Figura 18. mascaltepec de González, Cuentla, Hinton 6577 (fo- tograffa del holotipo, K!). Arbol caducifolio, hasta de 15 m de alto; tronco de 30-50 cm diám., corteza con placas cuadran- gulares, en individuos muy jóvenes ésta es rugosa: ramillas jóvenes de 1-4 mm diám., con abundante pubescencia blanco-amarillenta, formada de trico- mas estrellados, que se reduce y ennegrece con el tiempo; lenticelas blancas, de 0.5-1 mm, más no- torias en ramas viejas; yemas ovoides, de color cas- taño obscuro, de 1-8 mm de largo, con escamas coriáceas y pilosas; estípulas oblanceoladas, de 9— 13 mm de largo, membranosas, pilosas en márgenes y base, decíduas; hojas jóvenes con abundante pu- bescencia rojiza que cambia a amarillo en la mad- Quercus elliptica.—A. Rama. —B. Inflorescencia masculina. С. Tricomas. (Romero 3991.) urez; hojas maduras coriáceas, lanceoladas, ovado- lanceoladas, obovadas o elípticas, lámina 5-21 X 3-10 cm, ápice aristado, corto a largamente acu- minado, base obtusa, redondeada, cordada o aguda, con frecuencia asimétrica, borde aristado, dentado, a veces entero, revoluto, hasta con 7 aristas en cada lado, es frecuente que éstas se presenten en un sólo lado; las nervaduras se prolongan hasta las aristas, las primeras de 6 a 14 en cada lado, rectas o lig- eramente arqueadas, ascendentes, ramificandose en el borde; haz verde claro, lustroso, glabro excepto en la base y en la nervadura central, en hojas in- maduras es rojizo a verde obscuro, con abundante pubescencia formada por tricomas estrellados y simples glandulares, con la madurez la pubescen- Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de Mexico cia disminuye en cantidad y cambia de rojiza a blanco-amarillenta; nervadura central a veces algo elevada, nervaduras primarias y secundarias im- resas; envés lanoso tomentoso, con tricomas es- trellados de aproximadamente 1.5 mm de largo, con las ramas entrelazadas, epidermis lisa, con abun- dantes tricomas simples glandulares, de color ám- bar, el tomento cambia de blanco en hojas jóvenes, a amarillo en las maduras, nervaduras conspicua- mente elevadas; pecíolo de 0.7-2.9 X 1-2 mm, lanoso tomentoso, con frecuencia más ancho en la base; amentos masculinos de 3-10 cm de largo con muchas flores, raquis con abundante pubescencia, perianto pubescente de 1-2 mm de largo y de 2 o 3 mm diám., amarillento, con frecuencia los bordes ~ son rojos, estambres 6 о 7, exertos, anteras oblon- gas, glabras, apendiculadas, de 1-1.5 X 1 mm, con € frecuencia rojizas, filamentos hasta de 2 mm de largo: amentos femeninos de 1—14 cm de largo, con 1 a 6 flores, raquis pubescente; perianto de 2 X 1.5 mm, amarillento: estigmas 3 o 4, espatulados, de color rojo obscuro; frutos 1 a 4 en pedünculos de 6—7.5 mm: ctipula poculiforme a pateliforme, de 10-14 mm diám., de 4—7 as con ápices de color castafio, pubescentes; bellota mm de alto, escamas lax- globosa a comprimida, pared interna del pericarpo lanosa, de 6-15 mm de largo, de 9-13 mm Фат. con el ápice plano a umbonado. Figura 19. Discusión. Quercus hintonii se reconoce por sus hojas generalmente lanceoladas, aristadas, con el envés densamente tomentoso y epidermis lisa. Q. hintonii muestra similitud con Q. crassifolia y Q. dysophylla (ver descripción de Q. crassifolia). Distribución y hábitat. En México en el Estado de México. En bosques de Quercus—Pinus, se aso- cia con Quercus magnoliifolia, Pinus oocarpa, P. pringlei, Clethra mexicana y Juniperus flaccida, en altitudes de 1300-1950 m Fenología. Florece en marzo y los frutos mad- uran en junio a agosto. Nombres populares y usos. Encino prieto. Su madera es empleada localmente para la ela- boración de mangos de herramientas, vigas, postes de cerca, bancos rústicos y leña. Podría utilizarse para fabricar duelas de parquet, lambrín, muebles, durmientes, pilotes para minas y barricas (Arcia, — O. Estado de Méxi- Matuda & col. Ejemplares examinados. со: Ата сегсапѓаѕ de alepec, 29823 (CODAGEM); Clachic — 7 km al NE, Pineda 1048 (INIF); 4 km al S de Amatepec, Pineda 1052 (INIF): San Simón de Guerrero, km 83 Toluca-San Diego Cuentla, Rojas, Romero & — 3218 ep ТИТА); Mina de Agua, Rojas & Romero 3605 (ENCB, Cuentla, Rojas & гич 3606 (ЕМС 3 a A); Sultepec, > تو = = = 87 m , km 25 carretera a San Miguel кокк Torres 575 IZTA); Tejupilco, Cerro de Nanchi sarcía s.n. (IZTA); эч ое le la poblac ión ds Tejupilco, - NIF); 5 km al SW de Nanchititla, González PH ub Pefia Bonita, González 5399, 5444, 5 (MEXU); Potrero Chicc — 36, 5438 (MEXU); Reserva Ecológica de Nanchititla, — 318 (IZTA); 5 km al NE de Palos Prietos, Pineda 4 (INTE); Los Cuervos, Rojas & Romero 36 ENC P Te- mascaltepec de Gonz — Chorrera, Hinton he (ENCB); Pant ija, Hinton 6225 (ENCB); km 70 carretera Federal Rojas & T 3400 de Temascaltepec de jupilco, Rz Bead "E (IN 'anías, Matuda & ‹ sd GEM). ENCB, IZTA); 5 km с^ —. ' 29827. 31588 (CODA- 20. Quercus laurina Humboldt & Bonpland, PI. Aequinoct. 2: 32, pl. 80. 1809. TIPO: México. Hidalgo: Cerro de las Navajas, cercanías de Morán, Bonpland 4143 (isotipo, B!). Quercus — is Benth., Pl. Hartw. 56. 1840. TIPO: México. Hidalgo: Real del Monte, Anueg 2 ). Quercus ‘ourgaet Oersted ex оаа Мет Acad. Sei. 20: 2m pl. 366. 1924. TIPO: México. San Ni- e de México, m он 1013 Quercus lanceolata Humb. & Bonpl., Pl. Reauinoot. 2: 34, pl. 81. 1809. TIPO: México. Morán a Santa Rosa, Bonpland s.n. (B). Quercus caeruleocarpa, T em. Natl. Acad. Sci. 20: 163, pl. 321. 19 nd México. Distrito Federal: Contreras, 9 * Te Endlich 1365 a (B) Árbol de 10-30 m de alto, tronco de 50 ст diám. o más, corteza con grietas poco profundas y piezas chicas, de color gris obscuro; ramillas de color gris, de 1-2.5 cm diám., pubescentes al principio, des- pués glabrescentes, pubescencia de tricomas es- trellados pequefios; lenticelas menores de 0.5—3 mm de largo, de color claro u obscuro; yemas de 1.5—4 mm de largo, ovoides, agudas, de color cas- tafio, escamas gruesas con el margen apical ciliado; estípulas de 3—6 mm de largo, oblanceoladas o sub- agudas, escariosas, membranosas, decíduas; hojas jóvenes con pubescencia formada por tricomas es- trellados pequefios y simples dispersos en haz у envés, con el tiempo ésta se reduce a las nerva- duras central y primarias; hojas maduras rígidas, coriáceas, lustrosas, lanceoladas o elíptico-oblan- lámina (2-)5-11(-15.5) X 1.5-4(-6.5 cm, ápice agudo o acuminado y por lo general ar- ceoladas, istado, base redondeada, cordada, atenuada o cor- dada, borde entero o dentado, plano o revoluto, a veces ondulado, engrosado, con 1 a 5 aristas de cada lado, a veces se presentan sólo de un lado de la hoja, distribuidas en el tercio superior; nerva- duras primarias de 4 a 12 en cada lado, rectas о ligeramente arqueadas, ascendentes, se ramifican antes del margen; haz verde lustroso, todas las ner- vaduras pálidas y elevadas, nervadura central pu- 586 Annals of the Missouri Botanical Garden Figura 19. Quercus hintonii.—A. Rama. —B. Fruto. С. Tricoma. —D. Hoja. —E. Flores masculinas. (Romero 3604.) bescente en su base; envés lustroso, amarillo o a veces más pálido que el haz, pubescencia de tri- comas estrellados estipitados, restringida por k general a las axilas de las nervaduras primarias, donde se observan tricomas simples, blancos y es- casos tricomas glandulares de color ámbar, epider- mis papilosa, nervaduras elevadas; pecíolos de (2—) 25-15-22) mm de largo, de 0.5-1.5 mm diám., fin- amente pubescentes, glabrescentes; amentos mas- culinos de 3.54.5 em de largo, raquis negro con Volume 89, Number 4 2002 Romero et al. 587 Quercus en el Estado de México Figura 20. Quercus laurina.—A. Rama. —B. Fruto. (С. —C: Torres 1468; D, E: Romero 3321.) tricomas largos, perianto de 2-3 mm diám., con los lóbulos largamente ciliados, anteras de 1.5 mm de largo, filamentos de | mm de largo: fruto anual o bianual, solitario, en pares o en grupos de 3, sésiles o en pedúnculos de 3-12 mm de largo: cúpula hemisférica. de 10—15 mm diám., escamas leñosas, Tricomas. —D. Hoja. —E. Inflorescencias masculinas. no engrosadas en la base, pubescentes, de ápice obtuso: bellota ovoide, pared del pericarpo lanosa, de 7-19 mm de largo, de 7-12 mm diám., incluída un tercio de su largo en la cúpula. Figura 20. Discusión. Quercus laurina se reconoce por sus 588 Annals of the Missouri Botanical Garden hojas aristadas, el envés con pubescencia restrin- gida a las axilas de las nervaduras. Esta especie posee similitud morfológica con Q. affinis y se dis- tingue por poseer yemas florales de forma conoidal, base de las hojas cuneada y nervaduras planas; mientras que Q. laurina posee las yemas ovoides, base de las — atenuada o redondeada y nerva- duras elevadas. Distribución y hábitat. En México en los esta- dos de Distrito Federal, Guanajuato, Guerrero, Hi- dalgo, Jalisco, Estado de México, Michoacán, Mo- Querétaro, Zacatecas. En bosques de Pinus—Quercus, relos, Puebla, Oaxaca, Tlaxcala y Pinus— Abies, bosque mesófilo de montafia y encinares per- turbados, en altitudes de 2240-3150 m. Fenología. | Florece en abril y fructifica de junio a octubre. Nombres populares y usos. Chilillo, encino co- lorado, atlualpitzahul, encino blanco, encino lau- relillo, encino roble, encino xicatahua, tesmolera, encino hoja angosta, huitzalacate, encino prieto, encino uricua, encino chilillo. Se utiliza para bancos, muebles rásticos, cabos de herramienta, vigas de construcción, papel Kraft y fabricación de chapa (González, 1986) y como lefia. Ejemplares examinados. MEXICO. Estado bee Méxi- co: Acambay, cerro Detiña, шо 1468 (IZTA); Muy- teje, Quintero & Rojas 18, 22 (IZTA); Ar ne de Be- cerra, 5 km al NW LARUM Be ndita, Pineda 1347 (INIF); 4 `сатеса, barran Antonio — Ern 159 (ENCB); Alcansi, y & Mancera s.n. (ENCB); 1 km al E de San Antonio Zoyatzingo, ا‎ 2643 32 (ENCB); 1 km al E de San Antonio, base del Ixtaccihuatl, Xelhuantzi 5081 (I F); — cumbre del Cerro Chil- иса, Rzedowski 32611 (ENCB); Atlautla, Tecomazusco, camino de Atlautla, al paraje de s Santa Teresa, Herndndez 40 (INIF); Capulhuac, Cerro l km al W de San Miguel Almaya, Rodríguez s.n. (ЕЛ NC SB 6 al 5 de Coacalco, vertiente N de la е 30768 (ЕМСВ); Sierra det Guadalupe, € d ~ alco, Rzedowski 30774 (ENC 15 — al i rimas, geste s.n. Coatepec, Опе ra 42 (ENCB); 5 ES al E de Coatepec, Román 315 (ENC * Chalco, cerca de la colonia Agrícola M: vila Camacho, Cabrera s.n. (ENCB); Chicoloa- pan, 5 km al E de Coatepec, — 311 (ENCB); El Oro de Hidalgo, La Cima, Pérez 19 (IZTA); La Cima, Valdés 1 (1Z i : Sd Matuda 28101 (ENCB); Ixtapaluca, Estación Experimental de Enseñanza e Inves- + ión de Zoquiapan, е 1160 (ENCB); 1 km al » de Llano Grande, Rivas & Campos 13 (ENCB); 6 km 2s SW de Río Frío, Roe & Scott. 1440 (ENCB); Ji- quipilco, vertiente W de la Sierra de las Cruces, Alvarez 8, 15 (IZTA); km 15 carre : ra Jiquipilco-Villa del Carbón, Romero & por 3176 (IZTA); Jilotepec, San Bartolo, Ma- tuda & col. 29063 (ENC ^s Santa Ana — 3kma E de San — cen in, Fernández 36 (ENCB); Lerma, Santa María Tlalmimilolpan, cerca de Lerma, Franco 63 | 7 1 erra * Guadalupe. era m 2° (ENCB); cerca de calpan de — Villa Guarda, Herr 19] Salazar, Villanueva s.n. (ENCB); Nau- Alpina, Corona s.n. (ENCB); El (ENCB); Santiago Tepatlaxco, Jiménez 132 (IZT A); Los Remedios, Maarigal 74 (INIF); San Francisco Chimalpa, Ruiz 33 (1 Ocuilan de Ar- teaga de Arteaga, km 18 terracería Ocuilan de Arteaga— Cuernavaca, Castañeda, Lugo & Trejo 13 (VZTA); San Mar- tín de las Pirámides, Cumbre del Cerro Gordo, Rzedowski 18810 (ENCB); Sultepec, La Cieneguilla, — 163 (ENCB); San Miguel йш э y Torres 3 302 (IZTA); Te- 1 & Tejero s.n. (ENCB) zález, un al de X PASE IZTA); Comunal de Tequisquia- pan, Huerta T-8 (ENCB); Teotihuacán, Cerro Ge јат al de Teotihuac ‘in, Espinosa 635a (ENCB); Texcoco, 24 km al E de Texcoco, sobre brecha rumbo a Tlaloc, García s.n. (ENCB); 8 * al SE T Tequesquinahuac, Perino 3137 (ENCB); Сай: ada al 5 masc — tura 1123 (INIF); Timilpan, — de San Andrés, Ca- macho 248 (IZTA); Tlaln 1analco, (INIF); 4 km de San Rafael, Magafia 1057 (NIF): a Medellín 227 (INIF); 3 km al S de San Rafael, Ramá ENCB); Villa del Carbón, El Plan, González s.n. ( Т А), El Gato San Jerónimo, Rojas «€ Romero 3168 (IZTA); Villa Nicolás Romero, Transfi- "uración, Ocaña s.n. (УТА); Zacualpan, cerro La Corona, Машаа 30324 (CODAGEM). 21. Quercus mexicana Humboldt 9 Bonpland, Pl. Aequinoct. I: 35, pl. 82. 1809. TIPO: México. Guanajuato: Santa а Bonpland 4218 (B) — rugulosa Mart. & Gal., Bull. Acad. Roy. Sei. — Vol. 10, pt. П: 209. 1843. TIPO: México. edro y San Pablo, cerca de Real del Monte, XM 116 (BR — Árbol de 3-15 m de alto, corteza gris; ramillas de 1.5-2 mm diám., al principio con pubescencia formada de tricomas estrellados, pronto glabrae, corteza gris, con lenticelas menores de 0.5-2(-3) mm de largo, casi del mismo color que la corteza; yemas de (1-)2-3(-6) mm de largo, ovoides, ápice agudo, de color castaño, escamas pubescentes en sus márgenes; estípulas de 3-5 mm de largo, es- cariosas, lineares, caedizas, algunas persistentes por un tiempo cerca de las yemas apicales; hojas jóvenes del mismo color que las adultas; haz con el indumento formado de tricomas estrellados muy pequefios dispersos, envés con tricomas estrellados dispersos; hojas maduras decíduas, elípticas, lan- ceoladas u oblongas, coriáceas, lámina (2-)3-9(- 12) X (0.9—)1.5—3.5(—4.2) em, ápice agudo, suba- gudo o redondeado, con una arista de hasta 2.5 mm de largo, base ligeramente revoluto, engrosado, car- tilaginoso, borde entero; nervaduras de 6 a 12 en cada lado, ascendentes, casi rectas, se bifurcan en el ápice; haz verde obscuro, con algunos tricomas estrellados dispersos en la base de la hoja y en la Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México SAN S ESTA — [A es. RAH 9 B sp. ek A Figura 21. Quercus mexicana.—A. Rama. —B. Fruto. nervadura central, ésta impresa o ligeramente ele- vada, las secundarias impresas; envés ligeramente más pálido, con pubescencia de tricomas estrella- dos contortos, que se observan como puntuaciones, uniformemente distribuidas, epidermis papilosa: pecíolos pubescentes, de (2-)3-8 mm de largo, de 0.5-1.5 mm diám.; flores desconocidas; fruto anual, solitario o en pares, sésiles o en pedúnculos de 2— 9 mm de largo; cúpula hemisférica, de 10-13 mm diám., con las escamas delgadas, no engrosadas en la base, pubescentes, con el ápice subagudo o re- dondeado, márgenes por lo general glabros: bellota ovoide, pared interna del pericarpo lanosa, de 9— 15 mm de largo, de 8-11 mm diám., incluida un tercio o la mitad de su largo en la cúpula. Figu- ra 21. Discusión. (Quercus mexicana se reconoce porque sus hojas tienen bordes enteros, ápice ar- istado y envés con tricomas contortos que se ob- — C. Tricoma. (Quintero 30.) servan como puntuaciones. Quercus mexicana muestra similitud con Q. crassipes, esta ültima se distingue por no presentar los tricomas contortos a manera de puntuaciones. Distribución y hábitat. En México en los esta- dos de Chiapas, Distrito Federal Guanajuato, Hi- dalgo. Estado de México, Nuevo León, Puebla, Querétaro, San Luis Potosí, Tamaulipas y Veracruz. En bosques de Quercus y Quercus—Pinus, también en suelos erosionados y a orillas de arroyos, se aso- cia con Pinus montezumae, P. leiophylla, P. teocote y P. rudis, en altitudes de 2230—3050 m. Fenología. | Fructifica de julio a enero. Nombres populares y usos. Encino, encino te- zahuatl. Hasta el momento no se tiene conocimiento so- bre sus usos. Fjemplares examinados. MÉXICO. Estado de Méxi- co: Atizapán, 4 km al N de Atizapán, Castilla 2725 Annals of the Missouri Botanical Garden (IZTA); Coyotepec, Sierra de Alcaparrosa, Rojas & Romero 156 (IZTA); о Fraccionamiento La Herradu- ra Viel. ER s.n. ОМІР); Jilote pec. Ripe Matuda 26742 (CODAGEM); Las Peñas, Rojas € Romero 3076 (ZTA): Ojo de Agua, Rojas & Romero 2 — = mascalapa, Sierra de Guadalupe Р. 3 Padres, Ortíz s (IZTA); Cerro Gordo, Tejero & Castilla 634 zT А); Te. potzotlán, Arcos del Sitio, Ceballos 2 (IZTA); Villa del Carbón, cortina de la — San Luis Taximay, Tejero & Abundiz 2766 (ZTA): S in Luis Taximay, Tejero & Castilla 2723 (IZTA); Villa Nicolás Re m | km al E de Lan- zarote, Rzedowski 31877 (INIF); З km al NW de Cahu- асап, Rzedowski 32628 (INIF). 22. Quercus seytophylla Liebmann, Overs. Danske Vidensk. Selsk. Forh. Med- 180. 1854. TIPO: México. Oaxaca: Yalala a Yagochi, 144— 6 =3557 (C) Kongel. lemmers Arbeider: Liebmann Árbol de 8-20 m de alto, de 30-50 em diám.: ramillas de 1.5-3 mm diám., grises a castaño rojizo o negras, finamente pubescentes, con tricomas pe- queños estrellados; lenticelas pálidas, de 0.5-3 mm de largo; yemas de 1-4 mm de largo, ovoides o elípticas, escamas pubescentes con los márgenes ciliados, engrosadas en la base, de color castaño: estípulas lineares, pubescentes, caedizas, membra- nosas, de hasta 5 mm de largo; hojas jóvenes pu- bescentes, haz verde amarillento con tonos cafés, con numerosos tricomas pequeños, estrellados, se- mejantes a los que se observan en las ramillas, tri- comas de mayor tamaño se concentran en la ner- vadura central; hojas | maduras — coriáceas, generalmente obovadas, elípticas, oblanceoladas o lanceoladas, lámina (6.2—)8-14.5(-19.5) х (2.3-) 4—7(—8.5) em, ápice con un diente alargado, aris- tado, base obtusa o cuneada, a veces oblicua, borde engrosado, revoluto, cartilaginoso, con Та 7 dientes aristados a cada lado, casi desde la base de la hoja, aristas hasta de 4 mm de largo; nervaduras 5 a 9 primarias en cada lado, rectas о arqueadas, anas- tomosándose cerca del borde. pasando directa- mente al diente y luego a la arista, a. veces las divisiones de las primeras lo hacen también; haz verde grisáceo, no lustroso, glabro, excepto en la base de la nervadura central, generalmente los tri- comas ennegrecidos, nervaduras impresas, el as- pecto general es rugoso; envés densamente pubes- cente, blanquecino, que cambia a amarillento y que se ennegrece con el tiempo, indumento formado por tricomas estrellados con más de 10 rayos, sésiles, epidermis ampulosa y papilosa; nervaduras central y secundarias conspicuamente elevadas, blancas; pecíolos de (0.7—)1—4 em de largo, de 1-2 mm diám., se engrosan hacia la base, con fina pubes- cencia gris; amentos masculinos laxos, perianto de 1.5-2 mm diám., rlabros, anteras de 1.5 mm de B largo, filamentos de 1.5-2 mm de largo: fruto anual, en grupos de dos, sobre pedünculos de 5-15 mm de largo; cúpula hemisférica, de 10-12 mm diám., escamas delgadas, pubescentes, largamente cilia- das en los bordes, ápices redondeados a agudos, adpresas; bellota ovoide, de 8-10 mm diám., de 10 mm de largo, incluída de un medio a dos tercios de su largo en la cúpula. Figura 22. Discusión. Quercus scytophylla se reconoce por sus hojas dentado-aristadas, haz opaco y envés con pubescencia blanca. Distribución y hábitat. En México en los esta- dos de Durango, Guerrero, Jalisco, Estado de Méx- ico, Michoacán, Nayarit, Oaxaca, Puebla, Sinaloa y Sonora. En bosque de Quercus, bosque de Pinus— Quercus y bosque mesófilo de montaña, se asocia a Arbutus, Pinus oocarpa, Pinus teocote, en altitudes de 900-2500 m Fenología. Florece en febrero y fructifica en octubre. Nombres populares y usos. Encino prieto, en- cino blanco, encino rosillo. Su madera se utiliza para la elaboración de tra- bajos artesanales, así como fuente celulósica de pulpa para papel (González, 1986). MÉXICO. Estado de Méxi- co: Amatepec, entre Amatepec y Sultepec, Matuda & col. 29832 (CODAGEM); Nuevo Santo Toni: di los Plátanos, — E col. 29387 (ENCB); Sultepec, El PAREN: POINTE examinados. erca de Sultepec, Madrigal 17 (INIF); Las Trojas, ríguez 179 (МЕ); Sulte — Rodriguez 181 (ENCB): Tejupilco, dea Chico, 15 km a ' de Nanchititla, González et al. 5436 (E NCB): Рейа Bonita, Nanchititla, Ramos s.n. a NC B); 2 km al NW de Nanchititla, Rze- dowski 2 7 (ENCB); Temascaltepec de González, km 65 le ( — Toluca-Temascaltepec, Casillas (IZTA); Real de Arriba, Huerta, nos Rar & col. RA-6 (ENCB, INIF); presa La Carbonera, Nah 3652 (ENCB, INI); km 30 de la — ra federal 130, cercanfas de ‘ ZTA); Real de Arriba, Ramos «€ Rocha 96 (INIF); n 51 de la carretera Toluca-San Diego Cue tla, Rojas & МА 3212 (INIF); Valle de Bravo, Huerta & col. CV-23 (ENCB, INIF); Valle de Bravo, aeg 2604 de NCB); Zacualpan, km 6 de la carretera Zacualp: p е 152 (IZTA): Lac — Маш & are 30298, 30299 (E 3 de la carretera Campana de Plata-Subestación C oronas, Timo & Castilla 1734 (IZ TA). Casas Viejas, eer 23. Quercus urbanii Trelease, Proc. Amer. Phi- los. Soc. 60: 32, pl. 2. 1921. TIPO: México. Michoacán a Guerrero, Sierra Madre, 20 June 1899, E. Langlassé 1066 (B). Árbol de 4—10 m de alto, tronco de 20-30 em diám.; ramillas gruesas, con costillas, de 5-11 mm diám., densamente pubescentes, pubescencia amarillenta, gris-amarillenta o negra, formada por Volume 89, Number 4 2002 Romero et al. Quercus en el Estado de México Figura 22. tricomas estrellados estipitados con las ramas er- ectas y por abundantes tricomas glandulares sobre la epidermis, corteza castaño-rojiza a negra; len- ticelas inconspicuas debido a la pubescencia, en las ramas таз gruesas éstas miden hasta 4 mm de largo; yemas de 5—10 mm de largo, ovoides, color castaño, las escamas engrosadas en la base, gla- brescentes las exteriores, densamente pubescen- tes las interiores; estípulas de 7 mm de largo, de | mm diám., pubescentes principalmente en la base y bordes, persistentes en las yemas; hojas jóvenes pubescentes, haz cubierto por tricomas glandulares rojizos, tricomas simples dispersos y tricomas estrellados largos, estos últimos princi- palmente en los bordes, envés con pubescencia blanca formada por tricomas estrellados muy lar- gos: hojas maduras, rígidas, gruesas, pandurifor- mes, obovadas, suborbiculares, orbiculares u ova- Quercus scytophylla.—A. Rama. —B. Fruto. —JC. Tricomas. (Orozco 49b.) do-elípticas, lámina (12-)15-30 X (12-)17-34 cm, a veces más larga que ancha, ápice obtuso, a veces algo escotado, base profundamente cordada, borde revoluto, grueso, cartilaginoso, entero, den- tado u ondulado, con (5— — 10-20 aristas por lado, en las % partes superiores, a veces desde la base, aristas de hasta 4 mm de largo: nervaduras de 9 a 11 en cada lado, ascendentes, algunas forman directamente las aristas, la mayoría se ramifican cerca del borde; haz verde pálido, algo lustroso, rugoso, glabro, excepto en las nervaduras princi- pales y primarias en donde se encuentran tricomas glandulares simples y estrellados, nervaduras im- presas a ligeramente elevadas; envés con tomento de aspecto lanoso, amarillento, formado por tri- comas estrellados, estipitados con rayos de hasta 3 mm de largo, enredados, cubriendo uniforme- mente la epidermis papilosa y glandulosa, nerva- 592 Annals of the Missouri Botanical Garden Figura 23. duras elevadas; pecfolos de 3—4 cm de largo, de 2—4(—6) mm diám., tomentosos, se ennegrecen con la edad; amentos masculinos de 13-23 cm de lar- go, raquis densamente tomentoso, perianto de 6 mm diám., glabro, excepto los bordes de los lób- ulos que son ciliados, anteras apendiculares, de 2-2.5 mm de largo, filamentos de 2 mm de largo; amentos femeninos de 10 a 20 flores en pedün- culos gruesos, con tomento abundante de color amarillo; frutos en grupos de 5 a 10 o más, en Quercus urbanii.—A. Rama. —B. Inflorescencias masculinas. С. Tricoma. (Torres 197; González 5017.) pedünculos de 7.5-9 cm de largo, de 4-5 mm diám., con denso tomento amarillento que se ob- scurece y cae con el tiempo; cúpula de 8-12 mm diám., escamas pubescentes con los ápices redon- deados a truncados; bellota ovoide, pared interna del pericarpo lanosa, de 10 mm de largo, de 8— 10 mm diám., incluída de un tercio a un medio de su largo en la cápula. Figura 23. Discusión. Se reconoce por sus hojas grandes, Volume 89, Number 4 2002 Romero eta 593 Quercus en К Estado de México acucharadas, el tomento del envés amarillo y sus ramillas gruesas Distribución y hábitat. Еп México en los estados de Durango, Guerrero, Jalisco, Estado de México, Nayarit, Sonora y Zacatecas. En bosques de Pinus— Quercus y bosque mesófilo de montafia, asociado con Quercus scytophylla, Q. crassifolia, Q. obtusata, Q. laurina, Pinus montezumae, pseudostrobus, P. oocarpa y Alnus, en altitudes de 1400-2500 m. Fenología. Florece en diciembre y fructifica en octubre. Nombres populares y usos. Encino. Su madera se usa como leña (González, 1986); se sabe que también es utilizada para la elabora- ción de arados y los frutos sirven para alimentar ganado porcino (Vázquez, 1992). — examinados. MEXICO. _Estado de Méx co: Sultepec, entre Sultepec y el paraje El C E Madrigal 25 (ENCB, INIF); La Cieneguilla Rodríguez I (INIF); km = carretera Sultepec-San Miguel — polora Torres 197 (IZTA); km 18 carretera Sultey коо. Torres 428 ; Tejupilco, 5 5 km al Sw ER Nanchititla, — 5017 (ENCB); ea hormi- gueros, Nanchititla, ández 1747 (INIF); 12 km de El Corupo a Nanc t эң "Rojas & Romero 3955 EZTA): Puerto del Embocadero, 7 km al W de Luvianos, Rze- dowski 22126 (ENCB, INIF); Nanchititla, alrededores de a Población, Rzedowski 30307 (ENCB); Zacualpan, Zac- ualpan, Matuda 30314 (CODAGEM). Literatura Citada Aguilar, E. & S. Romero. 1995. cuatro especies de encino e descritas por War- 31: Estudio taxonómico de . 1979. Anatomía y RUM Físicas de la ra de Tres = m e En ncino del Estado de Méx- » Tesis. Esc. Nac a M. & J. aa. 1987. Estado de Michoacán, México. Centre d'E s caines et жин ре Instituto Nacional de In- vestigaciones Mia ales, México. . Estudio del Uso del Bosque para la Extracción d i Madera para Construcción de Ca- sas y Fabricación de Herramientas en una Comunidad Otomí. San Andrés o = de México. Tesis de Licenciatura, UNAM, о. Los Encinos — del Mexi- — A. 1938. Les chénes, monographie du genre Quer- s. Paul Lechevalier, Paris. Chin no, V. & Р. Jacques. 1986. Contribución al Conoci- miento de la Flora Medicinal de (denen, Puebla. Tesis de Lic ii iatura, De La Paz O., C. 1976. Características Anatómicas de Cinco Fins de México. Bol. Téen. Inst. Nac. Invest Forest. No. 46. México. 19 82. — tructura Anatomica de Cinco Especies del Gina ro Quercus. Bol. Técn. Inst. Nac. Invest. For- esl., México, No. 88. . 1985. Características Anatómicas de Siete Es- к: ies del Сёпего Quercus. Bol. Тёсп. Inst. Nac. Invest. . México, No. — V. 1986. Сони ion al Conocimiento del Gé- nero — (Fagaceae) en el Estado de Jalisco. Inst. Bot. Univ. Aut., Guadalajara, Méx xico. INEGI. 1987. Sintesis Geográfica. Nomenclator y Anexo Cartográfico del Estado de México. Secretaria de Pro- gramación y Aag a sto, México. Martínez, M. 1954. Los Encinos del Estado de México. Conia ba — Exploradora del Estado de México. Gobierno del Estado de México, Dirección de Agric 'ul- tura y Ganadería, Toluca, Estado de México, México. 1974. Flora Novo-Galiciana. Contr. Univ. 2 . & R. Mc Vaugh. 19 1972. The oaks (Quercus) de- sc led by Née (1801), and {е аА & Bonpland (1809), with comments on related species. Contr. Univ. 522. c classification of — Ann. 934. Infrageneri (Fagaceae) p b rH ation. of sectional. names Sci. Forest | — genus (шеси m in Mexico. Pp. 447— 458 en T. P. Ramamoorthy, . Lot & J. F EE Biological Diversity a Mexic ‘o: Origins and Distribution. Oxford Univ. Press, Oxford. The Tepehuán of — Utah Press, Salt La e 1. W. 1969. eir Material Culture. Univ. City. Romero, R., C. s & S. Gómez. 2000. Flores hermaf- roditas de — glaucoides Mart. & Gal. (Fagaceae) = el Estado de Michoacán, México. Acta Bot. Mex. 52: 9—54. — J. 1978. Vegetación de México. Ed. Limusa, éxico. e Rzedowski. 1979. Flora ded iy ү” p Méxic o 1: 104-114. Ed. Continental S Teese Ww. 1924. The American oaks. Mem. Natl. Acad. . 20: 1-55, pl. 1-420. El Género Quercus т en el Tesis, UNAM, México ушш. ү. 1993 Estado de Puebla: ANNALS OF THE MISSOURI BOTANICAL GARDEN: CHECKLIST FOR AUTHORS SCOPE The Annals publishes original articles in systematic botany and related fields. Papers whose purpose is the establishment of new nomenclatural plants and bry Rather, they should be submitted to Novon for consideration. (Noron phytes are not accepted. instructions are available on the Web at www.mobot.org or from the managing editor.) CONDITIONS FOR PUBLICATION Authors are urged to contact the — editor prior to —— manuscripts. Because the dis file will be used by the printer for мерна it is im- portant to des to the items listed in the format section of the checklist. Authors’ electronic files are formatted by the editors and are sent t tte/electronic the printer in electronic form (not as camera-ready page Manuscripts that have ge n reviewed before submission will be subject to the full normal review process initiated here. Manusc ‘ripts not properly — may be ише! for revision prior to review. It is the author's responsibility missing ашын SA nen su the editor does not pe огт library researc to е = pt 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 Page costs are $80 per page. although charges may sometimes be reduced. Changes in proof made by authors will be billed non-negotiably to the author at $3.50 per line changed. CORRES TUNNE, Amy McPherson, Managing Editor * Press, P.O. Box 299, St. US а — MBG Press, Monsanto Center, Blvd., St. Louis, MO 63110-229 Tel: (314) 577-5112; Fax: (314) 577-9594. Email: annals@mobot.org Louis, MO 63166-0299, 4500 Shaw www.mobot.org 1. General [] Text is in English or Spanish on numbered pages. LJ Manuscript is submitted on MS-DOS/WINDOWS 3% in. diskette, as Microsoft Word? [preferable] file. L] Three copies of double-spaced priis d manuscript, in- cluding abstract, legends, tables, bs imen lists, Li erature Cited, and footnotes, are enclosed. The printed manuscript is the same as the — on the dis- elte. [|] Photocopies of the figures are included with each copy of the manuscript. Copies of half-tone figures (photo- graphs) and plates must be of sufficiently good quality to show all the detail of the originals. Digital or dye- sublimation proofs of electronic graphic files are sub- entities in vascular O0 0 Û Û OO O OT mitted. [Send originals or photostats of originals when manuscript is accepted for publication. Authorities, abbreviated according to Brummit & Pow- ell's Authors of Plant Names, are pro names the f vided for species’ t time they are mentioned in the text; OR 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 2. Format of Text Right-hand — is not justified, and words are not hyphenated ther One space is ins — after periods, colons, question = marks, and all other punctuation mar First hs аге indented with a tab. Iwo hyphens with no space between them are used to indicate ranges of n bers 2500 m. A single hyphen is used in hyphenated words, for ex- ample, 9-nerved, oval-triangular. t sentences of paragrap umbers, for example, page num- ; measurements, 5—7 mm; altitudes, 1500— Common Latin words or phrases are not italicized (e.g. et al., i.e., sensu, etc. Only taxon names at the rank of genus and below are italicized. Caria accents, umlauts, and other diacritical marks have been included. All figures and tables are cited in the text and are numbered in the order in which they are to appear. 3. Style Recent issue of the Annals is used as a model. Chicago Manual of Style, latest edition, is used as a reference 4. First Page Footnotes are typed as double-spaced paragraphs on the first page. The first footnote contains a ments, including information on granting agenci herbaria that loaned specimens, and the name of he . The second footnote is the author's address. Ad- — for additional authors are each separate foot- notes. No other footnotes occur, except in tables, where unavoidable. artist Abstract & Key Words A one-paragraph abstract precedes text. Papers in Spanish have an English abstract in addition} to a Spanish resumen. The abstract is concise (1 paragraph) and includes brief statements about the paper's intent, materials and methods, results, and significance of findings. A brief list of ke y words immediately follows the Ab- stract. Volume 89, Number 4 2002 595 [] O O O O O O O O O O Taxonomic Treatment Species entries are organized as follows: Heading, Lat- in diagnosis (if necessary), description, distribution, vernacular name(s), discussion, specimens examined. The discussions are parallel and follow the same order, e.g., diagnostic characteristics, distinction from similar species, variation, distribution and ecology, nomencla- ture and typification, uses. One paragraph per basionym is used as follows: Taxon author, — citation, Spa citation, e.g ch- miedia latifol — s) Sa. Nishida, < Mistoun Bot. Gard. 86: 68 van Joni hil Nees, Syst. Laur.: 674. hu E: Peru. Locality not in- 1835 (fl), dar 1433 — Е!; еп). dicated, iso- ypes, BM!, E!, K!, LE not seen, OXF not se Synonyms based on different — are placed i in sep- arate paragraphs, each beginning with followed by other combinations (if appropriate). and the basionym, citation of the A brief Latin — for each new taxon is provided rather than a complete Latin description. s with infraspecific taxa: D ae and discussion are composite (incorporating all infraspe- For species wil cific taxa) and parallel with other species desc a Descriptions of и taxa are parallel with one another (in the same species). All — are listed under the ade infraspecific taxc Desc poe Descriptions are lle sin a given ank — easurements are metric. Hyp ade a 1 . Hyphens are used ir pare do extreme 2-)14.: 8.0(-31.9) cm long,” ee values are not expected: ovary with (2)4(6) loc * Length X : lamina 36.4— intermediat — are given in the following manner J.1-16 W he n Go ant, nomina nuda, misapplied names, and superfluous names are included in the discussion fol- lowing the description, but are not part of the formal synonymy. Citation of Types Exclamation points are used for specimens examined, and types not seen are indicated as such (e.g., ! US not seen Lectotype за, are included together with an in- dication of where and the thor. This arate is listed in the ie of the paper inue is making the lectotyp- ‚ were — the year, e Literature Cited. If ification, the phrase “here designated” is used. Tables Tables are neat, double-spaced, and clearly presented. In most cases the printer will typeset these from hard Captions are — double-spaced as paragraphs at the tops of the tables. Each table starts on a separate sheet. O O O O O O 9. Abbreviations Periods are used after all abbreviations (which are minimized) except metric measures, compass direc- and herbarium designations. When dates are given as part of collection information, three-letter month abbreviations are used, t for months with four letters, which are TE out in full. States are not abbreviated, and cities are spelled out. [St., lions, ouis, is acceptable. | Periodicals are abbreviated e to B-P-H (Bo- tanico-Periodicum- Renan and to B-P-H/S (Bo- — -Periodicum- = инат Хирот eviated according to Brummit : M рне мн of Plant Nan Book titles are abbreviated acc dis to Taxonomic ES erature, edition 2, but with initial letters capitalize ook t itles are spelled out in the Literature Cited. Ifa m does not appear in B-P-H or TL-2, or if Th uo are not available, its title is fully spelled out Herbaria are abbreviated according to Index Herba- riorum, edition eight (8). Abbreviated forms are not used for references in the text, except when citing the names of plants. If it necessary to cite a particular page in the text, the form Smith (1998: 12) i 7 s usec 10. Specimens Examined If many specimens were examined, those cited in the text are limited to ca. 1% manuscript pages index to specimens following If there are a large number, an examined is placed at the end of the paper, arra ` the Literature Cited. It is collector, followed by een number, follow 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. Specimens are cited in the text as follows: Additional specimens examined (or Selected specimens examine MEXICO xaca: Sierra San Pedro Nolesco, Talca. 12°37'N, 85° 14W. 950-1100 m, 3 Feb. 1987 (fl), Jer- gensen 865 (BM, G, K, US). [Dates and reproductive status are optional but are omitted from longer lists. | aged alphabetically by ed by Countries are run together in the same dide e.g.. COU E A RY A. Major political division: ... COU N- TRY 1 Major political division: . — par- асга — are used for major continental regions within major political divisions. Specimen Vouchers and Genetic Sequences If paper presents original data, associated herbarium vouchers are cited. [ Vouchers for seed and/or other col- lections should be included — зр Depen- dent on the paper, reference to the inal wild source may be required.| Vouc ‘hes are Шы cited for common names and uses taken from specimen labels Herbarium vouchers state the collector and number, herbarium in which the voucher is located, and a clear annotation that the material represents the voucher for the study in question [] Nucleic acid or protein sequences corresponding to 596 Annals of the Missouri Botanical Garden equal or greater than 50 nucleotides are ~ into an appropriate data bank, e.g., GenBank/EMBL. The accession numbers are provided before ons alion. ong sequences (exceeding two pages) will not be rou- tinely published. Author accepts responsibility for establishing the ac- curacy of information provided. O 12. Keys O s are — and have been checked carefully for consiste ney with the descriptions. Leads of each cou- IL plet are S Dichotomous keys are indented OO Infraspecific taxa are keyed separately, not in species keys 13. Literature Cited The Literature Cited contains all references cited in the text All entries in the Literature Cited are cited in the text. Spelling of author(s) name(s) and years of publication have been double-checked. O BB 0 All entries have been verified against — sources, especially journal titles diacritical marks, and spelling in languages pis than nadie sh. Periodicals are listed authors last name, d s). Year. ps title of article. Journal abbreviated as in B-P-H/S. Volume: pages. No parenthetical part tube ‘rs after lis numbers are given unless each part is paginated separate For more than one кз, his style is followed: au- thors last name, initial(s), second author's initial(s), last name & third author's initial(s), last Books appear as follows: author's last name, initial(s). Year. Full e Title (edited ii Editor), 3rd ed., Vol. 2. Publisher, City of Publicatio Citations i work “in prep.,” unpublished theses and dissertations, and similar references . accents, d as follows: O name. O O- to inaccessible sources have been eliminated or kept to a minimum. They are not necessarily included in the Literature Cit- ad Illustrations [] Author(s) name(s) and figure number(s) are written in pencil on the back of each figure or plate, on both — and review copies Seale bars appear on illustrations, photographs, and maps Magnifications/reductions are not indicated in ions, cap- — High- рЫ гога of the figures аге included with each с A the manuscript. Шо аге — to fit either column width (2 5/8 їп. or ca. 68 mm) or full page width (5% in. or са. 140 mm), or illustrations no larger than 11 X 16 in. ERE Volume 89, Number 4, pp. 453-596 of the ANNALS OF was published on December 9, 2( = 29 X 40 ст) tions is 54 X — . [Maximum size for printed illustra- X 8% in.] Oversized or delicate figures are submitted as photostats. "igures 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. [] Figures are grouped into composite plates when pos- sible: edges of р are abutted. No stripping is inserted between plate or figure seg- ments. [Printer will insert ‘ping C] Edges of figures are squa LJ Maps include reference to — and longitude and are bounded by a fine border. = О Se anning electron micrographs are free of conspicuous charging. Axes on graphs are all labeled. [] Captions provide all explanatory text. Captions are separate from other text, one paragraph for each group of figures, and following the style in current issues of the Annals, C] Symbols on maps are legible and reduction has been taken into consideration. ELECTRONIC ARTWORK—PRINTER’S SPECIFICATIONS . We accept line art, halftones and color figures on disk or via electronic delivery. Please include a printout or email of the file directory that includes the file name(s), size and kind of file. 2. You should discuss figure sizing specifications with the journal prior to disk submission, and the figures on disk should be the final size desired. 3. — пе files should be submitted in TIFF, EPS, or PDF file format. Applic — that are acceptable are Adobe Photoshop, Corel Draw, Adobe Illustrator and Adobe Freehand, but please, if possible, “Save As’ uf, eps, or pdf file. 4. All line art should be scanned at 1200 d.p.i. All вту- scale and color images should be scanned at 350 d.p Color images should be submitted in CMYK mode, not RGB or Indexed Color mode. Include ra screen and printer fonts used within the graphic file. Use only Adobe Type 1 postscript fonts. True Type fonts and System “bitmap” fonts should not be used. If the files have been — резе indicate what compression program has been used. The disk submission should be ace a by a hard copy of the figure that exactly matches what is on the disk in the event that the disk is unusable. ACCEPTABLE. ELECTRONIC 3% 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. Please contact Karen Ridgway, e-mail address: kridgway@ allenpress.com, for more information. ARTWORK MEDIA THE MISSOURI BOTANICAL GARDEN )02 ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 89 2002 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 2002 ISSN 0026-6493 VOLUME 89 2002 AGUILAR ENRIQUEZ, MARÍA DE LOURDES (see Silvia Romero Rangel. Ezequiel Carlos Rojas Zenteno & Maria de Lourdes Aguilar Enríquez) — — 551 ALBACH, DIRK C. (see Steven J. Wagstaff, Michael J. Bayly, Philip J. Gar- nock-Jones & Dirk C. Albach) |... 38 ALISCIONI, SANDRA S. Contribución a la Filogenia del Género Paspalum (Poaceae: Panicoideae: Paniceae) 504 AL-SHEHBAZ, IHSAN A. (see Marcus Koch & Ihsan A. Al-Shehbaz) |... 88 BAYLY, MICHAEL J. (see Steven J. Wagstaff, Michael J. Bayly, Philip J. Gar- nock-Jones € Dirk C. Albach) —. | 36 BRADFORD, JASON С. Molecular Phylogenetics and АВА ‘al Evolution in Cunonieae (Cunoniaceae) : — 49] CIALDELLA, ANA MARÍA & LILIANA MONICA GIUSSANI. Phylogenetic Relationships of the Genus Piptochaetium (Poaceae, Pooideae, Stipeae): Evidence from Morphological Data TM 305 CLAYTON, JASON R. (see Alan W. Meerow, Charles L. — Qin-Bao Li & Jason R. Clayton) 000000 — 0 CRACRAFT, JOEL. The Seven Great Questions of Systematic Biology: An Essential Foundation for Conservation and the Sustainable Use of Bio- diversity 127 CROSBY, MARSHALL R. Statistical Summary of Some of the Activities in the Missouri Botanical Garden Herbarium, 2001 303 DENHAM, SILVIA S., FERNANDO O. ZULOAGA & OSVALDO MORRONE. Systematic Revision and Phylogeny of Paspalum Subgenus Ceresia (Po- aceae: Panicoideae: Paniceae) — — < — < < =< — 337 DINERSTEIN, ERIC (see David M. Olson & Eric Dinerstein) 199 ERNEBERG, MARIANNE (see Richard N. Mack € Marianne Erneberg) |... 176 EZCURRA. CECILIA. El Género Justicia (Acanthaceae) en Sudamérica Aus- JJJ. 225 FIALKOWSKI, CAROL (see Debra K. Moskovits, Carol Fialkowski, Gregory M. Mueller & Timothy A. Sullivan) 153 FREIRE, SUSANA E., LILIANA KATINAS & GISELA SANCHO. Gochnatia (Asteraceae, Mutisieae) and the Gochnatia Complex: Taxonomic bia cations from Morphology — 524 GARNOCK-JONES, PHILIP J. (see Steven J. Wagstaff, Michael J. Bayly, Philip J. Garnock-Jones & Dirk C. Albach) oa 38 GIUSSANI, LILIANA MÓNICA (see Ana María Cialdella & Liliana Mónica Giussani) — ا‎ 305 GOLDBLATT, PETER & JOHN C. MANNING. Evidence for Moth and But- terfly Pollination in Gladiolus (Iridaceae—Crocoideae) — 110 GOLDBLATT, PETER & JOHN С. MANNING. Plant Diversity of the Cape Region of Southern Africa — 281 GUY, CHARLES L. (see Alan W. Meerow, Charles L. Guy, Qin-Bao Li & Jason R. Clayton) HAUK, WARREN D. A Review of the Genera ee nia and Potamoganos (Bignoniaceae) — JOE, TONY (see Gary Paul Nabhan, Patrick "an & Tony Joe) |... KATINAS, LILIANA (see Susana E. Freire, Liliana Katinas & Gisela Sanc ho) KOCH, MARCUS & IHSAN A. AL-SHEHBAZ. Molecular Data Indicate Com- plex Intra- and Intercontinental Differentiation of American Draba (Bras- sicaceae) — LI, QIN-BAO (see Alan W. Meerow, Charles L. Guy, Qin-Bao Li & Jason R. Clayton) — LOWRY, PORTER P, II (see Jun Wen, Porter P. Lowry ll, кыш. L. Walck & Ki-Oug Yoo) — a MACK, RICHARD N. € MARIANNE ERNEBERG. The United States Natu- ralized Flora: Largely the Product of Deliberate Introductions ... MANNING, JOHN C. (see Peter Goldblatt & John C. Manning) 0 MANNING, JOHN С. (see Peter Goldblatt & John С. Manning) MEEROW, ALAN W., CHARLES L. GUY, QIN-BAO LI «€ JASON R. CLAY- TON. Phylogeny of the Tribe Hymenocallideae NM eae) Based on Morphology and Molecular Characters ........ MORRONE, OSVALDO (see Silvia S. Denham, Fernando O. Zuloaga & & Os- valdo Morrone) MOSKOVITS, DEBRA K., CAROL FIALKOWSKI, GREGORY M. MUELLER & TIMOTHY A. SULLIVAN. Chicago Wilderness: a New Force in Urban Conservation MUELLER, GREGORY M. (see Debra K. Moskovits, Carol Fialkowski, Greg- ory M. Mueller & Timothy A. Sullivan) NABHAN, GARY PAUL, PATRICK PYNES & TONY JOE. Safeguarding Spe- cies, Languages, and Cultures in the Time of Diversity Loss: From the Colorado Plateau to Global Hotspots — OLSON, DAVID M. & ERIC DINERSTEIN. The Global 200: Priority Ecore- gions for Global Conservation |... PIMM, STUART L. The Dodo Went Extinct (and Other Ecological Myths) .. — PORTER, J. MARK (see Victor W. Steinmann & J. Mark Porter) ....... PYNES, PATRICK (see Gary Paul Nabhan, Patrick Pynes & Tony Joe) RAZAFIMANDIMBISON, SYLVAIN G. A Systematic Revision of Breonia (Ru- ylaceae—Naucleeae) RICHARDSON, P. MICK (see George E. Schatz & P. Mick Ric — — ROJAS ZENTENO, EZEQUIEL CARLOS (see Silvia Romero Rangel, Eze аи Carlos Rojas Zenteno & María de Lourdes Aguilar Enriquez) — 524 88 400 414 176 110 281 400 ROMERO RANGEL, SILVIA, EZEQUIEL CARLOS ROJAS ZENTENO & MARIA DE LOURDES AGUILAR ENRIQUEZ. El Género Quercus (Fa- gaceae) en el Estado de México SANCHO, GISELA (see Susana E. Freire, Liliana Katinas & Gisela Sancho) SCHATZ, GEORGE К. Taxonomy and Herbaria in Service of Plant Conser- vation: Lessons from Madagascar’s Endemic Families SCHATZ, GEORGE К. & Р. MICK RICHARDSON. Conservation, the 47th Annual Systematics Symposium of the Missouri Botanical Garden. Intro- duction: Conservation and the Future of Life SCHNEIDER, JULIO V., ULF SWENSON & GEORG ZIZKA. Phylogenetic Reconstruction of the Neotropical Family Quiinaceae (Malpighiales) Based on Morphology with Remarks on the Evolution of an Androdioe- cious Sex Distribution STEINMANN, VICTOR W. & J. MARK PORTER. Phylogenetic Relationships in Euphorbieae (Euphorbiaceae) Based on ITS and ndhF Sequence Data SULLIVAN, TIMOTHY A. (see Debra K. Moskovits, Carol Fialkowski, Gregory M. Mueller & Timothy A. Sullivan) SWENSON, ULF (see Julio V. Schneider, Ulf Swenson & Georg Zizka) ~~ VAN DER WERFF, HENK. A Synopsis of Ocotea (Lauraceae) in Central Amer- ica and Southern Mexico WAGSTAFF, STEVEN J., MICHAEL J. BAYLY, PHILIP J. GARNOCK-JONES & DIRK С. ALBACH. Classification, Origin, and Diversification of the New Zealand Hebes (Scrophulariaceae) WALCK, JEFFREY L. (see Jun Wen, Porter P. Lowry II, Jeffrey L. Walck & Ki-Oug Yoo) WEN. JUN, PORTER P. LOWRY II, JEFFREY L. WALCK & KI-OUG YOO. Phylogenetic and Biogeographic Diversification in Osmorhiza (Apiaceae) YOO, KI-OUG (see Jun Wen, Porter P. Lowry П, Jeffrey L. Walck € Ki-Oug 00 ZIZKA, GEORG (see Julio V. Schneider, Ulf Swenson & Georg Zizka) — ZULOAGA, FERNANDO О. (see Silvia S. Denham, Fernando O. Zuloaga & Osvaldo Morrone) їл yu — 524 145 125 64 (EE L m] — | 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 International Journal of Plant Sciences American Midland Naturalist Journal of Animal Ecology American Naturalist Journal of Applied Ecology Annals of the Missouri Botanical Garden Journal of Biogeography | Annual Review of Ecology and Systematics Journal of Ecology Biotropica Journal of Tropical Ecology Brittonia Journal of The Torrey Botanical Society Conservation Biology | Limnology and Oceanography Diversity and Distributions Missouri Botanical Gaiden Annual Ene | Ecological Applications New Phytologist Ecological Monographs . Paleobiology Ecology Quarterly Review of poer Evolution Systematic Biology Functional Ecology Systematic Botany Global Ecology and Biogeography Letters Information regarding JSTOR is available at http://www.jstor.org 120 Fifth Avenue, New York, NY 10011 — CONTENTS Phylogenetic Relationships in Euphorbieae (Euphorbiaceae) Based on ITS and ndhF Sequence Data Victor W. Steinmann & J. Mark Porter Molecular Phylogenetics and Morphological Evolution in Cunonieae (Cunoniaceae) C. Bradford Contribución a la Filogenia del Género Paspalum (Poaceae: Panicoideae: Paniceae) Sandra S. Aliscioni Gochnatia (Asteraceae, Mutisieae) and the Gochnatia Complex: Taxonomic Impli- cations from Morphology ...... Susana E. Freire, Liliana Katinas & Gisela Sancho El Género Quercus (Fagaceae) en el Estado de México ....... - Silvia Romero Rangel, — Ezequiel Carlos Rojas Zenteno y María de Lourdes Aguilar Enríquez Checklist T Rites Index to Volume 89 _ Cover illustration. — Breonia richardsonii Razafim.. d Es ik Misi