Volume 98, Number April 2011 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, in- dependent reviewers. Instructions t o Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.info. Editorial Committee Roy E. Gereau Victoria C. Hollowell Scientific Editor, Latin Editor, Missouri Botanical Garden Missouri Botanical Garden Ihsan A. Al-Shehbaz Beth Parada Missouri Botanical Garden Managing Editor, Gerrit Davidse Missouri Botanical Garden Missouri Botanical Garden Allison M. Brock Peter Goldblatt Associate Editor, Missouri Botanical Garden Missouri Botanical Garden Gordon McPherson Tammy Charron Missouri Botanical Garden Associate Editor, Charlotte Taylor Missouri Botanical Garden Missouri Botanical Garden Cirri Moran Henk van der Werff Press Coordinator, Missouri Botanical Garden Missouri Botanical Garden For subscription information contact Annals of the Missouri Botanical Garden, % Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2011 is $180 per volume U.S., $190 Canada & Mexico, $215 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www.mbgpress.info The Annals of the Missouri Botanical Garden (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. Postmaster: Send address changes to Annals of the Missouri Botanical Garden, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annuls are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Ab- stract/Global Health databases, ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. The full-text of Annals of the Missouri Botanical Garden is available online though BioOne™ (http:// www. bioone . org) . © Missouri Botanical Garden Press 2011 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and their environment, in order to preserve and enrich life. © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Volume 98 Annals m Number 1 of the 2011 Missouri Botanical Garden MORPHOLOGY-INFERRED PHYLOGENY AND A REVISION OF THE GENUS EMMOTUM (ICACINACEAE)1 Annals of the Missouri Botanical Garden Miers (1852) later described E. affine Miers and E. glabrum Benth. ex Miers and transferred Bentham’s species of Pogopetalum to Emmotum. Engler (1872) was the first to propose an infrageneric classification in Emmotum with two sections based on attributes of the trichomes on petals, stamens, and ovary. Section Longistyla Engl. (= Emmotum ) was defined by Engler based on the following characters: petals adaxially with red trichomes covering the whole surface, filament of the stamen with the base dilated and the apex attenuate, anthers oblong-ovate with a terete connec- tive, and the style longer than the ovary, whereas section Brevistyla Engl, has petals with trichomes at the base and apex, filaments of the stamens with the base and the apex dilated, anthers oblong-linear with the connective fleshy subtetragonous, and the style shorter than the ovary. Engler included E. glabrum in his section Brevistyla based on a plate published by Miers (1851-1861). This plate, which shows a plant with a short style and stamen with an anther as long as the filament, is more correctly referable to E. nitens (Benth.) Miers, and it is this species that is selected as lectotype for section Brevistyla (Duno et al., 2007). Engler further assigned E. orbiculatum (Benth.) Miers, a species with a short style of less than 0.5 mm long, to section Longistyla. Thus, some species assignments were inconsistent with the attributes proposed by Engler to define his sections. Van Tieghem (1897) described the family Emmo- taceae, which he placed close to Icacinaceae, and included the two genera Emmotum and Pogopetalum. Sleumer (1940) followed Engler’s infrageneric divi- sion of Emmotum, but elevated sections to subgenera. The most complete treatment of Emmotum thus far was published by Howard (1942a), in which he proposed four additional species: E. nudum R. A. Howard, E. conjunctum R. A. Howard, E. Jloribun- dum R. A. Howard, and E. ffilvum R. A. Howard. Howard correctly transferred E. glabrum to subgenus Euemmotum Sleumer, which laxly corresponded to Engler’s section Longistyla. Howard therefore con- cluded that the monotypic subgenus Brevistyla was of little use and proposed to abandon the infrageneric classification. He was probably unaware that E. orbiculatum was better referred to section Brevistyla, thus making this bispecific. There are also a few regional treatments of Emmotum (Engler, 1872; Carvalho et al., 1973; de Roon, 1994; Howard & Duno de Stefano, 1999), most of which closely followed Howard’s (1942a) classification scheme. Karehed (2001), using several molecular markers (ndhF, rbcL, atpB, and 18S ribosomal DNA [rDNA]), along with morphological and anatomical evidence, the Euasteridae I, including a large number of the genera traditionally included within that family. His results conclusively demonstrated that Icacinaceae are polyphyletic with genera scattered across several families and orders. The analyses identified two major clades within a narrowed circumscription of Icacinaceae, now included within the Garryales, in the Asteridae clade (Angiosperm Phylogeny Group, 2009). These were the Icacina A. Juss. group, composed mainly of pantropical lianas, and the Emmotum group, composed of mostly Neotropical trees, with the exception of Platea Blume, a genus with five Asiatic species. Here we present the results of a cladistic analysis based on morphological characters that address the following questions. (1) Is the genus Emmotum monophyletic? (2) If so, what are its synapomorphies and other delimiting characters? (3) What is the position of the genus relative to other Neotropical genera of the Emmotum group? (4) What are the phylogenetic relationships among species of Emmo- tum, especially in the context of the infrageneric classification proposed by Engler (1872)? TAXON ANALYSIS As a result of the taxonomic work of this investigation, 13 species of the genus Emmotum are recognized and included as terminal taxa in the ingroup. There are no previous studies exploring the monophyly of Emmotum. As mentioned, Karehed (2001) found the family Icacinaceae not to be monophyletic, and its circumscription was narrowed to include two main groups: the Icacina group and the Emmotum group. We chose members of this first & C. H. Thomps., Ottoschulzia cubensis (C. Wright ex Griseb.) Urb., Poraqueiba paraensis Ducke, and P. sericea Tul. In Karehed’s (2001) study, Poraqueiba Aubl. was sister to Emmotum, thus supporting the inclusion of this genus as a member of the outgroup. The genera Oecopetalum Greenm. & C. H. Thomps. and Ottoschulzia Urb. have been mentioned as morphologically and anatomically related to Emmo- tum and Poraqueiba (Howard, 1942a, 1942b). The more distantly related Neotropical species Calatola costaricensis Standi., a member of the Emmotum clade, was used to root the tree. This taxon is characterized by the autapomorphies of dioecious plants, with 4-merous, unisexual flowers (vs. monoe- cious plants, with 5-merous, bisexual flowers). If i i!j in 4^3^3 s s § § § s § s s,i § s s S22333323333332223 ,1 . __ J{ MMmm’m ii « - « f f s’ n $ s s s s s s s s £ o 5 5 5 5 5 5 5 ^ «- s 5 s 5 H s 5 5 it 3 s I J iis||f I I#if Hi mmim iliyjfei! Annals of the Missouri Botanical Garden Table 2. Data n Ottoschulzia ci E. conjunctum E.fagifoUum E. floribundum E. glabrum E. harleyi E. yapacanum 02 1011x01100002 10? 120? 3003 703300001000020 00000020000111000100071201112303000011000 00 101 100 110 11 12 00 1100 71201 112 3 03 00000 12 00 0111110011012110011007020001000 1000000000 01111100110121100110070200010001000000000 012111x0110121101011102212123212110111111 xl2 11 1011 10 12 110 10 11102202 12 3202011111111 01211121110121101011112102214101010111111 11210020000121101011102202123202011111111 xl2111xlllll21101011102202 123202011111111 012111x0110121101011102202123202011111111 01211121111121101011102212123212110111111 012111x1111121101011102202123202011111111 X1210010000L21101011102202 123202011111111 01211101111121101011112102214101010111111 01211101111121101011112102214101011111111 01211121111121101011112102214101011111111 11210010000121101011102202123202011111111 A morphological data matrix was obtained after careful examination of more than 200 herbarium specimens of Emmotum and the outgroup taxa from the following 25 herbaria: A, AAU, BM, CAS, ECON, F, G, GH, INPA, K, LBP, MA, MER, MEXU, MICH, MO, MY, NY, PMA, PORT, S, SCZ, US, VEN, and XAL. Specimens examined included the types of all described terminal taxa. Forty-one morphological characters were obtained from general vegetative and reproductive morphology. continuous, namely petiole length (character 5), 1am- (character 9), number of flowers (character 15)^ length ratio (character 28), style length (character 32), and gaps in the interspecific variation of the taxa (Table 1). All characters used in the analysis are listed and de- scribed in Appendix 1, and the data matrix is presented in Table 2. All characters were similarly weighted, and (nonadditive, Fitch parsimony [Fitch, 1971]). Terminology follows the Systematic Association Committee for Descriptive Biological Terminology (1962) for leaf shape and apex (characters 7, 23, 26, and 36). The trichome terminology used in this study (characters 3, 4, and 21) has been compared with a 1948). There is a wealth of anatomical and pollen information for the family (see Karehed, 2001, for bibliography), but most of these characters turned out outgroup relationships herein. PHYLOGENETIC ANALYSES The phylogenetic analyses were performed with the program NONA (Goloboff, 1993) in conjunction with a WinClada shell (Nixon, 2002). The parsimony ratchet algorithm (Nixon, 1999) was implemented to optimize the parsimony search with the following settings: number of iterations/repetitions = 200 and number of trees to hold/iteration = 1. Support for the clades recovered from the analyses was estimated with the use of 1000 jackknife iterations (JK) and the following search parameters: 30 search replicates (mult*30), three starting trees per replication (hold/ 3), and maximum number of trees set to 25,000. Additional support for these clades was evaluated with the use of the Bremer index or Bremer decay 20,000 trees as maxiium number held iJmemory (h 20000), the longest of them 10 steps longer than the the bootstrap [BS] 10 command). Resulis Of the 41 characters originally explored, the number of flower parts (character 12), sexuality of the flower (character 14), sepal aestivation (character 16), and Annals of the Missouri Botanical Garden petal connation (character 19) were not phylogeneti- cally informative (12% of the original data). These characters were thus excluded from consecutive analyses. The cladistic analysis resulted in a single most parsimonious tree (MPT) of 77 steps, with a consistency index (Cl) of 0.76 and a retention index (RI) of 0.85. The single MPT obtained (Fig. 1) unambiguously identifies Emmotum as a highly supported (JK = 100%, BS = 8) monophyletic group with nine synapomorphies and Poraqueiba as the sister taxon, with which it shares two synapomorphies. Discussion ROBUSTNESS OF THE PHYLOGENETIC HYPOTHESIS Similar morphological analyses at the species level show a similar number of morphological and/or anatomical informative characters, e.g., Asarum L. with 32 taxa (ingroup) and 37 characters (Kelly, 1997) and Monsonia L. with 25 taxa (ingroup) and 20 characters (Aldasoro et al., 2001). Some other analyses, with more characters, attempt to resolve relationships at the species and generic levels (e.g., Schulman & Hyvonen, 2003, for the genus Adelobo- trys DC., Melastomataceae). In this case, the data matrix included up to 117 morphological characters (91 informative). This high number of characters is due to the inclusion of a large number of outgroup taxa, up to 10 species in seven different genera, thus increasing the number of informative characters. The inclusion in our analysis of some lianoid genera such as Casimirella Hassl., Leretia Veil., or Mappia Jacq. could result in the addition of new morphological, anatomical, and palynological informative characters. However, the inclusion of these characters, although providing better resolution at the base of the tree, would increase neither the resolution nor the support to the taxa and clades in the ingroup, which is our main focus. It is unlikely that many additional, informative macromorphological characters could be discovered, rendering the search of micromorpholog- ical or molecular data necessary to provide further resolution of species relationships. The Cl of the single MPT, 0.76, falls near the higher end of the range normally obtained in similar, morphology-based analyses (Sanderson & Donoghue, 1989). These ranges of the Cl indicate a moderate level of homoplasy. The topology of the consensus tree, and character states with unambiguous optimization supporting relevant clades, are described below. appears as sister to a clade comprising the genera Poraqueiba and Emmotum. Ottoschulzia cubensis shows two autapomorphies: malpighiaceous hairs with unequal arms (character 3) and inflorescences with one to three flowers (character 15). There is also one autapomorphy for the genus that appears as non- informative: the petal joined at the base (character 19). For Ottoschulzia there are also other vegetative and floral character states that represent revertions (e.g., a reduction in the size of the petiole [character 5] and the lamina [character 6], type of apex [character 7], few, nonconspicuous secondary veins [characters 9 and 10], and small fruits [character 37]). On the other hand, the genus Oecopetalum shows one autapomor- phy, fruits with an acute apex (character 39), but this genus also shows several other unique floral charac- ters that were not included in the analysis (e.g., persistent calyx, petals with prominent midrib [unlike those of Poraqueiba ], and the anthers with lateral dehiscence). The inclusion of these characters (see Greenman & Thompson, 1914; Howard, 1942b) would not affect the topology, resolution, or support for the tree because they are autapomorphic. The status of Poraqueiba as sister taxon to Emmotum, as inferred by Karehed (2001), is supported by our results. Both genera have Amazo- nian or Guayanan distributions, while the other outgroups have Antillean and Central American distributions; two synapomorphies support this clade, namely articulated hairs (character 4) and number of flowers (character 15). PHYLOGENY OF EMMOTUM Emmotum is easily diagnosable and strongly supported by nine synapomorphies: red pigmentation of tissues upon pickling (character 2); white flowers (character 17); slightly developed petal texture (character 20); petals with cylindrical hairs (character 21); multilocular ovary (character 34); fruits small, less than 1.5 cm long (character 37), and with the apex shortly mucronate (character 39); slightly developed endocarp (eh. 40), and seed with a relatively large embryo (character 41). Within Emmotum, two clades are identified. These reprised as Sleumer’s (1940) subgenera. The first clade corresponds to Emmotum sect. Brevistyla (JK = 85%, BS = 2) and consists of four species found only south of the equator, principally in Brazil. Emmotum orbiculatum is from the Brazilian states of Amazonas this clade. Emmotum nitens grows in the cerrado of Brazil to Minas Gerais and Bolivia and is the most Volume 98, Number 1 2011 Duno de Stefano & Carnevali Phylogeny and Revision of Emmotum austral taxon of the clade. Two sets of populations formerly referred to E. nitens deserve recognition as separate species — E. amazonicum Duno & Carnevali, described below, is restricted to the state of Amazonas, Brazil, where it is partially sympatric with E. orbiculatum, while E. harleyi Duno is restricted to the Chapada Diamantina in Bahia, Brazil (Duno et ah, 2007). This Brazilian clade is characterized by two synapomorphies: filaments 1.5- 2 mm long (character 24) and the stamen connective slightly developed (character 30). These four taxa feature discontinuous indument with only two clusters of hairs (character 22), one at the base and another at the apex of the adaxial surface, but the character is not applicable to any member of the outgroup. The second clade corresponds to Emmotum sect. Emmotum and includes nine species (JK = 71%, BS = 2). The clade is characterized by three synapo- morphies: the apex of the fdament dilated (character 27), the anther three or four times longer than the filament (character 28), and the style 2.5-4 mm long (character 32). As here circumscribed, this clade corresponds to those species with densely lanate petals, stamens with filaments dilated at the base and attenuate at the apex, and the anthers ovate. This Desv. ex Ham. Within Emmotum sect. Emmotum, a clade made up of E. acuminatum (Benth.) Miers and E. floribundum (JK = 73%, BS = 2) is sister to the rest of the taxa. They have three synapomorphies: a dorsifixed anther (character 25), a diminute ring of tissue around the base of the ovary (character 31), and lateral styles (character 33). Another clade contains two subclades, one containing E. conjunc- tum R. A. Howard and E. fulvum, and another containing the other five species. On the other hand, E. affine is sister to E. conjunctum and E. fulvum. These two taxa are not supported by any synapo- morphy but by one character state reversion, a lamina with caducous pubescence (character 11). Both species occur on sandstone-derived substrates medium to tall forests on tepui slopes between 850 and 1700 m, while E. conjunctum grows in several kinds of savanna associations and forests, between 300 and 1400 m. Emmotum fagifolium is the sister taxon of a clade including three Guayanan species, E. glabrum, E. celiae R. A. Howard, and E. yapacanum R. A. Howard (JK = 97%, BS = 4) with four reversions: short petioles, 0.4-0.8 cm long (character 5), a small lamina, 2-5 cm long (character 6), lamina with only two or three pairs of secondary veins (character 9), and inconspicuous secondary veins (character 10). EVOLUTION OF SELECTED CHARACTERS Petals In the genus Poraqueiba the petals are fleshy and glabrous with two internal furrows separated by a usually well-developed median ridge. Emmotum, on the other hand, has petals slightly fleshy (character 20), densely indumented with cylindrical hairs (character 21), and lacking internal furrows and ridge. Stamens In the Emmotum group the Icacinaeae, the relevant evolutionary trends. In Poraqueiba, the stamens have fleshy, thickened filaments and anthers with wide, elongated connectives. An examination of the evolution of the features associated with the androecium in Emmotum indicates a trend toward an elongation of these parts. In Emmotum sect. Brevi- styla, the stamens are similar to those of Poraqueiba but with thinner connectives, anthers, and filaments. In Emmotum sect. Emmotum, the stamens become thin and longer, and the connective becomes ovate with no prolongation. The analysis shows that the different characters associated with the stamen morphology are involved in the topology of the single MPT. For example, Poraqueiba is supported by two characters (27 and 29) from stamen morphology and Emmotum sect. Emmotum is also defined by a synapomorphy associated with these characters (28). Also, in Emmotum sect. Emmotum, another clade (E. acuminatum and E. floribundum ) is supported by a stamen character (25); in this case, the anthers become dorsifixed. We believe many of these stamen’s characters are strongly correlated, e.g., relative size of the filament and the anther, shape of the anther, and shape and prolongation of the connective. One of the most important characteristics of the genus Emmotum is the multilocular ovary (character 34), which is not found in any other Icacinaceae. Another important character is the length of the style (character 32). There is a trend toward an elongation of the style in the evolution of the Emmotum group. Poraqueiba has an inconspicuous style; in Emmotum sect. Brevistyla, the style is slightly longer, while it is longest in Emmotum sect. Annals of the Missouri Botanical Garden Volume 98, Number 1 2011 Duno de Stefano & Carnevali Phylogeny and Revision of Emmotum IUCN Red List category. Emmotum conjunctum has a wide distribution, including areas that are protected and that are not densely populated. It is therefore listed as Least Concern (LC) according to IUCN Red List criteria (IUCN, 2001). Discussion. Emmotum conjunctum is sister to E. fulvum, but from a phenetic point of view it is also similar to E. fagifolium. Both species, E. conjunctum and E. fagifolium, have been treated in divergent ways by different authors. Howard (1942b) hypothesized a relationship of E. conjunctum with E. fulvum and E. affine. On the other hand, Steyermark (1966) consid- ered it conspecific with E. fagifolium, but de Roon (1994) kept both species as separate. The type collections of both species show remarkable differenc- mi 1: Volume 98, Number 1 2011 Duno de Stefano & Carnevali Phylogeny and Revision of Emmotum Distribution and ecology. Emmotum yapacanum is endemic to Venezuela. It has only been collected in Amazonas State. It grows on shrublands on sandstone-derived substrates between 800 and 1300 m elevation. IUCN Red List category. Emmotum yapaca- num is only known from three collections. It could be best listed as Data Deficient (DD). However, because it is restricted to isolated sandstone table inaccessible area, it could alternatively be listed at Least Concern (LC) according to IUCN Red List criteria (IUCN, 2001). Annals of the Missouri Botanical Garden STUDIES IN THE CLEOMACEAE I. Hugh H litis* Jocelyn C. Hall ,3 ON THE SEPARATE Theodore S. Cochrane,2 and Kenneth J. RECOGNITION OF CAPPARACEAE, CLEOMACEAE, AND BRASSICACEAE1 (Fig. 1; e.g., litis, 1957; Rodman et al„ 1993, 1994, 1996; Rollins, 1993; Hall et al., 2002). Until 1 Our special thanks are du< s-ss; Engler’s Syllabus der PSlanze, 30 Annals of the Missouri Botanical Garden Crateva tapia L. (Nicaragua, Chontales, Nee 28459 [WIS]). — A. Seeds, exterior views, surrounded by irregularly shaped, pulpy, very short, conical radicle sharply differentiated from the cotyledons by deep constrictions; the strongly asymmetrical connections of Volume 98, Number 1 2011 litis et al. Studies in Cleomaceae BRASSICACEAE Volume 98, Number 2011 litis et al. Studies in Cleomaceae 33 Annals of the Missouri Botanical Garden REVISION OF THE Thomas G. hammers 1 INFRAGENERIC CLASSIFICATION OF LOBEUA L. (CAMPANULACEAE: LOBELIOIDEAE) Volume 98, Number 1 2011 Lammers Revision of Lobelia Schonl. and Lobelia sect. Isolobus (A. DC.) C. B. Clarke, and the merger of Lobelia sect. Tylomium and Lobelia sect. Rhynchopetalum under the former Like Presl (1836), Post and Kuntze (1903) insisted that Rapuntium was the correct name for the genus. Their classification differed from that of Schonland (1889) in the segregation of Rapuntium sect. Trematocarpus Kuntze from Rapuntium sect. Tupa (G. Don) Kuntze, and Rapuntium sect. Haynaldia Kuntze from Rapuntium sect. Tylomium (C. Presl) Kuntze; inclusion of the genera Grammatotheca C. Presl, Monopsis, and Parastranthus as sections; and the renaming of Lobelia sect. Eulobelia and Lobelia sect. Hemipogon as Rapuntium sect. Cardinalis Kuntze and Rapuntium sect. Dortmanna (0. 0. Rudbeck ex Hill) Kuntze, respectively. The most highly structured classification of Lobelia was that of Wimmer (1943, 1953, 1968), which was derived from Schonland’s (1889) and was the first since de Candolle (1839) to assign every recognized species to an infrageneric taxon. A major innovation was the apportionment of the sections among three subgenera: Lobelia subg. Lagotis E. Wimm. (later corrected to Lobelia subg. Lobelia ), Lobelia subg. Mezleria (C. Presl) E. Wimm., and Lobelia subg. Tupa (G. Don) E. Wimm. Lobelia subg. Lobelia comprised two sections, Lobelia sect. Hemipogon (later corrected to Lobelia sect. Lobelia ) and Lobelia sect. Holopogon. The former was divided into Lobelia subsect. Trachyspermae E. Wimm. (which should have been corrected to Lobelia subsect. Lobelia) and Lobelia subsect. Leiospermae E. Wimm., the latter into Lobelia subsect. Cryptostemon E. Wimm. and Lobelia subsect. Delostemon E. Wimm. Lobelia subg. Mezleria likewise comprised two sections, Lobelia sect. Eumezleria E. Wimm. (later corrected to Lobelia sect. Mezleria ) and Lobelia sect. Para- mezleria E. Wimm. Lobelia subg. Tupa comprised six sections: Lobelia sect. Isolobus , Lobelia sect. Eutupa A. DC. ex E. Wimm. (later corrected to Lobelia sect. Tupa), Lobelia sect. Rhynchopetalum, Lobelia sect. Homochilus, Lobelia sect. Revolutella E. Wimm., and Lobelia sect. Galeatella E. Wimm.; of these, Lobelia sect. Tupa was divided into Lobelia subsect. Primanae E. Wimm. and Lobelia subsect. Haynaldianae E. Wimm. Some of the subsections were divided into greges and subgreges. Wimmer also used the symbol “§” to denote “kleinere Gruppen (ohne Rang)” (1953: 479), in one case within a grex, in others within a subsection. Many of the taxa in Wimmer’s (1943, 1953, 1968) classification were defined by single charac- ters. For example, Lobelia subg. Lobelia and Lobelia subg. Mezleria were distinguished from Lobelia subg. Tupa solely on habit: “saepe annuae, debilies vel suffrutices vel frutices” in the latter (with the proviso that Lobelia subg. Tupa sect. Isolobus actually resembled the other two subgenera in habit). Similarly, Lobelia sect. Holopogon and Lobelia sect. Lobelia were distinguished solely on the basis of whether all five or only the ventral pair of anthers were bearded at the apex. One’s confidence in the naturalness of a classification that embodies this sort of single-feature character- ization is frankly quite low. In this regard, the classification of Murata (1995) represented a marked improvement because of his attempt to find multiple correlations among function- ally unrelated characters (cf. Stuessy, 2009). Specific changes were: (1) narrowing the circumscription of Lobelia sect. Lobelia so that it encompassed only Lobelia subsect. Lobelia; (2) removal of Lobelia subsect. Leiospermae from Lobelia sect. Lobelia and recognition of most of it as Lobelia sect. Heyneana J. Murata; (3) treatment of the remainder of Lobelia subsect. Leiospermae as Lobelia sect. Dioicae (E. Wimm.) J. Murata in Lobelia subg. Mezleria; (4) division of Lobelia sect. Holopogon into Lobelia sect. Cryptostemon (E. Wimm.) J. Murata and Lobelia sect. Delostemon (E. Wimm.) J. Murata; (5) addition of Pratia sect. Pratia to Lobelia sect. Mezleria (which thus became Lobelia sect. Pratia (Gaudich.) J. Murata); (6) placement of Lobelia sect. Isolobus in Lobelia subg. Mezleria instead of Lobelia subg. Tupa; (7) addition of Pratia sect. Colensoa (Hook, f.) Baill. to Lobelia subg. Tupa as Lobelia sect. Colensoa (Hook, f.) J. Murata; and (8) removal of all non- Chilean species of Lobelia sect. Tupa to Lobelia sect. Colensoa. No ranks lower than section were used. comings. Only types and a few other exemplars were assigned to each section; one is left to guess where the remainder belong. (This is complicated by the fact that a few exemplars were said to belong to one section in the text and a different one in the accompanying table.) Additionally, some of the names used violate provisions of the ICBN. For example, at sectional rank Hemipogon has priority over Pratia, while Stenotium and Microcentron have priority over Heyneana. Characters and Their States 40 Annals of the Missouri Botanical Garden section, some of the more complex characters that have proven particularly efficacious in crafting the revised classification of Lobelia are explained and discussed, below are understood fully. is extremely heterogeneous in overall body e so than any other genus in the family, various species form a continuum from ; plants that undergo no secondary growth or lignification, to those that become lignified in their basal portions, to those that amass considerable wood throughout their axes, building Second, relative stoutness of the stems varies considerably, irrespective of the presence or absence of lignification. In many species, the stems are no more than 5 mm in diameter, even when some termed gracile. Stems of greater girth, which often are woody to some degree, are termed robust. Very robust stems with a broad parenchymatous or hollow pith are termed pachycaul (the gracile and less robust stems are contrasted as leptocaul). Third, the longevity of individual plants varies. Most species, whether herbaceous or woody, are iteroparous (polycarpic). Among species that are are typically annual or in a few cases biennial, while those with robust and especially pachycaul stems are plies Lesial. Fourth, almost every life form (Raunkiaer, 1934) is represented. Hemicryptophytes are perhaps the commonest, followed by geophytes and chamae- phytes; these all may be considered “herbaceous perennials.” The remaining herbaceous species are therophytes. Woody plants include both nanophaner- ophytes (subshrubs, shrubs, and treelets) and phanerophytes (large shrubs and trees). Most species are terrestrial, but a few are facultatively or obligately hydrophytic. INFLORESCENCE The basic reproductive unit in Lobelia is a solitary pedicellate flower in the axil of a leaf, typically a leaf near the apex of the stem. The pedicel frequently bears a pair of opposite or subopposite bracteoles, typically in the proximal two thirds; in many species, however, these are altogether lacking. When multiple flowers are produced (by far the norm), the sequence of bloom is always acropetal (centripetal). The stem is characteristically anaux- otelic, i.e., incapable of resuming vegetative growth after flowering ceases. If the leaves that subtend the flowers differ in neither size nor shape from ordinary foliage, the plants are said to bear solitary axillary flowers. More commonly, these distal floral leaves are i.e., they are bracts. As a result, the flowers are borne in a terminal raceme. Sometimes, the transition from foliage to bracts is very gradual; such “foliose racemes” may be mistaken for a stem with solitary axillary flowers, particularly when anthesis begins. Other times, the change from foliage to bracts is quite abrupt and a distinct peduncle may separate the distalmost leaf from the basal bract. In such “bracteate racemes,” the demarcation of the inflorescence is quite obvious. Variation in pedicel length can cause a raceme to resemble a spike (uniformly short pedicels), a corymb (pedicel length decreasing acropetally), or an umbel (rachis condensed). In some species, branches arise two species, the bracteate racemes become pseu- doaxillary via sympodial growth, i.e., though termi- nal, they are soon overtopped by vegetative growth from an adjacent axillary bud. As with habit, the appearance of the corolla varies enormously, first, the five lobes may all be relatively uniform in size, shape, and connation, i.e., mono- morphic. Alternatively, the three ventral lobes may be larger, broader, and connate for a greater distance than the dorsal pair, i.e., the lobes are dimorphic. In one small group of the latter, however, the central ventral lobe is conspicuously larger and wider than the two flanking it. Second, angular orientation of the lobes around the flower’s axis varies; this is best understood by mentally superimposing a clockface on the flower as seen in face view. In species with dimorphic lobes, the corolla is bilabiate: the two dorsal lobes more or less parallel one another in an upright posture, while the ventral three spread out as a trifid lip. In other words, the dorsal lobes are oriented at 12 o’clock or at 11 and 1 o’clock, while the ventral lobes fall at 4, 6, and 8 o’clock. In species with monomorphic lobes, greater variation is possible. Some species have a bilabiate corolla as described above, though the distinction between the dorsal and ventral lip may not be as pronounced as it is when the lobes are dimorphic. Far more common is for the dorsal lobes to arise in a slightly more ventral 41 42 Annals of the Missouri Botanical Garden develop a partial or complete wing around the circumference. CHROMOSOME NUMBER Published chromosome numbers for Lobelia were additions have been made by several authors (Knox & Kowal, 1993; Stace & James, 1996; Ruas et al., 2001; Murray et al., 2004) and a few overlooked counts discovered (Lepper, 1979, 1980, 1982). The base chromosome number (*) for Lobelia is readily interpreted as seven, making the numerous plants with 2 n = 14 diploid. From this base, several levels of polyploidy have evolved, including tetraploid (2 n = 28), hexaploid (2 n = 42), octoploid (2 n = 56), decaploid (2 n = 70), endecaploid (2 n = 77), dodecaploid (2 n = 84), tridecaploid (2 n = 91), and icosaploid (2 n = 140). Additional numbers (2 n = 12, 16, 18, 20, 22, 24, 26, 38) are interpreted as aneuploid, although allopolyploidy cannot be ruled out. Taxonomic Treatment Only those names validated by citing Lobelia as a per se. All other effectively published supraspecific names referable to the genus will be found under the section to which its type belongs. This includes the following names at generic rank: Calcaratolobe- lia Wilbur, Colensoa Hook, f., Dortmanna, Enchy- sia, Euhaynaldia Borbas, Galeatella (E. Wimm.) 0. Deg. & I. Deg., Haynaldia Kanitz, Holostigma G. Don, Holostigmateia Rchb., Hypsela C. Presl, Isolobus, Neowimmeria 0. Deg. & I. Deg., Petro- marula Belli ex Nieuwl. & Lunell, Piddingtonia, Pratia, Rapuntium, Rhynchopetalum Fresen., Speir- ema Hook. f. & Thomson, Trimeris, Tupa, and Tylomium C. Presl. An index to sectional assign- ments for all species included wiLhin Lobelia by Lammers (2007a) is provided in Appendix 1. Lobelia L., Sp. PL 2: 929. 1753. Cardinalis Riv. ex Fabr., Enum. 122. 1759, nom. illeg. sub Art. 52.1. Laurentia Adans., Fam. PL 2: 134, 568. 1763, nom. illeg. sub Art. 52.1 (Brummitt, 2000; but cf. Lammers, 1997). Mecoschistum Dulac, FI. Hautes-Pyrenees 459. 1867, nom. illeg. sub Art. 52.1. TYPE [designated by Hitchcock & M. L. Green in Anon., Nomencl. Prop. Brit. Bot. 184. 1929]: Lobelia cardinalis L. [Under Art. 10.5(b), this supercedes the earlier choice of Lobelia dortmanna L. (111. FI. N. U.S. (Britton & Brown) 3: 29. 1913) (cf. McNeill et al., 1987).] Plants perennial (hemicryptophytes, geophytes, chamaephytes, nanophanerophytes, or phanero- phytes) or annual (rarely biennial), sometimes pliestesial, 0.02-9 m tall, terrestrial or rarely hydrophytic or epiphytic. Roots fibrous or rarely tuberous, adventitious or the primary root persisting as a tap root. Stems gracile or robust, the latter sometimes pachycaul, herbaceous, suffruticose, or woody, prostrate, decumbent, ascending, or erect, simple to much branched, sometimes rhizomatous, stoloniferous, and/or apically or basally rosulate; latex acrid, viscous, white or rarely colored. Leaves alternate, simple, exstipulate, dillenid, dorsiventral (rarely centric), pinnately (rarely palmately) veined, sessile or petiolate; margin commonly variously toothed, less often entire, rarely lobed or parted. Flowers tetracyclic, perfect (rarely imperfect and the plants then dioecious or gynodioecious), zoophilous, chasmogamous with a specialized method of protan- drous secondary pollen presentation, resupinate, epigynous, zygomorphic, pedicellate, solitary in the axils of the upper leaves or these reduced in size, creating a terminal (very rarely pseudoaxillary) foliose or bracteate anauxotelic (very rarely aux- otelic) raceme, sometimes branching (paniculate), rarely spikelike, corymbose, or subumbellate in appearance; pedicels bibracteolate or ebracteolate, very rarely oligobracteolate. Calyx synsepalous, radially (rarely bilaterally) symmetric, adnate to lobes 5 (very rarely 4), valvate, typically triangular, rarely with a reflexed auriculate appendage in each sinus. Corolla early-sympetalous, typically for half or more of its length, bilaterally (rarely almost radially) symmetric, bilabiate or sub-bilabiate with 2 dorsal and 3 ventral lobes or unilabiate with 5 ventral lobes, typically some shade of blue or purple, often with undertones of rose or mauve, less often red, pink, magenta, orange, yellow, green, or white, concolorous or subtly (less often sharply) bicolorous or tncolorous, very rarely with a slender long or short ventral nectar spur; tube straight to arcuate, typically cleft almost to its base on the dorsal side, laterally fenestrate; lobes valvate, monomorphic, or dimorphic with the ventral 3 larger, sometimes with a palate or a pair of gibbosities on the ventral lip at the mouth of the tube. Stamens 5, antisepalous, inserted on rim of hypanthium or at very base of corolla tube, connate distally, forming an exserted (rarely included) and dorsally deflected staminal column free from (rarely adnate to) the corolla tube; anthers tetrasporangiate, dithecal, basifixed, in- trorsely dehiscent by longitudinal slits, the dorsal Volume 98, Number 1 2011 Lammers Revision of Lobelia 3 longer than the ventral 2, forming a ventrally oblique tube occluded at the orifice, the ventral pair or all 5 anthers bearded at apex with dense tufts of filiform hairs, the ventral pair sometimes with a single apical bristle each, or rarely all anthers nude at apex; pollen grains tricolporate or tricolpate, prolate, ellipsoid, psilate, binucleate or trinucleate when shed. Gynoecium syncarpous, bilocular; ovary completely or partially inferior (rarely almost superior through reduction of the hypanthium); placentae axile, large; ovules numerous, small, anatropous, unitegmic, tenuinucellate; style solitary, the apex; stigma bilobed, the lobes appressed and nonreceptive as the style grows through the anther tube, pushing out pollen, after which the stigmas spread and become receptive. Fruit a capsule, loculicidally dehiscent by an apical pair of triangular valves, or less often a fleshy or dry berry. Seeds small, numerous, black, brown, or golden, terete, trigonous, or lenticular (the last sometimes winged), rarely irregularly angular, cuboidal, or quadrate; testa reticulate or striate; embryo small, straight; endosperm copious, cellular, oily or rarely starchy, its formation ab initio cellular. Chromosome numbers. 2 n = 12, 14, 16, 18, 20, 22, 24, 26, 28, 38, 42, 56, 70, 77, 84, 91, 140 (Lepper, 1979, 1980, 1982; Knox & Kowal, 1993; Lammers, 1993; Stace & James, 1996; Ruas et al., 2001; Murray et al., 2004). Distribution. Almost cosmopolitan, native to (1) the New World, from southern Canada to Tierra del Fuego, including the Greater and Lesser Antilles, Turks and Caicos Islands, the Bahamas, Galapagos Islands, Juan Fernandez Island, and the Falkland Islands; (2) northern and western Europe, from Scandinavia and northern Russia to the Iberian Peninsula, including the Faroe Islands and the British Isles; (3) Africa, in Morocco and from Senegal to Sudan and South Africa, including the Azores, Madeira, St. Helena, Madagascar, and the Mascarene Islands; (4) southern and eastern Asia, from Oman to the Russian Far East south to Indonesia and Papua New Guinea, including the Kuril Islands, Japan, the Bonin Islands, the Ryukyu Islands, Taiwan, Hainan, the Philippines, the Andaman and Nicobar Islands, and Sri Lanka; (5) Australia and New Zealand, including Norfolk, Kermadec, and Chatham islands; and (6) the central Pacific, specifically Pitcairn, Rapa, and the Hawaiian Islands. The New World and Africa are each home to over 37% of the species; 12% of the species are found in Asia, 10% in Australia and New Zealand, and 3% in the Hawaiian Islands (Lammers, 2007a). The two species native to Europe are interlopers from boreal North America (Lobelia dortmanna ) and northwestern Africa ( L . urens L.), while the single species native to Pitcairn from southern Africa through Madagascar, Australia, Norfolk Island, New Zealand, Chatham Island, and Kermadec Island to Juan Fernandez Island and Chile. Etymology. The name honors Flemish herbalist Matthias de L’Obel (1538-1616), physician to William of Orange and James I of England, who in his Stirpium Adversaria Nova (1571) argued for the necessity of exacting observation in botany and medicine. Discussion. The circumscription of Lobelia em- braced here is that used in the recent accounts of the family in the World Checklist and Bibliography series (Lammers, 2007a) and the encyclopedic Families and Genera of Vascular Plants (Lammers, 2007c). This is nearly identical to that advocated by Schonland (1889) and Wimmer (1943, 1953, 1968); the sole difference is the inclusion of species previously assigned to Pratia (Moeliono, 1960; Adams, 1972; Wilbur, 1991; Murata, 1995; Lam- mers, 1998) and Hypsela (Chiapella, 1996; Lam- mers, 1999a). A key to distinguish the genus from the rest of the family was provided by Lammers (2007c). cessors (Wimmer, 1943, 1953, 1968; Murata, 1995) in the abandonment of subgenera; the sole rank utilized here is that of section. To recognize two or three subgenera would suggest some fundamental bifurcation or trifurcation very early in the evolution- ary history of the genus. The molecular phylogenies published thus far do not reveal any such event, and it does not seem likely that such a pattern will emerge with further sampling. The overall structure revealed by the most comprehensive of these (Knox et al., 2006; Antonelli, 2008) is of numerous reasonably well-defined clades with varying degrees of related- ness to one another. My opinion is that the best way to translate this sort of evolutionary structure into a traditional formal classification is to recognize multiple coordinate sections. In crafting them, I have placed primary emphasis on multiple correlations of phenotypic characters and on biogeographical coher- ence. To the extent that phylogenetic information is available, I have used it to inform dec n 45 46 Annals of the Missouri Botanical Garden Impares (see below). In recent molecular studies (Knox et al., 2006; Antonelli, 2008), the species of this section plus the next two fell into two clades at the base of the phylogeny. The former also indicated to classification, Grammatotheca and Monopsis might be assigned to this section. If so. the correct name would be a new combination based on Rapuntium sect. Xanthomeria. If Monopsis were retained as a distinct genus, the correct name would be a new combination based on Rapuntium sect. Grammato- theca (C. Presl) Kuntze. 2. Lobelia sect. Holopogon Benth., FI. Austral. 4: 122. 1869. Rapuntium sect. Holopogon (Benth.) Kuntze, Lex. Gen. Phan. 478. 1903. TYPE [designated by J. Murata, J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 15: 357, 1995]: Lobelia gibbosa Labill. 56? 359? 1948. TYPE ’[here designated]: Lobelia Plants annual or rarely perennial (hemicrypto- phytes), (2-)10-40(— 60) cm tall. Stems gracile, herbaceous, decumbent, ascending, or erect, simple to branched. Leaves sessile or petiolate. Flowers solitary in the axils of the upper leaves or these reduced in size, creating a terminal sometimes secund raceme (rarely a panicle); pedicels bibracteo- late near middle. Hypanthium sometimes dorsally distended. Corolla bilabiate, various shades of blue or purple, sometimes marked with yellow, 8-30 mm; tube straight or somewhat curved, cleft almost to its strongly dimorphic, almost as long as the tube to longer than the tube, spreading, the odd lobe larger than the lateral pair or dorsal pair. Staminal column exserted; anthers bearded with tufts of filiform hairs at apex of all 5 (very rarely all nude). Fruit a capsule. Seeds ovoid or ellipsoid, trigonous or lenticular; testa striate (Murata type D). Distribution. Endemic to Australia; 14 species. Included species: Lobelia andrewsii Lammers, L. dentata Cav., L. gibbosa Labill., L. gouldii W. Fitzg., L. heterophylla Labill., L. leichhardii E. Wimm., L. psilostoma E. Wimm., L. rarifolia E. Wimm., L. rhombifolia de Vriese, L. rhytidosperma Benth., L. simplicicaulis R. Br., L. tenuior R. Br., L. trigono- caulis F. Muell., L. winifrediae Diels. Discussion. The circumscription used here is essentially the same as that adopted by Bentham (1876), Schonland (1889), and Post and Kuntze (1903); it is supported by a recent phylogenetic analysis (Knox et al., 2006). Wimmer’s (1953, 1968) circumscription of Lobelia sect. Holopogon, however, was much broader, encompassing diverse extra- Australian species; the section as treated here is equivalent to his Lobelia grex Impares only. Murata (1995) did not recognize a taxon equivalent to this section, but merely included its species within his Lobelia sect. Pratia. Molecular phylogenetic studies (E. Knox, pers. comm.) indicate that the type of Isotoma (R. Br.) Lindl., Lobelia hypocrateriformis R. Br., logically might be referred to this section (though the sect. Hypsela (C. Presl) Lammers [Heenan et al 2008; Knox et al., 2008a]). If that species were to b affected. 3. Lobelia sect. Colensoa (Hook, f.) J. Murata, J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 15: 358. 1995. Colensoa Hook, f., Bot. Antarct. Voy. II. (FI. Nov.-Zel.) 1: 156. 1852. Pratia sect. Colensoa (Hook, f.) Baill., Hist. PL 8: 366. 1885. TYPE [sub Art. 37.3]: Colensoa phys- aloides (A. Cunn.) Hook. f. [= Lobelia phys- aloides A. Cunn.]. Plants perennial (chamaephytes), 0.3-1 m tall. Stems robust, suffruticose toward base, herbaceous Leaves long petiolate, prominently serrate. Inflores- cence a nodding 5- to 20-flowered pedunculate pseudoaxillary corymbose raceme; pedicels oligo- bracteolate. Corolla bilabiate, purple or blue, some- times very pale, 30-45 mm; tube suberect, dorsally cleft to base; lobes monomorphic, almost as long as pair nearly distinct, the ventral 3 forming a trifid lip. Staminal column exserted; anthers nude at apex. Fruit a dark blue, globose, fleshy or leathery berry. Seeds ovoid, terete; testa striate (Murata type D). Chromosome number. 2 n = 26 (Lammers, 1993). Distribution. Endemic to New Zealand’s North Island; monotypic. Included species: Lobelia phys- aloides A. Cunn. Discussion. Wimmer (1943, 1953, 1968) as- signed this section to Pratia and included not only Lobelia physaloides but also species referable to 1979, Annals of the Missouri Botanical Garden Practical references. The members of the Mexi- can and Central American subgroup may be identified using the key provided by Wilbur (1991). 11. Lobelia sect. Homochilus A. DC., Prodr. (DC.) 7: 383. 1839. Rapuntium sect. Homochilus (A. DC.) Kuntze, Lex. Gen. Phan. 478. 1903. TYPE [designated by J. Murata, J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 15: 369. 1995]: Lobelia laxiflora Kunth. Plants perennial (hemicryptophytes or chamae- phytes) or shrubs, (0.2-)0.5-3 m tall. Stems robust, herbaceous, suffruticose, or woody, ascending or erect, simple to branched. Leaves sessile or less often petiolate. Flowers solitary in the axils of the upper leaves or these reduced in size, creating a terminal raceme; pedicels ebracteolate or bibracteolate. Co- rolla bilabiate or sub-bilabiate (rarely unilabiate), magenta, purple, red, pink, orange, or yellow, concolorous or sharply bicolorous, 22-48(— 56) mm; tube straight, laterally fenestrate or entire; lobes much shorter than the tube, spreading (rarely deflexed). Anthers bearded with tufts of filiform hairs at apex on the ventral pair. Fruit a capsule. Seeds ellipsoid or oblong, terete; testa striate (Murata type D). Chromosome number. 2 n = 14 (Lammers, 1993). Distribution. Endemic to the New World, with five species from southern Arizona to southwestern Colombia, and Lobelia decurrens endemic to Peru. Included species: L. aguana E. Wimm., L. decurrens Cav., L. ghiesbreghtii Decne., L. guerrerensis Eakes & Lammers, L. heteroclita McVaugh, L. laxiflora Discussion. The circumscription used here is identical to that of Wimmer (1953, 1968), Murata (1995), and Lammers (2004), and is supported by the phylogeny of Antonelli (2008). Practical references. A key to these species and complete descriptions are provided by Lammers (2004). 12. Lobelia sect. Tupa (G. Don) Benth., Gen. PI. [Bentham & Hooker f.] 2: 552. 1876. Tupa G. Don, Gen. Hist. 3: 700. 1834. Lobelia [un- ranked] Tupa (G. Don) Heynh., Nomencl. Bot. Hort. 1: 473. 1840. Rapuntium sect. Tupa (G. Don) Kuntze, Lex. Gen. Phan. 478. 1903. Lobelia subg. Tupa (G. Don) E. Wimm., Ann. Naturhist. Mus. Wien 56: 364. 1948. TYPE [sub Art. 22.6]: Lobelia tupa L. Tupa sect. Eutupa A. DC., Prodr. (DC.) 7: 391. 1839, nom. invalid, sub Art. 21.3. Lobelia sect. Eutupa A. DC. ex E. Wimm., Ann. Naturhist. Mus. Wien 56: 365. 1948, nom. illeg. sub Art. 52.1. Lobelia subsect. Primanae E. Wimm., Ann. Naturhist. Mus. Wien 56: 365. 1948. TYPE [sub Art. 22.6]: Lobelia tupa L. Plants perennial (hemicryptophytes or chamae- phytes) or shrubs (rarely trees), 0.5-4(— 6.8) m tall. Stems robust, herbaceous, suffruticose, or woody, ascending or erect, simple to much branched; latex white, pale yellow, or tan. Leaves sessile. Flowers solitary in the axils of the upper leaves or these reduced in size, creating a terminal raceme; pedicels ebracteolate or bibracteolate. Corolla unilabiate, red, pink, or purple, rarely yellow, concolorous or subtly bicolorous, 15-65 mm; tube arcuate; lobes much shorter than the tube, deflexed, coherent at apex. Anthers bearded with tufts of filiform hairs at apex of ventral pair. Fruit a capsule. Seeds ovoid, ellipsoid, or oblong, terete; testa striate (Murata type D). Chromosome number. 2 n = 42 (Lammers, 1993). Distribution. Endemic to central Chile; four species. Included species: Lobelia bridgesii Hook. & Am., L. excelsa Bonpl., L. poljphylla Hook. & Am., L. tupa L. Discussion. The narrow circumscription em- braced here is identical to that of Murata (1995) and Lammers (2000), which finds strong support in the molecular phylogenies of Knox et al. (1993, 2008b), Knox and Palmer (1998), Koopman and Ayers (2005), and Antonelli (2008). Wimmer’s (1953, 1968) circumscription of the section was much broader, encompassing the species here treated as Lobelia sect. Speirema, Lobelia sect. Trimeris, and Lobelia sect. Tylomium, as well as some of those in Lobelia sect. Rhynchopetalum. The circumscription used here is equivalent to the “§ species chilenses” within his Lobelia subsect. Primanae. Practical references. A key to these species and complete descriptions are provided by Lammers (2000). 13. Lobelia sect. Trimeris (C. Presl) A. DC., Prodr. (DC.) 7: 357. 1839. Trimeris C. Presl, Prodr. Monogr. Lobel. 46. 1836. Rapuntium sect. Trimeris (C. Presl) Kuntze, Lex. Gen. Phan. 478. 1903. TYPE [sub Art. 37.3]: Trimeris C. Presl, 61 X iXJu ax NEWLY SEQUENCED NUCLEAR Cynthia M. Morton 2 GENE (XDH) FOR INFERRING ANGIOSPERM PHYLOGENY1 Gmmemles, Pro- 64 Annals of the Missouri Botanical Garden is crucial that phylogenetics produce a well-resolved Studies of broad angiosperm phylogenetics have centered on using chloroplast genes individually and in combination (Chase et al., 1993; Soltis et al., 1997, 2000; Savolainen et al., 2000; Hilu et al., 2003; Jansen et al., 2007). More recently, mitochondrial genes have been incorporated into basal angiosperm phylogenies ( rpoC2 , Qiu et al., 2006). Nuclear genes have hardly been used, and only the 18S gene has been extensively incorporated into multigene analy- ses investigating angiosperm phylogeny (Soltis et al., 1997). On its own, 18S has provided very limited resolution and low internal support for many major clades (Soltis et al., 1997). Studies of rbcL and atpB chloroplast genes have provided phylogenies with more resolution than individual genes; however, topological incongruence does occur between the genes (Savolainen et al., 2000). Analysis of the chloroplast gene matK provided a tree with greater resolution and more support for major clades, but topological incongruences were observed among matK, rbcL, and atpB. Combining chloroplast genes in multigene data sets for analysis has improved both resolution and internal support (Savolainen et al., 2000; Burleigh et al., 2009). Hilu et al. (2003) within the angiosperm trees and found the following supported percentages for these genes: rbcL (24%), atpB (14%), 18S (7%), rbcL/atpB (55%), and rbcL/ atpB/ 18S (83%). When using chloroplast genes, one must be a bit cautious because these genes are maternally inherited and, due to issues such as hybridization, the resulting phylogenetic reconstruc- tions may not reflect the “true phylogeny” or correct topology. The first angiosperm consensus tree based on combined chloroplast ( rbcL and atpB ) and nuclear (18S) genes was conducted by Soltis et al. (2000). Although 18S is a very slowly evolving gene, the combined chloroplast and nuclear consensus tree angiosperm phylogeny because it is based on both maternally and biparentally inherited genes. Howev- er, many of the interrelationships in the angiosperm and land plant phylogenies are still controversial, and it is imperative to use information from different biparentally inherited genes to clarify and enhance the results derived largely from chloroplast genes. Xanthine dehydrogenase is a member of the molybdenum hydroxylase famdy of enzymes, which are thought to play important metabolic roles in purine metabolism and hormone biosynthesis. Mo- lybdenum (Mo) is essential for (nearly) all organisms and functions in more than 40 enzymes by catalyzing diverse redox reactions. Only four of those enzymes have been found in plants: nitrate reductase; aldehyde oxidase(s); xanthine dehydrogenase, which is involved in purine catabolism and stress reactions; and sulphite oxidase (Mendel & Hansch, 2002). Among Mo enzymes, the alignment of amino acid sequences permits the definition of domains that are well conserved, although they are variable. This study amplified approximately 1265 bp of the xanthine dehydrogenase gene (X dh), which is located on chromosome 4, contig fragment no. 82, and is approximately 4083 bp long (information obtained by comparing published Arabidopsis and this study’s sequences). The sequenced segment lies from 358 to 1623 bp including intron 1-3 and exon 2-4. The Arabidopsis (DC.) Heynh. sequence contains 13 introns and 14 exons. Hesberg et al. (2004) found that Arabidopsis has a 46%^47% identity to human and bovine X dh and therefore is homologous to X dh from other organisms. This examination segment contains approximately 11 indel regions. To date, only one copy of the gene has been reported, except in silkworms and Arabidopsis, in which a second copy has been found. In Arabidopsis, Hesberg et al. (2004) reported that the physicochemical properties of AtXDHl and AtXDH2 might be nearly identical because of their high degree of sequence similarity (94%). Markers from the nuclear genome that can be used on a broad taxonomic scale without problems of paralogy are limited in number. The possibility of paralogy exists for all genes, not just nuclear ones, because gain and loss of genes by various processes are continuous and all genes have some probability of paralogy is that gene trees derived from samples of single genes from each species may not match the compare the angiosperm phylogeny is a good indicator of having paralogous genes. We have used this test of congruence among genes to evaluate duplications or losses within Xdh. This analysis provides the first angiosperm plant gene tree using Xdh. More recently other researchers have been exploring the utility of Xdh for “gymno- sperm” phylogenetics and have found it to be similar to inferences using other genes and to be highly informative (Peery et al., 2008). We compared the topology obtained from this gene and matK genes. This assessment was done to achieve two major goals: (1) compare maternally inherited genes to a biparental gene to see if the topologies are more, equally, or less resolved, Volume 98, Number 1 2011 Morton Xdh for Angiosperm Phylogeny supported, and congruent, and (2) examine and compare the topologies for basal angiosperms and eudicots among the data sets for a greater under- standing of seed plant relationships. Materials and Methods TAXON SAMPLING Several different publications were consulted for taxon sampling schemes, including Savolainen et al. (2000), Soltis et al. (2000), Hilu et al. (2003), and Qiu et al. (2006). In total, 253 species representing 164 families and 247 genera of angiosperms and gymnosperms were sampled, and all but one was for a list of taxa with accession data and GenBank numbers. DNA ISOLATION, GENE AMPLIFICATION, AND SEQUENCING Table 1. Primers used for the amplifies ed with a combination of the primers listed below, rimer 1299-GF, which was designed to amplify the Primer name Sequence *471F 5'-C A ATGTGC RTTTGTY ACYCCYGG-3' *486F 5,-ACYCCYGGKTTTRTIATGTCIATGTA-3, 497F V-TTGTGATGTCGATGTATGCA' 975R 5'-TGCTCCWGCAGCCATCCAGAG-3' *1362R 5;-TC W G ATATTGGACTAGC W GT -3; *1365R 5'-TTYAA RTCHGATATYGGACTRGC-3' *1296F S'-AARTGGTTY GC Y GGIAC ACARAT-3' 1299F S'-TGGTTTGCTGGGACACAAAT-S' 1299-1F 5,-TGGTT^GCHGGSAMHCARAT-3, 1299-GF S'-TGGTTrGCTGGVAVYCARATA' *1869R S'-CGAAAYTCMACCATTCCMCCAGGA' 1872R S'-CGAAAYTCMACCATTCCMCC-S' 1869-1R f^KAAWYTCMACCATTCCMCCAGG-3' 1354R 5,-CAGAAACTATCCATKICCC-3/ For most samples, total DNA was isolated from fresh or herbarium specimens, although some was acquired from previously extracted samples. Leaves were ground with a mortar and pestle and subse- quently extracted using the DNEasy plant DNA extraction Kit (Qiagen, Valencia, California, U.S.A.) following the manufacturer’s protocol. Specific poly- merase chain reaction (PCR) and sequencing primers were initially designed using the Arabidopsis plant sequences and other organisms available in GenBank and used by Rodrfguez-Trelles et al. (2003). Additional primers were redesigned as needed. See sequencing. Cloning was performed on selected taxa using the pGem T-easy vector system II kit (Promega Corporation, Madison, Wisconsin, U.S.A.) following the manufacturer’s protocol. PCR was performed with 10 pmol/L primers in 25-pL reactions. In general, it was conducted using the following program: 96°C for 1 min. followed by 34 cycles of 94°C for 30-60 sec., 48°-60°C for 1 min., 68°C for 60-80 sec., and finally 72°C for 7 min. For some taxa, the initial 96°C for 1 min. and the final 72°C for 7 min. were eliminated. Primers used for amplification of clones were M13F and M13R. For sequencing reactions, the Big Dye Terminator Sequencing 3.1 kit (Applied Biosystems, Foster City, California, U.S.A.) was used, and fragments were separated on an ABI3730 DNA Analyzer from Applied Biosystems. PHYLOGENETIC ANALYSIS the MP analysis, a heuristic search with PAUP* (Swofford, 2004) was completed using the following strategy: addition sequence random, number of replicates 100, tree bisection-reconnection branch swapping, MULTREE option on, and MAXTREE auto-increased by 100. For likelihood analyses, the program MODEL- TEST v. 3.6 (Posada & Crandall, 1998) was used to select the models of nucleotide evolution. The program GARLI v. 0.96 (Genetic Algorithm for Rapid Likelihood Inference; Zwickl, 2006) imple- menting the general time reversible (GTR) substitu- tion model, with the base frequency and proportion of The parsimony bootstrapping analyses (100 repli- cates) were conducted in PAUP* (Swofford, 2004) using simple taxon addition, one tree held at each step during stepwise addition, tree bisection-recon- nection branch swapping, MULTREE option on, and no upper limit of MAXTREE set. Among plants, only Arabidopsis has been previ- ously reported to have two copies of Xdh, and these copies have a high degree of similarity (94%). Hesberg et al. (2004) reported that the physicochem- ical properties of Arabidopsis AtXDHl and AtXDH2 Annals of the Missouri Botanical Garden might be nearly identical because of their high degree of similarity (94%). For this study, over 80 clones were sequenced representing 20 taxa. Taxa that had more than one copy had 91%-99.8% similarity among sequences, such as Coriaria sarmentosa G. Forst. (sp.) of the Cucurbitaceae having a 95% similarity among cloned sequences. Only Tacca J. R. Forst. & G. Forst. and Poly gala L. had similarities of 71% and 58%, respectively, among the cloned sequences. In both taxa, once the highly divergent sequence was removed, homology among the remain- ing sequences was 98% and 99%, respectively. COMPARING METHODS OF ANALYSIS Sequence variability Approximately 1265 bp were sequenced, resulting in 1557 aligned characters due to the insertion of gaps. Sequences were aligned manually using Sequencher 4.7 (Gene Codes Corporation, Ann Arbor, Michigan, U.S.A). Eleven indels ranging from 3 to 66 bp in length were inserted in the alignment. Of the aligned characters, 1187 (76%) were variable, 1068 (69%) were potentially parsimony-informative, and approximately 180 (12%) were missing. The Lransilionrlrans version ratio in X dh was 1.10, indi- cating that a high level of saturation was not reached (ratio of 0.4 and below is an indication of highly saturated sequences, according to Holmquist [1983]). Phylogenetic results Parsimony trees containing 253 taxa produced 21,585 trees of 31,337 steps in length, with a consistency index (Cl) = 0.11 and a retention index (RI) = 0.50 (940 hours used). Homoplasy levels are comparable with those from Soltis et al.’s (2000) three-gene data set (Cl = 0.12, 567 taxa) and Hilu et al.’s (2003) matK gene data set (Cl = 0.14, 374 taxa). The discussion is based on the likelihood tree because of its better resolution and support, and the results from the MP topology will be contrasted only when relevant. Rootstrap support of 50%-74% is considered low, 75%-84% moderate, and greater than 85% high (Chase et al., 2000). The likelihood tree (Figs. 1-8) has a total of 190 of the 253 nodes (71%) that of 50% or greater, and support levels averaged 71%. The likelihood tree depicts the angiosperms as monophyletic (80%) with Ceratophyllum (Ceratophyl- laceae) as a sister to the rest of the flowering plants, followed successively by Amborella (Amborellaceae), and Nymphaea L. (Nymphaeaceae) and an Austro- baileya C. T. White-Kadsura Juss.-Schisandra Michx. (Austrobaileyales) clade as sister to the remaining angiosperms. Acorus L. (73%) is sister to the remaining monocots (74%). A weakly supported magnoliid clade (42%) consists of Piperales and Canellales (56%) and Laurales and Magnoliales (90%). Acorus plus the remaining monocots, magno- liids, and Chloranthaceae sequentially diverge after the Austrobaileyales. Eudicots include a basal grade of Ranunculales and Proteaceae, Sabiaceae, Trocho- dendraceae, Buxaceae, Gunneraceae, and Dillenia- ceae and Santalaceae, which form a grade to the remaining eudicots. The grade of taxa containing Ranunculales and Proteaceae, Trochodendraceae, and Gunneraceae to the eudicots received 72%, 71%, and 72% support, respectively. The remaining eudicots are split into two clades. The first clade consists of the Ericales, Comales, and euasterids I and II (lamids and campanulids), i.e., the asterid clade (Asteridae, Fig. 5). The second clade consists of the following orders: Saxifragales, Myrtales- Caryophyllales-Cucurbitales, Crossosomatales, Ger- aniales, Rosales-Fabales-Fagales, Celastrales, Mal- pighiales, Brassicales-Malvales, Oxalidales, and Sapindales (Figs. 6-8). EAR D ER N NO R1V Early-diverging angiosperms have been the subject of intensive molecular studies (Chase et al., 1993; Soltis et al., 1997, 1999, 2000, 2005; Hoot et al, 1999; Mathews & Donoghue, 1999, 2000; Parkinson et al., 1999; Qiu et al., 1999, 2000, 2005, 2006; Barkman et al., 2000; Graham & Olmstead, 2000; Savolainen et al., 2000; Zanis et al., 2002, 2003; Borsch et al., 2003; Goremykin et al., 2003; Hilu et al., 2003; Aoki et al., 2004; Stefanovic et al., 2004; Lohne & Borsch, 2005). Recent studies have found the earliest diverging lineages of the angiosperms to be members of the Amborellaceae and Nymphaea- ceae, either as a clade or grade, followed by the Austrobaileyales (Savolainen et al., 2000, atpB ; Soltis et al., 2000, 2005; Hilu et al., 2003; Qiu et al., 2006). The Xdh gene had a topology supported by 81% bootstrap values for the monophyly of the ng of Ceratophyllum, followed by Amborella, Nymphaeaceae, and Austrobaileyales (Fig 1). The position of Ceratophyllum has varied widely in previous studies: sister to either all angiosperms (Chase et al., 1993; Savolainen et al., 2000 [rbcL, atpB/rbcL ]), sister to monocots (Zanis et al., 2002, 2003; Qiu et al., 2005; Soltis et al., 2005), sister to Chloranthaceae (Qiu et al., 2006), sister to eudicots Volume 98, Number 1 Morton 67 (Soltis et al., 2000; Hilu et al., 2003; Qiu et al., 2005, 2006), or sister to Acorns, which is sister to the Only one sample of Ceratophyllum was used in our study, which may influence the placement of this taxon in this analysis. The placement of Amborella and Nymphaeales followed by Austrobaileyales is in intensive studies of the earliest diverging lineages, using three mitochondrial, one nuclear, and four chloroplast genes (Qiu et al., 2006), inferences among gene data sets were shown. Comparison of the mitochondrial and chloroplast data sets found that mitochondrial genes supported Amborella-N ymphaeales, while chloroplast genes supported Amborella as sister to remaining angio- sperms. Depending on the data set, Ceratophyllum was either sister to Chloranthaceae, with a likelihood bootstrap value of 26%, or sister to eudicots, with a likelihood bootstrap value ranging from 38% to 51%. Only one nuclear gene was examined in the study from Qiu et al. (2006), and comparisons were not made with the mitochondrial or chloroplast topolo- gies. Previous studies, along with the present CHLORANTHACEAE The position of Chloranthaceae has varied in previous studies, from sister to all of the euangio- sperms (Qiu et al., 2006; Cantino et al., 2007), sister to Ceratophyllum (Qiu et al., 2006), sister to monocots (Hilu et al., 2003), sister to magnoliids 2002, 2003; Soltis et al., 2005), sister to magnoliids (Savolainen et al., 2000 [atpB data set]), or sister to monocots and magnoliids (Savolainen et al., 2000 [using the rbcL or atpB/rbcL data sets]; Soltis et al., 2000). Most of these placements have had very low to only moderate bootstrap support. This family exhibits numerous distinctive morphological characters and putative synapomorphies, and its fossil record can be traced back to the Early Cretaceous (Friis et al., 1986; Pedersen et al., 1991; Doyle et al., 2003). The Xdh likelihood tree found that Chloranthaceae, represent- ed here by two genera, forms a strongly supported clade (97%) that is sister to all monocots, magnoliids, and eudicots (Fig. 1). The most parsimonious tree Annals of the Missouri Botanical Garden (MPT) differs in that Chloranthaceae was sister to all ACORUS PLUS REMAINING MONOCOTS The Xdh sequences provided moderate support (73%) for the monophyly of the monocots, with Alismatales (Araceae and Tofieldiaceae) followed by Acorns as sister to the remaining monocots (Fig. 1). This topology has been found in most single and multigene analyses (Chase et al., 2000, 2006; Savolainen et al, 2000 [rbcL, atpB/rbcL ]; Soltis et al., 2000, 2005; Hilu et al., 2003). The monophyly of the Alismatales received high support (100%), along with the monophyly of the remaining monocots (100%). The internal structure of the monocots is represented by two clades. The first clade contains Dioscoreales (100%) and Pandanales (Cyclanthaceae and Yelloziaceae, 88%) with 71% support. The second clade includes sister to Asparagales (Xanthorrhoea- ceae, 100%), followed by Liliales (Smilacaceae), and two nested sister clades, one consisting of Poales (Bromeliaceae) and Arecales with weak support (32%) and the other of Zingiberales (Marantaceae and Strelitziaceae, 100%). These relationships are highly congruent with those suggested by Hilu et al. (2003). Volume 98, Number 1 Morton 70 Annals of the Missouri Botanical Garden Volume 98, Number 2011 Annals of the Missouri Botanical Garden was different than that in most other analyses. However, this position for Asarum was also found in the study by Hilu et al. (2003), but, with increased taxon sampling, Qiu et al. (2006) showed that Asarum regrouped with members of Aristolochia- ceae. Morphological characters clearly place Asarum in Aristolochiaceae, with aristolochic acids, sepals connate, filaments slightly adnate to the style, ovary inferior, and ovules numerous among their common traits (Judd et al., 2007). Limited taxon sampling may be influencing the placement of Asarum in this analysis. Increasing the sampling of Aristolochia- ceae, Piperaceae, and Saururaceae may result in different inferences. The Magnoliales formed a weakly supported clade (65%), and all families with more than one genus are monophyletic. The Annonaceae, Degeneriaceae, Eupomatiaceae, Himantandraceae, Magnoliaceae, and Myristicaceae have occupied various positions in most major analyses (Qiu et al., 1999, 2000, 2005, 2006; Zanis et al., 2002; Hilu et al., 2003). Xdh sequences provide evidence that Himantandraceae is sister to the remaining members of the order, followed by Myristicaceae, Degeneriaceae, and Magnoliaceae, which are sister to the clade Eupomatiaceae and Annonaceae. The Laurales formed a moderately supported clade (78%) with all multitaxa families remaining mono- phyletic, except for the Lauraceae. The nonmono- phyly of Lauraceae was also found by Qiu et al. (2006) using several taxa, but Chanderbali et al. (2001) examined additional taxa and found a clade containing Cryptocarya R. Br. to be basal in the Lauraceae. The placements of Atherospermataceae, Calycanthaceae, Gomortegaceae, Hernandiaceae, Lauraceae, Monimiaceae, and Siparunaceae of the Laurales have varied in most major analyses, and their relationships will remain unresolved until more examinations are completed (Qiu et al., 1999, 2000, 2005, 2006; Chanderbali et al., 2001; Zanis et al., 2002; Hilu et al., 2003). Although the magnoliid grouping has been found in many other molecular analyses, sound morpholog- ical or phytochemical characters to define the groups are lacking (Doyle & Endress, 2000). Soltis et al. (2005) cited some chemical compounds such as neolignan and P-type sieve-tube plastids for most of the families, and Stevens (2001, onward) mentions Volume 98, Number 2011 Annals of the Missouri Botanical Garden sister-group relationship of Dillenia L. (Dilleniaceae) and Osyris L. (Santalaceae) (45%) has been found in other studies (Soltis et ah, 2000, 2005), but the relationships of Dillenia are uncertain, with many lainen et al., 2000; Soltis et ah, 2000, 2003, 2005; Hilu et al., 2003). The relationships of the Santala- ceae are equally uncertain (compare Savolainen et al., 2000; Soltis et al., 2000, 2003; Hilu et al., 2003). and consists of the four major lineages Comales, Ericales, and euasterid I and II. In the Xdh tree the Ericales-Comales clade is strongly supported (97%, Fig. 4). The euasterids I and II (Fig. 5) formed poorly supported sister clades (63%). Although support values are difficult to compare, this topology is mostly in agreement with other multigene studies (Savolai- nen et al., 2000; Soltis et al., 2000). Ericales-Comales EUDICOTS CLADE I Asterids The asterid clade, sensu Olmstead et al. (1992, 1993) and Judd and Olmstead (2004), is monophy- letic (58% bootstrap support) in the Xdh tree (Fig. 5) The monophyly of Ericales received moderate support (83%) in the Xdh analysis, while the monophyly of the Comales lacks bootstrap support (31%, Fig. 4). The resolution and support for the Ericales clade were similar to those in other analyses, including Bremer et al.’s (2002) six-gene plastid 76 Annals of the Missouri Botanical Garden first clade includes Iteaceae sister to Grossulariaceae (100%). A similar topology was recovered by Fish- bein et al. (2001), Hilu et al. (2003), and Soltis et al. (2005). The second clade consists of a monophyletic Hamamelidaceae (100%), which has been document- ed in many other studies (Savolainen et al., 2000 [atpB/rbcL]; Soltis et al., 2000, 2003, 2005; Fishbein et al., 2001; Hilu et al., 2003). A third clade including Altingiaceae, Cercidiphyllaceae, Daphni- phyllaceae, Hamamelidaceae, and Paeoniaceae is well supported (100%). Several major families are missing, such as Saxifragaceae and Crassulaceae. Therefore, it is still premature to comment on the position of the taxa in the third clade. Soltis et al. (2007a) also were unable to recover stable relation- ships among the woody Saxifragales. The Saxifragales ( inscription contains a wide range of morpholog- ical diversity that departs from traditional classifica- tion. The nonmolecular synapomorphies have not yet been found and more research is needed for clarification. Myrtales-Caryophyllales-Cucurbitales The analyses recover a weakly supported (6%) clade consisting of Myrtales, Caryophyllales, and Cucurbitales (Fig. 6). This inference may be misleading or incorrect due to the weak support because no other analyses have grouped these orders together. In contrast, the MPT is like previous MP analyses, revealing Caryophyllales as sister to the eudicots, and Cucurbitales as sister to Fagales. In a variously placed at the base of the asterids (Hilu et al., 2003; Soltis et al., 2005), sister to the Saxifragales (Soltis et al., 2003), sister to the eudicots (Savolainen et al., 2000 [ rbcL , atpB/rbcL ]; Soltis et al., 2000), or sister to the asterids and rosids (Savolainen et al., 2000 [ atpB ]). Cucurbitales are typically found associated with a nitrogen-fixing clade, which includes Rosales, Fagales, Fabales, and Cucurbitales (Soltis et al., 1995, 1997, 2000, 2003, 2005; Savolainen et al., 2000 [ atpB , rbcL, atpB/rbcL ]; Hilu et al., 2003). The position of the Myrtales is usually nested within, or sister to, a clade consisting of Malvales, Brassicales, and Sapindales (Soltis et al., 1997, 2000, 2005; Savolainen et al., 2000 [atpB, rbcL, atpB/rbcL ]; Hilu et al., 2003). The Myrtales clade in the Xdh data set has two members with partial sequences that may be creating errors in their placement. In all three orders, the families are underrepresented; we sampled only two of the approximately 13 families of the Myrtales, seven of the approximately 29 families of the Caryophyllales, and three of the approximately seven families of the Cucurbitales. The differences between the likelihood genes in independent and combined analyses. Both families, are most likely contributing to this unusual placement of the Myrtales. Two families were sampled for the Myrtales (63%), Myrtaceae and Penaeaceae. Metrosideros Banks ex Gaertn., Myrciaria 0. Berg, and Psidium L. formed a monophyletic Myrtaceae with low support (63%, Fig. 6). This order has had various circumscriptions in the past, consisting of a number of families, and it would be premature to speculate on its composition, as well as relationships within the order and with other rosids. The monophyly of this order is in agreement with the results of Conti et al. (1996) based only on rbcL, Savolainen et al. (2000) based on atpB/rbcL, Soltis et al. (2000, 2005) based on rbcL/atpB/ 18S, and Hilu et al. (2003) based on matK. The Xdh sequence data strongly support the circumscription of the Caryophyllales (95%, Fig. 6). The monophyly of this order has been indicated in previous molecular studies (Savolainen et al., 2000 [atpB, rbcL, atpB/rbcL ]; Soltis et al., 2000, 2003, 2005; Cuenoud et al., 2002; Hilu et al., 2003). Two clades within Caryophyllales are recognized with Xdh data. Caryophyllales I (100%) includes Caryophylla- ceae and Nyctaginaceae, followed by a well-support- (100%). Caryophyllales’ II (39%) contains a clade of the carnivorous families Droseraceae and Nepentha- ceae (98%) sister to a clade of Plumbaginaceae and Simmondsiaceae (57%). The position of the Sim- mondsiaceae has varied, being included in the Caryophyllales II clade in some analyses (Savolainen et al., 2000 [rbcL/atpB]), while, in others, it has been included in the Caryophyllales I clade (Soltis et al., 2000, 2005; Cuenoud et al., 2002; Hilu et al., 2003). One of the most comprehensive studies of the Caryophyllales, by Cuenoud et al. (2002), indicates that the position of Simmondsia Nutt, lacks strong support, demonstrating that additional work is needed. Our results show Simmondsiaceae as part of the Caryophyllales II clade with strong support. Members of Caryophyllales have some non-DNA features supporting its basic circumscription, but many gaps still remain to be filled (for characters, see Nandi et al., 1998; Stevens, 2001, onwards). The Curcurbitales were defined by the Angiosperm Phylogeny Group (1998) and Angiosperm Phylogeny Group II (2003) as containing the following seven families: Anisophylleaceae, Begoniaceae, Coriaria- ceae, Corynocarpaceae, Cucurbitaceae, Datiscaceae, and Tetramelaceae. Begoniaceae, Coriariaceae, and Volume 98, Number 1 2011 Morton Xdh for Angiosperm Phylogeny Cucurbitaceae were included in this analysis (Fig. 6), which found the Curcurbitales to be monophyletic, although this was only weakly supported (72%). Other gene studies have also found the Curcurbitales to be monophyletic (Savolainen et al., 2000 [ atpB , rbcL, atpB/rbcL]; Soltis et al, 2000, 2003, 2005; Hilu et al., 2003; Zhang et al., 2006). Morphological and phytochemical synapomorphies for the Curcurbitales are still unclear; even the most recent study by Zhang et al. (2006), using 13 molecular markers and morphological characters, could not identify a Crossosomatales The Crossosomatales clade, here consisting of Staphyleaceae and Stachyuraceae, is well supported (100%, Fig. 6). Both the rbcL/atpB analysis (Savolai- nen et al., 2000) and the analysis of 18S, rbcL, and atpB (Soltis et al., 2000) contained a well-supported Crossosomatales clade. Stachyuraceae has previously been placed in the Violales, whereas Staphyleaceae has been placed in the Sapindales (Cronquist, 1981). Thus, another group of families previously placed in distantly related orders come together in a well- supported clade. Matthews and Endress (2005) investigated the floral characters and revealed potential new synapomorphies for a close relationship among the core member of Crossosomatales. Among the prominent floral features of Crossosomatales are solitary flowers, presence of a floral cup, imbricate sepals with outermost smaller than inner, pollen shortly stalked gynoecium, postgenitally united carpel tips forming a compitum, stigmatic papillae 2-cellular or more, ovary locules tapering upward, long integuments forming zigzag micropyles, cell clusters wilh bundles of long yellow crystals, mucilage cells, seeds with smooth, sclerified testa and without differentiated tegmen. However, more molecular and morphological work is needed to resolve both the composition of this clade and its position within the rosids. Geraniales Two taxa of Geraniales were sampled for the Xdh analysis, Geranium L. (Geraniaceae) and Greyia Hook. & Harv. (Melianthaceae) (Fig. 6). These two families are held together in a weakly clade (61%). These families were previously placed (Cronquist, 1981; Takhtajan, 1997) in different orders, and no known morphological synapomorphies hold the present order together. Geraniales have been found in previous analyses (Savolainen et al., 2000 [ atpB/rbcL ]; Soltis et al., 2000) to be sister to the Crossosomatales. While there is no support for that sister relationship here, likelihood values at the base of the grades leading up to the eurosids are low (12% or less) and rearrangements may be expected with additional sampling of taxa and inclusion/combina- tion with more data sets. Eurosid I In the analysis, the nitrogen-fixing clade (including Rosales, Fagales, Fabales, and Cucurbitales; Soltis et al., 1995), excluding Cucurbitales, is weakly sup- ported at the base of the eurosid I grade (20%), followed by Celastrales (88%) and Malpighiales (< 50%) (Fig. 7). Three nitrogen-fixing subclades received low to strong support and have no support as a clade (28%): Fabales (87%), Fagales (100%), and Rosales (65%). These subclades have been placed together with the order Cucurbitales in other single and multigene analyses (Savolainen et al., 2000 [atpB, rbcL, atpB/rbcL ]; Soltis et al., 2000, 2003; Hilu et al., 2003). The Fabales consist of well- supported Fabaceae (100%) and Quillajaceae. The well-supported clade of Fagaceae, Betulaceae, Myr- icaceae, and Juglandaceae. The Rosales are sister to Fabales and Fagales. This order consists of Rosaceae sister to the rest of the order followed by Rhamnaceae sister to Humulus L. (Cannabaceae), Ulmaceae sister to Celtis L. (Cannabaceae), and Moraceae sister to Urticaceae. The Cannabaceae are not monophyletic in the Xdh analysis because they group with members of the Rhamnaceae and Ulmaceae; however, the placement of Humulus is based on a partial sequence, which may have caused anomalies. Soltis et al. (2005) found similar clades including a nonmono- phyletic Cannabaceae in the Rosaceae. As mentioned earlier, the MP analysis found eurosid I to consist of Cucurbitales, Fabales, Fagales, Myrtales, and Ro- sales. The Xdh analysis supports (100%) the Celastrales clade (100%, Fig. 7). Most recent analyses support the order Celastrales, including various families placed next to the Malpighiales (Zhang & Simmons, 2006), Oxalidales (Hilu et al., 2003; Soltis et al., 2005), or the Oxalidales and Malpighiales (Soltis et al., 2000). The Xdh likelihood and MP analysis showed Celastrales in a well-supported clade sister to The Malpighiales form a weakly supported clade Malpighiaceae (100%), and part of the Euphorbia- ceae (73%, Fig. 7). The Euphorbiaceae are not monophyletic, as also found in the analyses by Soltis Ililll llllSISill! 'SlJiSiliitlil Jj III 111 : h J *1 1 I it 1 li ! filili !lfll!ii!!i!l!fflil!ll nn i ®!-l Iffliill'il !l •IllllilllS 'l llll I I 1 liil I jjjllll ! j ! 1 ! I in i i! ill j i * 1 i 3 i pisisjlipi#! | yiiiliiliiilii! il 1 n fiiSitt ill lii i Hl3siiS“sSsli5l I»Ii»IsJ3eJis II liUll lllfill I j i I iyaisl ill nil ini fill Hull I i f i I ! i n Si sllfifllisisiiiSsiiJi liiBi yiliilillsilPlill! j. j,i . f ,3 llllliili 1111 ! J I S i n | I „ | EssIlsisislsslSlilsl; jlilllilillllil illllllilllii ! 3 I 1 i’i n .1 gifts! 1 illllillilll If B - -S iJ Llijll ffliiiii li i ! J 1 1 I i REVISION OF PALEOTROPICAL MEGALASTRUM (DRY OPTERID ACEAE)1 Annals of the Missouri Botanical Garden Volume 98, Number 2011 POLLINATION SYNDROMES OF Petra Wester and Regine NEW WORLD SALVIA SPECIES WITH SPECIAL REFERENCE TO BIRD POLLINATION1 Volume 98, Number 1 2011 Wester & ClaBen-Bockhoff 1( Pollination Syndromes of New World Salvia Cultivated plants (83 species) were examined at the following places: (1) the Botanical Gardens Mainz (mainly wild collections; Germany), Hamburg (Ger- many), University of California Botanical Garden at Berkeley, Fullerton Arboretum, Rancho Santa Ana Botanic Garden, Univerity of California Riverside Botanical Gardens, The Huntington Botanical Gar- dens, Tilden Regional Park Botanic Garden (all California, U.S.A.), Chihuahuan Desert Gardens (El Paso, Texas, U.S.A.); (2) the nurseries E. Hiigin (Freiburg, Germany) and Native by Native Land- scapes (Johnson City, Texas, U.S.A.); and (3) the private gardens of F. Bemdt (Cochabamba, Bolivia), B. Clebsch (Santa Cruz, California, U.S.A.), and M. Dimmitt (Tucson, Arizona, U.S.A.). Fresh flowers were fixed in 70% ethanol for further investigations. Vouchers are deposited at MJG with some duplicates in B, BIGU, UC-JEPS, K, LPB, MEXU, TEX-LL, and UCR (Appendix 1). Herbarium specimens originated from AAU, ARIZ, B, BIGU, BM, C, CGE, COL, COLO, F, FLAS, G, GH, GOET, H, HAO, HUT, K, L, LD, LPB, MEXU, MICH, MJG, MO, NA, NY, OAX, P, RSA, S, SD, TEX-LL, UC-JEPS, UCR, US, W, WU, XAL, and ZEA (see Appendix 1). Specimens examined in the field or herbarium specimens were identified by means of the literature (see Appendix 1), type material, and with the aid of J. Wood (OXF), A. Vazquez (IBUG), A. Espejo (UAMIZ), H. Vibrans (CHAPA), M. Veliz (BIGU), C. Froissart (Olivet, France), and A. Sanders (UCR). Because there is no actual revision of Salvia subg. Calosphace (Benth.) Benth., we did not include species with an unclear taxonomic status (partly listed as nomen dubium by Epling, 1939, or not listed in the floras), unde- scribed species (e.g., species ined., or sp. A in Pool, 2007), and probable hybrids from cultivation (S. ianthina Otto & Dietr.). POLLINATOR OBSERVATIONS Observations of floral visitors including birds, bees, butterflies, hawkmoths, and flies were made in the field, botanical gardens (Berkeley, Rancho Santa Ana, Riverside, and Fullerton Arboretum), and a private garden (F. Bemdt, Bolivia; see also Wester & ClaBen-Bockhoff, 2006a, 2007). The birds were determined after Schuchmann (1999) and with the help of J. F. Ornelas and C. Gonzalez (both from Instituto de Ecologfa, Xalapa, Mexico), M. Ordano (Universidad Nacional Autonoma de Mexico, Mexico City, Mexico), O. Reyna (Universidad de Guadala- jara, Mexico), and G. Stiles (Universidad Nacional de Colombia, Bogota, Colombia). POLLEN TRANSFER SIMULATION To test the fit between flowers and pollinators and confirm the exclusion of certain visitors as pollina- tors, the process of pollen transfer was experimentally simulated. For this purpose, a museum skin or metal rod was inserted into the fresh flowers (see Wester & ClaBen-Bockhoff, 2006a, 2006b). The functionality of the lever mechanism, the morphometric needs for a successful pollen transfer, and the site and precision of pollen deposition were recorded at the living material. The skins were borrowed from the Coleccion Boliviana de Fauna, La Paz, Bolivia (CBF), and the Zoologisches Forschungsmuseum Alexander Koenig, Bonn, Germany (ZFMK). PLANT TRAITS To identify the pollination syndrome, floral structures were morphologically and morphometrical- ly investigated: The length of the flower was measured from the proximal end of the flower to the distal end of the proximal end of the flower to the flower entrance, and the upper lip from its distal end to the flower entrance. Because the position of the lower lip is important for excluding bees, it was classified as reflexed (bent downward more than 90°), deflexed (bent downward about 90°), or antrorse (bent downward less than 90° to 0°). It was noted when the lips are oriented close to each other (lengthening the tube) and when the than horizontally, oriented. The lower lip was called cup-shaped when its apical margins were incurved. Flower shape was classified as bilabiate when the was more dominant. Flowers with a combination of both shape characters were classified as intermedi- ate. Stielteller flowers (sensu Vogel, 1954) have long, slender tubes with lobes spreading at a right angle (salverform). Structures for nectar retention at the base of the flower were classified as definite constrictions (emarginations, folds) of the corolla, definite (papil- lae) or weak (ridges) outgrowths of the corolla wall and hairs, and posterior lever arms (part of the staminal connectives reaching the proximal part of the flower). Thecae exsertion indicates the distance between the distal end of the upper lip to the distal end of the 104 Annals of the Missouri Botanical Garden The corolla color of fresh flowers i after the CMYK color chart (C: cyan, M: magenta, Y: yellow, K [here S]: black; Kiippers, 1999). Color data from the literature are often not well defined. We therefore classified color descriptions as follows: red includes all colorations from scarlet, claret, roseate to includes crimson; lavender includes light violet; blue includes royal blue; and dark violet includes almost black. If more than one color is mentioned, the species is defined in different ways or has color morphs. If known, the most frequent color was used. In a second step the colors were grouped: (a) yellow and orange; (b) red, orange, and yellow; (c) red; (d) red, purple, and pink; (e) purple, pink, and lavender; (f) blue and dark violet; and (g) white. Flowers that were not assignable to these classes because they include combinations with brownish ochre, red- brown, violet, or pale violet were grouped separately. Floral nectar was measured as standing crop from cultivated plants (Botanischer Garten Mainz), usually in the morning of the first day of anthesis. Sugar concentrat in was determined using handheld refrac- tometers (N1 [Atago, Tokyo, Japan]: 0%-32 % sucrose percentage by weight [w/w]; N2 [Atago]: 28%-62% sucrose w/w; Eclipse 45-81 [Bellingham & Stanley, Kent, U.K.]: 0%-50% sucrose w/w). The volume of nectar was measured with a 25- pi microsyringe (ILS, Stiitzerbach, Germany). Data on distribution, altitude, habitat, and growth forms were based mainly on the literature (see Appendix 1) and completed by our own observations. Because the growth form concept is only partly adequate for (sub-)tropical plants, we only roughly classified the species, mentioning more than one growth form when needed. CLASSIFICATION Within the 602 examined New World sages, four groups can be distinguished. (1) Species with flowers that definitely exclude insects, but fit to birds. These are termed omithophilous and are found in 184 species (30.6%). (2) Flowers that definitely allow bees access to nectar and morphologically fit to them. They form the largest group with 351 species (58.3%) and are grouped as melittophilous. (3) Species with flowers that exclude bees but enable butterflies or long-tongued flies to reach the nectar and touch the reproductive organs. This group only includes one species that is assumed to be psychophilous. (4) The remaining species (10.9%) do not fit to one of these groups. They are either so variable that they allow several pollinator guilds access to nectar or are not sufficiently known to be grouped (see Appendix 2). In 17 species, hummingbirds were observed to pollinate the flowers, i.e., to touch pollen sacs and stigmas, while bees and butterflies were also attracted but did not pollinate. Based on these data, we identify them as omithophilous. They share the following (floral) traits (Table 1A, B): Total flower length/corolla tube length. The total flower length ranges from 13 cm in Salvia dombeyi Epling (Fig. 1A) to 7 mm in S. confertiflora Pohl (Fig. ID), and the length of the corolla tube ranges from 9 cm to 5 mm in these species, respectively (Figs. 2, 3). Regarding the geographic distribution, the longest flowers occur in the Andes (especially Pem, S. dombeyi ), in Brazil (also where the shortest flowers occur), and in Mexico and Honduras. Pollinator groups were inferred from visitor observations; floral traits including size, shape, color, nectar, and scent; the fitting between flower and pollinator; and the capacity of the flower to exclude nonpollinator groups. We defined the exclusion of bees from the nectar by long and/or narrow corolla tubes, the exclusion of bees and butterflies by lacking slender corolla tubes. Visitors are not pollinators if they do not touch the reproductive organs, e.g., if the stigma or pollen sacs are too far exserted to be touched by a bee. We are aware that during different seasons and at different localities the range of pollinators may vary, but nevertheless only animals that fit to the flower can act as pollinators. Flower shape. The flower shape is bilabiate ii species (18%; Fig. IB, L, M, 0, Q, R), tubular in species (59%; Fig. 1C-K, N), ; species (23%; Fig. 1A, flowers is mostly straight (parallel sides, e.g., in Salvia iodantha Femald or S. nervata M. Martens & Galeotti in Fig. 1C, I), elliptic in outline (Fig. 1G), or funnelform (e.g., in S. macrophylla Benth., S. pauciserrata Benth., or S. oppositiflora Ruiz & Pav. in Fig. IE, F, N). Stability of the corolla. The stability of the corolla ranges from being stiff as in, e.g., Salvia confertiflora, S. karwinskii Benth., or S. madrensis Seem. (Fig. ID, Volume 98, Number 1 2011 Wester & ClaBen-Bockhoff 105 Pollination Syndromes of New World Salvia Stability of the lower corolla lip. The stability of the lower corolla lip varies from stable, e.g., in Salvia confertiflora (Fig. ID) to weak, e.g., in S. patens Cav. Lower corolla lip length/position. The length of patens to 1 mm in S. nervata (Fig. II) and shows diverse positions, being mostly antrorse (e.g., S. dombeyi, S. gravida Epling, and S. atrocyanea Epling in Fig. 1A, L, M), but also deflexed (e.g., S. oaxacana Fernald in Fig. 10) to reflexed (e.g., S. lasiantha Benth., S. macrophylla, S. pauciserrata, and S. karwinskii in Fig. IB, E-G). The lower lips may be additionally cupped with incurved front margins (e.g., S. confertiflora and S. leucantha Cav. in Fig. ID, H). The lateral lobes may be oriented vertically (e.g., S. confertiflora . S. leucantha, and S. madrensis in Fig. ID, H, K). In flowers with weak, very short, reflexed, or deflexed lower lips, the landing of insects is impossible or difficult. The same applies for antrorse lower lips that lengthen the functional tubes (e.g., in S. confertiflora and S. leucantha in Fig. ID, H). Antrorse lip positions bent downward either slightly to almost a right angle may offer a potential landing platform; however, bees are excluded by tube length. In species with short flowers, the tubular shape and increased tube length caused by the antrorse lower Corolla color. The flowers in omithophilous Salvia species are always conspicuous. The color of the corolla is most often red (49%), but also orange, yellow, purple, pink, lavender, violet, and blue, and rarely brownish ochre or red-brown (Fig. 4). A weak color change occurs in S. graciliramulosa Epling & Jativa and S. spathacea Greene. The calyx is mostly green or sometimes weakly colored in portions. Especially in species with white flowers, the calyx can be more strikingly colored, e.g., mostly purple in S. leucantha (Fig. 1H), purple or white/whitish in S. tomentella Pohl, or light violet to whitish in S. divinorum Epling & Jativa. Several other species have colored calyces and bracts, e.g., red in S. confertiflora (rig. ID) and S. regia Cav., yellow in S. madrensis (Fig. IK), pink in S. involucrata Cav., sometimes purple in S. gravida and S. lasiantha (Fig. IB), dark red to purple or almost black in S. spathacea, violet-black in S. atrocyanea, or blue- black or whitish to light green in S. paramicola Fern. Alonso. Regarding the geographic distribution, there are no noteworthy characteristics in flower color, except a high percentage of red-flowered taxa in Brazil, whereas blue and dark violet flowers occur Central America. Nectar guides. Nectar guides are usually lacking, but in at least nine species they occur on the lower lip near the entrance and in at least 15 species they are sometimes present. Nectar guides are mostly white and sometimes dark red-blackish; both colors occur in Salvia exserta Griseb. The whitish stamens may contrast to a colored corolla, as for example in S. atrocyanea (Fig. 1M). In S. alborosea Epling & Jativa, Nectary length. The nectary length varies from 1 to 5 mm and some nectaries are very voluminous (Table 1A, B). Nectar retention structures. Nectar retention structures at the base of the corolla are rather common in omithophilous Salvia species. In almost all species they appear as slight lateral and horizontal narrowings of the corolla tube (not included in Table 1A, B). In addition, there can be strong lateral or abaxial constrictions (including folds), lateral ridges, papillae or posterior staminal lever arms. Toothlike structures of staminal < Salvia species of subgenus Calosphace, the staminal connectives often form toothlike structures near the abaxial side of their joints. They are relatively large in S. lasiantha, S. atrocyanea, S. exserta, or S. guaranitica A. St.-Hil. ex Benth., but small in S. dombeyi, minute in S. gravida, or even absent as in S. oppositiflora. In all omithophilous Salvia flowers, they do not occlude the flower entrance, because they are either arranged laterally (S. atrocyanea in Fig. 1M; S. exserta) or greatly distal (S. lasiantha in Fig. IB). Thecae position. The thecae position varies from being enclosed by the upper lip in 101 species (e.g., in Salvia dombeyi or S. atrocyanea in Fig. 1A, M) to being exserted as much as 43 mm (S. speciosa C. Presl ex Benth.) in 64 species (e.g., S. lasiantha, S. iodantha, S. macrophylla, or S. pauciserrata in Fig. 106 Annals of the Missouri Botanical Garden HI 1 f 3 S i S 1 i i g lg g 1 1 1 1 §s i Sg 1118 li il> ii 8-16 R1 [4-8] Cb 0 P [4-8] 5-8 Cb 0 0? Cb 0 PL 6-10 PL 4-9 Cb 3-9 3-9 PL? 0 PL 1-3 position* 1 . .222 j ^?22 * S | f , | j, - U , .L:. ! ..j-,;;: . . 3. -II=SS 3!t|I| ISJSJ 3 3333 3. it. Uhl-SHI hills I3|| 3 ij-.njf Mi numm mm nm i mmi »> numm mm mn i mna 1 1 llilffiliiyiilsl! Volume 98, Number 1 2011 108 Volume 98, Number 2011 Wester & ClaBen-Bockhoff 1( Pollination Syndromes of New World Salvia 110 Volume 98, Number 1 2011 112 14 u II' H I i II jl, »!' If* i ls|Si I i? Ill i S llsHl ]. ji III j i i is££%£ 3ooooo .oJS * .. ' g'ls g i „ % /" ^ S * S * * 4 tS 4 ## HI3S 3|SP urn *«« is? mn mm mn j ,.,ij j I 3|JiIIlljilj IIH1IM ill I E £ S S £ £$,& s£ S£ ScS EcS hi h o3 cl 333 m u 22 | gL 333 4 33 .33 33s m a p p ji in a p n m iii j m j l i3 m Wifi UMif ] ii ip Ip ip !* f f rw rrrrrri I h ills i 3§tf=Sf’s^"3g8eS8aS ■a i i % -s -a 3 33 3 S J 3 S -a s -g -a g £ -s -s i ! Ii!n i! I 2$ ^ Ss S‘ I fl i i i11 i i I1 i SIM I ! ■ . Hi iii 1 1 1 1 1 1 j i 1 ! * M i Hi f f HI M 116 118 Annals of the Missouri Botanical Garden li il> ill ii- if mimm &L s s'5s S a f If - » ■s -a -a r s -a -a s -a a a a a -a a a -a -a 3331 2 3 3 -3 I s 2 2 3 " 3 4«i? PP III II ! n 1 1 / mu* I| H i i la j ! i m 1 1 1 11 I 1 i 3 120 Volume 98, Number 1 2011 1 mm i i lit 1 1 1 I 1 1 ; Mi! ! >«« 1 li ph PE ph, ss MX sh EC ph MX sh, sometimes MX, US a a “a a § s' § S § 3 a 8 £ 8 8 £ a 9 a s -a -s -a -s g -a g g s-a-s II' 3 35312 3 - 2 3 l 2 I 3 2 5 3 3 If' * 3 !?•« *1* 1 m 24.4 ± 3.5 (18.8- 1 32.5), n = 58 r 24.5 ± 2.4 (17.8- 6.9 ± 4.1 (2.5- 31.2), n = 28 13.5), n = 5; m 1 H i, II I1 II II1 Is t 1 I 1 a I J . nun regnelliana Briq. ™MeWin^ardnere*H' B‘ rosei Femald rabrifaux Epling ntbriflora Epling mfula Kunth sagittate Ruiz & Pav. I 122 Annals of the Missouri Botanical Garden i 1; i 1 1 § a s li "i ph sh (sometimes ph sh sh phorss ph sh ph sh ph ph ph to sh ah b: !!<■ ’ 3a 3 35 3 3 3 s S 3 3 3 S 3 S fl> 3 3 3 JJ‘ - If - - - If II I :* 5* H til t I1!1 t }i I i s li 1 j 1, it 1 J i i ! i , 1 1 . 1 i i ii Pi i! i 1 1 1 Pi i H IP Volume 98, Number 1 2011 11 II* !fi 1|\ I i n m S • n 3 i £ 3 S !l £ 124 Volume 98, Number 1 2011 Wester & ClaBen-Bockhoff 125 Pollination Syndromes of New World Salvia IB, C, E, F). In 11 species the position of the thecae is variable (Table 1A, B). No noticeable flower Pedicels. The pedicels are mostly flexible, and their length ranges from 34 mm in Salvia dombeyi to being absent in S. spathacea. Growth form. As to the growth form, almost all bird-pollinated Salvia species (97%) are perennial herbs, subshrubs, or shrubs. Only S. exserta and S. subrotunda A. St.-Hil. ex Benth. are annuals, while S. coccinea Buc’hoz ex Etl. varies from annual to perennial. Three species (S. dombeyi, S. regia, S. sessei Benth.) are scandent subshrubs, shrubs, or sometimes small trees. Distribution. The omithophilous sages show a wide distribution in the New World, reaching from California to Chile and Argentina (Fig. 5, Table 1 A, B). Comparing the different countries, Mexico has the largest number of omithophilous species (57 to 58 spp.) and omithophilous endemics (41 to 42 spp.). However, considering the total number of Salvia species within the country (about 300 spp., the largest species number within a single country), the relative percentage of the omithophilous species is relatively low (about 19%) compared to other countries. Centers of diversity for the bird-pollinated species of Salvia are (Fig. 5): (1) the central and southern highlands of Mexico and Guatemala; (2) the northern part of the South American Andes (Colombia, Ecuador, Peru, the latter two with about half of the species being bird pollinated); and (3) the southern part of Brazil with more than 50% bird-pollinated species (especially southeastern Brazil, with 25 of the 35 Brazilian species, all endemic to the country). The omithophi- lous species in Cuba and Hispaniola (including Haiti and the Dominican Republic) are also all endemic. Relative to a country area, El Salvador, Guatemala, Haiti, and the Dominican Republic have the highest density of omithophilous sages by species number. Habitat. Bird-pollinated Salvia species occur in all habitats from rainforests to at least mesic habitats in deserts from ca. 200 to 4000 m in elevation. They mainly occur in the highlands of the American and in southeastern Brazil from ;a. 1500 to 3000 m, ca. 1000 to 1500 m. Systematics. In terms of systematics, most of the New World omithophilous sages belong to the widespread group Salvia subg. Calosphace (179 spp. in 56 sections). Three species (S. henryi A. Gray, S. roemeriana Scheele, S. summa A. Nelson) belong to Salvia sect. Heterosphace Benth. (Arizona to central Mexico), S. spathacea (California) to section Audibertia (Benth.) Epling, and S. pentstemonoides Kunth & C. D. Bouche (Texas) to section Eusphace Benth. (see Table 1A, B). MELITTOPHILOUS SPECIES Most of the New World sages allow bees to pollinate the flowers (58%; Appendix 2, Fig. 6A-I). Because their flowers correspond in their proportions with bees, provide a landing platform, have in general short corolla tubes, and deposit pollen on the bee’s body, we classify them as melittophilous. Altogether, 65 species of this group were seen as living plants (33 at natural localities, 54 cultivated in botanical gardens and nurseries). Pollinating bees were observed at Salvia stachydifolia C. Presl ex Benth., S. rypara Briq. subsp. platystoma (Epling) J. R. I. Wood, S. cuspidata Ruiz & Pav. subsp. bangii (Rusby) J. R. I. Wood, S. sophrona Briq., S. carduacea Benth., S. apiana Jeps.. S. mellifera Greene, S. leucophylla Greene, S. dorrii (Kellogg) Abrams, S. munzii Epling, S. columbariae Benth., and other species not listed in detail. Flowers of this group share the following traits: Corolla tube length. The corolla tube is short (< 2 cm), enabling the bees to reach the nectar at the base of the corolla. Corolla color. The color of the corolla is often blue and violet, but also white, pink, purple (Fig. 6A- I), or rarely yellowish. Nectar guides are often found at the flower entrance (Fig. 6C, G, H). Staminal connectives. A peculiar feature of many bee-pollinated Salvia species in subgenus Calo- sphace is the presence of distinct ventral teeth or barriers at the connectives. The latter may be large subsp. platystoma (Fig. 6D) and S. pusilla Femald, for example, the lever mechanism is triggered by pushing back these barriers. In species like S. amplifrons Briq. and S. cuspidata subsp. bangii, the ventral formations are laterally arranged. Pollen sacs. The pollen sacs of most species are enclosed by the upper lip (Fig. 6A-D, I). Exceptions (exposed pollen sacs) are found in all species of sections Echinosphace Benth. (Fig. 6E, G) and Audibertia (Fig. 6F, H), and in very few species of 126 Annals of the Missouri Botanical Garden reflexed lower lips and greatly exserted thecae. — E. S. macrophylla. — F. S. pauciserrata. — G. S. karwinsJcii, elliptic in outline paposanapL). ' P if folia, S. texana (Scheele) Torr., 5. rypara subsp. costae with thecae and pollen on the bird’s head. Scale bars: A-C, E, F, I-R = 1 cm; D, G, H = 0.5 cm. 128 ing only small amounts of hig] (e.g., Salvia stachy difolia, S. zz Volume 98, Number 1 2011 130 Annals of the Missouri Botanical Garden Volume 98, Number 1 2011 Wester & ClaBen-Bockhoff 131 Pollination Syndromes of New World Salvia more or less fruity scent of the corolla and the calyx was detected. Nectar of medium volume was found protruding 5-11 mm into the tube. Although several populations were studied at roadside localities in western Texas, no pollinators were seen. Nevertheless, we deduce from the floral traits and proportions, tube dimensions, and landing platform that butterflies and/or long-tongued flies are the pollinators of the species. REMAINING SPECIES The remaining New World Salvia species (Appen- dix 2) share characters with species of different groups and thus allow different pollinator guilds access to nectar (see Table 2) or are too little known to be characterized sufficiently. Species probably pollinated by bees and birds. Salvia purpurea Cav. ( Salvia subg. Calo- sphace; Table 2) has highly variable flowers with different corolla colors, flower sizes, and proportions of tube to lower lip length (sometimes even within one individual). Long-tubed flowers with relatively short lower lips allow birds access to the nectar, but a better landing platform for bees, may be pollinated by both birds and bees. Indeed, field observations in perching) and bees visit the flowers, being dusted with pollen. Bees were also found as nectar and hawkmoths stole nectar. Additonally, studies in Mexico showed that the activity of flower visitors varied. At some places hummingbirds were more frequent (Guadalajara); at other localities bees were more abundant (Xalapa). The yellow flowers of Salvia aspera M. Martens & Galeotti ( Salvia subg. Calosphace; Fig. 6L; Table 2) also vary in length. Although the larger flowers with a concentration may attract birds, the smaller ones may also allow bees access to nectar. Their lower lips are relatively narrow, but stable enough to present a landing platform for bees. The dark violet corolla of Salvia concolor Lamb, ex Benth. ( Salvia subg. Calosphace; Fig. 6K; Table 2) offers a large landing platform for bees at its lower lip. However, the long corolla tube probably excludes bees from nectar. The medium amount of nectar with its relatively low sugar concentration also indicates bird pollination. We assume that bees might reach the nectar in short-tubed flowers, but are excluded in longer Salvia mexicana L. (Salvia subg. Calosphace; Table 2) is a quite variable species with two varieties, S. mexicana var. minor Benth. and S. mexicana var. major Benth. In general, the varieties considerably differ in corolla size, in the proportion of tube to lower lip length, and in the lower lip position. Additionally, these characters and the corolla color differ within the varieties and overlap. The large flowers in both varieties and the flowers with a relatively short lower lip in comparison to the flower tube are clearly omithophilous as they exclude bees from nectar. Smaller flowers, especially those with long lower lips, are thought to be intermediate and/or pollinated by bees. Nectar data of the examined individuals in both varieties might be rather attractive to birds. Indeed, at the lighter violet flowers of S. mexicana var. major, hummingbirds were observed visiting the flowers, while bees stole nectar. The following three species have small- to medium-sized corolla tubes and large lower lips that clearly allow bees access to nectar. The blue Salvia scutellarioides Kunth (Salvia subg. Calo- sphace; Table 2) with white nectar guides on the lower lips and rather short corolla tubes has a medium volume of nectar with a relatively low sugar thecae that might not be touched by bees. Salvia discolor Kunth (Salvia subg. Calosphace; Table 2), with a dark violet (almost black) corolla contrasting pollinated flowers in its large lower lip, short- to medium-sized tube, and nectar of high sugar concentration. However, nectar guides are lacking and the large volume of nectar might also attract birds. Salvia retinervia Briq. (Salvia subg. Calo- sphace; Table 2; Fig. 60), with a dark violet tube of variable length, also has a relatively large volume of nectar, which might be attractive to birds. In Bolivia, large bees were observed to pollinate the species, whereas other bee species stole nectar. Species probably pollinated by birds and butterflies / long-tongued flies. In contrast to the above-men- tioned species, Salvia mohavensis Greene (Fig. 6M) and S. clevelandii (A. Gray) Greene (both Salvia sect. Audibertia; both Table 2) exclude bees due to their long and narrow corolla tubes and exserted thecae. Their medium volume of nectar with a medium to low sugar concentration corresponds to this finding. The blue to blue-violet flowers of Salvia clevelandu have a strong sweet scent that might be attractive to butterflies. Although butterflies were indeed ob- served at the flowers (Rancho Santa Ana Botanic Annals of the Missouri Botanical Garden 134 Volume 98, Number 2011 138 Annals of the Missouri Botanical Garden with pollen (Visco & Capon, 1970; Read, 1983; this study). Faegri and van der Pijl (1971) assumed S. lasiantha to be myiophilous without giving reasons. Although the species has relatively short corolla tubes, it appears omithophilous to us, with its usually deflexed or reflexed lower lips and exserted thecae (Wester & ClaBen-Bockhoff, 2007). Salvia cacaliifo- lia Benth. was listed as melittophilous by Cruden (1972), but hummingbirds were observed at the ClaBen-Bockhoff, 2007; see also Skutch, 1940). The flowers have short lower lips with front margins that are often oriented upward, making landing cumcult for insects. It is unlikely that bees are pollinators as they would hardly touch the exposed reproductive organs. Trelease (1882) discussed S. heerii Regel as being omithophilous and psychophilous. According lower lip prevent bees from landing on the flowers and from reaching the nectar, while the floral form and the large amount of nectar point to birds as a likely pollinator. He also mentioned a narrow passage beneath the connectives and white nectar guides as with this because the reduced lower lip does not offer a suitable landing platform for butterflies; we regard the species as typically omithophilous. We also do not share the opinion of Faegri and van der Pijl (1971: 179) that the flowers in S. coccinea are only “semi-omithophilous” because of their labiate shape. Correspondingly, we disagree with Reisfield (1987: 262, 279), who mentioned species with bilabiate flowers (e.g., S. julgens Cav. and S. patens) as being in a “transition phase” lithophily. Additionally, » the common view that omithophilous : characterized by stiff corollas, we also found weaker flowers to be bird pollinated, at least in species with short- to medium-sized flowers. This fits our observations that hummingbirds handle the flowers without damaging them (see also Snow & Snow, 1980). Stiff corollas are not absolutely needed to protect the flowers against visiting birds and robbing birds or bees, as stated by Proctor et al. (1996). Instead, these animals are also able to pierce strengthened corollas (e.g., S. grewiifolia S. Moore and S. leucantha, Wester & ClaBen-Bockhoff, 2007). We conclude that stiff corollas stabilize the flowers In addition to corolla shape and stability, corolla color of the omithophilous sages also differs from what has been suggested by other authors. According to Baumberger (1987), 52% of the omithophilous angiosperms have multicolored corollas (see also Porsch, 1924; Faegri & van der Pijl, 1971) (including nectar guides), but in sages the corolla is usually unicolored; it only contrasts with the calyx or leaves. The distribution of omithophilous New World sages does correspond to that of hummingbirds. In regions with a high number of omithophilous Salvia species, hummingbirds show their highest species Reisfield, 1987). Furthermore, as do bird-pollinated plants in general, omithophilous sages appear primarily at higher altitudes (Cmden, 1972; Fein- singer, 1983; Reisfield, 1987; Kay et ah, 2005). This corresponds to the equally high abundance of hummingbirds at similar altitudes (Cmden, 1972; Schuchmann, 1999; Altshuler et ah, 2004). The interdependence between bird-pollinated flowers and hummingbirds is also supported by a comparison of floral tube lengths (of omithophilous sages and omithophilous angiosperms in general) with bill lengths in hummingbirds. For example, both the range of bill lengths and tube lengths in general broadens with lower latitude (Baumberger, 1987; Schuchmann, 1999). Melittophily: Barriers in flowers as adaptation to bees. While omithophilous flowers definitely ex- clude bees, bee-pollinated flowers do not exclude birds. However, birds are more attracted by omitho- philous species because their flowers, for example, der Pijl, 1971; Stiles, 1981; Rodrfguez-Girones & 2004). The examined species of the bee-pollinated sages differ from the general picture of bee-pollinated flowers (Vogel, 1954) in nearly lacking a definite petal scent (exceptions in Salvia sections Audibertia and Echinosphace ). They share this character with many other Lamiaceae, which instead produce scent in their sepals and vegetative leaves. This scent is assumed to be used for pollination in insect- pollinated species, while a typical melittophilous scent is largely lacking (Mattem & Vogel, 1994). There are two unusual features related to floral organization worth noting in bee-pollinated species. First, more or less concealed or closed floral entrances due to connivent corolla lips appear in some species (e.g., in Salvia betonica, S. lavandu- loides ), resembling those in the tribe Antirrhineae Chav. (Plantaginaceae Juss.), for example. This blocking of the entrance protects pollen and forces the bees to handle the flower with physical force (see Westerkamp & ClaBen-Bockhoff, 2007). Also, there is often one additional staminal barrier in the flowers narrowing the entrance. As ClaBen-Bockhoff et al. Volume 98, Number 1 2011 Wester & ClaBen-Bockhoff 139 Pollination Syndromes of New World Salvia (2004a) found, this barrier may either be the sterile thecae (5. uliginosa ) of the two adjacent stamens or a ventral outgrowth of their sterile connective arms (5. rypara, Fig. 6D), illustrating that different morpho- logical structures result in the same functional constructions in the flower. Besides blocking the flower entrance, the large barrier contributes to the release of the lever mechanism as the bees encounter this barrier in their search for nectar (see also Reisfield, 1987). As the majority of the New World Salvia species belong to subgenus Calosphace, they have straight posterior lever arms that reach far behind the flower entrance. By producing ventral protrusions, a lever apparatus results that parallels the form of the curved levers, typical for many bee- pollinated species from the Old World (see also Hildebrand, 1865; Ogle, 1869; Correns, 1891; Himmelbaur & Stibal, 1932, 1933, 1934). Perhaps the bees to release the lever mechanism. In contrast, such an auxiliary construction is not needed in bird- pollinated flowers as birds are able to touch each kind of abutment with their long bills. Indications of psychophily in Salvia. The genus Salvia encompasses only a few species assumed to be psychophilous, e.g., S. scabra L. f. in South Africa (long-tongued flies assumed by Potgieter & Edwards, 2001) or S. nanchuanensis Y. Z. Sun in China (Wester, unpublished). Until now, no observations of flower-pollinator interactions exist in these species. This is also true for S. whitehousei (Fig. 6J), which we assume to be a further psychophilous species in the New World, with butterflies or psychoid insects (long- tongued flies) as pollinators. This species is only known from Texas and Coahuila (Mexico) and is little known. The upper lip does not have the typical hooded or galeate shape of most of the sages including other representatives of Salvia sect. Salviastrum (e.g., S. texana, Fig. 6B, C) and most of the Salvia species in subgenus Calosphace, and as incorrectly illustrated in the only published descrip- tion by Whitehouse (1949). Instead, the lips of S. whitehousei spread at a right angle to the long tube and thus the flower forms a Stielteller flower sensu Vogel (1954), being characteristic for psychophilous Intermediates or not clearly assignable spe- cies. We found “intermediate flowers” not clearly belonging to a specific pollination syndrome. They could not be grouped because of fragmentary or conflicting descriptions or because they share characters of several syndromes (for instance, bee- pollinated and bird-pollinated flowers). Because of evolutionary transitions, we expect that not all species can be classified into pollination syndromes. They illustrate natural variation and should not be taken as an argument against the validity of pollination syndromes. One species that provides an example of conflict- ing or missing data is Salvia purpurea, which is highly variable in color, size, and proportions. Dependent on the locality and time of observation, either hummingbirds, butterflies, or bees are reported to be more abundant (C. Gonzalez, M. Ordano & A. Luis-Martmez, pers. comm.). Different visitor groups were also found in S. clevelandii, which is reported to be often visited and pollinated by hummingbirds and also by small hawkmoths (Cox, 1981; B. O’Brien, pers. comm.). The yellow flowers of S. aspera are mentioned to be pollinated by bees (Ramamoorthy & Elliott, 1998), but it is possible that this conclusion refers to the yellow color taken as a typical melittophilous trait. Instead, bird-pollinated species also have yellow flowers. Difficulties are also caused by taxonomic conflicts. Some taxa may be recognized as synonyms or some species may include highly variable subspecies or varieties ( Salvia arenaria A. St.-Hil. ex Benth., S. amethystina Sm., S. camea Kunth, S. mexicana; see also Appendix 3). Salvia camea, for instance, includes formerly separate species that vary in flower (S. camea var. punicans (Epling) J. R. I. Wood & Harley) to melittophilous (sensu S. gracilis Benth.), forming a continuum with other taxa like a “ring of races” (Reisfield, 1987: 278). The appearance of different morphs within a species complicates their grouping. Salvia mohavensis, for instance, is distrib- uted in the southwestern United States and northern Mexico (Neisess, 1983; A. Sanders, pers. comm.). The widespread “normal” light blue Stielteller flowers combine omithophilous and psychophilous characters and are observed to be visited by hummingbirds and long-tongued flies (Fenster et al., 2004; B. O’Brien & P. Wilson, pers. comm.), with the latter classified as psychoid insects. Individuals of S. mohavensis in northern Mexico, isolated by 300 km, have flowers with a rather omithophilous form (broader corolla tube, lip orientation, color) and are reported to be visited by hummingbirds (A. Sanders, pers. comm.). If gene flow is disrupted between these different morphs, subspecies or even new species may evolve (cf. Johnson, 1997). Morphological intermediates sharing characters of two or three pollination syndromes can be interpreted in different ways. One interpretation is that the S: s, 3=fS \ StT co?« ES‘ R1 SantrSdosGSae2s ,^993, 1996. 5. ^IXr- eS“i939 fcrlhim 9^i\s'. cLpidamZh &PaZ. EUtsasss; »»swc5£ssss }SiS5S c/iionopep/ieuEplin::: ^l111..!'^’..^'*^ - >» „ 125551* (K); Reid Moran 20590 (SD); P. Wester 467, 485 (MJG, cult.). S. (MJG, cult.); Wiggins 5300* (GH, NY). S. chionophylla decumbens Alain: Liogier, 1973, 1994; Liogier 17932* 1973; Dieringer et al„ 1991; Schondube et al„ 2004; Wester (K). S. diamantina E. P. Santos & Harley: dos Santos & & ClaBen-Bockhoff, 2007; Pringle 4947 * (P); P. Wester 198, Harley, 2004. S. dichlamys Epling: Epling, 1939, 1940a; 283, 290, 302 (MJG, wild), 265 (MEXU, MJG, wild). S. Epling & Jativa, 1966; Martinez & Matuda, 1984; Hinton S*Hs3=SS= S:ills3^Sv52: 1998;"orti2-Mido a ah' 8^>(^;pd’^’[57^7 ^ (MIg! ?SSS£sS?(S;S Standi. & L. 0. Williams ex Klitg, Klitgaard, 2007. S. Epling: Epling, 1939; Hinton 5406* (NY). 5. dugesiana SS~S”£isi Sf -k %££&. Mi ^tlL ** ^ ^ *«. (C, congestifolia Epling: Epling, 1939; Epling & Toledo, 1943. elegans Vahl: Epling, 1939; Wagner, 1946; Dieringer et Epling, 1939; dos Santos, 1996. S. corrugata Vahl; Epling, MeVaugh 8297 (P); P. ' wlj? "skTwToalt.). S. SHBSSzSliSS SSB£ sSSES 1987. S. manaurica Fern. Alonso: Fernandez Alonso, 2002, Epling, 1939; Epling & Toledo, 1943; dos Santos, 2004. S. SKS3s£5SS :;=.s»ssi 1939. S- PurPurea Cav, Epling, 1939; Dieringer et^al, £5 Benth, Lagerheim, 1899; Epling, 1939; Fleming 36 (K); Hading 6827 (F); Jameson 711 * (K, NY, P). SSSSSSSi SSHSSsiSiSS: SaSHSSSs Macbride 3041 (F). 5. rhinosina Griseb.: Epling, 1939; 1941;’^% & Smith 23323* (UC-JEPS, NY, US); Wasshau- sen & Encamacion 568 (K). S. rhombifolia Ruiz & Pav.: 1991. S. richardsonii B. L. Turner: Turner, 2008c, 2008d; iBBSSSBrBB S. saccifera Urb. & Ekman: Epling, 1939; Liogier, 1994. 5. sacculus Epling: Epling, 1939. S. sagittata Ruiz & Pav, Epling, 1939; Dorst, 1956; Wood & Harley, 1989; ESSS5SH5S l<«u; Epling N Toledo. 1013: do. Sanlo. N Had,,. 2001: Arbo M. M. et al 5248 (K); Irwin et al. 25759 (K); R. L. £Sg.s£S 5S3 Harley) Fern. Alonso: Wood & Harley, 1989; Fernando 4532 * (NY), 4725* (P). S. secunda Benth, Epling, 1939, 1940a; dos Santos, 2004; Forzza et al. 2841 (K); Regnell III 932* (P); Vauthier 408* (G, L). 5. seemannii Femald: Gottsberger & G. Gottsberger 125-16471 (K); Hatschbach & Pereira 11442, 11445 (K); Hunt 6396 (K, P); S. Vogel 1956/ SiSSrBSSI setosa Fernald: Epling, 1939; Palmer 64* (K). S. setulosa Femald: Epling, 1939; Pringle 8403* (NY). S. shannonii Donn. Sm, Epling, 1939. S. sharpii Epling & Mathias: %ZL^^S^32LZ^. t pp |l£KS athacea^Gr^ne^’(^ant^^rm^^968- 444 PIG, wild), [4001 544 (MJG, cult.).’ S. speciosa C. (k speirematoides C. Wright: Epling, 1939; Wnghl 3657 * 38SSSS63S8& NY, P, U5X-U, US); Wood ,975 » 5070- (COL)^. s^TpTS^k Epling, 1939; Wood, 2007; P. Wester 35 (LPB, MJG, wild, Epling, 1939; Wright 3156* (I (K); C. 1/. T. 7WW & Barker 426* (GH, MO, NY, P). 5. Ruiz & Pavon, 1798; Epling, 1939; Macbride, 1960; Dombey Liogier, 1994; EW E-8657o* (NY). S. subglabra (Urb.) S’. s 142 (LPB, MJG, wild). S. summa A. Nelson: Walker & Elisens, 2001; P. Wester 373 (MJG, cult.). S. synodonta tehuacana Femald: Epling, 1939; Pringle 8587* (TEX- LL). S. tenella Sw.: Epling, ^1939; Liogier, 1994.JJ. Vahl: Epling, 1939; Wood & Harley, 1987; Wood, 2007; Pringle 8381*^ (TEX-LL); P. Wester 36 (LPB, MJG, wild).J. Epli Femald: Epling, sa»5s=ss Epling, 1939; Parry 760* (NY). S. univerticillata HS^SSSo SS aSSS-SsES s==s= Epling: Epling, 1939. 5. veronicifalia A. Gray ex S. £S»~2HatS sfiSs«H“£2s »55S^55St (HUT). 5. xanthotricha Harley ex E. P. Santos: dos Santos, S. zacualpanensis Briq, Epling, 1939. S. zacuapa- US)- s- Acknowledgment of Reviewers 3 98, Number 1, pp. 1-156 of I appeared in Volume 97, Number 4, p.^676 3 of Number 3 was r www.mbgpress.info CONTENTS Morphology-inferred Phylogeny and a Revision of the Genus Emmotum (Icacinaceae) Rodrigo Duno de Stefano & German Carnevali Femandez-Concha Studies in the Cleomaceae I. On the Separate Recognition of Capparaceae, Cleomaceae, and Rrassi caeeaCtJ "f* Hugh H. litis, Jocelyn C. Hall, Theodore S. Cochrane & Kenneth J. Sytsma Revision of the Infrageneric Classification of Lobelia L. (Campanulaceae: Lobelioideae) Thomas G. hammers Newly Sequenced Nuclear Gene ( Xdh ) for Inferring Angiosperm Phylc __ Cynthia M. Morton Revision of Paleotropical Megalastrum (Dryopteridaceae) Germinal Rouhan & Robbin C. Moran Pollination Syndromes of New World Salvia Species with Special Reference to Bird Pollination „ Acknowledgment of Reviewers __ Petra Wester & Regine Clafien-Bockhoff 37 63 101 156 Cover illustration, i . Howard, drawn by Bruno Manara. Volume 98 Number 2 Volume 98, Number 2 July 2011 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, in- dependent reviewers. Instructions to Authors are printed in the back of the last issue of each volume and are also available online • at www.mbgpress.info. Editorial Committee Roy E. Gereau Victoria C. Hollowell Scientific Editor, Latin Editor, Missouri Botanical Garden Missouri Botanical Garden Ihsan A. Al-Shehbaz Beth Parada Missouri Botanical Garden Managing Editor, Gerrit Davidse Missouri Botanical Garden Missouri Botanical Garden Allison M. Brock Peter Goldblatt Associate Editor, Missouri Botanical Garden Missouri Botanical Garden Gordon McPherson Tammy Charron Missouri Botanical Garden Associate Editor, Charlotte Taylor Missouri Botanical Garden Missouri Botanical Garden Cirri Moran Henk van der Werff Press Coordinator, Missouri Botanical Garden Missouri Botanical Garden of the Missouri Botanical Garden, % Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2011 is $180 per volume U.S., $190 Canada & Mexico, $215 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www.mbgpress.info The Annals of the Missouri Botanical Garden (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. Postmaster: Send address changes to Annals of the Missouri Botanical Garden, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Abstract/Global Health databases, Ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. The full-text of Annals of the Missouri Botanical Garden is available online though BioOne™ (http:// www.bioone.org). © Missouri Botanical Garden Press 2011 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Volume 98 Annals m Number 2 of the 2011 Missouri Botanical Garden A TAXONOMIC REVISION OF Sven Buerki, 2 Peter B. Phillipson,™ and Martin GOUANIA (RHAMNACEAE) IN ^ CallmandeP 5 MADAGASCAR AND THE OTHER ISLANDS OF THE WESTERN INDIAN OCEAN (THE COMORO AND MASCARENE ISLANDS, AND THE SEYCHELLES)1 Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 159 Morphology of Gouania The morphological characters that have been used in previous treatments of Gouania (notably Perrier de la Bathie, 1943) have formed a firm foundation for our work. However, coupled with these are new observations made in the herbarium and in the field. Patterns of morphological variation and ecogeo- graphic distribution were compared in order to circumscribe what we believe are biologically coherent taxa. A detailed discussion of the important morphological characters is given below. A key to the species allowing the or fruiting specimens is LEAVES AND STIPULES Perrier de la Bathie (1943, 1950) placed consid- erable emphasis on leaf morphology. We have investigated this in greater detail and conclude that variation in leaf shape, texture, color (in dried specimens), margin type, the abaxial surface indu- ment, and type of venation are particularly important. Most species of Gouania have ovate leaves, but exceptions include: G. callmanderi Buerki (elliptic to ovate-elliptic), G. humbertii H. Perrier (orbicular), G. mauritiana (lanceolate to deltoid), and G. laxiflora Tul. (cordate). The majority of the species have Phillipson & Callm. has distinctly thicker, more rigid coriaceous blades, while G. zebrifolia Buerki, Phil- lipson & Callm. has thinner, more flexible membra- nous blades. Leaves of most species are more or less concolorous when dried but are notably discolorous in others (e.g., G. phillipsonii Buerki). Leaf margins are generally entire, but some taxa have denticulate (' G . ambrensis Buerki, Phillipson & Callm.) or serrate (G. mauritiana) margins. The abaxial surface of immature leaves usually possesses some kind of indument, but this is often caducous, giving rise to glabrous mature leaves. In other species the indu- ment is persistent and sometimes very dense, variously distributed and composed of trichomes that differ in color, length, and orientation. The presence of strong tertiary veins arising from the secondary veins is also a valuable taxonomic character. Most of the species have at least one well-developed tertiary vein arising from the lowest pair of secondary veins (in addition to the fine reticulate network of tertiary veins that is present throughout the leaf), whereas only two species ( G . humbertii and G. lineata Tul.) completely and consistently lack conspicuous tertiary veins. Among all the taxa present in the region, G. taolagnarensis Buerki, Phillipson & Callm. has particularly well-developed venation, with even some conspicuous quaternary veins present arising from the first pair of tertiary veins. In this species the venation is more developed on one side of the leaf, causing the leaf base to be asymmetric. The reticulation is noticeably scalariform in most species; however, this pattern is inconspicuous in some taxa (e.g., G. pannigera Tul.). A dark brown triangular marking at the base of the stipule has been noted only in G. ambrensis. FLOWERS AND INFLORESCENCE Perrier de la Bathie (1943) made extensive use of the shape of the hypanthium and the form of the nectary disc in delimiting taxon. Our studies confirm the importance of these characters and have revealed other variations in these organs of which Perrier de la Bathie was apparently unaware, that we believe are also taxonomically useful. Most species have an obconic hypanthium, but Gouania cupuliflora Buerki, Phillipson & Callm., G. gautieri Buerki, Phillipson & Callm., and G. humbertii have a cupulate hypanthi- um. The nectar disc is always lobed, but the relative length of the lobes is highly variable. At one extreme, short lobes occur in G. aphrodes Tul. that reach only about one sixth of the length of the sepals, and at the other extreme the lobes of G. laxiflora are as long as or longer than the sepals. Moreover, the lobes of the disc within a single flower are markedly unequal in length in G. scandens (Gaertn.) R. B. Drumm. subsp. glandulosa (Boivin ex Tul.) Buerki, Phillipson & Callm. All species appear to have 3-lobed stigmas except those of G. cupuliflora, which are 2-lobed. The flowers may be sessile or pedicellate and grouped in indument of the peduncles and pedicels is variable. INFRUCTESCENCE AND FRUITS In this study, several characters based on the infructescence were used to circumscribe taxa, notably the distribution of the fruits that develop to maturity, the type and color of the fruit indument, and the immature fruit shape. In most species, the fruits are equally distributed along the infructescence (e.g., Gouania ambrensis, G. gautieri), whereas in G. perrieri Buerki, Phillipson & Callm. they are often aggregated in the proximal part and in G. laxiflora they are concentrated in the distal part. Furthermore, G. laxiflora is the only species having a pyriform Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 163 Figure 2. Distribution of Gouania species occurring in Madagascar, mapped on the bioclimatic zones (after Comet, 1974; Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 165 Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 167 tertiary veins. — E. Fruit. A, D drawn from the holotype Buerki et al. 61 (MO); B, C from Buerki et al. 55 (TAN); E from Bosser 28 (TAN). 170 Annals of the Missouri Botanical Garden 172 Annals of the Missouri Botanical Garden 174 Annals of the Missouri Botanical Garden (Paris) 11: 30. 1943, nom. nud. Gouania leguatii J. Gueho, Adansonia 18(4): 483. 1979, syn. nov. TYPE: Rodrigues Island. Cascade Victoire, Feb. 1941 (11.), R. Jauffret 129 (holotype, MAU not Kew 1916: 199. 1916, syn. nov. TYPE: Mozambique, Chupanga, 1860 (fr), Kirk s.n. (holotype, K!). Woody liana climbing to 4 m; stems dark green, drying brown, glabrescent, young shoots pubescent with ferruginous trichomes; stipules glabrous, 1(— 1.5) X ca. 0.25 mm; tendrils pubescent with caducous ferruginous trichomes. Leaves ovate with a cordate or rounded base; petioles (15-)25(— 35) mm, sparsely pubescent with caducous ferruginous tri- Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 177 chomes; leaf blades (4.5-)6(— 7.5) X (3-)4.5(— 5.5) cm; secondary veins alternate, 5 to 8 pairs, not reaching the margin but following it; tertiary i apparent, arising near the base of secondary veins, reticulation scalariform; abaxial surface glabrescent, veins puberulous with brown trichomes; midrib, secondary veins, and tertiary veins prominent; adaxial surface glabrous; margin entire to slightly denticulate; base cordate or sometimes rounded; apex acute. Inflorescences spindly, lax, puberulous with pale brown trichomes, (5.5-)10-13(-15) cm, consisting of glomerules subtended by stipulelike bracts, proximal hermaph- rodite flowers developing before the distal male flowers; glomerules pedunculate, puberulent with caducous, pale brown trichomes, accrescent, (l-)3 mm, 3- to 5-flowered; bracts small, glabrous on both surfaces, caducous, (0.5-)0.8 X ca. 0.2 mm; flowers pedicellate, pedicels accrescent, (l-)3.5 mm, pu- berulent with caducous, pale brown trichomes. Hypanthium obconic, becoming subglobose as the capsule develops; sepals ca. 2 X 0.75 mm, pale yellow to brown, outer surface puberulent, inner glabrous; petals ca. 1 X 0.5 mm, white; stamens slightly longer than the petal blades, with a pale 0.2 X 0.2 mm; disc flat, ca. 4 mm diam.; lobes 2-2.2 mm, rounded, generally longer or as long as the sepals; stigma with 3 linear lobes, those ca. 0.2 mm, style ca. 2 mm. Fruits 0 to 2 developing on each glomerule, spheroid (ca. 1.5 X 1.5 cm), glabrous, concentrated in the basal part of the infructescence; valves ca. 15 X 4 mm; immature fruits pyriform, glabrous; seeds brown, shiny, and slightly flattened on the inner surfaces, ca. 4 X 3 X 1 mm. Distribution and ecology. Gouania laxiflora is found in mainland Africa (Mozambique and Tanza- nia) (Drummond, 1966; Johnston, 1972: under G. scandens, see discussion below), the Mascarenes (Rodrigues Island), the Seychelles (Aldabra, lie Picard, Cosmoledo), the Comoro Islands (Mayotte), and western Madagascar (Fig. 2D). The species grows in semi-deciduous, deciduous, gallery, and secondary forests, generally on limestone, but sometimes on sand and gneiss at an elevational range from sea level to 350 m. IUCN Red List category. Within Gouania laxiflora has an EOO of 430,841 km2, an AOO of 360 km2, and 33 subpopulations, 13 of which occur within protected areas (Andohahela, Ankarana, Bemaraha, Beza Mahafaly, Namoroka, Tampoketsa d’Analamaitso, Zombitsy). It is thus assigned a preliminary status of Least Concern (LC) in Mada- and May. Discussion. Tulasne (1857) designated two syn- types of Gouania laxiflora, Boivin 3366 and Bernier 207. The sheet labeled Boivin 3366 at P is a collection of this species, while sheets with the same collection number at G and K bear material of G. aphrodes. To avoid possible confusion, we have lectotypified this species on Bernier 207, even though material of this collection is only known to be present at P. After careful study, we are convinced that this plant is conspecific with Gouania leguatii described from Rodrigues Island (Gueho, 1997). Material referred to this species has similar cordate leaves, disc lobes generally as long as or longer than the sepals, and the fruits glabrous and generally concentrated in the basal part of the infructescence. The same is true for the specimens cited in Flora Zambesiaca (Drummond, 1966) and in Flora of Tropical East Africa (Johnston, 1972) under G. scandens from the African mainland (see also discussion under G. scandens ). Gouania longipetala Hemsl. was described for an African species, based on two syntypes from Equatorial Guinea or Gabon ( Mann 17 from Bioko Island “Fernando Po” and Mann 1813 from Muni River “Kongui River”) and ley, 1868). Green (1916) realized that this material represents two different species and described G. mozambicensis to accommodate the East African entity based on the Kirk specimen, while retaining the name G. longipetala for the Central African species. Later Halle (1962) designated Mann 1 7 from Bioko as the “Type,” thus effecting lectotypification of G. longipetala. Kirk’s specimen from Mozambique is a specimen of G. laxiflora, and we include G. Gouania laxiflora is the only species shared between mainland Africa and Madagascar. Gouania laxiflora differs from G. perrieri, to which it is most similar, by its slightly denticulate leaf margin (vs. shallowly crenate margin in G. perrieri), lobes of the disc as long as or longer than the sepals (vs. lobes of the disc reaching one fourth of the B'S-? Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 181 pannigerd), its disc lobes reaching one fourth of the length of the sepals (vs. lobes of the disc reaching one sixth of the sepals in G. aphrodes, G. myriocarpa, and G. pannigera ). Furthermore, G. mauritiana appears to be endemic to Reunion Island, whereas G. aphrodes occurs in Madagascar and the Comoro Islands, and G. pannigera and G. myriocarpa are endemic to Madagascar. Additional specimens examined. REUNION ISLAND, s. loc., Barclay s.n. (K); £1. & imm. fr., Boivin 1383 (P); fl„ Boivin s.n. (K); fr., Commerson s.n. (K, P); 11., De L’Isle 605 (P); fl., Du Petit Thouars s.n. (P); fl., Herb. Colonial ministere de la marine s.n. (P); fl., Hilsenberg s.n. (K); fl., Dattier, Boivin 1383/2 (G, K, P); Cilaos, Cadet 724 (P), Cadet 1831 (P). 11. Gouania myriocarpa Tul., Ann. Sci. Nat., Bot. ser. 4, 8: 132. 1857, as “Guania .” Gouania mauritiana Lam. subsp. myriocarpa (Tul.) H. Perrier, Notul. Syst. (Paris) 11: 33. 1943. TYPE: Madagascar. Toamasina: Ambanivoules, cote orientale, s.d. (fr.), J. P. Goudot s.n. (holotype, G!; isotype, G!). Figure 8C-E. Woody liana climbing to 10 m; stems dark green, caducous ferruginous trichomes; stipules glabrous, 3(— 4) X ca. 0.75 mm; tendrils pubescent. Leaves ovate; petioles (5-)10-15(-20) mm, pubescent with ferrugi- nous trichomes; leaf blades (4.5— )5.5-6(— 7) X (2.5— )3(— 4) cm; secondary veins alternate, 4 to 5 pairs, not reaching the margin but following it; tertiary veins conspicuous, arising near the base of the lowest secondary veins, reticulation scalaiiform; abaxial sur- face glabrescent, pubescent with ferruginous trichomes on the veins; midrib, secondary veins, and tertiary veins prominent; adaxial surface glabrous; margin shallowly crenulate; base rounded, sometimes shallowly cordate; apex acute. Inflorescences congested, pubescent with hirsute ferruginous trichomes, (8-)12(— 25) cm, consist- ing of glomerules subtended by stipulelike bracts; glomerules sessile or shortly pedunculate, (2 to)4- flowered; bracts scarious, small, glabrous on both surfaces, ± persistent, (l-)2 X ca. 0.5 mm; flowers with a spindly pedicel (1-)1.5(— 2) mm, covered by a whitish indument. Hypanthium obconic, becoming subglobose as the capsule develops; sepals ca. 1.5 X 0.5 mm, reddish, outer surface pubescent to glabrous, inner glabrous; petals ca. 0.75 X 0.25 mm, reddish; stamens slightly longer than the petal blades, with a pale red filament and a 0.2 mm; disc flat, ca. 2 mm diam.; 1/6 as long as the se- pals; stigma with 3 linear lobes, those ca. 0.8 mm; style ca. 1 mm. Fruits 1 or 2 developing on each glomerule, small, spheroid (ca. 0.6 X 0.6 cm), glabrous, densely arranged throughout the infructescence; valves ca. 6 X 4 mm; seeds brown, shiny, ovoid, ca. 2 X 1 X 0.3 mm. Distribution and ecology. Gouania myriocarpa occurs in eastern Madagascar (Fig. 2D) in evergreen montane rainforest and secondary forests on basaltic rocks, gneiss, and laterite soils at an elevational range of (500-)800-l 00(— 1400) m. IUCN Red List category. With an EOO of 164,280 km2, an AOO of 216 km2, and 17 subpopulations, seven of which occur within protected areas (Andoha- hela, Anjanaharibe-Sud, Analamazoatra-Perinet, Maro- jejy, Masoala, Montagne d’Ambre, Ranomafana), Gouania myriocarpa is thus assigned a preliminary status of Least Concern (LC), according to IUCN Red List criteria (IUCN, 2001). Phenology. This species is known in flower from November to February and in fruit from March to July. Discussion. Tulasne described Gouania myrio- carpa together with six other species of Gouania in his treatment of the Rhamnaceae for Florae Mada- gascariensis (Tulasne, 1857). Later, adopting broad species concepts, Perrier de la Bathie (1943) reduced and G. pannigera ) to subspecies of G. mauritiana. We agree with Tulasne’s narrower delimitation and these taxa are again recognized at the species level. Furthermore, Perrier de la Bathie did not examine “Ambanivola, Goudot (in Herb. Delessert, non vu)” (Perrier de la Bathie, 1950: 46), and, as a result, he completely misinterpreted this taxon. In his treat- ments for G. mauritiana subsp. myriocarpa (Perrier de la Bathie, 1943: 33; 1950: 46), he included specimens that we now refer to G. phillipsonii, which we describe as new below. On the other hand, he referred specimens of true G. myriocarpa to his new variety, G. glandulosa var. breviloba (Perrier de la Bathie, 1943: 32; 1950: 43), a name that was not validly published. Gouania myriocarpa is morphologically most similar to G. lineata, both having glabrous fruits congested along their entire infructescences (see G. lineata for further comparison). Gouania myriocarpa can, however, be recognized easily among the Malagasy species by the presence of the ferruginous trichomes on the petioles, along the veins of the abaxial leaf surface, and the inflorescence. The lobes Volume 98, Number 2 2011 Volume 98, Number 2 2011 Buerki et al. Revision of Gouania (Rhamnaceae) 187 5344 (MO). y g P ) (T yp > sepals) and a relatively dense indument on their leaves. However, despite the diagnostic characters listed above, these species are not sympatric. Gouania phillipsonii occurs in forest mainly at mid- elevations on the eastern escarpment of Madagascar, whereas G. aphrodes occurs in the north at lower elevations, G. pannigera is known from the central region (often in degraded areas), and G. taolagnar- ensis is narrowly endemic to the region of Taolagnaro (Fort Dauphin) in the southeast. 192 Annals of the Missouri Botanical Garden Phillipson 64 (TAN). q ^ P J P g ment that covers the inflorescence and the flowers, immature fruits with a pale yellow indument that reddens with maturity, fruits with two or three transverse lines on the center of the wings, and the fruits concentrated at the distal part of the infructes- Gouania taolagnarensis is morphologically most similar to G. pannigera in the length of the lobes of the disc and the size and pubescence of the fruits, but differs in having ovate to cordate leaves, the margin of PHYLOGENETIC INFLUENCE ON Robyn J. Burnham2 and Caissa TWINING CHIRALITY IN LIANAS FROM AMAZONIAN PERU1 X X Volume 98, Number 2 2011 Burnham & Revilla-Minaya Twiner Chirality at Los Amigos, Peru Qmduff, 2004 Darwin, 1876; Hegarty & Caballe, 1991; Putz & Holbrook, 1991; Omduff, 2004 Baillaud, 1962; Putz, 1984; Ornduff, 2004 Baillaud, 1962; RJB, Gerrath et al., 2008 Ribeiro et al., 1999 Putz & Holbrook, 1991 Hegarty & Caballe, 1991 2005; RJB, pers. obs. Acevedo-Rodrfguez, 2005 Acevedo-Rodrfguez, 2005 Acevedo-Rodrfguez, 2005 Hegarty & Caballe, 1991; DeWalt et al., 2000; RJB, pers. obs. Croat, 1978; Hegarty & Caballe, 1991 Hegarty & Caballe, 1991 Acevedo-Rodrfguez, 2005 Pinard & Putz, 1994; thorough review of botanical perspectives on hand- edness terminology up to the early 1970s. We present here an exhaustive survey of a local climber flora in which handedness of apically twining species is recorded. In contrast to the globally oriented work of Edwards et al. (2007) in which they traversed a trail, counting the first 17 to 121 climbing stems (average of 81) at each of 17 sites (one of which was the Los Amigos site addressed here), we studied a geographically limited climber flora, including as many distinct taxa as possible. This difference be L ween the Lwo studies corrects for the probability of including large numbers of individuals of the same taxon when only the first-encountered, unidentified stems are evaluated, thus decreasing the phylogenetic breadth of the census. Our study addresses climbing plants exclusively from the Neotropics, and thus is also a departure from the excellent compendium of Omduff (2004) in which plants observed were largely under cultivation in botanical gardens and in which Asia, and Australia. Using our survey, we tested the hypothesis that distinct taxa within a flora are expected to twine 200 Annals of the Missouri Botanical Garden 202 Annals of the Missouri Botanical Garden censused within 20 m of another individual of the same taxon to avoid censusing clonal individuals. One to eight individuals w Mikania Willd. sericea Leonard amblybasis S. F. Blake acouci (Aubl.) A. DC. verrucosa (Willd. ex Roem. & Schult.) K. Schum. ex Markgr. sp. indet. 1 guTco Humb. & Bonpl. sp. indet. 1 ichybotrya Mull, i & Moldenke fagifolia (DC.) A. Juss. RJB 3433 RJB 3669 RJB 3713, RJB 3596 RJB 3611 RJB 3568, 3661 RJB 3491 RJB 3584 RJB 3542, RJB 3664 RJB 3587 RJB 3561 RJB 3670 RJB 3586 RJB 3738 RJB 3490, 2 RJB 3488 RJB 3583 RJB 3520 RJB 3688 RJB 3566 RJB 3715 RJB 3710, 3567 DEVELOPMENTAL ANATOMY AND MORPHOLOGY OF THE FLOWERS AND FRUITS OF SPECIES FROM GALIUM AND RELBUNIUM (RUBIEAE, RUBIACEAE)1 Karen L. G. De Toni2 , Volume 98, Number 2 2011 207 De Toni & Mariath Anatomy and Morphology of Galium and Relbunium subsections, and series. According to this author, the delimitations of the genera Relbunium and Rubia principally result from their respective geographic distributions, as well as differences in their growth forms, corolla, and fruit. Ehrendorfer (1955) suggest- ed that the closer morphological relationships between the American species of Galium and Relbunium indicated that Relbunium originated from Galium in the New World. Dempster (1978) viewed Relbunium as a clearly monophyletic group, even though its species are often confused with those of Galium. This author further affirmed that the species of Relbunium are easily distinguished when flowering or fruiting, observing that involucres are absent in Galium, whereas species in Relbunium exhibit four involucral bracts subtending each flower. However, according to Dempster (1978), vegetative r to both genera, supporting Relbunium a Galium, as first proposed by Endlicher (1839). In 1982, Dempster mentioned that the genus Relbunium should be limited to taxa with solitary and . Based on author transferred all of the species without these characteristics into the genus Galium. In her last work with South American Galium, Dempster (1990) returned to the proposals of Endlicher (1839) and defined Relbunium as a section of Galium. Plastid DNA sequence analyses of species of Rubieae (Natali et al., 1996) included Relbunium as R. hypocarpium (L.) Hemsl., but did not further address taxonomic issues. Bremer and Manen (2000) included 14 genera in their classification for the Rubieae and treated Relbunium separately from Galium. Galium comprises approximately 400 species (Judd et al., 1999), of which ca. 90 are found in the Neotropics (Andersson, 1992). The genus Relbunium comprises ca. 35 Neotropical species that are distributed from the southwestern United States (Porto et al., 1977). These species preferentially grow in open fields and forest borders in subtropical regions. Approximately 22 species are known for the southern regions of South America, including southern Brazil, Uruguay, and Argentina (Chamisso & Schlechtendal, 1828; de Candolle, 1839; Schu- mann, 1891; Ehrendorfer, 1955; Smith & Down, 1956; Rambo, 1962; Porto et al., 1977). Approxi- mately 39% of Relbunium species, representing over half of the described South American species of Relbunium, are included in the morphological survey Fruits and seeds are always important in taxonomic classifications of Rubiaceae, as can be seen in Schumann (1891), Bremekamp (1966), and Rob- brecht (1988). In addition to characterizing subfam- ilies and tribes, fruits and seeds have been used to distinguish genera, as in the case of Appunia Hook. f. and Morinda U. (Hayden & Dwyer, 1969). These infrafamilial morphological variations indicate a diversity of adaptations, especially in terms of seed dispersal (Bremer & Eriksson, 1992). Even though previous studies have variously treated Galium and Relbunium, there has been insufficient work focusing on their taxonomy that could integrate the anatomical, morphological, ecological, and molecular evidences. Few works describe the morphology and anatomy of the flowers, fruits, and seeds of these taxa in terms of their phylogenetic significance. According to Corner (1976), there is a need for extensive ontogenetic investigations of the pericarp, integument thickness, and the structure of the seed coat at the genus and species levels in the Rubiaceae. Consequently, the present study examines the structure and develop- ment of the flowers, fruits, and seeds within exemplar species of Galium and Relbunium to elucidate additional taxonomic characteristics toward the delimitation and relative ranks of these two groups/ We examined 13 species of Relbunium (sensu Ehrendorfer, 1955, modified by Porto et al., 1977) and two species of Galium, as well as two other species within the Rubioideae as outgroups. For examination of species within Relbunium, exemplars were chosen to represent all sections and subsec- tions recognized by Ehrendorfer (1955). All botan- ical material was deposited in the herbarium of the Rio de Janeiro Botanical Garden (RB), as shown in Table 1. All material destined for optical microscopic analysis was fixed immediately after collection in 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.2 (Gabriel, 1982), dehydrated in an ethanol series, embedded in Leica HistoResin (2- hydroxyethyl-metracrylate) (Nussloch, Germany), sectioned in a Shandon Hypercut microtome (Run- corn, Cheshire, U.K.) at 2-4 pm, and then stained with toluidine blue 0 0.05% (O’Brien & McCully, 1981). Observations (in bright field) were made using an Olympus BX50 optical microscope (Center Valley, Pennsylvania, U.S.A.) equipped with a CoolSnap-pro digital camera (Houston, Texas, U.S.A.). Histochem- ical analyses were performed using Sudan III to 208 Annals of the Missouri Botanical Garden Galium latoramosum Clos Relbunium equisetoides (Cham. & Schltdl.) R. humile (Cham. & Schltdl.) K. Schum. ,. Porto & Ehrend. m (Spreng.) Ehrend. i Ehrend. Brazil. Rio Grande do Sul: Ijuf, 24 Oct. 1984, R. Bueno s.n. (RB 444232); 30 Aug. 1904, A. Bommuller 132 (GH). Santa Catarina: Nova Teutonia, 3 Sep. 1944, F. Plaumann 120 (RB). Brazil. Rio Grande do Sul: Cayapava do Sul, 3 May 2002, K. De Toni s.n. (RB 444168); Cristal do Sul, 20 Aug. 2001, K. De Toni s.n. (RB 444206); Guaiba, 11 Jan. 2003, K. De Toni s.n. (RB 444228). Brazil. Rio Grande do Sul: Guaiba, 20 Feb. 2003, N. I. Matzembacher s.n. (RB 444212), K. De Toni et al. s.n. (RB 444210). (RB); Itatiaia, Mar. 2003, K. De Toni et al. s.n. (RB 4444211). Brazil. Rio Grande do Sul: Lavras do Sul, 4 May 2002, K. De Toni et al. s.n. (RB 444164). Brazil. Rio Grande do Sul: Cambara do Sul, 9 Mar. 2002, K. De Toni et al. s.n. (RB 444192); 7 Apr. 2002, K. De Toni et al. s.n. (RB 444173); 26 Dec. 2002, K. De Toni et al. s.n. (RB 444216); Guaiba, 11 Jan. 2003, K. De Toni et al. s.n. (RB 444227); Jaquirana, 6 Apr. 2002, K. De Toni et al. s.n. (RB 444177); Maquine, 26 Dec. 2002, K. De Toni et al. s.n. (RB 444217); Porto Alegre, 4 Oct. 2001, K. De Toni et al. s.n. (RB 444200); Tainhas, 29 Nov. 2001, K. De Toni et al. s.n. (RB 444197). Brazil. Rio Grande do Sul: Cambara do Sul, 9 Mar. 2002, K. De Toni et al. s.n. (RB 444188); 7 Apr. 2002, K. De Toni et al. s.n. (RB 444229). Brazil. Rio Grande do Sul: Barra do Ouro, 26 Dec. 2002, K. De Toni et al. s.n. (RB 444213); Cayapava do Sul, 3 June 2002, K. De Toni et al. s.n. (RB 444165, 444166); Cambara do Sul, 9 Mar. 2002, K. De Toni et al. s.n. (RB 444171); 29 Mar. 2002, K. De Toni et al. s.n. (RB 444178); 7 Apr. 2002, K. De Toni et al. s.n. (RB 444180); Porto Alegre, 6 Aug. 2001, K. De Toni et al. s.n. (RB 444207); Santo Antonio da Patrulha, 22 Aug. 2001, K. De Toni et al. s.n. (RB 444203); Sao Jose dos Ausentes, 10 Mar. 2002, K. De Toni et al. s.n. (RB 444184, 444185). Brazil. Rio Grande do Sul: Cambara do Sul, 9 Mar. 2002, K. De Toni et al. s.n. (RB 44189); 29 Mar. 2002, K. De Toni et al. s.n. (RB 444179); 7 Apr. 2002, K. De Toni et al. s.n. (RB 444170); 26 Dec. 2002, K. De Toni et al. s.n. (RB 444215); 5 Jan. 2004, K. De Toni et al. s.n. (RB 444160); Jaquirana, 6 Apr. 2002, K. De Toni et al. s.n. (RB 444175). Brazil. Rio Grande do Sul: Cambara do Sul, 29 Mar. 2002, K. De Toni et al. s.n. (RB 444181). Brazil. Rio Grande do Sul: Tainhas, 29 Mar. 2002, K. De Toni et al. s.n. (RB 444182); 29 Nov. 2002, K. De Toni et al. s.n. (RB 444196). Brazil. Rio Grande do Sul: Burro Novo, 6 Sep. 2004, K. De Toni et al. s.n. (RB 444230); Cayapava do Sul, 3 June 2002, K. De Toni et al. s.n. (RB 444167); Cambara do Sul, 9 Mar. 2002, K. De Toni et al. s.n. (RB 444187); 7 Apr. 2002, K. De Toni et al. s.n. (RB 444172); Jaquirana, 6 Apr. 2002, K. De Toni et al. s.n. (RB 444176); Sao Francisco de Paula, 29 Nov. 2001, K. De Toni et al. s.n. (RB 444198); 26 Dec. 2002, K. De Toni et al. s.n. (RB 444158); Sao Jose dos Ausentes, 10 Mar. 2002, K. De Toni et al. s.n. (RB 444186); Tainhas, 29 Nov. 2001, K. De Toni et al. s.n. (RB 444194). Brazil. Rio Grande do Sul: Cayapava do Sul, 4 June 2002, K. De Toni et al. s.n. (RB 444224). Brazil. Rio Grande do Sul: Cristal do Sul, 20 Aug. 2001, K. De Toni et al. s.n. (RB 444205); Lavras do Sul, 4 June 2002, K. De Toni et al. s.n. (RB 444163); Porto Alegre, 2 Aug. 2001, K. De Toni et al. s.n. (RB 444221); 16 Aug. 2001, K. De Toni et al. s.n. (RB 444222); Santo Antonio da Patrulha, 22 Aug. 2001, K. De Toni et al. s.n. (RB 444203); Sao Francisco de Paula, 9 Mar. 2002, K. De Toni et al. s.n. (RB 444193). Brazil. Rio Grande do Sul: Santo Antonio da Patrulha, 22 Aug. 2001, K. De Toni et al. s.n. (RB 444195). Brazil. Rio Grande do Sul: Porto Alegre, K. De Toni et al. 05 (RB). Brazil. Rio Grande do Sul: Porto Alegre, K. De Toni et al. 06 (RB). 210 Annals of the Missouri Botanical Garden mature ovule is anatropous, with a single (Figs. 2A, 3B, D, E). Flowers and fruits have a reduced pedicel, which is similar to a peduncle, but pedicel demonstrated varying lengths among the different species, being elongated (0. 7-1.1 mm) in R. humile (Cham. & Schltdl.) K. Schum. (Fig. 3F) and R. mazocarpum Greenm. (Fig. 3G). This structure was not observed for the two species of Galium examined, as their flowers did not have bracts (Fig. IN, 0). The ovary in Relbunium is composed of carpel mesophyll, covered by an epidermis, with one side facing toward the exterior and the other facing the carpel locule. The external epidermis may have secretory idioblasts in addition to pavement cells (Figs. 2A, 3A), as well as unicellular trichomes (Fig. 3B, F, H). The external periclinal walls are covered with a cuticle that is diversely ornamented. In five species, rounded epidermal cells were observed in the external epidermis of the ovary (Fig. 3D; Appendix 2, Volume 98, Number 2 2011 211 De Toni & Mariath Anatomy and Morphology of Galium and Relbunium <*11- mill im* <|ii il u ill ill u I !i ~ — t 1 1 1 1 *\\ I 1 /,' hut urn P<1II<11), as originating from Aluka (Aluka, 2009). Twenty African herbaria participated in this initiative and have contributed images of 51,822 specimens, which is ca. 18% of the total number of African specimen images accessible on the Internet via the JSTOR Plant Science portal (). Over 70% of the images for African taxa come from South African herbaria. Projects are now underway to increase the participa- tion of herbaria from other African countries, with a small grants program supported by the Andrew W. this. Images and associated metadata generated by participants in the API remain the property of nated through JSTOR Plant Science, an online environment, which is part of ITHAKA, a not-for- profit organization that aims to provide online access to an archive of international scholarly resources (). The API, however, did not intend to stop at scanning and disseminating images of type speci- mens. What started out as a purely taxonomic project soon emerged as a major initiative that can also be used to provide other, related information on plant diversity through the Internet. The preparation of electronic images of botanical artifacts is therefore only the first step in a process that has now developed into an indispensable electronic resource on African plants. The success of the API is a clear manifes- tation of the willingness of African herbaria to collaborate with their northern and southern col- leagues. This drive is facilitating the sharing of plant diversity information and the rapid and affordable dissemination thereof. By expediting this, African herbaria will enable their taxonomists, who are confronted by an immensely rich flora of over 50,000 species (Klopper et al., 2006), to overcome the digital divide and misconceptions about their ability to contribute in a meaningful way through embracing new technologies. A further significant contribution of the API is that broader environmental issues that have become the subject of dialogue between African and northern herbaria can now be addressed. In this regard, it and 4) build the competitive- ness of African countries and the continent.” Clearly, Africa is ready to assume responsibility for its own future and to combine its strengths in natural resources studies and to draw on its own innate abilities to manage these resources wisely for the benefit of all its peoples, indeed, for the benefit of the global village. It is very encouraging that plant diversity sciences have attracted the support of the Andrew W. Mellon Foundation as a supportive partner in helping Africa to tackle some of the impediments that resulted from decades of exploitation, civil wars, corruption, and political maladministration. These impediments seri- ously impact the ability of many African countries that are signatories to the Convention on Biological Diversity (CBD) to respond to challenges presented to As Africa takes another firm step toward develop- ing a continent that can take its rightful place in the global arena, the API clearly acts as a catalyst to drive related scientific ; the API, African plant es, and infrastructure o with most, if nc and dimensions of the API, progress is made in a stepwise manner. These steps, however, should be taken with the goal in mind to sustainably take ownership of the initiative and provide African taxonomists with the tools and knowledge to contribute toward, to manage, and to disseminate appropriate plant diversity information on the African flora. This sharing of information among participating countries, and ultimately with the broader botanical community, already contributes to a better under- standing of the continent’s significant botanical on the plant diversity of Angola (Figueiredo & Smith, 2008; Figueiredo et al., 2009; Smith & Figueiredo, 2010) and the continent’s lycophyte and fern flora (Roux, 2009). Such improved knowledge should further stimulate renewed interest in the documenta- tion of the continent’s botanical wealth that is imperiled by human pressures. This essential and primary knowledge is vital if the natural resources of 274 Annals of the Missouri Botanical Garden Table 1. African Plant Initiative r Table 1. Continued. ollow Thiers, 2009) arranged according to the number >f specimens imaged and uploaded to Aluka (2009), through API, innovation with access to significant resources undreamed of in the past and impossible to visualize before now. Building on such a strong foundation, African plant scientists now have an opportunity not only to benefit from the advantages of global scientific integration, but indeed to be equal partners in providing access to previously untapped environmental resources for their own benefit. As with all the admirable ing peer review process for African states, it is also imperative that African taxonomists collaborate on the API, in equal partnership with their northern Multilateral cooperation on botanical matters is now possible at both a regional and global scale. In this regard, as government budgets constrict as a result of social demands, private and corporate investment in biodiversity science becomes increas- ingly important. Additional initiatives aimed at further improving conditions conducive to joint i and South-North scientific cooperation uraged and supported at all levels. To development is of paramount human capital development was one of the primary aims of the Southern African Botanical Diversity EMBRYOLOGICAL EVIDENCE SUPPORTS THE TRANSFER OF LEITNERIA FLORIDANA TO THE FAMILY SIMAROUBACEAE1 278 Annals of the Missouri Botanical Garden which have often provided evidence of relationships among taxa ranging from species to families (Tobe, 1989). About 100 years ago, some of the embryo- logical characters of Leitneria were described by Pfeiffer (1912) for comparison with members of the Amentiferae; plants used for this study were cultivated at the Missouri Botanical Garden (MO). However, most of the characters were not docu- mented with the precision expected in modem works, and many other aspects of embryology are as yet undescribed. We now need exact information on the largest suite of embryological characters for critical comparison with other members of the Simaroubaceae. In this study, I confirm or correct Pfeiffer’s (1912) interpretations and document almost all of the embryological features of L. floridana. These are compared with characters of members of the Simaroubaceae, Meliaceae, and Rutaceae. I show the embryological relatedness of Leitneria within the Simaroubaceae, particularly with the genus Brucea and its relatives. Materials and Methods Staminate flower buds of Leitneria floridana at various stages of development were collected from trees cultivated at the Missouri Botanical Garden in 2000, 2001, and 2007 (voucher: U.S.A., Missouri, St. Louis City, Missouri Botanical Garden, cultivated, perhaps originated in Butler Co., MO, W. G. D’Arcy 3653 [MO 2426777]), and pistillate flower buds and fruits at various stages of development from trees in Butler County, Missouri, in 2000 and 2001 (voucher: U.S.A., Missouri, Butler Co., edge of swamp, S side of Neely ville Jet., W. G. D’Arcy et al. 4557 [MO 2155947]). Samples were fixed in five parts stock formalin/five parts glacial acetic acid/90 parts 50% ethanol (FAA). Flower buds and seeds were dehy- drated through an ethanol series and then embedded in Technovit 7100 (Kulzer, Wehrheim, Germany) for microtoming. Serial resin sections cut at a thickness of 5-7 pm were stained with Heidenhain’s hematox- ylin and mounted in Entellan (Merck, Darmstadt, Germany). The slides’ permanent mounts are housed at the Department of Botany of Kyoto University. ANTHERS AND MICROSPORES The staminate inflorescence is composed of about 40 to 50 flowers (or partial inflorescences) that are spirally arranged on the fertile axis (Fig. 1A). Each staminate flower, which we interpret as a 3-flowered cymule (Abbe & Earle, 1940; Abbe, 1974), bears about 10 stamens and is subtended by the primary bract (Fig. IB). The anther is tetrasporangiate (Fig. IB, F). Prior to maturation the anther wall comprises five or six cell layers, namely an epidermis, an endothecium, two or three middle layers, and a tapetum (Fig. 1C). The anther epidermis is persistent remains uncertain. The tapetum is glandular (Fig. ID, E) and its cells are usually multinucleate, with cells often appear to be fused. During maturation, the middle layers degenerate and the epidermal cells become somewhat enlarged. Cells of the endothecium become enlarged and develop fibrous thickenings (Fig. ID, F, G). Anther dehiscence takes place along longitudinal slits, with each slit common to two thecal microsporangia (Fig. 1G). unable to ascertain whether microspore mother cell meiosis is accompanied by simultaneous or succes- sive cytokinesis. Microspore tetrads are predomi- MEGAGAMETOPHYTE AND NUCELLUS The pistillate inflorescence is composed of 12 to 15 flowers arranged spirally on the fertile axis (Fig. 2A). Each flower has a single pistil with an elongate the upper side and a rudimentary ovule on the lower side in an ovarian locule (Fig. 2B). The fertile ovule is epitropous with a micropyle pointing upward; it is crassinucellale. In the absence of the youngest flower buds, I was unable to determine the number of archesporial cells. Each of the youngest buds available (collected 4 March 2000) had a single megasporocyte differentiated beneath a parietal tissue that was two cell layers in thickness (Fig. 2C). The parietal cells further divide periclinally, resulting in a thick parietal tissue (more than 10 cells thick) by the time the female gametophyte develops (Fig. 2F). The megasporocyte undergoes meiosis, resulting successively in a dyad (Fig. 2D), and then a linear tetrad of megaspores (Fig. 2E), as reported by Pfeiffer (1912). In the megaspore tetrad, the chalazal megaspore is functional and the three micropylar megaspores degenerate (Fig. 2E, F). The functional megaspore develops sequentially into a 2- (Fig. 2G), 4- (Fig. 2H), and 8-nucleate female gametophyte (Fig. 21), as reported by Pfeiffer (1912). Therefore, the mode of female gametophyte development corresponds to the Polygonum type (for diverse modes of female gametophyte development in Volume 98, Number 2 2011 Volume 98, Number 2 2011 Tobe Leitneria floridana (Simaroubaceae) 281 angiosperms, see Stuessy, 2009: 216). The organized mature female gametophyte is ellipsoidal, comprising an egg cell, two synergids, two polar nuclei, and three antipodal cells. The three antipodal cells are ephemeral (Fig. 21) and disappear fertilization. During megasporogenesis, the apical dermal cells of the nucellus divide periclinally to form a nucellar cap that is three cell layers thick (Fig. 2F), as reported by Pfeiffer (1912). At maturity, the ovule is hemitropous with the 3D). The nucellar tissue is very thick around the mature female gametophyte (Fig. 3D). At this stage, the parietal tissue lying above 1 gametophyte is up to 15 cell layers thick (Fig. 3D). It thickens further and persists through post-fertiliza- tion stages without being crushed by the enlarging female gametophyte (Fig. 4A, D, E); it may also persist in mature seeds. No set sequence of further development occurs in the chalaza, which may result, for example, in pachychalazy. A hypostase does not differenLiale. even in posl-ferlilizalion slages. No conspicuous obturator is formed. continue in the nucellus, so that a thick nucellar tissue is formed on the chalazal side (Fig. 4B). A circumscribed nucellar tissue eventually disappears as the seed grows, but a degenerating fragment may seeds (Fig. 41). Endosperm formation is of the nuclear type, and free nuclei are observed in the female gametophyte during early embryogenesis (Fig. 4C), as reported by the endosperm is cellularized centripetally (Fig. 4D). In mature seeds, the 10- to 30-cell-layered endo- sperm remains located outside the large embryo (Fig. 4H). embryogenesis in detail in this study, but fragmentary data on early and later embryogenesis indicate that the process progresses embryos (Fig. 4C-E). The suspensor is not especially large (Fig. 4E). The embryo in a mature seed is straight, large, and dicotyledonous (Fig. 4F). SEED AND SEED COAT INTEGUMENTS The ovule is bitegmic (Fig. 3A-D). The i integument is initiated first and the c much later and is always shorter than the inner (Fig. 3A-C). On the antiraphal side, both the inner and outer integuments are initially two cells thick (Fig. 3A), but through anticlinal divisions of inner most of their lengths (Fig. 3B-E). In contrast, on the raphal side, the inner integument is thicker, dividing to four cells thick (Fig. 3B). Neither the inner nor the respective thicknesses do not change throughout development. No vascular supply was found in either the inner or outer inlegumenl. The micropyle is formed by the inner integument to 4-nucleate stage. This integument grows beyond the apex of the nucellus within the confined space of the locule, so that its tip is extremely elongated and irregularly folded (Fig. 3D). ENDOSPERM AND EMBRYO Although I did not confirm pollen tube passage through the micropyle, there is no doubt that the pollen tube grows from the nucellar apex to the female gametophyte through the thick parietal tissue (Fig. 4A). In post-fertilization stages, cell divisions are about 27-30 mm long and 5-8 mm wide, each containing a single seed (Fig. 4F, upper right). The seed is enclosed by a hard endocarp composed of sclerotic cells (Fig. 4F). When the seed is still young, the coat comprises a 3- or 4-cell-layered tegmen (i.e., a developed inner integument) and a 4- or 5-cell- layered testa (i.e., a developed outer integument) (Fig. 4G). Among cells of the tegmen, those of the innermost layer, that is, the endotegmen, stain darkly, while cells of the outer layers are not specialized (Fig. 4G). Within the layers of the testa, the outermost layer cells are somewhat enlarged, while those of the other layers are unspecialized (Fig. 4G). Mature seeds are straight and exarillate. The testa is slightly multiplicative and becomes five to 11 cell layers thick (Fig. 4H, I). Although most of the cell layers of the testa persist, those of the tegmen degenerate (except for the innermost layer). Thus, the mature seed coat comprises a 5- to 11-cell-layered testa and a 1 -cell-layered tegmen (i.e., an endoteg- men) (Fig. 41). Cells of the exotesta are variously enlarged and contain scattered granular pigments. Cells of the mesotesta and endotesta are unspecial- ized, although Pfeiffer (1912: 198) observed pitted walls in all of them. Cells of the endotegmen have slightly thickened walls, as observed by Pfeiffer (1912). The seed coat may be categorized as testal (for seed coat terminology, see Comer, 1976; Schmid, 1986). Volume 98, Number 2 2011 spermum L.)13 Thickness of oi Thickness of oi . 11 cells thick 3 to 10 cells thick 2 to 40 < Annals of the Missouri Botanical Garden Volume 98, Number 2 2011 Tobe Leitneria floridana (Simaroubaceae) 287 Comer (1976), Nair Boesewinkel Bacchi (1943), Banerji & & Joseph (1957), (1960), Narayana Garudamma (1977, 1978, (1957), Wiger (1956, 1957), 1984), (1935) Ghosh (1966a, Boesewinkel & Bouman (1978), Comer (1976), Desai (1962), (1957), Mauritzon (1935, (1971), (1963), List & Steward (1965), (1936), Nair & Joseph (1960), Netolitzky (1926), Tobe & Peng (1990), van der Pijl (1957) * Data of Pfeiffer (1912) (Prakash et al., 1977). 7 The micropyle is formed by the inner integument in Naregamia alata (Nair, 1959a), Melia azedarach (Nair, 1959b), Sandoricum , 1978). nophylla and Pilocarpus racemosus Vahl. 'ersely in NepheliumL., or 288 Annals of the Missouri Botanical Garden nucellar cap; the hypostase is not differentiated, and the obturator is absent. The ovule is bitegmic, with both the inner and outer integuments initially two cell layers thick, soon thickening to three (or four) cell layers. The inner and outer integuments lack vascular bundles, and the outer integument is always shorter than the inner. The micropyle is formed by the inner integument alone; the inner integument is extremely elongate to beyond the apex of the nucellus, and is irregularly folded. Fertilization is probably porogamous in Leitneria Jloridana, with nuclear-type endosperm formation. The mature seeds are albuminous, with a 10- to 30- and the mature seed embryo is straight, large, and dicotyledonous. The mature seed is straight and exarillate; the mature seed coat is testal, comprising a 5- to 11 -cell- layered testa and a 1-cell-layered tegmen (endoteg- men); the other cell layers of the tegmen are crushed. Cells of the exotesta are slightly enlarged, containing granular pigments; cells of the mesotesta and endotesta are unspecialized, while those of the endotegmen are slightly thick-walled. with those from a wide range of members of the Simaroubaceae (19 genera), Meliaceae (50 genera), and Rutaceae (161 genera), using the family Sapindaceae s.l. (including Aceraceae and Hippo- castanaceae) as an outgroup. None of these families have been subjected to comprehensive embryological examination, and even for species investigated thus far, data are as yet fragmentary. Embryological data for the Simaroubaceae have been collected from 12 species of the genera Ailanthus, Brucea, Picrasma, Quassia L., Samadera Gaertn., and Simarouba Aubl. (Wiger, 1935; Nair & Joseph, 1957; Narayana, 1957; Nair & Sukumaran, 1960; Comer, 1976); those for the Meliaceae are from 18 genera in two major clades representing the subfamilies Cedreloideae ( Carapa Aubl., Cedrela P. Browne, Chukrasia A. Juss., Khaya A. Juss., and Swietenia Jacq.) and Melioideae (e.g., Aglaia Lour., Aphanamixis Blume, Chisocheton Blume, Cipadessa Blume, Dysoxylum Blume, Guarea F. Allam. ex L., Lansium Correa, Melia L., Naregamia Wight & Am., Sandoricum Cav., Trichi- lia P. Browne, Turraea L., and Walsura Roxb.) (Wiger, 1935; Gamdamma, 1956, 1957; Nair, 1958, 1959a, 1959b; Narayana, 1958; Nair & Kanta, 1961; Ghosh, 1966a, 1966b; Comer, 1976; Prakash et ah, 1977; Boesewinkel, 1981) within the family phylog- eny (Muellner et al., 2003); those for the Rutaceae are from 35 genera (Adenandra Willd., Aegle Correa, Agathosma Willd., Atalantia Correa, Boenninghau- senia Rchb. ex Meisn., Boronia Sm., Calodendrum Thunb., Chloroxylon DC., Choisya Kunth, Citropsis (Engl.) Swingle & M. Kellerm., Citrus L., Clausena Burm. f., Cneorum L., Coleonema Bartl. & H. L. Wendl., Dictamnus L., Dictyoloma A. Juss., Diosma L., Empleurum Aiton, Eriostemon Sm., Erythrochiton Nees & Mart., Flindersia R. Brown, Fortunella Swingle, Glycosmis Correa, Limonia L., Melicope J. R. Forst. & G. Forst., Murraya Koenig ex L., Phellodendron Rupr., Pilocarpus Vahl, Poncirus Raf., Ptelea L., Ruta L., Skimmia Thunb., Spathelia L., Triphasia Lour., and Zanthoxylum L.) (Mauritzon, 1935, 1936; Bacchi, 1943; Banerji, 1954; Johri & Ahuja, 1957; Desai, 1962; Narayana, 1963; Comer, 1976; Boesewinkel, 1977, 1978, 1984; Boesewinkel & Bouman, 1978); those for the Sapindaceae are from 20 genera ( Acer L., Aesculus L., Alectryon Gaertn., Allophylus L., Cardiospermum L., Cupania L., Diplopeltis Endl., Dodonaea Mill., Guioa Cav., Harpullia Roxb., Koelreuteria Laxm., Litchi Sonn., Magonia A. St.-Hil., Nephelium L., Paullinia L., Pometia J. R. Forst. & G. Forst., Sapindus L., Trigonachras Radik., Ungnadia Endl., and Xantho- ceras Bunge) (Guerin, 1901; Netolitzky, 1926; Mauritzon, 1936; David, 1938; Banerji & Chaudhuri, 1944; van der Pijl, 1957; Nair & Joseph, 1960; Khushalani, 1963; List & Steward, 1965; Haskell & Postlethwait, 1971; Comer, 1976; Tobe & Peng, 1990). The Rutaceae and Sapindaceae, which are features, particularly in seed and seed coat characters (Table 1). Thus, I collected embryological data for the Rutaceae from Cneorum (Boesewinkel, 1984), which represents the basal lineage within the family phylogeny (Muellner et al., 2007); data for the Sapindaceae were collected from Xanthoceras of the Xanthoceroideae (Guerin, 1901; Corner, 1976), which likely represents the first lineage to diverge within the entire family (Harrington et al., 2005). Comparisons show that Leitneria has many embry- ological features in common with the Simaroubaceae, Meliaceae, Rutaceae, and Sapindaceae. Even fea- tures that appeared distinctive to Leitneria and were expected to be of use in critical comparisons (e.g., occurred widely in all of the families compared. Such shared features do not provide evidence of affinities between Leitneria and any particular family or limited group of families. However, Leitneria either shared or are summarized in Table 2, and the distribution of Annals of the Missouri Botanical Garden Volume 98, Number 2 2011 Tobe Leitneria floridana (Simaroubaceae) 291 i with the Rutaceae (character 4). All of these seed and seed coat features are found in Leitneria. Thus, although an understanding of complete, Leitneria apparently fits best within the Among simaroubaceous genera for which embry- ological data are available, Brucea has the closest resemblance to Leitneria in that both genera form a l elongate, irregularly folded inner n Table 2). In addition to the two species of Brucea, micropyle formation has been examined in eight species of the genera Ailanthus, Quassia, and Samadera (see footnote 1 in Table 1). Wiger (1935: 83), who examined six species of Ailanthus, Quassia, and Samadera in addition to B. amarissima (Lour.) Merr., noted that “The further development in Brucea is rather interesting. The inner integument is tube-like and lengthened and by pushing against the placenta more or less folded, the micropyle thus being rather long and winding.” Nair and Sukumaran (1960), who also studied ovule development in B. amarissima, did not examine the micropyle, but a drawing of the mature ovule (Nair & Sukumaran, 1960: 176, fig. 28) clearly demonstrates that the micropyle is formed by an elongate, irregularly folded inner integument. A similar micropyle structure is also found in B. javanica (L.) Merr. (Comer, 1976: 459, fig. 536); Comer (1976: 253) reported that in B. javanica the inner integument elongates beyond the exostome and is often splayed out against the carpel wall. As noted above, a molecular sister-group relationship exists between Leitneria and the clade comprising Brucea, Soulamea, and Amaroria (Clayton et al., 2007). When ceae with MacClade vers. 3.04 software (Maddison & Maddison, 2005), micropyles formed from an elon- gate, irregularly folded inner integument were restricted to Leitneria and the clade containing Brucea, Soulamea, and Amaroria (Fig. 6). Thus, a morphological synapomorphy very likely unites Leitneria and the clade comprising Brucea, Soula- mea, and Amaroria (Fig. 6). We have no information on the micropyles of Soulamea and Amaroria, but they may share with Leitneria and Brucea a micropyle formed by an elongate, irregularly folded inner integument. The significance of this type of micropyle formation is not clear, but it may play a role in selecting pollen for fertilization, particularly in the anemophilous Leitneria. Seed coat stmcture may provide yet another synapomorphy for the clade containing Leitneria, Brucea, Soulamea, and Amaroria. In Leitneria the mature seed coat is composed of a testa with five to 11 cell layers and a tegmen of just one cell layer (i.e., an endotegmen). While the cells of the testa are little specialized, those of the endotegmen, which is the only persistent cell layer in the tegmen (character 11 in Table 2), stained darkly early in development and were somewhat thick-walled at maturity. According to Nair and Sukumaran (1960), cells of the endotegmen in B. amarissima are persistent and have thick walls and a dense accumulation of darkly staining contents. Thus, the seed coat structure of Leitneria is very similar to that of B. amarissima. Comer (1976) reported that all cells of the tegmen are crushed in B. javanica, Picrasma javanica Blume, Quassia indica (Gaertn.) Noot., and Simarouba spp., but this needs confirmation because observations on B. amarissima (Nair & Sukumaran, 1960) and B. javanica (Comer, 1976) are not concordant. Leitneria has a very thick parietal tissue in the nucellus that persists until late in the post-fertiliza- tion stages. In other angiosperms, the parietal tissue is generally destroyed by an enlarging female gametophyte before and after fertilization. The persistent thick parietal nucellar tissue of Leitneria may provide another key character for future comparisons with other Simaroubaceae. In conclusion, although previous morphology- based systematic treatments have not suggested a simaroubaceous affinity for Leitneria, my embryolog- ical studies showed that features of the genus are concordant with members of the Simaroubaceae, particularly Brucea, corroborating molecular evi- dence (Clayton et al., 2007). Extensive studies on the Simaroubaceae (particularly genera considered basal [e.g., Ailanthus, Castela Turpin, and Picrasma \ in the family phylogeny) would provide invaluable evidence for further assessment of the systematic position of Leitneria. Literature Cited Abbe, E. C. 1974. Flowers and inflorescences of the “Amentiferae.” Bot. Rev. (Lancaster) 40: 159-261. Abbe, E. C. & T. T. Earle. 1940. Inflorescence, floral Torrey Bot. Club 67: 173-193. Angiosperm Phylogeny Group. 1998. An ordinal classifi- Bot. Gard. 85: 531-553. § P Erratum J i| ill!! Ill i Missouri Bot PGard. 97: 457-467. J 11! HI 1 In Table 1, the incorrect GenBank numbers were Jj HI!! Ill I -1 ii ii inn in i j Ji- mu in i I £ >! nil! lil i ti Si i. Sills III I I 1 1} I I III \\ 1 II I iilii Hi i II 1 Si i| '1 I fell yin Volume 98, Nur www. mbgpr ess . info 157 i L. G. De Toni & Jorge E. A. Mariath Peter Goldblatt & Porter P. Lowry II Volume 98 Number 3 Volume 98, Number 3 September 2011 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, in- dependent reviewers. Instructions to Authors are printed in the back of the last issue of each volume and are also available online : at www.mbgpress.info. Editorial Committee Roy E. Gereau Victoria C. Hollowell Scientific Editor, Latin Editor, Missouri Botanical Garden Missouri Botanical Garden Ihsan A. Al-Shehbaz Beth Parada Missouri Botanical Garden Managing Editor, Gerrit Davidse Missouri Botanical Garden Missouri Botanical Garden Allison M. Brock Peter Goldblatt Associate Editor, Missouri Botanical Garden Missouri Botanical Garden Gordon McPherson Tammy Charron Missouri Botanical Garden Associate Editor, Charlotte Taylor Missouri Botanical Garden Missouri Botanical Garden Cirri Moran Henk van der Werff Press Coordinator, Missouri Botanical Garden Missouri Botanical Garden of the Missouri Botanical Garden, % Allen Mar- keting & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Subscription price for 2011 is $180 per volume U.S., $190 Canada & Mexico, $215 all other countries. Four issues per volume. The journal Novon is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www.mbgpress.info The Annals of the Missouri Botanical Garden (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. Postmaster: Send address changes to Annals of the Missouri Botanical Garden, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Abstract/Global Health databases, Ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. The full-text of Annals of the Missouri Botanical Garden is available online though BioOne™ (http:// www.bioone.org). © Missouri Botanical Garden Press 2011 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Volume 98 Annals m Number 3 of the 2011 Missouri Botanical Garden WHY ARE WE STILL PRODUCING PAPER FLORAS?1 298 Annals of the Missouri Botanical Garden (), the Biodiversity Heri- tage Library (), and the Botanicus Digital Library ( and JSTOR: ). Now, several scientific disciplines (e.g., medicine, astronomy, chemistry, and physics) are providing the public with pre-print access to articles (Tomaiuolo & Packer, 2000). Online, peer-reviewed, open-access electronic publication of journal articles has begun in the biological sciences (), even are not printed because of cost and space issues (Vision, 2010). For the most part, however, botany has lagged behind, despite current examples of online publications ( Constancea , , The Jepson Manual, , Phy- totaxa, , Phyto- neuron, , and Vulpia, ), and even earlier attempts at computerization (Krauss, 1973). Currently, the issue of valid and effective paperless publication of botanical names is under consideration, predicated by an International Associ- ation of Plant Taxonomy’s (IAPT) special committee report (Chapman et al., 2010). Even earlier, before pre-print status (or post-print, see Hamad, 2003a) of a manuscript, the internet enabled the shift from the trade model (i.e., traditional, restricted commercial publishing) to an online collaborative model for peer review (Hamad, 1996, 2003b; Wilson, 2001). Editors of botanical literature routinely seek scientific reviews by spe- cialists familiar with the concerned plant family and geographic region. To utilize a worldwide audience of reviewers and to connect authors and readers (Tenopir, 1995), taxonomic treatments should be provided online at all stages of preparation and revision (in certain cases after initial editing, e.g.. international floras where the author’s native lan- guage differs from that of the flora). paper and declining funding and bookshelf space. Yet for those who still wish to own a personal, printed copy of a particular book, print on demand (POD) is available (Black, 2011) and is becoming an economical alternative (e.g., ); however, these are not always of the same print quality and POD can contractually confound issues of copyright and intellectual property, though these problems are becoming less common. If the copyright Electronic paper (Heikenfeld et ah, 2011; e.g., ) is increasingly popular. Tablet platforms have exploded within the past year, e.g., Motorola’s Xoom with Android’s 3.0 Honeycomb competitive with Apple’s iPad. Botanists have not yet tapped the potential for floristic apps for smart phones and tablets as field Conventional floras are not designed for use in the field, yet many botanists want to be able to identify plants in the field. Imagine the possibilities of having a flora on a smart phone or tablet and not needing internet access as required by a search engine. Apps extend the taxonomic reach in the field. If the device internet as long as the flora is loaded. Additionally though, internet access can provide supplemental information (e.g., electronic keys, digitized photos and illustrations, specimen data, etc.). However, it is noted that some people still prefer carrying around books in which they can occasionally press a specimen within pertinent pages. Other issues include the readability of electronic display screens in bright daylight, the electricity requirements of devices (at least for charging or batteries), cost to produce computers as compared to paper, and long- term archiving (although the Internet Archive is noted here: ). Digital Floras When paper publications are digitized, previously unavailable functions, such as full text searching and hyperlinking, become possible (Rosatti & Duncan, 1995; Heidom, 2004; Kirkup et al., 2005; Kress & Krupnick, 2006). Online floras are now providing multiple-field searches into the data (see Brach & Song, 2006), which allow searching by taxon name or country, as well as searches through free text () and linking treatments to related information, photos and illustrations, maps, databases, checklists, and other floras (Funk, 2006). In addition to full text or “context searching,” digital floras also allow for updating with new revised treatments as they become available instead of having to wait for the next paper flora to be published. Data extractions of florulas for geograph- ical areas of special interest, e.g., at a country, state/ Volume 98, Number 3 2011 Brach & Boufford Moving Beyond Paper Floras 299 province, or county level, are facilitated by digitiza- tion (see Kartesz, 2011). Botanists are challenged to create interactive identification keys (see Dallwitz et al., 2000 onwards; Brach & Song, 2005, ; Kuoh & Song, 2005, http://www. efloras.org/key_page. aspx?set_id=l 0083 &flora_id= 1050; also e.g., Neotropikey: ). They can provide images to be utilized (e.g., Tropicos image search, ; Biodiversity of the Hengduan Mountains, ) by an online audience of students, researchers, and the general public worldwide. The Flora of China Project (FOC; ) provides online treatments for vascular plants at all manuscript stages () prior to paper publication of text volumes. Draft manuscripts are noted as such in the header, uploaded to the Web site as HTML files, and replaced with revised versions until considered final for volume publica- tion. Once a particular volume is sent to the printer, the final manuscripts for the volume are then databased (, see Brach & Song, 2006). Because of this dynamic accessibility, we regularly receive e-mail from the general public from specialists worldwide, suggesting corrections and inquiries for particular taxa. Also, after printed publication, the project’s editors and coauthors, and specialists outside of the project have the opportunity to submit corrections, which we have routinely incorporated into both the databased, published treatments accessible via an interface on and into the project’s database of corrigenda for published volumes. Newly published names not available in the printed version are added to the eFloras database and to the FOC Checklist (), also see Brach & Song, 2006). The checklist is maintained as a project within the Web-accessible Tropicos (). Because the database is regularly updated with corrections, flora users access a “dynamic” flora via the eFloras interface (while the family PDFs are currently maintained in “static” form so as to provide the original, published treatment). Questions for the future will include how to manage major revisions of families and genera after “print publication” of volumes, alternative treatments by authors differing in their taxonomic opinions, and authorship of revisions to the database years later. The FOC Project provides interactive identification keys (Brach & Song, 2005) in DELTA Intkey and NaviKey formats (). Electronic keys are available for large genera (generally 50 or more species, but also for some smaller genera). The DELTA data are archived and accessible () for updates and improvements. The keys can be modified (characters/states and items/taxa) using the DELTA Editor application, and DEVELOPMENTS IN DIGITIZING FLORAS Existing natural language descriptions can be parsed by software such as MARTT (Markuper of Taxonomic Treatments) (Cui, 2008) into XML language (demonstrated at ), thereby providing searchable databases online (). For new and revised editions of floras, digital floras should be created starting at the beginning of any new project. Programs such as DEscription Language for TAxonomy (DELTA, Dallwitz, 1980; Dallwitz et al., 1993 onwards; Brach & Song, 2005) provide tools (cf. the DELTA Editor: ) for inputting characters, character states, measurements, and counts into data matrices to then output as natural language descriptions (in which natural language descriptions read as a natural language and not as machine-generated) and inter- active identification keys that can be linked to related images. Open repositories for interactive key datasets (see ) may be used as examples in building new character sets and interactive keys and in creating new data conversion (Structure of Descriptive Data, SDD; DELTA, etc.) and input utilities, and also to provide an archive for the data. Also, digital image libraries of taxa can be implemented as part of the identification process (Agarwal et al., 2006; Hearn, 2009; also Wikispecies, and Wikimedia Commons, ). With pressing issues of cataloguing biodiversity, conservation, and sustainable use of resources, botanists are now challenged to routinely prepare and revise online treatments, including interactive identification keys and images for a worldwide audience of students and researchers. As botanists, TAXONOMIC STUDY OF GYMNOPOGON (POACEAE, CHLORIDOIDEAE, CYNODONTEAE)1 (1971), who 304 Annals of the Missouri Botanical Garden Volume 98, Number 3 2011 Annals of the Missouri Botanical Garden , 2.2-2.3 X 0 .3-0.4 n 310 Annals of the Missouri Botanical Garden ligules ca. 0.5 mm, Volume 98, Number 3 2011 314 Annals of the Missouri Botanical Garden 316 Annals of the Missouri Botanical Garden 317 , 3. 5-4.5 mm, 1- to 2- 3.5-1 mm, upper glume 3-1.5 mm, (0.3-)1.5-2 mm on the spikelets 2-flowered, lemmas with longer hairs, and eastem Bolivia, between 200 and 500 m elevation, glumes eurved. Boeehat and Vails (1990b) also Flowering was noted between May and July. J 66°30'W, Killeen 2588 (I 1 spikelets and long hairs in the lemma Egt John Sandiforthj E riff Stir 318 Annals of the Missouri Botanical Garden 322 Annals of the Missouri Botanical Garden Plants perennial, rhizomatous, rhizomes stoloni- form, thin (2-3 mm diam.), yellowish, glabrous, from which several aerial shoots develop; culms terete, 30-60 cm, slender, ascendent or decumbent, simple, erect; intemodes yellowish, inconspicuously striate, glabrous; nodes inconspicuous, slightly thickened, yellowish, glabrous; floriferous culms terete, yellow- ish, longitudinally striate and hispid at the apex, otherwise glabrous. Sheaths longer than the inter- nodes, finely striate, yellowish, pale toward the Volume 98, Number 3 2011 losengurtt, B„ B. R. Arrillaga de Maffei & P. Izaguirre d Artucio. 1970. Gymrwpogon P. Beauv. Pp. 241-244 ii t'c’ StLJlef 3 a, Botany Vol. 24, Pt. 2. Field T&sttaestszszvs .a. (3). Mi™; pLsOm' in,: Gatingen 2. (1), s.n. (3); Gehrt lllS^— zsp^sisz , B. R. Arrill. & (9). Aee 41102 (8a), 41330 (8a), 44332 (13), 46119 (13); Nkora385 (13); Nuttall s.n. (1). W iw; 3«3 (is). %z 2222,2 362 ;;r NUEVAS SECCIONES EN Daniel A. Gmliano ’ y Susana E. Freire ‘ BACCHARIS (ASTERACEAE, ASTEREAE) DE AMERICA DEL SUR1 KIP ca. 4:1 (vs. ca. 1:1), y los B. acaulis (Wedd. ex R. E. Fr.) Cabrera y B. nivalis minadas ca. 4:1 (vs. 6:1 a 8:1), corolas de las flores do (vs. . dos), proporcion de flores carpeladas:estaminadai 4:1 (vs. ca. 1:1), aquenios teretes glabros seccion Palenm (vease mas adelante). ?! £g?P 'lAA Annals of the Missouri Botanical Garden del papus de las flores carpeladas, si bien son uniseriadas, son ligera a moderadamente acrescentes y persistentes. En consecuencia, resultana neeesario efectuar estudios filogeneticos para establecer con precision la ubicacion sistematica de la seccion Helerolhalamus y de la aim seccion Subliguliflorae , asimismo incluida por Muller (2006, bajo el grupo informal de Baccharis polifolia ) en el subgenero Molina. Distribucion. Baccharis secc. Helerolhalamus comprende cineo especies propias del sur de Brasil, Uruguay y centro de Argentina y la region andina desde el sur de Peru hasta el norte de Chile y el noroeste de Argentina (Barroso, 1976; Giuliano, 2000, 2008; Muller, 2006). y mayormente adheridas entre si; cerdas del papus de las flores estaminadas con apice ensanchado. Dentro del subgenero Baccharis , la secccion Icma (Phil.) Giuliano se distingue por presentar los siguientes caracteres: subarbustos erectos; pelos no glandulares de los nidos con celula apical breve- mente flagelada; capitulos sesiles, solitarios en el extremo de las ramas y rodeados por un falso involucro de pequenas hojas, a veces agrupados formando una sinflorescencia corimbiforme; propor- cion de flores carpeladasreslaminadas ca. 2:1. Distribucion. Comprende una sola especie, en- demica de Argentina (Cabrera, 1974; Giuliano, 2000, 2008). ESPECIE INCLUIDA ESPECIES INCLUIDAS 1. Baccharis aliena (Spreng.) Joch. Muller, Syst. Bot. Monogr. 76: 305. 2006. Basionimo: Marshallia aliena Spreng., Syst. Veg. [SprengelJ (ed. 16) 3: 446. 1826. Heterothalamus alienus (Spreng.) Kuntze, Revis. Gen. PI. 3(3): 158. 1898. TIPO: Uruguay. Montevideo, s.d., Sello s.n. (holotipo, P foto en LP!). Observaciones. Malagarriga Heras (1957) y Barr- oso (1976) consideran a Rio Grande do Sul, Brasil, como la localidad tipica. 2. Baccharis boliviensis (Wedd.) Cabrera, Bol. Soc. Argent. Bot. 16: 256. 1975. 3. Baccharis hyemalis Deble, Balduinia 9: 20. 2006. 1. Baccharis gilliesii A. Gray, Proc. Amer. Acad. Arts, 5: 123. 1860-1862. Icma involucrata Phil., Anales Univ. Chile 41: 741. 1872. TIPO: Argentina. Mendoza, s.d., R. A. Philippi s.n. (holotipo, SGO no visto; isotipos, CORD foto en SI!, G foto serie F.M. n° 28510 en LP!). Baccharis gilliesii fue anteriormente incluida en la seccion Cuneifoliae (Giuliano, 2001), por poseer capitulos sesiles solitarios en el apice de las ramas, cada uno rodeado por un falso involucro de pequenas hojas, pero se diferencia de este taxon por presentar pelos uniseriados de los nidos pilosos con celula apical brevemenle flagelada y con paredes delgadas (vs. triangular y con paredes gruesas en seccion Cuneifoliae) y sinflorescencia corimbiforme (vs. racemiforme). 4. Baccharis psiadioides (Less.) Joch. Muller, Syst. Baccharis secc. Palenia (Phil.) Giuliano, comb, et Bot. Monogr. 76: 306. 2006. 5. Baccharis wagenitzii (F. H. Hellw.) Joch. Muller, Syst. Bot. Monogr. 76: 306. 2006. stat. nov. Basionimo: Palenia Phil., Anales Univ. Chile 90: 7. 1895. TIPO: Palenia delfmii Phil. [= Baccharis nivalis (Wedd.) Sch. Bip. ex Phil.J. Baccharis secc. Icma (Phil.) Giuliano, comb, et stat. nov. Basionimo: Icma Phil., Anales Univ. Chile 41: 740. 1872. TIPO: Icma involucrata Phil. [= Baccharis gilliesii A. Gray]. Esta seccion pertenece al subgenero Baccharis en virtud de los siguientes caracteres: corola de las flores carpeladas con apice 5-dentado; aquenios teretes, 10- a 12-costados, glabros, sin papilas, no visibles debajo del involucro a la madurez; cerdas del papus de las flores carpeladas dispueslas en dos series, acrescentes, caducas; corola de las flores estaminadas con limbo gradualmente diferenciado del tubo; apice del estilo obtrulado, con ramas breves La unica especie incluida en esta seccion presenta habito herbaceo, con tallos decumbentes o ascen- dentes; el indumento esta constituido por diminutos pelos no glandulares uniseriados con cuerpo de 4 6 5 celulas rectangulares anchas y celula apical breve y de apice obtuso o redondeado, y pelos glandulares biseriados, no agrupados en nidos, ubicados sobre las nervaduras de las hojas. Las hojas son sesiles, enteras, en apariencia uninervadas. Los capitulos son largamente pedunculados, solitarios en el apice de las ramitas; los involucros son acampanados; los receplaculos convexos, desprovislos de paleas; la proporcion de flores carpeladas:estaminadas es ca. 1:1. La corola de las flores carpeladas presenta apice 347 AN EVALUATION OF Sachiko Nishida* and Henk van der Werff » CLASSIFICATION BY CUTICULAR CHARACTERS OF THE LAURACEAE: A COMPARISON TO MOLECULAR PHYLOGENY1 Volume 98, Number 3 Nishida & van der Werff 349 201 1 Cuticular Characters and Phylogeny of the Lauraceae as Endlicheria Nees, Licaria Aubl., Nectandra Rol. ex Rottb., Pleurothyrium Nees, and Rhodostemono- daphne Rohwer & Kubitzki, are included within this complex. Chanderbali et al. (2001) sampled only the Ocotea complex representatively, within which seem- ingly natural groups of genera and parts of larger well-supported clades of this molecular study and their cuticular features, we investigate to answer the following three questions: (1) Do the cuticular characters vary within the clades recognized in Chanderbali et al. (2001)?; (2) Are cuticular features cuticular features hold promise for characterizing or identifying species or genera? We also discuss whether the species groups based on cuticular characters better agree with the clades in the molecular phylogeny or species grouped according to the traditional generic concepts. Materials and Methods Cuticles of 50 Neotropical species from 13 laurel genera (Aiouea Aubl., Aniba Aubl., Dicypellium Nees & Mart., Endlicheria, Kubitzkia van der Werff, Licaria, Nectandra, Ocotea, Paraia Rohwer, H. G. Richt. & van der Werff, Pleurothyrium, Rhodostemo- nodaphne, Umbellularia (Nees) Nutt., and Urban- odendron Mez) were examined. Their phylogenetic relationships as inferred from sequence variations of the nuclear genomes (ITS/5.8S) were identified by Chanderbali et al. (2001). Table 1 lists the sources of plant materials. Leaves were sampled from herbarium specimens at MO, using one leaf sample per species. Specific cuticular characters within a species remain constant (Nishida & Christophel, 1999; Nishida & van der Werff, 2007). The cuticles studied here were the cuticular membranes of the epidermis or stomatal complex that remained through the preparation, and the cuticular characters described were mostly features of the epidermal cells or stomatal complex whose impressions were preserved in the membrane (Chris- tophel & Rowett, 1996). The examination procedure followed that of Christophel et al. (1996), Nishida and Christophel (1999), and Nishida and van der Werff (2007). Samples (1 X 1 cm) were taken from near the left basal margin (adaxial surface up) of mature leaves. The leaf samples were soaked in 90% ethanol for ca. 18 hr., then placed in a test tube with 2 mL 30% H202 and 0.5 mL 90% ethanol. The test tubes were heated at 120°C in a heated dry block bath for about 2 hr. When the samples turned yellow, they were placed in 90% ethanol for ca. 10 hr. before rinsing in 2% ammonia (to adjust the pH), and were transferred to a Petri dish with double-distilled water (ddH20). The cellular contents of the sample leaves were removed with a fine artist’s brush. The cuticles were stained in 0.1% crystal violet for ca. 1 min., then mounted in phenol glycerin jelly on a slide and observed under a light microscope. Feature descrip- tions followed Wilkinson (1979), Christophel et al. (1996), Nishida and Christophel (1999), or Nishida and van der Werff (2007). The cuticles were also examined using an SEM. Sample preparation was the same as described above. Samples were dehydrated in a I-bulanol series, freeze-dried using a JFD-310 (JEOL, Tokyo, Japan) at 8°C, then coated with platinum, and observed under a JSM-6060B microscope (15 kY; JEOL). Results Table 2 lists the cuticular characters recognized in this study. Figures 1-3 show representative micro- graphs of cuticles in major cuticular groupings. All the species examined were hypostomatic. In addition to the absence or presence of stomata, cuticular features of epidermal cells differed within species between the adaxial and abaxial leaf surfaces. Among the cuticular characters often used in laurel taxonomy (e.g., Christophel & Rowett, 1996; Christophel et al., 1996), anticlinal wall straightness and stomatal features have varied considerably. Periclinal wall ornamentation and anticlinal wall thickness, in contrast, were uniform for most of the species examined, although both features can vary widely and are useful in distinguishing other laurel species (Christophel et al., 1996). The periclinal wall ornamentation of all of the species was smooth, except as seen in the abaxial leaf surfaces of Pleurothyrium cinereum van der Werff (wrinkled), Aniba cinnamomiflora C. K. Allen (papillose), and A. panurensis (Meisn.) Mez (papillose). The anticlinal walls of all of the species were more or less beaded. Following Wilkinson (1979) and Nishida and Christophel (1999), we classified anticlinal walls into three types: straight to slightly curved (SC); having loose, wide U-shaped curves of shallow amplitude (LC); or having tight U-shaped curves of shallow amplitude (TC). The main difference among these types was the frequency of curves. SC walls showed no obvious wave shape across one side of the cell but only with the walls straight to roundish (Figs. ID, E, G, H, J, K, M, N; 2G; 3B, D), whereas LC generally had one wave (Figs. 2H, J, K, M, N; 3E) and TC more than two waves per side (Figs. 1A, B; 2A, B, D, E; 3G, H, K, M, N). Annals of the Missouri Botanical Garden Three variations of stomatal features occurred among the examined species: lower stomatal ledge shape, surface appearance of the stomata observed via SEM, and evenness of the shape and size of the subsidiary cells. The stomatal ledges were the cutinized cell walls along the stomatal openings. The ledges, stained dark sharp edges and somewhat resembling a flying bat (BA; Figs. IB, E, H, K, N; 2E, H; 3K), narrow and Volume 98, Number 3 Nishida & van der Werff 351 201 1 Cuticular Characters and Phylogeny of the Lauraceae Epidermis anticlinal walls ID, E, G, H, J, K, M, N; 2G; 3B, D 2H, J, K, M, N; 3E 1A, B; 2A, B, D, E; 3G, H, 3A, J 2B; 3E 2K, N; 3B, H, N 1C, F, I, L; 2F, I, L; 31, L, 3C 3F lip-shaped (NL; Figs. 2B; 3E), or wide with round edges and appearing like a butterfly (BU; Figs. 2K, N; 3B, H, N). The surface appearance of the stomata was categorized into five types: circular and weakly to strongly protruding (C; Figs. 1C, F, I, L; 2F, I, L; 31, L, 0), strongly wrinkled (SW; Figs. 10; 20), weakly wrinkled (WW; Fig. 2C), papillose (PA; Fig. 3C), and lip-shaped or eyelid-shaped and protruding (LI; Fig. 3F). Circular shape varied from elliptical (e.g., Fig. 1L) to round (e.g., Fig. 31), and circles also appeared perfect (e.g., Fig. 30) o stomatal slit (e.g., Fig. 2L), depending on the species. The lip shape was distinguished by the protrusion of the central part of the stomatal surface, whereas the central part of other protrusions was generally depressed (compare Fig. 3F and Fig. 31). Strongly wrinkled surfaces markedly contrasted with the other cells of the stomata or epidermis (Figs. 10; 20), whereas weakly wrinkled surfaces had only shallow folding around the stomata (Fig. 2C). The subsidiary cells on each side of a stoma were more or less even, similar in size and shape (EV; Figs. IB, E, H, K; 2E, K, N; 3B, H, K, N), or uneven and dissimilar (UN; Figs. IN; 2B, H; 3E). The species with the even-shaped subsidiary cells, however, sometimes included stomata with slightly uneven subsidiary cells, tion more difficult. Although Christophel et al. (1996) referred the cells as a useful cuticular feature, we did not find any specialized cells for any species in this study. Discussion Chanderbali et al. (2001) have presented the most detailed phylogeny of the Ocotea complex to date, a study based on the ITS region. Such a phylogeny should be considered a theory of relationships until it the cuticular features of all species included in the ITS-based phylogeny offers a test (Fig. 4). In addition, we will discuss if and to which degree cuticular features can help in the identification of specimens belonging to the Ocotea complex. The first question we will address is if the cuticular characters vary within the clades recognized in Chanderbali et al. (2001). As can be seen in Figure 4, the answer is yes for some features and no for others. The character states of the anticlinal walls vary frequently within the clades. They are only constant in three clades consisting of two species each ( Endlicheria punctulata (Mez) C. K. Allen and Ocotea pauciflora (Nees) Mez; Pleurothyrium ciner- eum and P. insigne van der Werff; the two species of Urbanodendron ) and in the clade consisting of four species of Licaria. 352 Annals of the Missouri Botanical Garden Volume 98, Number 3 Nishida & van der Werff 353 201 1 Cuticular Characters and Phylogeny of the Lauraceae placed by Kubitzki (1982) in the group with a papillose lower leaf surface. Finally, the character states of the stomatal ledges did not vary at all within the clades. Thus, especially the characters of the phylogeny of the Ocotea complex proposed in Chanderbali et al. (2001). According to Christophel et al. (1996), epidermal (nonstomatal) cells have several additional characters supposedly useful for the taxonomy of the Lauraceae, including periclinal wall ornamentation, uniformity of thickness of anticlinal walls, uniformity of thickness of anticlinal walls, uniformity of cell size, and maximum cell dimension. These features are not included in Figure 4, but will be briefly discussed here. Periclinal wall ornamentation was not useful for the taxonomy of the Neotropical Ocotea complex, because almost all the species examined had smooth periclinal walls. Exceptions were for Pleurothyrium cinereum (wrinkled), Aniba cinnamomiflora (papil- lose), and A. panurensis (papillose), all of which broke within the clade. Uniformity of 354 Annals of the Missouri Botanical Garden cuticles (B, E, H, K, N), and SEMs of the insularis. — G-I. Kubitzkia mezii. — J-L, Cell size was also e Annals of the Missouri Botanical Garden reasonable to expect that these Licaria species will also have papillose stomata surfaces. On the other hand, if we consider combinations of features, we find features, while others cannot. Examples of clades defined by cuticular features are clade F (two Pleurothyrium species), clade G (three central American species of Ocotea ), and clade N (the Licaria species). Other clades lack a distinctive Nectandra species, with bisexual flowers) does not differ from clades A, B, C, and D (all species with unisexual flowers). Likewise, Umbellularia califom- ica (Hook. & Am.) Nutt, does not differ in cuticular features from Endlicheria chalisea Chanderbali or from Rhodostemonodaphne crenaticupula Madrinan. In general, cuticular features alone are not sufficient for assigning a specimen to a particular clade. The third question to be addressed is if cuticular features hold promise for identifying species or characterizing genera. One could hope that cuticular features might function as a morphological bar code, allowing identification of sterile specimens without having to resort to DNA analysis. Our results indicate that this is not likely to be the case. Of the largest Neotropical genus, Ocotea, 10 species with unisexual not greatly differ in cuticular characters. Likewise, clade E (four species of Nectandra ) is homogeneous and it is not possible to identify the species solely on their cuticular characters. Some genera can be identified by a particular combination of cuticular features; examples are Pleurothyrium (clade F), although should 1 Urbanodendron and Licaria (clade N). The two largest genera, Ocotea and Nectandra, were found to be polyphyletic in the DNA-based analysis and it is therefore not surprising that neither of those can be defined on the basis of cuticular features. Finally, do the cuticular features validate the DNA-based phylogeny proposed by Chanderbali et al. (2001)? The groups of species that can be correspond well with the clades found in the phylogenetic arrangement of the species (Fig. 4). Of course, groupings of species have also been proposed on purely morphological grounds. In most cases there is a congruence between groups based on morphology, those based on cuticles, and those based on phylogeny. The Ocotea helicterifolia (Meisn.) Hemsl. group was first recognized by Rohwer (1991) and accepted by van der Werff (1999). Mez (1889) already recognized the Ocotea species with unisexual flowers as a distinct group (as subgenus Oreodaphne Nees). Van der Werff (2002) distin- guished the O. insularis (Meisn.) Mez group (repre- sented only by O. insularis in this study). Ocotea subg. Dendrodaphne (Beurl.) Mez was also already recognized by Mez (1889). Rohwer (1993) revised the species of the genus Nectandra and discussed the relationships of the species. He commented that the N. coriacea (Sw.) Griseb. group (including N. purpurea (Ruiz & Pav.) Mez and N. salicifolia (Kunth) Nees) was not linked to any other group in Nectandra, and he questioned whether this group really belonged to Nectandra. Thus, the same species groups are found in the phylogeny, in the analysis based on cuticle characters, and in species groups based on floral and fruit characters. However, there are a few cases where cuticle/phylogeny groups differ from species groups based on morphology. One example is the close relationship between Aiouea costaricensis (Mez) Kosterm. and O. insularis, first found in the phylogeny of Chanderbali et al. (2001). Renner (1982), who revised the genus Aiouea, regarded it as closely related to Endlicheria or Aniba and Licaria, but a relationship with Ocotea had never been proposed. This is an example of species relationships found in the phylogeny that are supported by the cuticle data but conflict with species relationships that are based on flower and fruit morphology. A second example is the group including all species with unisexual flowers (End- licheria, Rhodostemonodaphne, and part of Ocotea) found in the phylogeny, which is supported by the cuticle data but is not found in relationships based on flower and fruit morphology. A third example is offered by the four species of Licaria. In the phylogeny and the cuticle analysis, these species form one group, but in the revision by Kurz (2000), three species groups are recognized on the basis of stamen characters. Two of these groups are repre- sented in our study: L. cannella (Meisn.) Kosterm. is part of one group and the other three Licaria species morphology are not found in the groups based on phylogeny or on cuticle characters. Thus, in two Cuticles of 50 Neotropical species belonging to the Ocotea complex sensu Chanderbali et al. (2001) were studied. Several species groups could be recognized EL GENERO GLANDULARIA (VERBENACEAE) EN ARGENTINA1 Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina hemisfeno sur, se encuentra en la Argentina, Bolivia, Brasil, Chile, Paraguay, Peru y Uruguay. En la Argentina hay una gran diversidad especffica, las que hasta la provincia de Santa Cruz, en el sur aproximadamente entre los 56°5' y los 72° de longitud oeste y desde los 22° y los 51° de latitud sur. Las especies crecen desde el nivel del mar, a lo largo de toda la costa Argentina hasta los 4600 m, en la provincia fitogeografica Altoandina en la provincia de Jujuy, segun el criterio de Cabrera y Willink (1973). Las especies habitan en lugares abiertos como campos graminosos o praderas, bordes de caminos, laderas de cerros, margen de selvas o bosques. Linnaeus funda en 1753 el genera Verbena. En 1791 Gmelin describe al genera Glandularia, sobre la base de Anonymus carolinensis T. Walter, hoy G. carolinensis J. F. Gmel. sinonimo de G. canadensis (L.) Nutt. Sin embargo diferentes autores considera- ron a Glandularia como una seccion (Walpers, 1845; Schauer, 1847; Briquet, 1895; Reiche, 1910; Perry, 1933) o como un subgenero del genera Verbena (Bentham & Hooker, 1876; Lewis & Oliver, 1961), por su parte Moldenke, entre 1940 y 1983, tampoco acepto a Glandularia como genera independiente de Verbena (Moldenke, 1940, 1941, 1942a, 1942b, 1946, 1947, 1948a, 1948b, 1949a, 1949b, 1950, 1953, 1955, 1956, 1958, 1961a, 1961b, 1962, 1963a, 1963b, 1964a, 1964b, 1964c, 1964d, 1964e, 1968, 1969, 1972, 1975, 1979, 1981, 1983; Moldenke & Moldenke, 1949). Autores que tratan a Glandularia como genera son Small en 1933, donde describe las especies que crecen al suroeste de Estados Unidos y Schnack y Covas (1944) y Schnack (1964), quienes dan a conocer por primera vez los argumentos por los cuales Glandularia debe tratarse como genera independiente considerando: numero basico de cromosomas, ploidfa, tamano de los cromosomas, anatonua del tallo, apendices glandulares conecti- vales, longitud del estilo en relacion con el ovario, desarrollo del conectivo en relacion con las tecas. Umber (1979), Botta y Poggio (1988), Botta \l989, 1990, 1992, 1993), O’Leary et al. (2007) y O’Leary y Peralta (2007). Respecto al tratamiento infragenerico del genera Glandularia, Troncoso (1974) establece las secciones Glandularia y Nobiles (Schauer) Trane., basandose en la morfologra floral. En 1978, Schnack y Covas agrupan las especies en dos subgeneros teniendo en cuenta el habito general y caracteres de los frutos: Glandularia y Paraglandularia Schnack & Covas. Posteriormente Umber (1979), revalua las sec- ciones creadas por Troncoso (1974) y pasa a la sinonimia de la seccion Glandularia la seccion Nobiles y establece, tomando en consideracion Botta en 1989 retoma el criterio de Schnack y Covas (1978) y recientemente Sanders (2001), vuelve a validar la clasifrcacion infragenerica con dos sec- ciones previamente establecida por Troncoso (1974). En la presente contribucion no se considera la clasificacion infragenerica, por no realizarse en esta oportunidad el estudio general del genera. Metodos de estudio, analizando ejemp algunos casos plantas vivas, consultandose las descripciones originales, ejemplares tipo y material de herbario de cada uno de los taxones. Los herbarios consultados, citados conforme a las siglas que figuran en Holmgren et al. (1990), fueron los siguientes: B, BAB, BAF, BM, BR, CAS, CONC, CORD, CTES, F, G, G-DC, G-BOIS, G-BU, HBR, ICN, K, LIL, LL, LP, MA, MBM, MO, MPUC, NY, P, PACA, S, SGO, SI, SP, TEX, UC, UPCB, US y W. En el Apendice 1 se brinda la lista completa de nombres cientrficos, rncluyendo los smonimos, en el Apendice 2 se da la lista de los nuevos sinonimos y en el Apendice 3 se da la lista completa de colectores de los ejemplares estudiados. observan Caracteres Morfologicos Las formas biologic Glandularia, (1957), se reunen prrncrpalmente en dos grupos: Hemicriptofitos, dentro de los cuales pueden encon- trarse las siguientes variantes. Hemicriptofitos ras- treros en donde las ramas vegetativas pueden desarrollar rarces adventicias en los nudos, las ramas florfferas son ascendentes como en G. peruviana (L.) Small, G. pulchella (Sweet) Trane., G. tenera (Sprang.) Cabrera y G. tomophylla (Briq.) P. Peralta. Hemi- criptofitos erectos, apoyantes o no, como G. sessilis (Cham.) Trane., G. stellarioides (Cham.) Schnack & Covas y G. tristachya (Trane. & Burkart) Schnack & Covas. El otro grupo representa los camefitos del tipo sufrutice como G. hassleriana (Briq.) Trane, y G. megapotamica (Spreng.) Cabrera & G. Dawson; algunas especies pueden formar grandes matas lignificadas en la base con ramas de renuevo anuales en la periferia como en G. araucana (Phil.) Botta, o Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 361 relacion al caliz pudiendo ser mayores, iguales o menores a la mitad de la longitud de este. FLORES Las flores son perfectas, levemente zigomorfas, con pedicelo muy breve. El caliz es tubular, 5-dentado, con dientes subiguales, en general los abaxiales mas desarrollados que los adaxiales. El caliz fructifero persistente es mas largo que las clusas y se presenta generalmente plegado y conlorlo a la madurez; la dehiscencia se produce por una lmea longitudinal. La corola es hipocraterimorfa, conspicua, blanca, blanco- crema, amarilla, rosa, lila a roja. Puede existir una graduacion en el color dentro de una misma especie e incluso dentro de la misma planta. El limbo es extendido, 5-lobado, con lobulos subiguales, general- mente emarginados, a veces recurvos; el tubo corolino esta bien desarrollado, exteriormente puede ser glabro (Glandularia aurantiaca , G. peruviana , G. radicata (Moldenke) Tronc. ex Mulgura, G. sulphurea (D. Don) Schnack & Covas), pubescente (G. andina , G. scrobiculata (Griseb.) Tronc., G. selloi (Spreng.) Tronc.), o en algunos casos glanduloso (G. phlogiflora , G. platensis ); interiormente presenta una banda longitudinal y abaxial de tricomas agudos, adpresos y relrorsos. ANDROCEO El androceo esta formado por cuatro estambres didinamos, insertos en la mitad superior del tubo corolino, con filamentos breves a veces el par abaxial exerto, anteras ovales, basifijas; pueden presentar apendices conectivales en el dorso, los cuales se clasifican en: (1) Apendices con pie y cabeza claramente desarr- ollado: el pie puede tener distinta longitud. Los apendices en general sobrepasan la longitud de las tecas y pueden sobresalir por la garganta corolar (Glandularia selloi , G. sulphurea , G. tenera ). Eigura 1C. (2) Apendices reducidos: presentan un pie y una ca- beza, pero son de menor longitud que los anteriores, no sobrepasan la longitud de las tecas ni sobresalen por la garganta corolar (Glandularia aurantiaca , G. cabrerae , G. macrosperma) . Figura IB. (3) Apendices sin pie: en estos casos solo esta desarrollada la cabeza del apendice, la cual se posa sobre el tejido conectivo, nunca sobrepasan las tecas ni la garganta corolar (Glandularia guaranitica Tronc., G. phlogiflora , G. megapota- mica ). Figura 1A. GINECEO El gineceo presenta el estilo filiforme, mas de tres veces el largo del ovario, estigma bilobado, lobulo anterior estigmatifero, a veces apenas exerto, base no ensanchada o apenas ensanchada. Ovario bicarpelar, bilocular, con un ovulo por loculo. Numero basico de cromosomas . x = 5 (Botta, 1989). Fruto. El fruto es seco y dividido en cuatro clusas subcilmdricas de seccion subtngona, de apice obtuso o redondeado como en Glandularia platensis , G. tomophylla y G. stellarioides , o mas o menos apiculado o rostrado como en G. flava , G. lilloana (Moldenke) Botta y G. tenera . La base de las clusas es angostada y sin dicho repliegue como en G. parodii , G. andina y G. sessilis . Fenologia Las especies de Glandularia , y en general del resto de las Verbenaceae, poseen una floracion prolonga- da, desde mediados de octubre hasta mediados de abril, para el hemisferio sur, aunque muchos ejemplares que crecen en zonas templado-calidas, la floracion puede comenzar desde el principio de la primavera y prolongarse hasta comenzado el invierno. Pueden observarse flores y frutos en un mismo ejemplar y en la misma florescencia, hacia el apice las flores y hacia la base los frutos. Tratamiento Taxonomico Glandularia se ubica, dentro de la familia Verbenaceae, en la tribu Verbeneae Schauer (Tron- coso, 1974). Esta tribu ha sido redefinida por Atkins (2004: 460), autora que la caracteriza por poseer fruto seco esquizocarpico, separado a la madurez en cuatro clusas; anteras con tecas paralelas, o apenas divergentes, a menudo con conectivo dilatado; estilo 2-lobado, lobulo anterior estigmatifero. Clave para Reconocer los Generos de la Tribu Verbeneae la. Estambres fertiles 2; caliz acrescente en el fruto Hierobotana Briq. lb. Estambres fertiles 4; caliz no acrescente en el fruto. 2a. Subarbustos o plantas en cojm; inflorescen- cias generalmente no ramificadas (a veces con ramificacion simple); clusas con base angos- tada; cromosomas x = 9 o x = 10. 3a. Conectivo de las anteras no sobrepasan- do las tecas Junellia Moldenke 3b. Conectivo de las anteras sobrepasando las tecas Mulguraea N. O’Leary & P. Peralta 364 Annals of the Missouri Botanical Garden 14b. 22b. Caliz con pubescencia hispida; apendices coneclivales generalmenle sobresaliendo de la garganta corolar; clusas mayores de 3 mm long. 24a. Lamina foliar con lobulos angostamente ovados, superficie adaxial glabra o escasa- mente estrigosa; clusas con apice redondeado 24. G. selloi 24b. Lamina foliar con lobulos fili formes, superficie adaxial hispida a adpresa; clusas con apice apiculado 18. G. parodii Bracteas menores a la mitad de la longitud del caliz. 25a. Ramas adultas glabras o glabrescentes (escasamente estrigosas). 26a. Ramas escasamente estrigosas; hojas y bracteas estrigosas. 27a. Dientes del caliz tubulados; apendices conecti vales generalmente sobresaliendo de la gar- ganta corolar; clusas apiculadas 29. G. lenera 27b. Dientes del caliz filiformes; apendices conectivales no sobresaliendo de la garganta corolar, no superando la longitud de las Lecas; clusas con apice redondeado 4. G. aristigera 26b. Ramas glabras o cuando jovenes puberulas; hojas generalmente glabras o escasamente estri- gosas; caliz glabro o hispidulo solo en la base. 28a. Sufrutice; hojas no carnosas, 3-lobadas, 3- 6 5-partidas a bipinnatipartidas; corolas amari- llas, crema o blancas. 29a. Sufrutice erecto; hojas angostamente ovadas, peciolo de 5-7 mm long.; clusas con apice obtuso 3. G. araucaria 29b. Sufrutice postrado; hojas linear a ovado-elipticas, sesiles; clusas con apice rostrado 9. G. flava 28b. Hemicriptofito rastrero; hojas subcarnosas triangular, 3- 6 5-sectas, bipinnatipartidas; corola lila o violeta 22. G. radicata 25b. Ramas maduras hispidas. 30a. Raquis de la florescencia no alargado en la fructificacion; bracteas lineares a angostamente ovadas; clusas de 2—2.5 mm long 8. G. dissecta 30b. Raquis de la florescencia alargado en la fructificacion; bracteas triangulares, ovadas o angosta- mente ovadas; clusas mayores de 2.5 mm long. 31a. Par superior de estambres con apendices conectivales reducidos, no superando la longitud de las tecas; clusas 5-6 mm long 13. G. macrosperma 31b. Par superior de estambres con apendices conectivales superando la longitud de las tecas; clusas menores de 4 mm long. 32a. Hemicriptofito rastrero; bracteas triangulares de 1.5—2 mm long.; clusas apiculadas 1. G. andalgalensis 32b. Sufrutices; bracteas ovadas a angostamente ovadas; clusas con apice obtuso. 33a. Sufrutice enano (5—10 cm); caliz y bracteas hispidas a estrigosas; corola pu- bescente, lila o blanca 16. G. microphylla 33b. Sufrutice ascendente (20-25 cm altura); caliz y bracteas hispido-glandulares; corola glabra, amarilla 28. G. sulphurea 1. Glandularia andalgalensis (Moldenke) P. Peralta, comb. nov. Basionimo: Verbena andalgalensis Moldenke, Phytologia 5: 227. 1955. TIPO: Argentina. Calamarca: Andalgala, Pampa del Arenal, 2700 m, mar. 1916, P. Jorgensen 1613 (holotipo, UC no visto; isotipos, LL no visto, LL foto SI!, SI 3671!, SI 3672!, UC foto SI!; US 11866!, US foto SI!). Figura 2. Hemicriptofito rastrero, muy ramificado; ramas hispidas; entrenudos de 0.5-0. 1 cm. Hojas sesiles, laminas de 0.7—0.85 X 0.3— 0.4 cm, 3- 6 5-partidas a bipinnatipartidas, triangulares, hispidas a estrigosas en ambas superficies, lobulos lineares, el medio de 8 mm, los laterales de 4 mm, apices oblusos, margenes algo revolutos. Inflorescencia en monobotrios. Flo- resceneias en espigas densas, multifloras (a veces 2 6 3 flores), raquis algo alargado en la fructificacion; entrenudo basal de 1—2 cm; bracteas de 1.5—2 mm, triangulares, superficie abaxial pilosa a estrigosa, margen ciliado. Flores eon caliz de 7—8 mm, cilmdrico, hispido en la mitad superior y estrigoso en la mitad inferior, dientes del caliz de 0.5—1 mm, triangulares; corola de 12 mm, violeta, azulada, rosada, glabra exleriormente, fauce pubescente; par superior de estambres con apendices conectivales conspicuos, superando las tecas y sobrepasando la garganta corolar; estilo de 7.5-8 mm. Clusas de 4 mm long., cilindricas, apice apiculado, base no angostada y con repliegue basal y transversal de la pared comisural; superficie externa retieulada, superficies internas lisas. Distribucion geogrdfica y habitat. Glandularia andalgalensis es endemica del Departamento Andal- gala de la provincia de Catamarca en Argentina, crece entre los 2700 y los 2900 m. Observaciones. Glandularia andalgalensis perte- nece al genero Glandularia por ser de habito rastrero, con conectivo de menor longitud que las tecas, apendices conectivales desarrollados, estilo mas de Ires veces la longitud del ovario, inflorescencia en monobotrios y corola hipocraterimorfa; es sobre esta base que se transfiere esta especie del genero Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 365 Verbena al genero Glandularia. Glandularia andal- galensis es afm a G. microphylla y G. sulphurea por la division de la lamina foliar, la pubescencia hispida, la presencia de apendices conectivales superando la longitud de las tecas y las clusas menores a 4 mm long.; se diferencia porque G. microphylla y G. sulphurea son sufrutices, tienen hojas pecioladas, bracteas ovadas y clusas con apice obtuso. Solo se han podido estudiar dos ejemplares: el material tipo y el ejemplar de Herrera 158 (SI). El ejemplar tipo fue determinado por Moldenke en un primer momento como Verbena dissecta Willd. ex Spreng., de la cual se diferencia claramenle por el alargamiento del raquis en la fructificacion. Material adicional examinado. ARGENTINA. Cata- marca: Dpto. Andalgala, sobre ruta 47, ca. la intersection con el Rio Ingenio o Potrerillos, Herrera 158 (SI). 2. Glandularia andina (Griseb.) P. Peralta, comb, et stat. nov. Basionimo: Verbena erinoides Lam. var. andina Griseb., Abh. Konigl. Ges. Wiss. Gottingen 19: 241. 1874. TIPO: Argentina. Calamarca: Vayas Allas, alpen von Belen, ene. 1872, P. G. Lorentz 597 (holotipo, GOET no visto, GOET foto SI!; isotipos, CORD no visto, CORD foto SI!). Glandularia santiaguensis Covas & Schnack, Revista Argent. Agron. 11: 92, fig. 2. 1944, syn. nov. Verbena santiaguensis (Covas & Schnack) Moldenke, Phytolo- gia 2: 150. 1946. TIPO: Argentina. Santiago del Estero: camino entre Santiago del Estero y Catamarca, 1 abr. 1944, B. J. C. Schnack 2111 (holotipo, SI 3808!). Hemicriptofilo raslrero, ramas floriferas ascen- dentes, radicantes en la base; ramas estrigosas con tricomas simples orientados haeia la base; entrenudos de 2—4 cm. Hojas con peciolo de 0.5— 1.7 cm, lamina de 2—5 X 1.5— 3.5 cm, 5-seclada, lobulos partidos a sectados, de contorno ovado-triangulares estrigosas en ambas, lobulos de apice agudo y margen revoluto. Inflorescencia en pleiobotrios heteroteticos o mono- botrios. Elorescencia en espigas densas multifloras, raquis algo alargado en la fructificacion; enlrenudo basal de 5-60 mm; bracteas de 3-4 mm, angosta- mente ovadas, superficie abaxial estrigosa, ciliadas en el margen. Elores eon caliz de 5—6.5 mm, estrigoso a hispido, dientes ca. 1 mm, aristados; corola de 8.5— 9 mm, escasamente pubescente en el exterior del tubo y sobre los lobulos, lila claro; par superior de estambres eon apendices conectivales, inclusos o no, superando la longitud de las tecas; estilo de 6 mm. Clusas de 2 mm, cilindricas, rectas, apice redondea- do, base no angostada y con repliegue basal y transversal de la pared comisural, superficie externa reticulada, superficie interna lisa. Iconografia. Covas y Schnack (1944: 93, fig. 2), sub Glandularia santiaguensis. N ombre vulgar. Malva de vaca, en Biurrum 427 (SI). Usos. Se la considera buena forrajera para ganado lechero, del ejemplar Salas , J. C. s.n. (SI). Distribucion geografica y habitat. Glandularia andina habita en las provincias del centro-norte de Argentina. Crece en bordes de camino, pastizales y campos, entre los 400 y 4000 m. Observaciones. La descripcion original de Glan- dularia andina se basa en un ejemplar cultivado a partir de semillas traidas de Santiago del Estero y cultivadas en Mendoza. Glandularia andina es similar a G. tomophylla por la pubescencia del caliz, tamano de las clusas y apendices conectivales; ambas se diferencian porque G. tomophylla presenta la lamina con los 2/3 superiores lobado y 1/3 basal sectado, hispida en la superficie abaxial y las bracteas de mayor tamano. Puede ser confundida ademas con G. cabrerae y G. cheitmaniana , por la division sectada de las hojas; estas ultimas se diferencian principalmente por presentar un habito ereeto y por los apendices conectivales breves no superando las tecas. Material adicional examinado. ARGENTINA. Cata- marca: Dpto. El Alto, Tapso, al lado del FC, Boelcke s.n. (SI); Dpto. La Paz, entre Recreo y Paso Cruz, ruta prov. 20, Troncoso 1755 (SI); Dpto. Poman, s. loc., Biloni 6247 (SI). Chaco: Dpto. Grab Giiemes, 1 km de Castelli en direccion E, por ruta que une Castelli-El Asustado, Fortunato 1274 (SI). Cordoba: Dpto. Cruz del Eje, Soto, Nicora 17911 (SI); Dpto. Ischillm, entre Avellaneda y Los Pozos, Hunziker s.n. (SI); Dpto. Punilla, San Esteban, Nicora 1526 (SI); Dpto. Rio Primero, entre la Estacion La Posta y Canada Honda, Sayago 2132 (SI); Dpto. Rio Seco, Sebastian Eleano, Luti 4104 (SI). Formosa: Dpto. Bermejo, entre Fortin Pilcomayo y Bajo Hondo, Cordini 17 (SI). La Rioja: Dpto. Chamical, entre Santa Lucia y el empalme con ruta nac. 79, Biurrum 235 (SI); Dpto. Grab Belgrano, ruta nac. 79, cantera de Lajas, a unos 10 km de Oita, rumbo a Olpas, Pagliari 804 (SI). Salta: Dpto. Cachi, Cuesta del Obispo, Piedra del Molino, Cabrera 30726 (SI); Dpto. Chicoana, Cuesta del Obispo, Cabrera 22009 (SI); Dpto. Grab Giiemes, Rio Juramento, junto al camino, Cabrera 29965 (SI); Dpto. Iruya, camino a Iruya, Kiesling 3573 (SI); Dpto. La Vina, ruta nac. 68, 6 km al S de La Vina, camino a Cafayate, Zuloaga 7888 (SI); Dpto. Melan, ruta nac. 34, 78 km N de Metan, Cor do s.n. (SI); Dpto. Rosario de la Frontera, ruta nac. 34, pocos km N de Rosario de la Frontera, Burkart 30581 (SI). San Luis: Dpto. Gdor. Dupuy, La Mira Pampa, Cano 2629 (SI); Dpto. La Capital, Alto Peneoso, Hunziker , J. H. 6467 (SI). Santiago del Estero: Dpto. Capital, Ciudad, campo ca. Rio Dulce, Ulibarri 1013 (SI); Dpto. Copo, Monte Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 367 Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 369 Figura 3. Glandularia aurantiaca var. glabra (Hicken) P. Peralta. — A. Detalle de la insereion foliar. — B. Caliz y brae tea. A, B de A. Ruiz Leal & F. A. Roig 15723 (SI). N ombre vulgar . Y erba meona, Malaq lawogo, Tapo lawogo, en ejemplar de Vuolo s.n. BACP 2049 (LP); Margarita celeste, en el ejemplar Slurzenegger 3048 (SI). Distribucion geograjica y habitat. Glandularia aristigera crece al norte, noreste y centro de la Argentina; tambien en suroeste de Bolivia (Cocha- bamba, Santa Cruz), en tres estados del sur de Brasil (Mato Grosso, Parana y Rio Grande do Sul) y en los deparlamenlos del norte, centro y sur de Paraguay; probablemente tambien crezca en Uruguay. Habita en zonas abiertas, modificadas, bordes de camino, humedos, afloramientos arenosos, lugares suelos rocosos, hasta los 1250 m. Fenologia. Glandularia aristigera florece de octubre a marzo, aunque puede llegar la floracion a mayo. Observaciones. El holotipo de esta especie de- positado en B esta deslruido (B foto F 17403 en SI!). Por tal motivo se elige como lectotipo el isotipo depositado en el herbario BM, por ser el mas completo segun notas de la Sra. Nelida Troncoso y corresponderse con el protologo. El ancho de los lobulos foliares es un caracter muy variable, a veces son angostos como los presentes en la especie aim Glandularia tenera , de la cual se diferencia principalmente por el apice rostrado de las clusas, los dientes subulados del caliz y los apendices conectivales delgados que superan la longitud de las tecas. Material adicional examinado. ARGENTINA. Chaco: Dpto. 1° de Mayo, Colonia Benitez, Schulz 268 (CTES); Dpto. 12 de Octubre, 14 km W de Grab Pinedo, ruta prov. 94, Krapovickas 17315 (SI); Dpto. Bermejo, Las Palmas, Schulz 2888 (CTES); Dpto. San Fernando, entrada a Resistencia a orillas del Rio Negro, Diaz s.n. (SI); Dpto. Tapenaga, Villa Angela, Rodrigo 2668 (SI). Cordoba: Dpto. Grab San Martin, Villa Maria, Hicken 308 (SI). Corrientes: Dpto. Capital, 5 km E de Laguna Brava, Krapovickas 15910 (SI); Dpto. Curuzu-Cuatia, Curuzu-Cuatia, Martinez Crovetto 8400 (SI); Dpto. Grab Paz, Lomas de Vallejos, Schinini 7011 (CTES, SI); Dpto. Goya, ruta nac. 126, 2 km N de ayo. Batelito, Krapovickas 22730 (SI); Dpto. Ituzaingo, 15 km E de ruta nac. 12, camino a San Carlos, Krapovickas 18053 (CTES, SI); Dpto. Monte Caseros, Orillas arroyo Timboi, Nicora 4956 (SI); Dpto. San Martin, Carlos Pelegrini, 8 km al N, ruta nac. 14, Krapovickas 20129 (SI); Dpto. San Miguel, ruta nac. 12, 24 km E de Km 1111, Krapovickas 14872 (SI, CTES); Dpto. Santo Tome, Arroyo Chimiray, Krapovickas 26143 (SI). Entre Rios: Dpto. Colon, Palmar de Colon, sobre ruta nac. 14, Troncoso 3767 (SI). Formosa: Dpto. Formosa, alrededores de la ciudad, Guaglianone 301 (SI); Dpto. Patino, El Descanso, Filipov 53 (SI); Dpto. Pilaga, Vuoto s.n. BACP 2049 (LP); Dpto. Pilcomayo, Siete Palmas, Rojas 9011 (SI); Dpto. Pirane, Pirane, Pierotti 72839 (SI). Jujuy: Dpto. Doctor Manuel Belgrano, s. loc., 370 Annals of the Missouri Botanical Garden Volume 98, Number 3 2011 371 Annals of the Missouri Botanical Garden w,' / / / / '✓/ • /, / ' / / ; ♦ / s m m */ / m 9 • E E Figura 6. Glandularia nana (Moldenke) Tronc. — A. Aspecto general. — B. Detalle de la pubesceneia del tallo. — C. Hoja, snperficie adaxial. — D. Detalle de la hoja, superfieie abaxial. — E. Flor y bractea. — F. Corola expandida mostrando la pubesceneia y la insercion de los estambres. — G. Estilo y ovario. — H. Estambre vista ventral. — I. Estambre vista dorsal. — -J. Clusa. A-J de Martinez 9693 (SI). Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 373 Volume 98, Number 3 2011 Dpto. Las Heras, Cerros y Pampa de la Palcata, Rim Leal Cocucci 2244 (CTES, SI). Neuquen: Dpto. Alumine, s. loc., rnytoiogia V. i«i. iVOd iNomDre reemptazan e Asp 157 (BAF); Dpto. Catan Lil, a 6 km de Las Cortaderas para Verbena radicam Gillies & Hook, ex Hook, hacia Carahuila, Correa 7904 (CTES, SI). Rio Negro: Dpto. var. glabra Hicken, Darwiniana 1: 66. 1923, 377 Volume 98, Number 3 2011 Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 381 por las hojas enteras, el no alargamiento del raquis en la fructificacion, apice de las clusas y los apendices conectivales sin desarrollo de pie; G. phlogiflora se deferencia de G. guaranitica por la pubescencia hirsuto-escabosa del tallo, el mayor tamano de las bracteas florales y los dientes de caliz subulados; mientras que G. megapotamica se diferencia por presentar tallos, hojas y calices ralamente estrigosos. Tambien suele ser confundida con G . peruviana y G. tweedieana , especies que se diferencian porque la inflorescencia se alarga nolablemente despues de la antesis, por la ausencia de glandulas conectivales y por tener flores rojas. Material adicional examinado . ARGENTINA. Buenos Aires: Pdo. Ensenada, Punla Lara, Fabris 3191 (LP). Corrientes: Dpto. Concepcion, Ea. El Fortin del Ibera, Pedersen 9078 (SI); Dpto. Grab Paz, Estero Malo, ruta nac. 5, 17 km E de Caa-Catf, Carnevalli 4260 (CTES); Dpto. Ituzaingo, Ea. San Pedro, Arbo 1113 (CTES); Dpto. San Cosme, 2 km E de Santa Ana, Schinini 13002 (SI); Dpto. San Martin, Rio Aguapey y ruta nac. 40, Tressens 4112 (CTES); Dpto. San Miguel, ruta nac. 12, Km 1100, Carnevali 2621 (CTES); Dpto. Santo Tome, Ea. Francisco, 23 km NW de Gdor. Virasoro, Krapovickas 17226 (SI). Entre Rios: Sin dpto., Della del Parana, rio Ceibo, Cabrera 1962 (LP). Misiones: Dpto. Apostoles, Escuela Agrotec. P. Centilini, Cabrera 28744 (SI); Dpto. Cainguas, Predio UNLP, Reserva Valle del Arroyo Cuna Pirn, Picada desde la Est. Biol, hasta el arroyo Liso, Biganzoli 1465 (SI); Dpto. Candelaria, pasando el puente de Santa Ana, rumbo hacia el rio, Barboza 422 (CTES, SI); Dpto. Capital, Posadas, Alboff s.n. LP 8498 (LP); Dpto. Leandro N. Alem, ca. 2 km N of Cerro Azul on ruta prov. 3, Oltnslead 2004-124 (SI); Dpto. Ldor. Grab San Martin, Predio UNLP, valle del Arroyo Cuna Piru, camino abandonado entre el Balneario y la ruta 7, Biganzoli 118 (LP, SI); Dpto. Obera, ruta 105 a 23 km al W de Obera, Romanczuk 718 (SI); Dpto. San Ignacio, Jardm America, Vanni 3856 (CTES). BRASIL. Rio Grande do Sul: Pelotas, Morro Redondo, Pedersen 12592 (CTES, SI). Santa Catarina: Sao Joaquim, banks of Rio Taimbe- zinho, 1 km E of Bom Jardim da Serra (Cambajuva), Smith 10203 (SI). PARAGUAY. Alto Parana: in regione fluminis Alto Parana, Fiebrig 6452 (SI). Caaguazu: 10 km N de Caaguazu, camino a Yhu, Krapovickas 12550 (CTES, SI). Central: Estero del Ypoa, 2 km W of Pindoty, Zardini 23186 (SI). Guaira: Colonia Independencia, region de Villa Rica, Rojas 4736 (SI). Itapua: Isla Yacireta, Pin 333 (CTES). Misiones: 10 km N de Ayolas, Schinini 25980 (CTES). Paraguari: Ybytiru, Montes 12989 (LP). 11. Glandularia hassleriana (Briq.) Tronc., Darwin- iana 19: 738. 1975. Basionimo: Verbena has- sleriana Briq., Bulb Herb. Boissier, ser. 2, 4: 1056. 1904, como “ hassleranaT . TIPO: Para- guay. [Cordillera?]: “prope Tobaty in palude”, Set., E. Hassler 6464 (lectotipo, aqui designado, G no vislo, G folo SI!; isotipos, K no vislo, K folo SI!, MO no visto, MO foto SI!, NY 0138270!, NY foto SI!, P no visto, S no visto, S foto SI!, SI 3791!, US no visto, US foto SI!). Figura 4. Sufrutice de hasta 1 m de alt., tallo cuadrangular hispido, entrenudos de 6—10 cm. Hojas con peciolo de 0.8-2. 5 cm, lamina de 3-10 X 0. 7-3.5 cm, eliptica a ovada, enteras, apice agudo, base atenuada, margen serrado, ambas superficies estrigosas. Inflo- rescencia en pleiobotrios heleroteticos o monobolrios. Florescencia en espigas densas multifloras, raquis alargado en la fructificacion; entrenudo basal de 2.5— 5 cm; bracteas de 9—10.5 mm, angostamente ovadas, ciliadas, estrigosas hispida sobre la vena media. Flores con caliz de 10—11 mm, hispido, sobre las costillas, reslo glabro, dientes de 3—4 mm; corola de 18 mm, rosada, violeta o lila, externamente glan- dulosa, fauce pubeseente; par superior de estambres con apendices conectivales, menores que las tecas, sin pie, inclusos en la corola; estilo de 17 mm. Clusas de 4—5 mm, apice rostrado, base no angoslada y con repliegue basal y transversal de la pared comisural, superficie externa reticulada hacia el apice, acana- ladas hacia la base, superficie interna papilosa. N ombre vulgar . Yerba de la china, margarita amarga, flor de vovia (Moldenke & Moldenke, 1949). Distribucion geografica y habitat. Glandularia hassleriana crece en el noreste de Argentina, y en el sur de Brasil, Paraguay y Uruguay. Habila sobre campos pantanosos, arenosos y humedos, suelos pedregosos, hasta los 300 m. Observaciones. Briquet (1904b) cita en el proto- logo tres sintipos ( Hassler 5550, Hassler 6401 , Hassler 6464), se elige como lectotipo el ejemplar de Hassler 6464 , porque coincide con el protologo de la especie. Briquet en la descripcion original (1904b: 1056), describe la inflorescencia como: “. . .Flores in spicas abbreviatas terminates. caracter que no se corresponde con los ejemplares tipo vistos, en ellos puede apreciarse claramente el alargamiento del raquis en la fructificacion. Briquet (1904b), agrega a la sinonimia de esta especie a Verbena phlogiflora Chod., el cual es un nombre inexistente. Glandularia hassleriana es una especie similar a G. nana por la morfologia foliar, apice de las clusas y par superior de estambres; esta ultima se diferencia por el tamano de las clusas (ca. 6 mm vs. 4—5 mm en G. hassleriana ), bracteas de menor tamano (ca. 4 mm vs. ca. 10.5 mm) y la pubescencia de los tallos hirsula-glandulosa (vs. hispido). Material adicional examinado. ARGENTINA. Corr- ientes: Dpto. Ituzaingo, 11 km N de San Carlos, Krapovickas 24989 (CTES); Dpto. Mburucuya, Ea. Santa Teresa, Cabrera 28190 (SI); Dpto. San Roque, Ea. Caaguazu, 11 km NE de Chavarria camino a Tacuaritas, potrero Plantel, aprox. 3 km al W del casco, Arbo 6873 (CTES); Dpto. Santo Tome, Garruchos, Krapovickas 21608 (SI). Misiones: Dpto. nlSl it Bot. 1: Volume 98, Number 3 Peralta & Mulgura 407 201 1 Glandularia (Verbenaceae) en Argentina H. 19410 (3), 19487 (5b), 19521 (9), 19862 (2), 20079 (33), 20416 (2); Bartoli 106/02-2 (13), 60/02-2 (13), 98/02-2 (3); Basualdo, L 2891 (4); Batista, L. R. s.n. ICN 26838 (29); Bazzi, G. R. 230 (4); Beck, G. 191 (16), 6659 (6), 7470 (4), 9773 (19), 22163 (19); Behn, F. 8031 (3); Belgrano, M. J. 260 (19), 261 (30), 273 (29), 306 (4); Bell, M. 3 (16); Beltrao, R. s.n. SMDB 774 (19); Beorchia, A. 74 (16); Bernard, L. 499 (29); Bernardello, L. M. 706 (19), 803 (29), 811 (19); Bianco-Cantero 1209 (33); Biganzoli, F. 35 (4), 91 (19), 118 (10), 153 (4), 163 (19), 344 (10), 821 (19), 1461 (19), 1463 (4), 1465 (10), 1661 (19), 1725 (4), 1729 (30); Biloni, J. S. 6169 (33), 6247 (2), 6451 (12); Biraben, M. 3 (33), 525 (5a), 526 (5a), 532 (5a), 977 (33), 5019 (4), s.n. CTES 180546 (7), s.n. LP 906920 (32), s.n. LP 906985 (32); Bissio 239 (19); Biurrum, F. 235 (2), 427 (2), 2546 (19), 2690 (32), 3609 (33), 3628 (21), 4504 (33), 5035 (33), 5189 (16), 5242 (21), 5369 (6), 5405 (33), 5486 (33), 5624 (16), 5665 (16), 5739 (16), 6202 (9), 6240 (33), 6242 (21), 6350 (21), 6557 (21), 7040 (33), 7215 (33); Bocco, M. E. 1183 (33); Bocher, T. W. 768 (9), 1311 (5b), 2070 (28a), 2138 (3), 2145 (28a), 2395 (22) ; Bocio, M. E. 824 (29); Boelcke, O. 915 (14), 1252 (32), 1253 (29), 1272 (29), 1320 (29), 4073 (22), 4860 (32), 4923 (8) , 4968 (32), 5140 (8), 6342 (8), 9360 (8), 9924a, b, c (28a), 10139 (5b), 10774 (9), 11236 (5b), 11777 (18), 11837 (21), 13611 (3), 13620 (3), 13747 (3), 15466 (9), 15675 (3), 15730 (3), 15799 (5a), 15842 (9), 15856 (9), 15858 (9), 16003 (21), 16643 (32), s.n. (2); Boffa, P. 148 (14), 330 (8), 1059 (19), 1093 (19); Bonhers, A. s.n. SGO 42528 (28a); Bonifacio, M. 703 (5a); Bonsim, 0. 3 (19); Bonzani, N. 95 (33); Bordignom, S. 900 (4), 703 HUI 779 (19); Bordon, A. O. s.n. CTES 408675 (17); Botta, S. M. 109 (21), 252 (19), 269 (29) , 270 (19), 272 (19), 275 (32), 291 (33), 292 (32), 293 (19) , 294 (29), 299 (33), 311 (7), 330 (7), 332 (MO, SI) (7), 334 (7), 335 (7), 336 (7), 338 (7), 339 (7), 340 (7), 341 (7), 342 (7), 343 (7), 344 (7), 345 (7), 346 (7), 347 (7), 350 (7), 352 (19), 357 (19), 358 (33), 370 (22), 371 (19), 402 (18), 405 (9) , 411b (18), 412 (18), 413 (18), 417 (18), 418 (18), 455 (5a), 457 (5a), 459 (5a), 461 (5a), 519 (33), 535 (5a), 552 (5a), 554b (13), 556 (13), 563 (5a), 687 (33), s.n. SI 26671 (9); Bottino, O. J. 16 (19); Bowes, N. D, 46 (29); Brescia 5075 (30) ; Bridarolli, A. 1002 (33), 1177 (33), 2564 (20), 3042 (30), 3232 1/2 (7); Brown 284 (19), 1521 (7); Buchtien, O. 433 (16); Buck, P. s.n. PACA 11535 (20), s.n. PACA 11605 (20) ; Buenanueva, J. s.n. (33); Burkart, A. 698 (29), 1005 (25), 1375 (19), 1424 (30), 2699 (19), 2780 (21), 3082 (8), 3616 (29), 3765 (14), 4331 (14), 4715 (21), 4797 (21), 4799 (21) , 5127 (14), 5377 (16), 7017 (14), 7019 (14), 7494 (33), 8235 (14), 8474 (32), 8490 (29), 8678 (29), 9791 (29), 9884 (15), 9932 (27), 10191 (8), 10413 (33), 10414 (21), 10722 (33), 10796 (22), 10928 (21), 11003 (21), 11446 (7), 11536 (23) , 11547 (12), 11640 (16), 12087 (33), 12546 (33), 12684 (21), 12748 (29), 12764 (32), 13193 (19), 13196 (19), 13199 (2), 13201 (2), 13202 (7), 13206 (21), 13364 (29), 13462 (19), 13828 (33), 13969 (21), 14081 (10), 14250 (9), 14251 (9), 14252 (9), 14269 (13), 14272 (13), 14274 (5a), 14277 (5a), 14877 (28a), 14903 (3), 15134 (14), 15226 (30), 15838 (21) , 17570 (4), 17859 (29), 17961 (19), 18149 (29), 18372 (32) , 18389 (4), 18396 (25), 18445 (29), 18561 (4), 18839 (25), 18871 (4), 19015 (29), 19168 (21), 19221 (18), 19469 (29), 19550 (30), 19556 (19), 19658 (11), 19987 (24), 20114 (33) , 20139 (7), 20789 (33), 20846 (22), 21369 (32), 21553 (24) , 21887 (29), 21890 (19), 22072 (19), 22073 (22), 22091 (22) , 22279 (29), 22280 (32), 22281 (19), 22284 (29), 22285 (29), 22286 (29), 22288 (32), 22289 (32), 22290 (32), 22292 (32), 22293 (29), 22380 (29), 22403 (14), 22699 (SI, SP) (19), 22703 (29), 22707 (32), 22708 (29), 23094 (29), 23455 (32), 23456 (32), 23457 (32), 23458 (32), 23460 (21), 23462 (32), 23466 (29), 23795 (32), 23797 (29), 23798 (21), 23801 (29), 23802 (32), 23803 (29), 23804 (29), 23806 (29), 23807 (32), 23808 (29), 23908 (29), 24014 (29), 24230 (29), 24231 (29), 24903 (29), 25165 (24), 25171 (24), 25177 (19), 25399 (29), 25402 (32), 25409 (29), 25428 (27), 25430 (29), 25431 (SI, SP) (32), 25432 (32), 25434 (29), 25638 (29), 25639 (29), 25640 (29), 25641 (21), 25642 (29), 26344 (29), 26353 (27), 26530 (19), 26547 (12), 26566 (19), 26597 (21), 26600 (2), 26606 (2), 26615 (29), 26616 (32), 26622 (32), 27052 (29), 27055 (14), 27057 (31), 27338 (29), 27350 (29), 27356 (29), 27358 (4), 27594 (32), 27756 (29), 27757 (32), 27758 (32), 27759 (29), 28061 (14), 28062 (29), 28065 (29), 28069 (21), 28071 (29), 28075 (29), 28077 (29), 28194 (19), 28194 (21), 28417 (29), 28420 (4), 28421 (10), 28815 (19), 28816 (29), 28817 (29), 28818 (29), 28820 (29), 28821 (29), 28822 (29), 28827 (32), 28830 (21), 28965 (29), 29014 (24), 29452 (25), 29459 (29), 29461 (29), 29462 (29), 29464 (29), 30019 (29), 30093 (29), 30094 (29), 30096 (29), 30102 (29), 30103 (14), 30106 (29), 30111 (27), 30114 (29), 30118 (29), 30119 (29), 30121 (29), 30570 (21), 30571 (21), 30581 (2), 30586 (19), 30587 (32), 30588 (32), 30591 (6), 30592 (2), 30594 (2), 30603 (7), 30609 (19), 30610 (7), 30611 (7), 30612 (7), 31055 (29), 31061 (29), 31065 (17), 31069 (30), 31070 (10), s.n. (29), s.n. SI 20461 (29), s.n. SI 27706 (29), s.n. SI 27708 (29); Burmeister, C. 25 (33); Butzke, A. 7758 (20), 7892 (10); Cabezas, V. s.n. SI 23191 (16); Cabral, E. 270 (19); Cabrera, A. L. 48 (5a), 509 (14), 1021 (33), 1099 (33), 1241 (14), 1584 (14), 1618 (14), 1962 (10), 2431 (14), 3226 (8), 3388 (19), 3401 (14), 3852 (21), 4891 (14), 5179 (29), 5278 (21), 5378 (14), 5473 (21), 6570 (21), 6671 (18), 7181 (32), 7195 (29), 7312 (19), 7772 (16), 7968 (19), 8058 (29), 8459 10228 (29), 10250 (19), 10352 (8), 10828 (32),’ 11142 (3)’ 11606 (19), 11957 (11), 12105 (16), 12763 (19), 13940 (16), 14514 (23), 14780 (29), 14835 (18), 14894 (18), 15228 (16), 15388 (16), 15474 (16), 15759 (7), 15763 (7), 15813 (7), 15830 (23), 15994 (23), 16031 (7), 16074 (23), 16083 (23), 16317 (33), 16414 (33), 16671 (33), 16705 (33), 16747 (33), 16967 (12), 17128 (8), 17276 (21), 17331 (19), 17625 (16), 17660 (16), 17719 (16), 17729 (16), 18082 (33), 18338 (12), 18658 (9), 18687 (3), 18693 (3), 18767 (18), 18952 (16), 19012 (16), 19128 (9), 19390 (18), 19402 (21), 19413 (18), 19433 (18), 19458 (21), 19547 (9), 19583 (18), 20142 (12), 20393 (33), 20780 (7), 20799 (23), 20852 (27), 20908 (21), 21053 (21), 21054 (33), 21205 (9), 21903 (9), 21904 (9), 21970 (9), 22009 (2), 22057 (21), 22205 (21), 22279 (7), 22568 (21), 23866 (7), 24026 (7), 24458 (16), 24508 (16), 24631 (33), 24782 (19), 25417 (33), 25418 (32), 25437 (33), 26473 (7), 27008 (33), 27041 (16), 27141 (33), 27147 (33), 27208 (22), 27257 (33), 27451 (16), 27469 (4), 27551 (6), 27626 (29), 27656 (29), 27751 (32), 27927 (7), 27974 (12), 28044 (23), 28090 (24), 28173 (29), 28190 (11), 28364 (27), 28488 (10), 28499 (11), 28530 (17), 28555 (17), 28560 (4), 28610 (4), 28627 (19), 28643 (30), 28737 (10), 28744 (10), 28767 (11), 29016 (32), 29030 (29), 29044 (30), 29180 (4), 29254 (29), 29258 (17), 29333 (17), 29363 (4), 29513 (18), 29520 (29), 29531 (33), 29549 (33), 29622 (33), 29633 (33), 29634 (19), 29638 (21), 29665 (21), 29672 (33), 29673 (21), 29696 (22), 29735 (2), 29903 (23), 29909 (23), 29965 (2), 29973 (19), 30369 (23), 30403 (7), 30497 (19), 30574 (16), 30726 (16), 30726 (2), 30753 (33), 30795 (22), 30849 (22), 30937 (12), 30942 (23), 31066 (33), 31169 (16), 31305 (33), 31321 (7), 31433 (7), 31554 (7), 31659 (16), 31889 (2), 32070 (7), 32079 (7), 32102 (23), 32208 (23), 32214 (7), 32270 (23), 32315 (29), 32452 (16), 32813 (18), 32897 (3), 32992 (9), 33015 (13), 33051 (13), 33202 (5a), 33272 (5a), 33465 (9), 33566 (6), 33602 (6), 33637 (32), 33908 (7), (19), 502 (19); Cano, E. 490 (32), 491 (30), 491 1/2 (29), 513 immmm 10S <1«>; C«e.te, E. 35 (15), 39 (2%), 157 65 (28a), s.n. SI s.n. PACA 35908 (20), s.n. PACA 50182 R. 8 (29), 2286 51274 (19); Escobar 55 (19); Escoiaro 38 SSS3SSHSS S =5SSS:="“ SSS3S mwmwm (32), s.n. (4), s.n. (7), s.n. (19), s.n. (19), s.n. (19), s.n. (19), iliilil Volume 98, Number 3 Peralta & Mulgura 409 201 1 Glandularia (Verbenaceae) en Argentina 912 (3); Gonzales, F. 444 (19); Grass!, D. 2227 (33); 5807 (5a), 7057 (19); Guaglianone, E. R. 183 (29), 196 (32)’ 210 (21), 221 (30), 222 (30), 223 (10), 224 (29), 225 (29), 230 (29), 252 (32), 286 (17), 295 (4), 301 (4), 421 (25), 436 (4), 526 (29), 581 (32), 807 (30), 889 (19), 903 (11), 904 (26), 1314 (29), 1459 (33), 1844 (21), 2184 (32), 2186 (4), 2216 (4), 2378 (33), 2508 (30), 2508 (30), 2629 (7), 2643 (23), 3236 (3), 3283 (19); Guerrero, A. 27 (5a), 40 (5a); Guglialmelli, L. s.n. SI 27712 (29), s.n. SI 27712 (29); Guinazu, G. R. 112 (33), p.p. 226 (9), p.p. 226 (27); Guitman, M. R. s.n. LP 929176 (21), s.n. LP 929179 (19); H. A. L. 5365 (16), 6527 (28a), 7211 (18), 7255 (18); Haene, E. H. 201 (18), 561 (18), 789 (33), 1078 (3), 1758 (16), 1989 (16), 2014 (16), 2089 (16); Hagelund, K. 8139 (11), 9564 (10) , 11680 (24), 13274 (10), 14128 (17); Harley, R. M. 28001 (19), 28011 (32), 28014 (4), 28039 (11); Hartwing, W. s.n. CONC 11552 (28a); Hassler, E. 192 (32), 1396 (19), 1466 (10), 1603 (4), 1680 (25), 11052 (11), 11312 (14), 12335 (32), 12411 (4); Hatschbach, G. 1993 (19), 3067 (20), 3138 (20), 4218 (4), 4369 (10), 7444 (4), 9946 (4), 14967 (20), 15008 (20), 15041 (20), 15051 (4), 15522 (20), 18321 (20), 18368 (4), 19774 (4), 22637 (20), 23670 (20), 23852 (4), 23884 (4), 25241 (4), 25738 (20), 26081 (20), 27219 (19) , 30760 (4), 30817 (4), 34547 (4), 35172 (10), 46164 (20) , 47610 (29), 49225 (4), 49380 (19), 50552 (19), 52326 (11) , 52352 (19), 57866 (19), 58800 (4), 71822 (19), 72415 (4), 72697 (19); Heidler s.n. PACA 11252 (19); Henz s.n. PACA 25970 (19), s.n. PACA 29591 (29), s.n. PACA 32539 (19), s.n. PACA 33438 (19); Herb. Inst. Fitotecnico 9 (23); Hernonem, S. s.n. (29); Herrera, S. M. E. 158 (1), 370 (33); Herter, W. G. 19 (19), 85 (32), 1805 (8), 181a (24), 181b (24), 181c (24), 7 50 A (21), 750B (21); Hicken, C. M. 21 (21), 33 (9) , 54 (9), 74 (8), 78 (21), 228 (32), 307 (19), 308 (4), 473 (19), 474 (29), 478 (29), 491 (29), 571 (29), 891 (29), s.n. (8), s.n. (21), s.n. SI 1407 (29), s.n. SI 27704 (29), s.n. SI 27705 (29), s.n. SI 27709 (29), s.n. SI 27711 (29), s.n. SI 3345 (19), s.n. SI 3347 (32), s.n. SI 3349 (19), s.n. SI 3354 (19), s.n. SI 3355 (19), s.n. SI 3366 (19), s.n. SI 3367 (29), s.n. SI 3369 (29) , s.n. SI 3370 (29), s.n. SI 3372 (29), s.n. SI 3373 (29), s.n. SI 3374 (29), s.n. SI 3375 (29), s.n. SI 3378 (29), s.n. SI 3380 (9), s.n. SI 3434 (29), s.n. SI 3435 (29), s.n. SI 3440 (21) , s.n. SI 3442 (21), s.n. SI 3443 (21), s.n. SI 3449 (32), s.n. SI 3450 (32), s.n. SI 3571 (2); Hieronymus, G. s.n. (33), s.n. (33); Hjerting, J. R 31 (23), 9471 (22); Holmberg, E. L. s.n. SI 10467 (7), s.n. SI 10449 (19); Huajardo, E. D. de 432 (18), 1715 (15), 2548 (15), 3013 (15), 3028 (9), 3114 (9), 3212 (28a), 3738 (15), 3761 (9), 4397 (9), 4734 (33); Huidobro, A. M. R. 3493 (32), 3565 (32), 3603 (32); Humbert, H. 21218 (7); Hunziker, A. T. 1212 (7), 2302 (19), 2568 (22), 3040 (16), 4584 (18), 5094 (33), 5107 (19), 5160 (22) , 9353 (29), 16107 (29), 16418 (29), 17940 (2), s.n. (2); Hunziker, J. H. 384 (18), 2148 (16), 2160 (19), 2334 (29), 3128 (13), 6463 (21), 6464 (33), 6467 (2), 7817 (16), 8110 (16), 8132 (16), 8312 (9), 8330 (8), 9785 (2), 9897 (33), 9898 (33), 9937 (29), 9947 (20), 9950 (4), 9951 (4), 9952 (4), 9956 (30) , 9963 (4), 9966 (30), 9975 (25), 9978 (32), 9979 (29), 10475 (16), 10660 (7), 10749 (29), 11300 (13), 11504 (21), 11510 (29), 11525 (21), 11562 (21), 11829 (10), 12022 (21), 12023 (21), 12061 (21), 12072 (21), 12742 (12), 12754 (19), 12778 (12), 13071 (7), s.n. (32); Hurrell, J. 4902 (32), 4913 (32), 5207 (32), 5217 (29), 5789 (29); Ibarrola, T. 341 (30); Illin, N. 1123 (5a), s.n. LP10001 (5a); Insaurralde, I. 1139 (10) ; Irgang, B. s.n. ICN 51770 (29), s.n. ICN 81222 (20); Iudica 49 (23), 120 (16), 186 (12); Izaguirre 9584 (8), 11887 (29); J. R. 60 LP10002 (5a); Jiles, C. 730 (28a), 3459 (28a); Jimenez, B. 1362 (14); Job, M. M. 1410 (19), 1414 (8), 1898 (33), 1901 (33), 1902 (33), 2949 (19), 2973 (9), s.n. LP 903550 (19), s.n. LP 905475 (32), s.n. LP 909080 (32), s.n. LP 909216 (32); Jorgensen, R 86 (12), 1026 (33), 1028 (32), 1297 (21), 1298 (23), 2465 (29), 2466 (32), 2469 (17), 2470 (4), 2477 (25), 2636 (30), 3769 (10), 3771 (4), 3772 (32), 3773 (25), 3774 (25), 86a (12); Jozami, J. M. 143 (32); Jurado, D. s.n. SI 27707 (29); Kaufmann 4954 (21); Kegler, A. 14972 (20); Keller, H. 417 (19), 418 (30), 2830 (20); Kermes, E. 3 (27); Kiesling, R. 1044 (33), 1065 (19), 1066 (33), 1127 (19), 1202 (33), 1224 (12), 1236 (23), 1242 (2), 1295 (16), 1360 (28a), 1564 (16), 1925 (19), 1926 (21), 2939 (33), 3034 (18), 3095 (33), 3115 (33), 3127 (3), 3281 (16), 3323 (33), 3327 (33), 3347 (33), 3538 (16), 3573 (2), 3674 (16), 3724 (19), 3877 (16), 3968 (7), 4032 (33), 4139 (3), 4214 (3), 4239 (33), 4398 (33), 4439 (3), 4459 (3), 4464 (33), 4536 (16), 4546 (16), 4577 (16), 4645 (16), 4655 (13), 4708 (33), 4780 (3), 4788 (3), 4798 (33), 4847 (33), 4874 (9), 4966 (33), 4993 (21), 5004 (33), 5071 (19), 5112 (19), 5204 (16), 5357 (21), 5426 (7), 5460 (33), 5480 (7), 5481 (33), 5482 (7), 5567 (7), 5568 (7), 5569 (7), 5571 (7), 5575 (7), 5576 (7), 5578 (7), 5579 (7), 5581 (7), 5582 (7), 5583 (7), 5584 (7), 5591 (7), 5592 (7), 5593 (7), 5594 (23), 5641 (23), 5767 (12), 5768 (12), 5860 (22), 5907 (2), 5916 (33), 6064 (33), 6254 (16), 6284 (3), 6314 (28a), 6317 (33), 6324 (33), 6496 (16), 6948 (28a), 6963 (28a), 6978 (13), 7068 (16), 7700 (16), 7861 (33), 7916 (16), 7941 (16), 8045 (16), 8071 (19), 8108 (18), 8187 (7), 8282 (21), 8316 (7), 8562 (28a), 8582 (16), (13) , 10154 (13), 10226 (9), s.n. (16); Killeen, T. 4261 (6); King, D. O. 246 (19); Klein, R. M. 4293 (20), 4406 (20), 4881 (20), 5084 (20), 5084a (20); Knob, A. 5674 (4), 6142 (14) ; Koslowsky, J. 12352 (5a); Krapovickas, A. 673 (32), 1592 (23), 1624 (7), 1741 (2), 1824 (9), 1908 (33), 2324 (4), 2427 (19), 2752 (29), 2924 (21), 3086 (8), 3317 (14), 5408 (28a), 6403 (29), 6408 (19), 6428 (21), 7433 (22), 11442 (32), 11668 (19), 12134 (4), 12550 (10), 13080 (4), 13085 (30), 13087 (32), 13184 (25), 13641 (30), 14581 (13), 14872 (4), 14985 (30), 15163 (19), 15492 (30), 15588 (29), 15838 (32), 15910 (4), 16015 (4), 16298 (24), 16357 (25), 16373 (30), 16725 (30), 16791 (26), 16794 (30), 17226 (10), 17315 (4), 17388 (2), 17516 (21), 17805 (29), 17824 (29), 17972 (10), 18053 (4), 18520 (32), 18939 (19), 19272 (6), 19436 (32), 19495 (6), 19561 (32), 19932 (29), 20129 (4), 20536 (22), 21073 (30), 21181 (26), 21300 (11), 21440 (19), 21608 (11), 21609 (26), 22284 (29), 22288 (19), 22360 (21), 22425 (9), 22730 (4), 22778 (29), 22995 (20), 23981 (30), 24227 (10), 24385 (31), 24989 (11), 25274 (25), 25731 (19), 25759 (32), 26004 (11), 26039 (32), 26114 (19), 26138 (30), 26138 (17), 26141 (26), 26143 (4), 26689 (7), 26701 (7), 27068 (32), 28715 (4), 28956 (11), 29027 (10), 30520 (2), 31011 (6), 33701 (11), 36784 (20), 36832 (11), 37538 (30), 38526 (19), 38648 (7), 39050 (6), 39276 (23), 41233 (11), 44340 (4), 47471 (33), 47561 (21), 20129a (29); Kristensen 313 (7); Kublmann, M. 3717 (31); Kuntze s.n. LP 8418 (33); Kurtz, F. 5605 (5a), 9691 (28a); Lagiglia, H. A. 953 (9), 1089 (18); Laite s.n. (20); Landrum, L. R. 8627 (32); Lanfranchi, A. E. 496 (14), 628 (33), 753 (14), 818 (33), 1070 (33), 1079 (29), 1550 (21), 1567 (33), 1638 (29), 1717 (19), 1737 (29), 1852 (29); Lavia, G. 2 (32); Lee Anderson, D. 1813 (33); Legname, P. R. 4036 (23), 4432 (12), 4075c (7), 5027c (23); Leon, R. 2509 (5a); Levi, U. 3214 (28a); Lichtenstein, J. S. de s.n. (19), s.n. SI 17476 (32), s.n. SI 17479 (29), s.n. SI 18104 (19), s.n. SI 27710 (29), s.n. SI 27713 (29), s.n. (24), s.n. SI 18102 (33); Lima, D. s.n. ICN 20957 (29); Linch s.n. (29); Lindeman, J. C. 2788 (4), 3717 (4), 3772 (19); Lizer s.n. SI 3376 (29); Llamas, A. de 142 (14), s.n. BAB 1542 (10); Looser, W. 72 (33), 117 (21); Lopez, A. G. 1 (14); 410 Annals of the Missouri Botanical Garden Lopez, N. 22 (16); Lourteig, A. 487 (19), 515 (19), 2030 (19); Lurvey, E. 3 (19), 215 (30); Luteyn, J. L. 6515 (16); Luti, R. 501 (13), 4104 (2), 4493 (33), 5879 (28a); Lynelo s.n. (29); M. M? 14 (3); Malms, J. s.n. PACA 86473 (29); Maldonado, R. B. 90 (33), 419 (2), 472 (32), 833 (6), 968 (33), 1018 (33), 1436 (33), s.n. (9), s.n. LP 48417 (9); Mallo, M. s.n. BAB 18521 (3); Malvarez, M. R. 609 (7), 1365 (32); Marazzi, B. BM 265 (4); Marches! 10045 (27); Marchett, F. 404 (19); Marchionni, J. M. 8 (4); Marchioretto, M. S. 58 PACA 87049 (19); Marin, G. 249 (10); Marin, 0. 11 (23); Marq 56 (21); Marquez 36 (18); Marquez, S. 43 (5b), 102 (9) ; Martinez Achembachi, G. 899 (33), 1000 (4); Martinez Crovetto, R. 443 (8), 2738 (29), 2739 (29), 3378 (7), 4937 (29) , 5527 (29), 5965 (30), 6088 (7), 6107 (29), 6151 (9), 6350 (19), 7250 (16), 8160 (4), 8400 (4), 8862 (4), 8976 (4), 9444 (4), 9498 (4), 10174 (4), 10390 (29), 10392 (29), 11114 (19), 11227 (4), 11392 (17), g 361 (10); Martinez, A. s.n. BAA 9409 (4), s.n. BAA 9459 (32), s.n. BAA 9553 (29), s.n. BAA <>57!! |32). s.n. BU <)0<)3 (17): Martinez. L. 124<>(> 134 '(10), 294 (19); Mauri, E. 84 (8); Mazzoti s.n. Herb. Inst! Sta. Catalina 641 (2); Medan, D. 670 (21); Medinaceli, M. 3 (16) , s.n. (6), s.n. (16); Medrano, G. 7631 (18); Meglioli, C. 160 (16); Meglioli, S. 27 (33), 60 (33); Merck, H. s.n. SI 3576 (14); Mereles, F. 1428 (4), 1980 (4), 8171 (10); Meyer, T. 139 (29), 143 (32), 926 (23), 2158 (31), 2409 (29), 2673 (32) , 8685 (7), 10154 (32), 13073 (33), 13243 (33), 13425 (33) , 13753 (33), 18051 (16), 34401 (16), 141a (29), 142 (29), 680 (17); Michel, R. 95 (19), 443 (6); Millan, A. R. 300 (29), 617 (33), 665 (21), 778 (33), 779 (19); Molfino s.n. (20); Mongieri 45 (10); Montero O. G. 12282 (28a); Montes, J. E. 1747 (10), 1825 (4), 2278 (4), 2279 (19), 2454 (19), 2521 (17) ! 12445 (30), 12502 (4), 12539 (30), 12570 (10), 12609 (10) , 12851 (11), 12961 (11), 12989 (10), 15434 (17), 15449 (26), 15464 (32), 15477 (4); Monticcelli, J. C74 (21), 1-6 (9), J-8 (29); Montiel, J. C. s.n. SI 27714 (29); Moreira Filho, H. 353 (19); Morel, I. 38 (17), 199 (32), 419 (17), 789 (32), 1008 (32), 1178 (17), 1493 (17), 2003 (32); Morello, J. 4046 (29), 5289 (33), 16931 (19); Morrone, O. 120 (32), 599 (4), 1119 (30) , 1134 (19), 1663 (4), 1868 (4), 2673 (16), 3032 (12), 3154 (6), 3429 (23), 3814 (23), 4469 (7), 4492 (12), 4612 (23), 4642 (7), 5209 (29), 5224 (29), 5232 (19), 5255 (27), 5302 (29), 5323 (19), 5354 (29), 5413 (21), 5471 (21), 5492 (29) , 5749 (3), 5783 (29); Moyano s.n. LP 10003 (5a); Mroginski, L. A. 2 (19), 464 (19), 501 (32); Mulgura, M. E. 1123 (29), 1129 (19), 1135 (18), 1291 (16), 1417 ’(23), 1564 (30) , 1693 (10), 1722 (4), 2027 (30), 2057 (4), 2119 (19), 2347 (19), 2849 (4), 2911 (4), 2928 (33), 3489 (29), 3490 (19), 3746 (4), 3758 (19), 3777 (4), 3786 (30), 3795 (19), 3813 (4), 3836 (10), 4042 (4), 3670 bis (19), 3671 bis (4); Muniez, A. A. 57 (14); Munoz 1306 (19), 1405 (19), 1410 (32) ; Munoz Pizarro, C. B-157 SGO 57745 (28a); Munoz, J. de D. 1277 (32), 1313 (32), 1379 (21), 1380 (32); Miisch, P. 990 (27), 1021 (27); Najera, J. 15 (4), 40 (29); Naranjo, C. 754 (32), 755 (19), 758 (19), 760 (19), 763 (32), 765 (19), 766 (18) ! 930 (9), 931 (9); Navas, J. R.’6 (3); Nee’, M. 33656 (6), 33730 ((>i, 35388 (32). 35722 H». 38087 <0i. 3005‘> (6). (25), 834 (25); Neuman, R. 394 (23); Neumeyer, J. J. 436 (3); Nicora, E. G. 138 (29), 140 (19), 141 (21), 147 (33), 200 (29), 543 (8), 653 (10), 811 (19), 873 (19), 971 (29), 1309 (33) , 1312 (21), 1317 (29), 1471 (33), 1526 (2), 1558 (19), 1981 (29), 2007 (19), 2014 (29), 2276 (19), 2316 (21), 2317 (21), 2323 (33), 2358 (18), 2379 (33), 2432 (19), 2454 (33), 2493 (33), 2557 (33), 2593 (19), 2640 (19), 2774 (33), 2860 (29) , 3279 (19), 3281 (29), 3285 (19), 3286 (19), 3290 (29), 3541 (32), 3542 (29), 4012 (9), 4248 (33), 4257 (21), 4351 (3), 4352 (18), 4568 (29), 4636 (29), 4664 (29), 4686 (29), 4758 (29), 4816 (2), 4817 (19), 4821 (2), 4956 (4), 6937 (8), 8142 (16), 8247 (16), 8295 (16), 8414 (18), 8469 (16), 8559 (16), 8569 (13), 8817 (16), 9756 (30), s.n. SI 17611 (19), s.n. SI 17640 (33), s.n. SI 17676 (33), s.n. SI 17724 (19), s.n. SI 17785 (33), s.n. SI 17834 (33), s.n. SI 17867 (29), s.n. SI 17911 (2), s.n. SI 18560 (32), s.n. SI 18625 (21), s.n. SI 19553 (18); Niglia, A. 4 (19); Novara, L. J. 506 (7), 835 (23), 2121 (19), 5623 (2), 5673 (6), 8244 (4), 9277 (19), 9403 (21), 9418 (19); Novatis, H. S. de s.n. BAB 70223 (17); Nunez, E 1187 (25); Obregozo 397 (7); Ocampo, E. 1429 (33); O’Donell, C. A. 340 (33), 493 (33), 610 (33), 893 (33), 1026 (3), 2237 (9), 4427 (33), 4467 (33); O’Dwyer, E. B. s.n. SI 26396 (14); Offermam, C. s.n. (9); Olea, D. 278 (19); Olmstead, R. 177 (3), 196 (5a), 2004-107 (29), 2004-108 (32) , 2004-118 (4), 2004-122 (30), 2004-123 (19), 2004-124 (10), 2004-148 (18), 2004-149 (18), 2004-153 (9), 2004-156 (16) , 2004-161 (18), 2004-172 (28a), 2004-183 (15), 2004-205 (15), 2007-5 (2), 2007-31 (16), 2007-58 (16), 2007-66 (19), 2007-70 (23), 2007-73 (7), 2007-84 (33); Orbea, R. s.n. SI 18724 (19), s.n. SI 18746 (18); Osten, C. 2748 (18), 5642 (24), 13571 (10), 19188 (21), 20754 (7); Otto, A. 6445 (5a), 8138 (5a), s.n. LP 12999 (9), s.n. LP 13520 (9); Pabst, G. 6093 (20); Pagliari, E. 804 (2); Palacios, M. A. 773 (20), 1920 (29), 4236 (33), 4267 (33); Palari 749 (32); Panigatti, A. 572 (29); Parker, H. L. s.n. SI 25668 (24); Parodi, D. 61 (8), s.n. (8), s.n. (8); Parodi, L. R. 14066 (33), 14069 (33); Paseoal, D. s.n. HUI 4478 (19), s.n. HUI 4860 (19); Pastore, A. I. 136 (19), 148 (29), 361 (33), 723 (10), 1163 (29), s.n. SI 3572 (14), s.n. SI 984 (29); Pastore, F. 33 (19), 56 (21), 148 (29), 983 (14), 1163 (29), 1199 (19), 2006 (21), 2007 (19), 2035 (22), 2036 (19), 2038 (19), 2057 (21), 2075 (33), s.n. SI 1210 (21), s.n. SI 984 (29); Pastrana 9156 (7); Pedelaborde, J. L. 12099 (29), 12106 (19), 12174 (29); Pedersen, T. M. 4069 (17), 4262 (25), 4484 (30) , 5234 (11), 5344 (25), 7171 (25), 9078 (10), 9134 (30), 9176 (29), 9238 (26), 9585 (19), 9627 (25), 9923 (33), 10188 (31) , 11445 (24), 11580 (21), 11672 (32), 12388 (30), 12556 (17) , 12592 (10), 12944 (30), 13292 (9), 13998 (17), 14829 (30), 15658 (21), 15908 (20), 16228 (21); Pensiero, J. 893 (32) , 1395 (32), 1547 (27), 1702 (32), 1703 (29), 2838 (32), 3366 (29), 3499 (32), 4178 (19), 4308 (32), 4381 (6), 4510 (23); Pena-Chocarro, M. 1824 (10); Peralta, M. 106 (21), 107 (29); Pereira, E. 7689 (4); Perez Moreau, R. A. 3687 (33) , s.n. BA 31/2243 (30); Perez Moreau, R. L. 3038 (9), 3190 (9), 3278 (13), 3407 (9), 3687 (33); Perez, B. 145 (19); Perez, L. 285 (4), 2567 (4); Perrone, V. R. 30338 (3); Petersen, E. 57 (12); Petrocelli, A. 36 (9); Pf.ster, A. 7294 (3); Philippi, F. s.n. SGO 42486 (28a), s.n. SGO 42524 (28a), s.n. SGO 54740 (28a), s.n. SGO 68382 (28a); Philippi, R. A. s.n. (3), s.n. (3); Piccinni, B. 2022 (33), 2565 (19), 3850 (19); Pieroth, S. 72839 (4); Pierotti, L. s.n. LIL 100804 (6); Pierotti, S. A. 1223 (23), 4123 (17), 5152 (33), 6556 (6), s.n. (33) , s.n. LIL 154762 (33), s.n. LIL 99460 (33), s.n. LIL 99587 (33), v 108 (33); Pin, A. 333 (10); Pivetta 977 (20), 968 PACA 59156 (19); Poggio 837 (32), 849 (29); Porto, M. L. s.n. ICN 9586 (19); Pott, A. 3866 (17), 6816 (4); Pozner, R. 193 (33); Prado 56 (29), 234 (29); Prina, A. 1203 (3), 1233 (3), 1341 (3), 1467 (3), 1520 (5b), 1569 (5b), 1743 (3), 2625 (3), 2635 (3), 2655 (3), 2700 (18); Proyecto Ventania 89 (21), 186 (21), 892 (21); Pujalte, J. C. 43 (14), 121 (16); Quarin, C. 513 (19), 655 (32), 1239 (19), 1240 (29), 1887 (32), 2124 (19), 2437 (30), 2642 (25), 2797 (10), 2863 (10), 3879 (32), 3881 (30); Rabinovich, D. 130 (19), 154 (29); Volume 98, Number 3 2011 Peralta & Mulgura Glandularia (Verbenaceae) en Argentina 411 Ragonese, A. 9 (32), 49 (29), 2003 (29), 2042 (19), 2397 (19) , 2606 (19), 2842 (19), 3007 (19), 3274 (19), 5006 (29), 8775 (18), 9404 (29), 9580 (29), 9616 (33), 9830 (33); Rambo, B. 444 (20), 9366 (26), 9950 (26), 32817 (20), 34719 (4), 37693 (20), 42606 (19), 43620 (10), 43780 (20), 51652 (4), 53317 (4), 53369 (17), 54059 (20), 55939 (20), 56245 (19), 59244 (20), s.n. PACA 444 (20), s.n. PACA 1120 (20) , s.n. PACA 4488 (20), s.n. PACA 4561 (20), s.n. PACA 4775 (20), s.n. PACA 8513 (20), s.n. PACA 8781 (20), s.n. PACA 8990A (19), s.n. PACA 8999 (20), s.n. PACA 9118 (19) , s.n. PACA 9165 (4), s.n. PACA 9544 (19), s.n. PACA 9958 (19), s.n. PACA 11262 (19), s.n. PACA 1137 (19), s.n. PACA 26992 (20), s.n. PACA 28579 (19), s.n. PACA 28583 (10), s.n. PACA 29550 (19), s.n. PACA 3038 (4), s.n. PACA 3042 (19), s.n. PACA 30980 (20), s.n. PACA 31723 (19), s.n. PACA 32817 (20), s.n. PACA 34719 (29), s.n. PACA 34729 (14), s.n. PACA 36417 (19), s.n. PACA 36419 (20), s.n. PACA 37345 (19), s.n. PACA 37498 (19), s.n. PACA 37693 (20) , s.n. PACA 38555 (20), s.n. PACA 38753 (19), s.n. PACA .39020 (20), s.n. PACA 3939 (4), s.n. PACA 39592 (19), s.n. PACA 39810 (19), s.n. PACA 3982 A (14), s.n. PACA 4242 (29), s.n. PACA 4273 (20), s.n. PACA 42781 (19), s.n. PACA 42941 (19), s.n. PACA 42959 (4), s.n. PACA 43282 (19), s.n. PACA 43350 (20), s.n. PACA 4336 (20), s.n. PACA 43524 (19), s.n. PACA 43620 (20), s.n. PACA 43778 (29), s.n. PACA 43780 (20), s.n. PACA 44020 (20), s.n. PACA 44104 (29), s.n. PACA 44165 (20), s.n. PACA 44275 (19) , s.n. PACA 46335 (19), s.n. PACA 51378 (29), s.n. PACA 52102 (14), s.n. PACA 52158 (20), s.n. PACA 53317 (4), s.n. PACA 54059 (20), s.n. PACA 54786 (19), s.n. PACA 55909 (19), s.n. PACA 55939 (20), s.n. PACA 56245 (19), s.n. PACA 56659 (20); Ramirez, J, R. 342 (32), RAP 338 (32) ; Rasp, A. E. 152 (3), 177 (9), 193 (18), 194 (18), 196 (18) ; Rau, G. s.n. SMDB 267 (19); Raviolo 7 (29); Rea Clavijo, J. 46 (16); Reales, A. 617 (7), 631 (7), 1444 (7), 1525 (7), 1922 (7), 1931 (7); Reca, A. R. 15 (16), 21 (16), 41 (16); Reiche, K. s.n. SCO 68376 (28); Reitz, P. R. 155 (19), 888 (19), 2233 (19), 4445 (32), 4523 (20), 10248 (20), 11306 (20) , 11372 (20), 12526 (20), 13335 (20), 16941 (20), 6404a (20), C 882 (17), cl279 (29), cl279e (29), el280e (HBR, LIL, SI) (29), cl281 (19), C1962 (19), c882 (19); Renaudeau 100 (19) ; Rentzell, I. von 1099 (27), 1100 (29), 1101 (21), 1102 (19), 18829 (33), s.n. SI 15231 (33); Renvoize, S. A. 2852 (19), 2853 (29), 3069 (11), 3111 (20), 3149 (11), 3414 (16), 3463 (23); Ribas, O. S. s.n. (29); Ricardi, M. 3939 (28a), 5689 (3); Rimbach 23 (9), s.n. LIL 31746 (9); Rocha, R. 482 (33) , 1391 (33), 2339 (8), 2340 (19), 2437 (32), 2668 (4), 2880 (9), 3693 (19); Rodriguez 297 (22); Rodriguez, F. M. 19 (7), 30 (10), 55 (32), 297 (16), 530 (10), 548 (30), 576 (32), 578 (4), 657 (19); Rodriguez, Y, 566 (14), 907 (19); Roig, F. A. 8151 (15), 8382 (15), 8908 (9), 13056 (3), s.n. (18) ; Rojas, T. 1882 (10), 1894 (17), 1895 (17), 2527 (25), 2528 (4), 2536 (4), 3395 (19), 3628 (25), 3864 (20), 4736 (10) , 5891 (30), 7647 (17), 8374 (32), 8991 (25), 9011 (4), 11785 (30), 12111 (32), 12868 (30), 14061 (17), 14169 (25), 14395 (4), 4819a (19); Romanczuk, C, 126 (29), 418 (32), 718 (10); Rosa, E. B. 11 (33); Rosengurtt, B. 491 (25), 1004 (24), 1005 (21), 1006 (24), 1009 (21), 1015 (24), 1110 (25), 1428 (19), 2116 (24), 2133 (21), 2351 (25), 8593 (29), 8858 (11) , 10484 (29), 21161/2 (8), B 2244 (32), bl012 (19), B- 182 (24), b749 (19), B-8042 (11), B-810 (21), b891 (19), PE- 5056 PACA 32950 (21), s.n. ICN 18829 (19), s.n. ICN 18830 (19) ; Rosillo, M. A. 23 (19); Rossato, M. 5346 (20); Rossi, J. B. 850 (33), s.n. LP 918399 (33); Rossow, R. 245 (13), 1518 (18), 4735 (14); Roth s.n. (5a); Rotman, A. D. 197 (33), 204 (2), 235 (2), 252 (2), 261 (19), 423 (16), 437 (33), 578 (7), 806 (6), 819 (6), 856 (21), 868 (19), 870 (7), 882 (7), 990 (2); Royo Pallares, O. 222 (19); Rua, I. de la 76 (19); Rugolo, Z, 832 (27), 1263 (21), 1268 (29); Ruiz Huidobro, A. M, 1176 (29), 1575 (21), .3079 (6); Ruiz Leal, A. 157 (13), 1167 (18), 1284 (22), 1507 (18), 3505 (3), 3627 (5a), 3628 (3), 4686 (3), 4825 (16), 5367 (3), 6913 (9), 6936 (9), 7093 (9), 7117 (9), 7468 (9), 7500 (9), 8497 (15), 8503 (9), 8556 (16), 8650 (3), 8960 (9), 9291 (9), 9455 (9), 9655 (9), 9790 (9), 15723 (5b), 211.37 (33), 21162 (19), 21270 (16), 21343 (22), 21899 (33), 21945 (9), 22009 (5a), 22141 (28a), 22230 (15), 22276 (29), 22319 (29), 22326 (21), 23248 (16), 23482 (22), 23619 (28a), 24005 (13), 24324 (5a), 24337 (9), 24450 (3), 24515 (9), 25496 (18), 25505 (9), 25539 (13), 25582 (13), 25698 (5a), 25716 (5a), 25758 (13), 25911 (18), 26037 (29), 26039 (19), 26042 (18), 26196 (9), 26293 (32), 26323 (22), 26329 (22), 26357 (13), 26394 (13), 26605 (5a), 26715 (9), 26753 (3), 26853 (9), 27181 (15), 28142 (22), 7093a (9), s.n. (16); Ruthsatz, B. 163 (16), 203 (16), 226 (16), 239 (16), 297 (16); Saa, J. s.n. CONC 126799 (28a); Salas, J. C. s.n. (2); Salgado, C. 30 (19); Sanchez, M. I. 300 (3); Sanzin, R. 132 (28a), 139 (33); Saravia, E. 993 (4); Saravia Toledo, C. 720 (7), 1405 (32), 2837 (19), 10599 (6), 10719 (23), 12899 (21), 12956 (19), 13117 (21), 14375 (23), 14622 (23); Sayago, M. 295 (21), 409 (21), 860 (21), 1470 (33), 1655 (33), 2132 (2), 2370 (19), 2511 (29), 2663 (33), 2683 (12); Scala, A. C. s.n. LP 8496 (14); Scappini, E. 1440 (33); Schajovskoy, S. 48 (3), 138 (9), 154 (3), 1 54/1 V (3), 38/III (3), 44/IV (3), 45/IV (3), 49/111 (9), 65/IV (3), 89/111 (13), s.n. (9), s.n. SI 26740 (13); Schiiiini, A. 445 (32), 1200 (4), 1673 (4), 3489 (10), 4028 (4), 4449 (32), 4476 (29), 4664 (29), 5109 (32), 5111 (32), 5143 (29), 5251 (19), 6805 (19), 6824 (19), 7011 (4), 7012 (19), 8308 (19), 8832 (32), 8891 (32), 9224 (4), 9576 (19), 9672 (4), 9866 (19), 10300 (11), 10373 (30), 11123 (10), 11599 (32), 11600 (29), 11778 (30), 11866 (30), 12753 (4), 13002 (10), 13209 (30), 14102 (30), 14940 (19), 15408 (10), 15480 (30), 15732 (10), 15821 (10), 16587 (4), 18509 (10), 20861 (26), 20884 (25), 20898 (10), 21632 (19), 21821 (20), 24817 (19), 25241 (11), 25830 (4), 25846 (17), 25980 (10), 26070 (30), 27579 (10), 27597 (11), 28329 (25), 31285 (19), 31286 (29), 31287 (19), 31321 (30), 32979 (23), 35058 (25), 1DB (20); Schnack, B. 7 (22), 21 (9), 476 (6), s.n. SI 26401 (7), s.n. SI 26402 (27), s.n. SI 26403 (27), s.n. SI 26404 (9), s.n. SI 26406 (16), s.n. SI 26409 (30), s.n. SI 26410 (11), s.n. SI 26411 (27), s.n. SI 26412 (32), s.n. SI 26413 (32), s.n. SI 26414 (29), s.n. SI 26415 (33), s.n. SI 26416 (33), s.n. SI 26417 (27), s.n. SI 26689 (10); Schreiter 3821 (6), 5013 (7), 5149 (12), 5151 (19), 6620 (21), 6679 (19), 11128 (12), 11464 (23); Schrottky, C. B. 13 (30), 15 (4), 25 (14); Schuask, B. s.n. (14); Schult 464 (19), 466 (20), 702 (19); Schulz, A. G. 246 (32), 266 (4), 268 (4), 269 (26), 270 (17), 285 (25), 783 (32), 1468 (26), 1475 (32), 1476 (19), 2887 (19), 2888 (4), 2892 (21), 3988 (23), 5267 (23), 5299 (6), 5337 (6), 5680 (4), 5991 (21), 6469 (19), 6469 (17), 6581 (21), 6869 (19), 6894 (11), 7440 (32), 7815 (17), 7945 (30), 7989 (30), 8802 (25), 8835 (25), 8892 (4), 8926 (4), 8962 (32) , 9098 (32), 9266 (2), 10034 (19), 11309 (17), 11456 (19), 16084 (25), 16321 (17), 16397 (17), 16593 (19), 17618 (17), 18691 (10), 18725 (10), 264 a (19), s.n. (19); Scliwabe, H. 561 (33), 796 (16), 924 (16); Scur, L. 501 (19); Sehnem, A. 3698 (19), 3882 (19), 7256 (19), 10791 (20), 12443 (4), 17089 (20), 9197 C (19); Seijo, G. 1132 (23), 1383 (21), 1425 (9), 1448 (9), 1456 (9), 1745 (21), 1746 (19), 1747 (33), 1809 (33) , 2153 (18), 2255 (3), 2672 (21); Selander, R. B. 88-A-7 (16); Semper, J. 4259 (3); Shwarz 789 (30); Silva, J. M. 3617 (4); S. Coll. 55 (23), 333 (8), 735 (19), 891 (29), 1044 (29), 1639 (15), 3420 (7), 4046 (29), 5856 (5a), 107 SI 28107 (6), 1969 SI 26407 (20), 6181 LIL 31349 (32), s.n. (16), s.n. (19), s.n. (4), s.n. (8), s.n. (9), s.n. (19), s.n. (19), s.n. (19), s.n. (32), 412 Annals of the Missouri Botanical Garden CNPO 1817 (20), s.n. CNPO 2865 (19), s.n. ICN 28852 (20), s.n. SGO 54747 (28a), s.n. SGO 42534 (15), s.n. SGO 54796 (15), s.n. SI 26401 (7), s.n. SI 3371 (29), s.n. SI 3379 (9), s.n. SI 3405 (15), s.n. SI 26419 (29); S. H. 614 (6); Slanis, A. 155 (12) , 167 (19), 193 (33); Sleumer, H. 2694 (32), 3542 (12); Smith, L. B. 8478 (20), 10203 (10), 11638 (20), 12592 (20), 13223 (20); Sobral, M. 463 (4), 1718 (30), 2018 (4), 2668 (20) , 3343 (21), 4439 (21), 6504 (20), 7697 (11), 8096 (20), 8981 (20), 9282 (29); Solbrig, 0. 156 (19); Solis Neffa, V. 639 (6), 813 (21); Solomon, L. C. 9996 (23); Sonego, R. s.n. HUI 5297 (19), s.n. HUI 5299 (19); Soriano, A. 482 (21), 668 (16), 710 (7), 862 (19), 927 (19), 1314 (3), 2121 (13), 2146 (13), 2147 (13), 2148 (13), 2270 (13), 3217 (5a), 3416 (13) , 3736 (13), 3872 (13), 4042 (9), 5049 (5a); Soza, V. 1813 (32) , 1819 (6), 1830 (16), 1842 (33); Spaholi, J. 39248 (29); Spegazzini, C. 10216 (10), 10298 (29), 10336 (32), 10410 (19), 10445 (25), 10559 (7), 10768 (7), s.n. (10), s.n. (19), s.n. s.n. (33); Spies s.n. PACA 37030 (19), s.n’. PACA 63207 (29)! Stehmann, J. R. 684 (19); Steibel 762 (29), 1059 (29), 3077 (9), 3363 (29), 3972 (18), 5167 (32); Steinbach, J. 2533 (4), 2673 (4); Strehl, T. 179 MPUC 638 (19); Stuessy, T. F. 6906 (13) ; Sturzenegger, O. BACP 2433 (17), BACP 2490 (17), s.n. BACP 3048 (4); Sxur, L. 157 (20); Taylor, C. M. 11459 (7); Tell s.n. (19); Teodoro Luis, I. 6 (ICN, SI) (19); Terribile, M. 157 (32), 366 (32), 448 (32), 476 (32), 492 (32), 739 (33); Theissen s.n. PACA 7848 (20), s.n. PACA 7850 (19), s.n. 25298 (20), s.n. PACA 25190 (19), s.n. PACA 25213 (19); Tirel, C. s.n. (22); Titos 770 (32); Tiyano 644 (29), 644a (33); Tolaba, J. A. 1957 (19), 2312 (6), 3455 (6); Tombesi, T. S. 37 (33); Torena, M. 1010 (2); Torgo, F. 21259 (19); Torres, A. 144 (20), 1017 (21); Torres, B. s.n. LP 918528 (16), s.n. LP 918537 (16); Torres, F. s.n. CONC 25157 (28a); Torres, L. M. 50 (9), 51 (28a); Trelles, J. 3352 (19); Tressens, S. G. 21 (19), 135 (32), 482 (10), 1713 (10), 1782 (30), 1874 (11), 1917 (26), 2677 (10), 3126 (25), 3819 (10), 4112 (10), 4113 (30), 4211 (25), 4362 (25), 6563 (10); Troiani, H. 3446 (19), 15021 (9), 15052 (13), 15107 (3), 15121 (9), 15813 (3), 15947 (5a); Troncoso, N. S. 128 (29), 170 (29), 172 (8), 179 (14) , 190 (29), 201 (19), 291 (29), 318 (19), 386 (21), 1063 (29), 1157 (29), 1219 (29), 1441 (29), 1652 (29), 1690 (29), 1755 (2), 1775 (2), 1791 (33), 1855 (33), 1921 (33), 1966 (29), 2055 (29), 2080 (29), 2207 (29), 2346 (32), 2641 (27), 2651 (27), 2768 (29), 2978 (29), 3036 (19), 3335 (27), 3767 (4), 3868 (29), 6022 (29), 19528 (33), 19532 (33), 19535 (33) , s.n. SI 17859 (29), s.n. SI 19530 (32), s.n. SI 20587 (21) , s.n. SI 27089 (29), s.n. SI 27116 (29), s.n. SI 27689 (8), s.n. SI 27711 (29); Ulibarri, E. A. 623 (19), 916 (33), 1013 (2), 1384 (33), 1479 (16); Ulibarri, U. 1780 (4); Umana, A. 26 (10); Urbani, M. 76 (32); V. Dellamea, M. 2939 (19); Valencia, J. s.n. SI 14924 (29), s.n. SI 14925 (19); Valentini, A. 47 (19), 501 (19), 503 (29); Valla, J. J. 18521 (3); Vallerini, M. C. 196 (3), 1488 (13), 1530 (13), 2080 (5a), 2156 (5a), 3571 (5a), 3630 (13), 3779 (5a), 3806 (13), 6339 (13); Vanni, R. 571 (10), 737 (19), 1203 (4), 1558 (19), 1573 (10), 1842 (4), 3041 (32), 3056 (10), 3856 (10), 4292 (19); Varocca, J. 17322 (23); Vattuone, I. C. 60 (14), 102 (19), 132 (21), s.n. (19); Vavrek, D. 287 (32); Vegetti 1006 (29); Velho, A. 16844 (20); Venturi, S. 4 (2), 14 (29), 378 (19), 1042 (2), 1930 (2), 2303 (2), 2651 (7), 2898 (7), 2922 (7), 3568 (12), 3579 (7), 3677 (33), 4063 (12), 4762 (2), 5027 (7), 5056 (23), 5132 (12), 5143 (21), 5264 (7), 5270 (19), 5329 (12), 7163 (2), 7164 (19), 7840 (33), 7976 (21), 8356 (12), 8777 (16), 9389 (16), 10752 (23), 14 (2da serie) (29), 33/ 6574 (32); Vidal, M. I. 44 (5a); Vignati, M. A. 158 (33), 196 (33), 289 (33), 290 (33), 524 (33), 592 (33), 969 (33), 7042 (21), 7066 (19), s.n. SI 14926 (16), s.n. SI 7060 (33), s.n. SI 7075 (15), s.n. SI 7077 (33), s.n. SI 7089 (15), s.n. SI 7089 (15), s.n. SI 7091 (21); Villa, E. 453 (32), 680 (32); Villafane, M. 323 (33); Villagran, C. 7808 (3); Villamil, C. B. 2046 (21), 2146 (21), 2395 (21), 2398 (29), 2489 (29), 2573 (19), 2580 (21), 2585 (21), 2639 (29), 2725 (9), 3131 (19), 3155 (21), 3232 (9), 4348 (29), 4382 (19), 4422 (19), 4618 (29), 7131 (29); Volponi, C. R. 26 (13), 307 (28a), 360 (21), 377 (32), 592 (33); von Ihering, H. 373 SP 20071 (4); von Rentzell, S. s.n. SI 19219 (33); Vuoto, R s.n. BACP 2049 (4); Walmich, C. 2244 (2); Wasum, R. 208 (19), 1178 (20), 1557 (20), 2042 (20), 2933 (19), 3298 (20), 6433 (19), 7391 (24), 8051 (20); Werner, D. 639 (16); West, H. s.n. LIL 281948 (7); Wolf 49 (32), 156 (4); Woolston, A. 306 (25), 731 (32); Xifreda, C. C. 343 (19), 5149 (19); Zabala, S. 285 (19); Zachia, R. 2749 (19), 2903 (19); Zanhali s.n. HAS 39248 (29); Zardini, E. M. 153 (21), 210 (8), 524 (19), 611 (29), 957 (4), 1145 (21), 2553 (4), 2730 (19), 3071 (19), 5359 (32), 6930 (4), 7115 (32), 7486 (30), 10808 (4), 13380 (32), 13773 (32), 14196 (4), 14217 (32), 14283 (32), 14534 (32), 14586 (32), 14887 (32), 15293 (32), 15564 (32), 15957 (19), 16127 (4), 16149 (25), 16237 (4), 16240 (32), 16417 (4), 16447 (4), 16685 (32), 16877 (32), 17202 (32), 17366 (4), 17441 (4), 17443 (4), 18128 (32), 18260 (32), 18567 (10), 18598 (4), 18633 (4), 18674 (32), 18836 (4), 19239 (32), 19999 (32), 20155 (19), 21312 (32), 21564 (32), 21774 (19), 22127 (32), 22263 (4), 22491 (4), 22773 (32), 22899 (4), 22901 (32), 22968 (32), 22972 (4), 23020 (32), 23024 (4), 23099 (32), 23186 (10), 23255 (32), 23317 (4), 23388 (32), 23389 (4), 23448 (32), 23543 (4), 23571 (4), 23577 (32), 23824 (32), 24191 (4), 24201 (32), 24322 (4), 24439 (32), 24442 (4), 24533 (4), 24649 (30), 24654 (19), 24823 (19), 24939 (4), 24943 (4), 25053 (4), 25333 (4), 27284 (32), 27646 (32), 27692 (32), 29395 (19), 58589 (19), 58674 (32), 58679 (11), 58696 (4), 59129 (4), 59148 (20), 59221 (10), 59380 (19), 59482 (29), 59727 (29), 59866 (4), 59910 (32), 59970 (19), 60255 (19), 60286 (19); Zoller, O. 5417 (13); Zuloaga, F. O. 589 (10), 645 (19), 646 (30), 968 (19), 1016 (11), 1254 (19), 1637 (23), 2321 (29), 2509 (7), 2565 (23), 2662 (7), 2726 (6), 3120 (29), 3252 (10), 3308 (30), 3649 (7), 3653 (7), 3662 (19), 4556 (6), 4912 (11), 5779 (19), 6012 (16), 6054 (16), 7625 (7), 7884 (19), 7888 (2), 8099 (4), 8103 (30), 8393 (4), 8420 (2), 9108 (12), 9161 (16), 9415 (2). CARACTERISTICAS Angelica Ramirez-Roa2 y Gerardo Varela ANATOMICAS DE HOJA Y FLOR Hernandez* CON IMPORTANCE TAXONOMICA PARA LA DELIMITACION DE CUATRO ESPECIES EN EL GENERO MOUSSONIA (GESNERIACEAE)1 416 Annals of the Missouri Botanical Garden Skog (1976), en su estudio sobre la tribu Gesnerieae, reporta e ilustra la presencia de domos estomatales y la morfologia de tricomas no glandu- lares. Para Rhytidophyllum tomentosum (L.) Mart, describe tricomas no glandulares uniseriados y simples, rodeados en la base por un anillo de celulas tricomas glandulares. Uno de los aspectos impor- tantes incluidos en el trabajo de Skog (1976), es la posible interpretacion ecofisiologica de la anatorma de las especies en relacion al ambiente. Este autor indica que en Rhytidophyllum Mart., los domos estomatales mas altos se encuentra en especies densamente pubescentes, que ocupan taludes ex- puestos y secos o en claros del bosque. Con base en estas observaciones, Skog (1976) sugiere que los domos estomatales permiten una transpiracion mas indumento denso de la superficie del enves. Las interpretaciones de Skog (1976) se contraponen con las de Wiehler (1970), quien correlaciono los domos estomatales con ambiente sombreados, o con las de Haberlandt o Biebl y Germ, quienes correlacionan los domos estomatales con habitats sombreados y humedos (revisado en Skog, 1976). Por otro lado, Skog (1976), con base en observa- ciones en Gesneria L., indica que la intensidad de luz podrfa determinar el niimero de capas celulares en el mesofilo y la diferenciacion del parenquima en empalizada y del tejido esponjoso. Este autor reporta, que en plantas de G. exserta Sw. creciendo en areas al descubierto y en pleno sol, observo un mesofilo claramente diferenciado en un parenquima en empalizada grueso, compuesto de uno a tres substratos celulares y un tejido esponjoso definido por varios substratos de celulas isodiametricas. Por el contrario, en plantas de G. reticulata (Griseb.) Urb. poco expuestas a sol directo, Skog (1976) observo una sola capa de parenquima en empalizada y una o dos capas de tejido esponjoso. Kvist y Skog (1992), en su revision del genera Kohleria, describen algunas caracterfsticas anatomi- cas del haz y del enves de la hoja que ilustra con (MEB). En este estudio los autores reportan los dos tipos de tricomas glandulares mencionados por Wiehler (1983) y los domos estomatales. Ademas, indican que los domos estomatales mas altos se encontraron en plantas con indumento denso, y explican que la elevacion de los estomas en domos estomatales facilitan el proceso de transpiracion, al permitir que los estomas se ubiquen por arriba de la capa de aire y humedad, la cual no se mueve debido al indumento. Previos estudios en K. spicata (Kunth) Oerst., especie que vive en habitats con exposition solar directa, muestra domos bien desarrollados contrastando notoriamente con los domos bajos de las especies del genera que crecen en condiciones de sombra. Kvist y Skog (1992) mencionan que aunque entre las especies del genera el indumento es variable, dentro de cada especie la variabilidad es consLanle. Rodrfguez-Flores y Skog (2008) al revisar el genera sudamericano Corytoplectus Oerst., utilizan la colo- ration del haz de las hojas como uno de los caracteres distintivos del genera y diferencias en los tipos de tricomas e indumento para delimitar especies. De hecho, en su revision taxonomica del genera, la clave de determination de especies, comienza precisa- mente con las diferencias en el indumento del haz de las hojas. Posteriormente, Ranurez-Roa et al. (2009) (endemica de Mexico) de Corytoplectus, confirman la importancia y utilidad del tipo de indumento y la superficie de la hoja para incorporar la especie en el genera y reconocer su afinidad con otras especies. Ademas, muestran las primeras imagenes del haz y el Tanto Denham (1949) como Wiehler (1983) llegan a la conclusion de que el tipo de tricomas en las gesneriaceas no varfan dentro de una misma especie, por lo que se pueden usar estas caracterfsticas para distinguirlas. Sin embargo, encontraron que el ambiente puede influir en la densidad y distribution de ellos en los organos de la planta. En lo que se refiere a las especies del genera Moussonia, existen pocos trabajos en donde se mencionan las caracterfsticas de la epidermis y sus derivados. Uno de ellos es la tesis doctoral de Miriam Denham (1949) en donde la autora hace una revision de literatura sobre los tricomas, y menciona solamente a M. deppeana y M. elegans. Denham (1949) menciona que Weiss en 1867 describe los tricomas de la primera especie como uniseriados, puntiagudos (senalados como tfpicos en la familia Gesneriaceae), con la base del tricoma bulbosa parcialmente sumergida en un collar de celulas epidermicas, y tricomas glandulares pequenos con una celula en el eje y con dos a cuatro celulas en la cabeza, las cuales secretan una sustancia ligera- Denham (1949), senala que en M. celulares de los tricomas estan abruptamente alargadas y, como consecuencia, se ven gemculadas. Otro trabajo que presenta datos sobre la anatomia de la epidermis de Moussonia, es la tesis doctoral de Hans Wiehler (1970), publicada en 1983. Este autor ;pidermis de los domos estomatales y Volume 98, Number 3 2011 Ramlrez-Roa & Varela Hernandez 41 7 Caracteristicas anatomicas en Moussonia estomas de M. deppeana, M. elegans y M. hirsutissima (C. V. Morton) Wiehler, y un tricoma glandular con el pie y las celulas epidermicas adyacentes en M. elegans (Wiehler, 1983: 92-93, figs. 176, 182-183) y senala que, M. deppeana y Kohleria tubiflora (Cav.) Hanst., presentan los domos estomatales mas altos y empalizada claramente diferenciados (Wiehler, 1983: 99, fig. 213). Wiehler (1983: 92-93, figs. 176, 182) las hojas tiene domos estomatales muy evidentes. Para M. elegans Wiehler (1983: 99, fig. 213) ilustra una de tres celulas, siendo la ultima, puntiaguda, y en la base alrededor del tricoma seis celulas epidermicas que elevan al tricoma solo un poco por arriba del nivel general de la epidermis. Las celulas de la base del tricoma, solo se diferencian del resto grandes y planas. Sin embargo estas celulas son diferentes en generos como Columnea L. (seccion Dalbergaria Tussac, tribu Episcieae), Kohleria, Achimenes Pers. y Capanea Decne. de la tribu Gloxinieae (Wiehler, 1983: 92-99, figs. 91, 93, 214-215), en donde son mucho mas grandes que el res to de las celulas epidermicas, largas y de apariencia redondeada y turgentes. Por ultimo, Ramfrez-Roa (2007a, 2007b) ha senalado las diferencias en la superficie del haz y enves de las hojas, tipos de tricomas y celulas en la base del pie de los tricomas, presencia de domos estomatales, y diferenciacion del parenquima en empalizada en algunas especies de Moussonia como M. adpressipilosa D. L. Denham ex Ramirez Roa, M. hirsutissima, M. larryskogii Ramfrez-Roa y M. strigosa (C. V. Morton) Wiehler. Ramfrez-Roa ha incluido, en sus estudios del genero, claves de identificacion que utilizan las caracteristicas morfo- caracteres han sido utiles para reconocer tanto a las especies de Moussonia ya mencionadas, como a M. ampla L. E. Skog, M. elegans, M. rupicola (Stand. & L. O. Williams) Wiehler, M. serrulata (C. V. Morton) Wiehler, M. fruticosa (Rrandegee) Wiehler y M. viminalis (Brandegee) Wiehler. Materiales y Metodos Las especies que se mcluyeron en este estudio fueron seleccionadas tomando en cuenta los si- guientes critenos: 1) que pudieran mostrar la variation de los caracteres anatomicos conocida hasta el momento para el genero Moussonia; 2) incluir a las dos especies con problemas de delimitacion, M. deppeana y M. elegans; 3) mostrar las posibles diferencias entre una de las especies con amplia sinonimia, M. elegans y M. jaliscana (S. Watson) D. L. Denham ex Ramfrez-Roa comb. nov. [= Isoloma jaliscanum S. Watson] uno de sus sinommos. Las observaciones preliminares y generales se hicieron exclusivamente en ejemplares de herbario, utilizando microscopio de diseccion (MD), con el objeto de reconocer los caracteres que por su constancia, pudieran servir para la delimitacion de las especies aquf estudiadas. De todos los ejemplares de herbario observados, solo se indican algunos representatives de cada taxon, los cuales estan citados en las descripciones correspondientes que se incluyen mas adelante. Una vez reconocidos los caracteres anatomicos y morfologicos que se tomarfan en cuenta para cada especie, se procedio a tomar las muestras para las observaciones anatomicas en el MER, con el fin exclusivamente de aumentar la resolution de las observaciones, observar con mas facilidad las estructuras, y complementar su description. Los ejemplares estudiados con el MEB, provienen de la coleccion del Herbario Nacional de Mexico (MEXU), y fueron los siguientes: Moussonia ampla (Panama. Van der Weiff 7255, MEXU-250256), M. deppeana (Veracruz, Mexico. Ventura 18120, MEXU-751039), M. elegans (Chiapas, Mexico. Breedlove 48663, MEXU-778741) y M. jaliscana (Jalisco, Mexico. Pringle 11072, MEXU-29588). Las muestras para ser observadas al MEB fueron tomadas en la parte media de hojas maduras, recortando dos rectangulos aproximadamente de 2 X 2 mm a partir del margen hacia la vena media, con el fin de tener representado tanto el haz como el enves. De la corola en antesis se tomaron porciones equivalente en tamano, ubicadas cerca de la garganta y hacia la base de la corola con el fin de observar la superficie adaxial y abaxial de la corola, tanto en la parte dorsal como en la ventral. Las muestras fueron observation en MEB, Hitachi S-2660N (Tokio, reportan en micrometros. El material de herbario observado en MEB no fue las micrograffas los tricomas, estomas y celulas en general tal como se observarfan con el microscopio de diseccion. Ademas como es sabido el trabajo taxonomico y el uso de las claves se realiza, la mayorfa de las veces, con material de herbario, esto es, deshidratado. La terminologfa morfologica se Volume 98, Number 3 2011 Ramirez-Roa & Varela Hernandez 419 Caracteristicas anatomicas en Moussonia Figura 1. Enves cle la hoja: lipo de relieve y domos. — A. Moussonia deppeana , relieve casi piano, estomas solo ligeramente elevados por los domos estomalales casi pianos. — B. M. elegans , relieve y estomas como el anterior. — C. M. ampla , relieve ampollado, estomas elevados evidentemente, ubicados sobre los domos en forma de ampollas; tricoma uniseriado glandular al frente y dos tricomas tipo hongo del lado izquierdo, observandose de ellos solo la glandula o la glandula con el pie. — D. M. jaliscana , relieve sinuoso, estoma sobre las ondulaciones. A de Ventura 18120 (MEXU 751039); B de Breedlove 48663 (MEXU 778741); C de van der Werff 7255 (MEXU 250256); D de Pringle 11072 (MEXU 29588). Figura 2. Corte transversal de la hoja. — A. Moussonia elegans , separacion de la epidermis del enves del parenquima esponjoso, a la altura del domo estomatal, poco evidente; no se aprecia claramente el parenquima en empalizada. — B. M. ampla , separacion de la epidermis del enves del parenquima esponjoso, a la altura del domo, muy evidente; parenquima en empalizada presente. Epidermis del enves hacia arriba, epidermis del haz hacia abajo, mesofilo entre las capas de epidermis, de la A de Breedlove 48663 (MEXU 778741); B de van der Werff 7255 (MEXU 250256). 420 Annals of the Missouri Botanical Garden de la hoja, pudiendose confundir con particulas de cera y en ocasiones, no son faciles de reconocer. En todos los taxa revisados en este trabajo bajo el MEB se observaron las cabezas de estos tricomas las cuales estomas (Fig. 1C del lado izquierdo). El tercer tipo es Densidad aparente de tricomas y estomas y tamano preliminar de los estomas. Las especies con relieve piano muestran una menor densidad aparente de tricomas (Fig. 1A, B) que las especies con relieve ampollado o sinuoso (Figs. ID, 2B). En lo que respecta a los estomas, tambien se observaron diferencias muy evidentes entre las especies, pero no dentro de cada especie. Por ejemplo, en Moussonia deppeana los estomas en MD se aprecian como “monticulos amarillentos” muy evidentes, los cuales estan densamente arreglados en la superficie; apreciandose incluso las celulas oclusivas y los ostiolos en algunos casos. Por el blanquecinas, laxamente distribuidos. En M. jalisca- na, con relieve sinuoso, se aprecia la cercama de los ampla los estomas se distribuyen sobre el domo estomatal, MEB indican que M. deppeana y M. ampla presentan el mayor y el menor numero de estomas por area, respectivamente, mientras que M. elegans y M. jaliscana muestras una densidad de estomas inter- media. La generalidad de este reporte sin embargo debe corroborarse con un la observation de un muestreo mas amplio. En cuanto al tamano de los estomas, M. deppeana tiene los estomas mas grandes que son de aprox. 40 pm, mientras que M. ampla son de aprox. 28.12 pm, en M. elegans aprox. 26.29 pm, y se observan los estomas mas pequenos en M. jaliscana con aprox. 24 pm. En MD los contornos de las paredes anticlinales de las celulas epidermicas se aprecian mas o menos claros mientras que en las paredes periclinales pueden ser papilosas proyectandose como muy pequenas protuberancias sin distinguirse mayor detalle. Cabe mencionar que la corola puede ser translucida cuando el grosor de la corola es delgada, como en Moussonia ampla o no translucida si la corola es gruesa como en las demas especies. Con base en MEB se describen cuatro tipos de celulas epidermicas, que se describen considerando las variaciones morfologicas de las paredes anti- clinales y periclinales de las celulas: 1) celulas con paredes anticlinales formando poligonos irregulares, organizadas densamente y con paredes periclinales papilosas (Fig. 3A); 2) celulas con paredes anti- clinales formando rectangulos irregulares y paredes periclinales lisas (Fig. 3B); 3) celulas con paredes anticlinales formando poligonos irregulares — con cinco o seis lados y paredes periclinales lisas (Fig. 3C); y 4) celulas con paredes anticlinales formando poligonos irregulares y paredes periclinales con una protuberancia angular — proyeccion tipo pico curvo (Fig. 3D). Moussonia jaliscana y M. deppeana muestras diferentes tipos celulares en la portion dorsal y ventral del tubo de la corola. En M. jaliscana la parte dorsal muestra el tipo celular 1 y la parte ventral el tipo celular 3 mientras que M. deppeana dorsalmente presenta el tipo celular 3 y ventralmente el tipo celular 4. En contraste M. ampla y M. elegans tanto respectivamente. Superficie del enves, domos estomatales y meso- filo. Las especies examinadas son arbustos terrestres de hojas generalmente delgadas y pubescentes, que se establecen en humedos de montana o en encinares humedos. Debido a la escasa information de las etiquetas de campo, es incierto si estas especies crecen sobre laderas sombreadas y humedas o en laderas expuestas y secas. De acuerdo con Wiehler (1983), especies c estomatales en la superficie del enves. De las especies estudiadas, Moussonia deppeana y M. elegans tienen hojas con indumenta denso y ampla con los domos evidentes muestra hojas con un 3 denso y un mesofilo con paren- quima esponjoso y empalizada bien definidos y grueso. Aunque las caractensticas morfologicas de M. ampla corresponden con la description de Skog (1976) y Kvist y Skog (1992) para especies que se establecen en ambientes mas expuestas, se desco- noce si la especie vive en areas sombnas o expuestas. Sin embargo otros estudios indican que la ilumina- Volume 98, Number 3 2011 Ramlrez-Roa & Varela Hernandez 421 Caracteristicas anatomicas en Moussonia paredes periclinales lisas. — -D. M. deppeana, con pliegue curvo. A de Pringle 11072 ( MEXU 29588); B de van der Werff 7255 (MEXU 250256); CyDde Ventura 18120 (MEXU 751039). cion intensa y deficiencia hfdrica podrh grosor del tejido en empalizada, seguramente deter- minando el incremento de la actividad fotosintetica (Fahn, 1978). Observaciones directas en el campo sugieren que cambios de humedad asociados a diferentes tipos de vegetacion, podrfa estar influyendo mas en las incidencia solar como menciona Skog (1976). Esta situacion podrfa estar ocurriendo en Moussonia larryskogii y M. hirsutissima (Ramfrez-Roa, 2007b). Ambas especies generalmente ocupan laderas prote- gidas en ambientes con diferentes regfmenes de humedad, M. larryskogii caracterizada por el enves casi piano crece en bosque tropical perennifolio (vegetacion con altos niveles de humedad), mientra crece en bosque humedo de montana es decir en una vegetacion con menos humeda que la anterior. Cabe aclarar que aunque las dos especies pueden ; los dos tipos de vegetacion en caracteristicas anatomicas. Aquf se reconocieron diferencias morfologicas utiles para distinguir Moussonia dep- peana de M. elegans, sin embargo, estos caracteres son informativos siempre y cuando se consideren ambas especies en sentido estricto. Las caracter- fsticas morfologicas que este par de especies pudieran tener en comun se explica por la ambigiie- dad para definir los lfmites morfologicos de M. elegans, problema que se atribuye a la amplia de M. elegans, al incluir varias especies en sinonimia (Wiehler, 1975a). A continuacion se presentan los caracteres anatomicos utiles para la distincion de estas dos especies. Los domos estomatales de Moussonia deppeana son muy visibles tanto en MD como a simple vista, estan relativamente juntos unos de otros, ocupando toda la superficie de la hoja, y presentan un color verde- r tamano y por li s evidentes 422 Annals of the Missouri Botanical Garden que los de M. deppeana. Ademas los domos estomatales en M. elegans estan dispersamente distribuidos y son de color verde palido, tal como el enves. A1 comparar el enves de las hojas en ambas especies en el MEB, M. deppeana presenta los domos estomatales mas grandes y mas densamente distri- icas, las especies tambien se pueden distinguir facilmente, M. deppeana tiene un indumenta veluti- no, en varias partes de la planta, incluyendo la corola, los lobulos de la corola son rectos y los lobulos del caliz son deltados a triangulares, mientras que M. elegans en sentido estricto, tiene un indumento viloso-piloso, corola laxamente pilosa con los lobulos expandidos y lobulos del caliz lanceolado subulados. Tambien se observo que la caracteiizacion de domos y mesofilos previamente descritos para Mous- sonia deppeana y M. elegans por Wiehler (1983), difiere de la nuestra. Nuestros estudios indican que los domos estomatales para M. deppeana ilustrados por Wiehler (1983, fig. 176) se parecen mas a los que aquf trabajo, el mesofilo observado para M. deppeana difiere de nuestras observaciones, Whieler (1983) presenta para esta especie un mesofilo con el parenquima en empalizada y esponjoso bien diferen- ciado con multiple capas, mientras que nosotros observamos un mesofilo con el parenquima de pocas capas celulares, distinguiendose solo con certeza, el tejido esponjoso. En cuanto a M. elegans, Wiehler (1983) muy probablemente pudo haber observado domos altos en alguna de las especies que el consideraba como sinonimos. Por ultimo, aquf reportamos diferencias entre M. elegans y M. jaliscana una de las especies incluidas en su sinonimia. En el enves de la hoja, el relieve piano de M. elegans, los domos estomatales pianos y los estomas grandes ay u dan a distinguirla de M. jaliscana. Ademas, la primera especie tiene, en la superficie adaxial de la corola celulas con paredes periclinales lisas y la corola i segunda las paredes son i, entre otras caracterfsticas. TRICOMAS Y CELULAS EPIDERMICAS ADYACENTES Moussonia elegans y M. jaliscana tambien se diferencian por los tricomas y las celulas alrededor de la base de los tricomas, la primer especie presenta en la base del tricoma un anillo de ocho a nueve celulas con paredes periclinales lisas y la segunda, un anillo de cinco celulas con paredes periclinales turgentes. Ademas, observaciones en MD, indican que en M. jaliscana las celulas en la base del tricoma presenta un contenido blanquecino, probablemente calcio, asf en la superficie del enves. El tipo de tricoma puede utilizarse para separar Moussonia ampla, quien tiene tricomas del tipo 1, largos con glandula apical, del resto de las especies aquf estudiadas. Wiehler (1983) menciono que este tipo de tricoma no es comun en la familia. Nosotros agregarfamos que tampoco en el genera, pues solo M. rupicola y Kohleria papillosa var. pendula C. V. Morton (= Moussonia) los presentan. Por el contrario, los tricomas tipo hongo al parecer son ampliamente distribuidos en el genera, observandose diferencias en el tamano y en la densidad relativa en la lamina de Existe aparentemente una correlacion entre trico- mas mas o menos espaciados y el relieve casi piano con domos estomatales pequenos, como en Moussonia deppeana (Fig. 1), o entre alta densidad de tricomas y domos estomatales altos caracterfsticos de los relieves sinuosos o tipo ampolla, como en M. jaliscana y M. ampla, respectivamente (Fig. 2B). Estas observa- ciones apoyan la hipotesis de Skog (1976) que propone que la presencia de estomas en domos estomatales altos es una estrategia para efectuar una traspiracion mas eficiente, especialmente en plantas con un numero elevado de tricomas. De acuerdo con Skog (1976) entre mas denso el indumento mas altos los domos estomatales. Sugerimos adicionalmente que el plegamiento de la epidermis al formar los domos estomatales consecuentemente tambien incre- menta la densidad de tricomas por unidad de area y la humedad en la capa del indumento. El tamano de los estomas permite distinguir a Moussonia deppeana de M. elegans, y M. elegans de M. jaliscana. Tamanos de estomas y conteos estomaticos parecen ser buenos criterios para distinguir especies dentro del genera. SUPERFICIE ADAXIAL DE LA COROLA Las celulas de la superficie adaxial de la corola pueden considerarse como otro criterio para separar las incluidas en este trabajo. Las modificaciones de las paredes periclinales de las celulas epidermicas, como son las papilas encontradas solo en Moussonia jaliscana y los pliegues en M. deppeana, ayudan a separar estos taxones de M. elegans. El grosor de la corola y la forma de las celulas epidemicas no habfan sido reportados antes de este 426 Annals of the Missouri Botanical Garden mas o menos gruesas, no translucidas, pilosas, con los lobulos expandidos, eroso-denticulados, y los estambres y estigma incluidos. Debido a la aparente semejanza en la forma y color de las corolas, algunas de las especies que se publicaron posteriormente a Moussonia elegans , fueron incluidas en ella por aulores como Fritsch (1913), Morton (1967), Gibson (1974), Wiehler (1975a) y Kvist y Skog (1992), por lo que esta ultima llego a tener una amplia sinonimia. Las especies y variedades que fueron incluidas en algun momento son: M. formosa Van Houtte ex Regel (Fritsch, 1913), M. cosLaricensis Klotzsch ex Oerst., M. papillosa Oerst. ex Hanst., Isoloma jaliscanum , Kohleria papillosa var. sericea Fritsch, K. collina Brandegee, K. pedunculata Brandegee, K. papillosa var. pendula y K. papillosa var. solilaria C. V. Morion. A1 comparar las caracteristicas distintivas encon- tradas en el prolologo con las que presenla la especie en sentido amplio, era evidente que se trataba de un complejo. Era la unica especie con amplia variacion tanto morfologica como anatomica, lo cual, al pretender incluirla en una clave de determinacion del genero, quedaba ubicada en diferentes partes de la clave, ya que presentaba, por ejemplo, diferentes tipos de indumento, el haz podria ser liso ( Moussonia elegans s. str.), densamente papilado (M. jaliscana ) o laxamente papilado (M. papillosa ), dicasios de 3 6 4 flores o incluso flores solitarias (Kohleria collina ), inflorescencias mas cortas que las hojas (K. collina) o mas largas (K. pedunculata ), lobulos del caliz linear- lanceolados (M. costaricensis ) u ovado-lanceolados (K. papillosa var. pendula) por ejemplo, corolas trans- lucidas (M. papillosa var. solitaria) o no (M. papillosa ), ovario densamente piloso (M. elegans s. str.) o casi glabro (K. pedunculata ), tubo del caliz obconico (K. collina) u obconico-eliptico (M. papil- losa ), etc. Cada sitio de la clave en general, podia corresponder a alguna de las especies de la sinonimia. Ademas, el amplio rango de variacion permitia que ejemplares que no podian ser ubicados en alguna otra especie fuera del complejo, fueran incluidos en ese concepto tan amplio que era M. elegans . Algunos de estos ejemplares ya han sido descritos recientemente como especies nuevas (Ram- lrez-Roa, 2007a, 2007b). Nomenclatura Se trato de localizar material original en el que se pudiera haber basado Decaisne para hacer la descripcion de esta especie, pero hasta el momento no se ha encontrado (Roxana Yockteng [P], com. pers.). Tambien, al comienzo del estudio del complejo, se penso en la posibilidad de epitificar la especie (Art. 9.7, McNeill et al., 2006). Sin embargo, se considera que el dibujo y la descripcion proporcionan las caracteristicas suficientes para reconocer a la especie. En este trabajo se presentan solamente las diferencias anatomicas y morfologicas que se encuentran en dos de las especies que formaban este complejo. Sin embargo, la decision tomada aqui para considerar a Moussonia elegans en sentido estriclo, se basa en las observaciones hechas tanto en los demas taxones que han formado el complejo segun Wiehler (1975a), asi como en las especies restantes del genero. La segregacion de las demas especies que han integrado el complejo y los cambios nomenclalurales pertinenles, sera presenlada en su momento en la revision del genero (Ramirez Roa, en preparacion). Ejemplares examinados. GUATEMALA. Alta Verapaz: near San Jose SE of Tactic, 1500 m, Standley 69659 (F). Baja Verapaz: Mtn. side N of Divide N of Santa Rosa, 1650 m, Standley 69878 (F). MEXICO. Chiapas: above Finca Cuxlepec, Breedlove 48663 (MEXU, MO); near municipal border betw. Bochil & Jitotol, 1300-1500 m, 18 Oct. 1963, D. L. & M. L. Denham 62/463 (US). 4. Moussonia jaliscana (S. Watson) D. L. Denham ex Ramirez-Roa, comb. nov. Basionimo: Isoloma jaliscanum S. Watson, Proc. Amer. Acad. Arts 25: 159. 1890. TIPO: Mexico. Jalisco: sandy damp walls of gullies, near Guadalajara, Nov. 1888, C. G. Pringle 1828 (holotipo, US- 0086832!; duplicados, BM!, F!, NY!, US- 01336337!). Arbustos de 1—2 m de alto, plantas vilosa, con tricomas amarillentos. Hojas casi isofilas lanceolado- elipticas, las grandes de 4-18 X 1.7-6. 5 cm, las pequenas de 4.9-8 X 2.4-4 cm, margen con 23 a 56 dienles, venas lalerales 7 a 9; peciolos 0.5— 3.5 cm de largo, vilosos; haz verde claro, generalmente densa- mente papilado en toda la superficie o hacia el margen, con la epidermis brillosa en algunas porciones, celulas epidermicas alrededor de los tricomas de varias hileras, diferentes a las demas celulas adyacentes; enves sinuoso, verde-amarillento, domos bajos, distribuidos de manera ± densa, un estoma por domo; mesofilo diferenciado en paren- quima esponjoso. Flores en cimas umbeladas de 4 flores, generalmente varias cimas hacia el apice de la planla; pedunculos de 0.8-6 cm de largo; bracteas 6— 13 X 0.5—3 mm, lineares, vilosas; pedicelos gruesos, 0.8-3. 2 cm de largo, vilosos, con tricomas amar- illentos; caliz con el tubo obconico triangular, 2-5 X 12-17. Standley, P. C. & L. 0. Wiliiarr Standi. & L.O. Williams. Pp. . Americanae. III. Ceiba 3. www. mbgpr ess . info Volume 98 Number 4 Volume 98, Number 4 May 2012 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, in- dependent reviewers. Instructions to Authors are printed in the back of the last issue of each volume and are also available online at www.mbgpress.info. Editorial Committee Ihsan A. Al-Shehbaz Victoria C. Hollowell Missouri Botanical Garden Scientific Editor, Gerrit Davidse Missouri Botanical Garden Missouri Botanical Garden Allison M. Brock Peter Goldblatt Associate Editor, Missouri Botanical Garden Missouri Botanical Garden Gordon McPherson Tammy Charron Missouri Botanical Garden Associate Editor, Charlotte Taylor Missouri Botanical Garden Missouri Botanical Garden Cirri Moran Henk van derWerff Press Coordinator, Missouri Botanical Garden Missouri Botanical Garden Roy E. Gereau Latin Editor, Missouri Botanical Garden the Missouri Botanical Garden, % Allen Market- ing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. E-mail: annals@allenpress.com. Subscription price for 2011 is $180 per volume U.S., $190 Canada & Mexico, $215 all other coun- is included in the subscription price of the Annals. annals@mobot.org (editorial queries) http://www.mbgpress.info The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Ab- stract/Global Health databases, ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. The full-text of Annals of the Missouri Botanical Garden is available online though BioOne™ (http:// www.bioone.org). © Missouri Botanical Garden Press 2012 The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). The Annals of the Missouri Botanical Garden (ISSN 0026-6493) is published quarterly by the Missouri Botanical Garden, 2345 Tower Grove Avenue, St. Louis, MO 63110. Periodicals post- age paid at St. Louis, MO and additional mail- ing offices. Postmaster: Send address changes to Annals of the Missouri Botanical Garden, % Allen Marketing & Management, P.O. Box 1897, Lawrence, KS 66044-8897. Volume 98 Annals m Number 4 of the 2011 Missouri Botanical Garden THE FERN GENUS POLYSTICHUM (DRY OPTERID ACE AE) IN COSTA RICA1 432 Annals of the Missouri Botanical Garden Among the prominent ferns in the high-montane terrestrial fern genus Polystichum Roth, known in English as “holly ferns.” Polystichum, one of the 10 largest fern genera (data from Smith et al. [2006]), is most diverse in subtropical regions of both the New World and Old World, but it is also diverse in tropical montane regions. In addition, the genus can be found in the alpine zone — in the Sino-Himalayan Region (sensu Takhtajan, 1986), in New Guinea, and in the Neotropics. About one third of the species in the genus is found in the American tropics, where there are three centers of diversity: the Greater Antilles (31 species [Mickel, 1997]), Mexico and Guatemala (18 species [Stolze, 1981; Mickel & Smith, 2004]), and the North and Central Andes (29 species [Kessler et ah, 2005; McHenry & Barrington, unpublished]). The taxonomy of the genus Polystichum remains unresolved in spite of substantial recent study (Roux, 2000; Little & Barrington, 2003; Driscoll & Barring- ton, 2007; Lu et ah, 2007; Li et al., 2008). It is possible to delineate a monophyletic genus Poly- stichum as long as Cyrtomium C. Presl s. str. is excluded; the species allied to C. balansae (Christ) C. Chr. are polystichums (Li et al., 2008). However, the subgeneric taxonomy is in chaos because (1) the key works on sections, by Daigobo (1972) and Roux (2000), are not based on a critical review of world been typified by allopolyploids (e.g., Polystichum sect. Aculeata Christ and Polystichum sect. Lasio- polystichum Daigobo); and (3) morphological conver- gence is rampant in the genus. In fact, given our character variation in the genus, it would be folly to attempt a subgeneric classification. Costa Rica lies in a unique geographic position in the American tropics, between the high-montane regions of Mexico and Guatemala to the northwest (Driscoll & Barrington, 2007). In this analysis, the polystichums found in Costa Rica fall into three groups. Polystichum speciosissimum (A. Braun ex Kunze) R. M. Tryon & A. F. Tryon is sister to all the other species. The remaining species fall into two monophyletic lineages. The first is an indusiate alliance (which I call the Mayan clade) including species endemic to Central America. The second is species endemic to the Central Andes. Identification of Polystichum species in the field and in the herbarium remains an extraordinary challenge, and herbaria continue to have high levels of misidentification in the genus. Forty years of fieldwork in Costa Rica have yielded a diversity of insights into the delimitation, habitat preference, and geographic provenance of the species in the country. In this paper, I seek to synthesize these data into a treatment of the Costa Rican species that will (1) serve as an identification guide, (2) provide insights into the natural history and evolutionary biology of these elegant and confusing plants, and (3) point the way to addressing the remaining problems. Materials a » Methods Herbarium materials from A, AAU, BM, CR, DS, E, GH, K, L, MO, NA, NY, P, UC, US, and VT formed the basis for this study. The materials at VT (for which there are often duplicates at CR) are the result of 21 excursions to Costa Rica between 1970 and 2011. I have made numerous observations in the field in Costa Rica that provide insight into the diversity and ecology of Costa Rican Polystichum species. An array of genetic analyses of Costa Rican polystichums from my lab (e.g., Barrington, 1990, 2003; Little & Barrington, 2003; Driscoll & Barring- ton, 2007) informs the morphological work reported . The General Cor s r ii ns provenance of the higher-elevation fern diversity in Costa Rica is a mix of species with two geographic patterns: species also found in the Andes and species that are common to the north in southern Mexico and Guatemala (Barrington, 2005). There is a tendency in the Costa Rican ferns for the paramo species to extend to the south and the montane-forest species to extend to the north (Barrington, 2005). Polystichum in Costa Rica fits these general patterns, but with some notable exceptions. Neotropical Polystichum is monophyletic, based on a molecular phylogenetic analysis of a sample of nine species included in a recent world phylogeny £ NATURE OF VARIATION I M POLYSTICHUM Four kinds of problems complicate the naming of Polystichum species: (1) phenotypic variation, (2) changes associated with increased size over an individual’s lifespan, (3) challenges presented by indument characters, and (4) hybridization. However, paucity of characters that are stable within species and labile between species is not a problem. There is a wealth of morphological characters to choose from (Little & Barrington, 2003), and close analysis yields allied species (e.g., Barrington, 2003). Volume 98, Number 4 2011 Barrington 433 Polystichum (Dryopteridaceae) in Costa Rica Phenotypic variation is rampant: characters such as petiole length, petiole color, and the curvature of pinnule margins are all labile in relation to variation in light levels. For instance, plants from the highly insolated habitats above treeline in the Cerro de la Muerte have much shorter petioles and much more recurved segments than genetically identical plants from the small shaded stream banks nearby. Unfortunately, these characters are prominent and have often been used as the basis for describing species in Polystichum. Variation with size is also common. For instance, scale color varies with size of leaf and thus presumably with age of the individual among fertile plants of Polystichum. The key variable is value in the sense of the Munsell color solid (e.g., Munsell, 1966); larger, older plants have darker scales than younger, smaller plants of the same species. Shape also varies with size: it is common for small fertile leaves of a species to differ in shape from larger leaves, especially in overall length:width ratio and in the length of the basal pinnae. Leaf dissection alfaroi (Christ) Barrington, P. hartwegii (Klotzsch) Hieron., and P. platyphyllum (Willd.) C. Presl — are Though critical to species identification in Poly- stichum, indument presents several challenging problems; most important is positional equivalence. An indument character must be scored from the same position on the same organ across the plants of the study set to ensure the homology inherent in the character. For instance, although the shape of pinna- rachis scales is one of the most powerful characters for distinguishing Polystichum species, the variation in scales lying in different positions on the pinna- rachis (base of the pinnules vs. between the pinnule attachments) can obscure the stable character states that divide species. Petiole scales present another problem of position. Since the scales at the base of the petiole are routinely different from those above more distal is critical. However, collections with leaves but no stems often do not have the material to ensure that the most proximal portion of the petiole was collected. To make matters worse, some of the most useful indument characters in Polystichum are vulnerable to loss in herbaria. Notable are the features of the cilia on the edges of petiole scales, which are powerful characters for distinguishing species. The more delicate cilia are especially vulnerable to being worn off during collection or routine handling of herbarium specimens, making the characters unscorable. Hybridization and polyploidy are common in Costa Rican Polystichum (Barrington, 1990, 2003). Three Costa Rican polystichums are allopolyploids, and four hybrids have been documented. Emblematic of the problems presented by hybrids in Costa Rica is the backcross hybrid between P. talamancanum Barrington and P. concinnum Lellinger ex Barrington, which is common in mixed populations with its tetraploid parent in disturbed terrain above treeline in the Cerro de la Muerte. The morphological obscured by the backcross hybrid. DIVERSITY, ECOLOGY, AND BIOGEOGRAPHY The genus Polystichum in Costa Rica comprises 12 species; nine of the 12 are diploid. Ecologically, two of the nine diploids (P. nudicaule Rosent. and P. speciosissimum) are characteristic paramo species growing above 3000 m. Five other diploid species are found in forests. These include two montane rain- forest species (forest classification follows Holdridge [1987] throughout), one of which has among the largest leaves in the genus (P. concinnum ), while the other is a diminutive species with unusually thin rhizomes for the genus (P. turrialbae Christ). The remaining three are to be found in premontane moist to wet forests (P. alfaroi, P. muricatum (L.) Fee, and P. hartwegii). Finally, two diploids (P. platyphyllum and P. dubium (H. Karst.) Diels) are characteristic of stream banks and earthen banks in premontane wet and rainforests in Costa Rica. The diploids from the forests are indusiate; those from paramo and stream bank are exindusiate. Biogeographically, the forest diploids are related to indusiate species endemic to Mexico and Central America, whereas most of the paramo and stream bank species are related to exindusiate species confined to the Andes. The single exception is P. speciosissimum, a paramo species disjunct from Mexico and Guatemala. Among the diploids, only the montane rainforest P. concinnum is endemic to the Talamanca massif; all others are either more widely distributed to the north (P. alfaroi, P. speciosissimum) or to the south (P. dubium, P. nudicaule) or are broadly distributed in the American tropics (P. hartwegii, P. muricatum, P. platyphyllum, P. turrialbae). The three remaining species in Costa Rica are allotetraploids, all from high-montane forests and paramos. Two are endemic: Polystichum lilianiae Barrington (high-montane) combines the heritage of the widespread P. turrialbae with an unknown second species, whereas P. talamancanum (paramo) com- 436 Annals of the Missouri Botanical Garden concinnum Figure 1. The Cosla Rican species of Polystichum. P. platyphyllum (Willd.) C. Presl. A-G drawn from D. S. Barrington 1973 (VT). P. dubium (H. Karst.) Diels. A, C, D, H, F drawn from A. Rojas & P. L. Pacheco 5075 (VT). P. concinnum Lellinger ex Barrington. A-F drawn from D. S. Barrington 724 (VT). P. muricatum (L.) Fee. A-F drawn from D. S. Barrington 1240 (VT). P. lilianiae Barrington. A-F drawn from D. S. Barrington 1926 (CR). P. turrialbae Christ. A-F drawn from D. S. Barrington 2050 (VT). — A. Sorus; scale bar = 1 mm. — B. Pinna-raehis scales; scale bar = 1 mm. — C. Leaf; scale bar =15 cm. — D. Basal petiole scale; scale bar = 5 mm. — E. Medial pinnule; scale bar = 5 mm. — F. Medial pinna; scale bar = 3 cm. — G. Lamina near apex; scale bar = 1 cm. — H. Rachis scales; scale bar = 2 mm. Volume 98, Number ■ 2011 Barrington Polystichum (Dryopteric 444 Annals of the Missouri Botanical Garden orbiculatum (Barrington, 1990). It seems likely that this species is unique in combining the genetic heritage of the Mayan clade and Andean clade. Selected specimens examined. COSTA RICA. Cartago: Villa Mills, 1.0 km N of Restaurant La Georgina, D. S. Barrington 577 [shade form] (CR). Limon: Cordillera de Talamanca, peak of Cerro Kamuk, 9°16/30//N, 83°02/W, 3350-3549 m, G. Davidse 26008 (CR, MO). San Jose: Cerro de la Muerte, summit of Cerro de la Muerte, “ox- trail,” offside of rd. to transmission towers, D. S. Barrington 1285 (CR). San Jose/Cartago: Cerro de la Muerte, roadcut; vertical rockface, D. S. Barrington 583 (CR); paramo region, Cerro de la Muerte, Cordillera de Talamanca, 3100 m, L. 0. Williams 24476 (F, NY, US); Cerro Chirripo, SW slopes, along trail from Canaan to summit, near la caverna, 9800- 10,300 ft., A. M. Evans , D. B. Lellinger & F. Bowers 110 (MO, US). mala. Formerly, the name P. fournieri was used for all of these species, including all Costa Rican plants, so the name is common on Costa Rican specimens but never (so far) correct. One feature distinguishing both of the species in Costa Rica from the Mexican P. fournieri is the droop-tip versus regularly uncoiling vernation, respectively, of the two. Polystichum turrialbae is a fairly common species growing at 2700—3200 m in oak forests of the Cordillera Central and the Cordillera de Talamanca in Costa Rica. The species is widely distributed from Mexico (Barrington, 1995; Mickel & Smith, 2004) to Bolivia (Kessler el ah, 2005). Polystichum turrialbae is diploid (Barrington, 2003); it is likely to lie in the indusiate Mayan clade based on its morphology. 12. Polystichum turrialbae Christ, Bull. Herb. Boissier, ser. 2, 6: 163. 1906. TYPE: Costa Rica. Heredia: Volcan Turrialba, 1905, K . Werckle s.n . (hololype, P!). Figure 1. Polystichum smithii Mickel & Beitel, Mem. New York Bot. Card. 46: 315. 1988. TYPE: Mexico. Oaxaca: summit of Cerro San Felipe, 18 km N of Rte. 175, 9500 ft., J. T. Mickel 7056 (holotype, NY!). Rhizome erect, ca. 2 cm diam., supported by adventitious roots. Fronds typically ca. 60 cm but reaching 90 cm, spreading; petiole scales ca. 1 cm, lanceolate, conform, stramineous or stramineous with a weakly developed fulvous center, the boundary between edge and center indistinct, petiole-scale cilia short, stiff, frequent. Laminae about twice as long as wide, acuminate, 2-pinnate and serrate, without a bulbil; pinnae acuminate, ca. 6 times as long as wide, attached at right angles to the rachis; the basal pinnae shorter than the next distal; pinna- rachis scales long-lanceolate, ochraceous; pinnules flat, attached at right angles throughout, the auricle poorly developed, the apex acute, spinules well- developed, the basal acroscopic pinnule of the medial pinnae longer lo much longer than the adjacent pinnule. Indusia over 1 mm in diam. Discussion. Polystichum turrialbae is one of two high-montane forest Polystichum species in Costa Rica with small leaves and thin rhizomes; the other is P. lilianiae. In both species, the rhizomes are erect- ascending and supported by adventitious roots. The leaves of P. turrialbae are less lustrous than those of P. lilianiae , and the basal pinnae are shorter than the next distal pair rather than longer. The petiole scales without a sharply demarcated central dark stripe also distinguish this species from P. lilianiae. The similar P. fournieri A. R. Sm. is endemic to montane rainforests of southern Mexico and western Guate- Selected specimens examined. COSTA RICA. Cartago: Fila Div., Cuerici, 2800 m, L. D. Gomez P. 2378 (DS, F, GH, NY, US). Heredia: N of Heredia, ca. 1 km beyond Porrosati, in ravine, 2100 m, D. B. Lellinger & J. J. White III 1680 (F, US). Limon: Cordillera de Talamanca, headwaters of unnamed W branch of Rio Teribe, betw. Rio Sini & continental divide at Cerro Bekom, 9°10/45//N, 83°03/30//W, 2500-2600 m, G. Davidse , G. Herrera Ch. & R. H. Warner 25740 (CR), 25761 (CR, UC). San Jose: Villa Mills, S of Pan American Hwy. about 1 km E of Hotel La Georgina, D. S. Barrington 696 (CR). Hybrids Polystichum concinnum Lellinger ex Barrington X P. speciosissimum (A. Braun ex Kunze) R. M. & A. F. This hybrid is fairly common above treeline on the Cerro de Talamanca in Costa Rica; I have encoun- tered it repeatedly on Cerro de la Muerte. It combines the pale green lamina and paler indument of Polystichum speciosissimum with the less dissected lamina of P. concinnum ; it is at least somewhat revolute. Barrington 705 (VT), the first documented plant of this hybrid, was a large plant when I first collected material from it in 1979, and it remains alive and prosperous as of 2011, having attained an age of at least 32 years. Selected specimens examined. COSTA RICA. Cartago: Cerro de la Muerte, roadcut on Pan American Hwy. near treeline, D. S. Barrington 705 (CR, VT). San Jose: Cerro de la Muerte, summit with radio towers, 3400 m, 9°33.5N, 83045.2rW, D. S. Barrington 1274 (VT). Polystichum concinnum Lellinger ex Barrington X R talamancanum Barrington This is the commonest of the hybrids encountered in Costa Rica. It is regularly found between 3000 and 3400 m along the Pan American Highway on the Cerro de la Muerte. The pinnae are ascendant like 446 Annals of the Missouri Botanical Garden Daigobo, S. 1972. Taxonomical studies on the fern genus Polystichum in Japan, Ryukyu, and Taiwan. Sci. Rep. Tokyo Kyoiku Daigaku, B 15: 57-80. Driscoll, H. E. & D. S. Barrington. 2007. Origin of Hawaiian Polystichum (Dryopteridaceae) in the context of a world phylogeny. Amer. J. Bot. 94: 1413-1424. Holdridge, L. R. 1987. Ecologia Basada en Zonas de Vida/ Life Zone Ecology. Libros y Materiales Educativos no. 83. Instituto Interamericano de Cooperation para la Agricultura, San Jose, Costa Rica. Kessler, M., A. R. Smith & M. Sundue. 2005. Notes on the genus Polystichum (Dryopteridaceae) in Bolivia, with descriptions of ten new species. Brittonia 57: 205-227. Li, C. X., S. G. Lu & D. S. Barrington. 2008. Phylogeny of Chinese Polystichum (Dryopteridaceae) based on chloro- plast DNA sequence data (trnL-F and rps4-trnS). J. PI. Res. 121: 19-26. Little, D. P. & D. S. Barrington. 2003. Major evolutionary events in the origin and diversification of the fern genus Polystichum (Dryopteridaceae). Amer. J. Bot. 90: 508- 514. Lu, J. M., D. S. Barrington & D. Z. Li. 2007. Molecular phylogeny of the polys tichoid ferns in Asia based on rbcL sequences. Syst. Bot. 32: 26-34. Mickel, J. T. 1997. A review of the West Indian species of Polystichum. Pp. 119-143 in R. J. John (editor), Holttum Memorial Volume. Royal Botanic Gardens, Kew. Mickel, J. T. & A. R. Smith. 2004. The Pteridophytes of Mexico. New York Botanical Garden, Bronx. Mickel, J. T., W. H. Wagner, Jr. & K. L. Chen. 1966. Chromosome observations on the ferns of Mexico. Caryologia 19: 85-94. Moran, R. C. 1987. Monograph of the Neotropical fern genus Polybotrya (Dryopteridaceae). Bull. Illinois Nat. Hist. Surv. 34: 1-138. Munsell, A. H. 1966. Munsell Book of Color: Glossy Linish Collection. Munsell Color Co., Baltimore. Petiver, J. 1712. Pteri-graphia Americana. Royal Society, London. Proctor, G. R. 1985. Lerns of Jamaica: A Guide to the Pteridophytes. British Museum (Natural History), Lon- don. Roux, J. P. 2000. The genus Polystichum in Africa. Bull. Mus. Nat. Hist. Mus. London, Bot. 30: 33—79. Smith, A. R. 1980. New taxa and combinations of pteridophytes from Chiapas, Mexico. Amer. Lern J. 70: 15-27. Smith, A. R. 1981. Llora of Chiapas, Part 2, Pteridophytes. California Academy of Sciences, San Lrancisco. Smith, A. R. & J. T. Mickel. 1977. Chromosome counts for Mexican ferns. Brittonia 29: 391—398. Smith, A. R., K. M. Pryer, E. Scheuttpelz, P. Korall, H. Schneider & P. G. Wolf. 2006. A classification of extant ferns. Taxon 55: 705—731. Stolze, R. G. 1981. Perns and fern allies of Guatemala. Part II: Polypodiaceae. Lieldiana, Bot., n.s., 6: 1-522. Takhtajan, A. 1986. Ploristic Regions of the World. University of California Press, Berkeley. Tutin, T., V. Heywood, A. Burges, D. Valentine & D. Moore. 1993. Llora Europaea. Vol. 1 Revised: Lycopo- diaceae to Platanaceae. Cambridge University Press, Cambridge. REVISION OF THE NORTH AMERICAN SPECIES OF GRINDEUA (ASTERACEAE)1 D. To Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 449 Cavanilles’s type was not found at MA, but there is a specimen at US annotated from Mexico, as s. coll., “Aster glutinosus sp. nov.” (US-00145634 ex BM) that was annotated by Steyermark (1 Mar. 1933) as “part of type collection of Aster glutinosus Cav.,” and that fits the protologue. We designate this specimen here as lectotype of A. glutinosus. Grindelia consists of annual pla neous floriferous stems such as (Steyerm.) G. L. Nesom (Palmer 10134, MO), G. confusa Steyerm. (Correll & Johnston 20165, TEX), and G. grandiflora Hook. (Cory 52450, PH), or of perennial plants with floriferous stems that arise from a rhizome (G. greenmanii Steyerm., Mueller 2173, MO; G. leptocarpa (De Jong & Beaman) Adr. Bartoli & Tortosa, Worthington 8963, NY; G. ohovatifolia S. F. Blake, Nesom & Morgan 6789, MO), a lignotuber (G. tumeri G. L. Nesom, Hinton 22268, TEX), or a rosette of leaves (G. camporum Greene, Howell 10827, MO) that originates proleptically from renewals at the base of stems from the previous year. Two aquatic taxa, G. leptocarpa and G. tricuspis, have noteworthy fistulose stems. Ueaves in Grindelia may be succulent, subcoria- ceous. or herbaceous; sessile or with blades that are attenuate into the pseudopetiole; entire, lobed, or pinnatisect. The blades may range from linear, elliptic, oblong, ovate, obovate, spatulate, to some- what triangular, or fiddle-shaped. Blade bases may be cuneate, cordate, or clasping, and sometimes auric- ulate; blade apices may be acute, acuminate, rounded, obtuse, or mucronate. Leaf margins may be crenate, denticulate, serrate, toothed, and some- times scabrous. Marginal teeth (Fig. 1) may be dull, ending in a group of sessile glandular trichomes, as in G. adenodonta (Wetter 605, NY), G. squarrosa (Pursh) Dunal (Engelmann 650, MO), and G. tumeri (Hinton 22268, TEX), or in a sclerotic tip as in G. arizonica A. Gray (Baker 682, GH) and G. camporum (Howell 10827, MO), or in a spinule as in G. lanceolata Nutt. (Bush 220, MO). Heads in Grindelia may be radiate or discoid, solitary, and sessile to pedunculate. The involucre may be hemispheric to urceolate or campanulate, and sometimes broadly urceolate or obconic. Phyllaries are disposed in four to eight series, subequal or graduated; filiform, linear, elliptic, ovate, or obovate, subulate (gradually tapering to the apex) or acumi- nate (tapering to the apex into a narrow or slender point), with the margin entire or scabrous; erect, spreading, curved, revolute, or reflexed. The phyllary base is usually sclerified; when subulate the apex flattened, with straight to curved tips; when acumi- nate, the acumen may be shorter than, equal to, or longer than the base, filiform, linear, triangular, or ovate, flattened to terete, erect to reflexed. Indumentum in Grindelia includes simple hairs, stipitate glandular trichomes, and sessile glandular trichomes that occur frequently in pits. Ray florets in Grindelia range from 12 to 60 per glandular trichomes. The ray ligule may be narrowly linear to elliptic, oblong or ovate, with the apex entire, lobed, or toothed. Disk florets are numerous, and the corolla may have a gradually or abruptly ampliate throat. Achenes in Grindelia may be yellowish to dull brown, light brown to stramineous or dark brown, 2- to 4-angled, prismatic or flattened, elliptic to globose or subquadrate to quadrate, sometimes obconic, and occasionally winged. The achene surface may be smooth, rugose, striate, barely tuberculate to tuber- culate, or with deep furrows or shallow incisions, with the achene apex rounded, truncate, with one or two caducous and varies from two to five subulate scales (greatest width ca. three times the radial thickness) to setiform awns (greatest width ca. two times the radial thickness), or bristles (width about equal to radial thickness) that are smooth to barbellate (sensu Strother & Wetter, 2006). Gametophytic chromosome numbers have been consistently reported as n = 6: Chambers et ah, 1998 (Grindelia integrifolia DC.); De Jong and Beaman, 1963 (G. leptocarpa and G. tricuspis); Dunford, 1986 (G. arizonica, G. camporum, G. fastigiata Greene var. fastigiata, G. fastigiata var. revoluta (Steyerm.) Adr. Bartoli & Tortosa, G. grandiflora, G. havardii Steyerm., G. lanceolata var. lanceolata, G. lanceolata var. texana (Scheele) Shinners, G. nana Nutt. var. nana, and G. suhalpina Greene); Lane and Li, 1993 (G. campomm, G. confusa, G. hallii Steyerm., and G. squarrosa ); Pinkava and Keil, 1977 (G. arizonica, G. grandiflora, G. greenmanii, and G. squarrosa f. squarrosa); Powell and Turner, 1963 (G. inuloides, G. pusilla (Steyerm.) G. L. Nesom, and G. robinsonii Steyerm.); Raven et al., 1960 (G. hallii); Semple et al., 1989 (G. inuloides). Sporophytic chromosome counts have been reported as 2 n = 12: Cronquist, 1994 (G. fastigiata, G. laciniata Rydb., G. nana, G. Annals of the Missouri Botanical Garden camporum Greene, M. A. Lane 3098 (PH). — B. Leaf tooth ending in a group of sessile glandular trichomes. Micrograph from G. tumeri G. L. Nesom, G. B. Hinton 1 7760 (TEX). — C. Leaf tooth ending in a spinule. Micrograph from G. lanceolata Nutt., M. A. Wetter 485 (SI). — D. Micrograph corresponds to B, but the tooth surface is covered by resins secreted by the glandular nana var. discoidea (Nutt.) A. Gray, and G. squarrosa ); Dunford, 1964 (G. camporum ); Dunford, 1969 (G. grandiflora, G. lanceolata var. texana, G. microcephala DC., and G. oolepis S. F. Blake); Lane, 1993 (G. hirsutula Hook. & Am. and G. humilis Hook. & Am.); Lane and Li, 1993 ( G . humilis ); McLaughlin, 1986 ( G . camporum ); Morton, 1981 ( G . nana and G. squarrosa ); Pinkava and Keil, 1977 (G. grandiflora and G. humilis ); Powell and Turner, 1963 (G. oxylepis Greene, G. oxylepis var. eligulata Steyerm., and G. squarrosa var. eligulata (Steyerm.) Adr. Bartoli & Tortosa); Powell and Powell, 1978 (G. havardii ); Pinkava and Keil, 1977 (G. squarrosa var. eligulata and G. squarrosa var. nuda (Alph. Wood) A. Gray); Raven et al., 1960 (G. hallii, G. hirsutula, G. humilis, G. squarrosa, and G. squarrosa var. nuda); Semple et al., 1989 (G. humilis and G. oxylepis ); Semple et al., 1992 ( G . camporum ); Whitaker and Steyermark, 1935 (G. camporum, G. Columbiana (Piper) Rydb., G. decumbens Greene, G. lanceolata, G. nana, and G. perennis A. Nelson); Zhao, 1996 (G. obcrvatifolia and G. tenella Steyerm.); Zhao and Turner, 1993 (G. grandiflora and G. pusilla). Mitotic chromosome counts also include 2 n = 24: Dunford, 1964 (G. camporum, G. hirsutula, and G. stricta DC.); Dunford, 1969 (G. lanceolata); Lane, 1993 (G. fraxinipratensis Reveal & Beatley, G. humilis, and G. humilis var. platyphylla (Greene) Adr. Bartoli & Tortosa); McLaughlin, 1986 ( G . camporum); Morton, 1981 (G. integrifolia); Semple et al., 1989 (G. subdecurrens DC.); Whitaker and Steyermark, 1935 (G. camporum, G. hirsutula var. brevisquama Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 451 Steyerm., G. humilis, G. maritima (Greene) Steyerm., var. platyphylla (Greene) Steyerm., and G. squarrosa G. rubricaulis DC. var. data Steyerm., G. rubricaulis var. nuda). 16a. 16b. 17a. 17b. 18a. 18b. 19a. 19b. 22b. Taxonomic Key to North American Species of Grindelia la. Aquatic plants; stems fistulose 2 lb. Terrestrial plants; stems not fistulose 3 2a. Stems 1 or 2, scarcely ramified; heads 2-3 cm diam.; involucre 4—7 X 10-15 mm; achenes 2-angled; pappus 2 to 5 bristles 20. G. leptocarpa 2b. Stems several, ramified; heads 3—4 cm diam.; involucre 10—15 X 15-20 mm; achenes 2- or 3-winged; pappus 5 to 12 bristles 39. G. tricuspis 3a. Phyllaries in subequal series 4 3b. Phyllaries in graduated or strongly graduated series 15 4a. Phyllaries subulate 5 4b. Phyllaries acuminate 8 5a. Perennial subshrubs with xylopodium; stems prostrate 12. G. hintoniorum 5b. Annual subshrubs; stems erect 6 6a. Plants 0.08-0.15 m; achenes sharply cut with transverse furrows 31. G. pusilla 6b. Plants 0.25-1.3 m; achenes smooth, rugose, or striate 7 7a. Leaves with sessile glandular trichomes, not in pits, and slipitate trichomes; achenes obconic, the inner flattened; pappus of 2 subulate scales 1 . G. adenodonta 7b. Leaves with sessile glandular trichomes in pits; achenes globose; pappus of 2 or 3 setiform awns 22. G. microcephala 8a. Leaves with teeth ending in a sclerotic tip or a curved spinule 9 8b. Leaves with teeth ending in a group of sessile glandular trichomes 13 9a. Teeth ending in a curved spinule 19. G. lanceolata 9b. Teeth ending in a sclerotic tip 10 10a. Phyllaries with margin scabrous; acumen filiform, terete or subterete 33. G. scabra 10b. Phyllaries with margin entire; acumen linear, triangular, elliptic, or narrowly ovate, flat or subterete 11 11a. Phyllary acumen linear-elliptic to narrowly ovate; pappus of 2 bristles 17. G. inuloides lib. Phyllary acumen triangular 12 12a. Plants rhizomalous; acumen shorter than the base; pappus of 2 or 3 bristles 9. G. greentnanii 12b. Plants not rhizomatous; acumen as long as the base; pappus of 2 to 5 curved, subulate scales 15. G. humilis 13a. Plants with xylopodium; stems prostrate with tips rising upward, not or few ramified; heads 1.4-2 cm diam.; pappus of 2 setiform awns 40. G. turneri 13b. Plants without xylopodium; stems ascendent or erect, ramified; heads 3-5 cm diam.; pappus of 2 or 3 bristles ... 14 14a. Outer phyllaries with acumen very long and narrow; achenes subquadrate, smooth or longitudinally striate with shallow transverse incisions at apex 24. G. nelsonii 14b. Outer phyllaries with acumen equal to or shorter than base; achenes elliptic, 2- or 3-angled, smooth or slightly rugose 29. G. palmer i 15a. Phyllaries acuminate 16 15b. Phyllaries subulate 36 Leaves with sessile glandular trichomes in pits 17 Leaves with sessile glandular trichomes not in pits 23 Leaves with teeth ending in a group of glandular trichomes 18 Leaves with teeth ending in a sclerotic tip 19 Achenes smooth, rugose, or slightly furrowed; pappus of 2 or 3 curved, setiform awns 34. G. squarrosa Achenes with prominent longitudinal furrows; pappus of 2 to 4 straight bristles 36. G. subdecurrens Acumen of phyllaries revolute; heads radiate or discoid 20 Acumen of phyllaries erect or curved, not reflexed or re volute; heads radiate 21 . . . 23. G. naua 20a. Heads pedunculate; involucre 7—10 X 7-10 mm; stems several, somewhat decumbent at the base . . . 20b. Heads sessile or subsessile; involucre 10—14 X 9—17 mm; stems 2 to 4, erect 6. G. fastigiala 21a. Heads 0.8-1. 4 cm diam.; phyllaries with acumen triangular, flat, erect; achenes with a truncate to knobby apex; pappus of 2 bristles 7. G. fraxinipratensis 21b. Heads 1.5-3 cm diam 22 22a. Heads 2-3 cm diam.; phyllaries with acumen terete or subterete, curved; achenes with a crown at the apex; pappus of 2 or 3 subulate scales 3. G. camporum Heads 1.5-2 cm diam.; phyllaries with acumen triangular, subterete, erect; achenes with apex truncate; pappus of 2 to 5 setiform awns 5. G. decumbens 23a. Stipitate glandular trichomes absent 24 23b. Stipitate glandular trichomes present 30 24a. Stems prostrate with tips rising upward; heads subsessile 21. G. macvaughii 24b. Stems erect; heads pedunculate 25 25a. Rhizomatous, perennial plants 32. G. robinsonii 25b. 26a. Not rhizomatous, annual, biannual, or perennial plants 26 Heads 1.5-2 cm diam 27 Annals of the Missouri Botanical Garden Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 453 Figure 2. Grindelia adenodonta (Steyerm.) G. L. JNesom. — A. Part of the involucre, inner phyllary series represented. — B. Achene, 2- or 3-angled, showing the rugose or striate surface. — C. Pappus, one of the two subulate scales. — D. Leaf tooth, with rounded apex, ending in a group of sessile, glandular trichomes. — E. Fertile habit. F-H. Representative phyllaries. — F. Outer — G. Middle series. — series. H. Inner series. Drawn from A. Ruth 284 (PH). or dull brown, obconic, 2- or 3-angled, the inner achenes flattened, 3-4.5 mm, rugose or striate, apices with a crown; pappus of 2 subulate scales, apex dilated, 5—6 mm, straight or somewhat curved, smooth. Chromosome number not known. Distribution and habitat. Grindelia adenodonta has been collected from thickets, prairies, and along streams, from 10 to 200 m in Texas (U.S.A.). Phenology. Grindelia adenodonta flowers from June to October. Volume 98, Number ■ 2011 1993: 541). 458 Annals of the Missouri Botanical Garden (1938) found ihis species different from the species monographed in 1934 and indicated his initial confusion. Discussion. Grindelia confusa is an annual spe- cies that differs from the other Mexican species of Grindelia by its simple hairs and stipitate glandular trichomes, its herbaceous, pinnatisect leaves, and its phyllaries with a long and terete acumen, arranged in (four or five) unequal series. Additional specimens examined . MEXICO. Chihuahua: Mpio. Valentin Gomez Farias, T. Lebgue & E. Estrada 3316 (TEX). Durango: in slightly saline meadows, 19 mi. NE of Durango, rle. 31, D. S. Correll & I. M. Johnston 20165 (TEX); 12.3 mi. W of town plaza of Guadalupe Victoria, on hwy. 40 & just W of Hernandez, M. A. Lane 2273 with R. Sanders (TEX); 23.2 mi. W of Guadalupe Victoria (town plaza) & 11.3 rd. mi. W of San Francisco, I. Madero on hwy. 40, M. A. Lane 2276 with R. Sanders (TEX). Tamaulipas: Matamoros, 16 July 1980, E. Guevara s.n. (XAL). 5. Grindelia decumbens Greene, Pit Ionia 3: 102. 1896. TYPE: U.S.A. Colorado: Cimarron, 30 Aug. 1896, E. L. Greene s.n. (holotype, NDG- 53581 digital image!; isolype, JNDG-53582 digital image!). Perennial subshrubs, 0.15—0.5 m, with conspicu- ous xylopodium; stems 3 to 8, prostrate with tips rising upward, ramified, glabrous. Leaves herba- ceous, sessile, entire or pinnatisect, narrowly oblong to spatulate; blade apex acute, margin entire or toothed, teeth ending in a sclerotic tip, simple hairs absent, sessile glandular trichomes in pits on both blade surfaces; basal leaves 2—8 X 0.6— 1.5 cm; cauline leaves clasping, progressively smaller. Heads radiate, pedunculate, 1.5-2 cm diam.; involucre hemispheric-campanulate, moderately resinous, 3— 11 X 6—15 mm; phyllaries in 5 or 6 unequal series, graduated, acuminate, linear-elliptic to narrowly oblong, reflexed, 3-7 mm, glabrous; acumen trian- gular, subterete, erect, 1—1.5 mm, with sessile glandular trichomes in pits. Ray florets 12 to 24; tube glabrous, ligule narrowly elliptic, 7-12 mm, apex entire or toothed. Disk florets with corolla 5—6 mm, with a gradually ampliate throat. Achenes yellowish or dull brownish, elliptic, 3- or 4-angled, 3-3.5 mm, somewhat striate, apex truncate or knobby; pappus of 2 to 5 setiform awns, 3—5 mm, curved, barbellate to barbellulate. Chromosome number: 2 n = 12 (Whitaker & Steyermark, 1935: 71). Iconography. Steyermark (1934: 540, fig. 18). Distribution and habitat . Grindelia decumbens has been found on plains, slopes, and stream banks, from 200 to 3000 m, in Colorado and northern New Mexico (U.S.A.). Phenology. Grindelia decumbens flowers from July to August. Etymology. The specific epithet is taken from the Latin adjective “decumbens” or “prostrate,” which refers to the habit of the species. Common name. Reclined gumweed (). Discussion. Grindelia decumbens is characterized by its prostrate stem bases, leaves with sessile glandular trichomes in pits on both blade surfaces, and the hemispheric-campanulate involucre with reflexed, acuminate phyllaries. Additional specimens examined. U.S.A. Colorado: T. 34 N, R. 3 W, M. Ownbey 1403 (GH); Conejos Co., dry pasture land near Rio Grande, turnoff to Los Pinos & Apache Canyon on State Hwy. 17, R. G. Walter 4644 (MO); Ouray Co., 13 mi. N of Ouray, J. A. Moore & J. A. Steyermark 3778 (MO); La Plata Co., Durango, C. F. Baker , F. S. Earle & S. M. Tracy 497 (GH); Mineral Co., Santa Maria Reservoir, NW of Creede, D. Walter & V. Walter 10742 (MO); Montrose Co., J. A. Moore & J. A. Steyermark 3780 (MO). New Mexico: Rio Arriba Co., Chama, M. Hebard 2 156 A (GH). 6. Grindelia fastigiata Greene, Pittonia 3: 102. 1896. TYPE: U.S.A. Colorado: Grand Jet., 27 Aug. 1896, E. L. Greene s.n. (holotype, NDG- 53589 digital image!; isotype, NDG-53586-88 three digital images!). Perennial subshrubs, 0.3-1. 5 m; stems 2 to 4, erect, ramified, glabrous. Leaves sessile, subcoria- ceous, oblong-spalulale to elliptic; blade apex acute or obtuse, margin entire, denticulate, or serrate, teeth ending in a sclerotic tip, glabrous, sometimes with margins scabrous, sessile glandular trichomes in pits on both blade surfaces; basal leaves 7-10 X 1.5—2 cm, blades attenuate into a pseudopetiole; cauline leaves clasping, 2.5-5 X 0.8-1. 5 cm. Heads radiate or discoid, sessile or subsessile; involucre hemi- spheric-campanulate, resinous, 10-14 X 9-17 mm; phyllaries in 5 or 6 unequal series, graduated, acuminate, linear-elliptic to elliptic, reflexed, 3-9 mm, glabrous; acumen narrowly ovate, terete or subterete, strongly revolule, 1.5-2 mm, with sessile glandular trichomes. Ray florets present or absent. Disk florets with corolla 5-6 mm, with a gradually ampliate throat. Achenes dark brownish yellow, 2- angled, elliptic, 3.5—5 mm, smooth or striate, apex truncate or somewhat knobby; pappus of 2 or 3 setiform awns, 3-3.5 mm, straight, barbellulate. Etymology. The specific epithet is taken from the Latin “fastigiatus,” which refers to the branches being clustered, parallel, and erect. Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 461 462 Annals of the Missouri Botanical Garden slriale or furrowed, apex knobby or forming a crown; pappus of 2 bristles, 3—4.5 mm, straight, smooth to barbellulate. Chromosome numbers: n — 6 (Pinkava & Keil, 1977: 681; Dunford, 1986: 301); 2 n = 12 (Dunford, 1969: 20; Pinkava & Keil, 1977: 681; Zhao & Turner, 1993: 650). Iconography. Steyermark (1934: 462, fig. 5). Distribution and habitat. Grindelia grandiflora has been collected from disturbed and sandy soils, in roadsides, open forests, prairies, and streams, from 300 to 1200 m, in Texas (U.S.A) and Coahuila (Mexico). Phenology. Grindelia grandiflora has been col- lected in flower from the beginning of July to October. Etymology. The specific epithet refers to the size of the heads, from the Latin “grandiflorus,” meaning “big flowers.” Common name. Many-ray gumweed (). Discussion. Grindelia grandiflora is character- ized by its annual habit and phyllaries with a filiform acumen. Asa Gray considered this species only a variety of G. squarrosa. Both species differ in the sessile glandular trichomes (located in pits in the leaves of G. squarrosa vs. not in pits in G. grandiflora). The acumen of phyllaries ranges from linear to narrowly triangular, terete to subterete in G. squarrosa (vs. the noteworthy filiform, subterete acumen in G. grandiflora). The pappus consists of curved, setiform awns in G. squarrosa (vs. smooth to barbellate bristles in G. grandiflora ), and the corolla of disk flowers has a gradually ampliate throat in G. squarrosa (vs. abruptly ampliate in G. grandiflora). Additional specimens examined. MEXICO. Coahuila: Arteaga, Exposicion E de la Sierra La Viga, El Zorrillo, J. A. Encina 821 with J. A. Portillo (ASU [barcode] 0018162); Sierra de las Alasanas al noroeste del ejido Mesa de las Tablas, J. A. Encina 826 with S. Cruz J. & E. Lopez (ASU [barcode] 0018163); Muzquiz, roadside swamp, T. Reeves & D. J. Pinkava PI 3009 (ASU [barcode] 0018164). U.S.A. Texas: Crockett Co., N part of Ozona, V. L. Cory 40706 (GH); Edwards Co., Ranch Exp. Station, V. L. Cory 5013 (MO); Fremont, laned county rd., line Sutton Co., 24 air mi. SSE of Sonora, V. L. Cory 52450 (PH); Rio Grande Co., Rio Grande valley below Donana, 1879, C. C. Parry et al. s.n. (GH); Sutton Co., SE part of Sonora, V. L. Cory 50544 (US); arroyo in S part of town at end of Poplar St., M. P. Dunford & C. C. Freeman 1157 (MO); E of Sonora, M. A. Wetter 1012 (NY); Tarrant Co., rocky hillsides, A. Ruth 64 (NY); Val Verde Co., along Devils River, 13 Sep. 1900, H. Eggert s.n. (MO-129986). 9. Grindelia greenmanii Steyerm., Ann. Missouri Bot. Gard. 21: 460. 1934. TYPE: Mexico. Coahuila: Lerios, mtn. section 15 leagues E of Saltillo, 10,000 ft., 10-13 July 1880, E. Palmer 471 (holotype, GH [barcode] 00008448!; iso- types, K [barcode] 000221374 digital image!, NY [barcode] 00169633 digital image!, PH [barcode] 00047857 digital image!, US [bar- code] 00931357 digital image!). Grindelia vetimontis G. L. Nesom, Phytologia 68: 330. 1990, syn. nov. TYPE: Mexico. Nuevo Leon: Mpio. Zaragoza, Cerro El Viejo, 15 mi. W Dulces Nombres, 18 Aug. 1948, F. G. Meyer & D. J. Rogers 2988 (holotype, M0- 193880!; isotype, F [barcode] 1587883F digital image!). Rhizomatous perennial subligneous subshrubs, 0.25—0.5 m; stems several, erect, ramified or not, with simple long hairs and generally also with stipitate glandular trichomes. Leaves herbaceous, sessile, oblong-elliptic, 2-11 X 0.5-1. 2 cm; blade apex obtuse to acuminate, margin denticulate to finely serrate, teeth ending in a sclerotic tip, simple hairs and stipitate glandular trichomes on both blade surfaces; basal leaves with blades attenuate into a pseudopetiole; cauline leaves clasping, progressively smaller. Heads radiate, sessile, 2—4 cm diam.; involucre hemispheric-campanulate, scarcely resin- ous, 12—14 X 22—25 mm; phyllaries in 4 or 5 subequal series, acuminate, herbaceous, linear to linear-elliptic, erect to spreading, 8-11 mm, gla- brous, margin entire; acumen triangular, flat to subterete, shorter than the base, 3-4 mm, with stipitate glandular trichomes. Ray florets 25 to 30; tube with stipitate glandular trichomes, ligule elliptic, 12—16 mm, apex entire. Disk florets with corolla 5—7 mm, with an abruptly ampliate throat. Achenes brownish, elliptic, 2-angled, flattened, 2.5—3 mm, smooth or slightly rugose, apex truncate; pappus of 2 or 3 bristles, 5—7 mm, smooth. Chromosome number: n — 6 (Pinkava & Keil, 1977: 681). Iconography. Steyermark (1934: 460, fig. 4). Distribution and habitat. Grindelia greenmanii grows in open forests of Pinus L. and Picea Link., from 2200 to 3650 m, in Coahuila, Nuevo Leon, and Tamaulipas (Mexico). Phenology. Grindelia greenmanii flowers from June to December. Etymology. Grindelia greenmanii was named in honor of Jesse More Greenman (1867—1951), curator of the Missouri Botanical Garden, who studied tropical flora, particularly from Mexico. Discussion. Grindelia greenmanii is different from other Grindelia species with a rhizomatous habit, such as G. obovatifolia and G. robinsonii , with its larger Volume 98, Number ■ 2011 468 Annals of the Missouri Botanical Garden Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 469 Distribution and habitat. Grindelia hirtella grows on calcareous hillsides, from 1200 to 1800 m, in the states of Oaxaca and Puebla (Mexico). Phenology. Grindelia hirtella flowers from Janu- ary to September. Etymology. The specific epithet is taken from the Latin “hirtus,” which refers to being covered with long, weak hairs. Discussion. Diagnostic features of Grindelia hir- tella include leaves with teeth that end in a sclerotic tip, with sessile glandular trichomes present on both blade surfaces, these trichomes not in pits; spreading or reflexed phyllaries arranged in graduated series, with a phyllary with subterete or terete acumen, and tuberculate achenes. The taxon was originally described as a variety of G. squarrosa , but in this species the teeth of the leaves end in a group of sessile glandular trichomes, the glandular trichomes of the epidermis are in pits, and the achenes are not tuberculate. Nesom (1990) considered Grindelia hirtella a variety of G. inuloides ; however G. hirtella is distinguished by its phyllaries arranged in unequal series, in contrast to the subequal series in G. inuloides ; in addition, the heads are smaller (1.5-2 cm diameter vs. 2.5—3 cm diameter) and the achenes are subquadrate and tuberculate (vs. elliptic and slightly flattened, to 3- or 4-angled, smooth or with transverse incisions). Additional specimens examined. MEXICO. Oaxaca: 8 km NE de San Gabriel, J. Gonzalez-Martmez 102 (MO); Oaxaca Valley, C. L. Smith 686 (MO). Puebla: vie. of San Luis Tultitlanapa, near Oaxaca, C. A. Purpus 2514 (MO). 15. Grindelia humilis Hook. & Arm, Bot. Beechey Voy., 147. 1833. TYPE: U.S.A. California: [probably from San Francisco Bay or Monterey Bay; see Gray, Geol. Surv. Calif. Bot. 1: 304. 1876.] s.d., Botany Beechey Voyage s.n. (holo- type, K [barcode] 250077 digital image!; isotype, UC [barcode] 292621 digital image!). Perennial shrubs or subshrubs, 0.3-0.6(— 1.5) m; stems 1 to several, erect or creeping, not ramified or with few branches, glabrous or hairy. Leaves succulent, sessile; blade apex acute or obtuse, margin entire or toothed, teeth ending in a sclerotic tip, simple hairs occasionally present along the nerves or at the margin, sessile glandular trichomes present on both blade surfaces, superficial or in pits; basal leaves narrowly obovale, 13—19 X 2—3 cm, attenuate into a pseudopetiole; blade apex acute or obtuse; cauline leaves linear to oblong, 5—9 X 0.8—2 cm, clasping, auriculate; blade apex acute. Heads radiate, pedunculate, 3.5—5 cm diam.; involucre hemispheric to hemispheric-campanulate, resinous, 12-20 X 17— 30 mm; phyllaries in 4 or 5 subequal series, acuminate, elliptic to narrowly ovate, 8—15 mm, erect or spreading, glabrous or with simple hairs, margin entire; acumen triangular, flat or subterete, as long as the base, with sessile glandular trichomes. Ray florets 20 to 60; tube glabrous, ligule narrowly elliptic, 18— 23 mm, apex entire. Disk florets with corolla 5—6 mm, with an abruptly ampliate throat. Achenes elliptic, light brown, 4-angled, 3—5 mm, smooth or rugose, apex truncate or oblique; pappus of 2 to 5 subulate scales, 2.5-3 mm, curved, smooth to barbellulate. Etymology. The specific epithet is taken from the Latin “humilis” for “low-growing,” which refers to the size of the plant. Discussion. Grindelia humilis is characterized by acuminate phyllaries, arranged in subequal, four or five series, with a triangular acumen as long as the phyllary base. The marginal teeth of its leaves end in a sclerotic tip, and the pappus consists of curved, subulate scales. The taxon is related to G. greenmanii and G. inuloides , sharing leaf teeth that end in a sclerotic tip and phyllaries arranged in subequal series. These two latter taxa differ by the acumen of the phyllaries. The acumen is also triangular but shorter in G. greenmanii , less than the basal segment; the acumen is narrower and linear-elliptic to narrowly ovate in G. inuloides. Both these species share a pappus of bristles, in contrast to the subulate scales seen in G. humilis. The type specimen for Grindelia humilis consists of only the apical portion of the plant and lacks the basal stem and leaves. This may be why Hooker and Arnott described the species as a dwarf herb with narrowly linear and entire leaves. Lane (1992) and Strother and Wetter (2006) included Grindelia humilis as a synonym of G. hirsutula , but the latter species differs from G. humilis by its subulate phyllaries arranged in graduated series. Two varieties are recognized for Grindelia humilis and can be distinguished by the following couplet. la. Stems erect; leaves with apex acute to obtuse, glandular trichomes not in pits 15a. G. humilis var. humilis lb. Stems creeping; leaves with apex rounded, glandular trichomes always in pits 15b. G. humilis var. platyphylla 15a. Grindelia humilis var. humilis. Grindelia stricta DC., Prodr. (DC.) 7: 278. 1838, syn. nov. TYPE: U.S.A. Alaska: Portum Mulgrave, T. Haenke 1873, syn. nov. Grindelia robusta Nutt. var. latifolia (Kellogg) Jeps., Man. FI. PI. Calif. [Jepson], 1020. SSrSSSSfS Rosa Island, 1872-1873, W. G. W. Harford s.n. 471 472 473 Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia All Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 479 482 Annals of the Missouri Botanical Garden 484 Annals of the Missouri Botanical Garden 486 Annals of the Missouri Botanical Garden Figure 11. Grindelia microcephala DC. — A. Part of the involucre, inner phyllary series represented. — B. Leaf tooth, ending in a group of sessile glandular trichomes. — C. Pappus, showing one of two or three setiform awns. — D. Achene, showing the smooth or slightly rugose surface, 3- or 4-angled. — E. Fertile habit. F-H. Representative phyllaries. — F. Outer series. — G. Middle series. — H. Inner series. Drawn from J. A. Moore & J. A. Steyermark 3004 (MO). tip. The taxon is related to G. fastigiata , sharing the arrangement and form of the phyllaries, the presence of pits in the leaves, and the ending of marginal teeth of the blades. Grindelia nana differs by the pedunculate smaller heads only 7—10 mm (vs. the larger sessile or subsessile heads in G. fastigiata, 10— 14 X 9—17 mm). Achenes are 4-angled in G. nana , but are 2-angled in G. fastigiata. Strother and Wetter (2006) included Grindelia nana in the synonymy of G. hirsutula , but both species are quite distinct. In G. hirsutula , the phyllaries are subulate (vs. acuminate in G. nana), the corolla of disk florets has a gradually ampliate throat (vs. abruptly ampliate in G. nana), and the pappus is comprised of two to four subulate scales (vs. two setiform awns in G. nana). We identified two candidate type collections by Nuttall from Vancouver. One sheet ( Nuttall s.n.) was found at BM and matches the protologue of Grindelia nana (with leaves serrate). The Nuttall s.n. collection seen at GH was more variable, with three stems with the leaves entire and one stem with serrate leaves. We designated a leclotype for G. nana from the more representative specimen at BM. The sheet at GH (00037702) with the serrate-leaved stem is regarded as a duplicate and isolectotype for G. nana var. nana. Annals of the Missouri Botanical Garden 1 mm ! # • * * ' a* i 4 9 * 1 cm ■■ 1 • .1 ■■ 1 •f t J • f Figure 12. Grindelia oaxacanci G. L. Nesom. — A. Part of the involucre, inner phyllary series represented. — B. Leaf tooth, C. Pappus, showing one of two setiform awns. — E. Inner series. — G. Middle series. — H. showing blunt teeth, with minute stipitale glandular trichomes on both surfaces. — — D. Achene, showing transverse incisions. E, G, H. Representative phyllaries. Outer series. — F. Fertile habit. Drawn from R. Merrill King 2631 (TEX). Volume 98, Number ■ 2011 494 Annals of the Missouri Botanical Garden Death Valley, California. He also collected speci- mens in Mexico and South America. Common name. Cabezona (). Discussion. Grindelia palmeri is a species with large radiate heads 3—5 cm in diameter, with the acuminate phyllaries in subequal, four or five series; the acumen of the phyllaries is triangular, flat, and reflexed, and the achenes are elliptic, 2- or 3-angled, and smooth or slightly rugose. It is similar to G. nelsonii , sharing the size of the heads, the arrange- ment of the phyllaries, and the pappus with two or three bristles. Grindelia palmeri differs by the form of the acumen of the outer phyllaries (vs. very long and narrow in G. nelsonii ) and the achenes (vs. subquadrate, smooth or longitudinally striate, with transverse incisions at the apex in G. nelsonii). X 12—20 mm; phyllaries in 6 or 7 unequal series, graduated, acuminate, elliptic to linear-elliptic, 3.5- 10 mm, erect or spreading; acumen subterete, 2—4 mm, in the upper phyllaries revolute to reflexed, with sessile glandular trichomes. Ray florets 20 to 30; tube glabrous, ligule narrowly elliptic, 8—12 mm, apex entire. Disk florets with corolla 6—7 mm, with a gradually ampliate throat. Achenes whitish to yellow, elliptic, 4-angled, 4-4.5 mm, rugose or striate, with a crown at the apex; pappus of 2 to 5 setiform awns, 2.5—4 mm, straight to curved, smooth to barbellulate. Chromosome number not known. Iconography. Steyermark (1934: 540, fig. 30), sub Grindelia howellii. Distribution and habitat. Grindelia paysonorum grows on disturbed soils, grasslands, open forests, and dry hillsides, from 900 to 1500 m, in Idaho and Montana (U.S.A.). Additional specimens examined. MEXICO. Colima: 3 km S of Michoacan-Colima border, D. Burch 5139 (MO). Michoacan: pasture, ca. 24 rd. mi. W of Jiquilpan, A. Cronquist 9759 (MO). San Luis Potosi: 22°N, C. C. Parry & E. Palmer 371 (MO); Mpio. Zaragoza, slopes, 10 km W of Santa Catarina on Mexican Hwy. 70, D. E. Breedlove 59348 with F. Almeda (TEX); microwave hill of Rte. 70 E of San Luis Potosi, W of Santa Catarina, D. J. Loockerman 40015 with J. A. Soule (TEX); 33 km E of San Luis Potosi or 1 km of Puerto de la Huerta on Hwy. 86, K. Roe , E. Roe & S. Mori 144 (TEX). 30. Grindelia paysonorum H. St. John, Res. Stud. State Coll. Wash. 1(2): 108-109. 1929, as “Paysonorum.” Grindelia nana Nutt. var. payso- norum (H. St. John) Steyerm., Ann. Missouri Bot. Gard. 21: 547. 1934. TYPE: U.S.A. Idaho: Nez Perce Co., dry ground, 900 ft., Lime Point, 9 May 1926, H. St. John 4361 (holotype, WS-48552!; isotype, RM [barcode] 0001124 digital image!). Grindelia howellii Steyerm., Ann. Missouri Bot. Gard. 21: 549. 1934, as "Howellii.” Syn. nov. TYPE: U.S.A. Idaho: Kootenai Co., on dry arid bluff tops, St. Maries River, 10 Aug. 1894, L. F. Henderson 2791 (holotype, GH [barcode] 00037697!; isotypes, RM [barcode] 0001122 digital image!, RM [barcode] 0001122a digital image!, US [barcode] 00008208 digital image!). Phenology. Grindelia paysonorum has been col- lected in flower from July to September. Etymology. The species was named in honor of Edwin Blake Payson (1893-1927), a botanist and professor of botany at the University of Wyoming (U.S.A.). Discussion. Diagnostic features of Grindelia pay- sonorum include the radiate heads, the acuminate phyllaries, in graduated, six or seven series, the somewhat clasping leaves, with stipitale and sessile glandular trichomes scattered across both blade surfaces. Grindelia howellii closely corresponds to these characters, also described from Idaho. Additional specimens examined. U.S.A. Idaho: Bene- wah Co., breaks of St. Maries River, S. J. Brunsfeld 79438 (NY); Idaho Co., at Lop of grade going down to Kamiah from Winona, Q. Jones 277 (NY); Lemhi Co., Salmon, P. Edwin B. & L. B. Payson 1883 (MO). Montana: Missoula Co., Holland Lake Rd., J. R. Pierce 1146 (NY); Powell Co., Ovando Valley, common along high water line of glacial ponds on Bandy Ranch SW of Upsata Lake, P. Lesica 5789 (NY); Greenough, 1 mi. E of Patonic turnoff, disturbed site, J. R. Pierce 1120 (MO). Perennial subshrubs, 0.6-0. 9 m; stems 1 or 2, erect, ramified, with stipitate glandular trichomes on the distal portion. Leaves herbaceous, sessile, oblong or elliptic, spatulate, 2.5—7 X 0.7— 2.2 cm, base cuneate, somewhat clasping; blade apex acute to obtuse, margin entire, serrate or denticulate, teeth ending in a sclerotic tip, stipitate and sessile glandular trichomes present on both surfaces. Heads radiate, pedunculate to subsessile, 2.5-3 cm diam.; involucre hemispheric-campanulate, resinous, 8-15 31. Grindelia pusilla (Steyerm.) G. L. Nesom, Phylologia 73(4): 327. 1992. Basionym: Grindel- ia microcephala DC. var. pusilla Steyerm., Ann. Missouri Bot. Gard. 21: 467. 1934. TYPE: U.S.A. Texas: betw. Frio & Nueces Rivers, on rd. to Laredo, 27/28 Jan. 1880, E. Palmer 469 (holotype, GH [barcode] 00008433!; isotypes, MO-126251 digital image!, NY [barcode] 00169617!, US [barcode] 00127643 digital image!). Figure 14. by Nesom [1990: 320], NDG 53647 digital 498 Annals of the Missouri Botanical Garden pappus of 2 or 3 setiform awns, curved, 3—5.5 mm, smooth to barbellulale. Etymology. The specific epithet is taken from the Latin “squarrosus,” meaning “rough, with scales, tips, or bracts projecting outward at about 90°, ” which refers to the spreading or reflexed phyllaries. Discussion. Grindelia squarrosa is characterized by its acuminate phyllaries in graduated series and by its leaves with sessile glandular trichomes in pits on both surfaces and marginal teeth that end in a group of sessile glandular trichomes. The taxon is similar to G. subdecurrens by the leaves that bear sessile glandular trichomes in pits on both surfaces, and the marginal teeth that end in a group of sessile glandular trichomes; the heads in both species have acuminate phyllaries arranged in four to six strongly graduated series. However, G. squarrosa differs by its achene surfaces being smooth, rugose, or slightly furrowed (achenes have prominent longitudinal furrows in G. subdecurrens ) and the pappus with curved, setiform awns (a pappus of two to four straight bristles in G. subdecurrens). Three varieties and two forms are recognized within Grindelia squarrosa. The three varieties are distinguished in the key couplets below, with the forms considered under the autonymic variety. la. Heads radiate 34a. G. squarrosa var. squarrosa lb. Heads discoid. 2a. Involucre 8 mm broad 34b. G. squarrosa var. eligulata 2b. Involucre 15—20 mm broad 34c. G. squarrosa var. nuda 34a. Grindelia squarrosa var. squarrosa. Heads radiate, 2. 2-3. 3 cm diam. Ray florets 22 to 36, tube glabrous, ligule narrowly elliptic, 7-15 mm, apex entire. Achenes smooth or rugose. Two forms are further distinguished in the couplet below. la. All leaves entire . . . 34a(l). G. squarrosa f. squarrosa lb. Basal leaves pinnatisect 34a(2). G. squarrosa f. pseudopinnatifida 34a(l). Grindelia squarrosa f. squarrosa. Grindelia perennis A. Nelson, Bull. Torrey Bot. Club 26: 355. 1899, syn. nov. TYPE: U.S.A. Wyoming: Sweet- water River, 27 July 1898, A. Nelson 4988 (holotype, RM [barcode] 0001126 digital image!; isotypes, MO [barcode] 714079 digital image!, NY [barcode] 00169621!, UTC-171663 digital image!). Grindelia serrulata Rydb., Bull. Torrey Bot. Club 31: 646. 1904, syn. nov. Grindelia squarrosa (Pursh) Dunal var. serrulata (Rydb.) Steyerm., Ann. Missouri Bot. Gard. 21: 482. 1934. TYPE: U.S.A. Colorado: Ft. Collins, 5000 ft., 10 Aug. 1891, J. M. Cowen s.n. (holotype, NY [barcode] 169625!). Grindelia squarrosa (Pursh) Dunal var. quasiperennis Lunell, Amer. Midi. Naturalist 3: 143. 1913, syn. nov. TYPE: U.S.A. North Dakota: Butte, 2 Aug. 1909, J. Lunell 1040 (lectotype, designated by Cronquist [citing annotated by Mark Wetter, 1994: 256], MIN not seen). Grindelia squarrosa (Pursh) Dunal f. depressa Steyerm., Ann. Missouri Bot. Gard. 21: 480. 1934, syn. nov. TYPE: U.S.A. Utah: alkali flat, 5 mi. W of Salt Lake City, 1300 m, 23 Aug. 1931, J. A. Moore & J. A. Steyermark 3768 (holotype, MO-1027218!). Leaves entire with blade margin nearly entire, sharply toothed or denticulate. Chromosome num- bers: n = n — 6 (Pinkava & Keil, 1977: 681, sub Grindelia squarrosa var. serrulata ; Lane & Li, 1993: 542, sub G. squarrosa var. serrulata ); 2 n = 12 (Morton, 1981: 361; Cronquist, 1994: 256). Iconography. Cronquist (1994: 259). Distribution and habitat. Grindelia squarrosa f. squarrosa grows on disturbed sandy soils, prairies, hillsides, and along streams, from 200 to 2900 m, in Arkansas, Colorado, Indiana, Iowa, Kansas, Massa- chusetts, Michigan, Montana, Nebraska, New Jersey, North Dakota, Pennsylvania, South Dakota, Texas, Utah, Wisconsin, and Wyoming (U.S.A.) and Coa- huila (Mexico). Phenology . Grindelia squarrosa f. squarrosa flow- C T T . P 1 ers from July to September. Common name. Curly cup gumweed (; )• Discussion. In Grindelia squarrosa f. squarrosa the leaves and margin of the blade are variable in shape. Grindelia perennis , with slender, nearly entire leaves (Nelson, 1899: 356), and G. serrulata , with sharply toothed leaves (Rydberg, 1904: 646), share the diagnostic characters of G. squarrosa. The same is accepted for the variety quasiperennis with denticu- late leaves that narrow toward the base (Lunell, 1913: 143), which was included in the synonymy of G. perennis by Steyermark. Fitting within the taxonomic variation encom- passed by Grindelia squarrosa , variety quasiperennis , variety serrulata , and form depressa established by Steyermark (1934) are included in the synonymy of G. squarrosa var. squarrosa , within form squarrosa. Strother and Wetter (2006) considered Grindelia perennis a synonym of G. hirsutula. Nevertheless, the type of G. perennis matches the diagnostic characters of G. squarrosa and differs from G. hirsutula in its acuminate phyllaries (vs. subulate in G. hirsutula ), its 500 Annals of the Missouri Botanical Garden Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 501 6.4 mi. S of the Coahuila— Nuevo Leon slate line, C. C. Freeman & M. A. Wetter 2054 (NY); Mpio. Galeana, near Rancho Aguililla, G. B. Hinton et al. 19581 (MEXU); N of Rancho Aguililla, G. B. Hinton 27658 (MEXU). 34c. Grindelia squarrosa var. nuda (Alph. Wood) A. Gray, Syn. El. N. Amer. 1: 118. 1884. Basionym: Grindelia nuda Alph. Wood, Bot. Gaz. 3: 50. 1878. TYPE: U.S.A. Oklahoma: Indian Territory, s.d., T. E. Wilcox s.n. (holotype, NY [barcode] 00169629!; isotype, NY [barcode] 00169620!). Figure 16. Grindelia aphanactis Rydb., Bull. Torrey Bot. Club. 31: 647. 1904, syn. nov. Grindelia nuda Alph. Wood var. aphanactis (Rydb.) G. L. Nesom, Phytologia 68: 323. 1990. TYPE: U.S.A. Colorado: Durango, on dry gravelly soil, 21 July 1898, C. F. Baker , F. S. Earle & S. M. Tracy 526 (holotype, NY [barcode] 169594!; isotypes, F [barcode] 0050283F digital image!, MICH [barcode] 1107420 digital image!, MO-714083 digital image!, NY [barcode] 169595!, RM [barcode] 0001113 digital image!, RM [barcode] 0001114 digital image!). Grindelia pinnatifida Wooton & Slandl., Contr. U.S. Natl. Herb. 16: 178. 1913, syn. nov. TYPE: U.S.A. New Mexico: Rio Arriba Co., open slopes, vie. of Chama, 2380-2850 m, 9 July 1911, P. C . Standley 6606 (holotype, US [barcode] 00127653 digital image!). Grindelia squarrosa (Pursh) Dunal f. angustior Sleyerm., Ann. Missouri Bot. Gard. 21: 481. 1934, syn. nov. [described under Grindelia squarrosa var. nuda]. TYPE: U.S.A. Texas: Floyd Co., in roadside marsh 16 mi. E of Lockney on Quitaque Rd., 23 Aug. 1921, R. S. Ferris & C. D. Duncan 3391 (holotype, MO- 902235!; isotypes, CAS [barcode] 7834 digital image!, GH [barcode] 00008441!, NY [barcode] 169629!). Heads discoid; involucre 15—20 mm broad. Stems generally solitary. Achenes longitudinally furrowed. Chromosome number: 2 n = 12 (Raven et al., 1960: 126, sub Grindelia aphanactis ; Pinkava & Keil, 1977: 681, sub G . aphanactis). Distribution and habitat. Grindelia squarrosa var. nuda has been collected from rocky and sandy soils, in prairies and open forests, in Arizona, Colorado, Kansas, New Mexico, Oklahoma, and Texas (U.S.A). The variety has also been documented in pine and juniper forests, from 1200 to 1950 m, in Chihuahua and Coahuila (Mexico). a monomorphic form in G. nuda and dimorphic (those of the outer florets 3-angled and the inner, 4-angled) in G. squarrosa. This difference in the variety nuda of G. squarrosa is due to the absence of ray florets that give rise to the 3-angled achenes. The type of Grindelia aphanactis , collected in the slate of Colorado, matches the diagnostic characters of the typical G. squarrosa with the exception of the absence of ray florets, as in G. squarrosa var. nuda. It differs from the Mexican variety eligulata of G. squarrosa by the size of the heads. In Grindelia pinnatifida the leaves are deeply toothed or pinnatisect, but the remaining characters correspond to those of the type of G. squarrosa var. nuda. Additional specimens examined. MEXICO. Chihuahua: ca. 27 (air) mi. SSW of Columbus, J. Henrickson 11213 (NY). Coahuila: Mpio. de Arteaga, Sierra Zapaliname, R. Hilton et al. 20291 (GH). U.S.A. Arizona: Yavapai Co., roadside near Prescott, J. A. Moore & J. A. Steyermark 3682 (MO). Colorado: Fremont Co., Coaldale, J. A. Moore & J. A. Steyermark 3782 (MO); Montezuma Co., Cortez, J. A. Moore & J. A. Steyermark 3775 (MO). Kansas: prairies, A. Finch 6 (MO); Cullison Pratt Co., M. W. Norris 197 (MO). New Mexico: Albuquerque Co., 19 Aug. 1883, H. H. Rusby s.n. (PH), M. E. Jones 4140 (PH); San Miguel Co., Espanola, Santa Fe to Las Vegas, 7 Sep. 1881, G. Engelmann s.n. (MO-130063); 15 mi. NW of Las Vegas, G. J. Goodman 2321 (MO); Union Co., along US 56/US 64, 3 km E of jet. with US 87 at Clayton Common but scattered in clay soil of ditch & adjacent pasture, M. A. Wetter 502 (MO). Oklahoma: Comanche Co., dry open ground, granite areas, Wichita Mins., Wichita Natl. Forest, E. J. Palmer 41965 (MO); Greer Co., near Willow, R. Stratton 332 (MO). Texas: Brown Co., dry open ground, Brownwood, E. J. Palmer 11111 (MO); Childress Co., Biology class 9014 (MO); Lipscomb Co., along Commission Creek, 3 mi. S of Higgins, forested area, D. S. Correll 30251 (MO); Jeff Davis Co., Davis Mtns., along rd. to Toyahvale, 20 mi. from Ft. Davis, 10 June 1931, J. A. Moore & J. A. Steyermark s.n. (MO- 1025879); Kendall Co., going W on Cascade, Caverns Rd., M. A. Wetter 603 (MO); Llano Co., Enchanted Rock, E. Whitehouse 9008 (MO); Potter Co., Amarillo, 1921, M. Hebard s.n. (PH). 35. Grindelia subalpina Greene, Pitlonia 3: 297. 1898. TYPE: U.S.A. Wyoming: Laramie, 9 July 1896, E. L. Greene s.n. (holotype, NDG-53650 digital image!). Phenology. Grindelia squarrosa var. nuda has been collected from May to October. Etymology. The specific epithet is taken from the naked,” which refers to the us,” meaning “ Latin “nud absence of ray florets. Discussion. Nesom (1990) considered Grindelia nuda and G. squarrosa (the typical variety) as two different species, based on the shape of the achenes, Grindelia platylepis Greene, Pittonia 3: 297. 1898. TYPE: U.S.A. Wyoming: Sherman, 29 July 1893, E. L. Greene s.n. (leclotype, designated here, NDG-53627 digital image!). Grindelia erecta A. Nelson, Bull. Torrey Bot. Club 26: 356. 1899. Grindelia subalpina Greene var. erecta (A. Nelson) Steyerm., Ann. Missouri Bot. Gard. 21: 501. 1934. TYPE: U.S.A. Wyoming: Laramie Hills, 11 Sep. 1898, A. Nelson 5306 (holotype, RM [barcode] 0001116 digital image!; isotypes, MO [barcode] 714074!, NY [barcode] 00169601!). 502 Annals of the Missouri Botanical Garden Volume 98, Number 4 2011 Bartoli & Tortosa Revision of North American Grindelia 505 38. Grindelia tenella Steyerm., Ann. Missouri Bol. Gard. 21: 464. 1934. TYPE: Mexico. Tamauli- pas: de Victoria a Tula, Nov. 1830, J. L. Berlandier 766 (holotype, GH [barcode] 00008459!; isotypes, MO [barcode] 714073!, NY [barcode] 00169645!). Annual herbs, 0.3-0. 6 m; stems 1, sometimes 2, erect, ramified, glabrous or with dense simple hairs close to the head. Leaves herbaceous or subcoria- ceous, oblong to obovate, 1.5—5 X 0.5— 1.1 cm, clasping; blade apex acute or obtuse, margin denticulate or serrate-denticulate, teeth dull or ending in a sclerotic tip or in sessile glandular trichomes, with simple hairs and sessile glandular trichomes on both surfaces. Heads radiate, pedun- culate, 2—2.5 cm diam.; involucre hemispheric, scarcely resinous, 6—8 X 9—13 mm; phyllaries in 4 or 5 unequal series, strongly graduated, acuminate, linear-elliptic, 3-4 mm, the outer bracts spreading and the middle bracts erect, with simple hairs; acumen sublerele to terete, erect or somewhat curved, 1.5-2 mm, with scattered sessile glandular trichomes; the upper phyllaries narrowly elliptic, 8 mm, erect, with short acumen. Ray florets 15 to 18; tube glabrous, ligule narrowly elliptic, 7—9 mm, apex entire. Disk florets with corolla 4—5 mm, with an abruptly ampliate throat. Achenes dull brown, subquadrate, dimorphic, 2 mm, strongly furrowed to tuberculate, the innermost flattened with longitudinal superficial nerves, apex truncate; pappus of 2 or 3 bristles, 4—5 mm, straight, smooth. Chromosome number: 2 n — 12 (Zhao, 1996: 261). Iconography. Steyermark (1934: 465, fig. 7). Distribution and habitat. Grindelia tenella has been collected from grasslands with oak, and oak forests, from 200 to 1800 m, in Nuevo Leon and Tamaulipas (Mexico). Phenology. Grindelia tenella has been collected in flower from June to late November. Etymology. The specific epithet is taken from the Latin “tenellus,” meaning “delicate,” which refers to its slender habit. Discussion. Grindelia tenella are annual herbs with radiate, pedunculate heads, with phyllaries having an acumen subterete to terete and arranged in graduated, four or five series; the achenes are strongly furrowed to tuberculate. The species is similar to G. grandiflora in that both are annual plants with sessile leaves, clasping, with sessile glandular trichomes on both blade surfaces, and have acuminate phyllaries arranged in graduated series. Grindelia tenella differs from G. grandiflora by the phyllary acumen that is subterete to terete (vs. filiform in G. grandiflora) and by the surface of the achenes (vs. striate or furrowed in G. grandiflora). Additional specimens examined. MEXICO. Nuevo Leon: Mpio. Monterrey, Fundicion, Fr. G. Arsene & B. Abbon 101 (MO); El Cercado, J. A. Duke 3974 (MO); Iturbide, Hwy. 31 to Los Tejotes, G. B. Hinton et al. 28206 (MEXU). Tamaulipas: low grassland, Hacienda Santa Engracia, V. H. Chase 7578 (MO); El Mirador, G. B. Hinton et al. 24524 (MEXU, MO); rd. from Santa Engracia toward Dulces Nombres, 4.5 rd. mi. from lowermost crossing of Mimbres Creek, 12.5 rd. mi. NE of Paraje Los Caballos, G. L. Nesom , M. Mayfield & J. Hinton 7452 (MO); on broad dry arroyo 19 km SE of Miquihuana on rd. to Palmillas, L. R. Stanford , K. L. Retherford & R. D. Northcraft 867 (MO); 10 km NW of El Progresso, which is 18 km NW of Ocampo, L. R. Stanford , K. L. Retherford & R. D. Northcraft 1 084 (MO); 5 km S of Hoja Verde, in arroyo, L. R. Stanford , L. A. Taylor 6 S. M. Lauber 2200 (MO). 39. Grindelia tricuspis (Sch. Bip.) Adr. Bartoli & Tortosa, comb. nov. Basionym: Olivaea tricuspis Sch. Bip., in Hooker’s 1c. PI. 12: 2—3, pi. 1103. 1876, non Olivaea tricuspis De Jong & Beaman, Brittonia 15: 89, fig. 1-9. 1963, nom. illeg. TYPE: Mexico. Jalisco: near Guadalajara, W. Schaffner 346 (holotype, P not seen; isotype, fragm. at US [barcode] 00124577 digital image!). Oligonema heterophylla S. Watson, Proc. Amer. Acad. Arts. 26: 138. 1891. Golionema heterophyllum (S. Watson) S. Watson, Bot. Gaz. 16: 267. 1891 [epithet ending corrected by S. Watson, 1891: 267, to neuter ending, cf. Art. 62.2c.]. TYPE: Mexico. Mexico: Atlacomulco, in shallow water, del Rio Hondo, 30 Aug. 1890, C. G. Pringle 3236 (holotype, CH [barcode] 00010745 digital image!; isotypes, CAS [barcode] 0002852 digital image!, F [barcode] 0050822 digital image!, F [barcode] 0050823 digital image!, GH [barcode] 00010746 digital image!, GOET [barcode] 001582 digital image!, JE [barcode] 00000622 digital image!, K [barcode] 000221400 digital image!, M [barcode] 0030095 digital image!, MEXU-33259 digital image!, MO-3726402!, NY [barcode] 00230805 digital im- age!, RSA [barcode] 0001438 digital image!, TEX [barcode] 00373648 digital image!, US [barcode] 00127629 digital image!). Aquatic rhizomatous herbs, 1—1.2 m; stems several, ramified, glabrous, fistulose. Leaves herbaceous, sessile; blade apex acute, with stipilate glandular trichomes when young; basal leaves sunken, lobed or pinnatisect, 4—10 X 0.5— 1.5 cm; basal cauline leaves elliptic, 4—10 X 0.5-1. 5 cm, margin entire or with 3 teeth near the apex, margin serrate, teeth ending in a sclerotic tip; upper leaves linear to elliptic, 1—3.5 cm, clasping, sometimes auriculate. Heads radiate, pedun- culate, 3-4 cm diam.; involucre hemispheric, resinous, 10-15 X 15—20 mm; phyllaries in 3 or 4 subequal 506 Annals of the Missouri Botanical Garden 508 Annals of the Missouri Botanical Garden 512 Annals of the Missouri Botanical Garden 2492 (17a); Bartram E. B. s.n. (15a); Beatley J. s.n. (7), s.n. (7); Berlandier J. L. 766 (38); 2057 (22); Bernard J. P. 58/263 (34a2), 58/510 (34a2); Bioletti F. T. s.n. (3); Biology class 9014 (34c); Blankinship J. W. s.n. (23a); Blumer J. C. 4 (33); Bolander H. N. 389 (13), 6493 (15b); Botany Beechey’s Voyage s.n. (13), s.n. (15); Bradley A. 477 (16a); Brandegee K. s.n. (15b); Breedlove D. E. 59348 (29), 63552 (9); Breitung A. J. 5676 (34al); Bremer K. 2325 (31); Brewer W. H. 850 (3); Brown H. E. 599 (23a); Brown 0. H. s.n. (34al); Brunsfeld S. J. 79438 (30); Burch D. 5139 (29); Burtt Davy J. 1383 (15a), s.n. (15a); Bush B. F. 220 (19a); Bush R. J. s.n. (34al). Candolle A. de s.n. (16); Canne-Hilliker J. M. 1941 (17c); Carter J. L. 1961 (34al); Chandler H. P. 5460 (10); Chas C. D. 36 (34al); Chase V. H. 7266 (32), 7376 1/2 (32), 7578 (38); Chiang F. 9170 (19d); Chittenden E. M. 104 (9); Christ J. H. 11478 (23a); Clemens J. 948 (1); Clokey I. W. 4323 (6a), 4324 (6a); Collom R. E. 463 (2); Colwell A. 85 (16a); Congdon J. W. s.n. (15a); Correll D. S. 20165 (4), 30251 (34c); Cory V. L. 5013 (8), 40706 (8), 50544 (8), 51421 (27), 52450 (8); Cotton J. S. 780 (23b); Cowen J. M. s.n. (34al); Craig C. F. 1774 (6b); Cronquist A. 9759 (29), 9811 (39). Demaree D. 9235a (15b), 10248 (3), 10424 (13), 10469 (3), 10522a (3), 10523 (13), 13192 (19d); Davidson A. 736 (2); De Jong D. C. D. 971 (20); Diaz Luna C. L. 2533 (36), 13168 (17a); Diwart F. M. s.n. (19a); Douglas D. 55 (13); Dreisbach R. R. 6967 (34al); Drushel J. A. s.n. (6a); Duke J. A. 3974 (38); Dunford M. P. 1156 (27), 1157 (8), 1173 (6a). Earle F. S. 509 (11); Eastwood A. 63 (6a), 280 (13), 794 (15b), s.n. (13); Edwin B P. 1883 (30); Eggert H. s.n. (8); Elmer A. D. E 4267 (13); Encina J. A. 821 (8), 826 (8); Engelmann G. 650 (34al), s.n. (34c); Ertter B. 10142 (13); Escobedo J. M. 2509 (17a); Eyerdam W. J. 1150 (15a), 1179 (15a), 1302 (15a), s.n. (15a), s.n. (15a). Femald M. L. 1078 (34al); Fernandez R. 4834 (17a); Ferris R. S. 3391 (34c); Finch A. 6 (34c); Findlay W. s.n. (34al); Fisher L. 37145 (32), 44165 (28); Fogg J. M. Jr. 22274 (34al); Franklin B. 4350 (6a); Freeman C. C. 2041 (34b), 2054 (34b), 2056 (40), 2137 (33); Friberg G. N. s.n. (15a). Galvan R. 3566 (36); Garza F. 4265 (9); Gates F. C. 14967 (34al), 18989 (34al); Ginzbarg S. 180 (9); Gonzalez-Martinez 102 (14); Goodman G. J. 765 (35), 2321 (34c), 3290 (34al); Gorman s.n. (15a); Grant J. M. s.n. (15a); Green P. B. 274 (35); Greene E. L. 974 (23a), s.n. (3), s.n. (5), s.n. (6), s.n. (6a), s.n. (13), s.n. (15a), s.n. (15a), s.n. (15a), s.n. (15b), s.n. (35), s.n. (35); Gregg J. 97 (34b), 354 (34b), 444 (28), 627 (28), s.n. (34b); Grimes J. 2278 (40); Guevara E. s.n. (4); Gutierrez R. s.n. (17a); Gutierrez S. 1 (17a); Guzman R. 953 (17a). Haenke T. s.n. (15a); Hall E. 266 (16a); Hansen G. 172 (15a), 1672 (15a); Harford W. G. W. s.n. (15a); Harrison G. J. 8334 (2); Hasse H. E. s.n. (3); Hebard M. 2156A (5), s.n. (34c); Hedgcock G. s.n. (34al); Heller A. A. 1820 (1), 1840 (15a), 3418 (23a), 5578 (15a), 5810 (15a), 7461 (13), 7542 (15a), 8117 (15a), 8575 (13), 10639 (34al), 10780 (15a), 12979 (23a), 13764 (15a), 15483 (3), 16312 (3), s.n. (15a); Henderson L. F. 432 (16a), 433 (15a), 434 (16a), 435 (23b), 1676 (15a), 2791 (30), 13987 (15a), s.n. (16a); Henrickson J. 5802a (28), 5988 (28), B6382 (36), 11213 (34c); Henry J. K. s.n. (15a), s.n. (15a), s.n. (15a); Hilton R. 20291 (34c); Hinton G. B. 2131 (26), 11919 (17a), 17760 (40), 18344 (9), 18618 (26), 18666 (12), 19481(40), 19581 (34b), 19831 (9), 19865 (9), 20024 (26), 20205 (9), 20291 (34b), 20398 (17a), 20518 (40), 20996 (9), 21161 (26), 21316 (26), 21434 (26), 22268 (40), 22398 (26), 22434 (9), 24382 (9), 24524 (38), 25335 (9), 25388 (9), 25402 (40), 25428 (9), 27192 (40), 27658 s.n. (3); Howe M. A. s.n. (15b); Howell J. T. 2786 (3), 4304 (13), 4352 (13), 4550 (3), 5201 (15a), 5221 (13), 5223 (13), 5329 (3), 5338 (13), 5346 (15a), 5473 (15b), 5496 (3), 5550 (3), 6566 (3), 7545 (15a), 8106 (15b), 8133 (3), 10796 (15a), 10805 (15a), 10827 (3), 10842 (3), 10849 (3), 10861 (15b), 11394 (13), 11414 (13), 11415 (13), 11436 (13), 11455 (13), 11456 (13), 11457 (13), 11462 (3), 11463 (13), 11648 (3), 11658 (15b), 11692(3), 11710 (15a), 11721 (3), 11723 (15a), 21491 (13), s.n. (16a); Humboldt F. W. s.n. (17a). Isle D. 478 (15a). Jackson R. C. 5097 (28); Johnston E. L. 335B (35); Johnston I. M. 8881 (11); Jones G. N. 3442 (15a), s.n. (15a), s.n. (15a); Jones M. E. 4140 (34c), s.n. (13); Jones N. 354 (23b); Jones Q. 277 (30); Joor J. F. s.n. (19a); Jowel J. 6574 (3). Keck D. D. 2943 (15b); Kellogg A. s.n. (13); Kenoyer J. M. 2778 (17b), s.n. (19b); King R. M. 2631 (25), 12699 (35), 12771 (35), Krai R. 27380 (28), 44003 (19a). Lane M. A. 2273 (4), 2276 (4), 2594 (21), 2689 (36), 2702 (36), 3088 (10), 3093 (3), 3097 (3), 3098 (3), 3138 (15a); Lebgue T. 3316 (4); Lesica P. 5789 (30); LeSueur D. H. 974 (28), 1016 (4); Lewis M. 40 (34); Lindheimer F. 255 (1), 418 (19d), 919 (1); Lodewyks M. C. 80 (19a); Loockerman D. J. 40015 (29); Lundell C. L. 18644 (19d); Lunell J. 1040 (34al); Lusk S. D. s.n. (31); Lyall D. s.n. (15a). Macoun J. 420 (15a), 421 (16a), s.n. (15a); Mahler W. F. 8653 (27); Marcus E. 29142 (15a), s.n. (22); Martinez R. J. 1281 (39); McKechnie B. E. 283 (11); McVaugh R. 12820 (21), 16004 (39), 16934 (39), 26625 (39); Mendez s.n. (36); Merrill King R. 2631 (25), 2934 (17b), 11914 (35), 12549 (35), 12781 (35); Metcalfe 0. B. 744 (2), 1302 (33); Metz M. C. s.n. (22); Meyer F. G. 1588 (15a), 2988 (9); Moore J. A. 3004 (22), 3043 (33), 3151 (11), 3183 (6a), 3607 (11), 3682 (15b)’ 3688 (23a), 3768 (34al), 3770 (6a), 3771 (6a)’ 3773 (2), 3775 (34c), 3778 (5), 3780 (5), 3781 (34al), 3782 (34c), 3784 (6b), 3785 (6b); Moore J. A. s.n. (23a), s.n. (34al), s.n. (34c); Mueller C. H. 896 (17a), 2173 (9), 2932 (26); Munz P. A. 11157 (3), 12916 (3), 17997 (10). Neff J. L. 8-19-91-3 (2); Nelson A. 4988 (34al), 5306 (35), 6865 (35), 6894 (34al); Nelson E. W. 6523 (37), 6864 (24); Nelson J. C. 1670 (16b); Nesom G. L. 4800 (9), 5281 (40), 6660 (28), 6789 (26), 7086 (26), 7105 (17b), 7149 (41), 7184 (40), 7189 (40), 7192 (40), 7194 (17b), 7198 (9), 7452 (38), 7587 (12), 7718 (26); Norris M. W. 197 (34c); Norton J. B. 4408 (3), 4410 (3); OwnbeyM. 1 403^(5). ^ ^ Palmer E. 128 (10), 163 (29), 174 (28), 248 (15b), 316 (34b), 469 (31), 471 (9), 472 (8), 520 (28); Palmer E. J. 4161 (19a), 10134 (1), 10249 (19d), 11111 (34c), 11295 (31), 12855 (19d), 29364 (19a), 31465 (19a), 33055 (19a), 33590 (1), 37831 (23a), 41965 (34c), 42014 (19d); Parish S. B. 4452 (3); Parks H. E. 0. 411 (15a), 24190 (15b); Parry C. C. 371 (29), s.n. (8); Patterson H. N. 235 (35); Peck M. E. 13359 (15a); Peirson F. W. 4825 (10), 8190 (3); Pennell F. W. 2581 (34al), 5478 (19d), 17006 (12), 24683 (6b); Perez E. 1632 (17a); Pierce J. R. 1120 (30), 1146 (30); Pilsbry H. A. s.n. (33), s.n. (33); Pitcher Z. s.n. (19); Powell A. M. 585 (17a), 616 (17a), 714b (17b); Pringle C. G. 748 (28), 3236 (39), 4805 (14), 6962 (17b), 9913 (17b); Purer E. A. 4994 (3), 5045 (3), 5462 (3), 5902 (3); Purpus C. A. 2466 (17b), 2514 (14), 4701 (28), 5151 (32), 5722 (32). Redfeam P. L. Jr. 33149 (19a); Redfield J. H. 2964 (6a); Reeves T. P 13009 (8); Reveal J. L. 1536 (7), 1887 (7), 2107 Volume 98, Number ■ 2011 Bartoli & Tortosa Revision of North American Grindelia 513 (7); Reverchon J. 471 (34al); Reyes J. A. 267 (36); Rigg G. B. s.n. (15a); R. M. A. 725 (15a); Roadhouse J. E. 65 (13); Rodriguez A. 1292 (17a); Rodriguez J. D. s.n. (17a); Roe K. 144 (29); Rollins R. C. 2691 (15a); Rose L. S. 33221 (13), 33315 (15a), 33348a (3), 33367 (15a), 40364 (15a); Rothrock J. T. 796 (2); Runyon R. 506 (27), s.n. (27); Rusby H. H. 206 (2), 1555 (35), s.n. (34c); Ruth A. 64 (8), 284 (1); Rydberg P. A. 9210 (18), 9692 (18). S. coll (35), Sandberg J H. 784 (23a); Sanders R. W. 1067 (17a); Savage T. E. s.n. (23a); Schaffner J. G. 738 (28); Schaffner W. 346 (39); Schmidt H. H. 864 (15a); Schmitt W. F. s.n. (15b); Semple J. C. 2708 (35); Selchell W. A. s.n. (15b); Shantz H. L. 597 (6a); Sheldon E. P. 10960 (23b); Smith C. L. 686 (14); Smith C. P. 1016 (15b); Smith L. C. 135 (25); Solbrig 0. 4452 (39); Spalding R. s.n. (23a); Spellenberg R. 4042 (28), s.n. (36); Standley P. C. 6606 (34c); Stanford E. E. 719 (3), 969 (15a); Stanford L. R. 461 (34b), 867 (38), 1084 (38), 2200 (38); St. John H. 4361 (30); Stephenson M. R. 126 (19d); Stewart R. M. 1702 (19d); Steyermark J. A. 24602 (19a), 65926 (34al); Stittler C. I. s.n. (19d); Stratton R. 332 (34c); Suksdorf W. N. 8 (3), 279 (13), s.n. (23b); Sundberg S. 1890 (26), 2911 (36), 3131 (40). Taylor M. 24 (12); Taylor M. S. 4256 (23a), 5189 (15b); Taylor T. M. C. 3009 (15a); Taylor W. R. 2802 (34al); Tenorio Lezama P. 6733 (17b), 20201 (25); Tharp B. C. 44451 (27); Thomas R. D. 20565 (19a), 20964 (19a); Thompson J. W. 4745 (23a), 4945 (23a), 7164 (23b), 7855 (15a), 9959 (15a), 10091 (15a), 11119 (23b), 17376 (23a); Tracy J. P. 4669 (23a); Trelease W. s.n. (19a); Turner B. L. 76-13 (17c), 15329 (17a). Umbach L. M. s.n. (23b). Van der Werff H. 6058 (10), 8707 (10); Ventura P. 15435 (17a); Villarreal J. A. 2312 (28), 2351 (26); Visher S. S. 3360 Walker H. A. 430 (15a); Walter D. 9195 (35), 10549(2), 10742 (5), 11546 (6b); Walter R. G. 4644 (5); Waterfall U. T. 13810 (36), 13869 (21), 15578 (36), 15641 (17a); Weber W. A. 2232 (23a); Welsh S. L. 12631 (6a); Wetter M. A. 77 (18), 305 (34al), 483 (19a), 485 (19), 502 (34c), 568 (6a), 590 (33), 603 (34c), 605 (1), 700 (2), 720 (18), 729 (6b), 737 (27), (6b), 925 (6b), 955 (2), 956 (19c), 961 (19c), 964 (19c)’ 969 (2), 1008 (19d), 1010 (19d), 1012 (8), 1031 (19c), 1033 (2), 2035 (27); Whitehouse E. 9008 (34c), 16425 (1); Whittemore A. T. 82-016 (36); Wilcox T. E. s.n. (34c); Wilder P. 4375 (15b); Wilkes Expedition s.n. (15a); Williamson C. S. s.n. (15a); Wolf C. B. 8101 (3), 8619 (15a); Wooton E. 0. 224 (33), 372 (33), 2568 (2), s.n. (33), s.n. (35); Worthington R. D. 8963 (20); Wright s.n. (8). Zamudio S. 2336 (17a), 2663 (17a); Zeller S. M. 956 (15a), s.n. (15a). SYSTEMATICS OF PATERSONIA (IRIDACEAE, PATERSONIOIDEAE) IN THE MALESIAN ARCHIPELAGO1 515 when viewed^ at high ratification (above 20X)^In ^ rhipidial spathes and Margins of some of the floral | by Stap£ an^ ibustmted d his^ 1894. paper, ? his description (1894: 241) used thedords “lobi | mm (1 inch) long in P. bomeelns and 22 mm (“11 I lin[es]”) long in P lowii. Not noted by Stapf, the | branched hairs (Table 1). Plants from western New Guinea, now West Papua, | 11 U mil ill, 1‘1 ! ! I' iisj, i; "J i¥ f; !i 1 1 3*i i j 5 L li i li T !'!■* J 1 J 1. ) I Mi !* i! y ii 3 pli !< Ill i 1 1 II I Jij ill 517 Annals of the Missouri Botanical Garden Annals of the Missouri Botanical Garden useful to examine Palersonia from Ml. Hamiguilan as it is some 800 km from Mt. Halcon, but both peaks are about 900 km from the next nearest site for the genus, the Kinabalu Massif. Given the geographic patterns of specialion in Palersonia , the Mt. Hami- guitan plants may well be a separate species. 6. Patersonia sumatrensis Goldblatt, sp. nov. TYPE: Indonesia. Sumatra: Gunung Leuser National Park, Gunung Leuser (as Mt. Losir), Gajolands, acid, stony open ground in min. healhland, 2100-2250 m, 29 Jan. 1937, C. C. van Steenis 8444 (holotype, L!; isotypes, CANB!, GH!, K!). Plantae Patersoniae lowii Stapf similes, sed ab ea foliis 3—3.5 mm latis marginibus laevigatis, spathis inaequalibus interna 40-45 mm longa externa 35-40 mm longa, carinis marginibusque spalharum laevigatis, Lubo perianthii laevi- galo alque ovario ca. 12 mm longo dislinguunlur. Tufted perennials 18—35 cm high; stems com- pressed, ?green, oval in section, 18—35 cm X ca. 1.5 mm, flexed above sheathing part of uppermost leaf and then usually slightly inclined. Leaves ca. 10 per shoot in 2 ranks, basal, mostly slightly exceeding flowering stem; blades 3—5 mm wide, when dry closely striate, surface (30X magnification) smooth or obscurely papillate ( de Wilde & de Wilde-Duyfjes 16390), margins smooth, rounded, hyaline, upper- most leaf sheathing the lower half (to 2/3) of stem, apex reaching beyond upper 2/3 of stem. Rhipidial spathes orange-brown, unequal, inner 40-45 mm, outer ca. 35-40 mm, spalhe keels and margins smooth; floral bracts dry-membranous, margins smooth. Llowers lilac-blue; perianth tube ca. 22 mm, smooth; outer tepals ovate, ca. 12 X 7.5 mm, inner tepals ‘/lacking; filaments united in a column ca. 1.5 mm; anthers ca. 2 mm; ovary ellipsoid, ca. 12 mm, smooth; style dividing at apex anthers, stigma lobes unknown. Capsules ellipsoid, ca. 20 mm long, smooth; seeds ellipsoid, smooth, glossy dark brown, obscurely striate, ca. 2 X 1 mm; micropylar caruncle white, flattened, ovate, appressed to surface ca. 0.5 mm long. Llowering recorded in January and April. Discussion. Endemic to the high mountains of northern Sumatra in Indonesia, Palersonia borneensis is evidently restricted to high elevations on Gunung (Mt.) Leuser in the Gunung Leuser National Park. Van Steenis, who made the first collection in 1937, noted that he found no trace of the plant on nearby mountains. It can be distinguished from other Malesian species of Palersonia by the absence of vestiture or any kind on the leaves, spathes, and floral bracts. Even the perianth tube is completely hairless. In one collection, de Wilde & de Wilde-Duyfjes 16390 , papillae are visible on the leaf surface at 30X magnification but not as clearly as in P. borneensis , the only other Malesian species with minutely papillate leaf surfaces. The general aspect of P. sumalrensis most closely approaches that of P. lowii , especially in height, but the somewhat narrower leaves usually slightly exceed the stem, which is weakly flexed above the sheathing portion of the uppermost leaf, a feature recalling the otherwise rather different P. inflexa of eastern Papua New Guinea. Paratypes. INDONESIA. Sumatra: Gunung Leuser Nature Reserve, W lop of Gunung Leuser, 2200-2700 m, 4 Apr. 1975, W. J. J. de Wilde & B. E. E. de Wilde-Duyfjes 16074 (fr.) (K, L); 13 Apr. 1975, W. J. J. de Wilde & B. E. E. de Wilde-Duyfjes 16390 (K, L); Mt. Leuser, Dillon Ripley 5.71. (PH). Literature Cited Amoroso, V. B., L. D. Ossioma, J. B. Arlalojo, R. A. Aspires, D. P. Capil, J. J. A. Polizon & E. B. Sumile. 2009. Inventory and conservation of endangered, endem- ic and economically important flora of Hamiguitan Range, southern Philippines. Blumea 54: 71-76. Beaman, J. H. & R. S. Beaman. 1998. The Plants of Mount Kinabalu, 3. Gymnopserms and Non-orchid Monocotyle- dons. Natural History Publications (Borneo), Kota Kinabalu, Indonesia. Cooke, D. 1986. Iridaceae. LI. Australia 46: 1-66. Geerinck, D. J. L. 1977. Iridaceae. Pp. 77—84 in C. Van Steenis (editor), Llora Malesiana, series 1, Spermatophy- ta, Vol. 8(2). Noordhoff-Kolff, Jakarta. Gibbs, L. S. 1917. Palersonia novo-guineensis Gibbs. P. 101 in LI. Arfak Mis. [Gibbs]. Taylor & Lrancis, London. Goldblatt, P. 1990. Phylogeny and classification of Iridaceae. Ann. Missouri Bol. Gard. 77: 607-627. Goldblatt, P. & J. C. Manning. 2008. The Iris Lamily: Natural History and Classification. Timber Press, Port- land. Goldblatt, P., V. Savolainen, 0. Porteous, I. Soslaric, M. Powell, G. Reeves, J. C. Manning, T. G. Barraclough & M. W. Chase. 2002. Radiation in the Cape flora and the phylogeny of peacock irises Moraea (Iridaceae) based on four plastid DNA regions. Molec. Phylogen. Evol. 25: 341-360. Goldblatt, P., A. Rodriguez, T. J. Davies, J. C. Manning, M. P. Powell, M. van der Bank & V. Savolainen. 2008. Iridaceae ‘Out of Australasia’? Phylogeny, biogeography, and divergence time based on plastid DNA sequences. Syst. Bot. 33(3): 495-508. Goldblatt, P., J. C. Manning, J. Munzinger & P. P. Lowry II. 2011. A new native family and new endemic species for the flora of New Caledonia: Patersonia neocaledonica (Iridaceae: Patersonioideae), from the Mt. Humboldt massif. Adansonia 33: 201-208. IUCN. 2001. IUCN Red List Categories and Criteria, Version 3.1. Prepared by the IUCN Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, United Kingdom. Keighery, G. J. 1990. Patersonia spirafolia (Iridaceae), a new species from south-western Australia. Nuytsia 7: 137-139. Merrill, E. D. 1907. Iridaceae. P. 268 in The Llora of Mt. Halcon, Mindoro. Philipp. J. Sci., C 2: 251-309. Volume 98, Number 4 2011 Goldblatt Patersonia (Iridaceae) in Malesia 523 Stapf, 0. 1894. Irideae. Pp. 241-242 in On the Flora of richness between Mediterranean floristic r London, Bot., ser. 2, 4: 69-264. Van Royen, P. 1979. The Alpine Flora of New Guinea. Valente, L. M., V. Savolainen, J. C. Manning, P. Goldblatt 881-892. ^ ^ 8 ^ Went, F. W. 1924. Iridaceae. P. 114 in L. F. de Beaufort, A. A. Pulle & L. Rutten (editors), Nova Guinea. Resultats des Expeditions Scientifiques a la Nouvelle Gurnee. Alan Grahc SEQUENCING NEW WORLD ECOSYSTEMS: COMPARISON OF THE CRETACEOUS AND CENOZOIC APPEARANCE OF HABITATS WITH BIOME- CHARACTERIZING PLANT GROUPS1 Volume 98, Number 4 2011 Graham Sequencing New World Ecosystems 525 communities (Graham, 2010b: table 7.1, 2011), and From these peaks in temperature, moisture, and now as an additional refinement; and 4) compare the distribution of tropical vegetation, the atmospheric appearance of habitats with the appearance of CO2 concentration declined, ranging from 1000 to lineages available to occupy these habitats that 1500 ppmv between 34 and 24 Ma (Middle to Late together helped shape the versions of the ecosystems Eocene; Pagani et ah, 2009), and temperatures began present today. Geologic and Climatic Environments In the Cretaceous the initial separation of South America from Africa started in the south at ca. 120 Ma and reached the north coasts of the continents ca. 100—90 Ma. Atmospheric CO2 concentration in the Maastrichtian was ca. 1000 parts per million by volume (ppmv), probably not exceeding 1300 ppmv, and global mean annual temperatures (MATs) were ca. 11.2°C warmer than at present (Andrews et ah, 1995). There were equable maritime climates, high sea levels, extensive epicontinental seas, flooded coastlines, low pole-to-pole thermal and biodiversity gradients with little evidence of significant or sustained glaciers, broad continental connections, an episodic descent from the hothouse interval of the Paleogene into the icehouse times of the Neogene. Between 49 and 44 Ma (Middle Eocene), spreading along the mid- Atlantic Ridge reached the North Atlantic disrupting the land bridge between North America and Europe at about the same time as temperatures began to cool. During this interval the prolo-Grealer Antilles emerged as an eastward- moving volcanic island arc that was colliding with the Bahamas Platform. Around 40 Ma the Rocky Mountains and the Andes Mountains were in an early period of uplift, and at 32 Ma South America separated from Antarctica to become an island continent. This initiated the cold Humboldt Current that began cooling the west coast of South America. Continental glaciation began on Antarctica near the i-i 1 i- . -i .• r 1 i Eocene/Oligocene boundary ca. 34 Ma (Bo et ah, and widespread distribution ot many plant and & J v ’ animal species. At the Cretaceous/Tertiary boundary 2009)’ and a droP 111 temperature at the end of the ca. 65 Ma, the asteroid impact (Nichols & Johnson, global Middle Miocene Climatic Optimum (MMCO) 2008) caused extinction of up to 50% of all species initiated alternating full-scale glacial and interglacial (more among the marine plankton) and altered the fluctuations on Antarctica at ca. 17 Ma (Andrill reptile (predator)-mammalian (prey) balance and Antarctic Geologic Drilling Project of the sea floor, other evolutionary dynamics. There was a slight and McMurdo Sound, ). Be- temporary lowering of global temperatures from ash tween ca. 10 and 6 Ma, the Andes Mountains that blocked sunlight, followed by rising temperatures attained the near-final half of iheii present maximum from increased CO2 and other greenhouse gas height, rising from an average of 2000 to 4000 m concentrations (Beerling, 2007) but with no long- (Oncken et ah, 2006, Graham, 2009a, 2010a. 88—97, term effect on climate (Crowley & North, 1991); and 2010b: 76-88). Around 5-3 Ma there was a Pliocene possibly some selection toward deciduousness. In the warm period and although sea levels may have risen Middle and Late Cretaceous (ca. 100-70 Ma), the bY ca- +20 m (), uplift and epicontinental seas and coastal waters began to accumulating volcanics in present-day southern retreat due to the uplift of continents and decreased Central America reunited South America with North formation of submerged, water-displacing plateaus. America through formation of the Isthmian Land Temperatures cooled by ca. 4°C (Zachos et ah, 2008) Bridge at ca. 3.5 Ma. Widespread continental as CO2 out-gassing waned from the slowing of plate glaciations began in the Quaternary, which was motion, and South America separated from North recently defined by the International Commission on America. In the Late Paleocene/Early Eocene (ca. 55 Stratigraphy and the International Union for Quater- Ma), a sudden increase in atmospheric CO2 concen- nary Research as beginning 2.6 Ma (see Giles, 2005). tration (Sexton et ah, 2011) to ca. 1300—1600 pppv These physical and climatic events acting on occurred, possibly involving emission from explosive evolutionary processes within lineages resulted in a vents in the Norwegian Sea as Greenland separated sequence of ecosystems beginning with eight versions from Europe, and there was a further increase in recognized for the Cretaceous. These were 1) the MAT, perhaps as much as 5°-6°C in the equatorial polar broad-leaved deciduous forest, 2) notophyllous regions (Head et ah, 2009). An early version of the broad-leaved evergreen forest, 3) paratropical rain- lowland Neotropical rainforest reached its maximum forest, 4) tropical forest, 5) aquatic, 6) herbaceous extent from an origin ca. 58—55 Ma (Wing et ah, freshwater bog/marsh/swamp, 7) mangrove, and 8) 2009), and tapir, crocodile, and Musaceae (banana- beach/strand/dune. One hundred million years later, like) lineages extended to within the Arctic Circle. events and processes have resulted in the 12 versions 526 Annals of the Missouri Botanical Garden currently recognized for the New World. These are 1) desert, 2) shrubland/chapparal-woodland-savanna, 3) grassland, 4) mangrove, 5) beach/strand/dune, 6) freshwater herbaceous bog/marsh/swamp, 7) aquatic, 8) lowland Neotropical rainforest, 9) lower to upper montane broad-leaved forest (deciduous forest), 10) o), and 12) nology, composition, proposed developmental sequence, and chronology of this 1999, 2010a, 2010b, 2011). The inventories on which they are based are included in the tables and appendices of Graham (1999, 2010a), respectively, and representatives are given in Appendix 1. Inventories After reviewing the geologic and climatic events contributing to the development of New World ecosystems the next step is to compile the represen- tative plant genera that characterize the 12 modem ones (Graham, 1999, 2010a, cf. references). The last step logically would be to consult an inventory of Cretaceous and Tertiary plants for the region. Ideally, the entries would have been assessed for accuracy of and would include the relevant literature and location of the types. Although compilations of the fossil record for individual genera and families are available (e.g., Juglandaceae, Manos et al., 2007; Rubiaceae, Graham, 2009b; Lythraceae, S. Graham, in prep.), no overall database exists. New discoveries rapidly render printed summaries of plant family histories, phylogeny, and biogeography obsolete even when published only shortly before (e.g, for the Lythraceae, S. Graham, personal communication, 2011). Several electronic and printed versions have been attempted (see Graham, 2010a: xiii) and others are underway, including a catalog of Cretaceous and Cenozoic vascular plants for the New World (; see Graham, 2012, in this issue), but it will be a long time before a dependably complete inventory is available. In the meantime, prominent genera in the plant biome that are part of modem New World ecosystems and their fossil representation (based on Graham, 1999, 2010a: app. 1 and 2, and references cited therein) are presented in Appendix 1. This compilation allows only a shadowy insight into the correspondence between habitat opportunity and the appearance of lineages suitable to occupy those habitats. Experi- ence has shown that it will be tempting to use the partial inventories and new discoveries of fossil plants to prematurely reinvent such concepts as tropical rainforest and environmental dynamics, the American biota, leaf/climate relationships, closure of the isthmian land bridge, and refugia. Even so, and even with the limitations of the database in its present state, the records currently allow refinements in the history of the plant component of New World ecosystems. As the inventory improves, a sharper image of this history is emerging. Ecosystem Development The sequence of appearance of Cretaceous and Tertiary ecosystems in the New World has been summarized in Graham (2010b: 174-198, 209-228, 235-267, 275-299, table 7.1; 2011). Elements (E in table 7.1) of most ecosystems appeared in the Cretaceous and Paleogene, but they coalesced into early versions (EV) and then into essentially modem versions (EM) in the Cretaceous (e.g., aquatic communities, lower to upper montane broad-leaved forests), Paleogene (coniferous forest, lowland Neo- tropical rainforest, modem New World mangroves), during and just after the MMCO (widespread grasslands, tropical dry forests, and related vegetation types), and in the Middle Miocene and Pliocene An influencing factor and another detail in this gradational development was the subtle, but none- theless recognizable, interaction between the appear- ance of suitable habitats and the appearance of modern lineages to occupy these habits, allowing Habitat Opportunities and Lineage Availability Some lineages defining the plant formations and associations of the New World ecosystems coevolved in concert with the relatively recent appearance of Hughes & Eastwood, 2006; Polylepis Ruiz & Pav., Simpson, 1986; the parrot genus Pionus, Ribas et al., 2007) in the Andean Highlands. Other progenitors were likely introduced periodically via migration and persisted when and if suitable habitats appeared. The subject of this study is to examine approximately when, for example, did Betula L. and Acer L JQuercus L. and Juglans L. (in eastern North America); Pinus L. and Quercus (Mexico, northern Central America); Poaceae (mid-continent North America, central Ar- gentina); Rhizophora L. (tropical and warm-temperate coasts); Pinus-Picea A. Dietr.-A/>ies MlW.-Tsuga Volume 98, Number 4 2011 Graham Sequencing New World Ecosystems 527 lowland Neotropical rainforest come together to (Endl.) Carriere— Alnus M'lW.-Belula—Popidus L.—Salix (namely, angiosperm-dominant) versions of present- L. (mountains of northwestern North America); day ecosystems. This is not unexpected given that Decodon J. F. Gmel. —Cabomba Aubl -Nymphaea L— wetland/aquatic seasonally unpredictable habitats, Pachira Aubl —Trapa L.-Typlia L. (various aquatic along with fluctuating dry sites in West Gondwana, habitats); and genera that would eventually fonn the are considered places of origin and/or early differen- tiation of the flowering plants. From Appendix 1 iL is constitute the present-day beech-maple/oak-hickory seen that CreLaceous and Paleogene lineages avail- associations of the deciduous forest formation; the able in the New World that took advantage of this pine-oak association of the mid-altitudes of northern habitat opportunity were the Araceae ( Pistia L.), Latin America; the oak— Weinmannia woods of the Cabombaceae ( Brasenites , extinct), Ceratophyllaceae Andean forest belt of Colombia; the grasses of the ( Ceratophyllum L.), Haloragaceae ( Obispocaulis , Tar- prairies and pampas; the modem Rhizophora- domi- ahumara, both extinct), Lemnaceae ( Limnobiophyl - nated mangrove communities; the coniferous and lum exLinct), LyLhraceae (j Decodon , Trapa), Nym- boreal forests of the high altitudes and latitudes; phaeaceae and plants wiLh nymphaealean affinities aquatic communities with a prominent angiosperm ( Aquatifolia , Paranymphaea, Pluricarpaella, all ex- componenL; and the modern rainforest? ThaL is, what tinct), Trapaceae (now part of Lythraceae; Hemitrapa, was the temporal relationship between the physical extinct; note that Trapa in pre-Miocene North opportunity provided by geologic and climatic events American floras [e.g., Barnett, 1989] may be Hemi- (namely , the presence of suitable habitats) and the trapa; the paleocomplex of related/similar genera availability of lineages to occupy those habitats as a needs revision; S. Graham, personal communication, development in defining modern New World ecosys- 2011), incertae sedis: lara (extinct), Quereuxia terns? Aquatic and brackish-water coastal conditions (extinct, Trapago ); several probable aquatics from were available long before the Cretaceous; dry to arid the Lower Cretaceous Crato Formation of Brazil seasonal habitats developed extensively during and (Mohr & Friis, 2000); in addition to the Azollaceae- after the MMCO; and cold, high-altitude environments Azolla Lam. and other older fern and allied groups. appeared in the Mio-Pliocene and Quaternary when The related freshwater herbaceous bog/marsh/swamp the later phases of Andean and other mountain uplift habitat was early occupied by Sphagnum L., combined with cooling trends in the Neogene. Selaginella P. Beauv., Equisetum L., and various However, Rhizophora did not appear in the New ferns, but the record of early angiosperms in this World until the beginning of the Middle Eocene; many environment is not well known, northern temperate deciduous elements, present since the Late CreLaceous and Paleogene in the high beach/strand/dune latitudes, did not move south in significant numbers to modernize the temperate forests of northern Latin This is another ancient habitat, but because America until the decline in temperature after the conditions for preservation are poor in the coarse, MMCO; and although Quercus is known from North shifting sand environment, the history of the America since the Late Cretaceous, it did not appear ecosystem is not well known, lhe earliest New World in northern South America until about 330 thousand records for the Aizoaceae (a family that includes, e.g., years ago (Kyr) to form the modem oak-Weinmannia the extant Sesuvium portulacastrum (L.) L.), Amar- community of the upper Andean forest. This study anthaceae (. Iresene P. Browne), Batidaceae (Bat is explores the temporal correspondence between habitat maritima L.), Convolvulaceae (Ipomoea pes-caprae opportunity and lineage availability as suggested in (L.) R. Br.), Fabaceae ( Canavalia maritima Thouars), Appendix 1, even as it presently stands, as one factor Poaceae ( Uniola paniculata L.), Polygonaceae ( Coc - in conceptualizing the sequence of development of coloba uvifera (L.) L.), and others that can range into modem New World ecosystems. the habitat may extend back into the Cretaceous (e.g., Fabaceae) and Paleocene (Poaceae), but fossil AQUATIC AND freshwater herbaceous bog/marsh/swamp records for these genera are unknown or are These habitats have been available since the origin of the lithosphere and were ready for colonization by angiosperms whenever they appeared (ca. 125 Ma or before). Aquatic and other freshwater environments were already occupied by a variety of angiosperm families in Early Cretaceous and Paleocene times; hence, they are among the oldest essentially modern comparatively recent, probably due, in part, to laphonomy. LOWER TO UPPER MONTANE BROAD-LEAVED FOREST (LATIN america)/deciduous forest (eastern north America) Terrestrial habitats extending from the coast to the moderate Cretaceous and Paleocene highlands were 530 Annals of the Missouri Botanical Garden temperature, increasing dryness, greater seasonality) provides a context for comparing the opportunity provided by increasing dry habitats with the availability of dry-adapted lineages. This evidence supports the intuitive assumption that dry to arid ecosystems would appear at about the same time as grasslands and fully develop somewhat later in the Neogene with increasing dryness (Graham, 1999: 266-268, 2010a: 261-262). Progenitors were present as individual elements and local early versions by the Middle Eocene (e.g., Green River Flora, Colorado/ Utah, ca. 45 Ma; Ephedra L. throughout the Americas), but soils, landscape, and climate com- bined to produce extensive, essentially modern versions in the Middle Miocene and especially in Late Miocene and Pliocene. It is also evident from Appendix 1 that fewer genera characterizing dis- tinctly drier vegetation are present in the Cretaceous and Paleocene floras than in the vegetation types previously discussed, and those in the later Paleo- gene often represent wide-ranging taxa found in other habitats (e.g., Acacia Mill., Aesculus L., Ilex L., Jacaranda Juss., Pinus, Quercus ). The families were present, but the genera forming the shrubland/ chaparral-woodland-savana mostly came later and coalesced in the Neogene. True deserts become most during the Late Miocene/Pliocene and during the interglacials of the Quaternary. The concept of a Neogene origin for modem deserts is the contribution primarily of Daniel Axelrod (1950, 1979, 1983). An essential component the relatively recent event. The assumption is currently being debated and higher elevations earlier in the Cretaceous and Tertiary are being proposed (Dokka & Ross, 1995; Wernicke et ah, 1996; Wolfe et al., 1997; House et al., 1998; Poage & Chamberlain, 2002; Retallack et al., 2004; Stock et al., 2004; Mulch et al., 2006; Schuster et al., 2006). This emphasizes the need to conceptualize the develop- ment of modem ecosystems as typically gradational and that concept is facilitated by the device of recognizing elements, early versions, and essentially modem versions appearing over time. TUNDRA AND ALPINE TUNDRA (PARAMO) the Quaternary beginning ca. 2.5 Ma, and alpine tundra/paramo, additionally, upon the elevation of highlands toward their maximum altitudes. Aster- aceae, Cyperaceae, Ericaceae, Fabaceae, Poaceae, Rosaceae, a were local and late in coming and the lowland and alpine tundra, along with modern deserts, are the most recent of the Earth’s ecosystems to appear. New World ecosystems has become possible only after the broad outlines of geologic, climatic, and biotic events and processes were established (Gra- ham, 1999, 2010a, 2010b, and references therein). In of studies, two such details have been considered: 1) the broad and overlapping sequence of appearance of the ecosystems (Graham, 2011); and 2) the corre- spondence between the appearance of habitats, and the presence of lineages to occupy these habitats, which further defines the sequence (present paper). The latter observation adds a nuance to concepts about ecosystem origin and evolution. It capitalizes on the baseline data consisting of identifications, integration with results from other lines of inquiry, and interpretation within the broadest context of ancillary information (namely, a systems approach; Graham, 2010b: 2—5). A guesstimated 95% or more of ecosystem evolution is determined by the forcing mechanisms of geologic events and climatic change acting on evolutionary processes. It is suggested here that a small but significant refinement additionally involves the correspondence between opportunities for colonization provided by physical alterations in the environment and the fortuitous availability of lineages resulting from the random processes of evolution and dispersion. In compiling this information it became evident that considerably better databasing of the paleobo- tanical record is needed. In Appendix 1 an asterisk (*) or double-asterisk (**) designates the genus or at least the family is reported in the fossil record. The others are characteristic genera of the ecosystems, and establishing a record for them would be especially useful in further clarifying lineage histo- Even in however, it is evident available long before there were angiosperm lineages to occupy them (aquatic, freshwater herbaceous bog/marsh/swamp, beach/ strand/dune). At other times the potential lineages or progenitors were present before there were suitable topographic and climatically defined habitats for 532 Annals of the Missouri Botanical Garden Graham, A. 2011. The age and diversification of terrestrial New World ecosystems through Cretaceous and Cenozoic time. Amer. J. Bot. 98: 336-351. Graham, A. 2011, et seq. Catalog and literature guide for the integration of molecular and morphological data sets. Syst. Biol. 56: 412-430. Martmez-Millan, M. 2010. Fossil record and age of the Asteridae. Bot. Rev. 76: 83—135. Cretaceous and Cenozoic vascular plants of the New World. Mohr, B. A. R. & E. M. Friis. 2000. Early angiosperms from , accessed 15 December 2011. Graham, A. 2012. Catalog and literature guide for the Lower Cretaceous Crato Formation (Brazil), a preliminary report. Internat. J. Plant Sci. 161(6 Suppl.): 5155-5167. Cretaceous and Cenozoic vascular plants of the New Mohr, B. A. R., M. E. C. Bernardes-de-Oliveira & D. W. Taylor. 2008. Pluricarpellatia , a nymphaealean angio- World. Ann. Missouri Bot. Gard. 98: 539-541. Head, J. J., J. I. Bloch, A. K. Hastings & J. R. Bourque. 2009. Giant boid snake from the Paleocene neotropics sperm from the Lower Cretaceous of northern Gondwana (Crato Formation, Brazil). Taxon 57: 1147-1158. reveals holler past equatorial temperatures. Nature 457: Morley, R. J. 2000. Origin and Evolution of Tropical Rain 715-717. Herendeen, P. S., S. Magallon-Puebla, R. Lupia, P. R. Crane & J. Kobylinska. 1999. A preliminary conspectus of the Allon Flora from the Late Cretaceous (late Forests. John Wiley & Sons, Chichester. Mulch, A., S. A. Graham & C. P. Chamberlain. 2006. Hydrogen isotopes in Eocene gravels and paleoelevation of the Sierra Nevada. Science 313: 87-89. Santonian) of central Georgia, U.S.A. Ann. Missouri Nich0ls, D. J. & K. R. Johnson. 2008. Plants and the K-T Bol. Gard. 86: 407-471. Boundary. Cambridge University Press, Cambridge. Hoffman, G. L. & R. A. Stockey. 1999.^ Geological setting 0ncken? 0„ G. Chong, G. Franz, P. Giese, H.-J. Gotze, V. Ramos, M. Strecker & P. Wigger (editors). 2006. The Andes, Active Subduction Orogeny. Springer, Berlin. Pagani, M., K. Caldeira, R. Berner & D. Beerling. 2009. The role of terrestrial plants in limiting atmospheric CO2 decline over the past 24 million years. Nature 460: 85— 88. Piperno, D. R. & H.-D. Sues. 2005. Dinosaurs dined on grass. Science 310: 1126—1128. Poage, M. A. & C. P. Chamberlain. 2002. Stable isotope evidence for a pre-middle Miocene rain shadow in the western Basin and Range: Implications for the paleo- topography of the Sierra Nevada. Tectonics 21: 1-10. Prasad, V., C. A. E. Stromberg, H. Alimohammadian & A. Sahni. 2005. Dinosaur coprolites and the early evolution of grasses and grazers. Science 310: 1177-1180. and paleobotany of the Joffre Bridge roadcut fossil locality (late Paleocene), Red Deer Valley, Alberta. Canad. J. Earth Sci. 36: 2073-2084. Hoorn, C. & F. Wesselingh (editors). 2010. Amazonia, Landscape and Species Evolution: A Look into the Past. Wiley-Blackwell, Chichester. Hoorn, C., F. P. Wesselingh, H. ter Steege, M. A. Bermudez, A. Mora, J. Sevink, I. Sanmartin, A. Sanchez-Meseguer, C. L. Anderson, J. P. Figueiredo, C. Jaramillo, D. Riff, F. R. Negri, H. Hooghiemstra, J. Lundberg, T. Stadler, T. Sarkinen & A. Antonelli. 2010. Amazonia through time: Andean uplift, climate change, landscape evolution, and biodiversity. Science 330: 927— 931. House, M. A., B. P. Wernicke & K. A. Farley. 1998. Dating topography of the Sierra Nevada, California, using Apatite (U-Th)/He ages. Nature 396: 66-69. Hughes, C. & R. Eastwood. 2006. Island radiation on a continental scale: Exceptional rates of plant diversifica- tion after uplift of the Andes. Proc. Natl. Acad. Sci. U.S.A. 103: 10334-10339. Renner, S. S., G. Grimm, G. M. Schneeweiss, T. F. Stuessy & R. E. Ricklefs. 2008. Rooting and dating maples (Acer) with an uncorrelated-rates molecular clock: Implications for North American/Asian disjunctions. Syst. Biol. 57: 795-808. Johnson, K. R. 1996. Description of seven common fossil Retallack’ G. J., J. G. Wynn & T. J. Fremd. 2004. Glacial- leaf species from the Hell Creek Formation (Upper Cretaceous: Upper Maastrichtain), North Dakota, South Dakota, and Montana. Proc. Denver Mus. Nat. Hist., Ser, 3, No. 12: 1-48. Johnson, K. R. & B. Ellis. 2002. A tropical rainforest in Colorado 1.4 million years after the Cretaceous-Tertiary boundary. Science 296: 2379-2383. Kappelle, M. & A. D. Brown (editors). 2001. Bosques interglacial-scale paleoclimatic change without large ice sheets in the Oligocene of central Oregon. Geology 32: 297-300. Ribas, C. C., R. G. Moyle, C. Y. Miyaki & J. Cracraft. 2007. The assembly of montane biotas: Linking Andean tectonics and climatic oscillations to independent regimes of diversification in Pionus parrots. Proc. Royal Soc., London, Ser. B, Biol. Sci. 274: 2399-2408. Nublados del Neotropico. Instituto Nacional de Biodi- Schuster, M., P. Duringer, J.-F. Ghienne, P. Vignaud, H. T. versidad, Santo Domingo. Little, S. A., S. W. Kembel & P. Wilf. 2010. Paleotemper- Mackaye, A. Likius & M. Brunet. 2006. The age of the Sahara Desert. Science 311: 821. ature proxies from leaf fossils reinterpreted in light of Sexton, P. F., R. D. Norris, P. A. Wilson, H. Palike, T. evolutionary history. PLoS ONE 5(12), published online doi: 10. 1371/journal. pone. 0015161. Luna- Vega, I. & S. Magallon. 2010. Phylogenetic compo- Westerhold, U. Rohl, C. T. Bolton & S. Gibbs. 2011. Eocene global warming events driven by ventilation of oceanic dissolved organic carbon. Nature 471: 349—352. sition of angiosperm diversity in the cloud forests of Simpson, B. B. 1986. Speciation and specialization of Mexico. Biotropica 42: 444-454. Manchester, S. R. & E. L. O’Leary. 2010. Phylogenetic distribution and identification of fin-winged fruits. Bot. Rev. 76: 1-82. Manos, P. S., P. S. Soltis, D. E. Soltis, S. R. Manchester, S.- H. Oh, C. D. Bell, D. L. Dilcher & D. E. Stone. 2007. Phylogeny of extant and fossil Juglandaceae inf erred from Polylepis in the Andes. Pp. 304-316 in F. Vuilleumier & M. Monasterio (editors), High Altitude Tropical Biogeog- raphy. Oxford University Press, Oxford. Stock, G., R. S. Anderson & R. C. Finkel. 2004. Pace of landscape evolution of the Sierra Nevada, California, revealed by cosmogenic dating of cave sediments. Geology 32: 193-196. 536 Annals of the Missouri Botanical Garden *Alchornea , *Allophylus, *Alnus Mill., *Alsophila R. Br., * Araucaria (Cretaceous, Mexico), **Arecaeeae, **Astera- ceae, Astro car yum, *cf. Banisteriopsis C. B. Rob., *Bauhinia L., Befaria Mutis ex L., * Bernardia Houst. ex Mill., * Blechnum (Cretaceous, Argentina), *Bombax , *Brunellia Ruiz & Pav., *Bucida , *Bursera , *Caesarea Cambess., Calophyllum , Ccircipa , *Carpinus L., * Catostemma Benth., *Cavanillesia Ruiz & Pav., *Cecropia (secondary, wide- spread), *Cedrela , *Ceiba , Ceratozamia Brongn., *Chamae- dorea , Charianthus D. Don., **chenoam (* Ire sine), Chlor- ophyllum, *Chry sophy llum L., *Clethra L., *Cleyera Thunb., *Clusia L., *Coccoloba, *Cordia , *Cornus L., * Cosmibuena Ruiz & Pav., * Cupania , *Cupressus L., * Cyathea Sm., *Cymbopetalum Benth., * Cyrilla Garden ex L., * Dacrydium Lamb. (Cretaceous, Argentina), Dacryodes Yahl-Sloanea L. (low-altitude wet forests, Antilles), * Danaea Sm., *Daphnop- sis, *Dendropanax Decne. & Planch., *Dennstaedtia Bernh. (Cretaceous, Argentina), *Desmanthus Willd., * Dichapetalum Thouars, Didymopanax, *Dioscorea L. type, *Diospyros L., cf. *Doliocarpus Rol., * Drimys , *Dryopteris Adans. (Cretaceous, Argentina), **Frieaceae (*Cavendishia Lindl. type), *Ery- thrina L. (widespread), * Eucommia Oliv. (extinct in the New World), Eugenia (as *Eugenia-Myrcia), Euterpe, **Fagaceae (*f Quercinium, Cretaceous, Mexico), **Fagaceae s.l. (*f Antiquacupula, Protofagacea , Cretaceous, Georgia, Heren- deen et al., 1999), *Fagus L., *Faramea, * Ficus (wide- spread), Freziera Willd., Gaussia H. Wendl., *Gleichenia Sm. (Cretaceous, Argentina), cf. * Glycydendron Ducke, *Grammi- tis Sw., *Guarea, *Guazuma Mill., *Gustavia L., **Hama- melidaceae (**j* Allonia , Cretaceous, Georgia, Herendeen et al., 1999), *Hampea/Hibiscus, *Hauya DC., * Hedyosmum Sw., * Heliconia , * Hemitelia R. Br-Cne midaria C. Presl, *Hiraea, *Hymenaea, * Hymenophyllum (Cretaceous, Argen- tina), * Ilex , * Jacaranda, *Jamesonia/Eriosorus Fee complex, **Juglandaeeae (as *^Monopites fossil pollen, referrable to Alfaroa, Engelhardia Lesch. ex Blume, Oreomunnea type), *Juglcms L., *Justicia L., * Lacmellea , *Laetia, **Lauraceae (* Persea type, Cretaceous, Mexico), * Laurelia Juss. (Creta- ceous, Argentina), *Liquidambar L., * Lisianthus L., *Lomar- iopsis-Stenochlaena J. Sm., * Lonchocarpus Kunth, *Lopho- sorici, *Luehea, * Lycopodium, *Lygodium, * Magnolia L., * Malpighia L., ** Malvaceae (as **j* Javelinoxylon , Cretaceous, Mexico), *Manicaria type, *Marcgravia L., Mastichodendron (Engl.) H. J. Lam, *Matayba, **Melastomataceae, *Meliosma, * Mimosa, *Mortoniodendron Standi. & Steyerm., * Myrcia , *Myrica L., *Norantea Aubl., *Ocotea, *Oreopanax Decne. & Planch., Ormosia Jacks., * Ostrya Scop., *Paullinia, *Pelto- phorum, *Peperomia Ruiz & Pav., Perrottetia Kunth, * Persea, *Petrea L., *Pinus, *Pityrogramma Link, **Platanaceae (Early Paleocene Castle Rock flora), *Platanus L., *Pleoden- dron Tiegh., **Poaceae, **Podocarpaceae-Taxodiaceae af- finities (Cretaceous, Mexico), *Podocarpus (Cretaceous, Argentina), *Populus, *Posoqueria Aubl., cf. *Pouteria, Prestonia R. Br., *Protium, *Prunus L., * P seudobombax , *Psidium L., *Psilotum Sw., *Psychotria, *Pteridium Gled. ex Scop., * Pteris , *Quercus L., *Rajania L., *Rauwolfia Ruiz & Pav., *Rhamnus, Richeria Vahl, *Roupalci Aubl., *Rourea Aubl., Roystonea 0. F. Cook, *Rubus L., *Sabicea Aubl., *Salix, Samcmea (Benth.) Merr., *Scimbucus L., *Scipium, Scheelea, *Securidaca L., *Selaginella, *Serjania Mill., *Smilax, Spathelia L., * Spathiphyllum , *Sphaeropteris/Tri- chipteris, *Stillingia, * Struthanthus Mart., Styrax L., *Sym- phonia, *Symplocos, * Synechanthus H. Wendl. type, *Tapir- ira Aubl., *Tecomci Juss., Terminalia (as *Combretum- Terminalia), *Ternstroemia Mutis ex L. f. (as *"\Ternstroe- mites, Early Eocene, North Dakota, U.S.A.), *Tetragastris, *Tetrorchidium Poepp., *Thyrsopteris Kunze (Cretaceous, Argentina), *Ticodendron Gomez-Laur. & L. D. Gomez (as *f Ferrignocarpum) , *Tilia L., **Tiliaceae (Early Paleocene Castle Rock flora), cf. *Tillandsia, * Tontelea Miers., * T ournefortia , *Trithrinax Mart., *Ulmus L., *Weinmannia L., *Zanthoxylum; open, secondary, clearings: * Alibertia A. Rich, ex DC., *Borreria G. Mey., *Byttneria Loefl., *Cuphea P. Browne, *Desmanthus, * Dicranopteris Bernh., * Passijlora L , **Roaceae, *Rajcmia L., *Smilcix L. SOUTH AMERICA/ANTARCTICA This includes tropical elements extending into the Andean forest belts; many (*, **) identifications are Late Cretaceous and Paleocene. *Acalypha L., *Acmopyle Pilg. (Cretaceous, Argentina, Chile), * Adiantum , * Alchomea , * Alnus , * Alsophila , * Anemia Sw., * Athyrium Roth, * Antrophyum Kaulf., * Araucaria, **Araucariaeeae (Cretaceous, Antarctica; * Araucaria, Arau - carites, foliage, Falcon-Lang & Cantrill, 2001; Araucariox - ylon, ’ Araucaripitys , wood), *Asplenium L., **Asteraceae, *Austrocedrus Florin & Boutelje, * Blechnum , Bocconia L., Bolax, Bonne tia, *Botrychium Sw., *Bravaisia, Brocchinia Schult. f., Brunellia, * Calophyllum, * Cedrela , * Ghrysophyllum , * Clethra , * Clusia , *Cnemidaria-Hemitelia, *Ctenis, * Dacry- dium, *Dennstaedtia, *Dicksonia, * Dicranopteris Bernh., *Drimys, *Dryopsis Holttum & P. J. Edwards, Empetrum, Escallonia Mutis ex L. f., Euterpe, * Ficus , *Fitzroya Hook. f. ex Lindl., *Gleichenia, * Guettarda , * Gunner a, Gyr anther a Pittier, Hedyosmum, Heliamphora Benth., * Heliconia, Heliocarpus L., *Ilex, **lsoetaeeae, *Isoetes, *Juglans, Laplacea Kunth, * Lecythis Loefl., *Libocedrus Endl., *Lippia, * Lonchocarpus , *Maytenus, * Meliosma , *Miconia Ruiz & Pav., Mora Benth., Murray a J. Konig ex L., *Myrica, * Nephrolepis Schott, * Nothofagus Blume (Cretaceous, Argentina), *0 cotea, *0reo- panax, *0smunda L., Palicourea, Pecopteris Brogn., P cruel- ty a, * Persea, **Poaceae, *Podocarpus (Cretaceous, Argentina, as *"fPodocarpidites, *^Gamerroites), *Polystichum Roth, *Protium, ^P run us, *Psychotria, *Pterocarpus, * Rhus , *Saco- glottis Mart., *Salix, *Sambucus, *Saurauia Willd., Sloanea, Stegolepis Klotzsch ex Korn, * Styrax, *Symplocos, *Tabebuia, Talauma Juss., **Taxodiaeeae (Middle Cretaceous, Antarctica, Athrotaxites , foliage, Falcon-Lang & Cantrill, 2001; * f T axodioxylon, wood), * Tibouchina Aubl., * Toxicodendron Mill., *Trema Lour., * Trigonobalanus Forman, * Turpinia Vent., Tyleria Gleason, * Urtica L., *Vaccinium L., *Vantanea, * Viburnum L., *Weinmannia, *Zamia. Deciduous Forest This refers to plant communities in eastern North America, mixed with riparian and coniferous forest in western North America. *Acer (fruits, Cretaceous, Kirk Johnson locality, Hell Creek Formation, South Dakota, see Renner et al., 2008: 798), cf. *' \Acer arcticum (Late Paleocene, Alberta, Canada, Sapinda- ceae, Hoffman & Stockey, 1999), *f Alismaphyllites (Late Paleocene, Alberta, Canada, **Alismalaceae, Hoffman & Stockey, 1999), * Alnus Mill. (Dillhoff et al., 2005), *f Amersinia (Late Paleocene, Alberta, Canada, **Cornaceae, Hoffman & Stockey, 1999), * Andromeda L. (e.g., Cretaceous, Greenland), Asimina Adans., **Asteraceae, *^Averrhoites (Late Paleocene, Alberta, Canada, extinct, **Oxalidaceae, Hoffman & Stockey, 1999), *f Beringiaphyllum (Late Paleocene, Alberta, Canada, **Cornaceae, Hoffman & Stockey, 1999), * Betula L. (Dillhoff et al., 2005), * Carpinus— Ostrya (Dillhoff et al., 2005, pollen), Volume 98, Number 4 2011 Graham Sequencing New World Ecosystems 537 *Carya Nutt, (putative, Dillhoff et al., 2005), *Castanea Mill., * Chaetoptelea Liebm. (Late Paleoeene, Alberta, Canada, Hoff- man & Slockey, 1999), * Chamaecy paris Spach, *Comus , * Dennstaedtia (Late Paleoeene, Alberta, Canada, Hoffman & Slockey, 1999), Deviacer (Late Paleoeene, Alberta, Canada, **Sapindaceae, Hoffman & Slockey, 1999), *Diospyros L. (Cretaceous, Greenland), * Fag us, *Fraxinus, * Ginkgo L. (Cretaceous, Alberta, Canada; Late Paleoeene, Alberta, Canada, Hoffman & Stockey, 1999), *Gleditsia L., * Glyptostrobus (Cretaceous, Alberta, Canada; Late Paleoeene, Alberta, Canada, Hoffman & Stockey, 1999), *Ilex, *f Joffrea (Late Paleoeene, Alberta, Canada, related to modem *Cercidiphyllum Siebold & Zucc., Hoffman & Stockey, 1999), * Juglans (Dillhoff et al., 2005), *Larix Mill., cf. Lemnospermum (Late Paleoeene, Alberta, Canada, **Araceae), *Lindera Thunb., *Liquidambar, *Liriodendron L, *^Macginicarpa (Late Paleoeene, Alberta, Canada, **Platanaeeae, Hoffman & Stockey, 1999), *Magnolia (Cretaceous, Greenland), * Metasequoia Hu & W. C. Cheng (Cretaceous, Alberta, Canada; Late Paleoeene, Alberta, Canada, Hoffman & Stockey, 1999), *Nyssa, *Osmunda (Upper Cretaceous; Late Paleoeene, Alberta, Canada, Hoffman & Slockey, 1999), *Pinus, **Plalanaceae (* f Erlinghorfia , *f. Plata nites, Johnson, 1996, Late Cretaceous/Lale Maastrich- lian Hell Creek Formation, North Dakota, South Dakota, Montana), * f Platananthus (Late Paleoeene, Alberta, Canada, **Plalanaceae, Hoffman & Stockey, 1999), *Platanus (Late Paleoeene, Alberta, Canada, Hoffman & Stockey, 1999), **Poaeeae, *Populus, *Pterocarya Kunth (Dillhoff et al., 2005), *Quercus (Dillhoff et al., 2005), *Salix, *Taxodium, *Tilia, *Ulmus (Early Middle Eocene, British Columbia, Canada, Dillhoff et al., 2005), Zingiber opsis (Late Paleoeene, Alberta, Canada, Zingiberaceae, Hoffman & Stockey, 1999). NORTHERN TEMPERATE ELEMENTS IN MID-ALTITUDE EASTERN MEXICO * Abies (Dillhoff et al., 2005), *Acer, *Alsophila, Aphanactis Wedd., **Asleraceae, * Berber is, Buddleja L., *Castanopsis (D. Don) Spach, *Catopsis, *Ceanothus , *Cercocarpus, * Chamaecyparis , *Cornus, * Crataegus, *Cupressus, *Cya- thea, *Drimys, **Ericaeeae (* Arbutus, *Arctostaphylos, Men- ziesia Sm., Pernettya, * Rhododendron L.), * Garry a Douglas ex Lindl., *Holodiscus, * Juniper us, *Libocedrus, *Mahonia, Myrcia (pollen as *Eugenia-Myrcia), *Oreopanax, Oxylobus (DC.) Moc. ex A. Gray, *Picea (modern distribution to the mountains of northern Mexico; fossils to Veracmz, Mexico, and Guatemala in the Neogene, Dillhoff et al., 2005), * Pinus (Dillhoff et al., 2005), **Poaeeae, *Podocarpus, *Prunus, Pseudotsuga Carriere, *Quercus, *Ribes L., * Sequoia Endl., *Smilax, *Tsuga (Dillhoff el al., 2005); moist/riparian: *Alnus, * Be tula, *Clethra, * Cornus , * Fraxinus , *Ilex, *Juglans, *()reopanax, * Phoradendron Nutt., *Platanus, * Polypodium L., *Populus, *Prunus, * P sittacanthus Mart., *Salix, * Struthanthus , *Styrax, *Symplocos, *Tillandsia, *Umbellularia (Nees) Nutt., Xylosma G. Forst.; highlands: * Artemisia (lower elevations), *Calocedrus Kurz l*Libocedrus, * Cupressus , **Cyperaceae, **Ericaceae ( Gaultheria L., Kalmia, * Rhododendron, *Vaccinium), Erigeron L., *Picea, * Pinus, **Poaceae, * Prunus, * Pseudotsuga, *Sequoiaden- dron J. Buchholz, * Sphagnum (upper elevations), *Taxus L., * Thuja , *Tsuga. APPALACHIAN CONIFEROUS FOREST * Abies, *Betula, **Ericaeeae (* Rhododendron), *Picea, * Pinus. GULF COASTAL (PINE) CONIFEROUS FOREST * Pinus. * Abies Mill., *Acer, *Alnus, *Carpinus, * Cornus, *Diospy- ros, *Fagus, *llex, *Juglans, * Liquidambar , * Lycopodium , * Magnolia, *Myrica, Os try a, * Pinus, *Platanus, *Populus, * Prunus, Pteridium, *Quercus, *Rubus, *Salix, *Sambucus, *Smilax, *Ulmus. WARM-TEMPERATE ELEMENTS WITH NORTHERN TEMPERATE ELEMENTS IN MID-ALTITUDE EASTERN MEXICO *Alfaroa (as *Alfaroa-Oreomunnea), *Clethra, * Cyathea , * Drimys , Eugenia (pollen as *Eugenia-Myrcia), * Hedy os- mum, * Heliconia , *Meliosma, Peperomia, *Persea, *Podo- carpus, *Weinmannia. Coniferous Forest BOREAL CONIFEROUS FOREST * Abies, *Alnus, *Betula (family Betulaceae and genus from several sites in northern North America; see Graham, 1999: 162—233; oldest records are from Middle Eocene Allenby Formation, British Columbia, Canada; Crane & Slockey, 1987), *Larix, ^ Pice a A. Dielr., * Pinus, 'J>'Populus, *Salix, * Sphagnum, * Thuja L., * Tsuga (Endl.) Carriere. WESTERN MONTANE/ALPINE CONIFEROUS FOREST Many putative identifications (*, **) have come together for the first time in the Middle Eocene of northwestern North America. PINE-OAK (MEXICO) CONIFEROUS FOREST * Alnus , *Cuphea, * Juglans, * Pinus, * Populus , *Quercus, * Struthanthus . Alpine Tundra/Paramo Many putative taxa (*) are from the Pliocene and Quaternary. NORTH AMERICA/ANTILLES/CENTRAL AMERICA Arenaria L., **Asteraceae, Cerastium L., Cirsium Mill., Draba L., Eryngium L., Gnaphalium L., **Juncaeeae {Luzula DC.), Lupinus L., Oxylobus, Phacelia Juss., *Plantago L. (e.g., Late Pliocene, Alaska), **Poaceae, * Potentilla L., Puya Molina, * Ranunculus L. (Quaternary records, and in the Pliocene of Colombia), Senecio (possibly among some pollen identified as *Asteraceae/Compositae). SOUTH AMERICA (INCLUDING SHRUBBY PLANTS OF THE PUNA) Acaena Mutis ex L., * Arenaria, **Asteraceae, Azorella Lam., *Baccharis, Buddleja, * Calceolaria L., **Cyperaeeae, Distichia Nees & Meyen., * Draba (North America), Escallo- nia, Espeletia Mutis ex Bonpl., * Fuchsia L., * Gaultheria L., *Gentiana L., Gynoxys Cass., Hedyosmum, *Hypericum, *Jamesonia, * Lupinus, *Miconia, *Montia L., ^ Plant ago, **Poaceae (Aciachne Benth., Calamagrostis Adans., Chus- quea Kunth [including Swallenochloci McClure], Deyeuxia 538 Annals of the Missouri Botanical Garden CATALOG AND LITERATURE Alan Graham2 GUIDE FOR CRETACEOUS AND CENOZOIC VASCULAR PLANTS OF THE NEW WORLD1 There is need for a sourci electronic form on fossil plants for purposes of systematics, biogeography, paleobotany, and paleo- ecology. The principal impediment to such projects has been experience with past efforts that proved too complex to sustain given the resources available (e.g., The International Organization of Palaeobotany’s [IOP] Plant Fossil Record, , later ; Boulter et al., 1991, Holmes, 1991). A number of new electronic broader information field such as the holdings of an institution focused on particular geographic regions (the Paleobotany Project of the Denver Museum of Nature and Science; ), woods (), pollen and spores (PalDat, the palynological database (), or Cenozoic angiosperms (the developing Cenozoic Angiosperm Database within the Paleobiology Database, ). The latter will attempt to compile fossil records for all organisms, for all time, and for all places. Numerous other institutions are placing images of fossil specimens online, and the National Science Foundation is encouraging these efforts in the United States through the program for Advancing Digitization of Biological Collections (ADBC). It is clear that until these mega projects are sufficiently well advanced to provide relatively complete records, there is need for consulted. The catalog announced here () fo- cuses on the plant fossil record as a guide to calibrating phylogenies, detecting biogeo- family assignment according to the Angiosperm Phylogeny Group (2009) classification, or if a form name is used, the biological affinities as cited by the author (if no affinity is cited at least to the level of family, the record is not included). In addition, the plant part or evidence on which the identification was made (including biochemical fossils and spectral patterns if derivation is cited to genus or family), age, Traditional nonelectronic sources currently exist, and although mostly out-of-date, they can be consulted as an initial guide to early reports and literature until other catalogs become more complete. These are Andrews (1970; with supplements by Blazer, 1975, and Watt, 1982), LaMotte (1952), Muller (1981), Archangelsky et al. (2000, et seq.), Germeraad et al. (1968), and Graham (1973, 1979, 1982, 1986, 1999, 2010: app. 1, 2). Examples from the catalog are presented below for the ferns, gymnosperms, and angiosperms. Filicineae f Conatiopteris schumanii (Cyatheaceae): Lower Cretaceous (Aptian), upper Chickabally Member, Budden Canyon Formation, near Cottonwood, Cal- ifornia, U.S.A., trunk, frond bases, adventitious roots; Lantz et al., 1999: 361: [Cladistic analyses “indicate that the new genus is nested among a paraphyletic assemblage of dicksoniaceous, lophorosiaceous, and metaxylaceous species that subtend a monophyletic Cyathaceae s.s.”] f Kuylisporites (f K. mirabilis, f K. separatus, f K. scutatus, fA. waterbolkii; Cyatheaceae): Cretaceous, Cenozoic, widespread, spores; Mohr & Lazarus, 1994: 765: [“. . . genus originated in the Northern Hemisphere in the late Cretaceous; K. waterbolkii . . . closely related to or identical with extant Cnemida- Chamaecyparis Spach sp. (Cupressaceae): Middle Eocene, Thunder Mountain flora, central ID, vegeta- tive axes; Erwin & Schom, 2005: 125: [“Radiomet- rically dated at 46-45 Ma, it represents one of the 1 Visit the catalog at . 2 Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A. alan.graham@mobot.org. doi: 10.3417/2011083 Ann. Missouri Bot. Gard. 98: 539-541. Published on 18 May 2012. Geol. J. Sci. REINSTATEMENT AND EXPANSION OF THE GENUS PERISTETHIUM (LORANTHACEAE)1 Volume 98, Number 4 2011 Kuijt 543 Reinstatement and Expansion of Peristethium is herewith proposed to realign the species involved in the C. archeri-S. leptostachyus intergeneric bridge as a separate genus, for which the name Peristethium is available. This includes the assignment of several species to Peristethium thus far resident in Struthan- thus ( S . aequatoris Kuijt, S. leptostachyus, S. polystachyus (Ruiz & Pav.) G. Don, and S. tortistylus Kuijt) or in Cladocolea (C. archeri, C. nitida Kuijt, C. peruviensis Kuijt, C. primaria Kuijt, and C. rorai- mensis (Steyerm.) Kuijt). Additionally, five newly discovered species from Andean South America are herein described and illustrated. As now constituted by these 15 species, the genus Peristethium demon- strates coherent, structural, and geographic integrity. Several features, seen in individual species, are unique or nearly so in New World Loranthaceae, even in the family generally, and include an inflorescence most often terminated by a single flower that is preceded by at least one pair of lateral 1 -flowered units called monads, an inflorescence bearing a partially caducous) scale leaves, and extremely small, mostly sessile and basifixed anthers that are positioned near the tips of the subtending petals. The combination of these features is unique in Loranthaceae. Peristethium includes both dioecious species as well as some with bisexual flowers; unusual in Neotropical Loranthaceae, some species are tetramerous and others hexamerous. In retrospect, it can be seen that Peristethium as a natural assemblage has been glimpsed in the past without being properly appreciated. The protologue of Struthanthus tortistylus Kuijt suggested an affinity to P. leptostachyum (Kunth) Tiegh. and P. poly- stachyum (Ruiz & Pav.) Kuijt, these species then still being placed in Struthanthus (as S. leptosta- chyus and S. polystachyus ; Kuijt, 2003a). A study of foliar sclerenchyma pointed ir suggesting taxonomic connections between P. ar- cheri (A. C. Sm.) Kuijt, P. leptostachyum, and the two new nomenclatural combinations presented here, P. primarium (Kuijt) Kuijt and P. roraimense (Steyerm.) Kuijt (as Cladocolea primaria and C. roraimensis; Kuijt & Lye, 2005). Parallel statements are to be found in the protologue of C. primaria (Kuijt, 1987a). The present treatment resolves such Van Tieghem’s protologue of Peristethium (Van Tieghem, 1895) is exceedingly sparse and based on the original Loranthus leptostachyus Kunth alone. He repeated Martius’s error by describing the flowers as bisexual, even though this had been corrected by Eichler (1868). The new genus was founded primarily on the several pairs of partly persistent, chaffy bracts that envelop the young inflorescence, but Van Tieghem, without providing any details, also stated that the stamens are of one type; in reality, they are always placed at two somewhat different heights. Van Tieghem’s errors (as applied to the only species then recognized) were repeated by Engler and Krause (1935), who added a further one, namely the threadlike shape of the filaments, all of which are said to be of equal length. As indicated in the generic diagnosis below, bisexual flowers are characteristic of at least one species not known to the above authors, and anthers are sessile or nearly so, threadlike filaments not being present. The rationale for recognizing Peristethium as a segregate from Struthanthus and as here circum- scribed is based on several structural characters. The most prominent among these is the above-mentioned development of several to many pairs of conspicuous chaffy scale leaves at the base and along the inflorescence axis, the axial ones (bracts) subtending triads and monad. Most of the scales along the fertile axis tend to be caducous, but basal scale leaves often persist. No species of Struthanthus possess such scale leaves, but several other, unrelated Neotropical Loranthaceae show similar but much smaller and mostly deciduous scale leaves along the lower inflorescence axes, as in Panamanthus Kuijt. In the Old World, a similar, independent evolution of modified scalelike leaves below the flowers has occurred, as in Diplatia Tiegh., Tolypanthus (Blume) Blume, and several other genera (Kuijt, 1981b). However, the scale leaves in these Old World genera are much larger, often brightly colored, and not be distinguished from the more inconspicuous pairs of prophylls at the base of axillary ramifications of & Steyerm. and P. sonorae (S. Watson) Kuijt (Kuijt, 2009). A second important generic character in Peri- stethium is the occurrence in each inflorescence of a single morphologically terminal flower, commonly The inflorescence is thus determinate. Paucity of materials has prevented me from confirming this aspect in some species accepted here for Peri- stethium. In at least one unrelated species, Struthan- thus deppeanus (Schltdl. & Cham.) G. Don in Mexico, the same determinate pattern exists (Kuijt, 1981b), but elsewhere in small-flowered genera in the A third important generic feature in Peristethium is the sessile or nearly sessile, exceedingly small 544 Annals of the Missouri Botanical Garden anther, inserted well above the middle of the petal (Kuijt, 1975b). In Mexico, it is also known in that bears it. Struthanthus characteristically has Struthanthus ; in fact, the generic synonym Spirostylis larger, versatile anthers on long, slender filaments. was based on this stylar twisting in what is now The other small-flowered genera in the Loranthaceae known as S. interruptus (Kunth) G. Don (Kuijt, possess entirely different androecia, especially Den- 1975a). Struthanthus tortistylus from Ecuador (2003a) dropemon (Blume) Rchb. (Kuijt, 2011a) and many was published before it was realized that contorted species of the genus Passovia H. Karst.; the latter styles are also characteristic of female P. polysta- have minute anthers, but can scarcely be held as chyum , a fact not previously mentioned in the related to Peristethium. The small size of the anthers, literature. One further species with strongly contorted and the prominent sterile anthers in male plants, for styles, an as yet undescribed species of Struthanthus example, of P. leptostachyum , can be problematic in from Peru, has recently been discovered. Thus we can the determination of sex, as shown by the errors made be assured that there are at least three instances in by Van Tieghem (1895) and Engler (1897) in which this curious, unexplained feature has evolved referring to that species as having bisexual flowers. independently in Neotropical, small-flowered genera. Nevertheless, species with bisexual flowers as well as The phenomenon is not known from other Lorantha- dioecious species exist in Peristethium. Further, the Ceae in either the New or Old Worlds, with the slender shape of the seedling in Peristethium , lacking solitary exception of Ileostylus micranthus (Hook, f.) a swollen haustorial pole, has also emerged as a Tiegh. of New Zealand (Barlow, 1966). The biological contrasting feature vis-a-vis most other small-flow- significance of contorted styles is not known, but it ered genera. Finally, the existence of foliar scleren- may bear a relation to the extraordinary behavior of chyma is a common denominator for those species the embryo sac known from Foranthaceae generally that have been investigated (Kuijt & Lye, 2005). The (see Kuijt, 1969, for a summary). profusion of stellate foliar sclereids is especially As now constituted, Peristethium also has con- remarkable in P. roraimense , where much of the leaf yincing geographical integrity: its species range from mesophyll consists of such cells. Amazonian Bolivia north through the Andes into The variation in inflorescence structure seen in Costa Rica, with two endemic, highly localized with the outliers on Mt. Roraima and the Pakaraima Moun- Peristethium is in general agreement evolutionary trends previously outlined for the tains (pig. 1). It is generally accepted that the upper Loranthaceae (Kuijt, 1981b). In that study, I regions of the lepuis of Venezuela and Guyana concluded that the evolution of their inflorescences environmentally correspond to the paramo life zone of led from monads to aggregation as triads, and that lhe norlhern Andes (Berry et al., 1995), and the this process begins at the base of the inflorescence. existence of the two rare species, P. roraimense and This latter tendency corresponds to the change from p mtldum (Kuijt) Kuijt, in the Mt. Roraima area is plants with bisexual flowers to a dioecious condition, not unexpected. The phylogenetic affinities of and that hexamery was derived from Letramery. With Peristethium with other small-flowered genera in the regard to Lriadization, we may see all stages in the Loranthaceae are at present unclear, although the process represented in extant species of Peristethium. coherence of Psittacanthinae, which also includes the For example, P. archeri , P. peruviense (Kuijt) Kuijt, large-flowered genera Aetanthus (Eichler) Engl, and and P. roraimense have strictly monadic inflores- Psittacanthus Mart., has received molecular affirma- cences. In P. primarium , P. palandense Kuijt, and lion (Vidal-Russell & Nickrent, 2008; Nickrent et al., several others, triads are present basally while 2010). However, the precise relationships among the monads are found subterminally and variably. various genera remain l0 be elucidated. Morpholog- Finally, P. confertiflorum may have only triads, with icjd mformallon, especially the occurrence of tetram- the additional, unanticipated production of some eroug flowers? determinate inflorescences, and mo- pentads. With the exception of P. primarium , all nadg lacking bracteoles, at present tetramerous species also exclusively have monadic rr- •, r n • , n- .1 ri j i r J attimty oi reristetnium with Ltaaocotea. inflorescences, thus combining two putatively ances- tral conditions. The curiously contorted style described and illustrated for Peristethium polystachyum and P. Peristethium Tiegh., Bull. Soc. Bot. France 42: 175. an Taxonomic Treatment tortistylum Kuijt (Kuijt) has equivalents in several other Neotropical genera. The majority of Mesoamer- ican species of Cladocolea show this feature, seen more strongly in the female flower than the male 1895. TYPE: Peristethium leptostachyum (Kunth) Tiegh. [= Loranthus leptostachyus Kunth, Nov. Gen. Sp. [H.B.K.] (quarto ed.) 3: 440. 1818 (1820)]. r