Phylogenetic Relationships within Ruellieae (Acanthaceae) and a Revised Classification

Erin A. Tripp; Thomas F. Daniel; Siti Fatimah; Lucinda A. McDade

Phylogenetic knowledge of the large plant family Acanthaceae has been greatly advanced over the last 2 decades. Studies have demonstrated the existence of several major lineages, most of which have been the focus of subsequent investigation. Missing among these is comprehensive study of the 48 genera currently classified in tribe Ruellieae, a pantropical lineage that includes several species-rich genera. We compared the number of validly published names to current estimates of species richness per genus in Ruellieae and found more than 2600 names available for ;1200 species. Using molecular data from two nuclear (ITSþ5.8S, Eif3E) and three chloroplast (trnG-trnR, trnG-trnS, psbA-trnH) markers, we test the placement of these 48 genera in Ruellieae, explore the monophyly of currently recognized taxa, and propose morphological features to diagnose major clades within the tribe. We were able to sample all but four of 48 genera, and all were resolved in Ruellieae except Zygoruellia. Many monospecific or oligospecific genera are nested within clades of more species-rich genera. We propose several new generic synonymies to reflect these results and insights from morphology. Finally, we present a revised classification of Ruellieae that contains seven subtribes. A solid phylogenetic hypothesis of relationships within Ruellieae contributes to progress in biology in three important ways: (1) it enables better assessment of trait homologies and thus characters upon which genera are delimited, (2) it contributes to the Global Strategy for Plant Conservation's initiative to document plant biodiversity, and (3) it facilitates cross-family comparative evolutionary analyses, including large-scale hypothesis testing of biogeographic patterns, clade size asymmetries, and differential diversification within Acanthaceae.

Phylogenetic Relationships within Ruellieae (Acanthaceae) and a Revised Classification Author(s): Erin A. Tripp, Thomas F. Daniel, Siti Fatimah, and Lucinda A. McDade Reviewed work(s): Source: International Journal of Plant Sciences, Vol. 174, No. 1 (January 2013), pp. 97-137 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/668248 . Accessed: 29/01/2013 06:43 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to International Journal of Plant Sciences. http://www.jstor.org This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Int. J. Plant Sci. 174(1):97–137. 2013. Ó 2013 by The University of Chicago. All rights reserved. 1058-5893/2013/17401-0008$15.00 DOI: 10.1086/668248 PHYLOGENETIC RELATIONSHIPS WITHIN RUELLIEAE (ACANTHACEAE) AND A REVISED CLASSIFICATION Erin A. Tripp,1 ,*,y Thomas F. Daniel,z Siti Fatimah,* and Lucinda A. McDade* *Rancho Santa Ana Botanic Garden, 1500 N. College Avenue, Claremont, California 91768, U.S.A.; yDepartment of Ecology and Evolutionary Biology and Museum of Natural History, University of Colorado, UCB 334, Boulder, Colorado 80309, U.S.A.; and zCalifornia Academy of Sciences, 55 Music Concourse Drive, San Francisco, California 94118, U.S.A. Phylogenetic knowledge of the large plant family Acanthaceae has been greatly advanced over the last 2 decades. Studies have demonstrated the existence of several major lineages, most of which have been the focus of subsequent investigation. Missing among these is comprehensive study of the 48 genera currently classified in tribe Ruellieae, a pantropical lineage that includes several species-rich genera. We compared the number of validly published names to current estimates of species richness per genus in Ruellieae and found more than 2600 names available for ;1200 species. Using molecular data from two nuclear (ITSþ5.8S, Eif3E) and three chloroplast (trnG-trnR, trnG-trnS, psbA-trnH) markers, we test the placement of these 48 genera in Ruellieae, explore the monophyly of currently recognized taxa, and propose morphological features to diagnose major clades within the tribe. We were able to sample all but four of 48 genera, and all were resolved in Ruellieae except Zygoruellia. Many monospecific or oligospecific genera are nested within clades of more species-rich genera. We propose several new generic synonymies to reflect these results and insights from morphology. Finally, we present a revised classification of Ruellieae that contains seven subtribes. A solid phylogenetic hypothesis of relationships within Ruellieae contributes to progress in biology in three important ways: (1) it enables better assessment of trait homologies and thus characters upon which genera are delimited, (2) it contributes to the Global Strategy for Plant Conservation’s initiative to document plant biodiversity, and (3) it facilitates cross-family comparative evolutionary analyses, including large-scale hypothesis testing of biogeographic patterns, clade size asymmetries, and differential diversification within Acanthaceae. Keywords: biodiversity, clade size asymmetry, Global Strategy for Plant Conservation, diversification, molecular, nomenclature, systematics, taxonomy, tropical. Introduction groups and of the delineation of those groups. For example, Manktelow et al. (2001) found evidence for a sixth major clade, and McDade et al. (2008) showed that a few genera had been misplaced in Scotland and Vollesen’s (2000) classification. Missing among phylogenetic studies is investigation of the large (;1200 species), pantropical Ruelliinae (sensu Scotland and Vollesen 2000) or Ruellieae, as referred to here and elsewhere (Tripp 2007). Although several clades within Ruellieae have been studied (Manktelow 1996; Carine and Scotland 2000; Moylan et al. 2004a; Schmidt-Lebuhn et al. 2005; Vollesen 2006; Tripp 2007; Tripp et al. 2009; Tripp and Fatimah 2012), to date there exists no broad-scale, comprehensive investigation of the 48 genera (table 1) treated in the group by Scotland and Vollesen (2000). Since Nees (1847b) published the first comprehensive classification of Acanthaceae, few major revisions have been presented (Lindau 1895; Bremekamp 1965; Scotland and Vollesen 2000). The comprehensive treatment by Scotland and Vollesen (2000) accepted 221 genera and circumscribed five major groups within Acanthaceae s.s. (i.e., the clade of plants with seeds born on modified woody, hooklike funiculi, termed retinacula, within explosively dehiscent capsules; Thunbergia, Mendoncia, Nelsonia, Avicennia, and relatives are not included in this ‘‘retinaculate clade’’ [fig. 1]). The retinaculate clade comprises the vast majority of species richness in Acanthaceae s.l. Four of the five retinaculate groups recognized by Scotland and Vollesen (2000) have been examined in at least one fairly densely sampled molecular phylogenetic study: (1) Justiciinae (Justicieae sensu McDade et al. 2000 and herein), (2) Acantheae (McDade et al. 2005), (3) Barleriinae (Barlerieae sensu McDade et al. 2008 and herein), and (4) Andrographiinae (Andrographidieae sensu McDade et al. 2008 and herein). These studies serve as tests of Scotland and Vollesen’s (2000) classification of Acanthaceae s.s. (¼Acanthoideae Link sensu Lindau 1895) into five 1 Circumscriptions of Ruellieae Nees (1847b) recognized 36 genera in his Ruellieae. Eight of these were retained by Scotland and Vollsen (2000) in their Ruelliinae; all others treated by Nees are now considered to be synonyms or are treated in other tribes. Bentham (1876) recognized 35 genera in his Ruellieae, 21 of which were retained by Scotland and Vollesen (2000). Lindau (1895) recognized 47 genera in his Contortae, 27 of which were retained by Scotland and Vollesen (2000). Bremekamp (1965) Author for correspondence; e-mail: etripp@rsabg.org. Manuscript received March 2012; revised manuscript received August 2012. 97 This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 98 INTERNATIONAL JOURNAL OF PLANT SCIENCES portant horticulturally (e.g., several Ruellia, Strobilanthes, Sanchezia) or are widely planted as street trees or living fences (e.g., Bravaisia, Trichanthera). Many species are also known for their usefulness in local traditions and medicines (e.g., an Indian Hygrophila is used as an aphrodisiac and sedative; Pawar et al. 2006; also see Manktelow 1996 for numerous uses of species of African Phaulopsis). Morphological Synapomorphies for Ruellieae Fig. 1 Summary phylogeny of the current understanding of relationships within Acanthaceae; tree reflects results of the most recent comprehensive phylogenetic study across the family (McDade et al. 2008). Strongly supported branches (99% Bayesian posterior probability) are in bold. Acanthaceae s.s. refers to plants with retinacula. Within Acanthaceae s.s., all plants have cystoliths except those in tribe Acantheae. Derived traits shared among Ruellieae but lacking in other Acanthaceae have been contemplated, but a consensus regarding definitive synapomorphies has yet to be reached, owing largely to the lack of comprehensive study across all members of the tribe. Possible synapomorphies for Ruellieae that have been advanced and debated include left-contort corolla aestivation, seeds with mucilaginous hygroscopic trichomes (fig. 2A–2C), the presence of a ‘‘filament curtain’’ (i.e., a physical partition within the corolla formed by the adnation of the filaments to the corolla; fig. 2D, 2E), unequal stigma lobes, or some combination of these traits (Grubert 1974; Scotland et al. 1995; Manktelow 2000; Manktelow et al. 2001; Moylan et al. 2004a; Tripp 2007). Although some of these characters have been studied extensively in multiple genera, others are poorly known, and character-state data have not been synthesized across Ruellieae. Taxon Delimitation within Ruellieae treated many of the same genera as Lindau (1895) in his Ruellieae; the remainder are currently recognized elsewhere or considered to be synonyms. However, as Bremekamp and Nannenga-Bremekamp (1948) earlier acknowledged, ‘‘As a large part of the genera belonging to the Ruellieae [sic] are still imperfectly known, a fully satisfactory subdivision of this tribe can not yet be given’’ (p. 3). From here forward, we consider 47 of Scotland and Vollesen’s 48 genera because recent study has demonstrated that one of them, Physacanthus, represents a heteroplasmic, intergeneric hybrid involving members of tribes Ruellieae and Acantheae (Tripp et al., forthcoming). Diversity and Distribution Species of Ruellieae are diverse morphologically, ecologically, and geographically. Plants of most species are herbs or woody shrubs, but some are lianas and others are trees (e.g., Bravaisia and Trichanthera; trees are otherwise uncommon in the family). Plants occupy a variety of tropical and subtropical (some temperate) habitats, ranging from rain forests to dry forests to arid shrublands to deserts. Mexico, Brazil, and Peru are centers of species richness in the New World (NW), and Tanzania, Madagascar, and southern to southeastern Asia are species-rich areas in the Old World (OW). Many Ruellieae are narrow endemics, yet others are widespread. In some African landscapes (e.g., Namibia), plants of Ruellieae are among the dominants and serve as a major source of fodder for herbivorous megafauna. In contrast, Ruellieae in the NW generally do not achieve such great population densities, although their presence in a large portion of tropical and subtropical NW habitats points to their ecological importance. Many Ruellieae are im- Within Ruellieae, genera have been delimited or variously assembled into subgroups based primarily on pollen type (Scotland 1992; Furness 1994, 1995; Daniel 1998; Tripp 2007) but also on corolla limb symmetry (Bentham 1876), presence of style-retaining trichomes inside corollas (Bremekamp 1944a), number of stamens (Nees 1832), presence or absence of anther appendages (Nees 1832), number of ovules (Anderson 1867; Wood 1994), projections on nectaries (Bremekamp 1938), and even trichome differences (Clarke 1908). Pollen morphology, in particular, has been emphasized ever since Lindau’s (1893, 1895) treatment across Acanthaceae. Of note is Scotland’s (1993) thorough study of pollen variation within Ruellieae, with pollen of 24 genera figured. Other studies have provided useful images of pollen from other Ruellieae, including Furness (1994, 1995), Furness and Grant (1996), Daniel (1998), Carine and Scotland (1998), Scotland and Vollesen (2000), Moylan and Scotland (2000), Schmidt-Lebuhn (2003), Vollesen (2006), Tripp (2007), Chen et al. (2009), Tripp et al. (2009), Greuter and Rankin Rodrı́guez (2010), and Tripp and Fatimah (2012). Of 47 genera in Ruellieae, pollen images have not been published, to our knowledge, for 15 of them. Challenges to and Motivations for Study One major hindrance to studying relationships within and evolution among Ruellieae is scarcity of relevant specimens. Inferences about genera that contain many species are made possible by the existence of ample accessible herbarium material. In Ruellieae, however, 13 of the 48 genera are monospecific, and another 24 are oligospecific (i.e., contain only two or three species). The vast majority of these mono- and oligospecific This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Table 1 Distribution and Diversity Information for the 48 Genera Classified in Ruellieae (Ruelliiinae sensu Scotland and Vollesen 2000) before This Study Ruelliinae sensu Scotland and Vollesen (2000) Distribution No. taxa (includes infraspecific) Acanthopale C.B. Clarke Aechmanthera Nees Apassalus Kobuski AF, MA IC NW 7 1 [Strobilanthes] 5 [Dyschoriste in part] Benoicanthus Heine & A. Raynal Blechum P. Brown Bravaisia DC. Brillantaisia P. Beauv. Brunoniella Bremek. MA NW NW AF, MA AA 3 6 [Ruellia] 3 12 7 Calacanthus T. Anderson ex Benth. Clarkeasia J.R.I. Wood Dischistocalyx T. Anderson ex Benth. Duosperma Dayton Dyschoriste Nees Echinacanthus Nees Epiclastopelma Lindau Eranthemum L. Eremomastax Lindau Eusiphon Benoist Hemigraphis Nees Heteradelphia Lindau Hygrophila R. Br. Ionacanthus Benoist Kosmosiphon Lindau Leptosiphonium F. Muell. Louteridium S. Watson Lychniothyrsus Lindau Mellera S. Moore Mimulopsis Schweinf. Pararuellia Bremek. Petalidium Nees Phaulopsis Willd. Physacanthus Benth. Polylychnis Bremek. Pseudoruellia Benoist Ruellia L. Ruelliopsis C.B. Clarke Sanchezia Ruiz & Pav. Satanocrater Schweinf. Sautiera Decne. Spirostigma Nees Stenosiphonium Nees Stenothyrsus C.B. Clarke Strobilanthes Blume Strobilanthopsis S. Moore Suessenguthia Merm. Trichanthera Kunth Trichosanchezia Mildbr. Zygoruellia Baill. Total IC AA, IC AF AF PAN AA, IC AF AA, AF, IC AF MA AA, IC AF PAN MA AF AA NW NW AF AF, MA AA, IC AF (1 India) AF, MA AF NW MA PAN AF NW AF AA NW IC (India) AA AA, IC AF NW NW NW MA 1 2 15 26 75 4 3 15 1 3 [Ruellia in part] 60 2 100 1 1 10 10 1 7 15 10 32 26 3 2 1 350 1 59 4 1 1 3 1 350 2 6 2 1 1 1185 Citation for no. taxa Herein estimated Wood and Scotland (2009) Brother Léon (1957); Greuter and Rankin Rodrı́guez (2010) Heine and Raynal (1968) Tripp et al. (2009) Daniel (1988) Sidwell (1998) Barker (1986); Moylan and Scotland (2000) Herein estimated Wood (1994) Champluvier and Senterre (2010) Vollesen (2006) Daniel (2004) Wood (1994) Vollesen (2008) Wood (1994); Hu et al. (2011) Ensermu (2006); Vollesen (2008) Benoist (1954); Tripp (2007) Bremekamp (1944a) Daniel and Figueiredo (2009) Vollesen (2008) Benoist (1940) Herein estimated Barker (1986) Daniel (1995) Herein estimated Vollesen (2008) Vollesen (2008) Hu et al. (2011) Obermeijer (1936) Manktelow (1996) Tripp et al. (forthcoming) Herein estimated Herein estimated Herein estimated Herein estimated Leonard and Smith (1964) Thulin (2007); Tripp and Fatimah (2012) Herein estimated Herein estimated Carine and Scotland (2000) Wong (1998) Moylan et al. (2004a) Milne-Redhead (1932) Schmidt-Lebuhn (2003) Daniel (2008) Herein estimated Herein estimated No. published names 17 5 5 3 32 7 49 7 2 3 28 29 166 16 3 199 3 4 168 2 158 1 1 11 11 5 9 33 12 50 40 6 4 1 835 3 69 7 2 1 10 1 624 7 10 4 1 1 2665 Note. Distribution codes are as follows: AF (mainland Africa excluding Madagascar), MA (Madagascar), IC (India through China), AA (Australia and/or Southeast Asia excluding China), NW (New World), and PAN (pantropical). There are a limited number of minor exceptions to distributions not reflected in the table (e.g., a few Ruellia occur in temperate [not tropical] habitats; a few Brillantaisia, one Phaulopsis, and one Heteradelphia occur off of the coast of mainland Africa [São Tomé]; etc.). The number of taxa (col. 3) in each genus was estimated using monographs, taxonomic treatments, or, where such works are lacking, estimates provided by authorities in nonmonographic literature or was herein estimated. A bracketed genus following the number of species indicates that the genus has, all or in part, been put into synonymy since Scotland and Vollesen (2000). The number of published names (col. 5) for each genus was compiled using the International Plant Names Index and Tropicos; thus, the count includes intraspecific names as well as names that, subsequent to their publication, have been combined or placed into synonymy. Superfluous names and subordinate autonyms are not included. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 100 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 2 Two of four synapomorphies for Ruellieae: seed hygroscopic trichomes (A–C) and a filament curtain (D–E). A, SEM of appressed hygroscopic trichomes on surfaces of seeds of Ruellia humilis (Tripp 14 [PH]). B, Photograph of hygroscopic trichomes covering entire surface of seeds of Ruellia nudiflora (Daniel 6293 [CAS]). C, Photograph of hygroscopic trichomes restricted to margins of seeds of Louteridium chartaceum (Daniel 5905 [CAS]). D, Internal view of longitudinal section of flower, opened along ventral surface and facing dorsal surface, showing filament curtain of Ruellia elegans. E, View of base (proximalmost portion) of corolla, showing filament curtain of Ruellia bourgaei. fc ¼ filament curtain, ov ¼ ovary, st ¼ style. genera also have restricted distributions within the geographic areas given in table 1, further diminishing the probability that such taxa are adequately represented in major herbaria. A solid understanding of the systematics of Ruellieae is crucial for three reasons. First, molecular phylogenies facilitate assessment of trait homologies and thus the synapomorphies upon which taxa are delimited. Like many other lineages in Acanthaceae, a disproportionate number of genera in Ruellieae occur in the OW (n ¼ 33) compared with the NW (n ¼ 11; three are pantropical), and more than half of the OW genera (17 of 33) contain three or fewer species (table 1). This striking pattern has yet to be evaluated in an evolutionary context. Lacking a phylogeny and comparative morphological study, we cannot test whether OW generic diversity is ‘‘inflated’’ versus whether current classifications reflect true differences in evolutionary history between the hemispheres, much less pose explanatory hypotheses for the pattern. Second, the Global Strategy for Plant Conservation (GSPC) grew out of the 1999 International Botanical Congress and was adopted by the Convention on Biological Diversity in 2002. Of 16 GSPC targets for 2010, the first was to produce ‘‘a widely accessible working list of known plant species . . . towards a complete world flora’’ (Paton et al. 2008). These authors analyzed progress to date toward target 1, noting good progress for bryophytes, ferns, and gymnosperms but relatively less for angiosperms. Of ;350,000 flowering plants, Paton et al. (2008) estimated a gap in coverage of ;178,000 species (50%). They identified the major families constituting this gap, listing 32 that contributed more than 1% to it, and determined that Acanthaceae was the sixth largest contributor. Partial achievement of GSPC target 1 for Acanthaceae will be reconciling the difference between the thousands of species in the family and the tens of thousands of names that exist for them. In Ruellieae, we here estimate that there are more than two times as many names available as taxa that are currently recognized (see the total row in table 1). Given the taxonomic working method (i.e., ‘‘splitting’’) of some previous Acanthaceae specialists, we suspect that a substantial number of genera and species in Ruellieae should be reduced to synonymy. However, paucity of knowledge regarding morphological This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 101 Table 2 Results of Alternative Phylogenetic Hypothesis Testing Hypothesis H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 H11 H12 H13 H14 lnL lnL Reject unconstrained constrained Difference Null Alternative Leptosiphonium not monophyletic (Chinese and PNG accessions not sister) Sanchezia not monophyletic (includes Suessenguthia) Hygrophila not monophyletic (two clades resolved) Heteradelphia not monophyletic (includes Eremomastax) Eremomastax not monophyletic (includes Heteradelphia) Mellera not monophyletic (includes Ionacanthus) Mimulopsis not monophyletic (includes Epiclastopelma) Dyschoriste not monophyletic (includes Apassalus) Strobilanthes not monophyletic (includes several other genera) Ruellia not monophyletic (includes several other genera) Zygoruellia allied to Whitfieldieae Leptosiphonium monophyletic Yes 41,556.44 41,598.89 42.45 .00 Sanchezia monophyletic Yes 41,556.44 41,763.56 207.12 .00 Hygrophila monophyletic Yes 41,556.44 41,589.77 33.33 .00 Heteradelphia monophyletic Yes 41,556.44 41,609.44 53.00 .00 Eremomastax monophyletic Yes 41,556.44 41,592.80 36.36 .01 Mellera monophyletic Yes 41,556.44 41,581.81 25.37 .00 Mimulopsis monophyletic No 41,556.44 41,558.38 1.94 .40 Dyschoriste monophyletic No 41,556.44 41,572.95 16.51 .05 Strobilanthes monophyletic Yes 41,556.44 41,598.96 42.52 .01 Ruellia monophyletic Yes 41,556.44 41,675.81 119.37 .00 Zygoruellia allied to Ruellieae Dischistocalyx sister to Acanthopale Satanocrater sister to Ruellia Yes 41,556.44 41,763.56 207.12 .00 Yes 41,556.44 41,615.99 59.55 .00 Yes 41,556.44 41,817.80 261.36 .00 Ruellieae not monophyletic Yes 41,556.44 41,598.89 42.45 .00 Dischistocalyx sister to Satanocrater þ (Acanthopale þ Ruellia) Dischistocalyx sister to Satanocrater þ (Acanthopale þ Ruellia) Ruellieae monophyletic P Note. Hypotheses H1–H10 test the monophyly of genera; H11 tests alternative placement of Zygoruellia in Ruellieae; H12–H13 test alternative resolution among the four genera that comprise Ruelliinae (see ‘‘Discussion’’ for rationale behind tests); and H14 tests an alternative hypothesis of nonmonophyly of Ruellieae. H1–H13 are positive constraint tests, whereas H14 is a negative constraint test. Note that for the purposes of H14, Zygoruellia was not considered part of Ruellieae. P values rounded to two decimals. PNG = Papua New Guinea. limits and relationships among genera precludes our assessment of this suspicion. An accurate estimate of species richness in Ruellieae, as a contribution to target 1 of GSPC, will be possible only after thorough investigation of phylogenetic and trait homology across the entire lineage. Finally, we have achieved a level of knowledge of the systematics of Acanthaceae, which are among the 12 or so most species-rich families of angiosperms, that will soon position researchers to undertake family-wide comparative analyses of trait evolution and species diversification. Yet we lack the relevant data in Ruellieae, a lineage that comprises at least 25% of species in the family. Delimiting monophyletic genera and reconstructing trait evolutionary history will be critical, for example, to test hypotheses regarding clade size asymmetries in lineages from different hemispheres, with different pollination systems, with different life strategies, and so on. At present, this large and important lineage is in danger of being left out of family-wide analyses. Table 3 Descriptive Information for Molecular Data Sets Used in This Study Eif3E No. accessions Aligned length No. variable characters (%) No. parsimony informative characters Pairwise divergence (all taxa) Pairwise divergence (Ruellieae only) Missing data (%) 37 1334 579 (43) 430 NA .02–.35 77 ITSþ5.8S 104 839 517 (62) 314 .00–.32 .00–.26 35 psbA-trnH 100 822 429 (52) 308 NA .00–.32 37 trnG-trnR 124 1164 555 (48) 332 .00–.13 .00–.10 22 trnG-trnS 99 1373 814 (59) 559 .00–.27 .00–.26 38 Combined 159 4644 2444 (53) 1605 .00–.29 .00–.29 42 Note. The 159 taxa in the combined analysis include Ruellieae and non-Ruellieae accessions as well as Sesamum. Sesamum was excluded from calculation of pairwise divergence (all taxa), and Zygoruellia was excluded from calculation of pairwise divergence (Ruellieae only). Percent missing data was high in the Eif3E and psbA-trnH data sets because non-Ruellieae accessions were scored as missing data owing to extremely high sequence divergence from Ruellieae (see text). Percent missing data was high in the other matrices because of PCR or sequencing failure. NA ¼ not applicable. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 102 INTERNATIONAL JOURNAL OF PLANT SCIENCES Objectives The goals of this study are fourfold: (1) to test the monophyly of Ruellieae, that is, to test whether Ruellieae as circumscribed by Scotland and Vollesen (2000) are resolved as a clade, via analyses of molecular sequence data from taxa representing all major lineages of Acanthaceae; (2) to build a phylogenetic framework in which morphological synapomorphies for Ruellieae and diagnostic features for clades within Ruellieae can be evaluated; (3) to use this phylogenetic framework to attempt to understand discrepancies in the number of OW versus NW genera and to produce a more accurate estimate of standing taxonomic diversity in Ruellieae; and (4) to present a revised classification of the genera and subtribes comprising Ruellieae. Importantly, research on the systematics of Ruellieae will permit the investigation of patterns of character evolution and correlates of diversification in multiple, pantropical lineages; it will ensure that clades are sufficiently understood so that they can be sampled in comparative studies of evolutionary processes across the family as a whole. Methods Taxon Sampling To test monophyly of Ruellieae, we sampled DNA from 43 of 47 genera treated in the group by Scotland and Vollesen (2000; i.e., excluding the 48th genus, Physacanthus; see above). For the remaining four genera (Calacanthus, Echinacanthus, Spirostigma, Stenothyrsus), we either were unable to obtain material or obtained only material that yielded DNA of insufficient quality to sequence. We also sampled 14 representatives from all other (non-Ruellieae) lineages in Acanthaceae, using McDade et al. (2008) as a guide (fig. 1). These other Acanthaceae included two Justicieae (Rhinacanthus, Mackaya), two Barlerieae (Barleria, Crabbea), one Andrographidieae (Cystacanthus), two Whitfieldieae (Camarotea, Chlamydacanthus), two Neuracanthus, two Acantheae (Aphelandra, Stenandriopsis), one Thunbergioideae (Mendoncia), one Avicennia, and one Nelsonioideae (Staurogyne). This broad sampling spans the phylogenetic diversity of Acanthaceae and permits alternative placement of genera that are putatively Ruellieae. We used Sesamum (Pedaliaceae) to root trees. To test monophyly of taxa within Ruellieae, we sampled multiple species of genera with more than one species, when possible. For monospecific genera, effort was made to include multiple accessions. We attempted to sample across the geographical range and morphological diversity for each genus and to include multiple accessions of genera that have not been previously studied using molecular data. The final data set contained a total of 159 accessions (appendix). Molecular Markers Fig. 3 Majority rule consensus tree derived from Bayesian analysis showing relationships among subtribes of Ruellieae. Thickened branches are supported by 95% Bayesian posterior probability (PP) and/or 70% maximum likelihood bootstrap (ML BS) values. Other major Five molecular markers were employed in this study: two nuclear and three chloroplast. The nuclear markers included lineages of Acanthaceae are also labeled and correspond to those shown in fig. 1. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 103 the ITS1þ5.8SþITS2 region and the low-copy, conserved orthologous Eif3E gene (Li et al. 2008), which functions in translation initiation and is utilized here for only the second time in Acanthaceae (Tripp and Fatimah [2012] used Eif3E in a small phylogenetic study of 12 taxa). Because a substantial portion of our molecular sequences were generated from degraded DNA extracted from herbarium material, we were able to amplify and sequence Eif3E from only a subset of the total taxon sample. Pilot analyses indicated that this subset was representative of major lineages of Ruellieae. The chloroplast spacers utilized were psbAtrnH, trnG-trnR, and trnG-trnS. Primers for ITSþ5.8S and the three chloroplast markers followed Tripp (2007), Tripp and Manos (2008), and Tripp (2010). Primers for Eif3E followed Li et al. (2008); the portion of this region amplified and sequenced spans part of exon 3, all of exon 4, part of exon 5, and the intervening two introns. Data Generation Some data used in this study were generated for previous studies (i.e., a small subset of the ITSþ5.8S, trnG-trnR, and trnG-trnS data). For data newly generated for this study (i.e., the vast majority of sequences from all five markers and nearly all sequences for Eif3E), total genomic DNA was extracted either from herbarium material or from silica-dried fresh leaf samples by the CTAB method (Doyle and Doyle 1987) with modifications or with DNEasy Plant Mini kits (Qiagen, Valencia, CA). Molecular targets were amplified using either GoTaq polymerase (Progmega, Madison, WI) or Phusion (Finnzymes; distributed by New England Biolabs, Ipswich, MA). PCR products either were directly sequenced (i.e., without a PCR product ‘‘cleanup’’ step) or were cloned following PEG precipitation using Topo TA kits (Invitrogen, Carlsbad, CA) and then sequenced. Cloning was necessary only for Eif3E, which is present in at least two copies in Acanthaceae (only one copy was utilized here; no cloning was deemed necessary for nrITS owing to single-banded PCR products). Multiple clones of Eif3E were sequenced for most individuals. Pilot analyses indicated that all clones from one individual were monophyletic, and the Eif3E data set was pruned to contain only one clone per individual. Cycle sequencing products were cleaned using sephadex, and sequences were generated by the authors on an Applied Biosystems 3130 capillary sequencer at Rancho Santa Ana Botanic Garden. We also used one sequence (not generated by us) available in GenBank (Stenosiphonium setosum). GenBank accession numbers are given in the appendix. Sequences were manually aligned using MacClade (ver. 4.08; Maddison and Maddison 2000), and alignments are available in TreeBASE (study S13187). Different character-exclusion regimes (i.e., with varying levels of conservatism with respect to re- Fig. 4 Most likely tree derived from GARLI analysis showing relationships among the subtribes of Ruellieae. Thickened branches are supported by 95% Bayesian posterior probability (PP) and/or 70% maximum likelihood bootstrap values. Other major lineages of Acanthaceae are also labeled and correspond those shown in fig. 1. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 104 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 5 Phylogenetic relationships among Erantheminae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). Taxonomic name changes proposed herein follow current names in brackets; these reflect the revised classification presented herein (see ‘‘Taxonomic Treatment’’). gions where homology assessment was problematic) were implemented to explore the effect of character exclusion on phylogenetic relationships. The final alignment excluded hypervariable regions (i.e., one portion of ITSþ5.8S that was difficult to align across all Acanthaceae), regions of known homoplasious inversions (one in psbA-trnH), regions with simple sequence repeats (one each in psbA-trnH and trnGtrnR, four in trnG-trnS), and regions with extremely long autapomorphic insertions (;300-bp insertion in Eif3E of Zygoruellia). Indels were left unmodified in all alignments (i.e., were not recoded as presence/absence characters). Phylogenetic Analyses This study explores relationships in a lineage in which 18 genera have never before been sampled for phylogenetic analysis (Apassalus, Baphicacanthus, Benoicanthus, Brunoniella, Dischistocalyx, Echinacanthus, Epiclastopelma, Heteradelphia, Ionacanthus, Kosmosiphon, Leptosiphonium, Lychniothyrsus, Pseudoruellia, Ruelliopsis, Sautiera, Strobilanthopsis, Trichosanchezia, Zygoruellia) likely in part owing to a paucity of herbarium material for most of them (see table 1; all are either monospecific or oligospecific). Specimens that do exist are often old and yield poor-quality, highly degraded DNA. Thus, Fig. 6 Phylogenetic relationships among Ruelliinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). Taxonomic name changes proposed herein or in earlier works follow current names in brackets; these reflect the revised classification presented herein (see ‘‘Taxonomic Treatment’’). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 105 For analyses of the concatenated matrix, Modeltest (ver. 3.7; Posada and Crandall 1998) and the Akaike Information Criterion were used to find the most likely model of sequence evolution. ML searches were conducted in GARLI (ver. 0.951; Zwickl 2006), and 100 ML bootstrap replicates were implemented to assess branch support. A GTRþGþI model of sequence evolution was applied to all partitions, along with empirical base frequencies. Two independent Bayesian searches of 5 million generations each using four heated chains were conducted in MrBayes (ver. 3.1.2; Huelsenbeck and Ronquist 2001). Trees were sampled every 1000 generations, and a 50% majority rule consensus tree was constructed using the last 1000 trees (all post burn-in) from both searches. Phylogenetic Hypothesis Testing Fig. 7 Phylogenetic relationships among Trichantherinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). our five molecular data sets were not parallel with respect to taxon sampling. In some instances, a given genus is represented only by one (Clarkeasia, Echinacanthus, Trichosanchezia) or two (Ionacanthus, Sautiera) markers. As such—and because a major goal of this study was to reconstruct relationships across all genera in Ruellieae, thus requiring inclusion of as many genera as possible—we analyzed a data matrix containing concatenated sequences from all five markers. Pilot analyses of the five individual matrices indicated no conflict in relationships: following Mason-Gamer and Kellogg (1996), maximum likelihood (ML) topologies were considered conflicting only if alternative resolutions were supported by 70% bootstrap support (BS; data not shown but matrices available from E. A. Tripp upon request). For Eif3E and psbA-trnH, because of extreme sequence divergence between putative members of Ruellieae and distant relatives in Acanthacae (i.e., the 14 representatives from nonRuellieae lineages in Acanthaceae; fig. 1), we analyzed sequence data only from putative Ruellieae. All other Acanthaceae were scored as having missing data for these two markers. Results from phylogenetic reconstruction led us to propose and test a number of alternative hypotheses regarding monophyly of genera (10 tests), placement of a genus in an alternative clade (one test), and alternative relationships among clades (three tests). In total, 13 ‘‘positive’’ constraint tests (H1–H13) were conducted (table 2). We also conducted a ‘‘reverse’’ or ‘‘negative’’ constraint analysis (cf. Tripp 2010) to ask whether a likelihood search that forced nonmonophyly of Ruellieae resulted in a significantly less likely topology (H14; table 2). Alternative topologies—that is, constraint trees—were constructed in MacClade (ver. 4.06), and GARLI (ver. 0.951) was used to find the most likely tree consistent with each constraint. No other relationships were resolved in the constraint trees (i.e., other relationships were constructed as polytomies) except that which pertained to the hypothesis under consideration. We used a one-tailed Shimodaira-Hasegawa test (Shimodaira and Hasegawa 1999; RELL optimization, 100 replicates) and compared our most likely unconstrained tree (null hypothesis) to the most likely tree from the constrained search (alternative hypothesis). Clade Size Asymmetries Our third goal was to use phylogenetic results to test two competing hypotheses to explain the excess of numbers of gen- Fig. 8 Phylogenetic relationships among Strobilanthinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names in phylogeny follow Scotland and Vollesen (2000). Taxonomic name changes proposed herein follow current names in brackets; these reflect the revised classification presented in earlier works (see ‘‘Taxonomic Treatment’’). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 106 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 9 Phylogenetic relationships among Hygrophilinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). era in the OW with respect to those in the NW (see ‘‘Challenges to and Motivations for Study’’): that the asymmetry reflects differences in taxonomic working methods (i.e., ‘‘splitting’’ in the OW, ‘‘lumping’’ in the NW) or that the asymmetry reflects real differences in evolutionary history of Acanthaceae in these two hemispheres. If the former is correct, we expect to see numerous OW genera nested within other genera. If the latter is correct, we expect to see multiple OW genera occupying isolated branches across the phylogeny of Ruellieae. Biogeography, Morphology, and Chromosome Numbers We synthesized data from the literature and from specimen occurrences in herbaria and herbaria databases to contribute a general biogeographic understanding of genera compris- ing Ruellieae. Using fresh, field-preserved, or herbarium material, we studied numerous morphological characters that have been previously used or are here explored for the first time to understand features that characterize genera in Ruellieae. Where possible, we searched for morphological features that can be used to diagnose clades and genera. Character evolution was not explicitly optimized on phylogenies. Using herbarium and/or fresh or field-preserved floral material, we studied pollen micromorphology using SEM. Following air drying for field-preserved materials, pollen grains were sputtercoated with gold and subsequently examined under SEM. Chromosome numbers for taxa included in our study were compiled from the literature—that is, no new counts were made for this study. All numbers reported here are meiotic, regardless of how they were originally published. Results Except as noted above for Eif3E and psbA-trnH, a minimalcharacter-exclusion approach was utilized after it was found that the various exclusion methods did not alter topological relationships. Prior to character exclusion, the combined matrix contained 5332 characters. After exclusion, the matrix contained 4644 characters. Of these, 1605 were parsimony informative. Descriptive information for data partitions as well as the combined matrix is provided in table 3. Of the 43 genera of putative Ruellieae, all but one were resolved in Ruellieae: the Malagasy endemic genus Zygoruellia was placed in Whitefieldieae. Morphology also supports a relationship of Zygoruellia to Whitfieldieae: pollen grains of the former (documented here for the first time) are lenticular and biporate with granular areas surrounding the two aper- Fig. 10 Phylogenetic relationships among Mimulopsinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (MLBS; below branch) values. Asterisk ¼ 100%. Generic names in phylogeny follow Scotland and Vollesen (2000). Taxonomic name changes proposed herein follow current names in brackets; these reflect the revised classification presented herein (see ‘‘Taxonomic Treatment’’). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) Fig. 11 Phylogenetic relationships within Petalidiinae derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). Taxonomic name changes proposed herein follow current names in brackets; these reflect the revised classification presented herein (see ‘‘Taxonomic Treatment’’). tures, like some species of Whitfieldia and Chlamydacanthus of Whitfieldieae (Manktelow et al. 2001; fig. 13). Within Ruellieae, seven major clades were resolved in both Bayesian and likelihood analyses and are given formal subtribal names below (figs. 3, 4). Although stem branch lengths for these seven clades were extremely short (fig. 4), all seven were supported by ML bootstrap (70% BS) and/or Bayesian posterior probability (95% PP). Bayesian and likelihood analyses returned nearly identical topologies, with differences minor and unsupported, except that one genus was unresolved along the backbone in the Bayesian tree but resolved within one of seven clades in the likelihood tree; however, this latter result was not supported. 107 Of the seven subtribes, Erantheminae was earliest diverging in our analyses. Derived with respect to Erantheminae was a clade comprising Ruelliinae sister to the remaining five tribes: Trichantherinae, Hygrophilinae, Mimulopsinae, Petalidiinae, and Strobilanthinae (i.e., the ‘‘TMPSH clade’’). These last five tribes (i.e., Trichantherinae through Strobilanthinae) formed three clades with relationships among them unresolved: Trichantherinae, Strobilanthinae, and (Hygrophilinae þ (Mimulopsinae þ Petalidiinae)). Relationships within each of the seven subtribes of Ruellieae (depicted in figs. 5–11) are discussed in detail in ‘‘Taxonomic Treatment.’’ Outside Ruellieae, our phylogenetic results largely corroborate those of previous studies (fig. 1) pertaining to relationships among other Acanthaceae (fig. 12). Acanthaceae s.l. is composed of Nelsonioideae, Avicennia, Thunbergioideae, and Acanthaceae s.s. Within Acanthaceae s.s., Acantheae is sister to a clade of plants bearing cystoliths, as in figure 1. Within the cystolith clade, Barlerieae, Andrographidieae, Neuracanthus, and Whitfieldieae together form a clade sister to Ruellieae þ Justicieae. Of the 10 Shimodaira-Hasegawa (SH) tests that constrained monophyly of genera that were not monophyletic in our analyses, eight were rejected, lending statistical significance to their nonmonophyly (P < 0:05, H1–H6, H9, H10; table 2). In contrast, forcing monophyly of Mimulopsis (P ¼ 0:40, H7) and Dyschoriste (P ¼ 0:05, H8) resulted in trees that were not significantly less likely than the most likely unconstrained tree (table 2). A likelihood search that constrained Zygoruellia within Ruellieae resulted in a significantly less likely topology (P < 0:05, H11; table 2). Searches that constrained Dischistocalyx þ Acanthopale (P < 0:05, H12; table 2) and Satanocrater þ Ruellia (P < 0:05, H13; table 2) resulted in significantly less likely trees. Finally, the negative constraint analysis that forced nonmonophyly of Ruellieae (exclusive of Zygoruellia) also resulted in a significantly less likely topology (P < 0:05, H13; table 2). Within Ruellieae, several monospecific and oligospecific genera were found to be nested with strong support in larger genera. Fig. 12 Phylogenetic relationships among other Acanthaceae (including Zygoruellia) derived from Bayesian analysis. Thickened branches are supported by 95% Bayesian posterior probability (BPP; above branch) and/or 70% maximum likelihood bootstrap (ML BS; below branch) values. Asterisk ¼ 100%. Generic names follow Scotland and Vollesen (2000). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 108 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 13 Pollen of Whitfieldeae, showing lenticular, biporate grains with granular areas surrounding the apertures. A, Zygoruellia. B, Chlamydacanthus. C, Camarotea. Images in B and C are reproduced from McDade et al. (2008). Discussion Morphological Synapomorphies for Ruellieae Monophyly of Ruellieae In this study, we used molecular data from five markers to test the placement of 43 of the 47 genera treated in Ruellieae by Scotland and Vollesen (2000; excluding Physacanthus). All but one of these is resolved with strong support in Ruellieae. In contrast, both molecular and morphological data indicate that Zygoruellia is allied to Whitfieldieae, and alternative placement of Zygoruellia in Ruellieae was rejected by molecular data (H12; table 2). Thus, Zygoruellia is excluded from Ruellieae as recircumscribed below. Monophyly of Ruellieae was strongly supported in all analyses, and an alternative test of nonmonophyly of Ruellieae resulted in a significantly less likely tree (H13; table 2). Within the broader context of all Acanthaceae, Ruellieae is strongly supported as part of the cystolith-bearing clade of retinaculate Acanthaceae (figs. 1, 3, 4). We were unable to obtain high-quality molecular data for Calacanthus, Echinacanthus, and Stenothyrsus and lacked access to any material of Spirostigma, thus preventing their inclusion in phylogenetic analyses. We have seen and studied the minimal material that we had access to for all four genera and feel confident that they should be treated in Ruellieae as discussed further below. Several traits have been hypothesized to represent synapomorphies for Ruellieae. Results from our study indicate that a filament curtain is lacking in several but not all of the genera that comprise the earliest-diverging lineage in the tribe, Erantheminae. Because homologous structures have not yet been documented in other Acanthaceae (with the possible exception of Glossochilus; Manktelow 2000), we hypothesize that the filament curtain (fig. 2D, 2E) evolved early in the history of Ruellieae but was lost once in one clade of Erantheminae. As such, we here tentatively consider the filament curtain to represent a synapomorphy for the tribe. This study also confirms that all members of Ruellieae share left-contort corolla aestivation, hygroscopic trichomes that cover seed surfaces (fig. 2A, 2B; these secondarily reduced to margins of seeds in some lineages [fig. 2C] or, very rarely, lost altogether), and bifid, filiform stigma lobes in which the dorsal lobe is equal to, shorter than, or completely reduced developmentally with respect to the ventral lobe. So far as is known among Acanthaceae s.s. and more specifically among cystolithbearing Acanthaceae, only Lankesteria and Whitfieldia of Whitfieldieae (fig. 1) share left-contort corolla aestivation with Ruellieae. However, Lankesteria and Whitfieldia possess three features that are lacking in all Ruellieae: concentric ridges on seed surfaces, granular areas that surround pollen This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) pores (these two traits are unique to Whitfieldieae), and capitate stigmas (Manktelow et al. 2001). Absent any other data to support a relationship of Lankesteria and Whitfieldia to Ruellieae, we here consider the trait combination of the filament curtain, left-contort aestivation, hygroscopic seed trichomes, and unequal to subequal filiform stigma lobes synapomorphic for Ruellieae, with loss of the filament curtain and seed trichomes in a few groups. We note that each of these characters is also found in other lineages of Acanthaceae (e.g., unequal, filiform stigma lobes in some Justicieae and Barlerieae; seed vestiture in some Barlerieae), but the combination of these traits appears to readily discriminate members of Ruellieae from members of the remaining lineages of Acanthaceae. Chromosome Number Chromosome numbers have been reported for 11 genera of Ruellieae (as here circumscribed); these span the full range of numbers known in the family, from n¼6 (Hygrophila; Singh et al. 1993) to n¼68 (Sanchezia; Singh 1951; Kaur 1970). Although 23 haploid chromosome numbers have been reported (n¼6, 8–22, 24, 28, 30, 32, 40, 42, and 68) among Ruellieae, n¼15 and n¼16 are both known for six genera, and n¼16 has been reported for the largest number of species. While basic chromosome numbers of x¼14 (Hemigraphis), x¼15 (Dyschoriste), x¼16 (Brillantaisia, Hygrophila, Phaulopsis), and x¼17 (Ruellia) have been proposed (Daniel 2000) or are likely for some genera of Ruellieae on the basis of known chromosome counts, a basic number of x¼8 might be proposed for the tribe. Of interest in this respect are the diverse counts for Strobilanthes s.l.: n¼8–11, 13–16, 20, 21, and 30 (Daniel and Chuang 1989; Iwatsubo et al. 1993). Biogeography Phylogenetic analyses presented here afford the opportunity to begin reconstructing geographical distribution histories for the various lineages that comprise Ruellieae. For example, all genera in Erantheminae, the basalmost subtribe, occur in Asia, Papuasia, and Malesia except Kosmosiphon, which is tropical West African. Within the TMPSH clade, all Trichantherinae are NW, all Strobilanthinae are OW, Mimulopsinae are entirely African, and both Hygrophilinae and Petalidiinae are OW and NW. Within Ruelliinae, basal groups are OW, from which there was at least one origin of NW Ruellia. Thus, in sum, if dispersal accounts for present-day distributions in Ruellieae (E. A. Tripp and L. A. McDade, unpublished data), at least four dispersal events from the OW to the NW are required. This is based on only a fraction of taxon sampling in some pantropical genera and thus can be construed as a conservative estimate. OW versus NW Generic Diversity Within Ruellieae, several monospecific and oligospecific genera (e.g., Ionacanthus, Epiclastopelma, Sautiera, Apassalus, Clarkeasia, Benoicanthus, Pseudoruellia, Lychniothyrsus, Polylychnis) were nested with strong support in larger genera. 109 These data support the hypothesis that, in Ruellieae, the inflation of OW generic diversity with respect to NW generic diversity is at least partly attributable to differences in taxonomic working methods of botanists (vs. differences in evolutionary histories). This result, together with insight from plant morphology as elaborated below, leads us to reduce several generic names to synonymy. How Many Species of Ruellieae? As summarized in table 1, we compared the total number of validly published names in Ruellieae (as tallied using the International Plant Names Index and Tropicos, including synonyms) to our estimate of the actual number of taxa in the tribe, which was based on our review of monographs, taxonomic treatments, and nonmonographic literature or was herein estimated. This exercise indicated that there are more than 2600 names available for what is more accurately probably 1200 or so taxa. Among genera in Ruellieae that contain more than three species, only six have been monographed or revised in their entirety within the last several decades (e.g., since 1970: Blechum [Tripp et al. 2009], Brillantaisia [Sidwell 1998], Duosperma [Vollesen 2006], Louteridium [Richardson 1972], Phaulopsis [Manktelow 1996], and Suessenguthia [Schmit-Lebuhn 2003]). Within these six genera, the ratio of the number of currently recognized taxa to the number of names that exist for them ranges from 19% to 90% (Blechum [6 recognized taxa/32 names], Brillantaisia [12/ 49], Duosperma [26/29], Louteridium [9/11], Phaulopsis [26/40], Suessenguthia [6/10]), suggesting no clear pattern from which one might otherwise extrapolate to estimate standing species diversity. As such, the estimate of ;1200 species represents our best current estimate of species richness in Ruellieae. Revised Classification of Ruellieae On the basis of data from this study as well as earlier perspectives, we propose a revised classification of Ruellieae (Ruelliinae sensu Scotland and Vollesen 2000) that includes 32 genera organized in seven subtribes plus five additional incertae sedis genera, for a total of 37 genera. This classification includes 34 genera previously recognized in Scotland and Vollesen (2000) plus three genera that have been described since that classification or are being described. It excludes Zygoruellia for reasons given above. Specific features uniting the clades comprising each subtribe as well as features that delimit genera within each subtribe are discussed below. Many aspects of our phylogenetic results and classification reflect generic assemblages recovered in Scotland’s (1991) cladistic study of genera of Ruellieae based on pollen morphology. We similarly emphasize pollen morphology (fig. 14) in the discussion of relationships among groups within Ruellieae (pollen variation in Ruellieae will be further explored in a forthcoming paper; E. A. Tripp et al., manuscript in preparation). Our morphological study of genera in Ruellieae revealed several characters, both synapomorphic and sympleisiomorphic, that support generic delimitation and phylogenetic patterns. However, in some cases (e.g., Brunoniella, Leptosiphonium) we do not yet know enough about plants to propose mor- This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Fig. 14 Pollen diversity in Ruellieae. A, Kosmosiphon azureus (Hepper 4127 [K]). B, Eranthemum nervosum (Singh 208 [NY]). C, Pararuellia delavayana (Bouff 42833 [CAS]). D, Brunoniella australis (Daniel 10065 [CAS]). E, Dischistocalyx grandifolius (Cheek et al. 5406 [MO]). F, Acanthopale pubescens (Jongkind & Rapanarwo 917 [MO]). G, Satanocrater paradoxus (Friis et al. 3147 [C]). H, Benoicanthus tachiadenus (Daniel 11024 [CAS]). I, Blechum grandiflorum (Breedlove 50544 [CAS]). J, Eusiphon geayi (Daniel 11048 [CAS]). K, Lychniothyrsus mollis (Santo 3084 [US]). L, Polylychnis fulgens (Mori 8698 [NY]). M, Pseudoruellia perrieri (Lebat et al. 2083 [K]). N, Ruellia bignoniiflora (Greenway This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) phological features that diagnose genera or support patterns of relatedness. Characters and character states are discussed in ‘‘Taxonomic Treatment.’’ Because this study represents the first attempt to comprehensively reconstruct relationships within this diverse tribe on the basis of molecular data, we recognize that some aspects of the classification are subject to future revision, especially those pertaining to large genera that have not yet been studied on a worldwide basis (e.g., Dyschoriste, Hygrophila; work in progress on Strobilanthes by R. Scotland, J. R. I. Wood, Y. Deng, and colleagues; work in progress on Ruellia by E. A. Tripp). Several smaller or geographically restricted genera (e.g., Dischistocalyx, Mellera, Mimulopsis, Petalidium) have also never been thoroughly investigated, and an improved understanding of diversity and variation in these groups will be achieved only via future revisionary studies. In sum, results from this study have allowed us to address the four objectives put forth in the introduction. The basic phylogenetic framework herein reconstructed for Ruellieae sets the stage for future investigation of character evolution, lineage diversification, and additional comparative analyses within the tribe and across Acanthaceae as a whole. Taxonomic Treatment Ruellieae Dumort. 1829. An. Fam. Pl.:23. Erantheminae Ruelliinae Trichantherinae Hygrophilinae Mimulopsinae Petalidiinae Strobilanthinae Erantheminae Nees Brunoniella Bremek. Eranthemum L. Kosmosiphon Lindau Leptosiphonium F. Muell. Pararuellia Bremek. & Nann.-Bremek. Pseudosiphonium ined. Ruelliinae Nees Acanthopale C.B. Clarke Dischistocalyx T. Anderson ex Benth. Ruellia L. (syn. nov.: Benoicanthus, Lychniothyrsus, Polylychnis, Pseudoruellia, Spirostigma) Satanocrater Schweinf. Trichantherinae Benth. & Hook.f. Bravaisia DC. Louteridium S. Watson 111 Sanchezia Ruiz & Pav. Suessenguthia Merxm. Trichanthera Kunth Trichosanchezia Mildbr. Hygrophilinae Nees Hygrophila R. Br. Brillantaisia P. Beauv. Mimulopsinae E. Tripp, subtr. nov., TYPE: Mimulopsis Schweinf., Verh. Zool.-Bot. Ges. Wien. 18:677. 1868. Plants perennial shrubs to small trees, some species pleitesial, leaf margins crenate to dentate, blades with few secondary veins, corollas generally with ‘‘herringbone’’ patterning in throats and retrorse trichomes on lower lip internally, pollen triporate to tricolporate, 12-pseudocolpate (secondarily decreased or increased in some taxa), germinal apertures surrounded by ‘‘sexine lips,’’ ovaries bearing eight or more ovules (not all of which necessarily mature into seeds), seeds with hygroscopic trichomes restricted to their margins (trichomes lacking completely in one taxon). Eremomastax Lindau Heteradelphia Lindau Mellera S. Moore (syn. nov.: Ionacanthus) Mimulopsis Schweinf. (syn nov.: Epiclastopelma) Petalidiinae Benth. & Hook.f. Duosperma Dayton Dyschoriste Nees (syn. nov.: Apassalus, Sautiera) Phaulopsis Willd. Petalidium Nees Ruelliopsis C.B. Clarke Strobilanthopsis S. Moore Strobilanthinae T. Anderson Clarkeasia J.R.I. Wood Hemigraphis Nees Stenosiphonium Nees Strobilanthes Blume Incertae sedis: Calacanthus T. Anderson ex Benth. Diceratotheca J.R.I. Wood & R.W. Scotland Echinacanthus Nees Sinoacanthus ined. Stenothyrsus C.B. Clarke Excluded taxa: Zygoruellia Baill. (belongs in Whitfieldieae Bremek.) Subtribe I: Erantheminae As here delimited, Erantheminae (fig. 5) is composed of Brunoniella, Eranthemum, Kosmosiphon, Leptosiphonium, Pararuellia, and a sixth genus to be described in a forthcom- & Kanuri 14012 [MO]). O, Louteridium purpusii (Breedlove 31613 [CAS]). P, Bravaisia berlandieriana (Pitzer & Mizquez 3437 [MO]). Q, Trichanthera gigantea (Madrigal et al. 874 [US]). R, Trichosanchezia chrysothrix (Diaz et al. 6954 [MO]). S, Sanchezia decora (Foster 8790 [MO]). T, Suessenguthia multisetosa (Wood 2157 [NY]). U, Brillantaisia stenolepis (Mhoro 35 [MO]). V, Hygrophila stricta (Maxwell 90-72 [MO]). W, Phaulopsis betonica (Love & Congclon 3157 [MO]). X, Ruelliopsis setosa (Smith 3107 [MO]). Y, Petalidium coccineum (Smooth 7819 [MO]). Z, Duosperma fimbriatum (Richards 19019 [NY]). AA, Strobilanthopsis linifolia (Harder et al. 3152 [MO]). BB, Dyschoriste hildebrandtii (Daniel et al. 9376 [US]). CC, Apassalus diffusus (Garcia & Pimentel 2544 [US]). DD, Mimulopsis arborescens (Tweedie 2156 [CAS]). EE, Mellera alata (Salubeni & Nachamba 4260 [MO]). FF, Epiclastopelma glandulosum (Mabberley 1144 [K]). GG, Ionacanthus calcarata (Humbert & Capuron 25789 [K]). HH, Eremomastax polysperma (Beguin & Beguin 1044 [MO]). II, Heteradelphia paulowilhelmia (Daniel 11137 [MO]). JJ, Baphicacanthus cusia (Shen 1958 [US]). KK, Clarkeasia parviflora (Parry 421 [K]). LL, Hemigraphis latebrosa (Singh 91017 [MO]). MM, Stenosiphonium cordifolium (Cooray 70020127R [CAS]). NN, Strobilanthes polyneuros (Wu et al. WP433 [MO]). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Fig. 15 Morphological diversity in Erantheminae. A, Kosmosiphon azureus (de Wilde s.n. [BR], Cameroon), showing paired leaflike bracts and long, thin corolla tube. B, Eranthemum austrosinense (Tripp & Deng 1535 [RSA], China), showing terminal spike with leaflike bracts and long, thin corolla tube. C, Pararuellia alata (collector/collection unknown [MO], China), showing lack of filament curtain (note lack of filament fusion) and X-shaped anthers (see text). D, Pararuellia glomerulata (Tripp & Deng 1594 [RSA], China), showing inflorescence macromorphology. E, Pararuellia delavayana (Tripp & Deng 1583 [RSA], China), showing infructescence macromorphology. F, Pararuellia alata (collector/collection unknown [MO], China), showing X-shaped anther. G, Pararuellia delavayana (Tripp & Deng 1583 [RSA], China), showing basal rosette habit. H, Brunoniella australis This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) ing paper (Pseudosiphonium ined.; G. Chunming et al., manuscript in preparation). Wood et al. (2012) described the new genus Diceratotheca and presented phylogenetic data that suggest its close relationship to Pararuellia. However, as we have not studied material of these plants, we opt here to include it as incertae sedis (see below) pending further studies that might confirm its placement within Ruellieae. Plants pertaining to Pseudosiphonium are from China and were previously known as either Ruellia venusta Hance or Leptosiphonium venustum (Hance) E. Hossain. However, Ruellia and Leptosiphonium cannot accommodate plants of Pseudosiphonium, as discussed below. Furthermore, our data reject an alternative hypothesis of monophyly of Papua New Guinean plants and Chinese plants that have been treated together in Leptosiphonium (H1, table 2); below, the name Leptosiphonium is applied only to plants from Papuasia. Erantheminae is the earliest diverging lineage within Ruellieae, to which the subtribe clearly belongs—all genera in it have left-contort aestivation, hygroscopic seed trichomes, and filiform stigmas. However, at present there are no known morphological synapomorphies to unite the six genera in Erantheminae. Plants of several but not all genera in this group are basal rosette forming herbs, and all have purple or white (to pale yellow) corollas (fig. 15). All genera in Erantheminae that we have examined have hygroscopic trichomes that cover the entire seed surface, suggesting that this state (vs. trichomes restricted to seed margins) may be ancestral within Ruellieae. In Erantheminae, three strongly supported clades are resolved (fig. 5): (1) Kosmosiphon þ Eranthemum; (2) Brunoniella þ Leptosiphonium; and (3) Pararuellia þ Pseudosiphonium. ML (fig. 4) but not Bayesian (figs. 3, 5) analyses resolved groups 2 and 3 as sister, but branch support was lacking (compare clade structure in fig. 4 vs. that in fig. 5). Interestingly, Eranthemum and Kosmosiphon have well-developed filament curtains (confirmed in the latter by I. Darbyshire), like the remainder of subtribes nested within Ruellieae, whereas other Erantheminae lack the structure. This complex trait is most likely synapomorphic for Ruellieae with, if ML results are correct (fig. 4 but genera not labeled), one loss in Erantheminae, along the branch that leads to Pararuellia, Pseudosiphonium, Leptosiphonium, and Brunoniella. However, the filament curtain may have instead evolved more than once in Ruellieae; character optimization studies are needed to discriminate between hypotheses. Owing to a lack of complete information about some genera in Erantheminae, our inferences regarding relationships within this subtribe should be considered tentative. Scant material in major herbaria has resulted in much of Erantheminae being relatively understudied, and the subtribe should be prioritized for future taxonomic work. Kosmosiphon þ Eranthemum. To our knowledge, a close relationship between Kosmosiphon and Eranthemum has not been postulated before, perhaps understandably given that the former is tropical West African and the latter occurs in Asia. In this study, the two genera were strongly supported as sister lineages (fig. 5). Kosmosiphon and Eranthemum 113 are united by their staminal configuration of 2 fertile stamens þ 2 staminodes as well as triaperturate ‘‘Wabenpollen’’ (¼‘‘honeycomb pollen’’; fig. 14A, 14B; Lindau 1893, 1895) with coarsely reticulate exine sculpturing. Plants of both genera also produce clavate fruits with four ovules per ovary and have similar long, tubular corollas suggestive of lepidopteran pollination (fig. 15A, 15B). Two genera in Ruelliinae (Satanocrater and Ruellia) also have triaperturate Wabenpollen, indicating that this type of exine sculpturing has evolved more than once in Ruellieae or is sympleisiomorphic with many subsequent shifts. Still, the combination of Wabenpollen, a 2 þ 2 staminal configuration, and four seeds per clavate capsule can be considered diagnostic for Kosmosiphon þ Eranthemum. Kosmosiphon (figs. 14A, 15A) is a poorly known, littlecollected monospecific genus endemic to tropical western Africa (Cameroon, Central African Republic, and Democratic Republic of the Congo). Kosmosiphon is readily identified by its two-flowered cymes in which each flower is subtended by a pair of highly distinctive petiolate, leaflike bracts (fig. 15A). No other Ruellieae have these paired bracts except the southwest African, arid-adapted Petalidium, to which Kosmosiphon bears no other resemblance. Other lineages of Acanthaceae (i.e., outside Ruellieae) have evolved similar bract morphologies (e.g., Chlamydacanthus of Whitfieldieae, some Dicliptera and Hypoestes of Justicieae). Indeed, Lindau (1895), in his protologue of Kosmosiphon, thought that the genus was related to Lankesteria (also of Whitfieldieae). Kosmosiphon is easily differentiated from its sister genus, Eranthemum, by these paired bracts and porate (vs. colporate) pollen apertures. We here consider the combination of Wabenpollen with porate apertures, the 2 þ 2 staminal configuration, and the paired petiolate, leaflike bracts to be diagnostic for Kosmosiphon. Two accessions of the genus were strongly supported as monophyletic in our study (fig. 5). Chromosome numbers have not yet been reported for the genus. The genus Eranthemum (figs. 14B, 15B), although more species rich and wide ranging than Kosmosiphon, is equally poorly understood. Species of Eranthemum occur from India to southeastern Asia. Like Kosmosiphon, plants of Eranthemum produce two-flowered cymes (sometimes reduced to solitary flowers), which are generally organized into terminal spikes. Unlike Kosmosiphon, however, inflorescences of Eranthemum are mostly terminal (fig. 15B), whereas those of Kosmosiphon are strictly axillary. Additionally, all specimens we have studied to date of Eranthemum have one rather than two leaflike (or subleaflike) bracts subtending each inflorescence or reduced inflorescence (i.e., a solitary flower; see Tripp 2010 for inflorescence terminology). Finally, Eranthemum has colporate pollen apertures (vs. porate in Kosmosiphon) that are concave in polar view, giving grains a trilobed appearance (not shown; E. A. Tripp and Ensermu K., unpublished data). However, a thorough revision of Eranthemum is needed to fully evaluate characters in the genus. We here consider the combination of Wabenpollen with colporate, concave apertures, the (Hosking 2952 [CANB], Australia), showing superficial resemblance to numerous purple-flowered species of Ruellia. I, Leptosiphonium cf. stricklandii (Daniel 6571 [CAS], Papua New Guinea), showing long, thin corolla tube. Photos by E. A. Tripp except H, by J. Hosking (used with permission), and I, by T. F. Daniel. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Fig. 16 Morphological diversity in Ruelliinae. A, Dischistocalyx thunbergiiflorus (Caleb et al. 245 [K], Cameroon), showing general macromorphology. B, Acanthopale pubescens (not vouchered, Zimbabwe), showing general macromorphology. C, Satanocrater ruspolii (Tripp & Ensermu 904 [RSA], Ethiopia), showing large and prominent glands on adaxial leaf surface. D, Satanocrater paradoxus (Tripp & Ensermu 906 [RSA], Ethiopia), showing bird-adapted flowers and calyces with lobes fused nearly to their apices. E, Ruellia humboldtiana (not vouchered, Duke This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 2 þ 2 staminal configuration, and the single bract per inflorescence to be diagnostic for Eranthemum. Three accessions of the genus were strongly supported as monophyletic in our study (fig. 5). Chromosome numbers of n¼11, 12, 15, 17, 19, 21, 22, and 42 have been reported for Eranthemum (Pathak et al. 1949; Narayanan 1951; Grant 1955; Nanda 1962; Kaur 1970; Love 1982; Bala and Gupta 2011). Pararuellia þ Pseudosiphonium and Brunoniellia þ Leptosiphonium. Within Erantheminae, ML analyses resolved Pararuellia þ Pseudosiphonium as sister to Brunoniella þ Leptosiphonium (these two clades were unresolved in Bayesian analyses; fig. 5). These four genera are apparently differentiated from all other Ruellieae by their lack of a floral filament curtain (fig. 15C)—that is, they lack a configuration in which tissue outgrowths from fused, adjacent filament pairs form a barrier, or ‘‘curtain,’’ that persists to the corolla base (see Mantelow 2000). However, we have not had access to sufficient material of Leptosiphonium to conclude with confidence that it lacks a filament curtain. Importantly, the generic protologue notes that the filaments are ‘‘connate in pairs towards their base’’ (Mueller 1886, p. 32), which may suggest the presence of a curtain. Further study of all genera in this group will permit full evaluation of filament curtain distribution in Erantheminae. In addition to putatively lacking filament curtains, Brunoniella, Leptosiphonium, Pararuellia, and Pseudosiphonium all produce cylindrical capsules with 8–16 ovules per ovary (with secondary increases to more than 20 in Leptosiphonium), instead of the four per ovule that characterizes Kosmosiphon þ Eranthemum. Pararuellia þ Pseudosiphonium. The sister-group relationship between the Chinese plants of Pararuellia (fig. 15C–15G) and Pseudosiphonium is strongly supported by DNA sequence data (fig. 5) and by a unique anther morphology not seen anywhere else in Ruellieae. Instead of the ‘‘normal’’ configuration of two thecae held side by side with the filament connective tissue clearly positioned dorsal to thecae, the thecae of Pararuellia and Pseudosiphonium extend outwardly in a 180° configuration from the filament connective tissue, which is expanded in these two genera relative to the rest of Ruellieae. We refer to this synapomorphic feature as ‘‘X-shaped’’ anthers (fig. 15F). Pollen of Pararuellia is triporate with equatorial apertures and is finely reticulate with microverrucae on exine ridges, as documented by Bremekamp (1964) and here confirmed by us (fig. 14C). We have not yet been able to examine the pollen of Pseudosiphonium, but its highly derived X-shaped anthers leave little doubt as to its close relationship to Pararuellia rather than Ruellia or Leptosiphonium, to which plants have been earlier attributed (see above). Plants of Pararuellia produce only basal leaves, forming a rosette habit (fig. 15G). In contrast, plants of Pseudosiphonium produce cauline leaves. Their rosette or cauline growth forms in combination with Xshaped anthers can be considered diagnostic for each genus. Chromosome numbers have not yet been reported for these genera, to our knowledge. 115 Brunoniella þ Leptosiphonium. At present, there are no known synapomorphies to unite the Australian endemic Brunoniella with the Papuasian Leptosiphonium other than molecular data (fig. 5). Pollen traits (e.g., both appear to be pantoforate) may provide such characters, but to date we have had insufficient floral material of each genus to assess pollen features as well as important characters of the corolla (e.g., presence or absence of filament curtains). Plants of both genera have ‘‘normal’’ (i.e., not X-shaped) anthers and occupy mesic environments. The somewhat-complex history of the genus Brunoniella has been discussed extensively in other publications (Bremekamp 1964; Barker 1986). Superficially, these plants (fig. 15H) resemble several diminutive species of Ruellia, which likely explains why some Brunoniella were originally described in Ruellia (Brown 1810, on the basis of amazingly well-preserved Banks and Solander collections made during Cook’s first voyage). As noted by Bremekamp (1964) and Moylan and Scotland (2000) and confirmed by our studies, pollen grains of Brunoniella are pantoforate and minutely tuberculate (fig. 14D), thus easily distinguishing the genus from others in Erantheminae. Chromosome numbers have not yet been reported for Brunoniella, to our knowledge. Leptosiphonium was described by F. Mueller to accommodate a taxon from Papua New Guinea, L. stricklandii, as ‘‘one of the most beautiful’’ discovered during a collecting expedition to the Strickland River Valley and one that its discoverers ‘‘hope to see ere long in ornamental culture’’ (Muller 1886, pp. 32–33). This reaction was presumably in response to its attractive flowers, which are long, tubular, and white to pale yellow in color (fig. 15I). Superficially, these plants also resemble some species of Ruellia with similarly long-tubed, pale-colored flowers, but molecular data presented here resolve that genus several clades removed from Erantheminae. At present, the genus is thought to contain ;10 species (Bremekamp and Nannenga-Bremekamp 1948) distributed in Papuasia (New Guinea and neighboring islands). However, species are poorly known, and Leptosiphonium is in need of revisionary investigation. Chromosome numbers have not yet been reported for the genus. Subtribe II: Ruelliinae As here delimited, Ruelliinae consist of four genera: Dischistocalyx, Acanthopale, Ruellia, and Satanocrater (fig. 6). Several additional genera are reduced to synonyms of Ruellia (see below). The four genera comprising Ruelliinae share two features: plants produce four stamens that have anthers with blunt bases, and they produce pollen grains that are spherical and lack colpi or nearly so (i.e., pores of some Satanocrater grains appear to be associated with a weakly developed colpus; Tripp and Fatimah 2012), although there have been at least two reversions to colporate pollen in Ruellia (Furness and Grant 1996; Tripp et al. 2009). Within Ruelliinae, Dischistocalyx is sister to a clade comprising the other three genera; within this latter clade, Sata- University Greenhouses), showing one of many different pollination syndromes in the genus (here, hummingbird pollination) and extrafloral nectaries with exudate. F, Pseudoruellia perrieri (not vouchered, Madagascar), showing distinctive succulent leaves. Photos A, D, and E by E. A. Tripp; B by B. Wursten (used with permission); F by G. Schatz (used with permission). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions Fig. 17 Morphological diversity in Trichantherinae. A, Louteridium donnell-smithii (Daniel 11337 [CAS], Guatemala), showing gibbous corolla and long-exserted reproductive parts. B, Bravaisia integerrima (Tripp & Lujan 519 [RSA], Venezuela), showing floriferous branch. C, Trichanthera corymbosa (Tripp & Lujan 520 [RSA], Venezuela), showing infundibular, red, densely pubescent corollas. D, Bravaisia integerrima (Tripp & Lujan 519 [RSA], Venezuela), showing tree habit and use as street tree. E, Trichanthera corymbosa (Tripp & Lujan 520 [RSA], Venezuela), showing tree habit and use as living fence. F, Bravaisia integerrima (Tripp & McDade 133 [DUKE], Costa Rica), showing prop roots and river edge habitat, with L. McDade for scale. G, Sanchezia sp. (not vouchered, Ecuador), showing tubular corolla. H, Suessenguthia multisetosa (Schmidt-Lebuh 966 [LPB], Bolivia), showing general macromorphology. Photo A by T. F. Daniel, photos B–G by E. A. Tripp, and photo H by A. Schmidt-Lebuhn (used with permission). This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 117 Fig. 18 Morphological diversity in Hygrophilinae. A, Brillantaisia grotanellii (Tripp & Ensermu 931 [RSA], Ethiopia), showing strongly bilabiate corolla, two staminodes, and membranous, yellow ‘‘hinge’’ toward corolla opening (below staminodes). B, Brillantaisia owariensis (not vouchered, Duke University Greenhouses), showing two staminodes and a laterally compressed upper lip, which forms a prominent hood. C, Hygrophila schulli (Tripp & Ensermu 927 [RSA], Ethiopia), showing four fertile stamens, an upper lip that forms only a weak hood, and prominent orange spines. Photos by E. A. Tripp. nocrater is sister to Acanthopale þ Ruellia, with support for all these relationships. In the absence of molecular data, closer relationships between Dischistocalyx and Acanthopale (fig. 14E, 14F) and between Satanocrater and Ruellia (fig. 14G, 14N) might have been hypothesized on the basis of pollen morphologies and other features (see below), but both of these hypotheses were rejected by our data (H12 and H13; table 2). Dischistocalyx (figs. 14E, 16A) is a little-studied African genus of ;15 species that are distributed from Nigeria to the Democratic Republic of the Congo (Bremekamp 1944b; Heine 1966; Champluvier and Senterre 2010). In our study, the three sampled species formed a strongly supported monophyletic group (fig. 6). As currently understood, species of Dischistocalyx share zygomorphic calyces, as implied by the genus name (i.e., ‘‘di’’ ¼ two, ‘‘schism’’ ¼ division), and echinate, triporate pollen grains with apertures arranged equatorially and each surrounded by a ring of echinae (this ring is visible in fig. 14E). All capsules of Dischistocalyx studied to date have more than 20 seeds (but see Bremekamp [1944b] and Heine [1966], who indicated that plants produce 10–20 ovules per ovary), and seeds have trichomes restricted to the margins. The genus, however, is severely in need of revisionary study because species are poorly characterized. Many herbarium labels indicate that plants of Dischistocalyx have fleshy leaves or are semisucculent and grow epiphytically near tree bases, both features that are uncommon among Acanthaceae. We consider the combination of zygomorphic calyces, triporate, and echinate pollen grains with a ring of echinae sur- rounding each aperture and seeds with marginal trichomes to be diagnostic for Dischistocalyx. Chromosome numbers have not yet been reported for the genus, to our knowledge. The genus Acanthopale (figs. 14F, 16B) has never been treated taxonomically in its entirety. The seven or so species are distributed primarily in Madagascar and tropical eastern Africa (i.e., from Ethiopia south to Tanzanzia; one species is disjunct in Cameroon) and are united by their pantoporate, echinate, spherical pollen grains (fig. 14F) and flattened, apically rounded uni- or bicellular trichomes that line inner corolla surfaces (see image in Tripp 2007). The four species sampled here form a strongly supported monophyletic group (fig. 6). Species of Acanthopale also have calyces that are actinomorphic or nearly so, with lobes that are frequently either obclavate in shape and/or slightly reflexed at the tip, ovaries with 2 ovules, and seeds with hygroscopic trichomes restricted to the margins. Species of Acanthopale are known to flower gregariously, with subsequent population dieback (Vollesen 2008). We consider the combination of pantoporate and echinate pollen, ovaries with two ovules, seeds with marginal trichomes, and four fertile stamens lacking anther appendages to be diagnostic for Acanthopale. Chromosome numbers have not yet been reported for the genus, to our knowledge. Although not resolved as sister taxa in this study, plants of Dischistocalyx and Acanthopale have long been thought to be closely related (cf. Bremekamp 1944b; Heine 1966). Both have porate pollen grains with echinate exine sculpturing (fig. 14E, 14F), a type that is otherwise absent in Ruellieae This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 118 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 19 Morphological diversity in Petalidiinae. A, Phaulopsis imbricata (Tripp & Ensermu 929 [RSA], Ethiopia), showing a complex, congested inflorescence. B, Ruelliopsis setosa (Tripp & Dexter 799 [RSA], Namibia), showing linear leaves. C, Ruelliopsis setosa (Tripp & Dexter 817 [RSA], Namibia) growing among boulders, showing ‘‘grass mimic’’ habit. D, Duosperma longicalyx (Tripp & Ensermu 888 [RSA], Ethiopia), showing papery bracts that enclose fruits. E, Petalidium cf. physaloides (Tripp & Dexter 844 [RSA], Namibia), showing papery bracts that enclose fruits. F, Petalidium linifolium (Tripp et al. 2031 [RSA], Namibia), with paired, showy bracts. G, Petalidium luteo-album (Tripp & Dexter 830 [RSA], Namibia), showing capsule with fracturing placentae. H, Petalidium lanatum (Tripp & Dexter 879 [RSA], Namibia), showing dense inflorescence heads and spinulose infructescences. I, Petalidium luteo-album (Tripp & Dexter 830 [RSA], Namibia), showing herringbone corolla patterning and paired, showy bracts. J, Dyschoriste thunbergiiflora (Tripp & Ly 937 [RSA], RSABG Greenhouses), showing calyx with partially fused lobes with hyaline margins (hy; labeled with an arrow). K, Sautiera tinctorum (Bowma 154 [L], Sri Lanka), showing strongly bilabiate corollas (ul ¼ upper lip; ll ¼ lower lip; labeled with arrows) and calyx with hyaline regions like Dyschoriste (hy; labeled with an arrow). Photos by E. A. Tripp. (note that Louteridium, Diceratotheca, and some Strobilanthes have verrucate to baculate to psilate or gemmate exine sculpturing [see Richardson 1972; Scotland 1993; Carine and Scotland 1998; Daniel 1998], but these genera are distantly related to Ruelliinae on the basis of the molecular data presented here; similar pollen morphology along with other features, such as corolla trichomes, probably also led earlier authors to conclude that Acanthopale was closely related to or synonymous with Strobilanthes [Baker 1884; Benoist 1967]). Dischistocalyx and Acanthopale also share the feature of having hygroscopic trichomes restricted to the margins of seeds, although this state has evolved multiple times elsewhere in Ruellieae. Future study with added taxon sam- pling may provide new insights into relationships between these two genera. Dischistocalyx and Acanthopale can be differentiated by calyx morphology (strongly zygomorphic vs. [sub]actinomorphic), ovule number per ovary (more than 20 vs. 2), pollen morphology (triporate with apertures surrounded by a ring of echinae vs. pantoforate and lacking echinae rings), and internal corolla vestiture (without vs. with flattened trichomes). As for Dischistocalyx and Acanthopale, Satanocrater and Ruellia have also been thought to be close relatives on the basis of pollen features (Scotland and Vollesen 2000), but the two were not resolved as sister taxa in this study (fig. 6). Plants of Satanocrater and Ruellia have Wabenpollen mor- This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 119 Fig. 20 Morphological diversity in Mimulopsinae. A, Mimulopsis alpina (not vouchered, Tanzania), showing general macromorphology. B, Mellera lobulata (Tripp & Ensermu 932 [RSA], Ethiopia), showing general macromorphology and stiff trichomes on lower corolla lip. C, Mellera nyassana (Bingham 10169 [MO], Zambia), showing stiff trichomes on lower corolla lip. D, Epiclastopelma glandulosum (Mabberley 1144 [K], Tanzania), showing putatively bird-adapted flowers and leaves with dentate margins. E, Ionacanthus calcaratus (Humbert & Capuron 25789 [K], Madagascar), showing putatively bird-adapted flowers and highly dissected leaves. F, Eremomastax speciosa (not vouchered, cultivated at Jardin Botanique de Lyon), showing unique corolla lobe arrangement. G, Heteradelphia paulojaegeria (Tripp & Ly 939 [RSA], cultivated at RSABG Greenhouses), showing standard corolla lobe arrangement and bird-adapted flowers. Photos by E. A. Tripp except A, by I. Darbyshire (used with permission), and F, by J. Haager (used with permission). phology in common (fig. 14G, 14N). The vast majority of species of Ruellia have porate pollen (Daniel 1998; Tripp 2007, 2010), although study of some African and Malagasy species (Furness and Grant 1996), as well as of one NW lineage of Ruellia (i.e., those formerly treated in Blechum; Tripp et al. 2009), indicated that some species have colporate pollen (fig. 14I). All Satanocrater studied to date have either porate or subcolporate pollen (Tripp and Fatimah 2012). Outside of Ruelliinae, the only other genera in Ruellieae with Wabenpollen are Eranthemum and Kosmosiphon (fig. 14A, 14B; Sharma and Vishnu-Mittre 1963), but those genera are resolved in a different clade with strong support. Thus, while morphological attributes (Wabenpollen plus four fertile stamens per flower) would seem to suggest that Satanocrater and Ruellia are sister taxa, molecular data presented here resolve Acanthopale as sister to Ruellia, which together are sister to Satanocrater. The genus Satanocrater contains four species that are united by the large, prominent glands (fig. 16C) that cover all plant surfaces (including external and internal portions of flowers) and their remarkable inflated calyces with lobes fused nearly to the apex (fig. 16D), thus making them appear tubular (Tripp and Fatimah 2012; the calyces split into segments following desiccation and/or immersion in water). These two features readily distinguish Satanocrater from all other genera in the family, although one other genus (Physacanthus) is similar in calyx morphology (Tripp et al., forthcoming). As such, it is not surprising that the six accessions of Satanocrater included in this study formed a strongly supported monophyletic group (fig. 6; similarly recovered in Tripp and Fatimah 2012). Plants of Satanocrater produce extremely aromatic essential oils that are detectable from both fresh and dried material, as well as during the DNA extraction process (E. A. Tripp and S. Fatimah, personal observation). We suspect that the large, dense glands that cover plant surfaces bear these aromatic compounds. Glands of Satanocrater are more than twice the diameter of similar glands of all other Acanthaceae studied to date (E. A. Tripp, unpublished data). Personal observations (E. A. Tripp, M. Thulin) indicate that two of the four species (fig. 16D) produce flowers that are visited by sunbirds, whereas the other two appear to be adapted to insect pollination. At least one species, S. paradoxus, has flowers that produce thin, black nectar, according to herbarium label data (Sebsebe and Ensermu 8705 [ETH]). Chromosome numbers have not yet been reported for the genus. The genus Ruellia (fig. 16E) contains perhaps 350 species that are distributed in both temperate and tropical areas worldwide but are concentrated in tropical portions of the Americas This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 120 INTERNATIONAL JOURNAL OF PLANT SCIENCES Fig. 21 Morphological diversity in Strobilanthinae, showing general macromorphology of several taxa. A, Baphicacanthus cusia (Tripp & Deng 1562 [RSA], China). B, Hemigraphis fluviatis (Tripp & Deng 1580 [RSA], China). C, Strobilanthes sp. (Tripp & Deng 1563 [RSA], China). D, Strobilanthes dyeriana (not vouchered, Duke University Greenhouses). E, Strobilanthes dimorphotricha (Tripp & Chen 1822 [RSA], China), showing contort corolla aestivation. Photos by E. A. Tripp. and, to a lesser extent, tropical and subtropical Africa (Tripp 2007). At present, no diagnostic features beyond pollen type (spherical, triporate, reticulate exine, i.e., Wabenpollen; fig. 14N) and a near-constant chromosome number of n¼17 are known for Ruellia (Tripp 2007). However, within Ruelliinae only Satanocrater shares the trait combination of Wabenpollen and four fertile stamens, but Satanocrater is highly distinctive morphologically (see above) and cannot be confused with Ruellia. Molecular data presented here indicate that several plants that have been ascribed to other genera are part of a clade that includes all sampled species of Ruellia: Benoicanthus (fig. 14H), Blechum (fig. 14I), Eusiphon (fig. 14J), Lychniothyrsus (fig. 14K; also see Tripp and McDade 2012), Polylychnis (fig. 14L), and Pseudoruellia (fig. 14M). Indeed, forcing monophyly of Ruellia exclusive of these other genera resulted in a significantly less likely tree (H10; table 2), and Ruellia s.l. (i.e., including the above-listed genera) was strongly supported in our study (fig. 6). Two of these, Blechum and Lychniothyrsus, were recently proposed as synonyms of Ruellia on the ba- sis of molecular data and morphological features (Tripp et al. 2009; Tripp and McDade 2012). Two others, Eusiphon and Polylychnis, were partially synonymized in Tripp (2007). Here, we reduce the remaining two genera, Benoicanthus and Pseudoruellia, along with remaining species in Eusiphon and Polylychnis, to synonymy. All aforementioned genera have the Ruellia pollen type (fig. 14H–14N). Multiple accessions of each segregate genus (Benoicanthus ¼ 3; Blechum ¼ 2; Lychniothyrsus ¼ 3; Polylychnis ¼ 2; Pseudoruellia ¼ 3; Eusiphon not sampled here) were monophyletic in our study (fig. 6). Plants previously ascribed to Benoicanthus are endemic to Madagascar; have attractive long, white flowers; and have half-fused calyx lobes that bear some resemblance to Eusiphon (¼Ruellia geayi (Benoist) E. Tripp, not sampled in this analysis). Heine and Raynal (1968) noted the similarity of Benoicanthus to Ruellia but distinguished the two by number of seeds per capsule and other features. However, those authors did not take into account the full range of variation of these traits in Ruellia. Plants previously ascribed to the monospecific Pseudoruellia, This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 121 Fig. 22 Macromorphology (A, C–E) and pollen (B) of Calacanthus grandiflorus, incertae sedis (A, C–E, not vouchered; B, collector unknown 7621 [US], India). Note general macromorphology and bilabiate corollas in A; bicolporate, polypseudocolpate pollen with sexine lips in B; putative corolla fold (cf; labeled with an arrow) in C; and the semifleshy leaves with crenate margins in D. Photos by S. Yadav (used with permission) except B, by E. A. Tripp. also endemic to Madagascar, are red flowered and have remarkable fleshy leaves (fig. 16F), which is otherwise unknown in Ruellia. Pseudoruellia was erected by Benoist (1962), but the sole species, Pseudoruellia perrieri, was first described in Ruellia (Benoist 1945). In that publication, Benoist noted that its distinctive calyx morphology was shared by other Malagasy Ruellia, concluding that ‘‘c’est un veritable Ruellia’’ (p. 5). Tripp (2007) discussed relationships of Eusiphon (Madagascar) and Polylychnis (Guianas) to Ruellia. The remaining combinations (i.e., those not proposed in Tripp 2007) are herein provided. The genus Spirostigma was described by Nees (1847a) on the basis of plants collected by Martius from ‘‘Araracoara,’’ which most likely refers to Araraquara of the Colombian Amazon (H. Esser, personal communication). We have seen no material ascribed to this name except the type specimen, an image of which was courteously provided to us by staff at M. Although not sampled for this study, examination of the traits of this plant permit inferences regarding its likely placement. Nees (1847a) described densely pubescent plants with unilocular anthers and capsules containing more than four seeds each. While the unilocular anthers suggest affinity to Acantheae (indeed, Nees treated Spirostigma in his Aphelandreae [¼Acantheae] and contrasted the genus to Stenandrium [also of Acantheae]), the number of seeds per capsule suggests affinity to some other tribe (Acantheae have four or fewer seeds per capsule). Examination of the type specimen clearly shows a plant with capsules containing more than four seeds; additionally, these seeds bear hygroscopic trichomes. These two features suggest that Spirostigma is better classified in Ruellieae than in Acantheae, as others have done (Lindau 1895; Scotland and Vollesen 2000). Most likely Nees erred in diagnosing anther thecae number. The dense stem and leaf pubescence and capsule morphology of the type specimen resemble features of several species of Ruellia from Amazonia (e.g., R. putumayensis Leonard). Indeed, Lindau (1895) treated Spirostigma within Ruellieae and among members with Wabenpollen (in the Neotropics, only Ruellia). More specifically, the distribution This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 122 INTERNATIONAL JOURNAL OF PLANT SCIENCES of trichomes on seed margins only further places this plant in the Physiruellia clade (Tripp 2007) of Ruellia. Below, we make the necessary combination into Ruellia. Further study of Colombian Ruellia is needed to determine whether other names (especially those described by Leonard [1958]) refer to taxa that are conspecific with those under discussion. Ruellia tachiadena (Heine & A. Raynal) E. Tripp, comb. nov. Basionym: Benoicanthus tachiadenus Heine and A. Raynal, Adansonia ser. 2. 8:192. 1968. Ruellia gruicolla Benoist Syn. Benoicanthus gruicollis (Benoist) Heine and A. Raynal Ruellia perrieri Benoist Syn. Pseudoruellia perrieri (Benoist) Benoist Ruellia longissima (Benoist) E. Tripp, comb. nov. Basionym: Eusiphon longissimus Benoist, Notul. Syst. (Paris) 15:6. 1955. Ruellia wasshauseniana E. Tripp, comb. et nom. nov. Basionym: Polylychnis ovata Wassh., Fl. Guianas, ser. A, Phanerogams 23:103. 2006. [The epithet ‘‘ovata’’ is already occupied in Ruellia; as such, a new epithet is needed, and ‘‘wasshauseniana’’ is proposed to commemorate Dieter Wasshausen for his substantial contributions to knowledge of South American Acanthaceae.] Ruellia radicans Nees Syn. Polylychnis radicans (Nees) Wassh. (incl. Polylychnis essequibensis Bremek., see Wasshause 2006) Ruellia hirsutissima (Nees) E. Tripp, comb. nov. Basionym: Spirostigma hirsutissima Nees, Fl. Bras. 9:84. 1847. Subtribes III–IV: Trichantherinae, Hygrophilinae, Petalidiinae, Mimulopsinae, and Strobilanthinae Among our most important findings is a strongly supported clade, the TMPSH clade, that contains five subtribes: Trichantherinae, Hygrophilinae, Petalidiinae, Mimulopsinae, and Strobilanthinae (fig. 3). This clade is united by the presence of porate or colporate and pseudocolpate pollen grains with the pores flanked by sexine lips sensu Scotland (1993). There have been secondary reversals of these characters in some of these subtribes (e.g., loss of pseudocolpi in one genus of Trichantherinae and some Strobilanthinae; pollen lacking sexine lips in one genus of Trichantherinae, some Strobilanthinae, and all Hygrophilinae). Wood et al.’s (2012) new genus Diceratotheca also has pollen with prominent sexine lips, but we are as yet uncertain as to its subtribal affinities (see ‘‘Incertae Sedis’’). So far as is known, however, pseudocolpi are not present in any Erantheminae or Ruelliinae and thus characterize only genera in the TMPSH clade. Note that pseudocolpi are also found in many Justicieae (fig. 1), which led Tripp et al. (forthcoming) to hypothesize this feature as synapomorphic for Justicieae þ Ruellieae. This hypothesis merits further evaluation in light of these results and in the context of more detailed phylogenetic results for Justicieae. Within the TMPSH clade, three lineages are part of a basal polytomy: Trichantherinae, Strobilanthinae, and Hygrophilinae þ (Petalidiinae þ Mimulopsinae), albeit without strong support for this third clade. Among these three lineages, all plants in Strobilanthinae and Hygrophilinae lack anther appendages, whereas plants in the other three subtribes have them. An important distinction regarding arrangement of apertures and pseudocolpi can be made among these five tribes. Whereas Trichantherinae, Hygrophilinae, and Strobilanthinae have colporate pollen (with numerous modifications to the colporate condition in Strobilanthes), Petalidiinae and Mimulopsinae have porate pollen (or brevicolporate in Dyschoriste). The clade containing Hygrophilinae þ (Petalidiinae þ Mimulopsinae) is characterized by plants with 12-pseudocolpate pollen, which we here hypothesize to be diagnostic for the group. However, there have been modifications of this number via both increases and decreases in pseudocolpi number in several groups (e.g., to 21 per grain in Ruelliopsis, 24 per grain in Hygrophila, 18–21 per grain in most species of Mimulopsis, and 48 per grain in some Dyschoriste, reduced to 0 in some Dyschoriste [this last part fide Scotland 1993]). Additionally, some Strobilanthes (of Strobilanthinae) also produce 12-pseudocolpate pollen, such that this trait has evolved in parallel in Ruellieae. A second feature that characterizes the vast majority of plants in Hygrophilinae þ (Petalidiinae þ Mimulopsinae) but that is by no means restricted to this group is herringbone patterning in corolla throats, which presumably functions as nectar guides. It is present, to our knowledge, at varying degrees of prominence in all 12 genera that comprise the three subtribes. Somewhat similar patterning is known elsewhere in Acanthaceae (e.g., many Justicieae). The sister-group relationship between Petalidiinae and Mimulopsinae supports earlier studies that have discussed genera in these two subtribes as belonging to a single group (such as Bremekamp’s [1965] Petalidiinae, which contained members of both Petalidiinae and Mimulopsinae as herein circumscribed). This is largely attributable to pollen features shared between Petalidiinae and Mimulopsinae, a combination of traits we here consider synapomorphic for the clade containing these two subtribes: pollen grains that are triporate (tricolporate in most Dyschoriste), have sexine lips surrounding the apertures, and are 12-pseudocolpate (secondarily altered in some genera). We here treat these genera as two distinct subtribes on the basis of number of ovules per ovary and distribution of hygroscopic trichomes on seed surfaces, as discussed below. Subtribe III: Trichantherinae As herein delimited, Trichantherinae contain six genera: Bravaisia, Louteridium, Sanchezia, Suessenguthia, Trichanthera, and Trichosanchezia. Within Trichantherinae, branching patterns, from earliest diverging to most nested, were as follows, mostly with strong support (fig. 7): Louteridium is sister to the remaining five genera, Trichanthera þ Bravaisia are together sister to the remaining three genera, and finally Trichosanchezia is sister to Sanchezia þ Suessenguthia. A close relationship of the last five genera has been previously hypothesized (Bentham 1876; Lindau 1895, in part; Daniel 1988; Scotland and Vollesen 2000). Trichantherinae are restricted to the Neotropics, occur in mesic environments, and are characterized by their shrub to large tree habits. Within Trichantherinae, pollination syndromes are diverse and range from bat to hummingbird to bee. Staminal configuration and condition is also variable in the subtribe and has been useful This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) for generic circumscriptions (Daniel 1988; Schmidt-Lebuhn 2003). Finally, pollen morphologies in Trichantherinae are among the most distinctive of all Acanthaceae. Pollen of Louteridium is unique and highly autapomorphic (fig. 14O), whereas another unique pollen type is synapomorphic for the clade containing the remaining five genera (fig. 14P– 14T). Except for molecular data, we know of no synapomorphies for Louteridium þ the other five genera. Louteridium is a highly distinctive genus distributed in primary tropical rain forests from Mexico south to Panama. All species in the genus produce large, gibbous, pale-colored (greenish or yellowish) corollas with long, exserted stamens and styles (fig. 17A). Greenhouse and field observations by us on three species indicate that flowers open at dusk and fall from plants in the morning. Thus, we here hypothesize bat pollination for Louteridium (also predicted by Manktelow [2000]) on the basis of morphology, timing of anthesis, and observations of bats visiting flowers (Daniel 1993). However, in other bat-pollinated Acanthaceae, hummingbirds are known to be frequent visitors and thus potentially important floral pollinators at crepuscular hours (Tripp 2010; L. A. McDade, personal observation). Indeed, Richardson (1972) hypothesized hummingbird pollination for Louteridium and reported observing bird visitors to a population of one species, as did Daniel (2010). Thus, plants of some species of Louteridium may be adapted to pollination by hummingbirds, bats, or both; alternatively, plants of some species may represent an evolutionarily intermediate stage between hummingbird and bat pollination, as predicted in Ruellia (Tripp 2010; also see Stebbins 1970; SanMartin-Gajardo and Sazima 2005). Field observations (Richardson 1972; Daniel 2010) on variation in flower color and visitation for L. donnell-smithii S. Watson lend some support to a hypothesis of intermediacy. Unfortunately, except for limited observations reported by Daniel (2010), to our knowledge no new information regarding field ecology of species of Louteridium has been published in the 4 decades since Richardson’s (1972) monograph. Collections are few, and detailed ecological studies are lacking. Without doubt Louteridium is ripe for further field study, and the exceptional large and conspicuous populations of some species of Louteridium observed by one of us (T. F. Daniel) would provide ideal study sites. In addition to the highly distinctive corolla morphology, species of Louteridium produce three-parted calyces, two or four fertile stamens (if two, then flowers also have two staminodes), and ovaries with more than 12 ovules. Pollen of Louteridium is large (more than 100 um), spherical, and pantoforate and has distinctly verrucate surface sculpturing (fig. 14O; Daniel 1998). No other Acanthaceae produce pollen of similar morphology and size. The ‘‘insulae’’ of pollen of the new genus Diceratotheca (Wood et al. 2012) are superficially similar to the verrucae of Louteridium, but the former has pollen with three equatorial pores (vs. pantoforate in Louteridium). The combination of three calyx lobes and distinctive pollen morphology is synapomorphic for Louteridium. Chromosome numbers have not yet been reported for the genus, to our knowledge. Bravaisia þ Trichanthera and Trichosanchezia þ (Sanchezia þ Suessenguthia). Bravaisia, Trichanthera, Trichosanchezia, Sanchezia, and Suessenguthia formed 123 a strongly supported clade in our study (fig. 7). These genera are united by shared pollen morphology (fig. 14P–14T), which is among the most distinctive of all forms within Acanthaceae and arguably among the most distinctive in angiosperms: grains are bicolporate (Scotland 1993; Daniel 1998) and have numerous bands of pseudocolpi (hence, ‘‘Rippenpollen’’ in Lindau 1895) between the apertures. The most remarkable feature of grains is the orientation of the pseudocolpi: those associated with a germinal aperature on one face are oriented 90° (i.e., perpendicular) to those on the opposite face. Sexine lips (cf. Scotland 1993) are also associated with each colpus (fig. 14P–14T). Multiple reports exist for this pollen type in the fossil record (Kuyl et al. 1955; Germeraad et al. 1968; Regali et al. 1974; Pocock and Vasanthy 1988), which are useful for divergence time estimations in Acanthaceae (E. A. Tripp and L. A. McDade, unpublished manuscript). Bravaisia þ Trichanthera. Among 221 genera in Acanthaceae (sensu Scotland and Vollesen 2000), the tree growth form is best developed in Bravaisia (fig. 17B, 17D, 17F) and Trichanthera (fig. 17C, 17E), with plants achieving heights of some 20 m (Daniel 1988). A few species in several other genera in the family (e.g., Aphelandra, Justicia, Mimulopsis, Populina, Ruellia) are also known for their large statures but are perhaps not as large as in Bravaisia and Trichanthera. In some regions of Latin America, Bravaisia is used as a street tree (fig. 17D), and Trichanthera is used as a living fence (fig. 17E). Two species of Bravaisia are commonly found in Neotropical mangrove habitats (Daniel 1988), and morphological features of these plants (e.g., aerial or prop roots; fig. 17F) reflect this adaptation. Moist habitat plants of Trichanthera also produce prop roots, which led Daniel (1988) to speculate as to its mangrove potential. The sister-group relationship between Bravaisia and Trichanthera was resolved in both Bayesian and ML analyses but in neither case was supported (fig. 7 and not shown). Bravaisia and Trichanthera are readily differentiated by gross floral morphology and, more specifically, staminal traits. Flowers of Bravaisia are short infundibular and white or pink with pigmented nectar guides (fig. 17B); those of Trichanthera are longer infundibular, deep red, and densely pubescent externally—hence the generic name (fig. 17C). Flowers of both genera produce four fertile stamens, but those of Bravaisia have appendaged thecae, whereas those of Trichanthera lack appendages. Filament curtains are of the ‘‘reduced’’ type in both genera, although this condition is apparently variable in Bravaisia (Manktelow 2000). Both genera have been observed to have flowers visited abundantly by bees and hummingbirds (E. A. Tripp, personal observation), and it has been reported that Trichanthera is also visited by bats (Rosales 1997). Chromosome numbers have not yet been reported for these genera, to our knowledge. Trichosanchezia þ (Sanchezia þ Suessenguthia). A close relationship of Sanchezia to Suessenguthia and Trichosanchezia has been noted (Leonard and Smith 1964; SchmidtLebuhn 2003) since the description of the latter two in the 1920s and 1950s, respectively (Sanchezia was described in the eighteenth century). Results here confirm a close relationship among the three (fig. 7). Pollen of Trichosanchezia, Sanchezia, and Suessenguthia is of the Bravaisia and Trichanthera This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 124 INTERNATIONAL JOURNAL OF PLANT SCIENCES sort (fig. 14R–14T; Daniel 1988; Schmidt-Lebuhn 2005). Unlike Bravaisia and Trichanthera, however, plants of these three genera grow to be large shrubs but never trees. Additionally, Bravaisia and Trichanthera occur primarily from Mexico to Colombia and Venezuela (Trichanthera also occurs in Brazil and Ecuador), whereas Trichosanchezia, Sanchezia, and Suessenguthia occur predominantly in Ecuador, Peru, and Bolivia, with some species of Sanchezia extending northward into Venezuela, Colombia, and southern Central America. From a macromorphological perspective, Trichosanchezia and Sanchezia resemble one another owing to their shared tubular, bird-adapted flowers (e.g., fig. 17G), but they are easily separated by staminal features (see below) and by the extremely dense golden-yellow stem pubescence of the former (hence the specific epithet of the monospecific Trichosanchezia, T. chrysothrix; ‘‘chryso’’ ¼ gold, ‘‘thrix’’ ¼ hair). Flowers of Suessenguthia (fig. 17H) are either infundibular (likely bee adapted) or tubular (likely bird adapted). Unlike Trichosanchezia, plants of Sanchezia and Suessenguthia often (but not always) have large, brightly colored bracts. The three genera, however, are readily differentiated by staminal features. First, whereas Trichosanchezia and Suessenguthia have four fertile stamens per flower, Sanchezia has 2 fertile stamens þ 2 staminodes. Second, anthers of Trichosanchezia lack appendages, whereas anthers of Suessenguthia and Sanchezia are appendaged. Thus, we postulate a morphological transition series from four to two fertile stamens and from anthers without to with appendages. Phylogenetic data presented here support this hypothesis, as do data from Schmidt-Lebuhn et al.’s (2005) study of Suessenguthia and Sanchezia. Data from the latter study indicated that Suessenguthia forms a paraphyletic group, from which Sanchezia is derived. Our study reveals a strongly supported clade containing these two genera, of which only Suessenguthia was monophyletic. Forcing monophyly of the four accessions of Sanchezia in our study resulted in a significantly less likely tree (H2; table 2). Further study with increased taxon sampling and more variable markers is needed to fully understand the evolutionary history of this group. While taxonomy within the small genera Suessenguthia (six species; Schmidt-Lebuhn 2003) and Trichosanchezia (one species) is relatively well understood, that of the large genus Sanchezia (;60 species) has long been problematic (D. Koenemann and E. A. Tripp, unpublished manuscript). Sanchezia is in critical need of revisionary work. We here consider the staminal features described above (two vs. four stamens and appendaged vs. lacking appendages) to uniquely characterize each of the three genera. Chromosome numbers have not yet been reported for Trichosanchezia or Suessenguthia. In Sanchezia, n¼40 has been reported for S. parvibracteata, and the highest chromosome numbers known for any Acanthaceae have been reported for S. nobilis: n¼;66 (Grant 1955) and n¼68 (Singh 1951; Kaur 1970). Such numbers likely reflect polyploid ancestry (Daniel 2000). Subtribe IV: Hygrophilinae As here delimited, Hygrophilinae contain only two genera: Hygrophila and Brillantaisia. A relationship between these has long been noted (Nees 1847b; Bentham 1876; Lindau 1895; Bremekamp and Bremekamp 1948; Scotland 1993; Furness 1994) owing to their polyspermous capsules, bilabiate corollas, and four-colporate pollen grains (fig. 14U, 14V), the last state otherwise unknown in Ruellieae. Sidwell (1998) explicitly explored the hypothesis of a close relationship between Hygrophila and Brillantaisia using morphological data and a limited outgroup sampling (five other genera); she recovered a clade containing Hygrophila þ Brillantaisia in which the former comprised a basal grade from which the latter were derived and monophyletic. Using a much expanded taxon sample here, phylogenetic analyses similarly support a close relationship between Brillantaisia and Hygrophila, and again results demonstrate that Hygrophila is nonmonophyletic, from which Brillantaisia is derived and monophyletic. We here consider the four-colporate pollen grains, in combination with unappendaged anthers, as diagnostic for the strongly supported Hygrophilinae (fig. 9). Brillantaisia (figs. 14U, 18A, 18B), as treated by Sidwell (1998), contains 12 named species (13 sensu Vollesen 2008) that are distributed throughout tropical Africa but that are most diverse in Guinean-Congolese forests. The genus is highly distinctive morphologically. Plants of all species have strongly bilabiate corollas in which the upper lip is laterally compressed and forms a prominent hood (fig. 18A, 18B), an androecium with 2 fertile stamens þ 2 staminodes, and slender, ellipsoid capsules containing a minimum of 12 seeds (up to 60 seeds per capsule in some species; Sidwell 1998). In addition, most species have broadly ovate leaves with winged petioles (seen in the background of fig. 18B). This trait combination easily differentiates Brillantaisia from all other genera in Acanthaceae, and results from our study indicate that the four sampled species form a strongly supported monophyletic group (fig. 9). An unpublished manuscript (K. Sidwell) reports, to our knowledge, the only documented case of a floral ‘‘lever-arm’’ mechanism during plant pollination in Acanthaceae. At the junction of the upper and lower corolla lips is an articulating membranous hinge. This hinge permits pivoting of both lips during bee visitation, which forces the two fertile anthers and style out of the upper hood and brings them into contact with dorsal surfaces of the pollinator. The presence of 2 fertile stamens þ 2 staminodes and the strongly hooded corollas have most frequently been used to distinguish Brillantaisia from Hygrophila (Furness 1994; Vollesen 2008), but see below for exceptions to stamen configuration. A chromosome number of n¼16 has been reported for two species of Brillantaisia (Mangenot and Mangenot 1958, 1962). Results from this study indicate that Hygrophila (figs. 14V, 18C) is nonmonophyletic (also see Sidwell 1998), and forcing monophyly of Hygrophila was rejected by our data (H3; table 2). Instead, species of Hygrophila formed two clades: one early diverging in Hygrophilinae, and one sister to Brillantaisia (fig. 9). Only the latter was strongly supported, and its sister relationship to Brillantaisia was also strongly supported. The three species in this latter group, H. cataracte, H. pilosa, and H. schulli, are African. In the other clade, H. costata and H. difformis are NW, whereas H. didynama is African and H. salicifolia is Asian. At present, we know of no morphological synapomorphies for either of these two clades. However, Furness (1994) noted some variation in pseudocolpi arrangement between OW and NW Hygrophila; thus, pollen This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) should be included in future study of the genus. As the name suggests, plants of Hygrophila usually occupy moist to wet habitats. Some species, such as the widespread weed H. polysperma of Asian origin, are aquatic; this species, which was originally spread by aquarium enthusiasts, has become invasive and noxious in several regions (Duke et al. 2000; Mora-Olivo et al. 2008). Like Brillantaisia, plants of Hygrophila produce polyspermous, thin-walled, ellipitical capsules. Unlike Brillantaisia, most species of Hygrophila have four fertile stamens, but some share the condition of 2 fertile stamens þ 2 staminodes (Vollesen 2008; Hu et al. 2011) with Brillantaisia (although not those that are in the clade that is sister to Brillantaisia). Other differences between the two genera include plant stature and corolla shape. Plants of most (but not all) species of Brillantaisia are large herbs or shrubs, whereas those of Hygrophila are usually smaller. As discussed above, the strongly bilabiate hooded corollas of Brillantaisia are highly characteristic of that genus. Although flowers of Hygrophila are typically bilabiate but not strongly hooded, some species, such as H. schulli, do produce flowers with weakly hooded upper lips (fig. 18C; E. A. Tripp, personal observation). Because Hygrophila is a large and exceptionally understudied genus (Clarke 1908), worldwide study of it is needed to fully understand trait variation, distribution, and species limits. At present, it is premature to propose new taxonomic concepts for the genus; future studies should consider the possibility of resurrecting one or more generic synonyms of Hygrophila (see Heine 1962). Chromosome numbers have been reported for 11 species referable to Hygrophila (n¼6, 12, 14–17, 22, and 32; Grant 1955; Miège 1962; De 1966; Verma and Dhillon 1967; Govindarajan and Subramanian 1983; Saggoo and Bir 1986; Khatoon and Ali 1993; Singh et al. 1993; Bala and Gupta 2011), with n¼16 being the mostly commonly reported number (eight species). Subtribe V: Petalidiinae As here delimited, Petalidiinae contain six genera: Duosperma, Dyschoriste, Phaulopsis, Petalidium, Ruelliopsis, and Strobilanthopsis. Two genera, Apassalus and Sauteria, are here considered synonyms of Dyschoriste. Nearly all species in Petalidiinae have four or fewer ovules per ovary (the only exception being Ruelliopsis, with an inferred secondary increase in ovule number to eight per ovary), and all species bear seeds with hygroscopic trichomes that cover their entire surface. These two features characterize Petalidiinae and distinguish the subtribe from its sister group, Mimulopsinae. An additional feature that helps to characterize Petalidiinae is presence of anthers with basal appendages, although these have been secondarily lost in some species of some genera in the subtribe. Our phylogenetic study resolved three strongly supported clades within Petalidiinae: Phaulopsis þ Ruelliopsis, Duosperma þ Petalidium, and Strobilanthopsis þ Dyschoriste (fig. 11). Although genera within Petalidiinae have long been associated with one another (reviewed by Manktelow 1996), this is the first proposal of these three sets of sister-group relationships, to our knowledge. Two genera of Petalidiinae have been revised taxonomically (Phaulopsis, Manktelow 1996; Duosperma, Vollesen 2006). 125 Phaulopsis þ Ruelliopsis. At present, we know of no morphological synapomorphies to unite Phaulopsis and Ruelliopsis, a relationship that was strongly supported in our analyses (fig. 11). Previous workers have associated the larger of the two genera, Phaulopsis, with Petalidium (Lindau 1895; Manktelow 1996). In contrast, close relationships between Ruelliopsis and other genera have not been proposed. Phaulopsis (figs. 14W, 19A) is a genus of upright or prostrate herbs, several of which become woody at the base with age. In this study, the five accessions sampled formed a strongly supported monophyletic group (fig. 11). Owing to the exceedingly complex and congested nature of inflorescences of Phaulopsis (fig. 19A), species identification has been challenging, making Manktelow’s thorough monograph (1996) an especially useful contribution. Phaulopsis is one of two genera in Ruellieae that is characterized by capsules with placentas and retinacula that, as a unit, fracture and separate from the capsule wall. The other genus in Ruellieae that shares this trait is Petalidium, but note that fracturing placentas have also evolved in a few other species in Ruellieae (e.g., Ruellia blechum, Ruellia erythropus) as well as in other genera of Acanthaceae (e.g., Dicliptera). Flowers of Phaulopsis bear four stamens, and anthers range from blunt to appendaged, with appendages to 0.5 mm in length. The four stamens per flower in combination with fracturing placentas and 12-pseudocolpate (rarely more; Manktelow 1996; Daniel and Figueriedo 2009) pollen with two areas of raised tectum per aperture (i.e., sexine lips) is here considered diagnostic for Phaulopsis (note that this combination would also characterize Petalidium except that pollen of the latter genus has four areas of raised tectum per aperture; see below). Manktelow (1996) summarized chromosome counts (mostly n¼17) of Phaulopsis, and Daniel and Chuang (1998) discussed them with respect to a basic number of x¼16 for the genus. Ruelliopsis (figs. 14X, 19B, 19C) is a little-studied yet highly distinctive genus restricted to arid southwestern Africa. Two accessions included in our study were resolved as sister (fig. 11). The genus was originally described to contain two species, R. setosa and R. mutica, but the latter was transferred to Hygrophila by Vollesen (forthcoming). A third species, R. damarensis, was described by Moore (1907) on the grounds of having a more tubular calyx and fewer ovules per ovary than R. setosa. We have seen only an image of the type specimen for this species (courteously provided by staff at BM); the plant does appear to have a slightly more tubular calyx than R. setosa (as noted in Moore’s protologue), but we are skeptical that its differences warrant species recognition. Additional material ascribed to this name should be studied to fully understand taxon limits within the genus. Ruelliopsis is unique within Petalidiinae, indeed nearly so in Acanthaceae, by its ‘‘grass mimic’’ growth form (fig. 19C). Plants are strongly stoloniferous, growing prostrate against the ground and producing tufts of leaves that are shaped like linear grass blades (fig. 19B). From a distance, one cannot differentiate a landscape dominated by Ruelliopsis from one dominated by grasses (fig. 19; E. A. Tripp, personal observation). Also of note is pollen of Ruelliopsis that has undergone a secondary increase in pseudocolpi number, from 12 to 18–21 (fig. 14X). Flowers of Ruelliopsis bear four stamens with appendaged anthers. The This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 126 INTERNATIONAL JOURNAL OF PLANT SCIENCES growth form in combination with stamen number and anther ornamentation—and 18–21-pseudocolpoate pollen—is here considered diagnostic for the genus. Chromosome numbers have not yet been reported for the genus. Petalidium þ Duosperma. Our phylogenetic results demonstrate a strongly supported sister relationship between Petalidium and Duosperma (fig. 11). Although to our knowledge this relationship has not before been hypothesized except in Tripp (2007), the genera share at least one synapomorphy: both produce only two ovules per capsule, with two aborted ovules sometimes also present (note that since Tripp [2007] we have determined that the original DNA extraction of Duosperma crenatum is contaminated, i.e., the species is not closely related to genera in Trichantherinae, as suggested in Tripp [2007]; see the corrected phylogenetic relationship of Duosperma crenatum herein presented [fig. 11]). Also potentially synapomorphic is that both genera retain fruits enclosed in papery bracts (fig. 19D, 19E) until a drenching rainfall event (E. A. Tripp, personal observation; Obermeijer 1936; Vollesen 2006); such events may take place only at intervals of months to years in regions with sporadic precipitation. Petalidium (fig. 19E–19I) is a remarkable genus of ;32 species almost entirely restricted to a small region of tropical and subtropical southwest Africa, bounded roughly by the Namib and Kalahari Deserts (curiously, one species, P. barlerioides, is from India). Species diversity is concentrated in the highly arid mountains of extreme northwestern Namibia and southeastern Angola. All species have two leaflike bracts that enclose flowers and fruits (fig. 19F, 19I). In many species of Petalidium, these bracts are modified and showy (e.g., fig. 19F, 19I). In other species of Petalidium, inflorescences are highly modified, with flowers and fruits occurring in extremely dense, compact heads; in these species, depending on the degree of inflorescence/infructescence density, the paired bracts subtending flowers may be obscured or may protrude beyond calyces and corollas and sometimes become spiny with age (fig. 19H). Like Phaulopsis, Petalidium also produces capsules with fracturing placentae (fig. 19G) and, like many other members of Petalidiinae (as well as Mimulopsinae), has flowers with herringbone corolla patterning (fig. 19I). In addition to the above-mentioned distinctive features, Petalidium is unique within Ruellieae by its pollen with four raised areas of tectum surrounding each aperture (i.e., elongate columellae above and below each aperture, in addition to flanking it; see the left-hand grain of fig. 14Y). In polar view, pollen of Petalidium is also triangular, although other Acanthaceae also have triangular pollen. All species of Petalidium have four stamens; in some species anthers are appendaged, whereas in others anthers lack appendages. Similarly, some species have four-parted calyces, while others are five-parted. Fieldwork in Namibia by the first author and colleague K. Dexter has documented many cases of sympatry among species with morphological intermediates present, which calls for research into reproductive isolating mechanisms and speciation within the genus (E. A. Tripp and K. Dexter, manuscript in preparation). The combination of paired, leaflike bracts; two-seeded capsules with fracturing placentae; triangular pollen in polar view with 12 pseudocolpi; and four areas of raised tectum serve as diagnostic features for this distinctive genus. Chromosome number has been counted only in the disjunct Indian species, P. barlerioides, but several counts have been made, all revealing n¼16 (Ellis 1962; Pal 1964; Bir and Saggoo 1981; Saggoo and Bir 1982). As for Phaulopsis (Manktelow 1996), knowledge of the genus Duosperma (figs. 14Z, 19D) has been augmented considerably by monographic treatment (Vollesen 2006). The 26 species are shrubby herbs of arid bushlands in tropical Africa. Species diversity is concentrated in southeastern Kenya to north-central Tanzania and again from southwestern Tanzania to northern Zambia and southern Democratic Republic of the Congo. Species of Duosperma produce four stamens per flower (rarely 2 fertile stamens þ 2 staminodes) and lack anther appendages. Vollesen (2006) reported that vegetative portions of plants are strongly odoriferous with a creosote-like smell, as are those of many other genera of African Acanthaceae (E. A. Tripp, personal observation). The two-seeded capsules lacking fracturing placentae and distinctive pollen (12 pseudocolpate with sexine lips) can be considered diagnostic for Duosperma. Chromosome numbers have not yet been reported for the genus, to our knowledge. Strobilanthopsis þ Dyschoriste. In our analyses, Strobilanthopsis and Dyschoriste were strongly supported as sister taxa (fig. 11); we are unaware of any earlier hypothesis of this relationship. Plants of both genera have ovaries with four ovules and four stamens with basally appendaged anthers. This last feature has been reported incorrectly as lacking for Strobilanthopsis (Burkill and Clarke 1900; Milne-Redhead 1932); close examination of numerous specimens reveals minute appendages. Strobilanthopsis (fig. 14AA) consists of two species, although only one of two names, S. linifolia, is in current use (the other species, S. prostrata, was described by Milne-Redhead [1932] as distinct from S. linifolia by its prostrate growth form and glabrous leaves; it is treated as a synonym of S. linifolia in Vollesen [forthcoming]). The two accessions of S. linifolia that we sampled for phylogenetic analysis formed a strongly supported clade (fig. 11). Strobilanthopsis is superficially reminiscent of Mellera (Mimulopsinae; see below) owing to its frequently obovate bracts and calyx lobes; this resemblance was similarly noted by Milne-Redhead (1932). However, numerous features of Strobilanthopsis differentiate it from Mellera: the former has entire (never distinctly crenate) leaves, has a maximum of four (vs. eight or more) ovules per ovary, has a regular pattern of pseudocolpi arrangement (vs. tectum bands between pseudocolpi fusing at poles), has 18 or more (vs. 12) pseudocolpi (fig. 14), and has inconspicuous (vs. extremely prominent) anther appendages. Strobilanthopsis is easily differentiated from its sister genus, Dyschoriste, by its calyx lobes that are free to the base (vs. fused for a quarter to, more commonly, half of their length) and that do not have hyaline margins (see below) and by its less conspicuous anther appendages. Traits that differentiate Strobilanthopsis from Dyschoriste are here considered diagnostic for Strobilanthopsis, although we note that we have not seen or studied any material attributed to S. prostrata. Plants of Strobilanthopsis are extremely variable, especially vegetatively (e.g., leaves range from linear to ovate); the genus occurs in tropical southern Africa from Zambia to Zimbabwe to northern South Africa. It has been reported on the labels of This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) herbarium specimens that some plants of Strobilanthopsis produce floral fragrances. This condition has been noted in the field for other African Acanthaceae (e.g., Ruellia and Petalidium; E. A. Tripp and K. Dexter, personal observations), whereas nearly all NW Acanthaceae lack floral fragrances detectable by the human nose. Chromosome numbers have not yet been reported for Strobilanthopsis. Dyschoriste (figs. 14BB, 19J), with ;60 species, is one of three pantropical genera in Ruellieae (table 1). Our phylogenetic analyses indicate that the genus is strongly supported as monophyletic only if Sautiera (fig. 19K) and Apassalus (fig. 14CC) are included in it (fig. 11). However, a hypothesis of monophyly of Dyschoriste exclusive of Apassalus (and Sautiera, which was not included in the test because of its basal phylogenetic position; see fig. 11) could not be rejected by our data (H8; table 2). Species of Dyschoriste are broadly distributed in tropical habitats of both Africa and the Americas, particularly in arid or seasonally dry environments. Plants are perhaps most easily recognized as belonging to Dyschoriste by their calyces: lobes are fused for at least a third of their length, with hyaline regions bordering each lobe (fig. 19J, 19K). Dyschoriste was described just 2 yr before Decaisne’s Sautiera (1834), which may explain why Decaisne made no comparison to Dyschoriste in his protologue of Sautiera (i.e., he might not have been aware of Dyschoriste). We have seen several specimens of the Timor-endemic and monospecific Sautiera and conclude that it is notable among Dyschoriste only by its distinctly bilabiate corollas (fig. 19K; similarly, C. B. Clarke made extensive notes on a specimen at Kew describing similarities of Sautiera to Dyschoriste). We therefore make the combination below. A larger taxon sample would permit exploration of whether this taxon is indeed basal among Dyschoriste. Kobuski (1928) described a genus of small, creeping plants from the Caribbean, Apassalus (‘‘without pegs’’), as distinct from Dyschoriste by its lack of anther appendages. However, many aspects of plants of Apassalus agree with Dyschoriste (e.g., calyx morphology including hyaline margins, number of seeds per capsule), and other authors have concurred that the generic distinction is unwarranted (Long 1970; Greuter and Rankin Rodrı́guez 2010; Wunderlin and Hansen 2010). As such and given its nested phylogenetic position within Dyschoriste (assuming synonymization of Sautiera; fig. 11), we make the remaining necessary combinations below. We consider the following combination of traits to be diagnostic for Dyschoriste: partially fused calyx lobes with hyaline margins, four stamens with appendaged anthers (lost in some species from the Caribbean and southeastern United States), four seeds per capsule, seeds fully covered by hygroscopic trichomes, and triaperturate, 12-pseudocolpate pollen with sexine lips. The partially fused calyx lobes with hyaline margins of Dyschoriste readily separate it from its sister genus, Strobilanthopsis. Most species of Dyschoriste, including those formerly treated in Apassalus, have chromosome numbers of n¼15 or 30 (Grant 1955; Daniel et al. 1990; Daniel 2000). A basic number of x¼15 appears likely for the genus. Dyschoriste tinctorum (Decaisne) E. Tripp & T.F. Daniel, comb. nov. Basionym: Sautiera tinctorum Decaisne, Nouv. Ann. Mus. Par. 3:383. 1834. [Decaisne cited Ruellia bilabiata 127 Rhwdt. as a synonym of Sautiera tinctorum, but the former name was never validly published.] Dyschoriste diffusa (Nees) Urb. Syn. Apassalus diffusus (Nees) Kobuski Dyschoriste cubensis Urb. Syn. Apassalus cubensis (Urb.) Kobuski Dyschoriste humistrata (Michx.) Kuntze Syn. Ruellia humistra Michx.; Apassalus humistratus (Michx.) Kobuski Dyschoriste parvula (Alain & Leonard) Greuter & R. Rankin Syn. Apassalus parvulus Alain & Leonard Subtribe VI: Mimulopsinae As here delimited, Mimulopsinae contains four genera: Eremomastax, Heteradelphia, Mellera, and Mimulopsis. Two other genera, Epiclastopelma and Ionacanthus, are here reduced to synonyms of Mimulopsis and Mellera, respectively. To our knowledge, all species in Mimulopsinae produce eight or more ovules per ovary (some plants—e.g., those that have been referred to Epiclastopelma—may have fruits with fewer fully developed seeds) and seeds with hygroscopic trichomes restricted to their margins (trichomes are lacking completely in one species of Heteradelphia; Daniel and Figuerido 2009). All genera in Mimulopsinae also characteristically produce leaves with crenate to dentate margins and few secondary veins. These three features characterize Mimulopsinae and distinguish the subtribe from its sister group, Petalidiinae, in which nearly all species produce ovaries with four or fewer ovules, seeds with trichomes that cover entire surfaces, and leaf features not as described above. However, as described below, although the subtribe itself is well marked morphologically, constituent genera are not well circumscribed. Mimulopsinae should be prioritized for future revisionary work; its small taxonomic size and restricted geographic distribution should facilitate such an endeavor. Mimulopsis þ Mellera. Our data resolved a weakly supported clade (fig. 10) consisting of Mimulopsis, Mellera, Epiclastopelma, and Ionacanthus. Within this clade, Mimulopsis can be considered monophyletic only if the oligospecific Epiclastopelma is included in it, but we were not able to reject an alternative scenario in which Mimulopsis exclusive of Epiclastopelma is monophyletic (H7; table 2). In contrast, Mellera was resolved as nonmonophyletic and consisted of two strongly supported clades, one of which contained the monospecific genus Ionacanthus; an alternative hypothesis of monophyly of Mellera (exclusive of Ionacanthus) was rejected by our data (H6; table 2). In this study, we consider Epiclastopelma to be synonymous with Mimulopsis and Ionacanthus to be synonymous with Mellera, and the combinations are made below. Plants that have been ascribed to these two genera are notable among their respective relatives for putative adaptations to bird pollination—species of Mimulopsis and Mellera are otherwise adapted to insect pollination. While some earlier authors have considered Mimulopsis (fig. 20A) and Mellera (fig. 20B) to be distantly related (Lindau 1895), recent authors (Vollesen 2008) have noted a close relationship between them. Several morphological traits support This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 128 INTERNATIONAL JOURNAL OF PLANT SCIENCES the close relationship recovered in phylogenetic analyses. Pollen of all species examined to date (Mimulopsis: 7 species; Mellera: 5 species), including new combinations proposed below, show subpolar fusion of the outermost pseudocolpi in each mesoporium, resulting in pseudocolpal ellipses (fig. 14DD– 14GG). This shared derived feature is not seen elsewhere in Ruellieae (although it is known elsewhere in Acanthaceae). Species of Mimulopsis and Mellera are also characterized by corollas with long, stiff, retrorse trichomes or spines that line the lower lip and/or throat (fig. 20C; these are secondarily lost in Epiclastopelma, and the internal corolla surfaces of Ionacanthus were not seen). Finally, plants of Mimulopsis and Mellera tend to have prominently petioled leaves, leaves with crenate to dentate margins (to highly dissected in plants attributed to Ionacanthus), and obovate or spathulate calyx lobes (fig. 20B). Species of Mimulopsis are primarily large shrubs to treelets that occupy evergreen, montane forests of tropical Africa, including Madagascar. Many if not most species in the genus are known to flower gregariously (Vollesen 2008), as occurs in other genera of Ruellieae (e.g., Acanthopale, Strobilanthes). Mimulopsis has been differentiated from Mellera primarily on the basis of features of anther ornamentation and corolla shape. Both genera produce four stamens, but the thecae of most Mimulopsis lack appendages except for the outermost locule of the longer pair of stamens; in contrast, all thecae of the four stamens are appendaged in most Mellera. In Mimulopsis, corollas tend to be subactinomorphic with a very short narrow unexpanded portion of the corolla tube (see McDade and Tripp 2007 for terminology) and a much expanded, nearly campanulate throat, whereas corollas of Mellera are generally more strongly zygomorphic and have a comparatively longer narrow unexpanded portion of the corolla tube (contrast fig. 20A vs. 20B). Although these differences readily differentiate most species of Mimulopsis and Mellera, there are exceptions to both that prevent strict assignment of character states to the two genera. Indeed, neither genus has ever been monographed, which will be necessary before unambiguous diagnostic traits for the two groups can be proposed. The oligospecific genus Epiclastopelma (figs. 14FF, 20D) is endemic to the southern Eastern Arc Mountains of Tanzania. It was described in the late nineteenth century by Lindau, who thought it was closely related to genera in his Petalidieae (e.g., Petalidium, Phaulopsis) rather than to Mimulopsis or Mellera, which he treated in his Strobilantheae and Hygrophileae, respectively (Lindau 1895, 1897). More recent authors have noted the similarity of Epiclastopelma to Mimulopsis (Vollesen 2008) but have opted to maintain the former as a distinct genus owing to its strikingly different corolla form and corollas that lack retrorse spines or trichomes (Vollesen 2008 also noted that plants produce only four seeds per capsule; however, ovaries can contain 12–16 ovules). Like Mimulopsis, plants of Epiclastopelma produce flowers with only the outermost locules of the two longer stamens appendaged, although this trait is apparently variable (Vollesen 2008). The two corolla differences—that is, shape and the lack of retrorse spines—most likely reflect adaptation to pollinators that differ from the rest of Mimulopsis; indeed, the Uluguru and Udzungwa mountain ranges to which plants of Epiclastopelma are restricted to are noted for their high diversity of sunbirds, which is reflected by bird pollination adaptation in other genera of Acanthaceae (Darbyshire 2008a). Because of this and because analyses consistently resolved Epiclastopelma as nested within Mimulopsis with strong support, we reduce the former to a synonym of the latter. Chromosome numbers have not yet been reported for Mimulopsis or plants ascribed to Epiclastopelma. Phylogenetic data presented here resolved two strongly supported clades of Mellera that together formed a basal grade with respect to Mimulopsis (fig. 10). As discussed above, species of Mellera tend to produce flowers in which all thecae of all four stamens are appendaged. However, those of some species are minute (e.g., M. submutica) or occur on only one of two anther locules. Given variability in this trait and others among species in this genus and considering the fact that neither Mellera nor its closest relative has been monographed, it is not surprising that the genus is not monophyletic. We here retain the two genera as distinct, pending further study to clarify species composition of lineages of Mellera. The monospecific genus Ionacanthus (fig. 20E), endemic to Madagascar, was described by Benoist (1940) and distinguished from Mimulopsis by its narrowed corolla that is spurred at the base. Like Epiclastopelma, flowers of Ionacanthus appear to be adapted to bird pollination, unlike flowers of its closest relatives; however, all other macromorphological plant features appear consistent with Mellera and Mimulopsis (we have not been able to study internal corolla structure or pollen morphology of Ionacanthus owing to a paucity of material). Because of this and our phylogenetic results, we synonymize Ionacanthus with Mellera, noting that future work on Mellera and Mimulopsis may further alter generic concepts (the name Mimulopsis has priority). Notably, diverse floral morphologies are currently recognized under single generic concepts in other Acanthaceae (e.g., Ruellia, Tripp and Manos 2008; Barleria, Darbyshire 2009; Tetramerium, Daniel 1986; Daniel et al. 2008). Chromosome numbers have not yet been reported for Mellera or plants previously ascribed to Ionacanthus. Mellera calcarata (Benoist) E. Tripp, comb. nov. Basionym: Ionacanthus calcaratus Benoist, Notul. Syst. (Paris) 9:65. 1940. Mimulopsis volleseniana E. Tripp & T.F. Daniel, nom. nov. Basionym: Epiclastopelma glandulosum Lindau, Bot. Jahrb. Syst. 22:114. 1895. [The combination Mimulopsis glandulosa (Lindau) Bullock was illegitimate because the epithet ‘‘glandulosa’’ was already in use; as such, a new epithet is needed, and ‘‘volleseniana’’ is proposed to commemorate Kaj Vollesen for his substantial contributions to knowledge of African Acanthaceae.] Mimulopsis macrantha (Mildbr.) E. Tripp, comb. nov. Basionym: Epiclastopelma macranthum Mildbr., Notizbl. Bot. Gart. Berlin-Dahlem 11:1079. 1934. Mimulopsis marronia (Vollesen) E. Tripp, comb. nov. Basionym: Epiclastopelma marroninum Vollesen, Fl. Trop. E. Africa Acanth. 1: 232. 2008. Eremomastax þ Heteradelphia. Phylogenetic results suggest that both Eremomastax (figs. 14HH, 20F) and Heteradelphia (figs. 14II, 20G) are nonmonophyletic (branches strongly supported), and alternative tests of monophyly were rejected This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) by the data (H4 and H5; table 2). Although similarities between the two genera have been noted by other authors (Scotland 1993; Daniel and Figueiredo 2009), their nonmonophyly is somewhat surprising owing to marked floral differences (fig. 20F, 20G). In Eremomastax, all five corolla lobes comprise the lower lip; thus, an upper lip is completely lacking (although not known elsewhere in Ruellieae, the striking corolla lobe arrangement seen in Eremomastax has evolved in other tribes of Acanthaceae, e.g., Acantheae; fig. 1). In contrast, flowers of Heteradelphia have the standard ‘‘two up, three down’’ lobe arrangement that characterizes the vast majority of other Acanthaceae. However, studies of Acanthaceae have demonstrated the evolutionary vagility of floral architecture within some genera (McDade 1992; Daniel et al. 2008; Tripp and Manos 2008). In this study, we defer making any nomenclatural changes in these two genera until more molecular data for more accessions can be added to phylogenetic analyses. As presently understood, Heteradelphia comprises two species of mesic habitats, one restricted to mainland tropical West Africa, and one endemic to the oceanic island of São Tomé. Eremomastax is monospecific and occurs geographically somewhat more broadly, from tropical West to tropical East Africa. Daniel and Figueiredo (2009: Heteradelphia) and Vollesen (2008: Eremomastax) provide modern generic circumscriptions of the two. Chromosome numbers have not yet been reported in either genus. Subtribe VII: Strobilanthinae There have been numerous attempts to make taxonomic sense of the diverse set of plant morphologies encompassed by the large subtribe Strobilanthinae (Bremekamp 1944a; Terao 1983; Moylan 1999; Venu 2007; see Carine and Scotland 2002 for a review), the other major species-rich subtribe in Ruellieae along with Ruelliinae. Scotland and Vollesen (2000) recognized six genera (Aechmanthera, Baphicacanthus [figs. 14JJ, 21A], Clarkeasia [fig. 14KK], Hemigraphis [figs. 14LL, 21B], Stenosiphonium [fig. 14MM], and Strobilanthes [figs. 14NN, 21C–21E]), and numerous additional genera had been proposed previously to accommodate the morphological variation (Bremekamp 1944a). Although a small number of groups within Strobilanthinae are ‘‘readily discernable’’ (Carine and Scotland 2000; Carine et al. 2004; Deng et al. 2006; Wood and Scotland 2009), apparently extreme levels of homoplasy prevent diagnoses of groups of species within the lineage (Carine and Scotland 2002; Moylan et al. 2002). This is especially reflected in pollen morphology in the subtribe Strobilanthinae (Carine and Scotland 1998; Wood and Scotland 2009), which encompasses a diversity almost as great as that contained among all Ruellieae. This diversity in pollen type has been discussed extensively elsewhere (Carine and Scotland 1998) and thus will not be elaborated on here (but see fig. 14JJ–14NN for a glimpse of this diversity). As argued convincingly by Carine and Scotland (2002) and Moylan et al. (2004a), the only practical solution to the taxonomy problem seems to be to place all segregate genera under an expanded concept of the genus Strobilanthes (i.e., Strobilanthes s.l.; see also Carine and Scotland 2002; Deng et al. 2007; Wood and Scotland 2009). Results from our phylogenetic study support this notion: the genus Strobilanthes 129 is grossly nonmonophyletic (fig. 8), and forcing monophyly of the six sampled accessions resulted in a significantly less likely tree (H10; table 2). If Aechmanthera, Baphicacanthus, Clarkeasia, Stenosiphonium, and Hemigraphis are included in Strobilanthes s.l., the genus (and sole constituent of the subtribe) is monophyletic and strongly supported. However, we defer making nomenclatural changes owing to the active research program on this group by others (R. Scotland, J. R. I. Wood, Y. Deng, and colleagues) who have already begun the process of a ‘‘nomenclatural cleanup’’ in this diverse lineage (Deng and Xia 2007; Wood and Scotland 2009). Of the five additional species resolved in Strobilanthinae in our study, names in Strobilanthes already exist for two of them: Aechmanthera tomentosa Nees ¼ Strobilanthes tomentosa (Nees) J.R.I. Wood and Baphicacanthus cusia (Nees) Bremek. ¼ Strobilanthes cusia Nees. Combinations for Clarkeasia, Hemigraphis, and Stenosiphonium still need to be made. The following combination of three traits is considered synapomorphic for species in Strobilanthinae: the presence of rows of hairs on the inner corolla wall as well as a ‘‘rugula’’ or a fold on the inner surface of the posticous corolla wall, both of which purportedly function to retain the style, and a ‘‘strobilanthoid’’ type of filament curtain, which comprises a transverse fusion of five filaments that form a ridge along the corolla wall (Manktelow 2000; Carine and Scotland 2002; Moylan et al. 2004b; note that the rugula of Strobilanthinae is generally not as prominent as nor is homologous to the rugula that characterizes some Justicieae; see fig. 1). Other traits have also been cited in attempts to delimit the group (e.g., filaments united to form a curtain, a bifid stigma with a reduced posterior lobe, simple or compound spicate or racemose inflorescences, and pollen apertures between the tectal ribs; Manktelow 2000; Moylan et al. 2004b), but these features occur commonly in other species of Ruellieae and other Acanthaceae. One fascinating aspect of the biology of Strobilanthes s.l. is the gregarious flowering that characterizes some species. Such ‘‘plietesial’’ species grow for 10–15 yr, flower synchronously, fruit, and then die. Wood (1994) documented and discussed mass flowering Strobilanthes in Bhutan. Plietesial species are known in other genera of Acanthaceae (e.g., Stenostephanus and Isoglossa of Justicieae; Daniel 2006; Darbyshire 2008b; fig. 1), but the phenomenon is best exemplified in Strobilanthes s.l. and other genera of Ruellieae (e.g., Acanthopale, Mimulopsis). Another interesting feature of some species of Strobilanthes s.l. is the thigmotrophic stigmas (Grant 1955). Just as Ruellia has diversified into numerous morphologies and niches in the NW, the genus Strobilanthes has done so in the OW. Future comparative studies of the evolutionary histories of these two groups may yield new insights into correlates of species richness in angiosperms. Numerous chromosome numbers (n¼8–11, 13–16, 20, 21, and 30) have been reported for more than 40 species of Strobilanthes (Daniel and Chuang 1989), with n¼16 being the most common number (18 species). Daniel and Chuang (1998) proposed x¼8 as the most likely basic number for the genus. Incertae Sedis Calacanthus. Calacanthus (fig. 22) was erected by T. Anderson in Bentham (1876) for a plant endemic to India This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 130 INTERNATIONAL JOURNAL OF PLANT SCIENCES with ‘‘flores magni, purpurei’’ (p. 1088; fig. 22A). It was a new generic name provided for Lepidagathis grandiflora, earlier described by Dalzell (1850) as ‘‘a species remarkable not only for its size (5–6 ft.), but also for its large and showy flowers, an inch and a half in length, of a purple colour, with two rows of bright yellow hairs down the centre of the lower lip; the leaves are 1 foot long and 4 inches broad. The flowers turn black in drying’’ (p. 138). On the basis of protologue overlap, it seems clear that the authors were writing about the same taxon. Although first described as related to members of Barlerieae (to which Lepidagathis belongs; fig. 1), later authors have associated the plant with Ruellieae on the basis of several features, including floral buds with contort aestivation (Anderson 1876; Lindau 1895; Scotland and Vollesen 2000). We agree with these latter assessments but have unfortunately not yet been able to locate any type material for this plant (Dalzell’s protologue indicates that plants were ‘‘Crescit in montibus Syhadree, lat. 16–19°; fl. Dec.’’). Despite this and the lack of ample or recent collections of Calacanthus grandiflora (thus no DNA sequences), numerous photos have been made of these plants in recent years. These provide information that is conflicting relevant to placement of Calacanthus among the seven subtribes of Ruellieae. We were also able to sample poorly preserved pollen from one collection. Pollen of Calacanthus is bicolporate (Scotland and Vollesen 2000) with sexine lips and pseudocolpate (fig. 22B). The pseudocolpi and sexine lips almost unambiguously associate the plant with the TMPSH clade. More specifically, the presence of compound apertures (colpori) suggests a relationship to Trichantherinae, Hygrophilinae, or Strobilanthinae. Whereas bicolporate grains in Ruellieae are otherwise known only in Trichantherinae, grains of Calacanthus lack the pseudocolpal arrangement (90° opposing faces) highly synapomorphic for bicolporate members of that clade. Other features of Calacanthus suggest a relationship to Hygrophilinae. Manktelow (2000) reported Calacanthus to have a ‘‘corolla fold’’ type of filament curtain, a subtype most frequent in Hygrophilinae (but also occurring in a few other genera of Ruellieae). Indeed, examination of photos of living plants of Calacanthus seems to suggest the presence of a corolla fold (fig. 22C) or hinge-like mechanism similar to that of Brillantaisia. The large, ovate leaves with winged petioles (fig. 22E) and bilabiate corollas (fig. 22A) are also reminiscent of Brillantaisia, as are the flowers that turn black upon drying and the herringbone patterning on corollas of Calacanthus (fig. 22D; also present in other subtribes). Still, other plant features of Calacanthus suggest a possible relationship to Strobilanthinae. Whereas Manktelow (2000) reported a corolla fold filament curtain type in Calacanthus, Anderson’s protologue (1876) seems to describe a ‘‘strobilanthoid’’ filament curtain type: ‘‘Stamina 4, didynama, supra medium tubum affixa, labiis multo breviora, filamentis basi per paria connatis et cum staminodio intermedio minimo dentiformi postice linea transversa v. membrane brevi connexis’’ (p. 1088). The semifleshy leaves of Calacanthus (fig. 22E) that have crenate margins are also highly reminiscent of Strobilanthes s.l. or Brillantaisia. In sum, features of this remarkable genus link it to Trichantherinae, Hygrophilinae, and Strobilanthinae of Ruellieae. Fresh material and DNA sequence data will help to place the genus with confidence. Diceratotheca. Wood et al. (2012) erected a new monospecific genus, Diceratotheca, to accommodate plants recently collected from Thailand. Plants of Diceratotheca are notable among Ruellieae by having anther thecae with two basal appendages each. Pollen grains of plants of Diceratotheca are spheroidal and triaperturate (with apertures arranged equatorially and surrounded by sexine lips) and have the autapomorphic trait of exine surfaces covered by conspicuous insulae. Plants of Diceratotheca almost certainly belong in Ruellieae owing to their left-contort corolla aestivation. However, further placement within Ruellieae is complicated by incongruity between molecular data and morphological features. Wood et al. (2012) used sequence data from four chloroplast markers to reconstruct relationships among a selection of genera in Ruellieae. Their results indicated that Diceratotheca was allied to Eranthemum or Pararuellia, both of Eranthiminae as herein circumscribed (fig. 5). However, so far as is known no Erantheminae possess appendaged anthers or pollen with sexine lips. Instead, both of these traits are frequent in the TMPSH clade. Wood et al. (2012) similarly acknowledged this discrepancy, noting that appendaged anthers characterize genera such as Dyschoriste, Mimulopsis, and Hygrophila and that pollen with sexine lips characterizes genera such as Epiclastopelma, Eremomastax, and Mellera as well as a few Strobilanthes. We defer placement of Diceratotheca in a subtribe of Ruellieae pending further study of material. Echinacanthus. The name Echinacanthus has been applied to plants from India and Nepal as well as China. The former is accurate, whereas the latter represents misapplication of the generic name. Chinese plants are not ascribable to Echinacanthus, and a new genus is to be proposed in a forthcoming paper (G. Chunming et al., manuscript in preparation) to accommodate those. The Indo-Nepalese plants of Echinacanthus have calyces with partially fused lobes and slightly hyaline margins, herringbone patterning inside corollas, four stamens bearing appendaged anthers, and seeds with trichomes covering their entire surface. These traits suggest placement of Echinacanthus in Petalidiinae, specifically near Dyschoriste. We have not seen pollen of Echinacanthus ourselves but, on a specimen at Kew, C. B. Clarke illustrated it as colporate (with pseudocolpi), which fits pollen of Dyschoriste but not other Petalidiinae. However, Lindau (1895) wrote ‘‘S. [presumably Saatgut] meist 8 im Fach’’ (p. 301), whereas most Petalidiinae have four or fewer ovules. Although it is likely that Echinacanthus will be placed in Petalidiinae, we take the conservative approach of not placing the genus in any subtribe until further data help to resolve the issue. We do, however, typify Echinacanthus below to prevent further confusion of application of the generic name. Echinacanthus attenuatus Nees, Pl. Asiat. Rar. (Wallich), vol. 3. 90. 1832. Lectotype selected here: Nepal, Wallich s.n. (GZU-000249910!; isolectotypes: E-00273504!, E-00273505!, K-Bentham!). The GZU specimen was selected over others because the ‘‘Nepalia’’ of the protologue appears on the sheet (the same is true of the K specimen) and has a fragment packet with floral material. Sinoacanthus ined. The genus Sinoacanthus is to be proposed for Chinese plants that have been attributed to Echinacanthus (G. Chunming et al., manuscript in preparation). Pollen of Sinoacanthus is tricolporate and autapomorphic in be- This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) ing polypseudocolpate: grains have up to 50 pseudocolpi (E. A. Tripp, unpublished data). Although the number of pseudocolpi does not help us to place the genus within a subtribal context, the compound nature of the apertures (colporate instead of porate) strongly suggests that the genus is allied to Trichantherinae, Hygrophilinae, or Strobilanthinae or to Dyschoriste of Petalidiinae. The number of germinal apertures (three) excludes placement in Trichantherinae and Hygrophilinae but agrees with placement in Strobilanthinae or near Dyschoriste. ML (but not Bayesian) analyses of data presented herein resolved Sinoacanthus as sister to Strobilanthinae, albeit without support (figs. 4, 8). However, Sinoacanthus has appendaged anthers (some plants with two appendages per thecae, like Diceratotheca), a condition unknown in Strobilanthinae but known in Dyschoriste, at least the single appendage per anther condition. We defer placing this genus in a subtribe until further study (G. Chunming et al., manuscript in preparation). Stenothyrsus. Clarke described Stenothyrsus from the Malaysian state of Perak in 1908. Although the genus is inadequately known and is poorly represented in European and other Western herbaria, it is highly distinctive morphologically by its long, terminal spicate inflorescences that reach up to 30 cm in length (E. A. Tripp, personal observation, and Clarke 1908), making it unlikely to be confused with any other Ruellieae. In the protologue, Clarke (1908) wrote that ‘‘this genus agrees in character very closely with the Tropical African genera Mellera and Eremomastax [Heteradelphia], but . . . otherwise as Hemigraphis’’ (p. 650). Bremekamp (1955) wrote that while studying material of Stenothyrsus at Kew, he discovered that flowers lacked the internal rows of trichomes that characterize the group (Strobilanthinae) to which Hemigraphis belongs. He also wrote that he studied pollen of Stenothyrsus and found it to be globose, with ‘‘four equatorial germ pores situated in the groves and of 24 or 28 meridional bands, all of the same size and shape and separated from each other by narrow groves, four of which contained the germ pores’’ (p. 651). This left him little doubt—and would otherwise leave us little doubt—that Stenothyrsus should be placed in Hygrophilinae. However, Clarke described the anther thecae as each containing ‘‘two spinules,’’ and appendaged anthers are otherwise unknown in Hygrophilinae but characterize other pseudocolpate subtribes, such as Petalidiinae, Mimulop- 131 sinae, and Trichantherinae. As such, we defer placing Stenothyrsus within any particular subtribe until further data are available to help resolve the issue. Acknowledgments We thank numerous colleagues for assistance that made this research possible: Ensermu Kelbessa and Mekbib Fekadu at ETH (fieldwork in Ethiopia and further collaboration); Deng Yunfei, Chen Fenglin, and Tang Huimin at IBSC (fieldwork in China and further collaboration); Shui Yi-Min at KUN (contribution of material of Pararuellia); and Ezekeil Kwembeya, Esmerialda Klaassen, Leevi Lanyeni, and Hendrina Hasheela at WIND (assistance with fieldwork in Namibia). We are especially indebted to our colleague Iain Darbyshire at K, who responded to and researched numerous queries, and to our colleague Kyle Dexter at E, for his extensive contributions to fieldwork in Namibia. Thanks also to Betsy Jackes and Graeme Cockes (material of Brunoniella); Robyn Barker and John Hosking (additional assistance with Brunoniella and living plant images); George Schatz at MO (photograph of Pseudoruellia); Carol Furness and Jayne Lilywhite at K (images of pollen of Kosmosiphon and Mellera insignis); John Hunnex at BM (image of Ruelliopsis damarensis type and information on Spirostigma); Christian Bräuchler, Franz Schuhwerk, and Hajo Esser at M (assistance locating type of Spirostigma); Shrirang Yadav at SUK (photographs of living Calacanthus); Jiřı́ Haager at Teplice Botanical Garden (photograph of Eremomastax); Alexander Schmidt-Lebuhn at CANBR (photograph of Suessenguthia); Bart Wursten and the Flora of Mozambique project (http://mozambiqueflora.com/; photograph of Acanthopale); S. Serata and E. Kelbessa (contributions to the SEM work conducted at CAS); Sirena Lee at SING (samples of Stenothyrsus); and Dan Koenemann, formerly at RSA (several contributions to project). Finally, we thank numerous curators and curatorial assistants at B, BM, BR, C, GZU, K, M, S, W, UPS, and Z for facilitating our visits to their herbaria. E. A. Tripp, L. A. McDade, and S. Fatimah were supported by NSF DEB-0919594 (to E. A. Tripp and L. A. McDade). T. F. Daniel was supported by NSF DEB-0743273 (to T. F. Daniel) and a fellowship from the Christensen Research Institute for fieldwork in Papua New Guinea. Appendix Voucher Information and GenBank Numbers Shown below are voucher information and GenBank numbers (Eif3E, ITSþ5.8S, psbA-trnH, trnG-trnR, trnG-trnS) for all specimens used in this study (note: Eif3E and psbA-trnH were analyzed only for taxa in Ruellieae in this study). Taxa are phylogenetically arranged from outgroups through Justicieae (see fig. 1). Outgroups: Sesamum indicum L., Jenkins 97141 [ARIZ], ITS: AF169853, trnGR: JQ7801019, trnGS: EU528998. Nelsonioideae: Staurogyne letestuana Benoist, NBG-B 200000119-77 [not vouchered], ITS: JX443805, trnGR: JQ7801020, trnGS: EU528999. Avicennia lineage: Avicennia bicolor Standl., Borg 10 [S], ITS: EU528877, trnGR: JQ780995, trnGS: EU528943. Thunbergioideae: Mendoncia phytocrenoides Benoist, Schönenberger 50 [K], ITS: AF169849, trnGR: JQ7801005, trnGS: EU528983. Acantheae: Aphelandra leonardii McDade, McDade 310 [DUKE] ITS: AF169761, trnGR: JQ780994, trnGS: DQ059287. Stenandriopsis guineensis (Nees) Benoist, cultivated at Kew, 1990-2299 [not vouchered], ITS: DQ028434, trnGR: JX443969, trnGS: DQ059258. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 132 INTERNATIONAL JOURNAL OF PLANT SCIENCES Barlerieae: Barleria lupulina Lindl., Daniel s.n. [not vouchered], ITS: AF169751, trnGR: JQ780996, trnGS: EU528946. Crabbea acaulis N.E. Br., Balkwill et al. 11649 [J], ITS: EU528885, trnGR: JQ7801000, trnGS: EU528953. Andrographideae: Cystacanthus turgida G. Nicholson, cultivated Kew 1996-479 [not vouchered], trnGR: JQ7801001, trnGS: EU528954. Whitfieldieae: Camarotea souiensis Scott-Elliot, Decary s.n. [US], trnGR: JQ780998, trnGS: JQ7801024. Chlamydacanthus euphorbioides Lindau, Daniel et al. 10445 [CAS], trnGR: JQ780999, trnGS: EU528951. Zygoruellia richardii-1 Baill. [formerly classified in Ruellieae], de Block et al. 1019 [BR], Madagascar, Eif3E: JX443744, ITS: JX443816, trnGR: JX443980, trnGS: JX444050. Zygoruellia richardii-2 Baill. [formerly classified in Ruellieae], de Block et al. 1019 [MO], Madabascar, Eif3E: JX443745, trnGR: JX443981, trnGS: JX444051. Neuracanthus lineage: Neuracanthus ovalifolius (Fiori) Bidgood & Brummitt, Friis et al. 5032 [K], ITS: EU528902, trnGR: JX443951, trnGS: EU528988. Neuracanthus umbraticus Bidgood & Brummitt, Daniel 6770.5 [CAS], ITS: EU528905, trnGR: JQ7801006, trnGS: EU528991. Justiceae: Mackaya bella Harv., Daniel s.n. [not vouchered], ITS: AF289796, trnGR: JQ7801003, trnGS: EU528979. Rhinacanthus gracilis Bojer ex Nees, Daniel s.n. [not vouchered], ITS: AF289766, trnGR: JQ7801009, trnGS: EU528995. Ruellieae: Acanthopale aethiogermanica Ensermu, Tripp & Ensermu 914 [RSA], Ethiopia, Eif3E: JX443714, ITS: JX443746, psbA-trnH: JX443817, trnGR: JX443902, trnGS: JX443982. Acanthopale confertiflora (Lindau) C.B. Clarke, Phillipson 2117 [MO], Madagascar, Eif3E: JQ763413, ITS: EF214470, psbA-trnH: JQ7801035, trnGR: EF214651, trnGS: JQ7801022. Acanthopale laxiflora C.B. Clarke, Polhill & Levitt 4985 [MO], Tanzania, ITS: EF214411, psbA-trnH: JX443818. Acanthopale pubescens C.B. Clarke, D’Arcy 17691 [MO], Equatorial Guinea, ITS: EF214477. Aechmanthera gossypina Nees, Stewart 17221 [MO], psbA-trnH: JX443819, trnGR: JX443903, trnGS: JX443983. Aechmanthera tomentosa-1 Nees [Strobilanthes tomentosa (Nees) J.R.I. Wood], Stewart 16063 [MO], India, psbA-trnH: JX443820, trnGR: JX443904. Aechmanthera tomentosa-2 Nees [Strobilanthes tomentosa (Nees) J.R.I. Wood], Nasii & Siddigi 2461 [CAS], Pakistan, trnGR: JX443905, trnGS: JX443984. Apassalus diffusus-1 Kobuski [Dyschoriste diffusa (Nees) Urb.], Garcia et al. 4412 [US], Dominican Republic, psbA-trnH: JX443821, trnGR: JX443906, trnGS: JX443985. Apassalus diffusus-2 Kobuski [Dyschoriste diffusa (Nees) Urb.], Pimentel & Garcia 513 [US], Dominican Republic, psbA-trnH: JX443822. Baphicacanthus cusia-1 (Nees) Bremek. [Strobilanthes cusia Nees], Shen 1958 [MO], Taiwan, ITS: JX443747, psbA-trnH: JX443823, trnGR: JX443907, trnGS: JX443986. Baphicacanthus cusia-2 (Nees) Bremek. [Strobilanthes cusia Nees], Shen 1855 [MO], Taiwan, psbA-trnH: JX443824, trnGR: JX443908, trnGS: JX443987. Benoicanthus tachiadenus-1 Heine & A. Raynal [Ruellia tachiadena (Heine & A. Raynal) E. Tripp], Croat 30623 [MO], Madagascar, ITS: JX443748, psbA-trnH: JX443825, trnGR: JX443909, trnGS: JX443988. Benoicanthus tachiadenus-2 Heine & A. Raynal [Ruellia tachiadena (Heine & A. Raynal) E. Tripp], Croat 30623 [US], Madagascar, ITS: JX443749, trnGR: JX443910. Benoicanthus tachiadenus-3 Heine & A. Raynal [Ruellia tachiadena (Heine & A. Raynal) E. Tripp], Daniel 11024 [CAS], Madagascar, Eif3E: JX443715, ITS: JX443750, trnGS: JX443989. Blechum brownei Juss. [Ruellia blechum L.], Sianca-Colin 1914 [MO], Mexico, ITS: EF214412, psbA-trnH: JX443882, trnGR: EF214601, trnGS: JX444039. Blechum costaricense Oerst. E. Tripp & McDade [Ruellia costaricensis (Oerst.) E. Tripp & McDade], Daniel et al. 6342 [DUKE], Costa Rica, ITS: EU812551, psbA-trnH: JX443883. Bravaisia integerrima-1 (Spreng.) Standl., cultivated Duke Greenhouses, ITS: EF214413, trnGR: EF214603. Bravaisia integerrima-2 (Spreng.) Standl., Tripp & Lujan 519 [RSA], Venezuela, ITS: JX443751, psbA-trnH: JX443826, trnGR: JX443911, trnGS: JX443990. Brillantaisia grottanellii Pic.Serm., Tripp & Ensermu 924 [RSA], Ethiopia, Eif3E: JQ763418, ITS: JX443752, psbA-trnH: JQ7801036, trnGR: JQ780997, trnGS: JQ780123. Brillantaisia madagascariensis T. Anderson ex Lindau, Daniel 10592 [CAS], Madagascar, ITS: JX443753, psbA-trnH: JX443827, trnGS: JX443991. Brillantaisia pubescens T. Anderson ex Oliv., ATBP 643 [MO], Uganda, Eif3E: JX443716, ITS: JX443754, psbA-trnH: JX443828, trnGR: JX443912, trnGS: JX443992. Brillantaisia vogeliana Benth., Daniel 11129 [CAS], São Tomé, ITS: JX443755, psbA-trnH: JX443829, trnGS: JX443993. Brunoniella australis-1 (Cav.) Bremek., Mecbold 3661 [M], Australia, psbA-trnH: JX443830. Brunoniella australis-2 (Cav.) Bremek., Daniel 10065 [CAS], Australia, psbA-trnH: JX443831, trnGR: JX443913, trnGS: JX443994. Brunoniella spiciflora (F. Muell. ex Benth.) Bremek., Daniel 10055 [CAS], Australia, ITS: JX443756, psbA-trnH: JX443832, trnGS: JX443995. Clarkeasia parviflora (T. Anderson) J.R.I. Wood, Chermsirivathana 1585 [K], Thailand, psbA-trnH: JX443833. Dischistocalyx grandifolius2 C.B. Clarke, Cheek et al. 5406 [K], Cameroon, ITS: JX443757, psbA-trnH: JX443834, trnGR: JX443914. Dischistocalyx hirsutus C.B. Clarke, Reitma et al. 2109 [MO], Gabon, trnGR: JX443915. Dischistocalyx thunbergiiflora T. Anderson, Nemba & Mambo 689 [MO], Cameroon, ITS: JX443758. Duosperma crenatum (Lindau) P.G. Mey., Zimba et al. 873 [MO], Zambia, trnGR: EF214604. Duosperma kilimandscharicum (C.B. Clarke) Dayton, Kindeketa et al. 1526 [MO], Tanzania, Eif3E: JQ763415, ITS: EF214415, psbA-trnH: JQ7801037, trnGR: EF214605, trnGS: JQ7801025. Duosperma longicalyx (Deflers) Vollesen, Tripp & Ensermu 888 [RSA], Ethiopia, Eif3E: JX443717, psbA-trnH: JX443835, trnGR: JX443916, trnGS: JX443996. Duosperma tanzaniense Vollesen, Lovett et al. 4115 [MO], Tanzania, ITS: JX443758. Dyschoriste albiflora Lindau, Luwiika et al. 580 [MO], Zambia, psbA-trnH: GQ995666, trnGR: EF214606, trnGS: GQ995605. Dyschoriste nagchana (Nees) Bennet, Tripp & Ensermu 933 [RSA], Ethiopia, psbA-trnH: JX443836, trnGR: JX443917, trnGS: JX443997. Dyschoriste oblongifolia Kuntze, Daniel 11763 [CAS], Florida, psbA-trnH: JX443837, trnGR: JX443918, trnGS: JX443998. Dyschoriste rubiginosa Ramamoorthy & Wassh., Daniel 5310 [NY], Mexico, psbA-trnH: JX443838, trnGR: JX443919, trnGS: JX443999. Dyschoriste verticillaris C.B. Clarke, Chase 1698 [NY], South Africa, psbA-trnH: JX443839, trnGR: JX443920, trnGS: JX444000. Echinacanthus longipes H.S. Lo & D. Fang [Sinoacanthus ined.], Harder et al. 4849 [MO], Vietnam, trnGR: JX443921. Epiclastopelma glandulosa Lindau [Mimulopsis volleseniana E. Tripp & T.F. Daniel], Mabberley 1144 [K], Tanzania, ITS: This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) 133 JX443760, psbA-trnH: JX443840, trnGR: JX443922, trnGS: JX444001. Eranthemum nervosum Dalzell & A. Gibson, Tripp 138 [DUKE], Costa Rica, Eif3E: JX443718, psbA-trnH: JX443841, trnGR: JX443923, trnGS: JX444002. Eranthemum tetragonum Wall., Maxwell 90-74 [MO], Thailand, Eif3E: JX443719. Eranthemum wattii Stapf, collector unknown [HERK?], Nepal, Eif3E: JX443720, psbA-trnH: JX443842, trnGR: JX443924, trnGS: JX444003. Eremomastax speciosa-1 (Hochst.) Cufod., Jongkind & Abbiw 1911 [US], Ghana, ITS: JX443761, psbA-trnH: JX443843, trnGR: JX443925, trnGS: JX444004. Eremomastax speciosa-2 (Hochst.) Cufod., Reitsma & Louis 1849 [NY], Gabon, psbA-trnH: JX443844. Eremomastax speciosa-3 (Hochst.) Cufod., Manktelow et al. 88 [UPS], country unknown, ITS: JX443762, psbA-trnH: JX443845, trnGR: JX443926, trnGS: JX444005. Hemigraphis cumingiana (Nees) Fern.-Vill., Yang et al. 8014 [MO], Taiwan, psbA-trnH: JX443846, trnGR: JX443927, trnGS: JX444006. Hemigraphis glaucescens C.B. Clarke, Maxwell 89-104 [MO], Taiwan, trnGR: JX443928. Heteradelphia paulojaegeria-1 Heine, Assi 17-169 [MO], Côte d’Ivoire, ITS: JX443763, psbA-trnH: JX443847, trnGR: JX443929, trnGS: JX444007. Heteradelphia paulojaegeria-2 Heine, Jacques-Georges 21949 [MO], Sierra Leone, ITS: JX443764, psbA-trnH: JX443848, trnGR: JX443930, trnGS: JX444008. Heteradelphia paulojaegeria-3 Heine, Bagshawe 1397 [US], location unknown, ITS: JX443765, trnGR: JX443931, trnGS: JX444009. Heteradelphia paulowilhelmia Lindau, Daniel 11464 [CAS], São Tomé, ITS: JX443766, psbA-trnH: JX443849, trnGR: JX443932, trnGS: JX444010. Hygrophila cataracte S. Moore, Bidgood et al. 4068 [MO], location unknown, ITS: JX443767, trnGR: JX443933. Hygrophila costata-1 Nees, Daniel & Pilz 9592 [MO], Honduras, Eif3E: JX443721, ITS: EF214419, psbA-trnH: JX443850, trnGR: EF214608. Hygrophila costata-2 Nees, Krapovickas & Cristobal 44012 [MO], Argentina, ITS: EF214418, trnGR: EF214608. Hygrophila didynama (Lindau) Heine, Smith 673 [MO], Zambia, ITS: JX443768, psbA-trnH: JX443851. Hygrophila difformis Blume, Hansen & Wunderlin 12893 [US], Florida, ITS: JX443769, psbA-trnH: JX443853, trnGR: JX443934. Hygrophila pilosa Raf., Gates 204 [NY], Zambia, Eif3E: JQ763419, ITS: JX443770, psbA-trnH: JQ7801038, trnGR: JQ7801002, trnGS: JQ7801026. Hygrophila salicifolia (Vahl) Nees, Pullen 8896 [US], Australia, Eif3E: JX443722, ITS: JX443771, trnGR: JX443935, trnGS: JX444011. Hygrophila schulli-1 (Buch.-Ham.) M.R. Almeida & S.M. Almeida, Tripp & Ensermu 927 [RSA], Ethiopia ITS: JX443772, psbA-trnH: JX443853, trnGR: JX443936, trnGS: JX444012. Hygrophila schulli-2 (Buch.Ham.) M.R. Almeida & S.M. Almeida, Jongkind & Schmidt 1727 [US], Ghana, ITS: JX443773, trnGS: JX444013. Ionacanthus calcaratus Benoist [Mellera calcarata (Benoist) E. Tripp], Humbert & Capuron 25.789 [BR], Madagascar, ITS: JX443774, trnGR: JX443937. Kosmosiphon azureus Lindau, Leeuwenberg 7800 [MO], Cameroon, psbA-trnH: JX443854, trnGR: JX443938, trnGS: JX444014. Kosmosiphon azureus-2 Lindau, Letouzey 10952 [BR], Cameroon, psbA-trnH: JX443855. Leptosiphonium venustum-1 (Engl.) Bremek. & Nann.-Bremek. [Pseudosiphonium ined.], Ceming 94829 [MO], China, trnGR: JX443939, trnGS: JX444015. Leptosiphonium venustum-2 (Engl.) Bremek. & Nann.-Bremek. [Pseudosiphonium ined.], You 9232 [NY], China, psbA-trnH: JX443856, trnGR: JX443940, trnGS: JX444016. Leptosiphonium cf. stricklandii F. Muell., Schodde & L.A. Cravan 4608 [US], Papua New Guinea, ITS: JX443775, psbA-trnH: JX443857, trnGR: JX443941, trnGS: JX444017. Louteridium chartaceum Leonard, Daniel & Butterwick 5905 [MO], Belize, Eif3E: JX443723, ITS: EF214420, psbA-trnH: JX443858, trnGR: EF214609, trnGS: JX444018. Louteridium conzattii Standl., Vazquez 17 [NY], Mexico, psbAtrnH: JX443859, trnGS: JX444019. Louteridium donnell-smithii S. Watson, Rees et al. 2 [MO], Belize, ITS: EF214421, trnGR: EF214610, trnGS: JX444020. Louteridium mexicanum Standl., Manriquez et al. 3758 [MO], Mexico, ITS: EF214422, psbAtrnH: GQ995626, trnGR: EF214611, trnGS: JX444021. Lychniothyrsus mollis Lindau [Ruellia ochroleuca Mart.], Hatschbach et al. 68442 [US], Brazil, Eif3E: JX443724, ITS: JX443776, psbA-trnH: JX443860, trnGR: JX443942, trnGS: JX444022. Lychniothyrsus mollis2 Lindau [Ruellia ochroleuca Mart.], Silva et al. 86 [US], Brazil, ITS: JX443777, psbA-trnH: JX443861, trnGR: JX443943, trnGS: JX444023. Lychniothyrsus mollis3 Lindau [Ruellia ochroleuca Mart.], Rodrigues et al. 80 [US], Brazil, ITS: JX443778, psbA-trnH: JX443862. Mellera lobulata-1 S. Moore, Nangoma/FAB 89 [NY], Malawi, ITS: JX443779, psbA-trnH: JX443863, trnGR: JX443944, trnGS: JX444024. Mellera lobulata-2 S. Moore, Tripp & Ensermu 923 [RSA], Ethiopia, ITS: JX443780, psbA-trnH: JX443864, trnGS: JX444025. Mellera lobulata-3 S. Moore, Bidgood et al. 4764 [CAS], Tanzania, psbA-trnH: JX443864, trnGR: JX443945. Mellera menthiodora-1 Lindau, Mwangoka et al. 4261 [MO], Tanzania, psbA-trnH: JX443866, trnGR: JX443946. Mellera menthiodora-2 Lindau, Bidgood et al. 4896 [CAS], Tanzania, ITS: JX443781, psbA-trnH: JX443867, trnGR: JX443947. Mellera nyassana S. Moore, Chapman & Chapman 8731 [MO], Zamiba, psbA-trnH: JX443868, trnGR: JX443948. Mellera submutica-1 C.B. Clarke, LaCroix 4792 [MO], Malawi, ITS: EF214424, trnGR: EF214613, trnGS: JX444026. Mellera submutica-2 C.B. Clarke, Richards 18210 [S], location unknown, Eif3E: JX443725, ITS: JX443782, psbA-trnH: JQ7801027, trnGR: JQ7801004. Mimulopsis arborescens C.B. Clarke, Poulsen et al. 1259 [MO], Uganda, ITS: EF214425, trnGR: EF214614, trnGS: JX444027. Mimulopsis excellens Lindau, Ewango 2179 [MO], Rwanda, ITS: JX443783, psbA-trnH: JX443869. Mimulopsis glandulosa Baker, Randrianjanaka & Zafy 193 [MO], Madagascar, ITS: JX443784, psbA-trnH: JX443870. Mimulopsis kilimandscharica Lindau, Gobbo 25 [MO], Tanzania, ITS: JX443785, psbA-trnH: JX443871. Mimulopsis lyalliana (Nees) Baron, Randrianaivo et al. 544 [MO], Madagascar, ITS: EF214426, trnGR: EF214615, trnGS: JX444028. Mimulopsis solmsii Schweinf., Tripp & Ensermu 934 [RSA], Ethiopia, ITS: JX443786, psbA-trnH: JX443872, trnGR: JX443949. Mimulopsis sp., Manktelow 91421 [UPS], location unknown, ITS: JX443787, psbA-trnH: JX443873, trnGR: JX443950, trnGS: JX444029. Pararuellia alata-1 H.P. Tsui, Zhiduan 960432 [MO], China, Eif3E: JX443726, ITS: JX443788, psbA-trnH: JX443874, trnGR: JX443952. Pararuellia alata-2 H.P. Tsui, Zhiduan 960476 [MO], China, Eif3E: JX443727, ITS: EF214434, psbA-trnH: JX443875, trnGR: EF214623, trnGS: JX444030. Petalidium canescens C.B. Clarke, Seydel 570 [NY], South Africa, psbA-trnH: JQ801040, trnGR: JQ7801007, trnGS: JQ801028. Petalidium lanatum C.B. Clarke, Seydel 4339 [US], Namibia, ITS: JX443789, psbA-trnH: JX443876, trnGR: JX443953. Petalidium luteoalbum A. Meeuse, Smook 7713 [MO], South Africa, Eif3E: JX443728, ITS: JX443790, psbA-trnH: JX443877, This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 134 INTERNATIONAL JOURNAL OF PLANT SCIENCES trnGR: EF214617, trnGS: JX444031. Petalidium ohopohense P.G. Meyer, Tripp & Dexter 849 [RSA], Namibia, ITS: JX443791, trnGR: JX443954, trnGS: JX444032. Phaulopsis angolana S. Moore, Champluvier 5230 [MO], Congo, ITS: EF214429, trnGR: EF214618, trnGS: JX444033. Phaulopsis imbricata-1 Sweet, Bidgood et al. 4589 [MO], Tanzania, Eif3E: JX443729, ITS: EF214430, trnGR: EF214619, trnGS: JX444034. Phaulopsis imbricata-2 Sweet, Tripp & Ensermu 929 [RSA], Ethiopia, ITS: JX443792, psbA-trnH: JQ7801041, trnGR: JQ7801008 trnGS: JQ7801029. Phaulopsis pulchella Mankt., Steiner 32 [MO], Tanzania, Eif3E: JX443730, psbA-trnH: JX443878, trnGR: JX443955, trnGS: JX444035. Phaulopsis rupestris (Nees) Lindau, Pettersson 427 [MO], Madagascar, ITS: JX443793, trnGR: JX443956. Polylychnis fulgens Bremek. [Ruellia fulgens (Bremek.) E. Tripp], Prance et al. 30665 [US], French Guiana, Eif3E: JX443732, ITS: JX443799, psbA-trnH: JX443884, trnGR: JX443961. Polylychnis ovata Wassh., Granville et al. number unknown [US], French Guiana, [Ruellia wasshauseriana (Wassh.) E. Tripp] ITS: JX443794, trnGS: JX444036. Pseudoruellia perrieri-1 (Benoist) Benoist [Ruellia perrieri Benoist], Phillipson 2988 [BR], Madagascar ITS: JX443795. Pseudoruellia perrieri-2 (Benoist) Benoist [Ruellia perrieri Benoist], Phillipson et al. 3432 [K], Madagascar, ITS: JX443796, psbA-trnH: JX443879, trnGR: JX443957, trnGS: JX444037. Pseudoruellia perrieri-3 (Benoist) Benoist [Ruellia perrieri Benoist], Daniel 10633 [CAS], Madagascar, psbA-trnH: JX443880, trnGR: JX443958. Ruellia amoena Sessé & Moc., Tripp et al. 1225 [RSA], Mexico, Eif3E: JX443731, ITS: JX443797, trnGR: JX443959. Ruellia asperula Lindau, Laurenio et al. 156 [US], Brazil, ITS: JX443798, psbA-trnH: JX443881, trnGR: JX443960, trnGS: JX444038. Ruellia humilis Nutt., Tripp 14 [PH], Pennsylvania, ITS: EF214508, psbA-trnH: GQ995632, trnGR: EF214678, trnGS: EU431038. Ruellia insignis Balf.f., Smith 566 [K, cultivated Duke Greenhouses], Socotra, ITS: EF2145113, psbA-trnH: JX443885, trnGR: EF214680, trnGS: EU431041. Ruellia primulacea Benth., Scarth-Johnson 85 [K], Australia, ITS: JX443800, psbA-trnH: JX443886, trnGR: JX443962. Ruellia sarukhaniana Ramamoorthy, Tripp et al. 1230 [RSA], Mexico, Eif3E: JX443733, ITS: JX443801. Ruelliopsis setosa-1 C.B. Clarke, Seydel 4182 [US], Namibia, psbA-trnH: JX443887, trnGR: JX443963, trnGS: JX444040. Ruelliopsis setosa-2 C.B. Clarke, Tripp & Dexter 790 [RSA], Namibia, Eif3E: JX443734, psbA-trnH: JX443888, trnGR: JX443964, trnGS: JX444041. Sanchezia lutea Leonard, McDade et al. 804 [DUKE], Panama, ITS: JX443802, trnGR: JX443965. Sanchezia peruviana (DC.) Rusby, Sanders et al. 17628 [DUKE], Costa Rica, ITS: JX443803, trnGR: JX443966, trnGS: JX444042. Sanchezia putumayensis Leonard, Vasquez et al. 12463 [MO], Colombia, ITS: JX443804, trnGR: JX443967. Sanchezia speciosa Leonard, Zak 3563a [DUKE], location unknown, Eif3E: JX443735, ITS: AF169835, trnGR: EU431005, trnGS: EU431010. Satanocrater fellatensis Schweinf., Friis et al. 6884 [C], Ethiopia, psbA-trnH: JQ7801042, trnGR: JQ7801010, trnGS: JQ7801030. Satanocrater paradoxa-1 Lindau, Gilbert et al. 7469 [MO], Ethiopia, Eif3E: JQ763420, trnGR: JQ7801011, trnGS: JQ7801031. Satanocrater paradoxa-2 Lindau, Tripp & Ensermu 906 [RSA], Eif3E: JX443736, psbA-trnH: JQ7801044, trnGR: JQ7801012. Satanocrater ruspolii Lindau, Tripp & Ensermu 904 [RSA], psbA-trnH: JQ7801046, trnGR: JQ7801015. Satanocrater somalensis-1 Lindau, Lavranos et al. 24660 [MO], Somalia, Eif3E: JX443737, psbA-trnH: JX443889, trnGS: JQ7801033. Satanocrater somalensis-2 Lindau, Thulin et al. 10098 [K], Somalia, Eif3E: JQ763421, psbA-trnH: JQ7801047, trnGR: JQ7801017, trnGS: JQ7801034. Sautiera tinctorum Decne. [Dyschoriste tinctora (Decaisne) E. Tripp], Schmutz 2939 [L], Timor, trnGR: JX443968, trnGS: JX444043. Stenosiphonium cordifolium-1 (Vahl) Alston, Iwarsson 662 [UPS], Sri Lanka, Eif3E: JX443738, ITS: JX443806, psbA-trnH: JX443890, trnGR: JX443970. Stenosiphonium cordifolium-2 (Vahl) Alston, Bremer & Bremer 911 [US], Sri Lanka, trnGR: JX443971. Stenosiphonium setosum T. Anderson, ITS sequence downloaded from GenBank: AY489379. Strobilanthes anceps Nees, Cramer et al. 6864 [US], Sri Lanka, ITS: JX443807, psbA-trnH: JX443891, trnGS: JX444044. Strobilanthes claviculata C.B. Clarke ex W.W. Sm., Sino-Amer (?) 1984 [US], China, trnGR: JX443972, trnGS: JX444045. Strobilanthes dyeriana Mast., cult. Duke Greenhouses [DUKE], Eif3E: JX443739, ITS: JX443808, psbA-trnH: JX443892, trnGR: JX443973, trnGS: JX444046 Strobilanthes forrestii Diels, Sino-Amer 1164 [US], China, ITS: JX443809, psbA-trnH: JX443893, trnGR: JX443974, trnGS: JX444047. Strobilanthopsis linifolia-1 (T. Anderson ex. C.B. Clarke) Milne-Redh., Smith 632 [MO], Zambia, Eif3E: JX443740, ITS: JX443810, psbA-trnH: JX443894, trnGR: JX443975, trnGS: JX444048. Strobilanthopsis linifolia-2 (T. Anderson ex. C.B. Clarke) Milne-Redh., Young 292 [US], location unknown, trnGR: JX443976. Suessenguthia barthleniana SchmidtLeb., Wasshausen & Wood 2278 [US], Bolivia, Eif3E: JX443741, ITS: JX443811, psbA-trnH: JX443895, trnGR: JX443977. Suessenguthia multisetosa (Rusby) Wassh. & J.R.I. Wood ex Schmidt-Leb., Vaquiata 5920 [MO], Bolivia, ITS: JX443812, trnGR: JX443978. Suessenguthia vargasii Wassh., Timaná & Rubio 2164 [MO], Peru, Eif3E: JX443742, ITS: JX443813, psbA-trnH: JX443896. Trichanthera corymbosa Leonard, Tripp & Lujan 520 [RSA], Venezuela, Eif3E: JX443743, ITS: JX443814, psbA-trnH: JX443897, trnGR: JX443979, trnGS: JX444049. Trichanthera gigantea-1 Humb. & Bonpl. ex Steud., cult. Duke Greenhouses [DUKE], ITS: JX443815. Trichanthera gigantea-2 Humb. & Bonpl. Ex Steud., Madrigal et al. 874 [US], Colombia, psbA-trnH: JX443898. Trichosanchezia chrysothrix-1 Mildbr., Chavez 98 [MO], Peru, psbA-trnH: JX443899. Trichosanchezia chrysothrix-2 Mildbr., Dı́az et al. 6954 [US], Peru, psbA-trnH: JX443900. Trichosanchezia chrysothrix-3 Mildbr., Chavez 98 [US], Peru, psbA-trnH: JX443901. Literature Cited Anderson T 1867 An enumeration of the Indian species of Acanthaceae. J Linn Soc Lond Bot 9:425–526. Baker JG 1884 Contributions to the flora of Madagascar. II. Monopetalae. J Linn Soc Lond Bot 20:159–236. Bala S, RC Gupta 2011 Acanthaceae, IAPT/IOPB chromosome data, 12. K Marhold, ed. Taxon 60:1784. Barker RM 1986 A taxonomic revision of Australian Acanthaceae. J Adel Bot Gard 9:1–286. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) Benoist R 1940 Descriptions de nouvelles Acanthacées Malgaches. Not Syst 9:65–73. ——— 1945 Descriptions de nouvelles Acanthacées Malgaches. Not Syst 12:1–16. ——— 1954 Les espéces du genre Eusiphon. Not Syst 15:5–6. ——— 1962 Nouvelles Acanthacées de Madagascar. Bull Soc Bot Fr 109:129–137. ——— 1967 Acanthacées. Pages 1–231 in H Humbert, ed. Flore de Madagascar et des Comores. Muséum National d’Histoire Naturelle, Paris. Bentham G 1876 Acanthaceae. Pages 1060–1022 in G Bentham, JD Hooker, eds. Genera plantarum. Vol 2. Reeve, London. Bir SS, MIS Saggoo 1981 Cytopalynology of certain Acanthaceae and Labiatae. J Palynol 17:93–102. Bremekamp CEB 1938 Acanthaceae. Rec Trav Bot Neerl 35:160–164. ——— 1944a Materials for a monograph of the Strobilanthinae (Acanthaceae). Verh K Ned Akad Wet Afd Natuurkd Sect 2 41:1–305. ——— 1944b Über Dischistocalyx T. And. ex Bth. und Acanthopale C.B. Clarke (Acanthaceae). Bot Jahrb Syst 73:126–150. ——— 1955 Notes on some acanthaceous genera and species of controversial position. Acta Bot Neerl 4:644–655. ——— 1964 On the systematic position of the Australian Nelsonias and Thunbergias and of the Ruellia species which by Domin were referred to Aporuellia Clarke. Proc K Ned Akad Wet 67:301–306. ——— 1965 Delimitation and subdivision of the Acanthaceae. Bull Bot Surv India 7:21–30. Bremekamp CEB, NE Nannenga-Bremekamp 1948 A preliminary survey of the Ruelliinae (Acanthaceae) of the Malay Archipelago and New Guinea. Proc K Ned Akad Wet 45:1–39. Brother Léon A 1957 Flora de Cuba. Vol 4. Dicotiledoneas: Melastomataceae a Plantaginaceae. No. 16 of Contribuciones Ocasionales del Museo de Historia Natural del Colegio de la Salle. Cultural, S.A.: Havana. Brown R 1810 Prodromus florae Novae Hollandiae et insulae VanDiemen. R Taylor, London. Burkill IH, CB Clarke 1900 Acanthaceae. Pages 1–262 and 506–515 (addenda) in WT Thiselton-Dyer, ed. Flora of tropical Africa. L Reeve, London. Carine MA, JM Alexander, RW Scotland 2004 A revision of the Strobilanthes kunthiana-group (Phlebophyllum sensu Bremekamp) (Acanthaceae). Kew Bull 59:1–25. Carine MA, RW Scotland 1998 Pollen morphology of Strobilanthes Blume (Acanthaceae) from southern India and Sri Lanka. Rev Palaeobot Palynol 103:143–165. ——— 2000 The taxonomy and biology of Stenosiphonium (Acanthaceae). Bot J Linn Soc 133:101–128. ——— 2002 Classification of Strobilanthinae (Acanthaceae): trying to classify the unclassifiable? Taxon 51:259–279. Champluvier D, B Senterre 2010 New taxa in Crossandrella, Dischistocalyx and Ascotheca (Acanthaceae) from Equatorial Guinea. Plant Ecol Evol 143:181–190. Chen W-H, Y-M Shui, Y-K Sima, R-M Zhang, Z-D Wei 2009 Pararuellia glomerata (Acanthaceae), a new species from Yunnan, China. Bot Stud 50:261–267. Clarke CB 1908 Family LXXXIX.—Acanthaceae. Pages 628–698 in G King, ed. Materials for the flora of the Malayan Peninsula. J Asiat Soc Bengal 74. Dalzell NA 1850 Contributions to the botany of western India. Acanthaceae. Pages 137–140 in WJ Hooker, ed. Hooker’s journal of botany and Kew Garden miscellany. Vol 2. Reeve & Benham, London. Daniel TF 1986 Systematics of Tetramerium (Acanthaceae). Syst Bot Monogr 12:1–134. ——— 1988 A systematic study of Bravaisia DC. (Acanthaceae). Proc Calif Acad Sci 45:111–131. 135 ——— 1993 Mexican Acanthaceae: diversity and distribution. Pages 541–558 in TP Ramamoorthy, R Bye, A Lot, and J Fa, eds. Biological diversity of Mexico: origins and distribution. Oxford University Press, New York. ——— 1995 Acanthaceae. Pages 1–158 in DE Breedlove, ed. Flora of Chiapas. Pt 4. California Academy of Sciences, San Francisco. ——— 1998 Pollen morphology of Mexican Acanthaceae: diversity and systematic significance. Proc Calif Acad Sci 50:217–256. ——— 2000 Additional chromosome numbers of American Acanthaceae. Syst Bot 25:15–25. ——— 2004 Acanthaceae of Sonora: taxonomy and phytogeography. Proc Calif Acad Sci 55:690–805. ——— 2006 Synchronous flowering and monocarpy suggest plietesial life history for Neotropical Stenostephanus chiapensis (Acanthaceae). Proc Calif Acad Sci 57:1011–1018. ——— 2008 Acanthaceae. Pages 185–190 in O Hokche, P Berry, O Huber, eds. Nuevo catálogo de la flora vascular de Venezuela. Fundación Instituto Botánico de Venezuela Dr. Tobias Lasser, Caracas. Daniel TF, TI Chuang 1989 Chromosome numbers of some cultivated Acanthaceae. Baileya 23:86–93. ——— 1998 Chromosome numbers of cultivated Acanthaceae and systematic implications. Pages 309–330 in P Mathew, M Sivadasan, eds. Diversity and taxonomy of tropical flowering plants. Mentor Books, Calicut. ——— 2010 Catalog of Guatemalan Acanthaceae: taxonomy, ecology, and conservation. Proc Calif Acad Sci 61:289–377. Daniel TF, TI Chuang, MA Baker 1990 Chromosome numbers of American Acanthaceae. Syst Bot 15:13–25. Daniel TF, E Figueiredo 2009 The California Academy of Sciences Gulf of Guinea expeditions (2001, 2006, 2008). VII. Acanthaceae of São Tomé and Prı́ncipe. Proc Calif Acad Sci 60:623–674. Daniel TF, LA McDade, M Manktelow, CA Kiel 2008 The ‘‘Tetramerium lineage’’ (Acanthaceae: Acanthoideae: Justicieae): delimitation and intra-lineage relationships based on cp and nrITS sequence data. Syst Bot 33:416–436. Darbyshire I 2008a Notes on the genus Dicliptera (Acanthaceae) in eastern Africa. Kew Bull 63:361–383. ——— 2008b Taxonomic notes and novelties in the genus Isoglossa (Acanthaceae). Kew Bull 64:401–427. ——— 2009 A reassessment of the status of Barleria sect. Cavirostrata (Acanthaceae) in tropical Africa, with a new species, B. richardsiae, described from the Tanzania-Zambia border region. Kew Bull 63:601–611. De A 1966 Cygological, anatomical and palynological studies as an aid in tracing affinity and phylogeny in the family Acanthaceae. I. Cytological studies. Trans Bose Res Inst Calcutta 29:139–175. Decaisne J 1834 Acanthaceae. Nouv Ann Mus Hist Nat 3:381–389. Deng YF, H Wang, S-S Zhou 2007 Two newly recorded species of Strobilanthes (Acanthaceae) from China. Acta Phytotaxon Sin 45:849–854. Deng YF, JRI Wood, RW Scotland 2006 New and reassessed species of Strobilanthes (Acanthaceae) in the flora of China. Bot J Linn Soc 150:369–390. Deng YF, N Xia 2007 Validation of the name Strobilanthes sarcorrhiza (Acanthaceae). Novon 17:154–155. Doyle JJ, JL Doyle 1987 A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochem Bull 19:11–15. Duke D, P O’Quinn, DL Sutton 2000 Control of hygrophila and other aquatic weeds in the Old Plantation Water Control District. Aquatics 22:4–8. Ellis JL 1962 Chromosome numbers in some members of Acanthaceae. Sci Cult 28:191–192. Ensermu K 2006 Acanthaceae. Pages 345–495 in I Hedberg, S Edwards, E Kelbessa, S Demissew, and E Persson, eds. Flora of This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions 136 INTERNATIONAL JOURNAL OF PLANT SCIENCES Ethiopia and Eritrea. Vol 5. National Herbarium, Addis Ababa, and Department of Systematic Botany, Uppsala University, Uppsala. Furness CA 1994 The pollen morphology of Hygrophila and Brillantaisia. Acta Bot Gall 141:267–278. ——— 1995 A pollen morphological study of Dyschoriste Nees and Chaetacanthus Nees (Acanthaceae: Ruellieae). Rev Palaeobot Palynol 84:331–345. Furness CA, MC Grant 1996 Pollen morphology of some Ruellia species (Acanthaceae) from Africa and Madagascar. Grana 35: 231–239. Germeraad JH, CA Hopping, J Muller 1968 Palynology of Tertiary sediments form tropical areas. Rev Palaeobot Palynol 6:189–348. Govindarajan T, D Subramanian 1983 Karyomorphological studies in South Indian Acanthaceae. Cytologia 48:491–504. Grant W 1955 A cytogenetic study in the Acanthaceae. Brittonia 8: 121–149. Greuter W, R Rankin Rodrı́guez 2010 Notes on some endemic Cuban species of Ruelliinae (Acanthaceae), on their seeds, pollen morphology and hygroscopic features. Willdenowia 40:285–304. Grubert M 1974 Studies on the distribution of myxospermy among seeds and fruits of Angiospermae and its ecological importance. Acta Biol Venez 8:315–551. Heine H 1962 Tropical African plants: XXVI*: some West African Acanthaceae. Kew Bull 16:161–183. ——— 1966 Acanthacées. Pages 45–55 in Flore du Gabon. Vol 13. Muséum National d’Histoire Naturelle, Paris. Heine H, A Raynal 1968 Benoicanthus Heine and A. Raynal (Acanthaceae), nouveau genre Malgachee. Adansonia 8:189–198. Hu J, Y Deng, JRI Wood, TF Daniel 2011 Acanthaceae. Pages 369–477 in Z Wu et al, eds. Flora of China. Vol 19. Science Press, Beijing. Huelsenbeck JP, F Ronquist 2001 MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17:754–755. Iwatsubo Y, M Mishima, N Naruhashi 1993 Chromosome numbers in four Japanese species of Strobilanthes (Acanthaceae). Chromosome Inf Serv 54:3–5 [original reference not seen]. Kaur J 1970 Chromosome numbers in Acanthaceae V. Sci Cult 36: 103–106. Khatoon S, SI Ali 1993 Chromosome atlas of the angiosperms of Pakistan. Department of Botany, University of Karachi, Karachi. Kobuski CE 1928 A new genus of the Acanthaceae. Ann Mo Bot Gard 15:1–8. Kuyl OS, J Muller, HT Waterbolk 1955 The application of palynology to oil geology with special reference to western Venezuela. Geol Mijnb 17:49–76. Leonard EC 1958 The Acanthaceae of Colombia. Contrib US Natl Herb 31:1–781 (3 vol). Leonard EC, LB Smith 1964 Sanchezia and related American Acanthaceae. Rhodora 66:313–343. Li M, J Wunder, G Bissoli, E Scarponi, S Gazzani, E Barbaro, H Saedler, C Varotto 2008 Development of COS genes as universally amplifiable markers for phylogenetic reconstructions of closely related plant species. Cladistics 24:727–745. Lindau G 1893 Beiträge zur Systematik der Acanthaceen. Bot Jahrb Syst 18:36–64. ——— 1895 Acanthaceae. Pages 274–354 in A Engler, K Prantl, eds. Die natürlichen Pflanzenfamilien 4(3b). VW Engelmann, Leipzig. ——— 1897 Acanthaceae africanae (Beiträge zur Flora von Afrika). Bot Jahrb Syst 22:114. Long RW 1970 The genera of Acanthaceae in the southeastern United States. J Arnold Arbor 51:257–309. Love K 1982 IOPB chromosome number reports LXXVI. Taxon 31: 574–598. Maddison DR, WP Maddison 2000 MacClade version 4: analysis of phylogeny and character evolution. Ver 4.08. Sinauer, Sunderland, MA. Mangenot S, G Mangenot 1958 Deuxième list de nombres chromo- somiques nouveaux chez diverses Dicotylédones et Monocotylédones d’Afrique occidentale. Bull Jard Bot Bruxelles 28:315–329. ——— 1962 Enqête sur les nombres chromosomiques dans une collection d’espèces tropical. Rev Cytol Biol Vég 25:411–447. Manktelow M 1996 Phaulopsis (Acanthacaee)—a monograph. Symb Bot Ups 31:1–183. ——— 2000 The filament curtain: a structure important to systematics and pollination biology in the Acanthaceae. Bot J Linn Soc 133:129–160. Manktelow M, LA McDade, B Oxelman, CA Furness, M-J Balkwill 2001 The enigmatic tribe Whitfieldieae (Acanthaceae): delimitation and phylogenetic relationships based on molecular and morphological data. Syst Bot 26:104–119. Mason-Gamer R, E Kellogg 1996 Testing phylogenetic conflict among molecular datasets in the tribe Triticeae (Gramineae). Syst Bot 45:524–545. McDade LA 1992 Pollinator relationships, biogeography, and phylogenetics. BioScience 42:21–26. McDade LA, TF Daniel, CA Kiel 2008 Toward a comprehensive understanding of phylogenetic relationships among lineages of Acanthaceae s.l. (Lamiales). Am J Bot 95:1136–1152. McDade LA, TF Daniel, CA Kiel, K Vollesen 2005 Phylogenetic relationships among Acantheae (Acanthaceae): major lineages present contrasting patters of molecular evolution and morphological differentiation. Syst Bot 30:834–862. McDade LA, SE Masta, ML Moody, E Waters 2000 Phylogenetic relationships among Acanthaceae: evidence from two genomes. Syst Bot 25:106–121. McDade LA, EA Tripp 2007 A synopsis of Costa Rican Ruellia (Acanthaceae), with descriptions of four new species. Brittonia 59: 199–216. Miège J 1962 Quatrième liste de nombres chromosomiques d’espèces d’Afrique Occidentale. Rev Cytol Biol Veg 24:149–164. Milne-Redhead E 1932 The genus Strobilanthopsis. Bull Misc Inform Kew 7:344–347. Moore S 1907 New or rare Acanthaceae from German south-west Africa. J Bot 45:226. Mora-Olivo A, TF Daniel, M Martı́nez 2008 Hygrophila polysperma (Acanthaceae), una maleza acuática registrada por primera vez para la flora Mexicana. Revista Mex Biodivers 79:265–269. Moylan E 1999 Hemigraphis (Nees) Acanthaceae from the Philippines: an example of species delimitation. Oxford Plant Syst 7:9–11. Moylan EC, JR Bennett, MA Carine, RG Olmstead, RW Scotland 2004a Phylogenetic relationships among Strobilanthes s.l. (Acanthaceae): evidence from ITS nrDNA, trnL-F cpDNA, and morphology. Am J Bot 91:724–735. Moylan EC, RT Pennington, RW Scotland 2002 Taxonomic account of Hemigraphis Nees (Strobilanthinae-Acanthaceae) from the Philippines. Kew Bull 57:769–825. Moylan EC, PJ Rudall, RW Scotland 2004b Comparative floral anatomy of Strobilanthinae (Acanthaceae), with particular reference to internal partitioning of the flower. Plant Syst Evol 249:77–98. Moylan EC, RW Scotland 2000 Hemigraphis neocaledonica Heine from New Caledonia is transferred to Brunoniella Bremek. Kew Bull 55:477–481. Mueller F 1886 Acanthaceae. Pages 32–34 in F Mueller, ed. Descriptive notes on Papuan plants, part I-IX. Boerhaave, Leiden. Nanda PC 1962 Chromosome number of some trees and shrubs. J Indian Bot Soc 41:271–277. Narayanan CR 1951 Somatic chromosomes in the Acanthaceae. J Madras Univ 21:220–231. Nees CG 1832 Acanthaceae. Pages 70–117 in N Wallich, ed. Plantae asiaticae rariores. Vol 3. Richter, London. ——— 1847a Acanthaceae. Pages 1–164 in C Martius, ed. Flora Brasiliensis. Vol 9. R Oldenbourg, Munich. ——— 1847b Acanthaceae. Pages 46–519 in AP de Candolle, ed. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions TRIPP ET AL.—PHYLOGENY AND RECLASSIFICATION OF RUELLIEAE (ACANTHACEAE) Prodromus systematis naturalis: regni vegetabilis. Vol 11. Treuttel & Wurtz, Paris. Obermeijer AA 1936 The South African species of Petalidium. Ann Transvaal Mus 18:151–162. Pal M 1964 Chromosome numbers in some Indian angiosperms. I. Proc Indian Acad Sci B 60:347–350. Pathak GN, B Singh, KM Tiwari, AN Srivastava, KK Pande 1949 Chromosome numbers in some angiospermous plants. Curr Sci 18:347. Paton AJ, N Brummit, R Govaerts, K Harman, S Hinchcliffe, B Allkin, EN Lughadha 2008 Towards target 1 of the Global Strategy for Plant Conservation: a working list of all known plant species—progress and prospects. Taxon 57:602–611. Pawar RS, AP Jain, SK Kashaw, AK Singhai 2006 Haematopoietic activity of Asteracantha longifolia on cyclophosphamide-induced bone marrow suppression. Indian J Pharm Sci 68:337–340. Pocock SAJ, G Vasanthy 1988 Cornetipollis reticulate, a new pollen with angiospermid features from Upper Triassic (Carnian) sediments of Arizona (U.S.A.), with notes on Equisetosporites. Rev Palaeobot Palynol 55:337–356. Posada D, KA Crandall 1998 Modeltest: testing the model of DNA substitution. Bioinformatics 14:817–818. Regali MSP, N Uesugui, AS Santos 1974 Palinologia dos sedimentos meso-cenozóicos do Brasil (I). Petrobrás 17:177–191. Richardson A 1972 Revision of Louteridium (Acanthaceae). Tulane Stud Zool Bot 17:63–76. Rosales M 1997 Trichanthera gigantea (Humboldt and Bonpland) Nees: a review. Livest Res Rural Dev 9:1–9. Saggoo MIS, SS Bir 1982 Cytological studies on certain Acanthaceae from central India. Proc Indian Acad Sci B 91:479–486. ——— 1986 Meiotic studies in certain members of family Acanthaceae from South India. J Indian Bot Soc 65:310–315. SanMartin-Gajardo I, M Sazima 2005 Chiropterophily in Sinningieae (Gesneriaceae): Sinningia brasiliensis and Paliavana prasinata are batpollinated, but P. sericiflora is not. not yet? Ann Bot 95:1097–1103. Schmidt-Lebuhn AN 2003 A taxonomic revision of the genus Suessenguthia Merxm. (Acanthaceae). Candollea 58:101–128. Schmidt-Lebuhn AN, M Kessler, J Müller 2005 Evolution of Suessenguthia (Acanthaceae) inferred from morphology, AFLP data, and ITS rDNA sequences. Org Divers Evol 5:1–13. Scotland RW 1991 A systematic analysis of pollen morphology of Acanthaceae genera with contorted corollas. Pollen Spores 44:269–289. ——— 1992 Systematics, similarity, and Acanthaceae pollen morphology. Bot J Linn Soc 109:529–541. ——— 1993 Pollen morphology of Contortae (Acanthaceae). Bot J Linn Soc 111:471–504. Scotland RW, JA Sweere, PA Reeves, RG Olmstead 1995 Higherlevel systematics of Acanthaceae determined by chloroplast DNA sequences. Am J Bot 82:266–275. Scotland RW, K Vollesen 2000 Classification of Acanthaceae. Kew Bull 55:513–589. Sharma BD, Vishnu-Mittre 1963 Contribution to the pollen morphology of the genera Eranthemum Linn. and Pseuderanthemum Radlkof. Proc Natl Inst Sci India B 29:520–526. Shimodaira H, M Hasegawa 1999 Multiple comparisons of loglikelihoods with applications to phylogenetic inference. Mol Biol Evol 16:1114–1116. Sidwell K 1998 A revision of Brillantaisia (Acanthaceae). Bull Nat Hist Mus London Bot 28:67–113. 137 Singh B 1951 Chromosome numbers in some flowering plants. Curr Sci 20:105. Singh RP, CBP Singh, J Kumar 1993 Meiotic studies in Hygrophila serpyllum T. Anders. Proc Indian Sci Congr 80:147. Stebbins GL 1970 Adaptive radiation of reproductive characteristics in angisoperms. 1. Pollination mechanisms. Ann Rev Ecol Syst 1:307–326. Terao H 1983 Taxonomic study of the genus Strobilanthes Bl. (Acanthaceae). Generic delimitation and infrageneric classification. PhD diss. Kyoto University. Thulin M 2007 Synopsis of Satanocrater (Acanthaceae). Nord J Bot 24:385–388. Tripp EA 2007 Evolutionary relationships within the species-rich genus Ruellia (Acanthaceae). Syst Bot 32:628–649. ——— 2010 Taxonomic revision of Ruellia sect. Chiropterophila (Acanthaceae): a lineage of rare and endemic species from Mexico. Syst Bot 35:629–661. Tripp EA, TF Daniel, JC Lendemer, LA McDade 2009 New molecular and morphological insights prompt transfer of Blechum to Ruellia (Acanthaceae). Taxon 58:893–906. Tripp EA, S Fatimah 2012 Comparative anatomy and morphology of the African genus Satanocrater (Acanthaceae). Am J Bot 99:967–982. Tripp EA, S Fatimah, I Darbyshire, LA McDade Forthcoming Origin of African Physacanthus (Acanthaceae) via wide hybridization. PLoS ONE. Tripp EA, PS Manos 2008 Is floral specialization an evolutionary dead-end? pollination system transitions in Ruellia (Acanthaceae). Evolution 62:1712–1737. Tripp EA, LA McDade 2012 New synonymies for Ruellia (Acanthaceae) of Costa Rica and notes on other Neotropical species. Brittonia 64:305–317. Venu P 2007 Strobilanthes Blume (Acanthaceae) in peninsular India. Botanical Survey of India, Kolkata. Verma SC, SS Dhillon 1967 In IOPB chromosome number reports XI. Taxon 16:215–222. Vollesen K 2006 A taxonomic revision of the genus Duosperma (Acanthaceae). Kew Bull 61:289–351. ——— 2008 Acanthacae. Pages 1–286 in HJ Beentje, SA Ghazanfar, eds. Flora of tropical East Africa. Pt 1. Royal Botanic Gardens, Kew. ——— Forthcoming Acanthaceae. In JR Timberlake, ed. Flora Zambesiaca. Vol 8, pt 1. Royal Botanic Gardens, Kew. Wasshausen D 2006 Family 156, Acanthaceae. Pages 1–141 in MJ Jansen-Jacobs, ed. Flora of the Guianas series A. Vol 23. Royal Botanic Gardens, Kew. Wood JRI 1994 Notes relating to the flora of Bhutan: XXIX Acanthaceae, with special reference to Strobilanthes. Edinb J Bot 51:175–273. Wood JRI, RW Scotland 2009 New and little-known species of Strobilanthes (Acanthaceae) from India and South East Asia. Kew Bull 64:3–47. Wood JRI, BP Williams, RW Scotland 2012 Diceratotheca, a new genus of Acanthaceae from Thailand. Kew Bull 67:1–9. Wong KM 1998 Patterns of plant endemism and rarity in Borneo and the Malay Peninsula. Pages 139–169 in CI Peng, PP Lowry II. Rare, threatened, and endangered floras of Asia and the Pacific Rim. Institute of Botany, Academica Sinica Monograph Series 16, Taiwan. Wunderlin RP, BF Hansen 2010 Atlas of Florida vascular plants. Institute for Systematic Botany, University of South Florida, Tampa. http://florida.plantatlas.usf.edu. Zwickl DJ 2006 Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD diss. University of Texas, Austin. This content downloaded on Tue, 29 Jan 2013 06:43:18 AM All use subject to JSTOR Terms and Conditions

RELATED PAPERS

RELATED TOPICS

Log In

Phylogenetic Relationships within Ruellieae (Acanthaceae) and a Revised Classification PHYLOGENETIC RELATIONSHIPS WITHIN RUELLIEAE (ACANTHACEAE) AND A REVISED CLASSIFICATION

Phylogenetic Relationships within Ruellieae (Acanthaceae) and a Revised Classification PHYLOGENETIC RELATIONSHIPS WITHIN RUELLIEAE (ACANTHACEAE) AND A REVISED CLASSIFICATION

Related Papers

RELATED PAPERS

RELATED TOPICS