American Journal of Botany 99(6): 967–982. 2012. COMPARATIVE ANATOMY, MORPHOLOGY, AND MOLECULAR PHYLOGENETICS OF THE AFRICAN GENUS SATANOCRATER (ACANTHACEAE)1 ERIN A. TRIPP2 AND SITI FATIMAH Rancho Santa Ana Botanic Garden, 1500 North College Avenue, Claremont, California 91711 USA • Premise of the study: Anatomical and morphological features of Satanocrater were studied to test hypotheses of xeric adaptations in the genus, which is endemic to arid tropical Africa. These features, together with molecular data, were used to test the phylogenetic placement of Satanocrater within the large plant family Acanthaceae. • Methods: We undertook a comparative study of four species of Satanocrater. Carbon isotope ratios were generated to test a hypothesis of C4 photosynthesis. Molecular data from chloroplast (trnG-trnS, trnG-trnR, psbA-trnH) and nuclear (Eif3E) loci were used to test the placement of Satanocrater within Acanthaceae. • Key results: Anatomical features reflecting xeric adaptations of species of Satanocrater included a thick-walled epidermis, thick cuticle, abundant trichomes and glandular scales, stomata overarched by subsidiary cells, tightly packed mesophyll cells, and well-developed palisade parenchyma on both leaf surfaces. Although two species had enlarged bundle sheath cells, a feature often implicated in C4 photosynthesis, isotope ratios indicated all species of Satanocrater use the C3 pathway. Molecular data resolved Satanocrater within tribe Ruellieae with strong support. Within Ruellieae, our data suggest that pollen morphology of Satanocrater may represent an intermediate stage in a transition series. • Conclusions: Anatomical and morphological features of Satanocrater reflect adaptation to xeric environments and add new information about the biology of xerophytes. Morphological and molecular data place Satanocrater in the tribe Ruellieae with confidence. This study adds to our capacity to test hypotheses of broad evolutionary and ecological interest in a diverse and important family of flowering plants. Key words: arid; C4 photosynthesis; carbon isotope ratio; Eif3E; herbarium specimen; molecular phylogeny; Ruellieae; Satanocrater; xeromorphic; xerophytic. With an estimated 4000+ species, Acanthaceae ranks among the 12 or so most diverse families of flowering plants. Over the last two decades, tremendous progress has been made in reconstructing the evolutionary history of this species-rich, geographically widespread, and ecologically important family (McDade, 1990, 1992; Scotland et al., 1995; Hedrén et al., 1995; McDade and Moody, 1999; McDade et al., 2000, 2005, 2008; Carine and Scotland, 2000, 2002; Manktelow et al., 2001; Moylan et al., 2004; Schmidt-Lebuhn et al., 2005; Kiel et al., 2006; Tripp, 2007; Borg et al., 2008; Tripp and Manos, 2008; Daniel et al., 2008; Tripp et al., 2009). Beyond contributions to basic knowledge of relationships within Acanthaceae, we are close to achieving an understanding of the family as a whole that will permit exploration of exciting evolutionary questions that require intensive taxon sampling in species-rich lineages. For example, because of the size, biogeographic range, and diverse ecologies in the family, Acanthaceae are ideal for exploring disparities in morphological, taxonomic, and phylogenetic diversity. Previous studies in Acanthaceae have had broad implications for systematic and evolutionary biology (e.g., McDade, 1990, 1992; Carine and Scotland, 2002; Tripp and Manos, 2008). Indeed, these and future comparative analyses are made possible by thorough sampling of all major lineages in a family. While most of these major lineages that comprise Acanthaceae s.l. (Fig. 1) have been targeted for phylogenetic study (e.g., Justiceae: McDade et al., 2000; Acantheae: McDade et al., 2005; Barlerieae: McDade et al., 2008; and Andrographideae: McDade et al., 2008), there have been no similar efforts to comprehensively test the phylogenetic placement of, and explore evolutionary relatedness among, the 48 genera (>1200 species) proposed to comprise the tribe Ruellieae (= Ruelliinae sensu Scotland and Vollesen, 2000). Morphological synapomorphies that have been hypothesized to unite all genera in Ruellieae include left-contort corolla aestivation, seeds with mucilaginous hygroscopic trichomes, and presence of a “filament curtain” (Grubert, 1974; Scotland et al., 1994; Scotland et al., 1995; Manktelow, 2000; Manktelow et al., 2001; Tripp, 2007). Of the 48 genera classified in Ruellieae, 25 occur in Africa, of which 20 are restricted to that continent (Tripp et al., unpublished manuscript). Thus, African Ruellieae represent a considerable (although understudied) fraction of the generic diversity in the tribe. African Ruellieae are of further interest because several are bird-pollinated, and bird pollination has been implicated in explaining diversification discrepancies among major 1 Manuscript received 19 July 2011; revision accepted 6 April 2012. The authors thank Ensermu Kelbessa (Ethiopia National Herbarium) for field assistance; Kaj Vollesen and Parmjit Bhandol (Kew Herbarium) for contributing pollen images to this study; Mats Thulin (Uppsala University) for contributing an image of living S. somalensis; Iain Darbyshire (Kew Herbarium) for identification of Kew specimens and assistance with loan preparation; curatorial staff at C, CAS, K, MO, and US for granting us access to study their collections; and Travis Columbus and Jeff Morawetz (Rancho Santa Ana Botanic Garden), John Freudenstein (Ohio State University), and two anonymous reviewers for suggestions that improved this manuscript. Funding for this project was provided by a U. S. National Science Foundation grant to E. A.Tripp. and L. A. McDade (DEB-0919594). 2 Author for correspondence (e-mail: etripp@rsabg.org) doi:10.3732/ajb.1100354 American Journal of Botany 99(6): 967–982, 2012; http://www.amjbot.org/ © 2012 Botanical Society of America 967 968 [Vol. 99 AMERICAN JOURNAL OF BOTANY Fig. 1. Summary phylogeny of current understanding of relationships within Acanthaceae, based on McDade et al. (2008). All branches are strongly supported. Acanthaceae s.s. refers to plants with retinacula. Within Acanthaceae s.s., all plants have cystoliths except those in tribe Acantheae. Previous classifications have treated Satanocrater in Ruellieae. geographic centers of Acanthaceae (Schmidt-Lebuhn et al., 2007; McDade et al., 2008b). One such African genus with bird pollination and one hypothesized to belong to Ruellieae (Scotland and Vollesen, 2000) is Satanocrater. Satanocrater Schweinf. (σατανος-, -κρατορας; devil’s bowl) includes four species that are distributed in tropical East Africa (Ethiopia, Kenya, and Somalia), with one species disjunct in tropical West Africa (Guinea; Fig. 2; Thulin, 2007). Species of Satanocrater grow characteristically on rocky limestone surfaces to alluvial red or black sandy soils in hot, high sunlight tree savannas and semiarid deserts, to open bushlands or shaded woodlands dominated by Acacia and Commiphora. Species of Satanocrater are attractive, large-flowered shrubs or perennial herbs with woody bases. The four species can be broadly grouped into two categories based on floral morphology (Fig. 3): (1) Satanocrater fellatensis Schweinf. and Satanocrater ruspolii (Lindau) Lindau have mauve to purple flowers with infundibular corollas suggestive of insect pollination; (2) Satanocrater paradoxus (Lindau) Lindau and Satanocrater somalensis (Lindau) Lindau have bright orange to red flowers that are strongly bent infundibular (for corolla shape terminology, see fig. 2C vs. D–H in Tripp, 2010), and both species are bird-pollinated (M. Thulin, Uppsala University, personal observations; E. A. Tripp, personal observations). Corollas of S. paradoxus are further modified in having an extremely reduced, ventralmost corolla lobe (hence the specific epithet). Perhaps the most salient morphological feature shared by all four species of Satanocrater is their remarkable, large, inflated calyces with lobes fused almost to the apex at anthesis (Fig. 3), but splitting into five segments postanthesis and when fruits mature. No other genus in the family possesses such calyces, with the exception of the genus Physacanthus Benth., which we know to be nested within Acanthaceae but has been variously classified into different lineages within the family (Tripp et al., unpublished manuscript). Fig. 2. Geographical distributions of Satanocrater fellatensis (䊏), S. ruspolii (夹), S. somalensis (•), and S. paradoxus (䉱) in Africa. Flowers and leaves, when crushed, give an aromatic and strong, spicy fragrance reminiscent of rosemary and thyme (personal observations: E. Tripp and S. Fatimah, study of herbarium material; E. Tripp, field study of S. paradoxus and S. ruspolii). This is unusual because plant fragrances, both floral and nonfloral, are thought to be rare within Acanthaceae. Herbarium labels also indicate that plants are used ethnobotanically, although specific information regarding traditional medicinal uses other than foliage used to spice teas is not elaborated upon. Native names of species (based on herbarium labels) include Elas (S. ruspolii); Alash, Alari, and Alah riger (S. somalensis); and Malabow (S. paradoxus). The overarching goal of the current study is to document anatomical, morphological, and molecular features of Satanocrater to permit our testing of two primary hypotheses: (1) that anatomy and morphology of all four species should reflect adaptation to the xeric environments in which plants occur and (2) that phylogeny reflects previous classification in Acanthaceae (Scotland and Vollesen, 2000), i.e., that Satanocrater belongs in Ruellieae. This study represents the first nontaxonomic investigation of the genus. There is a complex history of terminology applied to the anatomy and morphology of plants that occupy dry and/or hot environments (e.g., xerophytic vs. xeromorphic vs. xeroplastic; Thoday, 1933; Fahn and Cutler, 1992). In this study, we simply use the term xerophytic and its derivates (e.g., xerophyte) to refer to plants that occur on dry soils, live in dry atmospheres, are exposed to high levels of solar radiation, and have morphological or anatomical adaptations that reflect this mode of living (Wiesner, 1889). MATERIALS AND METHODS Specimen sampling—To test our two primary hypotheses, we studied 60 herbarium specimens (including three unidentifiable specimens due to June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER 969 Fig. 3. Flowers of the four species of Satanocrater. (A) Satanocrater paradoxus has orange corollas; photograph (by E. A. Tripp) shows the inflated calyx that characterizes the genus and the reduced fifth corolla lobe that characterizes this species. (B) Satanocrater somalensis has orange-red corollas; photograph (by M. Thulin) shows the inflated calyx. (C) Satanocrater ruspolii has purple corollas; inflated calyx present but obscured on this herbarium specimen (Burger 3330, US). (D) Satanocrater fellatensis has purple corollas; herbarium specimen (Friis et al., 6884, C) shows inflated calyx. inadequate material), representing multiple accessions of all four species of Satanocrater. Specimens were identified using Thulin’s (2007) dichotomous key. In addition, we obtained unpublished pollen images (by K. Vollesen and P. Bhandol) for three specimens whose identities were confirmed by our colleagues I. Darbyshire and K. Vollesen at Royal Botanic Gardens, Kew. All specimens examined (including those at Kew) are listed in Appendix 1. Anatomy—Although fresh material is preferred to herbarium material for anatomical study, there is no living material of Satanocrater available in cultivation to our knowledge. As such, we used herbarium material to study S. fellatensis and S. somalensis. Fieldwork by the first author made possible the collection of fresh material for S. paradoxus and S. ruspolii. Field-collected materials were preserved in FPA (a 1 : 1 : 18 ratio of 37% formaldehyde–propionic acid–70% ethanol). Mature leaves of S. fellatensis and S. somalensis were removed from herbarium specimens and softened in 1.5% Contrad 70 Solution for a minimum of 1 d, depending on leaf thickness (Schmid and Turner, 1977). After softening, leaves were rinsed thoroughly with deionized water to neutralize residual Contrad and then placed directly in FPA for a minimum of 2 d. All leaves, including those of S. paradoxus and S. ruspolii, were then dehydrated through a 30, 50, 70, 90, and 95% ethanol series for a minimum of 2 h per step. Between the 50% and the 70% ethanol series steps, mid portions of leaf samples were cut transversely for later embedding and transverse sectioning. Samples were transferred to 100% ethanol with 1% safranin (w/v) overnight, then placed in 100% ethanol (2 h). Leaf samples were then soaked in a 2 : 1 ratio of 100% ethanol to xylene (2 h), a 1 : 2 ratio of 100% ethanol to xylene (2 h), 100% xylene (2 h), 100% xylene (2 h), a 2 : 1 ratio of xylene to paraffin oil (2 h), a 1 : 2 ratio of xylene to paraffin oil (2 h), and then infiltrated in two 6-h changes of liquid paraffin. After infiltration, materials were embedded with paraffin using a Leica Histoembedder. Following the method of Columbus (1999), 10-µm leaf sections were prepared using an American Optical Co. Spencer Model 820 rotary microtome. Sections were mounted on slides, coverslipped, then stained following Ocampo and Columbus (2010), which is based on Sharman (1943). The slides are deposited at Rancho Santa Ana Botanic Garden and Royal Botanic Gardens, Kew. Slides were examined with Leitz Laborlux D and Leitz Wetzlar light microscopes and photographed using SPOT software version 4.1.1. Morphology—Leaf, seed, and pollen micromorphology of herbarium material was studied using scanning electron microscopy (SEM). Before SEM, dry seeds were photographed using an SZH-ILLD Olympus stereoscope. To examine seed surface vestiture, namely, the presence or absence of mucilaginous hygroscopic trichomes known in other Ruellieae, seeds were soaked in water and rephotographed with light microscopy. Mature leaves (both surfaces), seeds (either dried or redried after wetting), and pollen grains were sputter coated with gold using a PELCO SC-7 Auto Sputter Coater (Redding, California, USA). Structures were subsequently examined using an International Scientific Instruments DS-130 / WB-6 Scanning Electron Microscope. Light microscopy was also used to examine corolla aestivation and to search for the 970 [Vol. 99 AMERICAN JOURNAL OF BOTANY presence of a “filament curtain” (Manktelow, 2000), a structure formed by the adnation of filaments to the corolla wall that creates a partition and divides the corolla tube longitudinally into two compartments. See Manktelow’s (2000) monograph of the filament curtain for a detailed description of the four major types (phaulopsoid, corolla fold, reduced, and strobilanthoid). Carbon isotope ratios—Because preliminary anatomical data suggested that two of the four species of Satanocrater have enlarged bundle sheath cells, one feature that is part of a character suite associated with C4 photosynthesis, we prepared two samples per each of the four species (N = 8 in total) for 13C/12C carbon isotope ratio analysis. Approximately 4 mg of dried leaf tissue was pulverized in an automated tissue grinder and placed inside tin capsules (Costech Analytical Technologies, Valencia, California, USA). Samples were submitted to the Colorado Plateau Stable Isotope Laboratory (Flagstaff, Arizona, USA), which uses a Thermo Electron (Waltham, Massachusetts, USA) gas isotoperatio mass spectrometer. Isotope analyses were subjected to rigorous machine calibration checks, and several known standards (in addition to the carbon standard used to calculate δ13C values; see below) were run concurrently with analyzed samples. Carbon isotope abundances are expressed as the ratio of 13C/12C in a sample, which is compared to a known, international standard. The deviation of the sample’s ratio from that of the standard is expressed in parts per thousand, or ‰. The international carbon standard used in our study was Vienna Pee Dee Belemnite, which has an accepted absolute 13C/12C ratio of 0.0112372. The δ13C value of the sample is calculated as [13C/12C sample)/ (13C/12C standard) – 1] × 1000. We compared δ13C ratios derived from this study to the typical ranges of C3 (> −22‰ δ13C) and C4 (−10 to −16‰ δ13C) photosynthesis plants (O’Leary 1988). Molecular phylogenetics—To test the hypothesis that Satanocrater belongs to the tribe Ruellieae within Acanthaceae, we generated DNA sequence data from four markers: one nuclear (Eif3E, used here for the first time in Acanthaceae) and three chloroplast (trnG-trnS, trnG-trnR, psbA-trnH). To ensure that representatives of all known lineages of Acanthaceae were sampled, we used results from previously published phylogenies (McDade et al., 2000, 2008; Tripp, 2007) to guide taxon sampling (see Fig. 1 and Appendix 1). This broad sampling ensures our ability to place Satanocrater in a phylogenetic context with confidence. Not including samples of Satanocrater, a total of 22 accessions were used in the molecular study (Fig. 1): one outgroup (Sesamum), one Nelsonioideae (Staurogyne), one from the Avicennia lineage (Avicennia), one Thunbergioideae (Mendoncia), two Acantheae (Aphelandra, Stenandriopsis), two Barlerieae (Barleria, Crabbea), one Andrographideae (Cystacanthus), two Whitfieldieae (Camarotea, Chlamydacanthus), one from the Neuracanthus lineage (Neuracanthus), two Justicieae (Mackaya, Rhinacanthus), and eight Ruellieae (Acanthopale, Brillantaisia, Duosperma, Hygrophila, Mellera, Petalidium, Phaulopsis, Ruellia). All sampled genera of Ruellieae are largely (or entirely) African in distribution except Ruellia, which is most diverse in the neotropics. Nine accessions of Satanocrater were employed in this study: one S. fellatensis, four S. paradoxus, one S. ruspolii, and three S. somalensis. Individual data matrices included a total of up to 31 accessions (nine Satanocrater plus 22 other taxa). A combined data matrix included all 31 accessions, with data scored as missing if no sequence was successfully generated for a given locus. Total genomic DNA was extracted from leaf tissue sampled from herbarium specimens or from silica dried leaf material using either the CTAB method (Doyle and Doyle, 1987) with modifications or DNeasy Plant Mini Kits (Qiagen, Valencia, California, USA). Because DNA extracted from Satanocrater was particularly difficult to amplify, we did a secondary phenol–chloroform precipitation of the DNA to remove excess proteins and polysaccharides, following a modified nucleic acids isolation protocol (Chomczynski and Sacchi, 2006). Primer and PCR reactions conditions followed those described in previous publications for trnG-trnS (Tripp and Manos, 2008), trnG-trnR (Tripp, 2007), and psbA-trnH (Tripp, 2010). Degenerate primers used to generate data from nuclear Eif3E, a translation initiation factor, were those listed in Li et al. (2008). Most sequences of trnG-trnS, trnG-trnR, and psbA-trnH were generated previously for other studies except sequences for Satanocrater, which were newly generated for this study. All sequences of Eif3E were newly generated for this study. Because preliminary data indicated that there exist at least two copies of Eif3E in Acanthaceae, we cloned several taxa to construct a data matrix with only one of the copies. For these samples, PCR products were cleaned via polyethylene glycol precipitation, and TOPO TA Kits (Invitrogen Corp., Carlsbad, California, USA) were used for the ligation and recombination steps. Individual colonies were picked for subsequent “colony PCR”, which was followed by direct sequencing of these colony amplifications. New data reported in this study were generated on a 3130X Automated Genetic Analyzer (Carlsbad, California, USA) at Rancho Santa Ana Botanic Garden. Bidirectional sequences were assembled using the program Sequencher v.4.9 (Gene Codes Corp., Ann Arbor, Michigan, USA). All sequences are deposited in GenBank (Appendix 1). Preliminary analyses demonstrated that phylogenetic results were robust to different alignment methods (data not shown); sequences were manually aligned using the program MacClade v.4.06 (Maddison and Maddison, 2003) prior to final analysis (alignments available in TreeBASE, S12451). Data derived from psbA-trnH and Eif3E were extremely divergent across all Acanthaceae. Thus, phylogenetic analyses for these two markers were conducted on matrices pruned to contain only Satanocrater and members of Ruellieae. All taxa (i.e., including other Acanthaceae and the outgroup) were analyzed for trnG-trnS and trnG-trnR. Because data sets were not parallel in terms of taxon sampling, particularly with respect to presence or absence of accessions of Satanocrater (owing to difficulty of amplification) and whether to include (trnGtrnS, trnG-trnR) or exclude (psbA-trnH, Eif3E) genera outside of Ruellieae, we analyzed matrices both individually and in a combined framework. For individual matrix analyses, Sesamum was used to root trnG-trnS and trnG-trnR trees (based on McDade et al., 2008), and Brillantaisia + Hygrophila were used to root psbA-trnH and Eif3E trees (based in E. A. Tripp et al., unpublished data). In the combined analysis, Sesamum was used to root trees. Phylogenetic analyses were conducted using parsimony and maximum likelihood algorithms. Ambiguous alignment sites were excluded from analyses. Parsimony searches were conducted in the program PAUP* v.4.0b10 (Swofford, 2002) and included 500 random addition sequence replicates with tree-bisection-reconnection (TBR) branch-swapping, gaps treated as missing data, and the “amb-” option in effect for trnG-trnS, trnG-trnR, and the combined analysis. A semistrict consensus topology was constructed for all resulting most parsimonious trees. Maximum likelihood analyses were conducted in the program GARLI v.0.951 (Zwickl, 2006) using default settings, except empirical base frequencies were used. Branch support was assessed with a full heuristic parsimony bootstrap analysis in PAUP* using 1000 replicates, and 100 likelihood bootstrap replicates in GARLI. For the combined matrix parsimony bootstrap, no more than 1000 trees greater than the length of the shortest MP trees (see results) were saved per replicate. RESULTS Anatomy—Study of leaf anatomy and morphology indicated that all four species of Satanocrater possess features typical of plant adaptation to dry, hot environments. Some of these adaptations are seen in all four species (e.g., trichomes and/or peltate glandular scales, a thick epidermis and cuticle, tightly packed mesophyll cells), while others (e.g., stomata overarched by subsidiary cells, unifacial leaves, enlarged bundle sheath cells) are restricted to subsets of species. The selection of figures presented in this study reflects those that we feel best depict the majority of the features under discussion. Additional images, some of which depict some anatomical structures in better detail or under higher magnification, are available upon request from the first author. Indumentum—Two types of structures on leaf indumenta, (1) trichomes and (2) peltate glandular scales (PGS), were densely covering adaxial and abaxial surfaces of all four species of Satanocrater (Figs. 4–7). Transverse sections and SEM images demonstrate that S. fellatensis (Fig. 4) and S. paradoxus (Fig. 7) were covered by both types of structures, whereas S. ruspolii (Fig. 5) and S. somalensis (Fig. 6) were covered only by PGS. Transverse leaf sections show uniseriate trichomes consisting of two to eight cells. SEM and light microscope study indicate that these trichomes are nonglandular, although anatomical sections suggest that the distalmost cell is often enlarged with respect to more proximal cells (Fig. 7A; Satanocrater paradoxus). Transverse sections show that each PGS consists of (1) two basal cells, which are part of the epidermis, (2) one to two stalk cells (immediately superficial to the epidermis; these labeled in Fig. 5B) June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER 971 Fig. 4. Light (LM) and scanning electron micrographs (SEM) of Satanocrater fellatensis. (A, B) LM of transverse sections (TS) of leaf showing (A) various anatomical features, 125×; and (B) 10 secretory cells of PGS, 500×. (C) LM of leaf surface showing secretory materials of PGS, 64×. (D–F) SEM images of (D) leaf surface showing stomate overarched by subsidiary cells; (E) style showing PGS; (F) leaf surface showing trichomes and PGS. (A, B) from Friis et al., 10721, K. (C, D, F) from Friis et al., 6884, K, C. (E) from Adam 13815, MO. cy, cystolith; ep, epidermis; gc, guard cell; pgs, peltate glandular scale; pp, palisade parenchyma; sb, subsidiary cell; sc, secretory cells; sm, secretory material; tr, trichome. depending on stage of development, and (3) the head, which is composed of secretory cells that degenerate after secretion (Fig. 5B; S. ruspolii). Secretory head cells in Satanocrater can be seen in various stages of development (e.g., Fig. 4B vs. Fig. 5B). The secretory material is also observable on leaf surfaces of all four species through light microscopy (Figs. 4C, 5C, 6B, 7B). In addition to covering vegetative surfaces, PGS were found on reproductive structures including anthers (Fig. 5G), styles (Figs. 4E, 5H), and seeds (Fig. 9D, 9E). PGS across these surfaces were morphologically indistinguishable from those on leaf surfaces. Fig. 5. Light (LM) and scanning electron micrographs (SEM) of Satanocrater ruspolii. (A, B) LM of transverse sections (TS) of leaf showing (A) various anatomical features, 125×; and (B) one to two basal cells of PGS immediately surficial to the epidermis, 500×. (C) LM of leaf surface showing secretory materials of PGS, 64×. (D–F) SEM of leaf surface showing (D) subsidiary cells overarching a stomata; and (E, F) PGS. (G, H) SEM of (G) dorsal anther surface showing PGS, and (H) style showing PGS. (A, B, E) from Friis et al., 9894, C. (C, D, F) from Burger 3327, US. (G, H) from de Wilde, 6665, MO. bc, basal cell; cy, cystolith; ep, epidermis; pgs, peltate glandular scale; pp, palisade parenchyma; sb, subsidiary cell; sm, secretory material. Epidermis—In transverse section, all four species of Satanocrater were observed to have enlarged epidermal cells that are covered by a thick cuticle (Figs. 4A, 5A, 6A, 7A). the plane of the subsidiary cells (Figs. 4D, 5D, 6C); it is less clear that stomata reside on a different plane from subsidiary cells in S. paradoxus (Fig. 7C). All four species were frequently observed to have diacytic stomata in which the adjacent subsidiary cells are oriented perpendicular to guard cells (Figs. 7A & 8A). Stomata—SEM study indicated that at least three of the four species of Satanocrater, i.e., S. fellatensis, S. ruspolii, and S. somalensis, have stomata that are overarched by adjacent, subsidiary cells. In these species, guard cells can be seen interior to Parenchyma tissue—Three species of Satanocrater, S. fellatensis, S. ruspolii, and S. somalensis, have well-developed, elongated palisade cells on both adaxial and abaxial leaf surfaces (Figs. 4A, 5A, 6A), with each surface consisting of two to 972 AMERICAN JOURNAL OF BOTANY Fig. 6. Light (LM) and scanning electron micrographs (SEM) of Satanocrater somalensis. (A) LM of transverse section (TS) showing various anatomical features, 125×. (B) LM of leaf surface showing secretory materials of the PGS, 64×. (C–E) SEM of (C) open stomate overarched by subsidiary cells and (D, E) leaf surface showing PGS. (A) from Lavranos et al., 24660, MO. (B, D, E) from Bally and Melville 15679, K. (C) from Nuget 13, K. cy, cystolith; ep, epidermis; gc, guard cell; pgs, peltate glandular scale; pp, palisade parenchyma; sb, subsidiary cell. three cell layers. In contrast, abaxial leaf surfaces of S. paradoxus are differentiable from adaxial surfaces, with the former consisting of shorter, sometimes rounded parenchyma cells (Fig. 7A). In all four species, air space between mesophyll cells is extremely limited. Cystoliths—Anatomical sections of the four species of Satanocrater clearly show the presence of cystoliths—calcium carbonate or calcium oxalate crystals (Metcalfe and Chalk, 1950). These cystoliths are elongated and are enclosed in specialized cells, termed lithocysts, and their presence confirms the placement of Satanocrater within the cystolith clade of Acanthaceae (Fig. 1). In Satanocrater, cystoliths are concentrated primarily in two locations within the leaf: among the upper palisade cells (Figs. 6A, 7A) and surrounding the central vascular bundle (Figs. 4A, 5A, 6A). Cystoliths distributed among the adaxial palisade cells are generally oriented perpendicular to the plane of the leaf (e.g., Fig. 5A), whereas those surrounding the midvein are generally oriented parallel to it (e.g., Fig. 6A). [Vol. 99 Fig. 7. Light (LM) and scanning electron micrographs (SEM) of leaves of Satanocrater paradoxus. (A) LM of transverse sections (TS) showing various anatomical features, 125×. (B) LM of surface showing secretory materials of PGS, 64×. (C–E) SEM of (C) stomata and subsidiary cells and (D, E) SEM of surface showing trichomes and PGS. (A) from de Wilde 6657, MO. (B) from Friis et al., 3147, MO. (C– E) from Friis et al., 3147, C. cy, cystolith; ep, epidermis; gc, guard cell; pgs, peltate glandular scale; pp, palisade parenchyma; sb, subsidiary cell; tr, trichome. Bundle sheaths—Secondary vascular bundles in the leaves of S. paradoxus are surrounded with enlarged bundle sheath cells (Fig. 8A), one of several features that in other plants is associated with C4 photosynthesis. Enlarged bundle sheath cells are also observable, albeit in poorer quality sections, in leaves of S. somalensis (Fig. 8B). Morphology— Features pertaining to corolla aestivation, filament curtain structure, seed morphology, and pollen structure of Satanocrater all provide support for the placement of this genus in Ruellieae. However, most features do not differ qualitatively among species of Satanocrater. Corolla aestivation—Our examination of herbarium specimens indicates that Satanocrater has left-contort aestivation, which is consistent with all other genera in Ruellieae studied to date. Other tribes in Acanthaceae s.s. are characterized by different aestivation patterns (but see Discussion for two exceptions). June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER 973 Fig. 9. (A–F) Light and (G–I) scanning electron micrographs of (A, B, D, G–I) dry seeds, (C, F) wetted seeds and (E) dry seeds after having been wetted. (A–C) Satanocrater fellatensis from Berhauti 4158, K. (D–F) Satanocrater ruspolii from Kuchar 16903, K. (G) Satanocrater ruspolii from de Wilde 6665, MO. (H, I) Satanocrater ruspolii from Tripp & Ensermu 904, RSA. ht, helical thickenings; mt, mucilaginous trichomes; pgs, peltate glandular scale. (A, D) 20× and 25× (B, E) 64× (C, F) 20×. Fig. 8. Light micrographs of transverse sections of the leaves of Satanocrater paradoxus and Satanocrater somlensis showing enlarged bundle sheath cells. (A) S. paradoxus from de Wilde 6657, MO, 125×. (B) S. somalensis from Lavranos et al., 24660, MO, 125×. bs, bundle sheath; gc, guard cell; sb, subsidiary cell. Filament curtain—Our examination of corollas of Satanocrater indicates that all four species possess a filament curtain, consistent with all other genera in Ruellieae studied to date (Manktelow, 2000). Satanocrater paradoxus and S. somalensis have a reduced type of filament curtain (the former also documented in Manktelow, 2000), while S. ruspolii has a phaulopsoid type. Filament curtains in the fourth species, S. fellatensis, were intermediate between the reduced and phaulopsoid types. Seeds—Because the vast majority of herbarium sheets we examined lacked fruit and seed material, we were able to study seed morphology in only two of the four species. Seeds from S. fellatensis and S. ruspolii were similar in shape. Like all other Ruellieae studied to date, seeds of both species were either sparsely or densely covered by hygroscopic trichomes (Fig. 9). These trichomes, when wetted, formed a conspicuous, milky mucilaginous cloud (Fig. 9C, 9F). These hygroscopic trichomes were observed to have helical thickenings (Fig. 9G–9I), which uncoil when wetted. In a dry state, the seed indumentum appears as tightly packed golden-brown trichomes (Fig. 9A, 9D). Pollen—The pollen of all four species of Satanocrater is weakly colporate to porate, three aperturate, and spheroidal in shape (Figs. 10–13). A distinct albeit weakly developed colpus associated with each pore can be seen in Figs. 12C, 13D; other views show a porate aperture, but no obvious colpus (Figs. 10B, 11B). Exine sculpturing is coarsely reticulate and of the “Wabenpollen” sort from Lindau’s (1893, 1895) treatment of pollen morphology in Acanthaceae. Photosynthetic pathways— Leaves from all four species of Satanocrater produced δ13C values ranging between −25‰ and −28‰ δ13C, which is consistent with C3 photosynthesis. The two δ13C values generated per species were generally close in value: S. fellatensis (−28.21‰, −28.64‰), S. paradoxus (−27.19‰, −27.76‰), S. ruspolii (−25.79‰, −26.73‰), S. somalensis (−25.77‰, −25.86‰). Molecular phylogenetics— The final alignments for molecular analyses consisted of 861 characters for trnG-trnS (193 parsimony-informative), 920 characters for trnG-trnR (154 parsimony-informative), 593 characters for psbA-trnH (84 parsimony-informative), 935 characters for Eif3E (120 parsimonyinformative), and 3309 characters (551 parsimony-informative) for the combined analysis. Parsimony searches resulted in 15 most parsimonious (MP) trees of length 608 for trnG-trnS (CI = 0.77), 2 MP trees of length 498 for trnG-trnR (CI = 0.77), 17 MP trees of length 190 for psbA-trnH (CI = 0.79), 1 MP tree of length 440 for Eif3E (CI = 0.76), and 185 MP trees of length 1744 for the combined analysis (CI = 0.82). All phylogenetic analyses under both parsimony and likelihood criteria resolved accessions of Satanocrater as monophyletic and strongly supported as nested within Ruellieae (100% parsimony bootstrap [PBS], 100% likelihood bootstrap [LBS]; Figs. 14–16). Within Ruellieae, Satanocrater was consistently resolved as closely related to two genera (strongly supported by three of the 974 AMERICAN JOURNAL OF BOTANY [Vol. 99 Fig. 10. Scanning electron micrographs of polar and equatorial views of pollen of Satanocrater fellatensis from Adam 13815, MO. Fig. 12. Scanning electron micrographs of equatorial views of pollen of Satanocrater somalensis. (A) From Bally and Melville 15679, K. (B, C) From Gillett 22329, K. (D) from Gillett 22998, K. (B–D) Images taken by K. Vollesen and P. Bhandol, Royal Botanic Gardens, Kew. four markers), Ruellia and Acanthopale, which have been shown to be close relatives in an earlier study (Tripp, 2007). Within Satanocrater, analyses of the three chloroplast markers provided little to no resolution among accessions; branches that were resolved were, in general, extremely short and received no bootstrap support (Figs. 14, 15). The one exception is the likelihood analysis of trnG-trnS data, which provided weak support (75% LBS) for a sister group relationship between two accessions of S. somalensis that were successively sister to two accessions of S. paradoxus (the first with 75% LBS, the second not supported); sister to all of these was S. felletensis. We were unable to obtain sequence data for S. ruspolii for trnG-trnS. Parsimony and likelihood analyses of the nuclear Eif3E data set (Figs. 14, 15) resolved a strongly supported clade (100% PBS, 100% LBS) of two accessions of S. somalensis, which in turn was strongly supported as sister to one accession of S. paradoxus (99% PBS, 97% LBS). Sister to S. paradoxus + S. somalensis was one accession of S. ruspolii. We were unable to Fig. 11. Scanning electron micrographs of polar and equatorial views of pollen of Satanocrater ruspolii. (A, B) From Friis et al., 9894, C. (C, D) Images by K. Vollesen and P. Bhandol, Royal Botanic Gardens, Kew, from Kuchar 17604, K. Fig. 13. Scanning electron micrographs of equatorial views of pollen of Satanocrater paradoxus. (A, B) From Friis et al., 3147, C. (C, D) Images taken by K. Vollesen and P. Bhandol, Royal Botanic Gardens, Kew, from Friis et al., 1006, K. June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER 975 Fig. 14. Semistrict consensus topologies of most parsimonious tree(s) from analyses of trnG-trnS, trnG-trnR, psbA-trnH, and Eif3E data. Parsimony bootstrap values above branches, likelihood bootstrap values below (branches thickened if either value ≥ 70%). Gray boxes mark the tribe Ruellieae. Labeled on trnG-trnS cladogram are (1) the major lineages that comprise Acanthaceae (see Fig. 1) and (2) three pollen types found in Ruellieae, based on present taxon sampling. obtain sequence data for S. fellatensis for Eif3E. Analyses of combined data produced qualitatively similar results: a strongly supported clade of Satanocrater (that is strongly supported in Ruellieae), but poor (Fig. 16A) to no (Fig. 16B) resolution among species within Satanocrater. There was weak bootstrap support (78% LBS) for polyphyly of S. paradoxus in the maximum likelihood analysis. In sum, two markers (trnG-trnS and Eif3E) provided some support (weak in trnG-trnS, strong in Eif3E) for the two redflowered, bird-pollinated taxa (S. paradoxus and S. somalensis) being more closely related to one another than either is to the purple-flowered taxa (i.e., to S. fellatensis with trnG-trnS or to S. ruspolii with Eif3E); neither the trnG-trnS nor the Eif3E matrices contained both of the purple-flowered taxa, preventing our assessment of relationships between them. DISCUSSION In the present investigation, we studied anatomical, morphological, and molecular features of Satanocrater to test 976 AMERICAN JOURNAL OF BOTANY [Vol. 99 Fig. 15. Most likely tree derived from GARLI analyses of trnG-trnS, trnG-trnR, psbA-trnH, and Eif3E data. Parsimony bootstrap values above branches, likelihood bootstrap values below (branches thickened if either value is ≥ 70%). Gray boxes mark the tribe Ruellieae. Labeled on trnG-trnS phylogram are (1) the major lineages that comprise Acanthaceae (see Fig. 1) and (2) three pollen types found in Ruellieae, based on present taxon sampling. two primary hypotheses: one ecological and one evolutionary. First, results demonstrate that all species of Satanocrater possess traits indicative of adaptation to xeric environments. Two of the four species (S. paradoxus and S. somalensis) possess enlarged bundle sheath cells similar to those of plants with C4 photosyntheis, but carbon isotope ratios indicate that all four species use the C3 photosynthetic pathway. In this study, we obtained valuable anatomical information about two species for which we had access only to herbarium collections (S. fellatensis and S. somalensis). Our results reconfirm the value of museum holdings when fresh, field-collected material is unavailable. Second, morphological and molecular data confirm the phylogenetic placement of Satanocrater within the tribe Ruellieae of Acanthaceae. The current study contributes new knowledge to the evolution of Acanthaceae, to poorly known acanths with Old World bird pollination, and to xerophyte biology. To our knowledge, Satanocrater has never been studied from an ecological or evolutionary perspective. Indeed, a major hindrance to the present investigation, especially the molecular study, was that none of the species is available in cultivation. Moreover, collections of the genus are very few across major herbaria, particularly for S. fellatensis and S. somalensis. Although materials available to us were sufficient for the purposes June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER 977 Fig. 16. (A) Most likely tree and (B) semistrict consensus of most parsimonious trees of analysis of combined data. Parsimony bootstrap values above branches, likelihood bootstrap values below (branches thickened if either value ≥ 70%). Gray boxes mark the tribe Ruellieae. Labeled on most likely tree are (1) the major lineages that comprise Acanthaceae (see Fig. 1) and (2) three pollen types found in Ruellieae, based on present taxon sampling. of testing our two primary hypotheses, recalcitrant DNA derived from herbarium accessions precluded generation of data from additional nuclear markers with potentially greater molecular variation. Future study of Satanocrater would benefit greatly from access to fresh plant material. Heat and xeric adaptations—Over the past century, several significant works have contributed to a clearer understanding of xerophyte biology (Delf, 1915; Maximov, 1931; Thoday, 1933; Shields, 1950, 1951). These studies emphasized plant structural modifications that are associated with reducing water loss and/or increasing total photosynthetic rate and efficiency in organisms that inhabit hot and dry environments. Fahn and Cutler (1992) explored distributions of these modifications across 71 seed plants and, given that not all modifications occur together in a particular plant, found that some were more common than others. For example, thickened cuticles and a thickened epidermis were present in over two thirds of the 71 taxa, whereas unifacial leaves (palisade parenchyma on both adaxial and abaxial surfaces) and sunken stomata were present in only a third of the taxa. Data indicating that habitat and anatomy are strongly correlated in xerophytes continue to accumulate in recent years (e.g., Burrows, 2001). The present study adds to such knowledge. Epidermis and cuticle—It has long been recognized that the leaf epidermis, together with waxy cuticles of varying thicknesses, serve a vital role in regulating water loss as well as protecting tissues from excessive sunlight and other biological or nonbiological intruders (Martin and Juniper, 1970; Mauseth, 1988). In particular, excessive sunlight (especially ultraviolet and IR wavelengths) can damage plant tissues by overheating cell cytoplasm and bleaching chlorophyll. An epidermis helps to buffer underlying tissue from photon damage, and rigid cell walls (characteristic of epidermises of xeric-adapted plants) exhibit greater resistance to collapse from negative turgor pressure (Beckett, 1997). In all four species of Satanocrater, we observed large epidermal cells with thick cutinized outer cell walls (similarly seen in another xerophytic Acanthaceae, e.g., Blepharis ciliaris; Akhani et al., 2008). Trichomes and peltate glandular scales (PGS)—Together with cuticular waxes, trichomes and other indumentum coverings can serve as protection against excessive sunlight and water loss via transpiration and evaporation (Esau, 1977; van der Merwe et al., 1994; Stenglein et al., 2005). In this study, we found that nearly all plant surfaces of the four species of Satanocrater are covered by trichomes and/or PGS. Remarkably, PGS were observed covering reproductive structures such as seed surfaces as well as styles, anthers, and corollas. To our knowledge, these have not yet been reported on reproductive structures of other Acanthaceae. Because seed, anther, and corolla glands were indistinguishable from those on leaf surfaces, we hypothesize that reproductive and vegetative PGS are homologous in Satanocrater, but evolutionary–developmental studies are needed to assess this primary homology statement (de Pinna, 1991). Although the dense trichome and PGS covering plant surfaces is consistent with a hypothesis of heat and xeric adaptations, we cannot rule out alternative functions of these traits. In other plants, trichomes have an antiherbivory function (Valverde et al., 2001), and PGS have been shown to function in the production of aromatic compounds (Lassak and McCarthy, 1983; Gersbach, 2002; but note that some have hypothesized that aromatic and ethereal oils may lower evaporation and transpiration rates, reviewed in Fahn, 1990). In Satanocrater, all portions of plants are pungent when crushed. 978 AMERICAN JOURNAL OF BOTANY [Vol. 99 Although the presence of glands is likely sympleisiomorphic in Acanthaceae (Daniel, 1990; Ezcurra, 1993; McDade and Tripp, 2007; Akhani et al., 2008; E. A. Tripp et al., unpublished manuscript), the homology of glands of Satanocrater to those of other Acanthaceae remains to be tested, especially given that those of Satanocrater are up to twice the size of glands of other genera (E. A. Tripp and S. Fatimah, unpublished data). Parenchyma—Among the most prominent leaf features of xerophytes are lower surface area to volume ratios and reduced extracellular air spaces (Schimper, 1903; Maximov, 1929; Weaver and Clements, 1929). In xerophytes, this is accomplished by reducing cell size and increasing the proportion of palisade to spongy parenchyma, sometimes with a complete loss of the latter (McDougall and Penfound, 1928; Weaver and Clements, 1929). Increasing surface area and reducing extracellular air space within leaves results in more efficient and more rapid gas exchange as well as water transport, which enables increased photosynthesis when water is available for use (Mauseth, 1988; Fahn and Cutler, 1992). Additionally, when water is unavailable, small cells withstand negative turgor pressure better than large cells. Three of the four species of Satanocrater (S. fellatensis, S. ruspolii, S. somalensis) have unifacial leaves with extremely reduced air spaces. This is in marked contrast to most other genera in Ruellieae (M. Fekadu, Addis Ababa University, and E. A. Tripp, unpublished data) that have bifacial leaves with distinct palisade and spongy layers. Unlike these three species of Satanocrater, S. paradoxus has an abaxial parenchyma layer that is differentiable from the adaxial layer, but the abaxial cells are, in general, still tightly packed. Although all four species occur in arid environments of sub-Saharan Africa, Satanocrater paradoxus primarily grows in partially shaded, Acacia-Commiphora woodlands, whereas the others primarily grow in open, exposed bushlands (Fig. 17; Vollesen, 2008; E. A. Tripp, personal observation; but apparently variable in S. fellatensis based on herbarium label data). It is possible that anatomical differences between S. paradoxus and the other three species can be attributed to habitat differences. Whether unifacial leaves are ancestral within Satancrater, with one subsequent loss in S. paradoxus, or whether the condition has evolved multiple times, requires phylogenetic information. Although multiple accessions of the four species were not sampled exhaustively for this study, data from Eif3E and, to a lesser extent, trnG-trnS, provide some evidence that the redflowered taxa S. paradoxus and S. somalensis are derived with respect to the purple-flowered taxa S. fellatensis and S. ruspolli, but branch support was strong only in the nuclear analysis. The combined analysis resolved S. ruspolii as sister to the other three taxa (ML results only). If these phylogenetic hypotheses approximate true evolutionary history, then we can conclude that unifacial leaves are likely ancestral within Satanocrater. Stomata—Ample research on xerophytes has documented the presence of stomata that are sunken into furrows, grooves, or crypts, or are overarched by some other protective structures such as subsidiary cells or trichomes (Fahn and Cutler, 1992; but see Jordan et al., 2008 for examples of complexity). Sunken stomata, or stomata protected by adjacent cells, trichomes, and/or cuticular outgrowths, are shielded from the drying effects of wind, thereby reducing passive water loss (i.e., evapotranspiration) compared to conditions experienced by stomata occurring on the leaf surface plane. Furthermore, the air space formed by sunken stomata (or stomata otherwise protected) helps to moderate surface Fig. 17. Typical habitat of (A) Satanocrater ruspolii and (B) S. paradoxus. (A) Open and exposed bushlands. (B) Semishaded woodlands (S. paradoxus in foreground). temperature and maintain atmospheric moisture surrounding the stomata (van der Merwe et al., 1994). These features combine to increase photosynthetic rates in xerophytes by permitting longer periods of open stomata. Similar to the pattern described above with leaf parenchyma, SEM study indicated that three of the four species of Satanocrater (S. fellatensis, S. ruspolii, S. somalensis) have stomata that are overarched by adjacent subsidiary cells whereas stomata of the fourth species, S. paradoxus, are not clearly on a different plane from surrounding cells. Again, one hypothesis to explain this pattern is the distribution of plants of S. paradoxus primarily in shaded woodlands vs. the open bushland habitats typical of the other three species. However, these results are based on a limited number of observations, and natural history including habitat preferences of species of Satanocrater, particularly for S. fellatensis, is scarce. Preliminary phylogenetic information regarding relationships among species presented here suggest that overarching subsidiary cells may be ancestral with respect to nonoverarched cells. As has been found in other Acanthaceae (e.g., Inamdar et al., 1983; Akhani et al., 2008), subsidiary cells were often diacytic in all species of Satanocrater except S. fellatensis, for which poor preservation state of leaf material precluded better observation of its stomatal features. Bundle sheath cells—Many structural and biochemical modifications have evolved in association with xerophytism. June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER Among the most studied modifications are those pertaining to the photosynthetic pathway. C4 photosynthesis, for example, has been documented in ca. 7500 angiosperms encompassing 45 evolutionary origins in 19 families. These taxa are concentrated in five geographic regions, one of which is Africa between the Tropic of Cancer and Tropic of Capricorn (Sage, 2004). Despite the widespread occurrence of C4 photosynthesis both phylogenetically and geographically, there is only one documented case of C4 photosynthesis among the 221 genera presently recognized (Scotland and Vollesen, 2000) in Acanthaceae (Blepharis of tribe Acantheae; Akhani et al., 2008), and only one other occurrence in the large order Lamiales (ca. 25 000 species), to which Acanthaceae belong (Sage, 2004). C4 photosynthesis is often accompanied by the presence of enlarged bundle sheath cells, in combination with several other anatomical features such as an orderly arrangement of mesophyll cells surrounding bundle sheaths, i.e., Kranz anatomy. In this study, we observed enlarged bundle sheath cells in at least one species, Satanocrater paradoxus, and possibly also in S. somalensis, but cellular preservation state was poor in the latter species owing to utilization of herbarium material, which prevented unambiguous assessment of this feature. Thus, plant anatomy prompted our inquiry into the possibility of C4 photosynthesis in Satanocrater. However, δ13C values indicate that plants of this genus use the C3 rather than C4 photosynthetic pathway. Because preliminary phylogenetic data on species relationships in Satanocrater suggest that S. paradoxus and S. somalensis may be each other’s closest relatives, similar bundle sheath structures may be attributable to phylogenetic history. Comparative morphology, anatomy, and phylogeny—Our second goal in this study was to test Scotland and Vollesen’s (2000) classification of Acanthaceae in which they treated Satanocrater within the tribe Ruellieae. To test the hypothesis, we used newly synthesized data derived from our morphological characterization of the genus plus molecular phylogenetic data derived from all major lineages of Acanthaceae. Both data sources provided strong support for the placement of Satanocrater within Ruellieae. These data contribute new information on trait distributions across Acanthaceae. They also suggest new hypotheses for species relationships within Satanocrater (see Conclusions). Synapomorphies that have been proposed to unite the 48 genera in Ruellieae include left-contort corolla aestivation, the presence of a filament curtain, and seeds with mucilaginous, hygroscopic trichomes. Among Ruellieae, no pattern of corolla aestivation other than left-contort has been documented. Although there are two instances of left-contort aestivation outside of Ruellieae in Acanthaceae s.s. (Lankesteria and Whitfieldia of tribe Whitfieldieae, Fig. 1), the feature appears to strongly characterize Ruellieae. Given a lack of additional evidence that would otherwise suggest a relationship of Satanocrater to Whitfieldieae, the presence of left contort corolla aestivation in Satanocrater supports its placement in Ruellieae. Similarly, our finding of a filament curtain in Satanocrater (originally documented by Manktelow, 2000) supports the placement of Satanocrater in Ruellieae. Although the filament curtain has generally been discussed as a discrete trait, for example, with phaulopsoid and reduced representing two of the four different types, our study of S. fellatensis suggests there may exist character state intermediates. Regardless, the genus as a whole is polymorphic with respect to curtain type: reduced curtains characterize the bird-pollinated species, and phaulopsoid curtains characterize one of the two putatively insect-pollinated species. Similar variation 979 (but within a single curtain type, i.e., phaulopsoid) has also been observed in species of Ruellia that have different pollination syndromes (E. A. Tripp, personal observation). The occurrence of mucilage-producing cells on seed indumenta that expand and become sticky when wetted has been termed myxospermy (Grubert, 1974). Although most Acanthaceae have some sort of vestiture on seed surfaces, usually in the form of trichomes or scales, the vast majority of species with mucilaginous trichomes belong to Ruellieae. Still, hygroscopic trichomes have been documented in three of the other major lineages of Acanthaceae including Acantheae, Barlerieae, and Justicieae (Fig. 1; Grubert, 1974; Gutterman and Witztum, 1977). Although the mucilaginous seed trichomes of Satanocrater fellatensis and S. ruspolii (no seeds seen for the other two species) is consistent with the placement of the genus in Ruellieae, the scattered occurrence of this trait in other lineages of Acanthaceae calls into question the homology of myxospermy in Acanthaceae. Further research on the distribution of this trait throughout the family is needed. Among the tremendous breadth of variation in traits across Acanthaceae, there is perhaps no trait better studied and more used taxonomically than pollen morphology (Lindau, 1895; Scotland, 1992; Daniel, 1998). At present, there is no known pollen synapomorphy for uniting all genera in Ruellieae, but many groups within the tribe have been studied and are well-characterized by pollen type (Scotland, 1993; Furness, 1994, 1995; Carine and Scotland, 1998; Tripp, 2007; Tripp et al., 2009). All species of Satanocrater have round, weakly colporate to porate, triaperturate pollen with a coarsely reticulate exine. The only other genera in Ruellieae with this coarse reticulate (honeycomb) pollen, or “Wabenpollen” (Lindau, 1893, 1895), are Eranthemum and Ruellia, which preliminary molecular data indicate are distantly related in Ruellieae (Tripp et al., unpublished manuscript). Like Satanocrater, pollen of Eranthemum is also round and tricolporate, but individual muri of Eranthemum grains have a single (rarely up to four), large, sexine verruca (Sharma and Vishnu-Mittre, 1963), whereas muri of Satanocrater apparently lack sexine verrucae. Pollen grains of Ruellia either lack verrucae or have one to numerous small verrucae (Daniel, 1998; Tripp, 2007), but these do not reach the size of those of Eranthemum. Most Ruellia have triporate, not tricolporate pollen, although broad geographical sampling by Furness and Grant (1996) and by Tripp et al. (2009) indicated that pollen grains of two African and a few American species are tricolporate. The presence of four fertile stamens in Satanocrater further lends support to its closer relationship to Ruellia than to Eranthemum, which is part of a clade containing two fertile stamens plus two staminodes (E. A. Tripp et al., unpublished manuscript). Ruellia is part of a larger clade that includes other genera (such as Acanthopale, Figs. 14–16) with triporate, round grains, whereas most of the rest of Ruellieae consists of tricolporate genera with prolate grains and pseudocolpi (Figs. 14–16). Thus, pollen of Satanocrater resembles Ruellia in shape and tectum sculpturing (and by lacking pseudocolpi), but shares the colporate apertures of genera outside the Ruellia clade. We hypothesize that pollen of Satanocrater is intermediate between the Ruellia clade and the other major clades within Ruellieae. Phylogenetic analysis that included a limited number of other genera in Ruellieae support plausibility of this hypothesis. In preliminary analyses that included a larger sampling of genera in Ruellieae (E. A. Tripp et al., unpublished data), Satanocrater and Ruellia are resolved as close relatives, with Ruellia derived with respect to Satanocrater. 980 AMERICAN JOURNAL OF BOTANY Conclusions— As predicted, all four species of Satanocrater possess anatomical and morphological features consistent with xerophyte adaptations. Satanocrater is one of a small handful of ca. 221 genera in Acanthaceae to have been studied anatomically. The present investigation represents an important step forward in a full synthesis of the evolution of this large and important plant family. Second, results from our morphological and molecular study confirm previous hypotheses that Satanocrater belongs in Ruellieae. Within Ruellieae, pollen morphology and molecular information suggest a close relationship of Satanocrater to Ruellia (and Acanthopale), and we hypothesize that pollen of Satanocrater may reflect an intermediate stage within a transition series from colporate to porate and pseudocolpate to nonpseudocolpate pollen types in Ruellieae. On the basis of the evidence from our anatomical and morphological study, we here postulate three competing hypotheses of evolutionary relatedness among species of Satanocrater. First, S. somalensis and S. paradoxus may be sister taxa (at the exclusion of S. fellatensis and S. ruspolii) based on the presence of enlarged bundle sheath cells (H1). Second, S. fellatensis + S. ruspolii + S. somalensis may form a clade (at the exclusion of S. paradoxus) based on their possession of unifacial leaves with stomata overarched by subsidiary cells (H2). Third, S. somalensis + S. paradoxus may be sister taxa, and S. fellatensisi + S. ruspolii also sister taxa, based on pollination syndromes, with bird pollination in the former pair and insect pollination in the latter pair (H3). Although we did not exhaustively sample multiple accessions of all four species of Satanocrater to reconstruction species relationships, DNA sequence data from Eif3E (and to a lesser degree trnG-trnS) lend some support to H1 and H3, but not to H2. However, studies across multiple other groups of plants have demonstrated amply that the traits involved in H1–H3 are frequently subject to convergent evolution (e.g., Cerros-Tlatilpa and Columbus, 2009; Yamaguchi et al., 2010). In Acanthaceae, pollination syndromes in particular are especially labile (Tripp and Manos, 2008; Daniel et al., 2008). Future studies with increased specimen and molecular locus sampling, ideally with access to fresh material, will help to discriminate among the above hypotheses of relationships within Satanocrater. LITERATURE CITED AKHANI, H., M. GHASEMKHANI, S. D. X. CHUONG, AND G. E. EDWARS. 2008. Occurrence and forms of Kranz anatomy in photosynthetic organs and characterization of NAD-ME subtype C4 photosynthesis in Blepharis ciliaris (L.) B. L. Burtt (Acanthaceae). Journal of Experimental Botany 59: 1755–1765. BECKETT, R. P. 1997. Pressure–volume analysis of a range of poikilohydric plants implies the existence of negative turgor in vegetative cells. Annals of Botany 79: 145–152. BORG, A. J., L. A. MCDADE, AND J. SCHÖNENBERGER. 2008. Molecular phylogenetics and morphological evolution of Thunbergioideae (Acanthaceae). Taxon 57: 1–12. BURROWS, G. E. 2001. Comparative anatomy of the photosynthetic organs of 39 xeromorphic species from subhumid New South Wales, Australia. International Journal of Plant Sciences 162: 411–430. CARINE, M. A., AND R. W. SCOTLAND. 1998. Pollen morphology of Strobilanthes Blume (Acantahceae) from southern India and Sri Lanka. Review of Palaeobotany and Palynology 103: 143–165. CARINE, M. A., AND R. W. SCOTLAND. 2000. 68 Taxa and 32 characters: Resolving species relationships using morphological data. In C. M. Morton, M. M. Harley, and S. Blackmore [eds.], Pollen and spores: Morphology and biology, 365–384. Royal Botanic Garden, Kew, UK. CARINE, M. A., AND R. W. SCOTLAND. 2002. Classification of Strobilanthinae (Acanthaceae): Trying to classify the unclassifiable? Taxon 51: 259–279. [Vol. 99 CERROS-TLATILPA, R., AND J. T. COLUMBUS. 2009. C3 photosynthesis in Aristida longifolia: Implication for photosynthetic diversification in Aristidoideae (Poaceae). American Journal of Botany 96: 1379–1387. CHOMCZYNSKI, P., AND N. SACCHI. 2006. The single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction: Twenty-something years on. Nature Protocols 1: 581–585. COLUMBUS, J. T. 1999. Morphology and leaf blade anatomy suggest a close relationship between Bouteloua aristidoides and B. (Chondrosium) eriopoda (Gramineae: Chloridoideae). Systematic Botany 23: 467–478. DANIEL, T. F. 1990. New, reconsidered, and little-known Mexican species of Ruellia (Acanthaceae). Contributions from the University of Michigan Herbarium 17: 139–162. DANIEL, T. F. 1998. Pollen morphology of Mexican Acanthaceae: Diversity and systematic significance. Proceedings of the California Academy of Sciences 50: 217–256. DANIEL, T. F., L. A. MCDADE, M. MANKTELOW, AND C. A. KIEL. 2008. The “Tetramerium lineage” (Acanthaceae: Acanthoideae: Justicieae): Delimitation and intra-lineage relationships based on cp and nrITS sequence data. Systematic Botany 33: 416–436. DE PINNA, M. C. C. 1991. Concepts and tests of homology in the cladistic paradigm. Cladistics 7: 367–394. DELF, E. M. 1915. The meaning of xerophily. Journal of Ecology 3: 110–121. DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. ESAU, K. 1977. Anatomy of seed plants, 2nd ed. John Wiley, New York, New York, USA. EZCURRA, C. 1993. Systematics of Ruellia (Acanthaceae) in southern South America. Annals of the Missouri Botanical Garden 80: 787–845. FAHN, A. 1990. Plant anatomy, 4th ed. Pergamon Press, Oxford, U.K. FAHN, A., AND D. F. CUTLER. 1992. Xerophytes. Gebruder Borntraeger, Berlin, Germany. FURNESS, C. A. 1994. The pollen morphology of Hygrophila and Brillantaisia (Acanthaceae: Ruellieae). Acta Botanica Gallica 141: 267–278. FURNESS, C. A. 1995. A pollen morphological study of Dyschoriste Nees and Chaetacanthus Nees (Acanthaceae: Ruellieae). Review of Palaeobotany and Palynology 84: 331–345. FURNESS, C. A., AND M. C. GRANT. 1996. Pollen morphology of some Ruellia species (Acanthaceae) from Africa and Madagascar. Grana 35: 231–239. GERSBACH, P. V. 2002. The essential oil secretory structures of Prostanthera ovalifolia (Lamiacceae). Annals of Botany 89: 255–260. GRUBERT, M. W. 1974. Studies on the distribution of myxospermy among seeds and fruits of Angiospermae and its ecological significance. Acta Botanica Venezuelica 8: 315–551. GUTTERMAN, Y., AND A. WITZTUM. 1977. The movement of integumentary hairs in Blepharis ciliaris (L.) Burtt. Botanical Gazette 138: 29–34. HEDRÉN, M., M. W. CHASE, AND R. G. OLMSTEAD. 1995. Relationships in the Acanthaceae and related families as suggested by cladistic analysis of rbcL nucleotide sequences. Plant Systematics and Evolution 194: 93–109. INAMDAR, J. A., D. C. BHATT, AND G. S. CHAUDHARI. 1983. Structure and development of stomata in some Acanthaceae. Proceedings of the Indiana Academy of Sciences 92: 285–296. JORDAN, G. J., P. H. WESTON, R. J. CARPENTER, R. A. DILLON, AND T. J. BRODRIBB. 2008. The evolutionary relations of sunken, covered, and encrypted stomata to dry habitats in Proteaceae. American Journal of Botany 95: 521–530. KIEL, C. A., L. A. MCDADE, T. F. DANIEL, AND D. CHAMPLUVIER. 2006. Phylogenetic delimitation of Isoglossinae (Acanthaceae: Justicieae) and relationships among constituent genera. Taxon 55: 683–693. LASSAK, E. V., AND T. MCCARTHY. 1983. Australian medicinal plants. Mandarin Publishing, Victoria, Australia. LI, M., J. WUNDER, G. BISSOLI, E. SCARPONI, S. GAZZANI, E. BARBARO, H. SAEDLER, AND C. VAROTTO. 2008. Development of COS genes as universally amplifiable markers for phylogenetic reconstructions of closely related plant species. Cladistics 24: 727–745. June 2012] TRIPP AND FATIMAH—ANATOMY AND MORPHOLOGY OF SATANOCRATER LINDAU, G. 1893. Beitrage zur Systematik der Acanthaceen. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 18: 36–64. LINDAU, G. 1895. Acanthaceae. In A. Engler and K. Prantl [eds.], Die natürlichen Pflanzenfamilien, vol. 4, part 3b, 274–354. Engelmann, Leipzig, Germany. MADDISON, D. R., AND W. P. MADDISON. 2003. MacClade: An analysis of phylogeny and character evolution, version 4.06. Sinauer, Sunderland, Massachusetts, USA. MANKTELOW, M. 2000. The filament curtain: A structure important to systematics and pollination biology in the Acanthaceae. Botanical Journal of the Linnean Society 133: 129–160. MANKTELOW, M., L. A. MCDADE, B. OXELMAN, C. A. FURNESS, AND M.-J. BALKWILL. 2001. The enigmatic tribe Whitfieldieae (Acanthaceae): Delimitation and phylogenetic relationships based on molecular and morphological data. Systematic Botany 26: 104–119. MARTIN, J. T., AND B. E. JUNIPER. 1970. The cuticles of plants. Edward Arnold, New York, New York, USA. MAUSETH, J. D. 1988. Plant anatomy. Benjamin/ Cummings, Menlo Park, California, USA MAXIMOV, N. A. 1929. The plant in relation to water. A study of the physiological basis of drought resistance. George Allen & Unwin, London, U.K. MAXIMOV, N. A. 1931. The physiological significance of the xeromorphic structure of plants. Journal of Ecology 19: 273–282. MCDOUGALL, W. B., AND W. T. PENFOUND. 1928. Anatomy of deciduous forest plants. Ecology 9: 349–353. MCDADE, L. A. 1990. Hybrids and phylogenetic systematics I. Patterns of character expression in hybrids and their implications for cladistic analysis. Evolution 44: 1685–1700. MCDADE, L. A. 1992. Pollinator relationships, biogeography, and phylogenetics. Bioscience 42: 21–26. MCDADE, L. A., T. F. DANIEL, AND C. A. KIEL. 2008. Toward a comprehensive understanding of phylogenetic relationships among lineages of Acanthaceae s.l. (Lamiales). American Journal of Botany 95: 1136–1152. MCDADE, L. A., T. F. DANIEL, C. A. KIEL, AND K. VOLLESEN. 2005. Phylogenetic relationships among Acantheae (Acanthaceae): Major lineages present contrasting patters of molecular evolution and morphological differentiation. Systematic Botany 30: 834–862. MCDADE, L. A., C. A. KIEL, T. F. DANIEL, AND E. A. TRIPP. 2008b. Biogeography of the Acanthaceae. South African Journal of Botany 74: 358. MCDADE, L. A., S. E. MASTA, M. L. MOODY, AND E. WATERS. 2000. Phylogenetic relationships among Acanthaceae: Evidence from two genomes. Systematic Botany 25: 106–121. MCDADE, L. A., AND M. L. MOODY. 1999. Phylogenetic relationships among Acanthaceae: Evidence from non-coding trnL-trnF chloroplast DNA sequences. American Journal of Botany 86: 70–80. MCDADE, L. A., AND E. A. TRIPP. 2007. A synopsis of Costa Rican Ruellia (Acanthaceae), with descriptions of four new species. Brittonia 59: 199–216. METCALFE, C. R., AND L. CHALK. 1950. Anatomy of the dicotyledons, vol. I. Clarendon Press, Oxford, U.K. MOYLAN, E. C., J. R. BENNETT, M. A. CARINE, R. G. OLMSTEAD, AND R. W. SCOTLAND. 2004. Phylogenetic relationships among Strobilanthes s.l. (Acanthaceae): Evidence from ITS nrDNA, trnL-F cpDNA, and morphology. American Journal of Botany 91: 724–735. OCAMPO, G., AND J. T. COLUMBUS. 2010. Molecular phylogenetics of suborder Cactineae (Caryophyllales), including insights into photosynthetic diversification and historical biogeography. American Journal of Botany 97: 1827–1847. O’LEARY, M. H. 1988. Carbon isotopes in photosynthesis. BioScience 38: 328–336. SAGE, R. F. 2004. The evolution of C4 photosynthesis. New Phytologist 161: 341–370. SCOTLAND, R. W. 1992. Systematics, similarity and Acanthaceae pollen morphology. Botanical Journal of the Linnean Society 109: 529–541. SCOTLAND, R. W. 1993. Pollen morphology of Contortae (Acanthaceae). Botanical Journal of the Linnean Society 111: 471–504. SCOTLAND, R. W., P. K. ENDRESS, AND T. J. LAWRENCE. 1994. Corolla ontogeny and aestivation in the Acanthaceae. Botanical Journal of the Linnean Society 114: 49–65. 981 SCOTLAND, R. W., J. A. SWEERE, P. A. REEVES, AND R. G. OLMSTEAD. 1995. Higher level systematics of Acanthaceae determining by chloroplast DNA sequences. American Journal of Botany 82: 266–275. SCOTLAND, R. W., AND K. VOLLESEN. 2000. Classification of Acanthaceae. Kew Bulletin 55: 513–589. SCHIMPER, A. F. W. 1903. Plant geography upon a physiological basis. Clarendon Press, Oxford, UK. SCHMID, R., AND M. TURNER. 1977. Contrad 70, an effective softener of herbarium materials for anatomical study. Taxon 26: 551–552. SCHMIDT-LEBUHN, A. N., M. KESSLER, AND I. HENSEN. 2007. Hummingbirds as drivers of plant speciation? Trends in Plant Science 12: 329–331. SCHMIDT-LEBUHN, A. N., M. KESSLER, AND J. MÜLLER. 2005. Evolution of Suessenguthia (Acanthaceae) inferred from morphology, AFLP data, and ITS rDNA sequences. Organisms, Diversity & Evolution 5: 1–13. SHARMA, B. D., AND VISHNU-MITTRE. 1963. Contribution to the pollen morphology of the genera Eranthemum Linn. and Pseuderanthemum Radlkof. Proceedings of the National Institute of Science, India, B 29: 520–526. SHARMAN, B. C. 1943. Tannic acid and iron alum with safranin and orange G in studies of the shoot apex. Stain Technology 18: 105–111. SHIELDS, L. M. 1950. Leaf xeromorphy as related to physiological and structural influences. Botanical Review 16: 399–447. SHIELDS, L. M. 1951. Leaf xeromorphy in dicotyledon species from a gypsum sand deposit. American Journal of Botany 38: 175–190. STENGLEIN, S. A., A. M. ARAMBARRI, M. C. MENENDEZ-SEVILLANO, AND P. A. BALATTI. 2005. Leaf epidermal characters related with plant’s passive resistance to pathogens vary among accessions of wild beans Phaseolus vulgaris var. aborigineus (Leguminosae-Phaseoleae). Flora 200: 285–295. SWOFFORD, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer, Sunderland, Massachusetts, USA. THODAY, D. 1933. The terminology of “xerophytism”. Journal of Ecology 21: 1–6. THULIN, M. 2007. Synopsis of Satanocrater (Acanthaceae). Nordic Journal of Botany 24: 385–388. TRIPP, E. A. 2007. Evolutionary relationships within the species-rich genus Ruellia (Acanthaceae). Systematic Botany 32: 628–649. TRIPP, E. A. 2010. Taxonomic revision of Ruellia sect. Chiropterophila (Acanthaceae): A lineage of rare and endemic species from Mexico. Systematic Botany 35: 629–661. TRIPP, E. A., T. F. DANIEL, J. C. LENDEMER, AND L. A. MCDADE. 2009. New molecular and morphological insights prompt transfer of Blechum to Ruellia (Acanthaceae). Taxon 58: 893–906. TRIPP, E. A., S. Fatimah, I. Darbyshire, AND L. A. MCDADE. In Review Origin of African Physacanthus (Acanthaceae) via wide hybridization. TRIPP, E. A., AND P. S. MANOS. 2008. Is floral specialization an evolutionary dead-end? Pollination system evolution in Ruellia (Acanthaceae). Evolution 62: 1712–1737. TRIPP, E. A., L. A. McDade, S. Fatimah, AND T. F. Daniel. In Review. Phylogenetic relationships within Ruellieae (Acanthaceae), and a revised classification. VALVERDE, P. L., J. FORNONI, AND J. NÚÑEZ-FARFÁN. 2001. Defensive role of leaf trichomes in resistance to herbivorous insects in Datura stramonium. Journal of Evolutionary Biology 14: 424–432. VAN DER MERWE, A. M., J. J. A. VAN DER WALT, AND E. M. MARAIS. 1994. Anatomical adaptations in the leaves of selected fynbos species. South African Journal of Botany 60: 99–107. VOLLESEN, K. 2008. Acanthaceae (Part I). In H. J. Beentje and S. A. Ghazanfar [eds.], Flora of tropical East Africa, 1–286. Royal Botanic Gardens, Kew, UK. WEAVER J. E., AND F. E. CLEMENTS. 1929. Plant ecology. McGraw-Hill, New York, New York, USA. WIESNER, J. 1889. Biologie der Pflanzen. A. Hölder, Wien, Austria. YAMAGUCHI, T., S. YANO, AND H. TSUKAYA. 2010. Genetic framework for flattened leaf blade formation in unifacial leaves of Juncus prismatocarpus. Plant Cell 22: 2141–2155. ZWICKL, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, University of Texas at Austin, Austin, Texas, USA. 982 AMERICAN JOURNAL OF BOTANY APPENDIX 1. Voucher, information for all specimens examined. Appendix is arranged phylogenetically (see Fig. 1). Three specimens were unidentifiable based in scant herbarium material: Becket & White 1591 (K); Hemming 1708 (K); and Kuchar 16882 (K). Taxon – Voucher (Herbarium), GenBank: trnG-trnS, trnG-trnR, psbA-trnH, Eif3E, if voucher used in molecular study) Outgroups: Sesamum indicum L. – Jenkins 97-141 (ARIZ), EU528998, JQ78010019. Nelsonioideae: Staurogyne letestuana Benoist – NBG-B 200000119-77 (BR), EU529126, JQ7801020. Avicennia lineage: Avicennia bicolor Standl. – Borg 10 (S), EU528943, JQ780996. Thunbergioideae: Mendoncia phytocrenoides Benoist –Schönenberger 50 (K), EU528983, JQ7801005. Acantheae: Aphelandra leonardii McDade – McDade 310 (DUKE), DQ059287, JQ780994; Stenandriopsis guineensis (Nees) Benoist – K1990-2299 (K), DQ059258, JQ7801021. Barlerieae: Barleria lupulina Lindl. – Daniel s.n. (CAS), EU528946, JQ780996; Crabbea acaulis N.E. Br. – Balkwill et al. 11649 (J), EU528953, JQ7801000. Andrographideae: Cystacanthus turgida G. Nicholson – collector unknown 1996-479 (K), EU528954, JQ7801001. Whitfieldeae: Camarotea souiensis Scott-Elliot – Decary s.n. (US), JQ7801024, JQ780998; Chlamydacathus euphorbioides Lindau – Daniel et al. 10445 (CAS), EU528951, JQ780999. Neuracanthus lineage: Neuracanthus umbraticus Bidgood & Brummitt – Daniel 6770.5 (CAS), EU528991, JQ7801006. Justiceae: Mackaya bella Harv. – Daniel s.n. (CAS), EU528979, JQ7801003; Rhinacanthus gracilis Klotzsch – Daniel s.n. (CAS), EU528995, JQ7801009. Ruellieae: Acanthopale confertiflora (Lindau) C.B. Clarke – Phillipson 2117 (MO), JQ7801022, EF214651, JQ7801035, JQ763413; Brillantaisia grotanellii Pic. Serm. – Tripp & Ensermu 924 (RSA), JQ7801023, JQ780997, JQ7801036, JQ763418; Duosperma kilimandscharica (C.B. Clarke) Dayton – Kindeketa et al. 1526 (MO), JQ7801025, EF214605, JQ7801037, JQ763415; Hygrophila pilosa Raf. – Gates 204 (NY), JQ7801026, JQ7801002, JQ7801038, JQ763419; Mellera submutica C.B. Clarke – Richards 18210 (S), JQ7801027, JQ7801004, JQ7801039, JQ763416; Petalidium canescens C.B. Clarke – Tripp & Dexter 882 (RSA), JQ7801028, JQ7801007, JQ7801040, JQ763414; Phaulopsis imbricata Sweet – Tripp & Ensermu 929 (RSA), JQ7801029, JQ7801008, JQ7801041, JQ763417; Ruellia longipedunculata Lindau – Wood 13750 (US), EU431045, EF214686, GQ995633, JQ763412; Satanocrater fellatensis Schweinf. – Adam 13815 (MO); Berhaut 4158 (K); Boulos & Getahun 11812 (K); Burger 3301 (K); Friis et al. 6884 (C, K), JQ7801030, JQ7801010, JQ7801042 [specimen at C used for molecular study]; Friis et al. 10721 (K); Schweinfurth 1306 (K); Satanocrater ruspolii (Lindau) Lindau – Bally 10111 (K); Bidgood et al. 4982 (K); Burger 3301 (K); Burger 3327 (US, K); Burger 3330 (US); de Wilde 6665 (MO); Gilbert et al. 8237 (K); Gillett 4483 (K); Glover & Gilliland 803 (K); Friis et al. 9894 (C); Friis et al. 1020 (K); Friis et al. 11117 (K); Kuchar 16903 (K); Kuchar 17604 (MO); Kuchar 17604 (MO, CAS); Thomson 16 (K); Thomson 17 (K); Tripp & Ensermu 904 (RSA), JQ7801015, JQ7801046, [Eif3E sequence not acceptable into GenBank because length of 198 is too short]; Satanocrater somelensis (Lindau) Lindau – Bally 7288 (K); Bally & Melville 15679 (K); Beckett 639 (K), JQ7801018, JQ7801048, JQ763422; Hemming 1401 (K); Gillett 22998 (K); Gillett et al. 22329 (K); Gollerette 319 (K); Lavranos et al. 24660 (MO), JQ7801033, JQ7801016; Nuget 13 (K); Philips s.n. (K); Thulin et al. 10098 (K), JQ7801034, JQ7801017, JQ7801047, JQ763421; Satanocrater paradoxus (Lindau) Lindau – Bally & Smith B14916 (K); Bally & Smith B14581, (K); Brockman 188 (K); Brockman 189 (K); Brockman 195 (K); Jess 422 (K); de Wilde 6657 (MO); de Wilde 7312 (MO), JQ7801012, JQ7801044; Gillet & Hemming 24245 (K); Gilbert 3370 (K); Gilbert et al. 7469 (MO, K), JQ7801031, JQ7801011, JQ7801043, JQ763420 [specimen at MO used for molecular study]; Friis et al. 3147 (C, MO, K); Friis et al. 10991 (K); Kuchar 16153 (K); Thulin & Warfa 5345 (K); Thulin & Bashir 6927 (K); Tripp & Ensermu 906 (RSA), JQ7801014, JQ7801045; Weiland 1341 (MO), JQ7801032, JQ7801013.