Evolutionary Ecology

, Volume 29, Issue 3, pp 355–377

Convergence of anti-bee pollination mechanisms in the Neotropical plant genus Drymonia (Gesneriaceae)

Original Paper

Abstract

The neotropical plant genus Drymonia displays a remarkable variety of floral shapes and colors. One feature that is particularly important to coevolution with pollinators involves the variable shapes and widths of corolla tubes. To evaluate the evolutionary context for changes in corolla shape, we constructed a phylogeny of 50 of the 75 species of Drymonia using molecular markers from plastid (trnK-matK) and nuclear regions (ITS and ETS). Mapping tube shapes on the phylogeny supports open, bell-shaped (campanulate) corolla shape as the ancestral character state for Drymonia, with multiple independent origins of constriction in the corolla tube. Corollas with constrictions take one of three tube shapes: a constricted flower opening and throat with a large, expanded pouch on the lower surface (hypocyrtoid); a constricted flower opening and throat lacking an expanded pouch on the lower surface (urceolate); or a constricted opening and throat where the sides of the corolla appear laterally compressed. Fieldwork demonstrates euglossine bees (mostly Euglossa spp. and Epicharis spp.) visit campanulate corollas while hummingbirds visit corollas that are constricted. Results support eight independent origins of constricted corolla tubes from ancestors with campanulate corolla tubes: 3 hypocyrtoid clades, 3 laterally compressed clades, and 3 urceolate clades (one of which represents a shift from a hypocyrtoid ancestor). Constricted corollas are associated with shifts from the ancestral condition of poricidal anther dehiscence, which presents pollen to pollinators in multiple small doses, to the derived condition of longitudinal anther dehiscence, which presents all pollen to pollinators simultaneously. The association of hummingbird pollination with constricted corolla tubes suggests that narrowing evolved as a barrier mechanism that prohibits the visitation of flowers by bees.

Keywords

Convergence Drymonia Gesneriaceae Hypocyrtoid corollas Laterally compressed corollas Pollination biology Poricidal anther dehiscence Urceolate corollas 

Introduction

Flowers of neotropical plants of the family Gesneriaceae have diversified into a remarkable array of colors and shapes (Fig. 1), suggesting a diverse coevolutionary history with pollinators. Few pollination studies or species-level phylogenies exist for the group. An understanding of the ecology and evolution of their flowers has been further hampered by a confusing classification system and lack of monophyly for many of the traditionally recognized genera because species were frequently shifted between poorly defined genera and genera were shifted between poorly defined tribes (Hanstein 1854, 1856, 1859, 1865; Fritsch 1893–1894; Martius 1829; Ivanina 1965, 1967; Wiehler 1973, 1983; Burtt and Wiehler 1995; Möller and Clark 2013; Weber et al. 2013). Recent molecular-based studies have begun to clarify phylogenetic relationships and circumscribe monophyletic genera (e.g., Smith and Atkinson 1998; Smith et al. 1997, 1998, 2004a, b; Smith and Clark 2013; Zimmer et al. 2002, Clark et al. 2006, 2012; Möller and Clark 2013; Weber et al. 2013). The objective of this study is to provide an evolutionary and ecological context for understanding the evolution of the narrowing of corolla tubes in Drymonia Mart. and how this feature may function as a barrier mechanism to pollinators.
Fig. 1

Corolla shape variation evaluated in Drymonia. a, b Bell-shaped (campanulate) in Drymonia brochidodroma. c, d Laterally compressed in Drymonia multiflora. e, f Pouched (hypocyrtoid) in Drymonia teuscheri. g, h Urn-shaped (urceolate) in Drymonia urceolata. Photos from field collections by John L. Clark (a, b J.L. Clark et al. 6354; c, d J.L. Clark et al. 12499; e, f J.L. Clark et al. 6369; g J.L. Clark 10006; h J.L. Clark 9005)

Drymonia is one of the largest genera of Neotropical Gesneriaceae, with 75 species (Weber et al. 2013; Möller and Clark 2013). Martius (1829) circumscribed Drymonia on the basis of a leafy calyx and large corolla, but these features are also found in many other closely related genera. More recently Moore (1955) characterized Drymonia from other Gesneriaceae by the presence of poricidal anther dehiscence (Fig. 2a). Instead of undergoing longitudinal dehiscence (Fig. 2b), with thecae splitting fully open along the length and presenting all pollen simultaneously, the thecae in Drymonia open by a short basal pore which slowly releases pollen throughout anthesis (Fig. 2c, d). Wiehler (1983) aptly described these poricidal anthers as “salt-shaker-like.” In bud, the four anthers are grouped coherently around the style, with their pore-like thecae facing inward, and become connate along the length of their margins as they mature. Prior to anthesis, the curvature and the differential length of the filament pairs invert the anther structure by turning it upside down (i.e., rotating 180°), causing the basal pores to face upwards before they open. During anthesis, the strategically placed anthers are thus able to pour or “shake” their powdery pollen grains through the pores onto visitors when they tip the structure over as they enter the flower (Fig. 2f). Steiner (1985) noted that gland-tipped trichomes located inside the corollas of Drymonia serrulata (Jacq.) Mart. exuded oil that played a role in promoting the adhesion of pollen grains to the body of Epicharis bees (family Apidae, subfamily Apinae). The ‘salt-shaker’ structure releases the pollen in doses, such that it takes five to eight visits to fully empty the anthers (Wiehler 1983).
Fig. 2

Plant-pollinator interactions and anther dehiscence. a Poricidal anther dehiscence in Drymonia killipii (scale in mm). b Longitudinal anther dehiscence in Columnea medicinallis. c, d Poricidal anther dehiscence in Drymonia urceolata.eEuglossa bee captured and photographed from recently visited flower of Drymonia ecuadorensis, Rio Palenque Science Center. fDrymonia ecuadorensis visited by Euglossa bee. gDrymonia collegarum visited by Tawny-bellied Hermit (Phaethornis syrmatophorus). Photo a by Richard W. Dunn; bf by John L. Clark and G by Murray Cooper (a R.W. Dunn s.n. from cultivated material; b J.L. Clark et al. 10006; c J.L. Clark et al. 6906; dJ.L. Clark 10006.e, f = From Rio Palenque Science Center, Ecuador, No voucher specimen; g = El Pahuma Orchid Reserve, Ecuador)

The majority of flowering plants have anthers that open by splitting longitudinally along the entire locule. In contrast, anthers that dehisce poricidally represent less than 10 % of flowering plants (Buchmann 1983). Poricidal anthers are almost entirely associated with vibratory pollen collection (“buzz pollination”) by bees (Buchmann 1983). The transfer of pollen grains in Drymonia flowers is not facilitated by vibrations and is therefore unique among taxa with poricidal anther dehiscence.

The remarkable array of corolla shapes and colors (Fig. 1) across Drymonia has resulted in a confusing taxonomic history. The traditional or pre-phylogenetic circumscription of Drymonia was limited to species with campanulate corollas (Wiehler 1983; Moore 1955), as featured in Fig. 1a, b. However, more recent molecular phylogenetic work demonstrated that the genus was paraphyletic and necessitated the transfer of species into Drymonia from other genera, including discordant taxa that were previously in Alloplectus Mart., Nautilocalyx Linden ex Hanst., and Paradrymonia Hanst. (Clark 2005; Clark et al. 2006, 2012). These changes resulted in the addition of many species with different corolla shapes, making Drymonia one of the most morphologically variable clades in the family. The convoluted taxonomic history of Drymonia demonstrates why relying on floral traits can be problematic as the basis for generic circumscriptions. A classification system based on pollination syndromes, or convergent floral adaptations to different pollinator types (Faegri and van der Pijl 1979; Fenster et al. 2004), will not necessarily reflect phylogenetic relationships. In the case of Drymonia, the campanulate corolla and ‘salt-shaker-anthers’ likely represent adaptations to bee pollination (Wiehler 1983; Steiner 1985). Molecular phylogenies were important precursors to studies such as the present on Drymonia because they helped define monophyletic units. In contrast, traditional classifications exemplified by Drymonia would recognize most of the non-bee pollinated flowers in other genera (e.g., Alloplectus, Paradrymonia, or Nautilocalyx).

Another good example of an artificial circumscription is the gesneriad genus Hypocyrta Mart. The genus is no longer recognized, but mentioning it here helps understand the evolutionary plasticity of corolla shapes in the Gesneriaceae, as the defining character of Hypocyrta was the corolla shape: specifically, a constricted flower opening and throat, with a large, expanded pouch on the lower surface (Fig. 1e, f). Some of the 44 species previously classified as Hypocyrta are now classified in Drymonia, and the rest nest in seven other genera including Besleria L., Codonanthe (Mart.) Hanst., Corytoplectus Oerst., Nematanthus Schrad., Pachycaulos J.L. Clark and J.F. Sm., Paradrymonia, and Pearcea Regel. Phylogenetic methods based on molecular sequence data have greatly facilitated the classification of the family by discarding artificially recognized genera such as Hypocyrta and defining monophyletic genera that are morphologically diverse such as Drymonia.

In the present paper, we combine ecological and phylogenetic approaches to evaluate the evolution of corolla shapes across Drymonia. Flowers in the genus can be classified into four general shapes based on the width of the corolla tube and the flower opening. Bell-shaped (campanulate) corollas have a relatively constant width from the base through the throat and flower opening (5.6–17.0 mm), with lobes flaring out around the opening (Fig. 1a, b). Most Drymonia flowers with bell-shaped corollas are yellowish-green to white, and they tend to have fimbriate margins on the lobes. Pouched (hypocyrtoid) corollas are defined by a constricted throat and flower opening (3.0–4.2 mm) and an often greatly expanded pouch on the lower surface (9.1–15.4 mm, Fig. 1e, f). Pouched corollas tend to have yellow tubes that contrast with bright reds on the lobes. Urn-shaped (urceolate) corollas are apically constricted like pouched corollas (to a narrowest corolla width of 4.0–4.8 mm), but lack the ventral pouch (Fig. 1g, h). Finally, laterally compressed corollas have throat widths of approx. 4.0–10.0 mm, similar to those of bell-shaped corollas, but the flower openings appear pinched into narrow “key-holes” (3.0–4.5 mm wide; Fig. 1c, d).

We hypothesize that open bell-shaped flowers are pollinated primarily by bees (Fig. 2e, f), while the three constricted flower shapes (pouched, urn-shaped, and laterally compressed) are pollinated primarily by hummingbirds (Fig. 2g) and that constricted flower openings represent adaptations to prevent access by bees to nectar and pollen. Grant and Grant (1968) originally proposed that such an association between narrow openings and hummingbird flowers serves to reduce bee visitation. Here we evaluate directionality of shifts in corolla shape and provide an initial assessment of pollination syndromes by developing a species-level phylogeny of the genus, recording pollinators of focal flowers for each of the shape classes, and surveying the literature for additional pollination records. We also map shifts in anther dehiscence (poricidal vs. longitudinal) and discuss implications for pollination.

Materials and methods

Taxon sampling and outgroup selection

Fifty-nine species were sequenced for the trnK-matK of plastid DNA (cpDNA), the internal transcribed spacer (ITS) region and the external transcribed spacer (ETS) region of 18S-26S of nuclear ribosomal DNA (nrDNA). The ingroup included 50 of 75 Drymonia species. This research represents the most comprehensive phylogenetic taxon sampling to date for Drymonia. Most species were photographed in the field and determinations were verified with herbarium voucher specimens, photographs, and literature. The study of type specimens was necessary for the identification of many Drymonia species and was carried out in conjunction with an ongoing monographic revision. Some Drymonia species, such as D. serrulata, are common roadside weeds in South and Central America, but most are local endemics that are only found in intact forests. Extensive fieldwork for the present study was necessary because many Drymonia in the analyses are only known from one or two localities (e.g., D. decora J.R. Clark and J.L. Clark, D. ignea J.L. Clark, D. peltata (Oliver) H.E. Moore, D. submarginalis Gómez-Laurito and Chavarría, and others). All taxa have fertile voucher specimens archived at the Smithsonian Institution’s U.S. National Herbarium (US) and The University of Alabama Herbarium (UNA). A complete list of samples, voucher specimens with locality, and GenBank accession numbers is provided in Appendix 1.

Outgroup samples were chosen on the basis of previous phylogenetic studies of Gesneriaceae and Episcieae (=subtribe Columneinae) (Clark et al. 2006, 2012). Given our focus on Drymonia, we limited outgroups to species in closely related genera from the core Episcieae clade outlined in Clark et al. (2012). Specifically, we used Alloplectus aquatilis C.V. Morton, Columnea (2 spp.), Corytoplectus congestus (Linden ex Hanst.) Wiehler, Crantzia cristata (L.) Scop. ex Fritsch, Glossoloma (2 spp.), Neomortonia rosea Wiehler, and Pachycaulos nummularia (Hanst.) J.L. Clark and J.F. Sm. (Appendix 1).

DNA extraction, amplification, and sequencing

Most genomic DNAs were isolated from silica-dried leaf material collected in the field. Leaf samples were ground using a ThermSavant FastPrep FP120 cell disrupter (Qbiogene, Carlsbad, CA). DNA was isolated using the Qiagen DNeasy™ DNA isolation kit (Qiagen, Valencia, CA).

Templates of the nrDNA internal transcribed spacer region (ITS) were prepared using the primers ITS5HP (Suh et al. 1993) and ITS4 (White et al. 1990). Additionally, the reverse and forward of the internal primers ITS2 and ITS3 (White et al. 1990) were used to obtain double stranded DNA sequence of the entire ITS region. Templates of the nrDNA external transcribed spacer region (ETS) were prepared using the primers18S-ETS (Baldwin and Markos 1998) and ETS-B developed for Mimulus (Phyramaceae) by Beardsley and Olmstead (2002). Templates of the cpDNA trnK-matK were prepared using primers trnK1 (CTAACTCAACGGTAGAGTACTCG) and matK (CTCCTGAAAGATAAGTGG).

Polymerase chain reaction (PCR) amplifications followed the procedures described by Baldwin et al. (1995) utilizing Taq DNA polymerase (Promega, Madison, WI). To reduce within-strand base pairing that can result in interference with Taq polymerase activity, we found it essential to use 5 % DMSO and 5 % BSA in PCR reactions for ETS and ITS. The PCR products were electrophoresed using a 1.0 % agarose gel in 1× TBE (pH 8.3) buffer, stained with ethidium bromide to confirm a single product, and purified using PEG 8000 (polyethylene glycol) in 2.5 M NaCl under the conditions described in Johnson and Soltis (1995). Direct cycle sequencing of purified template DNAs was performed by the Nevada Genomics Center (University of Nevada, Reno, NV).

DNA chromatograms were proofed, edited, and contigs were assembled using Sequencher 3.0 (Gene Codes Corporation, Ann Arbor, MI). Sequences were deposited in GenBank (Appendix 1).

Alignment

All sequences were aligned in the multiple sequence alignment program MUSCLE (Edgar 2004). Because the sequences were not highly divergent in the ingroup (i.e., Drymonia), it was possible to make minor adjustments to minimize overlapping gaps. This approach allowed for single-site and multiple-site gaps to be treated with equal weight (Simmons and Ochoterena 2000). All regions were easily aligned and none were excluded from the analyses.

Assignment of corolla shape and assessment of anther dehiscence

The corollas of Drymonia were assigned to one of the following four corolla shapes: (1) bell-shaped (campanulate; Fig. 1a, b), (2) laterally compressed (Fig. 1c, d), (3) pouched (hypocyrtoid; Fig. 1e, f), and (4) urn-shaped (urceolate; Fig. 1g, h). Anthers were assigned as poricidally dehiscent (Fig. 2a) or longitudinally dehiscent (Fig. 2b). Some species with urn-shaped and laterally compressed corollas present an initial poricidal stage, which then rapidly develops into longitudinal dehiscence; we coded these as longitudinal in the character state reconstructions. Each taxon was carefully evaluated in the field or from herbarium specimens. All corolla shapes were photographed with images readily available on the first author’s website (www.gesneriads.ua.edu).

Drymonia is strongly supported as nesting in a clade with Glossoloma, Columnea, Alloplectus, and Neomortonia rosea (Clark et al. 2006, 2013). Compared to Drymonia, the flowers of Columnea, Glossoloma, and Alloplectus have a more elongate tubular or bilabiate corolla and are not readily assigned to one of the four shapes outlined above. The corolla shapes of Neomortonia rosea, Pachycaulos nummularia, and Corytoplectus congestus could be assigned to one of the above corolla shapes, but it would be conjectural to evaluate them in a phylogenetic context here because they are not the focus of the present study and including them would require extensive taxon sampling of additional outgroups. Characters were unordered (Fitch 1971) and character evolution analyses were performed in Mesquite, version 4.08 (Maddison and Maddison 2011) where parsimony optimization using the unordered states assumption was implemented. The character states were mapped onto the Bayesian consensus tree obtained in the molecular analyses (Fig. 3).
Fig. 3

Majority rule Bayesian inference tree shown with support values indicated for Maximum likelihood bootstrap and parsimony bootstrap. Based on three molecular markers (ITS, ETS and trnK-matK spacer). Numbers correspond to Bayesian posterior probabilities/maximum likelihood bootstrap/parsimony bootstrap. Nodes that collapse in the strict consensus tree are indicated by (“*”) at the base of the branch. Independent origins of longitudinal anther dehiscence (ingroup only) is shown by diagrams to right of taxon names. All other in-group taxa have poricidal anther dehiscence

Test of incongruence

The incongruence length difference test (ILD: Farris et al. 1994) was performed as the partition homogeneity test implemented in PAUP*4.0 b10 (Swofford 2003) with 1,000 bootstrap replicates (using a heuristic search, simple addition, and no branch swapping). The cpDNA, ITS, and ETS were each treated as separate partitions. As the ILD has been shown to indicate incongruence where none exists (Dolphin et al. 2000; Yoder et al. 2001; Barker and Lutzoni 2002; Dowton and Austin 2002), bootstrap analyses were performed on each partition separately to assess areas of conflict and to determine if any conflict was strongly supported (Mason-Gamer and Kellogg 1996; Seelanen et al. 1997).

Phylogenetic analyses

The parsimony analysis was performed to completion using a two stage heuristic search in PAUP* 4.0 (Swofford 2003). The first stage of the analysis was done using the following settings: 1,000 random addition cycles, holding 10 trees of equal length at each step; tree bisection-reconstruction (TBR) branch swapping with no more than 10 trees saved for each rep; MULTREES option not in effect. The second stage of the analysis was performed on all trees in memory with the same settings, but with the MULTREES option in effect. Two searches with altered settings did not find shorter trees; these included a search with 10 random addition cycles holding 1,000 trees at each step, and one with 1,000 random addition cycles holding 100 trees at each step.

Additional tree searches were done using the parsimony ratchet analysis with NONA (Goloboff 1999) and Winclada (Nixon 2002). Ten separate tree searches were conducted using the following settings: 200 iterations per search, one tree held for each iteration, 132 characters sampled (10 % of the total), and amb = poly-(only considers unambiguous support). The total evidence analysis was swapped to completion, but analyses of individual datasets were limited to 100,000 trees. Multiple ratchet searches were performed in WinClada as suggested by Nixon (1999) since the ratchet option can sometimes get stuck on suboptimal “islands” and it is therefore better to perform more separate searches with fewer iterations than one larger search with more iterations.

Clade robustness was evaluated in PAUP* with a bootstrap analysis (Felsenstein 1985). We used 1,000 heuristic bootstrap replicates with the following settings: 10 random addition cycles; tree bisection-reconstruction (TBR) branch swapping with no more than 10 trees saved for each replicate.

The parsimony analyses and clade support were evaluated for each individual dataset (ETS, ITS, trnK-matK), a combined molecular dataset, and a total evidence analysis. Conflict between datasets was evaluated by comparing incongruence of strongly supported clades from individual datasets (e.g., ITS vs. ETS; ITS vs. trnK-matK; and nrDNA vs. trnK-matK).

Bayesian inference analyses were conducted using MrBayes 3.2.2. (Ronquist et al. 2012) implemented in the CIPRES web portal (http://www.phylo.org/; Miller et al. 2009). Models of nucleotide substitution for each partition were assessed with JModeltest 2.1.3. (Darriba et al. 2012). The best-fit models, selected using the Akaike information criterion (AIC), were TPM2uf + G for ETS, GTR + I + G for ITS, and GTR + G for trnK-matK. Two independent Markov Chain Monte Carlo (MCMC) analyses, each with 20 million generations, were run. The analyses were started from random trees, sampling each 1000th generation. Convergence of the two independent MCMC runs was analyzed in Tracer v1.5 (Rambaut and Drummond 2007) and the first 25 % of trees were discarded as burn in before the posterior distribution was sampled. The remaining trees from both runs were used to construct a majority-rule consensus tree, and the posterior probability (PP) values were used as indicators of robustness.

Maximum Likelihood analyses were conducted in RaxML v7.6.6 (Stamatakis 2006; Stamatakis et al. 2008) implemented in CIPRES web portal (http://www.phylo.org/; Miller et al. 2009), and clade support was estimated by performing bootstrap with 1,000 replicates.

Ancestral reconstruction

Standard parsimony character optimization was implemented in Mesquite 2.75 (Maddison and Maddison 2011) to reconstruct the ancestral state for corolla tube shape and anther dehiscence. For the reconstruction we used the BI consensus topology derived from the total evidence data set, and considered the characters unordered and equally weighted. This method finds the ancestral states that minimize the number of changes required to explain the distribution of character states observed on the phylogeny (Maddison and Maddison 2011).

Flower visitation and pollination modes

To identify floral visitors, flowers from the six species of Drymonia were videotaped with Sony Digital Camcorders. With three cameras, we were able to videotape three different flowers (from three different plants) at any given time. Each camera was placed on a tripod approximately 2 m away from the focal flower and covered with a modified 2-L plastic bottle to protect it from rain. Flower visitor surveys were selected to focus on at least one species for each of the four different corolla shapes of Drymonia (Table 2). Examples of each corolla shape are shown in Fig. 1 and the coding of corolla shape for the entire matrix is provided in Table 1. The following six Drymonia species were videotaped: Drymonia affinis (Mansf.) Wiehler and D. hoppii (Mansf.) Wiehler (laterally compressed); D. dodsonii (Wiehler) J.L. Clark and D. tenuis (Benth.) J.L. Clark (pouched); D. ecuadorensis Wiehler (bell-shaped); and D. urceolata Wiehler (urn-shaped). A total of 164 h of filming (16–48 h per taxon) were performed in 2009, 2010, and 2011 in three localities in Ecuador. Visits were considered legitimate when the visitor entered the corolla and made contact with the anthers (Table 1).
Table 1

Pollinator survey information outlining taxon, distribution, source of observation (literature citation if not generated for this study), and corolla shape

Taxon

Distribution

Source of observation

Pollinator

Corolla shape

Ingroup (50)

 Drymonia affinis (Mansf.) Wiehler

Central and South America

This study (Ecuador)

Bird

Laterally compressed

 Drymonia alloplectoides Hanst.

Central and South America

  

Campanulate

 Drymonia ambonensis (L.E. Skog) J.L. Clark

Central America

  

Hypocyrtoid

 Drymonia atropurpurea Clavijo and J.L. Clark

Ecuador and Colombia

  

Campanulate

 Drymonia brochidodroma Wiehler

Ecuador and Colombia

  

Campanulate

 Drymonia candida Hanst.

Bolivia, Brazil, Colombia, Ecuador, Peru

  

Campanulate

 Drymonia chiribogana Wiehler

Ecuador

  

Campanulate

 Drymonia coccinea (Aubl.) Wiehler

South America

  

Laterally compressed

 Drymonia coccinea (Aubl.) Wiehler

South America

  

Laterally compressed

 Drymonia conchocalyx Wiehler

Central America

Feinsinger et al. (1987)

Bird

Laterally compressed

 Drymonia coriacea (Oerst. ex Hanst.) Wiehler

Central and South America

  

Hypocyrtoid

 Drymonia crassa C.V. Morton

Colombia and Venezuela

  

Campanulate

 Drymonia crenatiloba (Mansf.) Wiehler

Ecuador

  

Urceolate

 Drymonia decora J.R. Clark and J.L. Clark

Costa Rica

  

Campanulate

 Drymonia dodsonii (Wiehler) J.L. Clark

Colombia and Ecuador

This study (Ecuador)

Bird

Hypocyrtoid

 Drymonia doratostyla (Leeuwenb.) Wiehler

Bolivia and Peru

  

Laterally compressed

 Drymonia ecuadorensis Wiehler

Ecuador

This study (Ecuador)

Bee

Campanulate

 Drymonia foliacea (Rusby) Wiehler

South America

  

Campanulate

 Drymonia folsomii L.E. Skog

Costa Rica and Panama

  

Campanulate

 Drymonia hoppii (Mansf.) Wiehler

South America

This study (Ecuador)

Bird

Laterally compressed

 Drymonia ignea J.L. Clark

Ecuador

  

Urceolate

 Drymonia killipii Wiehler

Colombia and Ecuador

  

Campanulate

 Drymonia laciniosa Wiehler

Ecuador and Peru

  

Campanulate

 Drymonia lanceolata (Hanst.) C.V. Morton

Central and South America

  

Campanulate

 Drymonia longifolia Poepp.

South America

  

Campanulate

 Drymonia macrantha (Donn. Sm.) D.N. Gibson

Central America

  

Campanulate

 Drymonia macrophylla (Oerst.) H.E. Moore

Central and South America

  

Campanulate

 Drymonia microphylla Wiehler

Panama

  

Campanulate

 Drymonia mortoniana (Oerst.) H.E. Moore

Costa Rica

  

Campanulate

 Drymonia multiflora (Oerst. ex Hanst.) Wiehler

Central America

Stiles and Freeman (1993)

Bird

Laterally compressed

 Drymonia ovatifolia J.L. Clark

Central and South America

Enrique (1998)

Bee

Campanulate

 Drymonia parviflora Hanst.

Costa Rica and Panama

  

Campanulate

 Drymonia peltata (Oliver) H.E. Moore

Costa Rica

  

Campanulate

 Drymonia pendula (Poepp.) Wiehler

South America

  

Laterally compressed

 Drymonia pilifera Wiehler

Costa Rica and Panama

  

Campanulate

 Drymonia pulchra Wiehler

Ecuador

  

Campanulate

 Drymonia rhodoloma Wiehler

Colombia and Ecuador

  

Campanulate

 Drymonia rubra C.V. Morton

Costa Rica and Panama

  

Campanulate

 Drymonia rubripilosa Wiehler

Costa Rica

  

Laterally compressed

 Drymonia serrulata (Jacq.) Mart.

Central and South America

Steiner (1985)

Bee

Campanulate

 Drymonia stenophylla (Donn. Sm.) H.E. Moore

Central America

  

Campanulate

 Drymonia strigosa (Oerst.) Wiehler

Central America

Enrique (1998)

Bee

Campanulate

 Drymonia submarginalis Gómez-Laur. and M.M. Chavarría

Costa Rica and Nicaragua

  

Campanulate

 Drymonia tenuis (Benth.) J.L. Clark

Colombia and Ecuador

This study

Dziedzioch et al. (2003)

Bird

Hypocyrtoid

 Drymonia teuscheri (Raymond) J.L. Clark

South America

Dziedzioch et al. (2003)

Bird

Hypocyrtoid

 Drymonia turrialvae Hanst.

Central and South America

Dressler (1968)

Bee

Campanulate

 Drymonia urceolata Wiehler

South America

This study

Dziedzioch et al. (2003)

Bird

Urceolate

 Drymonia warszewicziana Hanst.

Central and South America

  

Campanulate

 Drymonia sp. 1—JLC 6863

Venezuela

  

Urceolate

 Drymonia sp. 2—JLC 8366

Ecuador

  

Urceolate

Results

DNA sequencing and alignment

Amplifications were successful for all regions for all individuals, with some exceptions (Appendix 1). Of the three regions sampled, ITS provided the most parsimony-informative substitutions (142 or 51 % of the combined three regions; Appendix 2). The trnK-matK spacer provided the least number of parsimony-substitutions (20 or 7.2 % of the combined three regions; Appendix 2). The aligned matrix for the full analysis contained 1,689 basepairs; of these, 1,176 were constant and 235 were uninformative. The outgroups of the analysis contributed 53 of the 278 parsimony informative substitutions. There were no ambiguously aligned sites excluded from the analysis. Sequence divergence in the ingroup was relatively conserved and no informative indels required scoring. The complete list of gene regions and statistics is provided in Appendix 2.

Tests of incongruence

The incongruence length difference (ILD) tests found no significant discordance between the ITS and ETS datasets (P = 0.100) or between the combined nrDNA (ITS concatenated with ETS) and trnK-matK (cpDNA) datasets (P = 1.00 and 0.500 respectively). Therefore, we combined these three datasets in a total evidence analysis.

Phylogenetic analyses

The parsimony analysis for the combined dataset resulted in 208 trees of length 1,230 (CI = 0.55, RI = 0.66, and RC = 0.36). The strict consensus of these trees is congruent with the tree shown in Fig. 3. Minor exceptions include polytomies for some of the crown clades. Clades that collapse in the strict consensus tree that are resolved in ML or BI are indicated by asterisks in Fig. 3.

Pollination modes

Results of videotaping were consistent with an association between constricted corollas and hummingbird pollination (Table 1). The species with bell-shaped corollas, D. ecuadorensis, was visited solely by euglossine bees (Fig. 2e, f) at a rate of 3.29 visits per hour. No bee visits were recorded to the five species with constricted openings (urn-shaped, pouched, and laterally compressed), while hummingbirds visited these flowers at rates of 0.04 to 0.36 visits per hour. Literature surveys further support this pattern; bees have been observed pollinating the bell-shaped D. aciculata Wiehler (observations by Dressler, cited in Steiner 1985), D. serrulata (Steiner 1985), D. strigosa (Enrique 1998), D. turrialvae Hanst. (Dressler 1968), and D. ovatifolia J.L. Clark (as Nautilocalyx panamensis (Seem.) Seem.; Enrique 1998), while hummingbirds have been observed pollinating the laterally compressed D. conchocalyx Hanst. (Feinsinger et al. 1987) and D. multiflora (Oerst. ex Hanst.) Wiehler (Stiles and Freeman 1993) as well as the pouched D. teuscheri (Raymond) J.L. Clark (Dziedzioch et al. 2003).

Discussion

Bayesian, maximum likelihood, and parsimony analyses all resulted in congruent topologies with high levels of node support, and corolla shapes were unambiguously optimized on the inferred phylogeny for Drymonia (Fig. 3). This optimization strongly supports the convergent evolution of corolla shapes in Drymonia, with multiple independent origins of the three types of constricted corolla tubes from campanulate ancestors (cf., legend in Fig. 3).

Traditional circumscription of Drymonia

The results presented here are congruent with previous phylogenetic studies that support the non-monophyly of traditional Drymonia (Clark et al. 2006; 2012). The traditional circumscription of Drymonia is artificial because it relies on corolla shape and anther dehiscence, traits that reflect pollination syndrome rather than evolutionary relatedness. Below we discuss shifts from campanulate corollas to each of the three types of constricted corollas in greater detail.

Pollination modes

Results of the pollinator observations and literature survey support the hypothesis that corolla constriction evolved multiple times in association with bird pollination. Species with the ancestral bell-shaped corollas are pollinated by bees, while species with pouched, urn-shaped, or laterally compressed corollas are pollinated by hummingbirds (Table 1). Euglossine and Epicharis bees found to pollinate bell-shaped Drymonia have thorax widths of 5–10 mm (Steiner 1985). The narrow openings of pouched (3.0–4.2 mm), urn-shaped (4.0–4.8 mm), and laterally compressed flowers (3.0–4.5 mm at the top of the throat and 1.0–2.5 mm near the base of the throat) effectively prevent access by bees to the pollen and nectar rewards of these flowers, while visitation results demonstrate they are not narrow enough to serve as barriers to hummingbird bills. Thus narrow corollas in Drymonia serve as a ‘barrier trait’, preventing or reducing visitation by bees and other insects.

We hypothesize that selection to reduce loss of pollen and nectar to insects was the main driver of evolutionary narrowing of corollas. Alternatively, constricted corollas may have evolved primarily to better guide hummingbird bills and increase precision and consistency of pollen placement. Temeles et al. (2002) demonstrated that flower width is an important factor when considering the coevolution and specialization of hummingbirds and flowers (also see Grant and Temeles 1992, Muchhala 2007). Future experimental work would be useful in evaluating the relative importance of bill-corolla fit vs. insect exclusion in the repeated evolution of constricted corollas across Drymonia.

Along with corolla shape, the shifts in primary pollinators across Drymonia are also associated with changes in anther dehiscence. The ancestral condition of poricidal dehiscence (“salt-shaker anthers”) is independently lost in six clades, each of which is also associated with constricted corollas and hummingbird pollination (Fig. 3). For some species in these clades, including D. urceolata (Fig. 2c, d), D. rubripilosa Kriebel and D. multiflora, an initial poricidal stage can be detected before anthers fully rupture via longitudinal dehiscence (prior to anthesis). The presence of a vestigial poricidal dehiscence stage further supports the conclusion that these represent recent shifts from ancestors with poricidal anthers. We suggest that the shifts in anther dehiscence represent adaptations to more effectively present pollen to each pollinator type. To maximize male fitness, selection should favor placing specific amounts of pollen on each visitor, with the optimal dose size depending on visitation frequency and visitor behavior (Harder and Thomson 1989; Thomson and Thomson 1992; Castellanos et al. 2006). Thus, because bees tend to have high visit rates and frequently groom excess pollen off their bodies (Harder 1990), bee-pollinated plants should present their pollen in numerous small doses to as many bees as possible, while for hummingbirds, pollen would be more effectively dispersed in fewer, larger doses (Thomson et al. 2000, Castellanos et al. 2006). In line with this scenario, the hummingbird-adapted Drymonia species with narrow corollas have much lower visitation rates than the bee-adapted D. ecuadorensis (Table 2). The ‘salt-shaker’ poricidal anthers in Drymonia likely evolved to slowly dose pollen to multiple bees, while also avoiding ‘over-dosing’ individual bees and triggering grooming behavior. Longitudinally-dehiscent anthers present few, larger doses to the infrequent hummingbird visitors.
Table 2

Results from videotaping flower visitors to six Drymonia species representing four corolla shapes

Drymonia species

Corolla shape

Number of plants filmed

Total hours filmed

Total visits

Visits per hour

Bees

Birds

Bees

Birds

D. affinis

Laterally compressed

3

16

0

1

0

0.06

D. dodsonii

Hypocyrtoid

4

47.7

0

17

0.00

0.36

D. ecuadorensis

Campanulate

3

32.5

107

0

3.29

0.00

D. hoppii

Laterally compressed

3

17

0

1

0

0.06

D. tenuis

Hypocyrtoid

5

26.7

0

1

0.00

0.04

D. urceolata

Urceolate

4

24

0

3

0

0.13

Pouched corollas

Pouched corollas (Fig. 1e) have three independent origins within Drymonia (Fig. 3); one in Central America and two in the northern Andes of South America (Fig. 3). Additional independent origins of pouched corollas occur in other New World genera such as Besleria, Columnea, Gasteranthus, Nematanthus, Pachycaulos, and Paradrymonia.

The possible adaptive function of the enlarged pouches at the base of the corolla is unclear. One possibility is that they may serve as an ‘overflow chamber’ for the accumulation of nectar (sensu Wolf and Stiles 1989), however we consider this unlikely as we have never found nectar in the pouches when flowers were dissected in the field. Wiehler (1983) suggested that pouches serve as a “target enlargement;” an increased visual display that aids in long-distance attraction of hummingbirds. Many Drymonia inflorescences include brightly-colored bracts, thus the pouches could also function to create a ‘bi-colored display’ (sensu Willson and Thompson 1982).

Urn-shaped corollas

There are three independent origins of urn-shaped corollas (Fig. 1g, h) in Drymonia and they are all located in South America (Fig. 3). It is noteworthy that urn-shaped corollas share recent common ancestors with bell-shaped taxa (e.g., Drymonia turrialvae and D. foliaceae + D. ovatifolia) and pouched taxa (e.g., D. dodsonii, D. tenuis, and D. teuscheri). This study sampled nearly all of the currently known species of Drymonia with urn-shaped corollas, including two species that are potentially new to science (JLC 8366 and JLC 6863).

Laterally compressed corollas

Laterally compressed corollas (Fig. 1c, b) within Drymonia have three independent origins; one in South America and two in Central America (Fig. 3). The clade that includes Drymonia rubripilosa and D. multiflora comprises species from Central America, and D. conchocalyx is also from Central America. The clade that includes Drymonia pendula, D. doratostyla, D. coccinea, D. hoppii and D. affinis comprises species from the western slopes of the northern Andes and western Amazonia. The South American clade of laterally compressed flowers is the only example in the genus where constricted corollas have retained the ancestral condition of poricidal anther dehiscence. All of the species with laterally compressed corollas are epiphytic and most have brightly colored bracts relative to sister-clades. Laterally compressed corollas are also found in most species of Glossoloma (Clark 2009) and in Nematanthus, for which hummingbird pollination is well documented (Franco and Buzato 1992; Sazima et al. 1995; Buzato et al. 2000; San Martin-Gajardo and Santana Vianna 2010).

Notes

Acknowledgments

The authors would like to thank Alex Monro from The Natural History Museum in London (BM) for sharing leaf material of Gesneriaceae. This study was supported by grants from the National Science Foundation (DEB-0949270 and DEB-0841958 to JLC). Fieldwork was greatly facilitated by the following undergraduate students from The University of Alabama: Cassandra L. Coleman, Seema Kumar, and Laura A. Frost. Lucas McDonald from Hillcrest High School (Tuscaloosa, AL) also helped in collecting data during a 2011 expedition to Ecuador. Murray Cooper and Richard W. Dunn are gratefully acknowledged for contributing images. We express our appreciation to two anonymous reviewers for useful comments that improved an earlier version of the manuscript.

References

  1. Baldwin BG, Markos S (1998) Phylogenetic utility of the external transcribed spacer (ETS) of 18S–26S rDNA: congruence of ETS and ITS trees of Calycadenia (Compositae). Mol Phylogenet Evol 10:449–463CrossRefPubMedGoogle Scholar
  2. Baldwin BG, Sanderson MJ, Porter JM, Wojciechowski MF, Campbell CS, Donoghue MJ (1995) The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Ann Missouri Bot Gard 82:247–277CrossRefGoogle Scholar
  3. Barker FK, Lutzoni FM (2002) The utility of the incongruence length difference test. Syst Biol 51:625–637CrossRefPubMedGoogle Scholar
  4. Beardsley PM, Olmstead RG (2002) Redefining Phrymaceae: the placement of Mimulus, tribe Mimuleae, and Phryma. Am J Bot 89:1093–1102CrossRefPubMedGoogle Scholar
  5. Buchmann SL (1983) Buzz pollination in angiosperms. In: Jones CE, Little RJ (eds) Handbook of experimental pollination biology. Van Nostrand Reinhold, New York, pp 73–113Google Scholar
  6. Burtt BL, Wiehler H (1995) Classification of the family Gesneriaceae. Gesneriana 1:1–4Google Scholar
  7. Buzato S, Sazima M, Sazima I (2000) Hummingbird-pollinated floras at three Atlantic forest sites. Biotropica 32:824–841CrossRefGoogle Scholar
  8. Castellanos MC, Wilson P, Keller SJ, Wolfe AD, Thomson JD (2006) Anther evolution: pollen presentation strategies when pollinators differ. Am Nat 167:288–296CrossRefPubMedGoogle Scholar
  9. Clark JL (2005) A Monograph of Alloplectus (Gesneriaceae). Selbyana 25:182–209Google Scholar
  10. Clark JL (2009) The systematics of Glossoloma (Gesneriaceae). Syst Bot Monogr 88:1–128Google Scholar
  11. Clark JL, Herendeen PS, Skog LE, Zimmer EA (2006) Phylogenetic relationships and generic boundaries in the tribe Episcieae (Gesneriaceae) inferred from nuclear, chloroplast, and morphological data. Taxon 55:313–336CrossRefGoogle Scholar
  12. Clark JL, Funke MM, Duffy AM, Smith JF (2012) Phylogeny of a Neotropical clade in the Gesneriaceae: more tales of convergent evolution. Int J Plant Sci 173:894–916CrossRefGoogle Scholar
  13. Darriba D, Taboada GL, Doallo R, Posada D (2012) jModelTest 2: more models, new heuristics and parallel computing. Nat Methods 9:772CrossRefPubMedGoogle Scholar
  14. Dolphin K, Belshaw R, Orme CDL, Donald L, Quicke J (2000) Noise and incongruence: interpreting results of the incongruence length difference test. Mol Phylogenet Evol 17:401–406CrossRefPubMedGoogle Scholar
  15. Dowton M, Austin AD (2002) Increased congruence does not necessarily indicate increased phylogenetic accuracy: the behavior of the incongruence length difference test in mixed-model analyses. Syst Biol 51:19–31CrossRefPubMedGoogle Scholar
  16. Dressler RL (1968) Pollination by euglossine bees. Evolution 22:202–210CrossRefGoogle Scholar
  17. Dziedzioch C, Stevens AD, Gottsberger G (2003) The hummingbird plant community of a tropical montane rain forest in southern Ecuador. Plant Biol 5:331–337CrossRefGoogle Scholar
  18. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797CrossRefPubMedCentralPubMedGoogle Scholar
  19. Enrique ERA (1998) Resources foraged by Euglossa atroveneta (Apidae: Euglossinae) at Union Juárez, Chiapas, Mexico. A palynological study of larval feeding. Apidologie 29:347–359CrossRefGoogle Scholar
  20. Faegri K, van der Pijl L (1979) The principles of pollination ecology. Pergamon, OxfordGoogle Scholar
  21. Farris JS, Källersjö M, Kluge AG, Bult C (1994) Testing significance of incongruence. Cladistics 10:315–319CrossRefGoogle Scholar
  22. Feinsinger P, Beach JH, Linhart YB, Rusby WH, Murray KG (1987) Disturbance, pollinator predictability, and pollination success among Costa Rican cloud forest plants. Ecology 68:1294–1305CrossRefGoogle Scholar
  23. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791CrossRefGoogle Scholar
  24. Fenster CB, Armbruster WS, Wilson P, Dudash MR, Thomson JD (2004) Pollination syndromes and floral specialization. Annu Rev Ecol Evol S 35:375–403CrossRefGoogle Scholar
  25. Fitch WM (1971) Toward defining the course of evolution: minimum change for a specific tree topology. Syst Zool 20:406–415CrossRefGoogle Scholar
  26. Franco ALM, Buzato S (1992) An assemblage of hummingbird-pollinated flowers in a montane forest in southeastern Brazil. Bot Acta 109:149–160Google Scholar
  27. Fritsch K (1893–1994) Gesneriaceae. In: Engler A, Prantl K (eds) Die Nat. Pflanzenfam., vol 4 (3b). Engelmann, Leipzig, Germany, pp 133–185Google Scholar
  28. Goloboff P (1999) NONA, version 2. Published by the author, TucumánGoogle Scholar
  29. Grant KA, Grant V (1968) Hummingbirds and their flowers. Columbia University Press, New YorkGoogle Scholar
  30. Grant V, Temeles EJ (1992) Foraging ability of rufous hummingbirds on hummingbird flowers and hawkmoth flowers. Proc Natl Acad Sci 89(20):9400–9404CrossRefPubMedCentralPubMedGoogle Scholar
  31. Hanstein J (1854) Die Gesneraceen des Königlichen Herbariums und der Gärten zu Berlin, nebst Beobachtungen über die Familie im Ganzen I. Abschnitt. Linnaea 26:145–216Google Scholar
  32. Hanstein J (1856) Die Gesneraceen des Königlichen Herbariums und der Gärten zu Berlin, nebst monographischer Uebersicht der Familie im Ganzen, II. Abschnitt. Gattungen und Arten. Erstes Stück. Die Niphaeen und Achimeneen. Linnaea 27:693–785Google Scholar
  33. Hanstein J (1859) Die Gesneraceen des Königlichen Herbariums und der Garten zu Berlin, nebst monographischer Uebersicht der Familie im Ganzen, II. Abschnitt. Gattungen und Arten. Zweites Stück. Die Brachylomateen. Linnaea 29:497–592Google Scholar
  34. Hanstein J (1865) Die Gesneraceen des Königlichen Herbariums und der Gärten zu Berlin, nebst monographischer Uebersicht der Familie im Ganzen, II. Abschnitt. Gattungen und Arten. Drittes Stück. Die Eugesnereen, Rhytidophylleen, und Beslerieen. Linnaea 34:225–462Google Scholar
  35. Harder LD (1990) Pollen removal by bumble bees and its implications for pollen dispersal. Ecology 71:1110–1125CrossRefGoogle Scholar
  36. Harder L, Thomson JD (1989) Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. Am Nat 133:325–334CrossRefGoogle Scholar
  37. Ivanina LI (1965) Application of the carpological method to the taxonomy of Gesneriaceae. Notes Roy Bot Gard Edinburgh 26:383–402Google Scholar
  38. Ivanina LI (1967) The family Gesneriaceae (The Carpological Review). Komarov Bot. Inst, Leningrad, USSR, 126 ppGoogle Scholar
  39. Johnson LA, Soltis DE (1995) Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Ann Missouri Bot Gard 82:149–175CrossRefGoogle Scholar
  40. Maddison WP, Maddison DR (2011) Mesquite: a modular system for evolutionary analysis. Version 2.75. http://mesquiteproject.org. Accessed March 2014
  41. Martius CFP (1829) Gesneriaceae. Nova Genera et Species Plantarum, vol 3. Impensis auctoris, Munich, pp 27–73Google Scholar
  42. Mason-Gamer RJ, Kellogg EA (1996) Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Syst Biol 45:524–545CrossRefGoogle Scholar
  43. Miller MA, Holder MT, Vos R, Midford PE, Liebowitz T, Chan L, Hoover P, Warnow T (2009) The CIPRES Portals. CIPRES. 2009–08–04. http://www.phylo.org/sub_sections/portal. Accessed March 2014
  44. Möller M, Clark JL (2013) The state of molecular studies in the family Gesneriaceae. Selbyana 31:95–125Google Scholar
  45. Moore HE (1955) Drymonia macrophylla. Baileya 3:109–112Google Scholar
  46. Muchhala N (2007) Adaptive tradeoff in floral morphology mediates specialization for flowers pollinated by bats and hummingbirds. Am Nat 169:494–504CrossRefPubMedGoogle Scholar
  47. Nixon KC (1999) The Parsimony Ratchet, a new method for rapid parsimony analysis. Cladistics 15:407–414CrossRefGoogle Scholar
  48. Nixon KC (2002) WinClada, version 1.00.08. Published by the author, Ithaca, New YorkGoogle Scholar
  49. Rambaut A, Drummond AJ (2007) Tracer v1.4: MCMC trace analyses tool. http://beast.bio.ed.ac.uk/Tracer. Accessed October 25, 2013
  50. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol 61:539–542CrossRefPubMedCentralPubMedGoogle Scholar
  51. San Martin-Gajardo I, Santana Vianna JR (2010) Pollination of Namatanthus brasiliensis: an epiphytic Gesneriaceae endemic to the southeastern Atlantic forests of Brazil. Selbyana 30:216–220Google Scholar
  52. Sazima I, Buzato S, Sazima M (1995) An assemblage of hummingbird-pollinated flowers in a montane forest in southeastern Brazil. Bot Acta 109:149–160CrossRefGoogle Scholar
  53. Seelanen T, Schnabel A, Wendel JF (1997) Congruence and consensus in the cotton tribe (Malvaceae). Syst Bot 22:259–290CrossRefGoogle Scholar
  54. Simmons MP, Ochoterena H (2000) Gaps as characters in sequence-based phylogenetic analyses. Syst Biol 49:369–381CrossRefPubMedGoogle Scholar
  55. Smith JF, Atkinson S (1998) Phylogenetic analysis of the tribes Gloxinieae and Gesnerieae (Gesneriaceae): data from ndhF sequences. Selbyana 19:122–131Google Scholar
  56. Smith JF, Clark JL (2013) Molecular phylogenetic analyses reveal undiscovered monospecific Genera in the tribe Episcieae (Gesneriaceae). Syst Bot 38:451–463CrossRefGoogle Scholar
  57. Smith JF, Wolfram JC, Brown KD, Carroll CL, Denton DS (1997) Tribal relationships in the Gesneriaceae: evidence from DNA sequences of the chloroplast gene ndhF. Ann Missouri Bot Gard 84:50–66CrossRefGoogle Scholar
  58. Smith JF, Kresge M, Møller M, Cronk QCB (1998) The African violets (Saintpaulia) are members of Streptocarpus subgenus Streptocarpella (Gesneriaceae): combined evidence from chloroplast and nuclear ribosomal genes. Edinb J Bot 55:1–11CrossRefGoogle Scholar
  59. Smith JF, Draper SB, Hileman LC, Baum DA (2004a) A phylogenetic analysis within tribes Gloxinieae and Gesnerieae (Gesnerioideae: Gesneriaceae). Syst Bot 29:947–958CrossRefGoogle Scholar
  60. Smith JF, Draper SB, Hileman LC, Baum DA (2004b) Evolution of GCYC, a Gesneriaceae homolog of CYCLOIDEA, within Gesnerioideae (Gesneriaceae). Mol Phylog Evol 31:765–779CrossRefGoogle Scholar
  61. Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22:2688–2690CrossRefPubMedGoogle Scholar
  62. Stamatakis A, Hoover P, Rougemont J (2008) A fast bootstrapping algorithm for the RAxML web-servers. Syst Biol 5:758–771CrossRefGoogle Scholar
  63. Steiner KE (1985) The role of nectar and oil in the pollination of Drymonia serrulata (Gesneriaceae) by Epicharis bees (Anthophorideae) in Panama. Biotropica 17:217–229CrossRefGoogle Scholar
  64. Stiles FG, Freeman CE (1993) Patterns in floral nectar characteristics of some bird-visited plant species from Costa Rica. Biotropica 25:191–205CrossRefGoogle Scholar
  65. Suh Y, Thien LB, Reeve HE, Zimmer EA (1993) Molecular evolution and phylogenetic implications of internal transcribed spacer sequences of ribosomal DNA in Winteraceae. Am J Bot 80:1042–1055CrossRefGoogle Scholar
  66. Swofford DL (2003) PAUP*: phylogenetic analysis using parsimony (* and other methods), Version 4.0. Sinauer Associates, Sunderland, MassachusettsGoogle Scholar
  67. Temeles EJ, Linhart RB, Masonjones M, Masonjones HD (2002) The role of flower width in hummingbird bill length-flower length relationships. Biotropica 34:68–80Google Scholar
  68. Thiers B (2013) [continuously updated]: Index herbariorum: a global directory of public herbaria and associated staff. New York Botanical Garden. http://sweetgum.nybg.org/ih/. Accessed March 2013
  69. Thomson JD, Thomson BA (1992) Pollen presentation and viability schedules in animal-pollinated plants: consequences for reproductive success. In: Wyatt R (ed) Ecology and evolution of plant reproduction: new approaches. Chapman and Hall, New York, pp 1–24Google Scholar
  70. Thomson JD, Wilson P, Valenzuela M, Malzone M (2000) Pollen presentation and pollination syndromes, with special reference to Penstemon. Plant Species Biol 15:11–29CrossRefGoogle Scholar
  71. Weber A, Clark JL, Möller M (2013) A new formal classification of Gesneriaceae. Selbyana 31:68–94Google Scholar
  72. White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T (eds) PCR protocols: a guide to methods and applications. Academic Press, San Diego, pp 315–322CrossRefGoogle Scholar
  73. Wiehler H (1973) 100 Transfers from Alloplectus and Columnea (Gesneriaceae). Phytologia 27:309–329Google Scholar
  74. Wiehler H (1983) A synopsis of the Neotropical Gesneriaceae. Selbyana 6:1–219Google Scholar
  75. Willson MF, Thompson JN (1982) Phenology and ecology of color in bird-dispersed fruits, or why some fruits are red when they are “green”? Can J Bot 60:701–713CrossRefGoogle Scholar
  76. Wolf LL, Stiles FJ (1989) Adaptations for the ‘Fail-safe’ pollination of specialized ornithophilous flowers. Am Midl Nat 121:1–10CrossRefGoogle Scholar
  77. Yoder AD, Irwin JA, Payseur BA (2001) Failure of the ILD to determine data combinability for slow loris phylogeny. Syst Biol 50:408–424CrossRefPubMedGoogle Scholar
  78. Zimmer EA, Roalson EH, Skog LE, Boggan JK, Idnurm A (2002) Phylogenetic relationships in the Gesnerioideae (Gesneriaceae) based on nrDNA ITS and cpDNA trnL-F and trnE-T spacer region sequences. Am J Bot 89:296–311CrossRefPubMedGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • John L. Clark
    • 1
  • Laura Clavijo
    • 1
  • Nathan Muchhala
    • 2
  1. 1.Department of Biological SciencesThe University of AlabamaTuscaloosaUSA
  2. 2.Department of BiologyUniversity of Missouri-St. LouisSt. LouisUSA

Personalised recommendations