Hummingbird Flowers Research Articles

Evolutionary convergence on hummingbird pollination in Neotropical Costus provides insight into the causes of pollinator shifts.

The evolution of hummingbird pollination is common across angiosperms throughout the Americas, presenting an opportunity to examine convergence in both traits and environments to better understand how complex phenotypes arise. Here we examine independent shifts from bee to hummingbird pollination in the Neotropical spiral gingers (Costus) and address common explanations for the prevalence of transitions from bee to hummingbird pollination. We use floral traits of species with observed pollinators to predict pollinators of unobserved species and reconstruct ancestral pollination states on a well-resolved phylogeny. We examine whether independent transitions evolve towards the same phenotypic optimum and whether shifts to hummingbird pollination correlate with elevation or climate. Traits predicting hummingbird pollination include small flower size, brightly colored floral bracts and the absence of nectar guides. We find many shifts to hummingbird pollination and no reversals, a single shared phenotypic optimum across hummingbird flowers, and no association between pollination and elevation or climate. Evolutionary shifts to hummingbird pollination in Costus are highly convergent and directional, involve a surprising set of traits when compared with other plants with analogous transitions and refute the generality of several common explanations for the prevalence of transitions from bee to hummingbird pollination.

DNA Barcoding Reveals Generalization and Host Overlap in Hummingbird Flower Mites: Implications for the Mating Rendezvous Hypothesis.

AbstractHummingbird flower mites are assumed to monopolize single host plant species owing to sexual selection for unique mating rendezvous sites. We tested the main assumption of the mating rendezvous hypothesis-extreme host specialization-by reconstructing interactions among tropical hummingbird flower mites and their host plants using DNA barcoding and taxonomic identifications. We collected 10,654 mites from 489 flowers. We extracted DNA from 1,928 mite specimens and amplified the cytochrome c oxidase I (CO1) DNA barcode. We analyzed the network structure to assess the degree of generalization or specialization of mites to their host plants. We recorded 18 species of hummingbird flower mites from three genera (Proctolaelaps, Rhinoseius, and Tropicoseius) interacting with 14 species of plants. We found that generalist mites are common, and congeneric mite species often share host plants. Our results challenge the assumption of strict specialization that supports this system as an example of mating rendezvous evolution.

A lever action hypothesis for pendulous hummingbird flowers: experimental evidence from a columbine.

Pendulous flowers (due to a flexible pedicel) are a common, convergent trait of hummingbird-pollinated flowers. However, the role of flexible pedicels remains uncertain despite several functional hypotheses. Here we present and test the 'lever action hypothesis': flexible pedicels allow pendulous flowers to move upwards from all sides, pushing the stigma and anthers against the underside of the feeding hummingbird regardless of which nectary is being visited. To test whether this lever action increased pollination success, we wired emasculated flowers of serpentine columbine, Aquilegia eximia, to prevent levering and compared pollination success of immobilized flowers with emasculated unwired and wire controls. Seed set was significantly lower in wire-immobilized flowers than unwired control and wire control flowers. Video analysis of visits to wire-immobilized and unwired flowers demonstrated that birds contacted the stigmas and anthers of immobilized flowers less often than those of flowers with flexible pedicels. We conclude that flexible pedicels permit the levering of reproductive structures onto a hovering bird. Hummingbirds, as uniquely large, hovering pollinators, differ from flies or bees which are too small to cause levering of flowers while hovering. Thus, flexible pedicels may be an adaptation to hummingbird pollination, in particular due to hummingbird size. We further speculate that this mechanism is effective only in radially symmetric flowers; in contrast, zygomorphic hummingbird-pollinated flowers are usually more or less horizontally oriented rather than having pendulous flowers and flexible pedicels.

Atypical Flowers Can Be as Profitable as Typical Hummingbird Flowers.

In western North America, hummingbirds can be observed systematically visiting flowers that lack the typical reddish color, tubular morphology, and dilute nectar of "hummingbird flowers." Curious about this behavior, we asked whether these atypical flowers are energetically profitable for hummingbirds. Our field measurements of nectar content and hummingbird foraging speeds, taken over four decades at multiple localities, show that atypical flowers can be as profitable as typical ones and suggest that the profit can support 24-h metabolic requirements of the birds. Thus, atypical flowers may contribute to successful migration of hummingbirds, enhance their population densities, and allow them to occupy areas seemingly depauperate in suitable resources. These results illustrate what can be gained by attending to the unexpected.

Global patterns of interaction specialization in bird–flower networks

AbstractAimAmong the world's three major nectar‐feeding bird taxa, hummingbirds are the most phenotypically specialized for nectarivory, followed by sunbirds, while the honeyeaters are the least phenotypically specialized taxa. We tested whether this phenotypic specialization gradient is also found in the interaction patterns with their floral resources.LocationAmericas, Africa, Asia and Oceania/Australia.MethodsWe compiled interaction networks between birds and floral resources for 79 hummingbird, nine sunbird and 33 honeyeater communities. Interaction specialization was quantified through connectance (C), complementary specialization (H2′), binary (QB) and weighted modularity (Q), with both observed and null‐model corrected values. We compared interaction specialization among the three types of bird–flower communities, both independently and while controlling for potential confounding variables, such as plant species richness, asymmetry, latitude, insularity, topography, sampling methods and intensity.ResultsHummingbird–flower networks were more specialized than honeyeater–flower networks. Specifically, hummingbird–flower networks had a lower proportion of realized interactions (lower C), decreased niche overlap (greater H2′) and greater modularity (greater QB). However, we found no significant differences between hummingbird– and sunbird–flower networks, nor between sunbird– and honeyeater–flower networks.Main conclusionsAs expected, hummingbirds and their floral resources have greater interaction specialization than honeyeaters, possibly because of greater phenotypic specialization and greater floral resource richness in the New World. Interaction specialization in sunbird–flower communities was similar to both hummingbird–flower and honeyeater–flower communities. This may either be due to the relatively small number of sunbird–flower networks available, or because sunbird–flower communities share features of both hummingbird–flower communities (specialized floral shapes) and honeyeater–flower communities (fewer floral resources). These results suggest a link between interaction specialization and both phenotypic specialization and floral resource richness within bird–flower communities at a global scale.

The dynamics of hummingbird dominance and foraging strategies during the winter season in a highland community in Western Mexico

AbstractCoexistence between ecologically similar species can promote competition for resources. Hummingbirds (Aves: Trochilidae) are nectarivorous birds that usually coexist in sympatry, and visit flowers using different foraging behaviors and levels of aggressiveness as a strategy to diminish resource competition. Here, we describe the dynamics in territorial and dominance behaviors in a hummingbird community inhabiting a highland during winter in Western Mexico. We explored in natural conditions how foraging strategies and dominance status of the hummingbirds was influenced by community species composition, sex, age and size of the individuals, floral abundance, and nectar resource availability. The community studied was composed of 11 species (four residents, three altitudinal migrants, three latitudinal migrants), and all possible combinations of dominance and territoriality were found. Differences in the dominance status and foraging behavior were related to the species, sex, age and body size of the individuals, as well as the number of flowers in the patches, and the abundance of the migratory species over time. The aggressive and territorial species preferred the places with more flowers, and started the majority of the interactions attacking even birds that did not visited the flowers. The subordinate non‐territorial hummingbirds visited floral patches of different sizes, did not start aggressive interactions and when they were involved in one, this occurred during their visits to the territorial hummingbird's flowers. Hummingbird dominance changed during the winter and at the end of the season, when the latitudinal migrant species left our study site, the resident species were more abundant, dominant and territorial. Dominance status and foraging behavior, together with floral preferences and the spatial distribution of nectar resources, acted as mechanisms organizing this hummingbird community.

Number of hummingbird visits determines flower mite abundance on hummingbird feeders.

Members of several genera of mites from the family Melicharidae (Mesostigmata) use hummingbirds as transport host to move from flower to flower, where they feed on pollen and nectar. The factors that influence hummingbird flower mite abundance on host plant flowers are not currently known. Here we tested whether hummingbird flower mite abundance on an artificial nectar source is determined by number of hummingbird visits, nectar energy content or species richness of visiting hummingbirds. We conducted experiments employing hummingbird feeders with sucrose solutions of low, medium, and high energy concentrations, placed in a xeric shrubland. In the first experiment, we recorded the number of visiting hummingbirds and the number of visiting hummingbird species, as well as the abundance of hummingbird flower mites on each feeder. Feeders with the highest sucrose concentration had the most hummingbird visits and the highest flower mite abundances; however, there was no significant effect of hummingbird species richness on mite abundance. In the second experiment, we recorded flower mite abundance on feeders after we standardized the number of hummingbird visits to them. Abundance of flower mites did not differ significantly between feeders when we controlled for hummingbird visits. Our results suggest that nectar energy concentration determines hummingbird visits, which in turn determines flower mite abundance in our feeders. Our results do not support the hypothesis that mites descend from hummingbird nostrils more on richer nectar sources; however, it does not preclude the possibility that flower mites select for nectar concentration at other spatial and temporal scales.

Hummingbird color preference within a natural hybrid population of Mimulus aurantiacus (Phrymaceae)

AbstractHummingbird flowers are typically red in color but the reasons for this are not well understood. Relatively few studies have examined hummingbird flower color preferences under natural conditions in which flower color varies within a species. We recorded hummingbird visitation rates to flowers that vary in color from yellow to red in a natural hybrid population between red‐ and yellow‐flowered Mimulus aurantiacus subspecies. We also examined whether there were any correlations between color and flower size or nectar content. Finally, we reviewed the literature on hummingbird color choice tests using feeders and flowers. There were no correlations in this population between flower color and flower size, nectar volume, or sugar concentration. Nevertheless, hummingbirds undervisited the two most yellow color classes, overvisited orange flowers, and visited the two most red color classes in proportion to their frequency in the population. While Hummingbirds preferred flowers expressing red pigments to those that did not, the flowers with the most red hue were not the most attractive, as has been observed in similar studies with other species of Mimulus. While feeder studies generally fail to show hummingbird preference for red, all studies using flowers, including those that control all floral traits other than color, find consistent preference for red. Experiments are suggested that might help disentangle hypotheses for why hummingbirds exhibit this preference.

When hummingbirds are the thieves: visitation effect on the reproduction of Neotropical snowbell Styrax ferrugineus Nees & Mart (Styracaceae)

The spectrum of floral visitors associated with a particular plant is frequently larger than predicted by the traditional concept of floral syndromes and the role that unpredicted visitors play in plant reproduction deserves attention. Hummingbirds are frequently recorded visiting flowers with distinct floral syndromes, especially in some hummingbird flower poor ecosystem such as the Cerrado. In this study we investigated the effect of frequent hummingbird visits on the reproduction of melittophilous Styrax ferrugineus. The flowers were visited by many different insect groups and visits by hummingbirds were frequent, especially early in the morning when nectar availability was higher. Nectar parameters varied considerably during the flower life span, and was probably affected by the temperature variation during the day. Hummingbird exclusion experiments showed no effect in the fruit-set of the plant. The inability of hummingbirds to deplete all of the nectar produced in this mass-flowering plant, and the existence of another resource (pollen) for the primary pollinator (large bees), are possible reasons why there was no perceptible hindrance of plant reproduction by the hummingbirds. Although no effect in the fruit-set was observed, the use of non-ornithophilous flowers must be important for hummingbirds, which are specialized nectar consumers.

Floral resources and hummingbirds on an island of flooded forest in Central Brazil

Nectar availability for hummingbirds was studied for two and a half years in an island of flooded forest in Uberlândia, MG, in the Cerrado region, the Neotropical savanna areas in Central Brazil. Flowering individuals of ten plant species from eight families that offered resources potentially used by hummingbirds were recorded along transects crisscrossing the area, which were surveyed every other week. The number of flowers was estimated for each plant. The volume and concentration of nectar were measured for a sample of flowers of each species, and the data were used to estimate the caloric content available per flower and the general resource availability in the area throughout the study period. Focal observations of hummingbird visits were carried out on blooming individuals of all plant species. Hummingbird flowers occurred throughout the study period between January 2001 and June 2003. Only three of the studied plants did not present clear ornithophilous features. Six hummingbird species were recorded in the area, but total nectar availability in the area was seldom sufficient for the daily energy requirements of a single bird. Therefore, it can be inferred from the study that these hummingbirds need to move among the mosaic of different plant physiognomies in the region in order to obtain enough energy for their survival. These movements may also contribute to increasing gene flow and regeneration of plant populations isolated in forest patches as the one studied here.

Effects of Pollen in Lobelia laxiflora (Lobeliaceae) Long-Lived Flowers on Fecundity of Tropicoseius chiriquensis (Acari: Mesostigmata: Ascidae)

Abstract Pollen and nectar are the most important floral rewards in plant—pollinator interactions. These rewards are voraciously consumed by hummingbird flower mites (Acari: Mesostigmata: Ascidae), and nectar and pollen feeding by mites have a strong impact on plant—hummingbird interactions. Tropicoseius chiriquensis (Acari: Mesostigmata: Ascidae) feed on nectar and pollen of Lobelia laxiflora H.B.K. (Lobeliaceae) flowers, and within the floral corolla they mate and lay eggs. Yet the benefits of nectar and pollen feeding in terms of survival and fecundity of flower mites are quite unexplored. Here, we evaluate the fecundity and numerical response of T. chiriquensis to the availability of pollen in long-lived protandrous flowers of L. laxiflora. Under field conditions, unmanipulated flowers with pollen had more mites (adults, nymphs, larvae, and eggs) than emasculated flowers without pollen, Numbers of mites fluctuated throughout the flower life span. More adult mites were found for the first 2 d of the flower (staminate phase) and decreased for the last three days of life of the flower (pistillate phase). In contrast, we found higher number of larvae in the pistillate phase. In a laboratory experiment, the fecundity of flower mites was 4 times as high when pollen was available as that when flower mites were not provided with pollen. They completed their life cycle in approximately a week, the life span of L. laxiflora flowers. Our results suggest that pollen availability in L. laxiflora long-lived flowers could influence the population dynamics of T. chiriquensis by having a positive impact on numbers and fecundity of the mites.

IS FLORAL SPECIALIZATION AN EVOLUTIONARY DEAD‐END? POLLINATION SYSTEM TRANSITIONS INRUELLIA(ACANTHACEAE)

Pollination systems frequently reflect adaptations to particular groups of pollinators. Such systems are indicative of evolutionary specialization and have been important in angiosperm diversification. We studied the evolution of pollination systems in the large genus Ruellia. Phylogenetic analyses, morphological ordinations, ancestral state reconstructions, and a character mapping simulation were conducted to reveal key patterns in the direction and lability of floral characters associated with pollination. We found significant floral morphological differences among species that were generally associated with different groups of floral visitors. Floral evolution has been highly labile and also directional. Some specialized systems such as hawkmoth or bat pollination are likely evolutionary dead-ends. In contrast, specialized pollination by hummingbirds is clearly not a dead-end. We found evidence for multiple reverse transitions from presumed ancestral hummingbird pollination to more derived bee or insect pollination. These repeated origins of insect pollination from hummingbird-pollinated ancestors have not evolved without historical baggage. Flowers of insect-pollinated species derived from hummingbird-pollinated ancestors are morphologically more similar to hummingbird flowers than they are to other more distantly related insect-pollinated flowers. Finally, some pollinator switches were concomitant with changes in floral morphology that are associated with those pollinators. These observations are consistent with the hypothesis that some transitions have been adaptive in the evolution of Ruellia.

A new genus of flower-dwelling melicharid mites (Acari: Mesostigmata: Ascoidea) phoretic on bats and insects in Costa Rica and Brazil

The genus Spadiseius gen. nov. of the ascoid family Melicharidae is described, based on all instars of two newly described species in Costa Rica and adults of four undescribed species in Costa Rica and Brazil. These mites undergo their life histories quickly in spike or spadix inflorescences of their plant hosts in lowland tropical rainforests, where they apparently feed on nectar or pollen. One inflorescence may bear hundreds of the developing mites in all instars. Pollen is found on all instars of these mites, which may act as miniature pollinators of their plant hosts. Adult males are strongly sexually dimorphic, with bizarrely modified setae peripherally on the body dorsum and on dorsal surfaces of the legs. These structures are thought to be used competitively for mates, in much the same way as in males of the related hummingbird flower mite genera Rhinoseius and Tropicoseius. Spadiseius calyptrogynae sp. nov. lives in inflorescences of a palm, Calyptrogyne ghiesbreghtiana, and has a phoretic association both with perching bats of the genus Artibeus, which are major pollinators of this plant, and scarab beetles of the genus Lagochile, which feed destructively on flowers of this plant. Spadiseius spathiphyllae sp. nov. lives in inflorescences of an aroid, Spathiphyllum friedrichsthalli, and is phoretic on meliponine bees of the genus Trigona that pollinate this plant. Polymorphism among adult males occurs in both species. In male S. calyptrogynae a “warty” ornamentation of the dorsal shield seems to be correlated with extreme elongation of certain body and leg setae, while in male S. spathiphyllae such disproportionate setal elongation seems correlated with more robust individuals of greater body and leg dimensions. Several newly recognized morphological attributes are noted among species of Spadiseius, including apically forked salivary styli.

Evolutionary associations between nectar properties and specificity in bird pollination systems

A long-standing paradigm in biology has been that hummingbirds and passerine birds select for different nectar properties in the plants they pollinate. Here we show that this dichotomy is false and a more useful distinction is that between specialized and generalized bird pollination systems. Flowers adapted for sunbirds, which are specialized passerine nectarivores, have nectar similar to that of hummingbird flowers in terms of volume (approx. 10-30 microl), concentration (approx. 15-25% w/w) and sucrose content (approx. 40-60% of total sugar). In contrast, flowers adapted to generalized bird pollinators are characterized by large volumes (approx. 40-100 microl) of extremely dilute (approx. 8-12%) nectar with minimal sucrose (approx. 0-5%). These differences in nectar traits are highly significant even when statistical analyses are based on phylogenetically separate pairwise comparisons between taxa. We present several hypotheses for the association between nectar properties and specificity in bird pollination systems.

The Effect of Hummingbird Flower Mites on Nectar Availability of Two Sympatric Heliconia Species in a Brazilian Atlantic Forest

Hummingbird flower mites feed and reproduce in flowers of host plants pollinated by hummingbirds, and use the nostrils and bill of the hummingbird to move from plant to plant. These mites compete with the pollinator for the nectar produced by flowers. An investigation was made of the relationship between the pattern of nectar production and the effects of hummingbird flower mites in the flowers of two sympatric species of Heliconia (Heliconiaceae). Nectar production was sampled by carrying out two experiments: 2-hour intervals and accumulated nectar. Flowers with and without mites were used in both experiments. Exclusion of mites increased nectar production, especially in accumulated daily production (a maximum of 49 % more nectar). Both Heliconia species had the same pattern of nectar production: a high concentration in the morning, which was progressively reduced as the day passed. This pattern of nectar production coincides with the behaviour of the pollinator, which makes more frequent visits in the morning, as observed in a previous study. The results suggest that the impact of mites on nectar availability of Heliconia is more important with regard to total volume of nectar produced irrespective of flower longevity. A high variation among individuals in nectar produced in the populations was also observed. Hummingbird flower mites strongly affect availability of nectar for hummingbirds.

Fenologia reprodutiva, biologia floral e polinizadores de duas espécies simpátricas de Marantaceae em um fragmento de Floresta Atlântica, Nordeste do Brasil

Foi analisada a floração, a frutificação e a biologia da polinização em duas espécies de Marantaceae: Ischnosiphon gracilis (Rudge) Koern. e Stromanthe porteana A. Gris. As observações foram realizadas em populações naturais no Parque Estadual Dois Irmãos (8º7'30" S e 34º52'30" W), um remanescente de Floresta Atlântica em Pernambuco. Nas duas espécies foi verificado padrão fenológico contínuo, com diferentes picos de floração e frutificação. As inflorescências em I. gracilis produziram 14,4 ± 3,4 flores e 1,3 ± 0,6 frutos, enquanto em S. porteana, produziram 125,4 ± 14,8 flores e 7,4 ± 4,9 frutos. Foram verificadas baixa razão pólen/óvulo e reduzida produção natural de frutos nas duas espécies. Em I. gracilis, a concentração de açúcares no néctar foi alta (26%-32%), característica de flores visitadas por abelhas e em S. porteana o néctar foi menos concentrado (20%), situação comum em flores visitadas por beija-flores. Ischnosiphon gracilis é polinizada por três espécies de abelhas Euglossini (Euglossa sp., Eulaema bombiformis e E. cingulata), enquanto S. porteana é polinizada por uma espécie de abelha Euglossini (Eufriesea surinamensis) e por duas espécies de beija-flores (Phaethornis ruber e Amazilia versicolor). As diferenças observadas entre as flores das duas espécies evitam partilha e competição por polinizadores, o que pode garantir a manutenção delas em seu habitat. Entretanto, a longo prazo, a baixa produção de frutos pode afetar a estrutura da população, diminuindo o sucesso reprodutivo das duas espécies.

Ascid mites (Acari: Mesostigmata: Ascidae) from Costa Rican hummingbirds (Aves: Trochilidae), with description of three new species and a key to the Proctolaelaps belemensis species group

Eleven species of hummingbird flower mites of the genera Proctolaelaps Berlese, 1923, Tropicoseius Baker & Yunker, 1964 and Rhinoseius Baker & Yunker, 1964 were recorded from hummingbirds in Costa Rica. Three new species of the genus Proctolaelaps are described and figured, i.e. P. threnetes Dusbabek & Literak, sp. nov., P. naskreckii Dusbabek & Havlicek, sp. nov. and P. chalybura Dusbabek & Capek, sp. nov. A key for identification of females and males of the P. belemensis species group is included. Five species of the genus Tropicoseius and three species of the genus Rhinoseius are recorded, three of these species are recorded from Costa Rica for the first time.

Down‐facing flowers, hummingbirds and rain

Flower orientation, the angle between a flower’s main axis and the horizontal, is an important but understudied flower trait. I explore the relationship between flower orientation, hummingbird pollination, and rain, and review evidence for the adaptive significance of a down‐facing orientation as a mechanism to avoid flower flooding. A screening of a tropical and a temperate flora showed that hummingbird flowers more commonly face down than insect flowers. In the temperate forests of southern South America, genera pollinated by hummingbirds and genera with down‐facing flowers—irrespective of pollination mode—increase along a strong E‐W precipitation gradient across the Andes. These patterns suggest that flowers with a down‐facing orientation are associated with pollinators that, like hummingbirds, remain active during rain and that this type of flower orientation may constitute an adaptation against the deleterious effects of nectar dilution and pollen rainwash. Still scarce experimental evidence confirms that a down‐facing morphology may prevent flower flooding. However, other morphological traits, like corolla closure and a narrow tubular shape, could also be effective in hindering nectar dilution and pollen rainwash.

Flower Color, Hummingbird Pollination, and Habitat Irradiance in Four Neotropical Forests1

ABSTRACTPollinator visual systems differ considerably among broad groupings such as bees, bats, and birds, and it has been proposed that factors shaping the diversity of flower color in tropical plants include differences in pollinator perceptual abilities. Within the pollinators of the Neotropics, one major difference between taxa is that hummingbirds perceive color well across a broad range of wavelengths from 300–660 nm whereas most bees perceive color well over a narrower range spanning 300–550 nm. Thus, hummingbirds can see red and other long‐wavelength reflection much better than bees. Another factor that could potentially influence flower color in tropical forests is the difference in light availability among habitats such as gaps, canopy, and forest understory. I conducted a survey of floral color in four Neotropical forests using a portable spectroradiometer to provide an unbiased measure of color reflectance. The primary pollinator agents and light habitats of each plant species were classified using primary literature or accounts of direct observations by experts. Flower color was not influenced by differences in light availability between open and closed habitats but was influenced by pollinator visual systems. Specifically, insect flowers reflected across a broad range of the spectrum but hummingbird flowers reflected mostly long‐wavelength light (typically median wavelength <585 nm). Thus, hummingbird‐pollinated flowers are not tuned specifically to hummingbird color sensitivity but instead may decrease conspicuousness to bees and other insects that have poor visual sensitivity to long‐wavelength color.

Systematics and evolution of tribe Sinningieae (Gesneriaceae): evidence from phylogenetic analyses of six plastid DNA regions and nuclear ncpGS

For nearly all species in the three genera of tribe Sinningieae (Gesneriaceae), Sinningia, Paliavana, and Vanhouttea (mostly in southeastern Brazil) plus 10 outgroups, we have sequenced six non-coding DNA regions (i.e., plastid intergenic spacers trnT-trnL, trnL-trnF, trnS-trnG, atpB-rbcL, and introns in the trnL and rpl16 genes) and four introns in nuclear plastid-expressed glutamine synthetase gene (ncpGS). Separate and combined analyses of these data sets using maximum parsimony supported the monophyly of Sinningieae, but the genera Paliavana and Vanhouttea were found embedded within Sinningia; therefore a new infrageneric classification is here proposed. Mapping of pollination syndromes on the DNA-based trees supported multiple origins of hummingbird and bee syndromes and derivation of moth and bat syndromes from hummingbird flowers. Perennial tubers were derived from perennial stems in non-tuberous plants.