Biotic Pollination: How Plants Achieve Conflicting Demands of Attraction and Restriction of Potential Pollinators

In flowering plants, pollination, transfer of pollen grains from the anther to the stigma, is a pre-requisite for seed set and plays a critical role in their reproductive success. Over 85% of the flowering plants are pollinated by animals (biotic pollination) and the remaining are pollinated by wind or water. Amongst animals, insects are the major pollinators. Birds and bats are the other important pollinators. Plant pollinator interactions are largely mutualistic, but there are a good number of species that achieve pollination by deceit. Biotic pollination essentially involves a sedentary partner (plant) and a mobile partner (animal). This imposes two conflicting demands on plants. 1) They have to develop effective devices to attract animals to visit their flowers in a sustainable way and use them effectively for pollination services. 2) They have to apply some degree of discrimination to restrict the number of animal species visiting the flowers. In the absence of such discrimination, all animal species that reside in the habitat may visit all synchronously flowering plant species and bring about extensive hetero-specific pollination. As pollination success depends on the transfer of pollen to conspecific stigma, heterospecific pollination compromises reproductive fitness of plant species. Most of the studies on pollination ecology have so far concentrated on floral attraction; only limited studies have been carried out on the details of restriction of floral visitors. This chapter gives a brief introduction on the origin of biotic pollination and its significance in diversification of flowering plants, and discusses our current understanding on how plants have been able to achieve conflicting demands of attraction and restriction of floral visitors for efficient pollination.

Flowers are the units of sexual reproduction that involves a series of sequential events—production of functional pollen grains (male partners) and ovules (female partners), transfer of pollen grains from the anther to the stigma (pollination), pollen-pistil interaction leading to the entry of pollen tubes into the ovules and delivery of male gametes near the female gamete, fertilization, development of fruits and seeds, dispersal of seeds, and growth of seedlings into adults. Pollination is one of the most critical events in sexual reproduction of seed plants and plays a crucial role in both fundamental and applied aspects of reproductive biology. Pollination is the only means of gene flow between conspecific plants and populations, and is thus the basis of recombination. Successful pollination is a prerequisite for fruit and seed development, and thus plays an important role in reproductive success of both cultivated and wild plant species. Pollination limitation reduces the yield of cultivated species (Roubik 1995; Knight et al. 2005; James and Pitts-Singer 2008) and often acts as a driving force for species vulnerability in natural habitats (Spira 2001; Wilcock and Neiland 2002; Biesmeijer et al. 2006; Potts et al. 2009; Shivanna 2012a; Burkle et al. 2013; Tylianakis 2013). Information on pollination ecology is essential for the release of genetically transformed plants (Armstrong et al. 2005). Pollination services are needed to sustain pollinators particularly those that depend exclusively on floral resources (nectar/pollen).  

Pollination is simply the deposition of pollen grains from an anther to the stigma. Based on the source of pollen, pollination is of three types: i) autogamy: transfer of pollen grains from the anthers to the stigma of the same flower, ii) geitonogamy: transfer of pollen grains from the anthers to the stigma of another flower of the same plant or another plant of the same clone and iii) xenogamy: transfer of pollen grains from the anther to the stigma of another non-clonal plant.

Sexuality of the flower and of the plant affects pollen flow within and between flowers/plants. The flowers may be hermaphrodite (producing both functional pollen and ovules) or male (producing only functional pollen) or female (producing only functional ovules). The plant may be bisexual (producing only hermaphrodite flowers) or monoecious (each plant producing both male and female flowers) or dioecious (each plant producing only male or only female flowers). Thus, all three types of pollinations can occur in plants producing hermaphrodite flowers; only geitonogamy and xenogamy can occur in monoecious species while only xenogamy is possible in dioecious species. Expression of male and female phases in hermaphrodite flowers may be synchronous or show temporal (dichogamy) or spatial (herkogamy) separation. When the female phase in dichogamous flowers is expressed earlier than the male phase, the condition is called protogyny and when the male phase is expressed earlier than the female phase, it is called protandry. In herkogamous flowers, the anthers and the stigma are spatially separated and are located at different levels in the flower. Herkogamy and dichogamy prevent autogamy but are not effective in preventing geitonogamy. Many hermaphrodite species are self-incompatible that prevents both autogamy and geitonogamy completely and permits only xenogamy.

Over 85% of the flowering plants are pollinated by a variety of animals (biotic-pollination) and the remaining by wind or water (Ollerton et al. 2011). Thus, biotic pollination provides one of the crucial ecosystem services both in agricultural fields and natural habitats. These services are threatened by habitat destruction and global warming which lead to reduction of pollinator diversity and abundance. Several recent studies have highlighted decline of pollinators and also parallel decline of plant species (that depended obligatorily on biotic-pollination) with their pollinators (Biesmeijer et al. 2006; Aguilar et al. 2006; Memmott et al. 2007; Potts et al. 2009; Garibaldi et al. 2013; Burkle et al. 2013 see also Tylianakis 2013).

Plant-pollinator interactions are largely mutualistic and result in reciprocal benefits. It is a form of ‘biological barter’ and involves exchange of resources of the plant such as pollen and nectar with the services of the pollinator (Ollerton 2006). There are, however, a good number of species that achieve pollination through deceit without offering any rewards to the pollinators. The ultimate strategy of plants is to deliver and receive conspecific pollen with minimum allocation of resources while that of the pollinator is to harvest maximum reward with minimum use of energy and time.

Evolution of biotic pollination provides several advantages to plants over wind and water pollination: i) biotic-pollination is more efficient as animals seek out conspecific flowers and often transport pollen for longer distances, ii) it can thrive even in habitats with minimum wind as in tropical forests with closed canopy and iii) plant species need to allocate lesser resources for pollen production than wind/water-pollinated species, as the latter involves extensive pollen wastage (see Pellmyr 2002).  Many evidences indicate that abiotic pollination by wind and water is a derived condition from biotic pollination: i) Anemophily is present in families such as Poaceae, Cyperaceae, Betulaceae and Juncaceae that are highly evolved and whose ancestry can be traced to animal-pollinated species (Faegri and van der Pijl 1971). ii) Anemophily is rare in tropical rain forests where pollinating animals are abundant. iii) Several wind-pollinated species show a combination of wind- and biotic-pollination. For example species of Salix produce nectar and their flowers are pollinated by insects; they also produce large amount of pollen that are carried by wind. (iv) Insect pollination seems to be ancestral in basal angiosperms which represent the first flowering plants to diverge from the ancestral angiosperms (Hu et al. 2012). Anemophily seems to have evolved as an adaptation to overcome the scarcity of pollinators. Hydrophily seems to have evolved in response to aquatic habitat which is considered to be a derived one for higher plants (Faegri and van der Pijl 1971). Most of the aquatic plants in which the flowers emerge above the water level are pollinated by animals similar to their terrestrial ancestors. Only those aquatic species in which flowers remain inside the water have developed devices to use water for pollen transport.

Two Conflicting Demands of Biotic Pollination: Attraction and Restriction of Pollinators

Animal-mediated pollination essentially involves a sedentary partner (plant) and a mobile partner (animal). This imposes two conflicting demands on plants. 1) They have to develop effective devices to attract suitable animals to visit their flowers in a sustainable way and use them effectively for pollination services. This requires substantial investment from plants in the form of attractants and rewards. 2) Plants have to apply some degree of discrimination to restrict the number of animal species visiting the flowers of each plant species. In the absence of such discrimination, all animal species that are present in the habitat may visit all synchronously flowering plant species. This would result in (i) unhealthy competition between plant species, (ii) enormous wastage of pollen and (iii) extensive heterospecific pollination (with pollen of other species). These features seriously compromise reproductive fitness of plant species (see Pellmyr 2002). Therefore, plants have to devise ingenious ways to achieve these conflicting demands of attraction and restriction of floral visitors for efficient pollination. Studies on pollination ecology so far have concentrated largely in understanding the details of attraction; only limited studies have been carried out to understand the details of restriction. Even the reviews and books on pollination ecology hardly discuss the restriction aspect of pollination, except some aspects of morphological filters (Pellmyr 2002; Roubik et al. 2005; Dafni et al. 2005; Bronstein et al. 2006; Waser and Ollerton 2006; James and Pitts-Singer 2008; Anonymous 2009; Willmer 2011; Patiny 2012). In this review an attempt is made to give a brief introduction on the origin of biotic pollination and its significance in diversification of flowering plants and pollinators, and bring together available information on how the plants have been able to achieve these conflicting demands of attraction and restriction of pollinators to flowers. As the literature in the area is extensive, the coverage tends to be subjective with an emphasis on recent studies.

Origin of Biotic-Pollination

Biotic pollination evolved in two groups of Gymnosperms (Cycadales and Gnetales) as early as the Permian period (about 250 million yrs. ago) and is present even in several extant species of these groups. Beetles, flies, thrips and wasps are the most common pollinators in Gymnosperms (Seymour et al. 2004; Terry et al. 2007; Marler 2010). Fragrance is the basis of attraction of pollinators. The chemical composition of their fragrance is similar to general herbivore deterrents (Pellmyr and Thien 1987). It is suggested that early attractants for pollinators evolved from herbivore deterrents (Pellmyr et al. 1991). The pollinators frequently utilize male cones for nursery purpose (for mating and obtaining food for larvae and adults) and female cones mimic olfactory cues of male cones to attract pollinators. In some species, the pollination drop present at the tip of the ovule may serve as a reward for pollinators. Some cycads show obligate mutualism with their specialist pollinators, and have evolved mechanisms to generate heat (4–12ºC above ambient temperature) in the cones. The heat increases the emission of volatiles and improves pollinators’ attraction (Terry et al. 2007; Marler 2010).

In Angiosperms, animal-mediated pollination was prevalent from the beginning of their evolution in the early Cretaceous period (over 100–120 million yrs. ago). The oldest fossil of the flower discovered in the early Cretacieous (about 120 million yrs. ago) is related to extant Nymphaeales.   Another fossil found in the upper Cretaceous (about 90 million yrs. ago) is similar to the extant Victoria species. Based on molecular phylogeny, earliest angiosperms, termed basal angiosperms, representing the first flowering plants to diverge from the ancestral angiosperms, have been recognized (Soltis et al. 2000). Pollination ecology of many species of basal angiosperms has been studied (Pellymr and Thien 1987; Endress 2010, 2011; Thien et al. 2009). The flowers of basal angiosperms are generally small, simple and radially symmetrical. Bisexual flowers are more common and all of them are protogynous. Most of the species show apocarpous (free carpels) condition and well-defined style is generally absent. They seem to attract generalized pollinators prevalent in animal-pollinated Gymnosperms and show many similarities to the pollination system in Gymnosperms (Pellymr and Thien 1987). Beetles and flies are the major pollinators and pollen is the major reward; many of them show floral thermogenesis (use the search bar above for more about this). Bee pollination is found only in Nymphaeaceae. Abiotic pollination is rare in basal angiosperms.

Animal Partners

Amongst animals, insects are the principal pollinators. Hymenoptera (bees, wasps and ants), Lepidoptera (butterflies and moths), Coleoptera (beetles) and Diptera (flies) are the major orders of insects involved in pollination. In recent years, thrips (Thysanoptera) have been reported to bring about pollination in a number of species (Ananthakrishnan 1993; Williams et al. 2001; Garcia-Fayos and Goldarazena 2008). Thrips are very small (< 1 mm long) and are poor flyers; they can, however, be carried for long distances by the wind. Thrips largely bring about autogamy and geitonogamy. Birds and bats are the other important pollinators. They are generally confined to tropical and sub-tropical areas that provide floral resources throughout the year. Bird-pollinated (ornithophilous) species have been reported in over 60 families (see de Wall et al. 2012). In the New World, hummingbirds are the principal bird pollinators and in the Old World and Pacific regions, sunbirds and sugarbirds are the major bird pollinators. In Australia the honeyeaters and in Hawaii the honeycreepers are the main bird pollinators. Bird-pollinated species are more prevalent in Australia when compared to other continents. According to an old survey (Ford et al. 1979) 111 species of birds have been reported to visit flowers of about 250 species of plants. Only two families of bats, Pteropodidae which occurs throughout tropical and subtropical regions of the Old World (including Australia and Pacific islands) and Phyllostomidae which inhabits tropical and subtropical regions of the New World contain species that are specialized in nectar feeding (see Fleming et al. 2009). Old World bats are larger with shorter tongues and do not hover whereas those of the New World are smaller with longer tongues and hover. Bat-pollination (chiropterophily) has been recorded in over 500 species belonging to 67 families and 28 orders of tropical and semitropical plant species (Gibson 2001; Flaming et al. 2009). Bat and bird pollinations are considered to be derived conditions and have evolved independently in many advanced families of flowering plants. Some unusual pollinators have also been reported in some plant species: cockroaches (Nagamitsu and Inoue 1997—Uveria), mice (Wester et al. 2009 —Pagoda lily), squirrels (Tandon et al. 2003—Butea), snails (Sarma et al. 2007—Volvulopsis) and lizards (Olesen and Valido 2003; Ortega-Olivencia 2012—Scrophularia; Hansen et al. 2006—Trochetia).

Bumblebees and honeybees have trichromatic vision with UV, blue and yellow as primary colors and their spectral range extend from about 550 nm to 336 nm. Human vision is also trichromatic (blue, green and red as primary colors) and the spectral range is confined to visible wavelengths (from 400 nm to about 700 nm). All Lepidoptera and flies so far examined have the ability to see objects in UV range (see Kevan 1983; Kearns and Inouye 1993). Those insects which can distinguish UV from longer wavelengths have a separate UV absorbing pigment. Some butterflies can see red while others cannot. Honeybees are also red-blind. Hummingbirds and a few other birds also see in the near UV region of the spectrum (Chen et al. 1984). Bats are generally red-blind although they can distinguish white vs black even in extremely low light (Gibson 2001).

Pollinating insects may be social or solitary (Anonymous 2012). All species of ants and termites, and some species of bees (honeybees, bumblebees and stingless bees) and wasps are social insects. Truly social (eusocial) insects are characterized by: i) all individuals of a colony share a common nest site, ii) reproduction in the colony is restricted to only one or a few females, iii) the individuals of a colony are made up of overlapping generations and iv) brood care within the colony is a co-operative function. Most of the other bees involved in pollination (such as Xylocopa, Megachile, Osmia and Nomia) are solitary. Solitary bees do not form colonies. A solitary female constructs her own nest, stores food for her brood in the form of pollen or nectar and then dies or departs without further care to her offspring. The adult life of these bees is generally short, spanning for a few weeks. They nest alone or in aggregations.

Invertebrate pollinators have limited memory and their visits are largely confined to just one type of flower at a given time (Minckley and Roulston 2006). In a complex flower with hidden rewards, insects have to learn by trial and error the best approach to harvest the reward with minimum time. When they change host species, they have to learn again to handle the second flower (Lewis 1986; see also Raine and Chitka 2007). Pollen-collecting bees generally have a limited foraging range and frequently commute between the nest and the host plants to unload the pollen sample. Bats have stronger memory than bees. Nectar-feeding bats can retain up to 40 food locations in their working memory, facilitating their foraging in tropical nocturnal environments (Winter and Stich 2005). Vertebrate pollinators are homoeothermic and need a continuous supply of food to maintain their high metabolic rate. They are able to visit flowers even during low temperature conditions under which insects are unable to fly. Insects, on the other hand, are able to survive long periods of unfavorable conditions in larval stages or as hibernating adults. Vertebrate pollinators and some butterfly species can cope with unfavorable conditions during some parts of the year through local and regional migrations (see Abrahamczyk et al. 2011).

Diversification of Flowering Plants and Pollinators

There was a dramatic increase in the number of angiosperm species soon after their origin in the Cretaceous period. Darwin recognized this increase and termed this diversification as an “abominable mystery” (see Friedman 2009). Paleontological evidences indicate: i) slow diversification of insect pollinator groups (although they arose in the early Mesozoic or even before) until they became associated with angiosperms and ii) rapid diversification of the angiosperms and pollinating insect groups simultaneously in the latter half of the Cretaceous and tertiary period. On the basis of these evidences evolutionary biologists suggested that biotic pollination acted as a catalyst for reciprocal diversification of pollinating insects and flowering plants in the latter half of the Cretaceous period (see Crane et al. 2000; Pellmyr 2002; Magallon and Castillo 2009; Smith et al 2011). A recent study suggests that the rise of bees coincided with the largest flowering plant clade, the eudicots which comprises 75% of the flowering plants (Cardinal and Danforth 2013; see also Cappellari et al. 2013).

Some evolutionary biologists, however, argue that biotic pollination acted not as an exclusive factor but as a co-factor in the diversification of angiosperms (Sanderson and Donoghue 1994; see Pellmyr 2002). This argument is based on: i) angiosperm diversification increased not at the beginning of their evolution when animal pollination arose but rather at a later stage and ii) animal-pollinated Gymnosperm groups (such as Zamia, Welwitschia and Gnetum) have not diversified more than their wind-pollinated sister groups.

Several evidences indicate that apart from biotic pollination, evolution of the flower and some of its innovations seem to be the contributing factors for the evolutionary diversification of flowering plants. The visual and fragrance components of the flower and the secretion of the nectar as a carbohydrate-rich reward increased dramatically the attraction of pollinators and their sustainability, thus increasing pollination efficiency of flowering plants when compared to the cones of Gymnosperms. The other innovations of the flower that seem to have contributed to the diversification of angiosperms are the evolution of the carpel and floral zygomorphy. Evolution of the carpel, as the result of covering of the ovule(s) with a sporophytic tissue differentiated into the ovary (enclosing the ovules), style and stigma, had profound effects on the operation of sexual reproduction of angiosperms. Unlike in Gymnosperms, pollen grains in Angiosperms do not have direct access to the ovules in which female gametes are located; they land on the stigma placed far away from the ovules. Evolution of the pistil resulted in the addition of another step in pre-fertilization events—growth of pollen tubes through the tissues of the stigma and style—before pollen tubes enter the ovules and release the male gametes for fertilization. All the carpels of the flower together are referred to as the pistil. In most of the eudicots and monocots, all the carpels of the flower are united (syncarpous) with a well-defined style. The transmitting tissue (through which pollen tubes grow) of the styles of individual carpels is also fused in syncarpous pistils enabling pollen tubes originating from one carpel to cross-over to the ovules of other carpels (see Armbruster et al. 2002; Endress 2010). Thus, pollen tubes originating from any stigma or a part of the stigma can enter the ovules of all the carpels in syncarpous pistils. This increases the efficiency of pollination and the extent of seed set even under pollination constraint.

More importantly, the pistil screens pollen grains for quality and compatibility before pollen tubes reach the ovules (see Shivanna 2003). As the number of pollen grains that land on the stigma is generally many times more than the number of ovules available for fertilization, pollen grains are subjected to intense competition during pollen germination and pollen tube growth in the pistil. Those pollen grains that germinate early and pollen tubes that grow faster in the style, enter the ovules earlier than the slow growing pollen tubes and effect fertilization. Pollen grains which germinate later and pollen tubes with slower growth rate are eliminated in this competition. Since cross-pollination takes place to a limited or greater extent in most of the species, there is considerable genetic variability in the pollen population on which pollen competition can operate. There is no scope for such a competition among male gametes in lower groups of plants including gymnosperms. The adaptive significance of pollen competition in increasing the fitness of the progeny is documented in several species through experimental manipulation of pollen competition (Mulcahy 1979; Mulcahy and Mulcahy 1987; Davis et al. 1987; Schlichting et al. 1987; Tejaswini et al. 2001; Lankinen and Madjidian 2011). Seedlings resulting from intense pollen competition show more vigorous and uniform growth when compared to those resulting from no competition or limited competition.

The pistil also plays a critical role in the breeding system of angiosperms. Self-incompatibility (SI) controlled by one or a few genes, each with two or multiple alleles, is one of the common outbreeding devices in plants (de Nettencourt 2001). When the SI allele present in the male gamete matches that of the female gamete, fertilization is prevented. In lower groups of plants where male and female gametes come in direct contact with each other, SI cannot prevent completely the fusion of gametes produced by the same individual. For example S1 and S2 gametes produced by the S1 S2 self-incompatible individual can fuse although fusion of S1with S1 and S2 with S2 gametes is prevented (see Shivanna 2003). In flowering plants SI recognition is established between the gametophytic pollen (each carrying one of the S alleles) and the sporophytic tissues of the pistil (carrying both the S alleles of the parent). In S1S2 plants, for example, pollen grains carrying both S1 and S2 alleles are recognized and inhibited in the pistil. Evolution of the pistil has made self-incompatible flowering plants the most efficient out-breeders since all male gametes that originate from the same plant or any other plant with the same genotype are effectively prevented from fertilization.

Yet another floral innovation that appeared in some major lineages in Cretaceous period was the evolution of bilateral zygomorphic flowers from radial actinomorphic flowers (Friedman 2009; Mach 2012; Vamosi and Vamosi 2012). As zygomorphic flowers have one plane of symmetry, effective pollen transfer depends on the accuracy in functional fitness between the flower and the pollinator leading to specialized pollination systems (Sargent 2004). Such specialized interactions facilitate reproductive isolation and speciation. Phylogenetic analyses have shown that lineages with zygomorphic flowers tend to contain more number of species than their actinomorphic sister groups (Sargent 2004; see Vamosi and Vamosi 2012). Thus, biotic pollination together with the evolution of the flower with the pistil and further floral elaboration in the form of zygomorphy would have acted as catalysts for the reciprocal diversification of flowering plants and pollinating insects.

The details of floral morphology in the basal angiosperms are in agreement with the above concept (Thien et al. 2009). As pointed out earlier, flowers in basal angiosperms are radially symmetrical and the carpels are largely free (apocarpous) with ill-defined style. In apocarpous species pollen grains that land on the stigma of a carpel cannot cross-over to the ovules located in other carpels of the flower. This reduces pollination efficiency and the extent of seed set under deficient pollination environment as the seed set is confined only to pollinated carpel. Also, short style prevalent in basal angiosperms is likely to be less effective in screening pollen quality when compared to eudicots and monocots with well-developed styles (see Mulcahy 1979; Armbruster et al. 2002). These floral features explain limited diversification of basal angiosperms.

Floral Attractants

Plants have to invest considerable resources to attract pollinators and sustain their visits by providing rewards to the visitors. Floral attractants advertise the presence of rewards that provide motivation for animals to visit flowers in a sustainable way. Flowers and inflorescences and their fragrance act as advertisers through visual and olfactory cues (Dafni et al. 2005). Flowers exhibit a remarkable diversity in colors, sizes, shapes, scents, and sexual systems; no other plant organ can match flowers in their diversity. This has enabled them to attract a range of animals to the flowers and use them for pollination services. A standard method used to study differential effects of the color and fragrance in pollinator attraction is to hide the color of the flower by covering the flowers/inflorescences with a non-transparent material allowing the scent to escape through fine holes and record pollinator responses to the fragrance (Dobson et al. 2005). To study the effect of fragrance, the flowers/inflorescences are sealed with hermetic transparent paper to prevent scent emission and study the responses of pollinators to the color. The responses of the above two treatments are compared with un-manipulated flowers (controls). Several investigators have used artificial flowers to study the behavior of floral visitors (Kearns and Inouye 1993; Johnson and Dafni 1998; Cresswell and Smithson 2005; Leonard and Papaj 2011). Visual cues are prepared from colored paper folded suitably by incorporating the landing area, if necessary; sugar solution is provided as the reward. Many animals forage readily from artificial flowers with the most basic similarity to actual flowers (Raguso et al. 2002).

Visual Cues

Visual cues are in the form of size, shape and color of the pollination unit (flower/inflorescence). Floral colors are essentially pigment-based; anthocyanins, anthoxanthins, carotenoids, flavones, flavonols and betaxanthins impart a wide range of colors to the flowers (see Pellmyr 2002). Amongst different floral organs, petals are the major attractants; occasionally other parts of the flower such as bracts become conspicuous.

Nectar guides. Flowers with hidden nectar generally have contrasting patterns on the petals, termed nectar guides, pointing toward the source of the nectar. The size and shape of nectar guides are highly variable (Medel et al. 2003). Many studies have shown that nectar guides do help the visitors to locate the source of nectar. A white mutant of blue-flowered Delphinium nelsonii resulted in the loss of nectar guides. Bumblebees avoided such mutant flowers; even when they visited such flowers, they took longer time to find the nectar site (Waser and Price 1983). Flowers of Linum pubescens (Linaceae) are pollinated by the bee-fly, Usia bicolor (Johnson and Dafni 1998). In flower models with nectar guides, the pollinator tended to follow the lines toward the center of the flower while on the plain models they showed undirected behavior, often moving to the edge of the model. Similarly bumblebees (Bombus impatiens) discovered the rewards more quickly on artificial flowers with star-like pattern when compared to plain flowers (Leonard and Papaj 2011).

In a recent study, using natural flowers, Hansen et al. (2012) studied the role of nectar guides in an iris, Lapeirousia oreogena. The flowers of this species have nectar guides in the form of white narrow markings pointing towards the narrow entrance of the long corolla tube and long-proboscid nemestrinid fly is its sole pollinator. Painting of the nectar guides with ink that matched the color of the corolla background dramatically reduced proboscis insertion into the corolla tube, although it did not affect the approaches of the flies to the flowers from a distance. As expected, removal of nectar guides significantly reduced pollen export as well as fruit set.

Change of color in older flowers and its role in pollinator attraction. In a number of species, older non-rewarding flowers change color (instead of senescing) and are retained on the plant for several days. This phenomenon has been reported in over 200 species belonging to 74 families (Gori 1983; Weiss 1991). In all the 26 such species investigated (Weiss 1991) fresh flowers offered nectar and pollen rewards, and older flowers contained little or no nectar and lacked pollen. Retention of older flowers increased plant’s attractiveness to pollinators from a distance. However, the pollinator discriminated the color of the flower from a close range and mostly visited rewarding flowers. In Lantana camara, a classical example of post-pollination changes in color, the flowers are yellow on the day of anthesis and offer pollen and nectar to the pollinator. They turn orange and then red on subsequent days. Red flowers are retained on the plant for several days although they do not offer rewards. In a caged experiment, Weiss (1991) reported that the pollinator butterfly (Agraulis vanilla) visited yellow inflorescences significantly more often than red ones. Discriminating ability of several pollinators (belonging to Diptera, Hymenoptera and Lepidoptera) between rewarding and non-rewarding flowers was recorded in a number of other color-changing plant species (Weiss 1991).

Olfactory Cues

Olfactory cues are in the form of volatile fragrance compounds emitted by flowers. Floral fragrance, in most of the species, is produced by petals but in several species other floral organs, particularly pollen grains (Dobson 1988; Falara et al. 2013), also contribute to the fragrance. The fragrance is in the form of complex mixtures of a large number of volatile compounds, which give each species a characteristic fragrance. Maximum emission of fragrance generally coincides with the activities of their pollinators (Ando et al. 2001) and fragrance emission often follows endogenous rhythm (Dudareva et al. 1999). For example, methyl benzoate is one of the most abundant fragrance compounds in flowers of snapdragon, tobacco and petunia. Its emission follows circadian rhythm (Kolosova et al. 2001); in diurnally pollinated snapdragon, maximum emission occurs during the day coinciding with the forging activity of its pollinator (bumblebees) and in nocturnal pollinated tobacco and petunia, maximum emission of methyl benzoate is during the night. However, nocturnally pollinated flowers of Clarkia breweri in which S-linalool (acyclic monoterpene) is a major component of the fragrance (Dudareva and Pichersky 2000) and Dianthus inoxianus in which aliphatic 2-ketones and sesquiterpenoids are the major components (Balao et al. 2011) do not show such differences in emission rate between the day and night. In Australian Chiloglottis orchids the levels of active compounds, chiloglottones, remain stable not only during the day and night but also over the lifetime of the flower lasting 2 to 3 wk. (Falara et al. 2013). In these orchids, UV-B light is required for their synthesis (Falara et al. 2013).

In general, the fragrance acts as long-distance attractant and at closer range both color and fragrance act synergistically to guide the visitor to the flower. In an interesting study on long-distance attraction of fragrance, (Ackerman 1983; see Williams 1983) took a canoe on a lake in Costa Rica and exposed floral odor of an orchid species, pollinated by a euglossine bee, at a distance of 50, 200 and 1000 m from the shore, and recorded the arrival of euglossine bees up to 1000 m to the odor source. For insects involved in nocturnal pollination, fragrance is the major or the sole attractant (Jurgens et al. 2002; Balao et al. 2011).

Analysis of floral fragrance. The chemical composition of floral fragrance is one of the most extensively investigated areas of floral biology since long because of their commercial value in perfume industry. The role of floral volatiles in pollinators’ attraction is comparatively recent. Earlier studies on the composition of floral fragrance and the role of its constituents in attracting pollinators have been reviewed by Williams (1983) and Knudsen et al. (1993). The fragrance is generally extracted through headspace extraction and analyzed using gas chromatography and mass spectrometry (GC-MS). In the head-space extraction method (Dobson et al. 2005), the flower, intact on the plant is sealed inside a glass chamber and its emitted volatiles are collected by continually purging the glass chamber through a polymer mesh that binds the volatiles. The volatiles are then extracted from the polymer with an organic solvent.

There is great variation among species in chemical composition of floral scents, and the number and relative concentration of constituent volatiles. Floral scents are complex mixtures of small volatile molecules largely made up of monoterpenoid (such as linalool, limonene, myrcene, geraniol, cineole and menthol), sesquiterpenoid, phenylpropanoid, and benzenoid compounds (see Williams 1983; Knudsen et al. 1993, 2006; Dudareva and Pichersky 2000). In many species, fatty acid derivatives and a range of other chemicals containing nitrogen or sulphur are also present.

In a number of species, the responses of pollinators to various fragrance compounds have been analyzed by using either electroantennogram, a technique commonly used to measure the output of the antenna of an insect to the brain for the given odor and/or a bioassay. In the bioassay, insects are exposed to floral scents or fractions of floral scents or synthetic chemicals, and the responses of insects are studied in a cage of suitable size. Chemicals are generally applied to filter paper discs which are placed in glass vials. Some workers have used killed female insect dummies for the assay in a few orchid species that show sexual deception (Ayasse et al. 2003). To remove innate odor from the dummies, they are Soxhlet-extracted in dichloromethane, dried and fixed on insect pins. Such odorless dummies are impregnated with various odor samples to be tested and offered to the pollinators (Ayasse et al. 2003).

The mixture of volatiles emitted by each species is different; no two species, even if they are closely related, have been shown to produce identical mixture of volatiles. The pollinator is attracted to some of them but not to all of them. For example, Huber et al. (2005) analyzed floral fragrance compounds in two closely related orchids, Gymnadenia conopsea and G. odoratissima. They identified 45 volatiles in the flowers of G. conopsea of which three were physiologically active. In G. odoratissima, 44 volatiles were identified of which seven were physiologically active. The specificity of fragrance of a species is established not by individual fragrance compound but a combination of compounds. Insects are able to distinguish complex mixtures of floral scent volatiles from different species and respond accordingly. The emission of scent is markedly reduced after pollination (Schiestl et al. 1997).

Only a limited number of studies have been carried out on the fragrance of vertebrate-pollinated plants. Bat-pollinated flowers generally emit strong fruity smell. Bats are sensitive to odors of esters, alcohols, aldehydes and aliphatic acids (Gibson 2001). Analysis of floral fragrance of 11 bat-pollinated species showed 49 compounds of which 11 were sulphur compounds (Bestmann et al. 1997). In a recent study, two aliphatic ketones, 3-hexanone and 1-hexane-3-one, have been shown to be dominant compounds in the scent of a rodent pollinated species, Cytinus visseri (Johnson et al. 2011). The aliphatic ketone-rich scent in rodent-pollinated plants is in contrast to insect-pollinated plants in which terpenoids, aromatic or non-ketone aliphatic compounds are the dominant scent compounds (Bestmann et al. 1997).

In general, the response of pollinators to floral scent is based on innate abilities of the pollinators superimposed by learning experience (Andersson and Dobson 2003). The fragrance bouquets of flowers elicit strong, innate species-specific attraction to pollinators. However, many pollinators can learn the fragrance bouquets of other species through their association with the reward (Riffell 2011). This learning ability provides a means of flexibility to the pollinators in a changing floral habitat; it enables pollinators to harvest alternative sources of rewards when their preferred species are not in flowering (see Riffell 2011).

Floral Rewards

Flowers offer a range of rewards to pollinators. These include food sources such as pollen, nectar and seeds as well as specialized non-food rewards such as oils and resins, fragrance and brood sites (see Armbruster 2012).

Pollen and Nectar

Pollen and/or nectar are the major rewards for pollinators in most species of flowering plants. Many floral visitors such as honeybees, bumblebees and stingless bees depend solely on floral resources. For such bees, pollen and nectar serve as the exclusive food source not only for adults but also for their larvae. For other floral visitors that feed on other resources also, floral resources only complement their diet. Pollen is highly nutritious and is a rich source of proteins, vitamins, amino acids and minerals (Schmidt and Buchmann 1992; Roulston and Cane 2000). In several species pollen grains, apart from serving as rewards, emit volatiles and attract pollinators (Dobson 1988; Dobson and Bergstrom 1996, 2000). Pollen-foraging insects use pollen odors not only to discriminate between plant species but also between rewarding and non-rewarding flowers on the basis of pollen availability; this enables pollinators to restrict their visits to the rewarding flowers (Schmidt 1982; Dobson et al. 1999). For example, in nectarless Rosa rugosa pollinated by bumblebees, application of pollen volatiles, particularly geranyl acetate to emasculated, freshly opened flowers increased landing frequency of bees significantly (Dobson et al. 1999). When anthers were exchanged between first-day flowers (with pollen reward) and second-day flowers (without pollen reward), flowers with first day anthers received more bee visits.

The nectar is an aqueous solution made up of sugars (largely of sucrose, fructose and glucose) generally ranging from 15 to 45% and small amounts of amino acids (see Nicolson and Thornburg 2007; Heil 2011). Secondary metabolites such as alkaloids and phenolics are also present in the nectar of a range of species (Baker 1977). The nectar contains bacteria (Fridman et al. 2012; Alvarez-Perez et al. 2012) or yeast (de Vega et al. 2009; Herrera et al. 2009) in a number of species. Yeast metabolism in the nectar is thought to contribute to the floral scent (Pozo et al. 2009). Interestingly, the proportion of plant species that contain yeasts in the nectar was found to be highest in those pollinated by birds while those visited only by Hymenoptera showed the lowest values (de Vega et al. 2009). Recently de Vega and Herrera (2013) reported significant changes on sugar composition in flowers visited by nectarivorous ants; it contained significantly more glucose and fructose, and less sucrose when compared to the nectar of ant-excluded flowers. These changes were correlated with the density of yeast cells in the nectar indicating that changes in nectar composition is brought about by yeasts transported by ants.

As the nectar and pollen are harvested by visiting animals, the quality and quantity of available rewards get depleted during the life of the flower. Many pollinators are able to discriminate nectar rewarding flowers from non-rewarding flowers by making use of floral cues that honestly indicate nectar availability. Change of color of older flowers mentioned earlier is one such cue. The nectar in many species is colored and has so far been documented in 67 taxa belonging to a wide taxonomic and geographic ranges (Hansen et al. 2007). The color as well as its intensity is highly variable between species. Pollinators can assess the presence and quantity of nectar in the flowers based on visual cues. Dark brown nectar of Aloe vryheidensis is due to the presence of phenolic compounds (Johnson et al. 2006). Dark purple nectar of Leucosceptrum canum has been shown to be due to the presence of an anthocyanidin, 5-hydroxyfl avylium (Zhang et al. 2012). As 5-hydroxyfl avylium inhibits the growth of bacteria and fungi, it may have a role in preventing growth of bacteria and fungi in the nectar of long-lived flowers (96 h in this species). Two Mauritian plant species, Trochetia boutoniana and T. blackburniana, with colored nectar are pollinated by Phelsuma ornata and P. cepediana geckos, respectively (see Hanson et al. 2006). P. ornata geckos prefer colored nectar over clear nectar in artificial flowers (Hanson et al. 2006). Thus, the colored nectar in these species increases its visibility and acts as a foraging signal to increase pollination efficiency.

Floral nectar of several species is scented (Raguso 2004). Nectar-foraging bees are able to discriminate between flowers that contain nectar from nectar-depleted flowers based on nectar fragrance. In a study on solitary Osmia bees on Penstemon flowers, Howell and Alarcon (2007) reported that nectar-containing flowers received several times more bee visits when compared nectar-depleted flowers. Bees in which antennae were covered with non-toxic silicon (which blocked the olfactory capabilities of bees) did not discriminate between the two types of flowers indicating that bees are able to discriminate nectar-rewarding flowers from nectar-depleted flowers based on nectar volatiles.

Many other pollinator species are able to discriminate nectar-containing and nectar-depleted flowers on the basis of other floral cues. The flowers of Datura wrightii pollinated by a hawkmoth (Manduca sexta) show above ambient emissions of up to 200 ppm CO2 peak at dusk soon after they open (when the flowers are rich in nectar) but floral CO₂ diminishes rapidly after pollinator visit, although visual and fragrance cues persist for a much longer time (Guerenstein et al 2004). The pollinator moth seems to detect such differences with their CO2 sensing organ. In a dual choice assay using artificial flowers, moths preferred to feed from the flowers emitting higher levels of CO2 over those emitting ambient level of CO2 (Thom et al. 2004). Another instance of such floral cue that honestly indicates nectar availability has been reported in night blooming Oenothera cespitosa pollinated by another species of hawkmoth (Hyles lineate). The relative humidity of the headspace of the flowers of this species reaches about 4% above ambient relative humidity during the first 30 min after flower opening, although the other floral traits such as visual and fragrance cues may persist for up to 24 h (von Arx et al. 2012). As floral humidity gradients are largely produced by the evaporation of the nectar, these gradients indicate the amount of reward available to floral visitors. Hawkmoth pollinators can distinguish the differences in RH; they consistently visited flowers with higher humidity levels when compared to those with ambient humidity levels (von Arx et al. 2012). These abilities of pollinators to discriminate rewarding flowers from non-rewarding flowers increase the fitness of both the plants and pollinators.

Fragrant Compounds

Many Neotropical orchids (over 600 spp.) produce fragrant compounds largely terpinoids and aromatics to attract as well as reward male euglossine bees (over 130 spp.). Euglossine bees collect fragrant compounds and store them in their modified hind legs (Dressler 1982). The bees are thought to use these compounds to produce sex pheromones that are released to attract females. Attraction is species-specific in many of these orchids. Species of Bulbophyllum orchid which do not produce nectar are pollinated by fruit flies (Bactrocera spp.) (See Tan 2006). Methyl eugenol is a major fragrance compound in several Bulbophyllum spp. Methyl eugenol attracts its pollinator, male Bactrocera fruit flies, even at a very low concentration. The number of flies visiting the flowers in the morning hours was high and the number dwindles in the afternoon. This is correlated with the release of high methyl eugenol peak in the morning and absence of detectable peak from 14.00 hr. onwards in the afternoon (Tan et al. 2002; Tan 2006). Methyl eugenol is eventually converted into sex pheromone and is released by males to attract females (see Tan 2006). Thus, the fragrance in these orchids acts as an attractant as well as a reward. Apart from orchids, fragrance has been reported to act as attractant as well as reward in some six other Neotropical species such as Anthurium (Araceae), Gloxinia (Gesneriaceae) and Crinum (Liliaceae) (Dresler 1982; Pellmyr 2002; Armbruster 2012).

Oils and Resins

Several plants belonging to some 11 families such as Malpighiaceae, Krameriaceae and Scrophulariaceae provide oils as reward. Solitary bees harvest oils and use them for nest building and as a provision for their larvae (Buchmann 1987; Steiner and Whitehead 1991; Pauw 2006; Renner and Schaefer 2010; Armbruster 2012). Species of Dalechampia (Euphorbiaceae) and Clusia (Clusiaceae) and a few orchids produce resins as rewards. Solitary bees use them for nest building and for antifungal purpose (Pemberton and Liu 2008; Armbruster 2012).

Young Seeds

Many species reward pollinators with nutrients for their larvae in the form of seeds. Fig and fi g wasps, and Yucca and yucca moths are classical examples of providing seeds as rewards for pollinators (nursery pollination). Pollination in Ficus and Yucca is highly specialized; each plant species is generally pollinated by a specific species of insect (see Cook and Rasplus 2003; Machado et al. 2005). These pollination systems are examples of obligate relationships as neither the plant nor the pollinator is able to reproduce without the other. They represent insect-plant co-evolution. A few exceptions to this obligate mutualism in figs, however, have been reported recently (see Cruaud et al. 2012; Wang et al. 2013). Senita cactus (Lophocereus schottii) and Senita moth (Upiga virescens) is another example of nursery pollination; it is not, however, an obligate system as two species of halictid bees also bring about pollination of Senita cactus to a lesser extent (Fleming and Holland 1998). Seed reward has also been reported in a few other species (see Armbruster 2012). In nursery pollination systems the pollinators lay their eggs in the ovary of a proportion of flowers and the larvae consume developing seeds after they emerge. Viable seeds develop from uninfected ovaries/ovules.

Extensive studies have been carried out on the fig and fig-wasp mutualism. Most of the Ficus species produce monoecious flowers in special inflorescences called syconia. The female wasps enter the receptive syconium through the terminal pore, termed ostiole; they bring about pollination with pollen collected from male flowers of the syconium visited earlier, and lay their eggs in a proportion of female flowers. The larvae of the wasp feed on the gall formed in oviposited flowers and the remaining pollinated flowers develop into seeds. The emergence of adult wasps developed from the larvae coincides with the maturation of male flowers. The wingless male wasps are short-lived; they mate with the females and cut an exit hole in the syconium for the females to escape. The females loaded with pollen come out through the exit tunnel and enter another receptive inflorescence (which is in the female phase) through the ostiole to reproduce; they bring about pollination during their movement inside the syconium.  The trees in receptive phase generally occur at low densities. As the adult female wasps live only for a few days, it is critical that the female wasps fly to a receptive syconium usually of a different tree within a short time. Reproductive success of both fi g and wasp depends on transmission of a very strong signal by receptive syconia (Harrison and Shanahan 2005). Several studies have shown that volatile compounds emitted by the receptive syconia are responsible for attraction of their specific pollinator (Khadari et al. 1995; Proffit et al. 2008, 2009). The syconia emit volatile compounds attractive to pollinating wasps only during the period of receptivity. Pentane extracts of receptive syconia of Ficus carica have been shown to attract its pollinator Blastophaga psenes from distances of at least 5 m in the field (Hossaert-McKey et al. 1994). Even non-receptive syconia when their ostioles were painted with pentane extracts of receptive syconia elicit the entry of pollinator wasps (Hossaert-McKey et al. 1994).

Several studies have shown species-specificity of fragrant compounds and their role in attracting specific pollinators. The fragrance emitted by receptive figs of three sympatric Ficus species, namely, F. hispida, F. racemosa, and F. tinctoria could be clearly distinguished by the composition of their fragrance (Proffit et al. 2008, 2009). Specific odors of these three species were tested for attraction of the pollinator (Ceratosolen solmsi marchali) of one of the fig species, F. hispida. Behavioral bioassays showed that the pollinator was attracted only to the specific odor of F. hispida but not to the fragrance of other Ficus species (Proffit et al. 2009). In another dioecious fi g, Ficus semicordata. 4-methylanisole was shown to be the predominant (94–98%) volatile compound emitted by receptive male and female syconia (Chen et al. 2009). This compound was absent in receptive volatiles emitted by two other sympatric fi g species, F. racemosa and F. hispida. 4-methylanisole attracted its pollinator, Ceratosolen gravelyi, in a wide range of concentrations. The fragrance was totally absent in the volatiles emitted by syconia 4 days after pollination. Chemical blends lacking 4-methylanisole were unattractive to C. gravelyi. Receptive syconia of two non-host fi g species, F. racemosa and F. hispida, repelled C. gravelyi. These results on different fi g species clearly indicate that active volatiles are the species-specific signal compounds that attract their obligate pollinator to the host syconia at the receptive stage.

Recent studies of Wang et al. (2013) have indicated that apart from long range volatiles, contact cues after the wasp lands on the syconium also plays a role in determining pollinator specificity.

There is an optimum balance between the number of wasps that enter a syconium and available resources. Entry of too many wasps into the syconium results in oviposition in most of the ovaries; this would lead to eventual collapse of the syconium due to the development of very few or no seeds, compromising the fitness of the plant as well as the wasp species. Several hypotheses have been put forward to explain the stability of this mutualism by preventing wasps from overexploiting the figs (see Dunn et al. 2008). According to one of the hypotheses, the wasps seem to have developed mechanisms to discriminate wasp load in the syconium on the basis of the number of wings left behind by the wasps that have already entered the syconium at and around the ostiole (Ramya et al. 2011). In F. benghalensis and F. microcarpa, Ramya et al. (2011) have reported a strong correlation between the wing load on the ostiole and the wasp load in the syconium. The wasps preferentially entered those syconia from which the wings were removed rather than those on which the wings were retained. Adding wings artificially to the empty syconia also deterred the wasps from entering them, suggesting that residual wings serve as negative feedback regulators for the entry of wasps. In another fi g species, F. racemosa, however, wasps did not use leftover wings as cues. A rapid post-pollination decrease in pollinator attractant as has been reported in some other species may act as a cue and limit pollinator visits, thus minimizing overexploitation of ovaries by mutualistic wasps.

The fi g-fi g wasp pollinator mutualism is also exploited by several parasites of pollinating fi g wasps. The parasites generally do not enter the syconium but insert their ovipositors into the syconium from outside and deposit their eggs in some of the flowers; the offspring of the parasites feed pollinator larvae already present in the galls. For more information on pollinator and parasitoid interaction in figs please refer Nedft and Compton (1969); Dunn et al. (2008); Jousselin et al. (2008) and Ghara et al. (2011) and references therein.

Similarly, each species of yucca is pollinated by a specific species of moth (Tegeticula or Parategeticula) (see Powell 1992; Pellmyr and Huth 1994; Pellmyr 2003). The scent from virgin flowers of the host Yucca glauca was sufficient to attract its obligate pollinator Tegeticula yuccasella (Svensson et al. 2011). A female yucca moth mates with a male in yucca flower and collects pollen in its specialized mouthparts. She carries pollen to another flower, lays its egg in the ovary wall and deposits pollen grains on the stigma. The larva after its emergence in the ovary consumes only a proportion of the seeds from the developing fruit and the remaining seeds develop normally. After the fruits dehisce, the larvae drop to the ground, burrow into the soil and construct cocoons. The larvae remain in cocoons during the winter. Following spring rains, adult moths emerge from the cocoon. By this time yucca plants would be in flowering and initiate new pollination cycle. Some pupae may remain dormant in the soil for more than a year.

Nocturnal Pollination

In a number of species, pollination takes place at night. As visual attraction is not reliable during the night, nocturnally pollinated species relay more on olfactory cues. Nocturnally pollinated flowers are generally drab in color but emit strong smell. Some species show both diurnal and nocturnal pollination (Young 2002; Dar et al. 2006; Muchhala et al. 2009; Dafni et al. 2012). Moths and bats are the most important nocturnal pollinators. Moths have superposition compound eyes, where each rhabdom receives light through a wider aperture consisting of hundreds of facets. This increases the number of photons caught in dim-light conditions and thus facilitate nocturnal vision (Warrant 2004).

Bees in general are diurnal; their foraging activity begins at dawn and ends at dusk. However, several bees have developed abilities to forage in dim light. Many of them such as Xenoglossa fulva (Linsley et al. 1955), Xylocopa tabaniformis (Janzen 1964), Ptiloglossa guinnae (Roberts 1971) and Megalopta genalis (Warrant et al. 2004) are able to forage during crepuscular (twilight) periods. Some bee species such as Apis dorsata and Apis mellifera (Dyer 1985) have been reported to forage nocturnally on moonlit nights. However, Xylocopa tranquebarica has been shown to be a true nocturnal bee (Somanathan and Borges 2001; Kelber et al. 2006; Somanathan et al. 2008); it can fly and navigate even during the moonless parts of nights and so far is the only known bee to be able to do so. Unlike typical nocturnal insects such as moths which have superposition compound eyes, all bees including the nocturnal species, have apposition compound eyes that are well adapted to diurnal vision. X. tranquebarica has also been shown to exhibit color vision under moonlight, twilight, and even starlight conditions and is the only insect with apposition eyes to exhibit color vision under such dim light conditions. Opposition eyes of X. tranquebarica are larger with large facets and very wide rhabdoms which make them very sensitive to dim light (Somanathan et al. 2009).

Bats involved in pollination generally have longer tongues when compared to insectivorous species. Bat-pollinated flower syndrome includes nocturnal anthesis, drab coloration and musty, fetid odor. Their flowers are generally placed away from the foliage at the tips of branches or are borne directly on the trunk or branches (cauliflory). In many deciduous trees flowering is initiated after defoliation. Flowers of bat-pollinated species may be tubular or radially symmetrical and produce relatively large amounts of hexose-rich nectar (Flemings et al. 2009). In flowers of bat-pollinated Ochroma species as much as 7–15 ml of nectar has been reported (Faegri and van der Pijl 1971). Amongst floral traits, nocturnal anthesis, placement of flowers away from foliage and production of larger amount of nectar are found in almost all bat-pollinated flowers (Fleming 2009). Bats are long distance pollen dispersers. They use vision, olfaction and echolocation to locate flowers. They have keen sense of smell that helps in long-distance location of flowers. Bat-pollinated species generally occur in low density and bats play a crucial role in maintaining genetic continuity of their population (Fleming et al. 2009). New World Bats (Phyllostomidae) can produce ultrasonic sound to locate flowers. The sound is reflected by the petals of bat-pollinated flowers and bats have the ability of recognize this reflected sound (echolocation). The erect petals of the flowers of many species such as Mucuna holtonii have been shown to reflect sound pulses produced by echolocation bats that visit the flowers for their copious nectar (von Helversen and von Helversen 1999). Echolocation ability is not developed in most of the old world bats (Pteropopodidae); they depend largely on olfactory and visual cues to locate the flowers. The eyes of all bats are well adapted to low illumination. Old World fruit bats have color vision whereas New World bats can see only in black-and-white.

Pollination by Ants: A Rare and Ineffective Pollination Syndrome

Ants are among the most abundant insects on earth and visit flowers frequently. Also, the fossil record shows that ants were present during the explosive radiation of the angiosperms in the late cretaceous period even before the presence of bees (Beattie 1985). Surprisingly, ant pollination has not evolved as a major pollination syndrome in angiosperms. So far, ant pollination has been documented in less than 20 species of herbs and shrubs (Shivanna 2010). Although ants are active 24 hr. a day, ant-pollination does not seem to have evolved in any of the night blooming plant species (Beattie et al. 1984; Beattie 1985; Sharma et al. 2009; see Shivanna 2010). Interestingly, ant-pollination has not been reported so far in any tree species. The following are some of the hypotheses put forward to explain the rarity of ant pollination: 1. Ants bring about largely geitonogamous self-pollination as their movements are usually restricted to within the plants. 2. Secretion of antibiotics from metapleural glands present in most of the ant species (as a defense against bacteria and fungi) reduces pollen viability. 3. Pollen transfer is not efficient as their body surface is smooth and pollen grains do not adhere well to their bodies. 4. Ants groom their bodies too frequently with the result very little pollen is left on the body for transfer to the stigma.

Most of the observations recorded in various plant species do show that the movements of ants are restricted within the plant and most of the pollinations are geitonogamous. They may move between plants and bring about cross-pollination infrequently only in small herbaceous species that grow in high density in small patches (Hickman 1974). Studies on the effects of antibiotics show differential potency among ant species and differential sensitivity among pollen species (Beattie et al. 1984). Thus, inhibition of pollen function by exposure to the surface of ants may provide only a partial explanation for the rarity of ant pollination systems. Smoothness of the body surface of ants is not universal. Many ant species are as hairy as bees or are covered with bristles suitable for carrying pollen load. Although pollen load recorded on ants are generally low (10–30), pollen load of over 200 grains per ant has been recorded in a few species. Many investigators have discounted the suggestion that ants groom their body frequently since bees also groom their body frequently (Beattie et al. 1984). Thus, available evidences support, to some extent, the first two hypotheses to explain the rarity of ant pollination.

There is very little information on floral attractants for which ants are responsive. Although some investigators have suggested that floral scent could play a key role as floral attractant, there are no experimental evidences. The nectar acts as the reward for ants. For most of the species, ants are not exclusive pollinators; they share pollination services with several winged insects. For example, Balanophora flowers are visited by a variety of flying insects besides ants (Peakall and Beattie 1989). However, specialized, exclusive ant-pollination has been reported in two Australian orchids. Leporella fimbriata (Peakall 1989) is exclusively pollinated by male winged ants of Myrmecia urens. The other orchid, Microtis parviflora (Peakall and Beattie 1989) is pollinated by flightless worker ants of Iridomyrmex gracilis, through sexual deceit.

Non-Mutualistic Pollinations

In a number of plant species, pollination is achieved through non-mutualistic interaction (Dafni 1984; Renner 2006; Bernklau 2012). Such plants advertise the presence of reward without offering the reward and achieve pollination through deceit. Non-mutualistic interactions have evolved in all major groups of flowering plants. Orchids form the most important group of non-mutualistic pollination; about one third of the orchids (over 10,000 species) are estimated to be deceptive (Renner 2006; Jersakova et al. 2006). Most of the species producing reward-less flowers are pollinated by insects; however, two of the reward-less species have been reported to be bird-pollinated and one bat-pollinated (Renner 2006). Many animal visitors also harvest the rewards of the flower without effecting pollination. Such visitors are referred to as pollen or nectar robbers. Nectar robbers generally enter the flower from the side by piercing the corolla tube and harvest the nectar without coming in contact with the anthers and stigma. Pollen robbers forage the pollen from the anthers but do not come in contact with the stigma.

Food Deception

Plants of non-rewarding species (mimic) coexist with rewarding species (model) and the flowers of the mimic resemble those of the model. Floral visitors draw rewards from the model but do not discriminate against non-rewarding flowers of the mimic. Food deceptive orchids generally attract both male and female pollinators. In many food deceptive orchids, floral color seems to play more dominant role than the extent of morphological similarity of the model and the mimic (Gigord et al. 2002; Streinzer et al. 2010). A South Africa orchid, Disa ferruginea (Johnson 1994; Newman et al. 2012), for example, depends on a butterfly species, Aeropetes tulbaghia for pollination. In the Cape Province (western part of its range) a red flowered form of this orchid occurs and attracts butterfly visits by imitating red, nectar-producing flowers of Tritoniopsis triticea (Iridaceae). The butterfly does not discriminate between the nectarless orchid flower and the nectar-producing model flower in sympatric populations. Interestingly, in the Langeberg Mountains (eastern part of its range) an orange-flowered form of D. ferruginea occurs and mimics the orange, nectar-producing flowers of Kniphofia Uvaria (Asphodelaceae). The pollinator butterfly preferred red flowers in the western part where its main nectar model has red flowers while in the east it preferred orange flowers, where its main model has orange flowers. These results indicate that the flower color in D. ferruginea is adaptive and driven by local color preference of its pollinator. Several studies have indicated that the food deception is largely mediated by visual signals and olfactory signals do not seem to play any major role. Studies of Galizia et al. (2005) on a food deceptive orchid, Orchis israelitica and its model Bellevalia flexuosa (Liliaceae) showed that fragrance compounds in the mimic were quite different and weak when compared to the model. The odor of the model and the mimic elicited distinct activity patterns in the brain of its pollinator bee, indicating that the bee can easily distinguish the flowers of the model and the mimic on the basis of their odors. As the bee does not discriminate between the flowers of the two species, odor obviously does not play any role in signaling. In another study in which both the mimic and the model are orchids, Salzmann et al. (2007) reported that both the model (Anacamptis coriophora) and the mimic species (A. morio) emit complex odor bouquets. Several components of the fragrance of the model triggered electrophysiological responses in olfactory neurons of honey-bees and bumble-bees. The scent of the mimic, however, was too weak to elicit any electrophysiological responses (Salzmann et al. 2007), negating the role of olfactory signals in deception. The fragrance of the model, however, may still play a role in long distance attraction of the pollinators to the location where the model and the mimic grow.

Sexual Deception

A large number of orchids achieve pollination by sexual mimicry. Most of the species showing sexual mimicry are pollinated by species of Hymenoptera. Pollination through sexual deception is often highly specific and attracts a particular species of pollinator. Odor signal plays a dominant role in sexual mimicry (see Galizia et al. 2005). Flowers of non-rewarding species mimic visual, olfactory and tactile cues of conspecific females of the pollinator. Olfactory cues are similar to sex pheromones of receptive females. For example, sexually deceptive Australian orchid, Chiloglottis trapeziformis attracts males of its pollinator wasp, Neozeleboria cryptoides, by emitting a unique volatile compound, 2-ethyl-5-propylcyclohexan-1, 3-dione which is also produced by female wasps as a sex pheromone (Schiestl et al. 2003). The visitor lands on deceptive flowers and tries to copulate (pseudo-copulation) and during this process, brings about pollination. In sexual deception, pollinators are first attracted to the fragrance signals from a distance but in close vicinity they seem to be guided exclusively by visual signals (Streinzer et al. 2009).

Kullenberg (1956) was the first to show the role of fragrance in sexual deception in orchids. Until then it was thought that sexual mimicry in orchids was mediated only by visual cues. Kullenberg (1956) covered the flowers Ophrys lutea with a piece of cloth for a few hours and showed that the cloth attracted its pollinator, Adrena bees. Subsequent studies confirmed that both visual and volatile chemicals are involved in attraction of male bees. Flowers showing sexual mimicry produce complex odor bouquets of a large number of volatiles that are similar to those produced by the females of their pollinators (see Ayasse et al. 2003). Ophrys is one of the well-investigated orchids showing sexual mimicry. Different species of Ophrys produce distinctive qualitative and quantitative blends of floral fragrance compounds. The fragrance compounds produced by Ophrys flowers of several species have been shown to be similar to those found in the mandibular and/or Dufour’s glands of their pollinators (see Williams 1983; Borg-Karlson 1990). Ayasse et al. (2003) studied the details of fragrance in the flowers of O. speculum and its pollinator, Campsoscolia ciliata. In field trials, they used female dummies as control. The male insects hardly responded to odorless female dummies but were attracted in high frequency to freeze-killed virgin females and intact O. speculum flowers; Labellum extracts of O. speculum flowers and cuticular extracts of virgin females. They identified eight electro-physiologically active compounds. Of these, 9-hydroxydeconic acid was the major one in both the labellum and the insect. In Ophrys sphegodes, pollination has been shown to result in changes in odor components (Ayasse et al. 2000; Schiestl et al. 1997, 2003) and thus the attractiveness of pollinated flowers to its pollinator male bees (Andrena nigroaenea) decreases. Pollinated flowers showed a significant increase in farnesyl hexanoate. Flowers treated with farnesyl hexanoate were found to be less attractive to males indicating that this chemical acts as a repellent and guides pollinators to un-pollinated flowers (Schiestl and Ayasse 2001). In another study (Stokl et al. 2009), the odor compounds were analyzed in three species of Ophrys, O. lupercalis, O. bilunulata, and O. fabrella. All the species grow sympatrically and are pollinated by three species of Andrena. These Ophrys species use the same odor compounds for pollinator attraction, but in different proportions. Thus a change in the concentration of the constituents of floral odor can result in the attraction of a new pollinator species that acts as an isolation barrier towards other sympatrically occurring Ophrys species (Stokl et al. 2009).

Brood Site Deception

Several species mimic brood sites and attract insects whose larvae feed or lay their eggs on decaying organic matter such as dung, decaying feces and carrion. Brood site mimicry has been documented in species belonging to some ten families such as Annonaceae, Araceae and Aristolochiaceae (see Wiens 1978; Bernclau 2012). Their flowers mimic the odors of dung and/ or carrion to attract coprophilous beetles and flies that oviposit or feed on dung or carrion. These odors are very unpleasant to humans. The odors are composed of sulfide compounds, ammonia, alkylamines, cadaverine and putrescine. Fecal-like odors are also produced by skatole and indole compounds (Meeuse 1978; Dettner and Liepert 1994).

In many species showing brood site mimicry, the odors are enhanced by the production of heat (thermogenesis). Floral thermogenesis appears to be more common in basal angiosperms and Eumagnoliids when compared to other angiosperms. It has been reported in 3 families (Nymphaeaceae, Schisandraceae and Illiciaceae) of basal angiosperms, 6 families (Annonaceae, Araceae, Arecaceae, Aristolochiaceae, Cyclanthaceae and Magnoliaceae) of Eumagnoliids, and only 2 families (Nelumbonaceae and Refflesiaceae) of Eudicots (see Thien et al. 2009). Detailed studies have been carried out on Victoria amazonica, a thermogenic basal angiosperm (Prance and Arias 1975) in which beetles are the major pollinators. Flowers are protogynous; they are in the female phase on day 1 and in the male phase on day 2. Flower buds open in the evening of day 1, and produce heat (>11ºC above ambient) and strong odor that attract beetles. Flowers close by the next morning trapping the beetles followed by gradual loss of heat and odor. The flowers gradually turn from white to deep purplish-red. By evening of day 2, the stamens release pollen and the flowers reopen. The escaping beetles pick up pollen and fly to another flower of day 1 which is thermogenic and scented. When beetles are trapped inside the flowers, they consume starchy stylar appendages. High temperature inside the flower also seems to act as an energy reward for the beetles in maintaining their body temperature at endothermic level (Seymour and Matthews 2006). In general, thermogenesis and trapping mechanism are associated with food deceptive species which do not provide any rewards to the visitor. But in N. amazonica the flowers seem to provide rewards in the form of nutrients and heat energy (Seymour and Matthews 2006).

In species of Aristolochia (Murugan et al. 2006; Trujillo and Sersic 2006) the flowers are differentiated into an expanded limb, a long narrow tube of various lengths and a basal expanded utricle. The inner surface of the tube is lined with downward-pointed hairs which facilitate the entry of insects into the floral chamber where the sexual organs are located but not their exit. The flower produces an odor similar to decaying plant tissues, mimicking the natural oviposition substrate of their pollinators. Flowers are protogynous and the stigma is receptive at the time of flower opening. The pollinators (mostly Dipterean flies) are attracted by the color and odor of the flower, enter the floral chamber through the narrow tube and are entrapped for 24–48 h as the downward-pointed hairs on the inner surface of the floral tube prevent their exit. By the time anthers dehisce and the insects get coated by the pollen, the hairs in the tube become flaccid and start senescing allowing the flies to escape. The flies, coated with pollen enter freshly opened flower with receptive stigmas and bring about pollination.

The flowers of some species of Aristolochiaceae, Araceae, Orchidaceae and Saxifragaceae are pollinated by fungus gnats. The flowers or floral parts of many such species mimic oviposition sites (gill fungi) of fungus gnats (the larvae of which feed on gill fungi) (Vogel and Martens 2000; Okuyama et al. 2004 and references therein). Some of them also have trap mechanism. The visual mimicry in combination with emission of odor characteristic of fungi attracts fungus gnats which oviposit and bring about pollination. The pollination in Mitella (Saxifragaceae) species differs from other fungus gnat-pollinated species in that the fungus gnats do not lay their eggs in the flower and seem to consume small amount of the nectar available (Okuyama et al. 2004).

Aphidophagus hoverflies (Syrphidae) generally lay their eggs in places of aphid infestation as their larvae feed on aphids. The orchid Epipactis veratrifolia is exclusively pollinated by five species of aphidophagous hoverflies (Ivri and Dafni 1977) and the authors suggested that the flowers in this species mimic the shape and color of aphids to attract syrphid flies. However, recent studies (Stokl et al. 2010) have shown that the flowers produce α- and β-pinene, β-myrcene and β-phellandrene (which are similar to the alarm pheromone released by several aphid species). The pheromones attract aphidophagus hoverflies. The flies bring about pollination during egg deposition and nectar feeding. As these pheromones are not species specific (they attract several species of syrphid flies) and also the flies consume nectar, Stokl et al. (2010) prefer to call it generalized mimicry rather than typical mimicry.  Evolution of rewardless flowers and their stability has been discussed by some investigators (Pellmyr 2002; Renner 2006). Pollinators constantly encounter rewardless flowers during most of the foraging bouts even within the rewarding species. This regularity in encountering rewardless flowers makes pollinators follow flexible foraging. Because of this foraging flexibility pollinators do not discriminate more strongly against rewardless species (Pellmyr 2002; Renner 2006). It has been suggested that transiently rewardless flowers may have facilitated the evolution of rewardlessness as a stable strategy because their presence lowers the strength of selection on pollinators to consistently discriminate against lack of rewards (Renner 2006).

Restriction of Pollinators

Any natural community is made up of a number of plant species and a range of animal species, many of which are potential pollinators. However, each plant species attracts only a proportion of potential pollinators to visit its flowers in a sustainable way and use them for pollination services but prevents the visit of several others present in the community. So far studies on pollination ecology have largely been on floral attractants and rewards. However, for an efficient pollination system, restriction of visits to a reasonably limited number of pollinator species is important to ensure their visits to largely conspecific flowers. Although there are several studies on the factors which act as filters to prevent the visits of some animal species, there is hardly any discussion on this aspect in reviews on pollination ecology. Available evidence indicate that restriction of floral visitors may act at different levels—morphology of the flowers, species–specific fragrance, and quantitative and/or qualitative features of the nectar and of pollen.

Morphological Filters

The role of morphological elaboration of floral traits that permit some species of pollinators (and prevent some others) was known since long. Faegri and van der Pijl (1971) referred to such adaptations that permit only a particular group of animals as ‘harmony’ between the visitor and the flower. Lack of harmony prevents floral visitors. One such floral elaboration is the evolution of bilateral symmetry. It enables the flowers to guide the approach pattern of the visitors to harvest the rewards efficiently. Flowers of such species are visited only by those animal species that are able to find and harvest the rewards. Another type of elaboration is the evolution of the flowers with a long corolla tube or a spur in which nectar is located. Plant species with such flowers are very common. In southern Africa for example, long-tube flowers have been reported in about 200 spp. of at least 10 families. A large number of orchids have spurs of various lengths often up to 40 cm. In flowers with long corolla tubes, only insects that have the proboscis of matching length or birds with matching beak length can harvest the reward. Pollinator species with different tongue lengths tend to specialize on plant species with matching spur/corolla tube lengths. Such correlation has been highlighted in several studies on plant-pollinator communities (Pleasants 1983; Pellmyr 2002; see also More et al. 2007; Santos-Gally et al. 2013). In a bumblebee community with different tongue lengths (short, medium and long), a relationship between the corolla length of the plant species and the tongue-length of the bees that visit them has been reported even in earlier literature (Pleasants 1983). Such plant-pollinator relationship reduces competition by restricting the visits of the bees of various tongue lengths to flowers of matching corolla length. The bees with tongue-length shorter than corolla length would not visit the flowers of long corolla tubes since they cannot reach the nectar. Avoidance of the bees with long tongue-length to visit the flowers with shorter corolla tubes has been explained on the basis of less efficient foraging (of long-tongued bees on short corolla tubes) when compared to those with appropriate tongue length. Another reason put forward to explain this avoidance is the lower net profit to long-tongued bees in visiting flowers of short corolla tube as they have to compete with short-tongued bees which would reduce the nectar reward to unprofitable levels (see Pleasants 1983; More et al. 2007).

Pollination Syndromes

The role of morphological traits of flowers and their relationship to the pollinators has been elaborated in the form of pollination syndromes (Faegri and van der Pijl 1971). Since the time of Darwin, the traditional concept on the evolution of pollination systems has been toward specialization leading to a greater refinement in making the pollinator and the flower mutually inter-dependent. Various pollination syndromes such as beetle-pollinated, bee-pollinated, butterfly-pollinated and bird-pollinated were identified and described (Table 1). A syndrome is a combination of phenotypic traits associated with attraction and utilization of specific types of animals for Pollination syndromes do not form closed compartments and do not represent exclusiveness of the respective pollinators to the flowers of a particular syndrome. They also do not indicate the presence of all the specific traits of a syndrome in all plant species of that syndrome. Pollination syndromes simply represent traits that are over-represented in flowers that attract specific types of pollinators; they do not exclude other visitors (Faegri and van der Pijl 1971; Pellmyr 2002).

Specialization in floral traits leading to pollination syndromes was explained on the basis of selection pressure exerted by the pollinators. Different pollination syndromes are often present in closely related species. For example, Petunia axillaris (flowers white, nocturnally scented) is pollinated by nocturnal hawkmoth (Manduca sp.) while P. integrifolia (flowers colored and unscented) is pollinated by diurnal bee (Hexantheda sp.) (Ando et al. 2001). Similarly, unrelated plants also show similarities in floral traits; such similarities are attributed to evolutionary convergence resulting from selection by shared types of pollinators (Faegri and van der Pijl 1971).

Many recent studies have shown that in a majority of species, flowers are visited by a number of animal species and the advanced level of specialization is seen only in a limited number of species. Based on these studies traditional concept of pollination syndromes has been questioned (Waser et al. 1996; Waser and Ollerton 2006). However, many pollination biologists (Pellmyr 2002; Fenster et al. 2004; Bronstein et al. 2006; Fleming et al. 2009; Mitchell et al. 2009) argue that the concept of pollination syndromes helps in understanding the mechanisms of floral diversification and the convergence of floral form across angiosperms pollinated by similar pollinators. The concept of generalization and specialization is used in terms of the number of pollinators a plant species attracts. Most of the plants are pollinated by two or more species, indicating selection against excessive specialization; this gives pollination assurance even under spatio-temporal variation in pollinator abundance.

One of the major limitations of assessing the number of pollinators of a given plant species is the lack of data on actual pollinators based on pollen transfer. In some plant species, the number of animal species visiting the flowers is quite large. Over 70 species of insects have been reported to visit the flowers of Andira inermis (Fabaceae) (Frankie et al. 1976; see Bawa 1990). In such plant species true pollinators, on the basis of pollen transfer, are likely to be much less. Some of them may be casual visitors to explore the availability of resources and some others may predate on insects that visit the flowers; and yet others may rob the pollen and/or nectar without bringing about pollination. Further, the frequency of some of the visitors may be too low to have any impact on pollination efficiency of the plant species. The number of effective pollinators on the basis of pollen transfer has been documented only in a limited number of species. In one of the Syzygium species, S. heyneanum, out of 23 species of floral visitors’ only three species turned out to be effective pollinators (Giby Kuriokose, P.A. Sinu and K.R. Shivanna, unpublished). In the absence of such detailed information, records of the number of visitors to flowers may be misleading in assessing the number of pollinators of plant species and the extent of its specialization/generalization. In most of the plant species with unusually large number of visiting species, the number of pollinators is likely to be much less than the number of visitors reported. Generalization and specialization also applies to pollinators. Generalists are those that forage the flowers of a number of plant species and specialists are those that forage the flowers of a limited number of species. Specialization is not generally in the evolutionary interest of flower-visiting animals. Many animal species specialize when their preferred species is flowering but in the absence of their preferred species, utilize a wide range of plant species.

Co-evolution

Guided by Darwin’s concept of co-evolution (reciprocal selection between plant and pollinator), earlier pollination biologists interpreted most of the diversification of plants and pollinators as the result of co-evolution. Subsequently it became apparent that co-evolution is largely restricted to a limited number of species which show obligate relationship between plant species and its pollinator such as figs and fi g-wasps, yucca and yucca-moths (Pellmyr 2002; Bronstein et al. 2006), and several orchids. In these highly specialized pollination systems in which each plant species is pollinated by just one animal species, the specificity of pollinators is apparently the result of mutual adaptation of the plant and pollinator species. Attraction of species-specific pollinator is largely mediated through morphological elaboration of the flower and/or species-specific fragrance. Other species do not visit the flowers because these lack matching flower structure or specific fragrance.

The well-recognized example of co-evolution, first proposed by Darwin, is the length of floral spurs, and the length of insect tongues and beaks of birds (see Pellmyr 2002). While explaining the exceptionally long nectar spur (40–50 cm) of the star orchid in Madagascar, Angraecum sesquipedale, Darwin proposed that a co-evolutionary ‘race’ (between spur length of flowers and tongue length of insects) acted as a driving force for the directional increase in the length of plant’s spur (where the nectar is concealed) and the tongue of its pollinator. Floral visitors whose proboscis is longer than the length of the spur/corolla tubes of its host flower, rob the nectar without effecting pollination (as the insect need not push the proboscis deep enough to bring its mouthparts in contact with the anthers and the stigma). For effective pollination it is advantageous for the plant if the length of the spur/corolla tube exceeds the length of its pollinator tongue so that the pollinator is forced to push the proboscis deep into the spur thus bringing the mouthparts in contact with the anthers and the stigma. This would induce directional selection for longer spurs in the population. Evolution of deeply concealed nectar induces reciprocal selection in tongue length in the pollinators. This process can continue as long as genetic variation is available in both the partners. Correlation of plant fitness with the length of the spur was supported by studies of Nilsson (1988) in an orchid, Platanthera bifolia, in which spur length ranges from 30 to 50 mm. He shortened spur length by pushing the nectar upward and tying the spur with a thread and studied the responses of its pollinator moth. When the spur length was shorter than the length of the proboscis, the moths pushed the proboscis only up to the level of the spur necessary to forage the reward; their mouth parts did not come in contact with the sexual organs of the flower which is necessary for the removal of pollinaria and its deposition on the stigma. Shortening of the length of the spur proportionately reduced both the removal of pollinaria (male fitness) and their deposition on the stigma (female fitness) of visited flowers. Longer the spur length better was the removal of pollinaria and their deposition.

In contrast to many examples documenting the effects of selection pressure exerted by pollinators on plants, plant-mediated selection on pollinators is limited except in those highly specialized obligate pollination systems. This is largely because of the difficulty of measuring life-time fitness of pollinators which are mobile. Also plant-mediated selection on animals is indirect since the plants do not exert direct effects on the pollinator’s gene flow but only on their food intake (see Pellmyr 2002). However, plant-mediated selection on animals has been correlated to differences in bill length and shape in male and female individuals of the purple-throated Carib (Eulampis jugularis) (Temeles et al. 2000). In this species the female has significantly longer and curved bill when compared to the male. The males primarily visit nectar-rich patches of Heliconia caribaea which has short, straight flowers corresponding to their bills, whereas females are the primary visitors to flowers of Heliconia bihai, which has long, curved flowers matching the size and shape of their bills. On the basis of these studies sexual dimorphism in E. jugularis has been suggested to have evolved in response to differences in the floral traits of their primary floral resources. Because of this partitioning, heterospecific pollen transfer is avoided. Their subsequent studies provided evidence to indicate that differences in bill length and curvature between sexes in hermit hummingbirds are also associated with differences in their food plants (Temeles et al. 2010).

Strong selection on bill size has also been documented in Hawaiian honeycreeper in response to dietary shift (Smith et al. 1995). The honeycreeper, Vestiaria coccinea, used its long curved beak to access nectar in tubular lobelias. These lobelias have now become extinct or very rare. The bird has now changed its diet to the nectar of Metrosideros polymorpha, the flowers of which lack corolla. Comparison of the museum specimens of V. coccinea collected before 1902 with the extant individuals revealed a significant reduction in bill length in extant specimens.

Restriction of Pollinators in ‘Open Flowers’

In ‘open flowers’, which are generally symmetrical without any structural elaboration, pollen and nectar are exposed and practically any animal that visits such flowers can harvest both pollen and nectar rewards. Even in such open flowers, there is some degree of specificity; all animals present in the habitat do not visit the flowers. Syzygium cuminii (our unpublished observations) for example has a typical open flower. Apis dorsata and one species of ant are the major visitors to the flowers; a few other insects such as Apis cerana, Trigona sp., a few wasps and a butterfly make occasional visits but their frequency is too low to have any impact on pollination efficiency of the species. Flowers of Passiflora spp. also have brightly colored open flowers with exposed pollen and nectar. However, the number of floral visitors reported in different species of Passiflora is also limited (Varassin 2001; Shivanna 2012b). In a study on pollination ecology of annual weeds, Shivanna (unpublished observations) observed that Apis cerana was a regular and frequent visitor to the open flowers of Triumfetta rhomboidea for pollen, but they never visited open flowers of Abutilon indicum, Urena lobata and Melochia corchorifolia, which were flowering sympatrically next to T. rhomboidea. These studies and several others clearly show that although the rewards in such open flowers are accessible to any visitor, only a limited number of animal species visit the flowers in spite of the presence of a large number of insets in the habitat. In many such species, the number of animal species visiting flowers may not be more than those flowers with some structural specialization. In such species with open flowers, morphological filters do not operate; however, they have other filters to restrict the number of visiting species. Several studies have shown that the fragrance, nectar and even pollen may act as filters to restrict the number of species to visit their flowers. These filters may operate in specialized flowers also.

Fragrance Filters

Most of the studies on floral fragrance are concerned with their role in attracting floral visitors. As pointed out earlier, the composition of fragrance is unique to each plant species and attracts a specific pollinator or a group of pollinators. The fragrance acts as a filter in many plant species as their flowers lack suitable fragrance to attract other species. As pointed out earlier, in species with obligate specialization, fragrance appears to be the main filter. Several studies have highlighted the role of fragrance in repelling some floral visitors. Omura et al. (2000) reported that only a few species of insects visit the flowers of Osmanthus fragrans despite their strong scent and vivid coloration. A butterfly, Pieris rapae, a potential visitor, never visits the flowers of this species. They showed that isopentane fraction of floral volatiles, particularly γ–decalactone, is a strong repellent to this butterfly. Willmer et al. (2009) have reported that floral volatiles act as repellents to ants in some species of Acacia. A meta-analysis of 18 studies on the response of animals to floral scents by Junker and Bluthgen (2010) have also highlighted the dual function of floral scents; obligate floral visitors are attracted to floral scent while those which are facultative and antagonists are repelled by floral scents.

Nectar Filters

Studies on nectar have highlighted largely the role of its nutritive components, sugars and amino acids, as rewards for the visitors. The amount of nectar present in the flower and its sugar concentration are, to some extent, correlated with the type of animals visiting the flowers (Baker and Baker 1983). Bee-visited flowers generally have lower amount of nectar with higher sugar concentration while bat- and bird-visited flowers have higher nectar volume with lower sugar concentration. These features are well recognized and form a component of pollination syndromes.

Although the presence of non-nutritive metabolites such as alkaloids and phenolics in the nectar was known since long (Baker 1977), only limited number of studies are available on their role in attracting or repelling floral visitors (Stephenson 1981; Adler 2000; Adler and Irwin 2005; Raguso 2004; Wright et al. 2013). In an earlier study Stephenson (1981, 1982) showed that secondary compounds of the nectar of Catalpa speciosa are effective in filtering the visitors. The floral nectar of this species contains iridoid glycosides, catapol and catalposide, which adversely affected potential nectar thieves (ants and a skipper butterfly, Ceratomia catalpa). The legitimate diurnal bee pollinators were not affected by these glycosides. Some South African species of Aloe produce dark brown nectar with a bitter taste because of the presence of phenolic compounds in high concentration (1.2–1.5 mg/ml). Detailed studies of Johnson et al. (2006) on one such species, A. vryheidensis clearly indicated that the main effect of phenolics in the nectar is to repel inefficient-/non-pollinators. Bulbuls and white eye, which are effective pollinators are not affected by the bitter taste of nectar. Bees and sunbirds which are not effective pollinators do not even approach the flower as they do not like dark brown nectar. Naive honeybees and sunbirds, however, approach nectar during initial visits but avoid it completely in subsequent visits. Thus, the dark phenolic component of the nectar functions as a floral filter by attracting some animals and deterring others by its taste. In another study Liu et al. (2007) have shown that buck wheat nectar phenolics in combination with 15–35% sugar syrups attracted Apis cerana but acted as deterrents below or above this sugar concentration range. Such synergism between phenolics and sugar may provide a novel mechanism for plants to preferentially select some pollinators and to reduce energy investment in nectar.

Presence of alkaloids in plants is quite common and they act largely as deterrents to herbivores. As mentioned earlier, the nectar of some species contains alkaloids such as nicotine and caffeine. In Nicotiana atenuata pollinated by a moth and hummingbirds, presence of nectar nicotine decreased visitation time and the volume of nectar removed for both the pollinators, but increased the number of their visits (Kessler and Baldwin 2006). Interestingly, pollinators removed nearly 70% more nectar from nicotine-silenced plants (through genetic transformation) in which nectar lacked nicotine completely when compared to control plants. On the basis of these results they hypothesized that nectar repellents optimize the number of visits per volume of nectar produced, allowing plants to keep their nectar volumes small. The flowers of Coffee and Citrus contain low concentration of caffeine which does not deter their pollinator bees (Wright et al. 2013; see also Chittka and Peng 2013). Interestingly, the presence of caffeine in the nectar has been shown to significantly enhance the memory of bees; honeybees rewarded with caffeine were three times more likely to remember the reward when compared to honeybees rewarded with sucrose alone.

Floral nectar of several species is scented. The nectar scent compounds in some species have been shown to be a subset of the compounds emitted by the surrounding floral tissues, while in some others they contain unique scent compounds compared to the floral tissues (Raguso 2004). In a detailed study, Kessler and Baldwin (2006) analyzed secondary metabolites in floral and nectar fragrance of Nicotiana atenuata and studied their responses on their pollinators (one moth and two hummingbird species) and a nectar robber (one ant species). Various components of the fragrance showed positive or negative effects on different animals tested. Compounds from the same biosynthetic class tended to evoke similar responses. Benzyl acetone attracted the pollinators (moth and hummingbirds). Methyl salicylate repelled hummingbirds and ants but attracted moths. Although studies on nectar filter so far are limited, they have clearly shown that the nectar can act as one of the effective filters favoring some floral visitors and deterring others (Kessler and Baldwin 2006, 2011).

Pollen Filters

Nutritional quality of pollen is highly variable; some of them lack several essential nutrients and some are poor in proteins (Roulston and Cane 2000; Rasmont et al. 2005) and yet others contain secondary compounds which are repellent or toxic to insects (Pimentel De Carvalho and Message 2004; see Hargreaves et al. 2009; Sedivy et al. 2011). There is much evidence to indicate that pollen can act as a filter to select floral visitors.

Several studies have analyzed pollen loads of bee species to check the extent of their host specificity. Pollen load analyses of 35 species of the genus Chelostoma (Megachiledae) revealed that 33 species to be pollen specialists at the level of plant family or genus (Sedivy et al. 2008). In another study on pollen load analyses, Muller and Kuhlmann (2008) reported that 26 species of the genus Colletes (Colletidae) out of the 60 species analyzed, were specialists at the level of plant family, subfamily or genus and the remaining 34 species were pollen generalists to varying degrees, visiting the flowers of up to 15 plant families. Fourteen of the specialist species were found to harvest exclusively or predominantly the pollen of Asteroideae. However of the 34 pollen generalist species Asteroid pollen formed a very small proportion (<3 27="" a="" acids.="" adaptations="" although="" amino="" and="" asteroideae="" at="" based="" be="" bees="" bl="" both="" bumblebees.="" bumblebees="" choices="" collected.="" collected="" colonies="" constrained.="" contain="" content="" developed="" did="" differed="" digest="" digestibility="" essential="" feeding="" flowers.="" foragers="" generalist="" generally="" have="" higher="" honeybees="" host="" hosts="" in="" investigated="" leonhardt="" load="" loads="" located="" majority="" may="" more="" not="" nutritional="" o:p="" obviously="" of="" on="" only="" patterns="" physiological="" physiologically="" plant="" pollen.="" pollen="" protein="" quality="" range="" recent="" remaining="" same="" seems="" select="" seven="" site="" species="" study="" the="" their="" they="" thgen="" this="" those="" thus="" to="" unfavorable="" use="" visited="" were="" which="" with="">

Several experimental studies have shown that bees, when offered pollen of several species, prefer pollen of some species over others. In one of the early studies, Schmidt (1982) reported that honeybee species, Apis mellifera, preferred pollen of almond over desert broom, saguaro (Carnegiea sp.), and Creosote (Larrea sp.). In another set of experiments, the pollen of maple was preferred over that of cottonwood (Populus sp.), dandelion (Taraxacum sp.) and pine. The amount of pollen removed by A. mellifera from preferred species was significantly higher when compared to pollen of others. Thus this intrinsic ability of the floral visitors to discriminate pollen of different species can act as a filter.

Some studies have been carried out on the effects of feeding host and non-host pollen to bee species on the development of their larvae. Pollen of Sinapis arvensis (Brassicaceae) and Echium vulgare (Boraginaceae) failed to support larval development of Colletes bee species specialized on pollen of Campanula (Praz et al. 2008a). Similarly, pollen of Asteraceae and Ranunculaceae permitted larval growth of only those bee species that are specialized to harvest pollen from plants belonging to these families; their pollen failed to support larval growth of other bee species. Recognition of host pollen in specialized species seems to be genetically determined (Praz et al. 2008b). They compared the floral preferences of one of specialized bee species, Heriades fruncorum (Megachiledae, specialized on Asteraceae pollen), reared on host and non-host pollen. Bees restricted pollen collection to host flowers irrespective of whether they were reared on host or non-host pollen. Recently, Sedivy et al. (2011) studied larval performance of two generalist solitary bees, Osmia bicornis and O. cornuta on pollen diet of four plant species. The larvae of O. bicornis developed well on pollen diet of Ranunculus but failed to develop on Echium pollen; it was the reverse for the larvae of O. cornuta. Larvae of both the species developed well on pollen of Sinapis, but failed to develop on pollen of Tanacetum vulgare (Asteraceae). These studies clearly indicate that palatability of pollen can act as an effective filter to restrict the number of floral visitors. Pollen of non-host species may hamper the digestion of the larvae and the bees seem to have adapted their metabolism to digest pollen of their host species.

Concluding Remarks

Enormous information has accumulated over the years on pollination ecology but our understanding on several aspects in this field is far from complete (Mayer et al. 2011). Although biotic pollination evolved in a few gymnosperms and is present in several extant species, it is not refined and may not be more efficient than wind pollination prevalent in a majority of gymnosperms. Lack of specialization in attracting a wider range of pollinators resulting in inefficient pollination may be one of the primary reasons for lack of diversification in entomophilous gymnosperms. Evolution of the flower in angiosperms with nectar as a reward enhanced floral attraction leading to marked improvement in the frequency and consistency of the pollinators’ visits. Apart from enhanced pollination efficiency, other innovations of the flower particularly the pistil, and evolution of zygomorphy with its further elaboration leading to floral specialization seems to have contributed to reproductive success and dramatic diversification of angiosperms.

The central focus on biotic pollination is the ability of plants to achieve conflicting demands of attraction and restriction of animal species for pollination services. A major limitation of the enormous data accumulated on biotic pollination is the lack of clear distinction between the pollinator(s) and floral visitors in most of the investigations. It is important for pollination biologists to confirm pollinators and their efficiency amongst floral visitors on the basis of pollen transfer to the stigma. Also, biotic pollination is dynamic and shows temporal and spatial variations. It is desirable to study pollination at different locations and during the entire duration of the flowering period for a clear understanding of pollinators’ variability in time and space and its impact on pollination success. Studies on biotic pollination have become more complicated in the light of the reports that many insect pollinators, in the absence of preferred flowers in the habitat, are able to exploit other floral resources by associative learning of floral odors and/or colors with the reward (Riffell 2011). There are even reports of honeybees visiting the flowers of wind-pollinated species to harvest their nutritive pollen source when insect pollinated plants are not available (Keller et al. 2005; Hocheri et al. 2012).

Our understanding on various floral attractants and their role in pollinators’ attraction, and the details of various floral rewards is comprehensive. Considerable information is also available on deceptive pollination. Restriction of floral visitors is as important as their attraction for an efficient pollination system to work. The role of morphological and fragrance filters in specialized flowers that limit the number of animal species with matching features to harvest the reward and consequently eliminate others that are incapable of harvesting the rewards are well documented. The factors responsible for restricting the number of visitors to ‘open flowers’ seems to be more subtle. Although several recent studies have indicated the role of fragrance, nectar and pollen acting as filters in restricting animal visitors, such studies are limited and they need to be extended to a larger number of species. The role of nectar components particularly the secondary metabolites, which was largely ignored so far, needs to be investigated with more vigor. The role of pollen, its quality, digestibility and toxicity seem to play an important role in attracting/ restricting potential pollinators. Interestingly most of the pollen species that have been found to be unfavorable for the development of larvae of several bee species happen to have flowers with fully accessible pollen (Sedivy et al. 2011 and references therein) indicating that pollen filter may be primarily responsible for limiting the number of visitors to open flowers. Available evidence clearly indicates that a combination of floral traits —floral morphology, its fragrance, the quality and quantity of both nectar and pollen rewards—seem to be involved in achieving the dual function of attraction and restriction of floral visitors. The role of each of these may vary between species. Hopefully future studies would try to dissect these functions with appropriate techniques for a better understanding of the role of different components of the flower, the most innovative organ of angiosperms, in achieving the dual functions.