| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
Department of Botany and Zoology, Stellenbosch University, Private Bag X1, Matieland, 7602, South Africa
Received for publication May 26, 2005. Accepted for publication March 16, 2006.
ABSTRACT
The long-standing notion of pollination syndromes, which postulates that plants form recognizable groups according to pollinator type, has been challenged recently on the basis of apparent widespread generalization in pollination systems. As a test of the pollination syndrome concept, I examined the pollination biology of a group of 15 orchids that share a recognizable syndrome of floral features that includes yellow-green coloration, oil secretion, pungent scent, shallow flowers, and a September peak in flowering. The orchids occur in sympatry in the Cape Floral Region of South Africa. According to the pollination syndrome concept, the similar floral features of this group indicate a shared pollinator. To test this prediction, I observed pollinators on Pterygodium alatum, P. caffrum, P. catholicum, P. volucris, Corycium orobanchoides, and Disperis bolusiana subsp. bolusiana. They shared a single species of pollinator, the oil-collecting bee, Rediviva peringueyi. Female bees collected oil from the lip appendage using modified front tarsi. The orchids reduce interspecific reproductive interference through differences in pollinarium length or the use of mutually exclusive pollinarium attachment sites on the body of the bee. The results are contrary to the expectation of generalization in pollination systems and suggest that pollinators play an important role in mediating selection on floral traits.
Key Words: conservation • convergent evolution • Corycium • Diseae • Disperis • Euglossine bees • oil-secreting flowers • pollination syndromes • Rediviva • renosterveld • Pterygodium • South Africa
Darwin identified the relationship between plants and their pollen vectors as a model system for the study of adaptation, and made the study of orchid pollination his first work after the publication of The Origin (Darwin, 1862
). In the century of observation of pollination systems that followed, the interpretation that flowers are adapted to particular pollinators found strong support and crystallized in the concept of "pollination syndromes" (Vogel, 1954
; van der Pijl, 1961
). According to the pollination syndrome view, the flowers of most angiosperms are sufficiently specialized for pollination by particular types of animals for there to be recognizable floral "syndromes"—suites of convergent floral characters that are adapted to specific classes of pollinators (Fenster et al., 2004
). Thus, one may distinguish between, for example, bird-flowers and bee-flowers on the basis of differences in floral reward, flower color, scent, phenology, and morphology (e.g., Faegri and van der Pijl, 1979
; Proctor et al., 1996
). As demonstrated by Darwin (1862)
in his famous prediction of the long-tongued pollinator of a Malagasy orchid, pollination syndromes are logically extended into predictions of pollinator type on the basis of floral features.
More recently, the notion that floral traits conform to a pollination syndrome and represent an adaptive response to a particular pollinator or set of pollinators has been challenged because specialization of the pollination system is postulated to result in greater variance in reproductive success across years and thus ought to be selected against (Waser et al., 1996
; Ollerton, 1998
). The idea that generalization might be the more usual evolutionary outcome of plant–pollinator interactions, with specialization occurring only rarely, found support in comparisons across North American floras (Waser et al., 1996
). In an analysis of a European flora, the same authors found that although flowers are clustered in "phenotype space," there is no strong association with visitor types as pollination syndromes would predict. However, when the same data set was reanalyzed, this time excluding nonpollinating flower visitors and grouping pollinators into functional groups (e.g., long-tongued bees), then widespread specialization for pollination by these functional groups was detected (Fenster et al., 2004
). While the extent of pollination specialization in north temperate regions is controversial, there appears to be strong evidence for extreme pollination specialization in tropical and south temperate regions, which, unlike the northern hemisphere, escaped recent glaciation. Johnson and Steiner (2000
, 2003
), cite several studies from the tropics and southern hemisphere which highlight floral radiations and specialization onto different pollinators. The reality is that very little is known about the pollination systems of the vast majority of plants in the species-rich developing countries of the world. From a conservation perspective, it is precisely in these rapidly changing environments that the need to understand the extent of ecological dependency of plants on particular pollinators is greatest.
The Cape Floral Region of South Africa is a global hotspot of plant diversity (Goldblatt and Manning, 2002
), from which a number of specialized pollination systems have been described (e.g., Vogel, 1954
; Rebelo, 1987
; Johnson, 1994
; Manning and Goldblatt, 1996,
1997
; Pauw, 1998; Goldblatt et al., 2001
; Johnson et al., 2001
). One of the most fascinating pollination systems of southern Africa is that involving oil-secreting flowers and oil-collecting bees of the genus Rediviva (Melittidae). Only female Rediviva collect floral oil, which is probably used mainly as a larval food (Vogel, 1974
; Cane et al., 1983
; Buchmann, 1987
; Vinson et al., 1997
). Early indications are that the system is extensive and encompasses some 140 species and 14 genera of oil-secreting plants and 24 species of Rediviva (Vogel, 1984
; Manning and Brothers, 1986
; Steiner, 1989
; Steiner and Whitehead, 1996
; Whitehead and Steiner, 2001
; Manning and Goldblatt, 2002
; Pauw, 2005
). The plants are in the families Scrophulariaceae (sensu lato), Orchidaceae, and Iridaceae.
Subgroups of similar plant species can be recognized within the extensive oil-bee pollination system. The one examined here includes 15 oil-secreting orchids that share the following syndrome of floral features: pale yellow-green flowers without extensive black markings; secretion of floral oil as a pollinator reward; characteristic pungent scent; flowering period 15 August to 25 October peaking in September; flower depth 5–8 mm (Fig. 1a–n). The species occur in close association with one another in the lowlands of the Cape Floral Region and include members of three genera (Pterygodium, Corycium, and Disperis). According to the pollination syndrome concept, the similar floral features of this group indicate a shared pollinator. My aim was to test this prediction through extensive field work. I started by investigating the syndrome of floral features in more detail. I report here on the flower visitors and describe how the unusual structures of the flowers can be interpreted as adaptations for pollination by a single species of pollinator, the oil-collecting bee Rediviva peringueyi. Then, I discuss the outcome of breeding system experiments, which are informative about the role of pollinators in determining seed set, and describe the biogeography of the plants and their pollinators. I end by discussing the implications of these findings in the context of current ideas about generalization and specialization of pollination systems, and consider the conservation implications of specialization.
|
Study species and sites
Initial investigation of the pollination biology of the very common oil-secreting orchid Pterygodium catholicum led to the realization that several of its floral features (discussed earlier) are shared with co-occurring species. This group of species includes Corycium orobanchoides, Disperis bolusiana subsp. bolusiana, D. capensis var. capensis, D. circumflexa subsp. circumflexa, D. cucullata, D. villosa, D. xduckittiae, Pterygodium alatum, P. caffrum, P. cruciferum, P. hallii, P. inversum, P. platypetalum, and P. volucris. The orchid taxa included in this study are named according to Linder and Kurzweil (1999)
. In variable taxa, populations that deviate from the syndrome of shared features were by definition excluded. The orchids are currently all included in subtribe Coryciinae (Linder and Kurzweil, 1999
).
The study was conducted in the coastal lowlands of the Western Cape Province in an area bound by Cape Point, Caledon, Ceres, Piketberg, and Saldanha. Within this area, the study species occur on clay soil derived from shale or granite. The climate is Mediterranean, and the vegetation is a fine-leafed, fire-prone, asteraceaeous shrubland known as renosterveld (Low and Rebelo, 1996
). In all the study species, aboveground parts are present only during the wet winter and spring. Flowering was most prolific in the first spring after a fire in the preceding summer. In later successional stages, flowering declines and the plants produced only leaves in the shade of bushes. Most of the species multiply vegetatively through the production of sister tubers at the end of roots and thus form extensive, clonal colonies. The richness of the group varied among study sites around a median of three species (max. = 7; min. = 1; N = 18). The species were most common on south-facing slopes and the edges of seeps. Populations of the different species often abutted and replaced each other in a predictable sequence along a moisture gradient. The only other spring-flowering, oil-secreting species at the sites were Diascia spp. and Hemimeris racemosa (Houtt.) Merrill. Both genera are fire-stimulated, annual Scrophulariaceae. Hemimeris racemosa was common at most sites, and Diascia spp. were present at several sites, but common at only one.
Thirty-nine study sites from throughout the entire distribution range of these taxa were selected and visited between August and October of 1997 to 2005. At each site, I focused on the pollination biology of the study species, but to provide a context for this study, I also observed pollinators on all other encountered oil-secreting plant species, particularly the common H. racemosa.
Floral phenology, advertisement, and reward
Flowers or inflorescences were collected, dissected, described, and photographed. Voucher specimens were pressed or preserved in FAA (formaldehyde, acetic acid, 70% ethanol at 5 : 5 : 90, v/v/v) and placed in the Bolus Herbarium, University of Cape Town. Spectral reflectance curves were obtained for a selection of the study species to objectively assess their apparent similarity in color. Measurements were taken using an Ocean Optics (Dunedin, Florida, USA) USB2000 spectrometer with its OOIBase32 Spectrometer Operating Software and a UV/VIS 400-µm fiber optic reflection probe, which was held 5 mm away at 45° to the petal surface. An Ocean Optics PX-2 pulsed xenon light was used. All ambient light was excluded during measurement and calibration. Calibration was performed using white Ocean Optics WS-1 Diffuse Reflectance Standard (with the light source on) and black velvet (with the light source off). For each species, five individuals were sampled and the curves averaged. The measurement was taken on the outside of the lateral petal or outside the dorsal sepal, whichever presented the largest advertisement.
To determine which floral parts were responsible for scent secretion, floral parts of P. catholicum were separated and placed in vials and the strength of the scent compared subjectively. Sudan IV (Coleman & Bell, Norwood, Ohio, USA) was used to confirm the presence of floral oil in the study species. The crystals dissolve in oil, staining it red; the crystals do not dissolve in aqueous solutions, such as nectar, which remain unstained. Crystals were sprinkled onto the liquid film, which was visible on some floral parts under magnification with a light microscope. The lip appendage of P. catholicum was sectioned, and the cellular structure was examined microscopically.
Pollination observations
Pollinator visitation rate was quantified for P. catholicum and P. alatum by observing ~10-m2 patches of the focal plant species for 15-min intervals from a distance of 2 m. Visitation rate was calculated as visits per flower per hour. Quantitative observations were restricted to warm (>20°C), windless hours between 1030 and 1530 hours. Pollinators were netted, sexed, identified according to Whitehead and Steiner (2001)
, and examined microscopically for the presence of pollinaria. Pollinaria were identified to species level by comparison with reference collections from the study sites. The sites of pollinarium attachment on the body of the bee were named according to Michener (1944)
.
Breeding systems
Breeding system experiments were conducted on the two most common orchid species, P. catholicum and P. alatum. To determine whether P. catholicum is capable of autonomous self-pollination, insects were excluded by enclosing budding plants on the Darling Hills in wire cages over which fine gauze bags were stretched (pollinator-excluded: N = 54 flowers, 16 plants; unmanipulated: N = 341 flowers, 66 plants). Pterygodium alatum plants were similarly enclosed at Malmesbury (pollinator-excluded: N = 42 flowers, 12 plants; unmanipulated: N = 238 flowers, 36 plants). To determine whether P. catholicum is self-compatible, hand-pollination experiments were conducted at Rondebosch Common using pollinator-excluded plants. Pollinaria were collected using a bent dissecting needle and dabbed onto the stigma, which was accessed by freeing the tip of the lip appendage from the dorsal sepal (self-pollination: N = 31 flowers; cross-pollination N = 32 flowers). Capsule set was recorded, rather than seed set, because seeds are minute and numerous and are released from the capsules soon after maturation.
Biogeography
To determine the extent of geographical congruence between the plants and their pollinators, plant distributions were plotted at a quarter-degree-square resolution using the records from the Bolus Herbarium University of Cape Town and the National Herbarium Pretoria. The plant database consisted of 594 records of the study species from 1830 to the present. Pollinator distributions were plotted using the data published in Whitehead and Steiner (2001)
, supplemented by my own collections in the South African Museum.
RESULTS
Floral advertisement and reward
The apparent similarity in color of the yellow-green flowers was objectively confirmed in a subset of species by matching patterns of spectral reflectance in the ultraviolet to infrared range (Fig. 2). The Sudan test confirmed the presence of floral oil in 14 of the 15 study species. A 15th taxon, an undescribed yellow-green form of D. capensis var. capensis, lacked oil (Fig. 1o).
|
|
|
|
Functional floral morphology
Female R. peringueyi were observed to use a rubbing motion of their front tarsi to collect floral oil, visible as a thin film in twin depressions on the apex of the lip appendage of P. alatum (Fig. 3a). The tarsus of the left front leg was placed in the right depression, and simultaneously the tarsus of the right front leg was placed in the left depression. While collecting oil with their front tarsi, the bees were observed to hang onto the flower, which tilts slightly downward. The tarsi of the hind legs gripped the margins of the lip, while the middle legs fixed the bee in a hanging position. The tarsal hooks of the middle legs hooked onto holdfasts at the base of the lip appendage (Fig. 4b). The holdfasts are provided by the sharp rails, which run down the vertical edges of the lip appendage. To reach the holdfasts, the tarsi of the middle legs slotted into channels formed between the rostellum arms and the lip appendage. The viscidia lie against the outer walls of these channels and face inwards toward the lip appendage, such that they became attached to the dorsal surfaces of the tarsi (Fig 1b). In P. alatum, the viscidia are set ~0.5 mm back from the holdfasts so that they coincide with the position of the distitarsus (tarsal segment 5). The pollinaria were extracted from the thecae (anther sacs) when the bee withdrew from the flower. Free from the flower and attached to the bee, the caudicle was observed to bend through ~150° to bring the tip of the distal end of the pollinarium around to the posterior edge of the tarsus. The flexing occurred within 1–3 s. In P. alatum, the stigma is located close to the viscidia on the tips of the rostellum arms (Fig. 4b). The short distance between the viscidium and the stigma corresponds with the unusually short caudicles of P. alatum. When a pollinarium was manually dabbed into the stigma, a proportion of the massulae were observed to adhere to the stigma, and several were broken free when the pollinarium was withdrawn. Visits to P. alatum and the other Pterygodium species lasted 2–5 s.
|
In P. catholicum, the pungent scent (described variously as "lemon," "soap," "wax," "new car," and "angry millipede") was found to be secreted primarily by the petals. In this species, the oil is held in narrow vertical grooves on the abaxial side of the tip of the lip appendage. These grooves, lead down into channels, which run vertically inside the lip appendage to midway down its length. The grooves and channels were lined by specialized secretory cells, corresponding to epithelial elaiophores described by Vogel (1974)
. In P. catholicum, pollinator behavior was observed to be similar to that described for P. alatum, but the visiting bee was largely hidden inside the galeum (hood) formed by the petals and dorsal sepal. The galeum presumably limits illegitimate lateral and dorsal access to the oil-secreting lip appendage. When the bee is in the oil-collecting position, the small magenta dots on the tip of the dorsal sepal coincide with the position of the eyes. Similar magenta dots or areas of dots occur on the perianth segments immediately above the oil-secreting region in P. volucris, P. hallii, Corycium orobanchoides (Fig. 1d, e, k), and H. racemosa (Pauw, 2005
). These areas of dots coincide with the position of the eyes of the foraging bee and might serve to guide and stimulate the oil-collection process. A functionally important difference with P. caffrum and P. alatum is the fact that the viscidia are set ~2.5 mm back from the holdfast and lie in the bottom of the channel at the base of the lip appendage (Fig. 4a). Thus located, the viscidia attach to the narrow posterior surface of the distal half of the basitarsus (Fig. 3b). The stigma faces upwards and is hidden on the adaxial side of the lip appendage. The pollinaria are correspondingly long, but were not observed to undergo flexing when withdrawn from the thecae. After pollination, the flowers of P. catholicum turn red in some populations.
I was not able to observe in detail the pollinator behavior on P. volucris. The unusual site to which pollinaria were attached on captured bees (ventral surface of last abdominal segment, Fig. 3c) suggests the following scenario. The petals close around the lip appendage to form a long, narrow galeum, but in contrast to the previous three species, there is dorsal access into the galeum via a circular aperture between the tips of the petals and the dorsal sepal (Fig. 1d). The bee inserts its front legs through this aperture and reaches down for the oil which pools in the cup at the tip of the lip appendage (Fig. 4d). It holds on by hugging the outside of the petals with its middle legs, which slot into the constriction near the top of the galeum. The hind tarsi fit into short channels, which are located on either side of the viscidia and are formed between tooth-like callosities and the lateral lip lobes. In this position, the ventral surface of the last abdominal segment makes contact with the viscidia, which face outward and toward the underside of the abdomen. The paired callosities on the lip, which do not occur in the other species considered here, presumably ensure the exact placement of the hind tarsi, and hence the abdomen.
The rare P. inversum flowered at three sites amongst the other species discussed here and set seeds successfully, although pollination was not observed. In P. inversum, the flowers are hyper-resupinate and the oil is secreted from the base of the lip appendage as well as from the lip, which is in the dorsal position (Fig. 1j). The oil pools in the concave lip and is held in place by trichomes on the base of the lip and lip appendage (Fig. 4f). The structure of the flowers suggests that the bee hangs onto the slightly downward facing flowers by hugging the narrow constriction between the lip and the lip appendage with its middle legs, while the front legs gather oil from the lip. The hind tarsi would contact the viscidia as they feel for a holdfast on the smooth lip appendage. Thus, hyper-resupination of the flowers of P. inversum (and P. hallii) allows the use of the hind legs of the pollinator, a novel pollinarium attachment site.
In Corycium orobanchoides the lip appendage diverges into two processes, which are tucked into twin pouches at the back of the galeum. The galeum is too narrow to accommodate the bee, and unlike the condition in P. volucris, there is no dorsal access into the galeum (Fig. 1e). Visiting bees were observed to remain outside the flower and to reach for the lip appendage in the back of the galeum by inserting both their front legs into the twin channels that run past the rostellum. The viscidia are located along the length of the channels ~5 mm anterior to the tips of the lip appendage. Thus when the tarsi of the front legs are in contact with the tips of the oil-secreting lip appendage, the large viscidia become attached to the ventral surface of the distal half of the tibia of the front legs. Bees were apparently holding on by hooking the tarsal hooks of the middle legs onto the small lateral projections of the lip. The hind legs secured a purchase on neighboring flowers in the densely packed inflorescence.
The behavior of R. peringueyi on the flowers of Disperis bolusiana subsp. bolusiana is similar to that observed in C. orobanchoides and corresponds with descriptions by Steiner (1989)
of pollination in other Disperis species. The oil is secreted from the tip of the lip appendage. The deeply saccate lateral sepals of Disperis, often assumed to contain the reward (e.g., Linder et al., 2005
), do not contain oil and function primarily to protect the rostellum arms and viscidia when the flower is in bud (Fig. 1f). The oil-secreting lip appendage is housed inside a deep and narrow galeum, and only the front legs of the bee enter the flower via the twin entrances.
Observation of oil-collecting bee behavior suggested that the distance over which the bee has to reach with its front legs in order to gather the reward is an important functional measurement (Vogel and Michener, 1985
; Steiner and Whitehead, 1991
). In P. alatum, P. caffrum, and P. catholicum, this distance (5 mm, 5 mm, and 7.5 mm, respectively) is more or less equivalent to the length of the lip appendage because the holdfast for the middle legs is located at the base of the appendage and oil is secreted at its tip. In D. bolusiana subsp. bolusiana, D. circumflexa subsp. circumflexa and Corycium orobanchoides, the bee remains outside the flower and reaches in only with its front legs. In these species, the functionally important distance (8 mm, 6 mm, and 7.5 mm, respectively) is equivalent to the depth of the galeum formed by the lateral petals and dorsal sepal. These distances henceforth are referred to as flower depth.
Breeding systems
In P. catholicum, pollinator exclusion reduced capsule production to 0%, whereas 98 ± 1% (mean ± SE) of control flowers set seed capsules. Similarly in P. alatum, pollinator exclusion reduced capsule set to 0%, whereas 68 ± 5% of control flowers set seed capsules. In all the other orchid species studied, the possibility of habitual autogamy or apomixis was less conclusively ruled out by (1) the frequent observation of individuals or flowers that failed to set seed after flowering and (2) the lack of a mechanism for transferring pollinaria to the stigma. Although there is no mechanism for autonomous self-pollination, P. catholicum flowers do produce seed capsules after being hand-pollinated with their own pollen (80 ± 10% capsule set if cross-pollinated by hand; 76 ± 11% capsule set if self-pollinated by hand).
Biogeographical analysis
Up to 10 members of the study group co-occurred at the quarter-degree scale, and the distribution of these plant species was closely matched by the known distribution of the oil-collecting bee R. peringueyi (Fig. 5).
|
The Rediviva peringueyi pollination guild and floral syndrome
The results presented here are consistent with the prediction of the pollination syndrome concept that a group of 15 plants, which share a syndrome of floral features, owe their similarity to the fact that they are adapted for pollination by a shared pollinator. This finding is contrary to the expectation of generalization in pollination systems and a consequent lack of recognizable syndromes of floral features (Waser et al., 1996
). The theoretical expectation of generalization follows from the observation that specialization of the pollination system is likely to result in greater variance in reproductive success across years and thus ought to be selected against (Waser et al., 1996
). In the pollination system described here, variance in pollination success amongst years and sites is extreme (Pauw, 2004
), but specialization has nevertheless occurred.
Pollination was observed in five of the species in the group, and in every case the oil-collecting bee R. peringueyi was the sole pollinator (Fig. 3, Table 1). Partial evidence provided here, in combination with the findings of Steiner (1989)
, indicates pollination by R. peringueyi in two additional members of the group, Disperis bolusiana subsp. bolusiana and D. villosa (Tables 1, 2). On the strength of these observations, pollination by R. peringueyi is predicted in the remainder of the species on which, largely due to their rarity, no pollinators were observed. Pollination by deceit is predicted for the yellow-green form of D. capensis var. capensis, which lacks floral oil (Fig. 1o). Together, these 15 plant species are suggested to constitute a pollination guild (sensu Manning and Goldblatt, 1997
).
The lack of a molecular phylogeny for the Coryciinae makes it difficult to assess the relative importance of shared ancestry vs. convergent pollinator-driven selection in creating the uniformity that is evident in the R. peringueyi pollination guild. Indications are that all of the species in the guild have their sister species outside the guild (A. Pauw, unpublished data), suggesting a very prominent role for pollinator-driven selection. An example is the superficially similar pair Pterygodium alatum and P. caffrum (Fig. 1b, c). On the basis of floral features, they are currently considered to be sister species (Kurzweil et al., 1991
). If, however, emphasis is placed on features other than those involved in advertisement, the placement of P. alatum in Pterygodium becomes doubtful. The unusual location of the stigma lobes sets this species part form Pterygodium, while the black discoloration of pressed plants and the basal rosette of keeled and pointed leaves suggest a closer relationship with Ceratandra and Corycium. If this relationship is proven, then the similarity of the flowers of P. alatum and P. caffrum is testimony to the power of convergent selection driven byR. peringueyi.
The pollination mechanisms employed by the various species in the R. peringueyi pollination guild are complex and precise and are a classical illustration of extreme adaptation and specialization. The pollinator reward is floral oil. It is secreted by the unique lip appendage, which characterizes the subtribe Coryciinae, and has aptly been referred to as "the most bizarre floral structure known in the orchids" (Kurzweil et al., 1991
, p. 1) (Fig. 4). In Pterygodium the appendage also provides the holdfast onto which the middle or hind legs of the bee hook, thus forming the central functional axis of the flower. Female R. peringueyi were observed to collect oil from the lip appendage with a rubbing motion of the front tarsi (Fig. 3a). Tarsal segments 2–5 of the front legs of R. peringueyi are coated in plumose hairs, which absorb the oil (Whitehead and Steiner, 2001
). Floral oil is not consumed by adult oil-collecting bees, but is collected by female bees, who use it as a larval food (Vogel, 1974
; Cane et al. 1983
; Vinson et al., 1997
). Thus, the fact that only female bees were captured on the flowers of the study species confirms the finding that oil, rather than nectar, is the reward.
Various morphological contrivances of the orchid flowers ensure that R. peringueyi is unable to collect the reward without making contact with the sticky viscidia, which attach the pollinaria precisely to the body of the bee (Figs. 1, 3). Bees flew between orchid inflorescences with as many as 27 attached pollinaria belonging to up to three species of orchids (Table 1) and were observed to alight on the flowers in a position that would result in contact between the pollinaria and the stigma. Reproductive interference between orchid species is reduced by the use of at least nine mutually exclusive pollinarium attachment sites (Fig. 1). Only two species, P. caffrum and P. alatum, were found to use the same part of the bee's body. Interspecific pollen transfer between these two co-occurring species is apparently prevented by differences in the length of the pollinaria. The pollinaria of P. caffrum are twice as long as those of P. alatum, and the stigma is correspondingly distantly located. The structural diversity of lip appendages in the group is considerable, as illustrated by Pterygodium (Fig. 4), and translates into a diversity of pollinarium attachment sites (Fig. 1). The precise pollination mechanism described here is essential for sexual reproduction—the exclusion of insects from the flowers of P. catholicum and P. alatum reduced seed set from high levels in control plants to zero.
Biogeography provides further evidence for a close match between plant and pollinator (Fig. 5). The distribution range of the study species is closely matched by that of R. peringueyi (Fig. 5). Temporal congruence is equally striking. The flowering time of the species coincides precisely with the brief flight period of R. peringueyi, from 14 August to 19 October (Whitehead and Steiner, 2001
).
Factors contributing to specialization
The pollination system described here is unusual in having only a single species of pollinator. The combination of a specialized reward (floral oil) and a depauperate fauna of oil-collecting bees sets the stage for the evolution of highly specialized plant–pollinator relations. Only six Rediviva species co-occur with the members of the R. peringueyi pollination guild in the lowlands of the southwestern Cape (Whitehead and Steiner, 2001
), while co-occurring oil-secreting plants comprise only Hemimeris racemosa and Diascia spp. (both Scrophulariaceae).
In addition to these biogeographical limitations, the unusual traits of the members of the R. peringueyi pollination guild may further limit the breadth of its pollinator fauna. In this study and that of Whitehead and Steiner (2001)
, the five Rediviva species which co-occur with R. peringueyi were captured only on oil-secreting Scrophulariaceae. They did not visit the members of the R. peringueyi pollination guild, nor did they carry pollinaria. Specialized advertising may be the first filter that limits the breadth of the pollinator fauna. The inconspicuous, pale yellow-green flowers and pungent scent of the R. peringueyi pollination guild contrasts with the scentless, brightly colored flowers of the Scrophulariaceae, such as Hemimeris racemosa, which attract the full range of Rediviva species (Pauw, 2004
).
Structural features of the flowers of the R. peringueyi pollination guild may be the second filter. At 12 mm, R. peringueyi is one of the largest Rediviva in the lowlands of the southwestern Cape. The smaller Rediviva species are apparently unable to access the reward provided by the members of the R. peringueyi guild with ease. Morphometrics suggests that these smaller Rediviva would be unable to reach the oil on the distal tip of the lip appendage. In the short-spurred Scrophulariaceae, the oil is more easily accessible. The exception is the very long-legged R. micheneri, which enters the range of the R. peringueyi pollination guild in the north. Its disproportionately elongated front legs (16.7 mm) are much longer that those of R. peringueyi (11 mm) and are specialized for collecting oil from the very long-spurred Diascia longicornis (Whitehead and Steiner, 2001
). Rediviva micheneri co-occurred with the R. peringueyi pollination guild at two sites but was not observed to be involved in the pollination of the group.
In contrast with the high degree of specialization exhibited by the members of the guild, their pollinator has relatively broad floral preferences. Apart from visiting the members of the R. peringueyi pollination guild, R. peringueyi also gathers oil from various short-spurred Scrophulariaceae. The most frequently visited of these was Hemimeris racemosa, which is pollinated by several Rediviva species (Pauw, 2005
), while short-spurred members of the genus Diascia were of secondary importance (Whitehead and Steiner, 2001
). In addition to oil, R. peringueyi also collected nectar and pollen from additional plant species (see Results: Flower visitors, visitation rate, and pollen loads), all of which were visited by a broad spectrum of insects. Thus, while the R. peringueyi pollination guild does form a clearly defined compartment within the plant–pollinator interaction web, this compartment is nevertheless linked into the broader pollination web by interactions between the pollinator and a core of generalist plant species. Multiple visitors are involved in the pollination of these relatively generalist, linking species, thus it seems unlikely that R. peringueyi in particular has played a significant role in the evolution of their floral features, which are morphologically quite distinct from the R. peringueyi pollination guild.
Conservation of a specialized pollination system
The biggest challenge in this study was the scarcity of suitable study sites. About 80% of lowland vegetation has already been transformed by urbanization and agriculture (Heijnis et al., 1999
). What remains are scattered fragments of natural habitat, mostly less than 1 ha in size. In many of these fragments, the absence of R. peringueyi and repeated pollination failure in the entire guild was recorded. We have probably already lost the chance to understand the intriguing flowers of species such as P. cruciferum, which persists in fewer than five remnants of natural vegetation where they seldom, if ever, receive pollinator visits. In contrast with the pollination systems of the north temperate regions, which almost invariably involve several ecologically equivalent pollinator species (Waser et al., 1996
; Fenster et al., 2004
), the pollination system described here is dependent on a single insect species. This presents a challenge for conservation because of the low level of ecological redundancy means that the loss of R. peringueyi may trigger linked extinctions amongst the plants in the R. peringueyi pollination guild. It seems unlikely that the R. peringueyi pollination guild will persist in a modern, cultural landscape without unique conservation planning.
FOOTNOTES
1 The author thanks the Bolus Herbarium, the National Herbarium Pretoria, and P. Linder for providing species distribution data, B. Liltved for the use of photographs and for help in locating study sites, M. Cherry for the use of a spectrometer, and the National Geographic and the NRF for funding. This article is dedicated to the memory of Vincent Whitehead. 
LITERATURE CITED
Buchmann S. L.. 1987. The ecology of oil flowers and their bees. Annual Review of Ecology and Systematics 18: 343-369.[CrossRef][ISI]
Cane J. H. Eickwort G. C. Wesley F. R. Speilholz J.. 1983. Foraging, grooming and mate-seeking behaviors of Macropis nuda (Hymenoptera, Melittidae) and use of Lysimachia (Primulaceae) oils in larval provisions and cell linings. American Midland Naturalist 110: 257-264.[CrossRef][ISI]
Darwin C.. 1862. On the various contrivances by which British and foreign orchids are fertilized by insects Murray, London, England.
Faegri K. van der Pijl L.. 1979. The principles of pollination ecology, 2nd ed Pergamon, Oxford, UK.
Fenster C. B. Armbruster W. S. Wilson P. Dudash M. R. Thomson J. D.. 2004. Pollination syndromes and floral specialization. Annual Review of Ecology Evolution and Systematics 35: 375-403.[CrossRef][ISI]
Goldblatt P. Manning J. C.. 2002. Plant diversity of the Cape region of southern Africa. Annals of the Missouri Botanical Garden 89: 281-302.[CrossRef][ISI]
Goldblatt P. Manning J. C. Bernhardt P.. 2001. Radiation of pollination systems in Gladiolus (Iridaceae: Crocoideae) in southern Africa. Annals of the Missouri Botanical Garden 88: 713-734.[CrossRef][ISI]
Heijnis C. E. Lombard A. T. Cowling R. M. Desmet P. G.. 1999. Picking up the pieces: a biosphere reserve framework for a fragmented landscape: the coastal lowlands of the western Cape, South Africa. Biodiversity and Conservation 8: 471-496.[CrossRef][ISI]
Johnson S. D.. 1994. Red flowers and butterfly pollination in the fynbos of South Africa. In M. Arianoutsou and R. H. Groves [eds.] Plant–animal interactions in Mediterranean-type ecosystems Kluwer, Dordrecht, Netherlands.
Johnson S. D. Pauw A. Midgley J.. 2001. Rodent pollination in the African lily Massonia depressa (Hyacinthaceae). American Journal of Botany 88: 1768-1773.
Johnson S. D. Steiner K. E.. 2000. Generalization versus specialization in plant pollination systems. Trends in Ecology & Evolution 15: 140-143.
Johnson S. D. Steiner K. E.. 2003. Specialized pollination systems in southern Africa. South African Journal of Science 99: 345-348.[ISI]
Kurzweil H. Linder H. P. Chesselet P.. 1991. The phylogeny and evolution of the Pterygodium–Corycium complex (Coryciinae, Orchidaceae). Plant Systematics and Evolution 175: 161-223.[CrossRef][ISI]
Linder H. P. Kurzweil H.. 1999. Orchids of southern Africa A. A. Balkema, Rotterdam, Netherlands.
Linder H. P. Kurzweil H. Johnson S. D.. 2005. The southern African orchid flora: composition, sources and endemism. Journal of Biogeography 32: 29-47.[CrossRef][ISI]
Low A. B. Rebelo A. G.. 1996. Vegetation of South Africa, Lesotho and Swaziland Department of Environmental Affairs & Tourism, Pretoria, South Africa.
Manning J. C. Brothers D. J.. 1986. Floral relations of four species of Rediviva in Natal (Hymenoptera: Apoidea: Melittidae). Journal of the Entomological Society of Southern Africa 49: 107-114.[ISI]
Manning J. C. Goldblatt P.. 1996. The Prosoeca peringueyi (Diptera: Nemestrinidae) pollination guild in southern Africa: long-tongued flies and their tubular flowers. Annals of the Missouri Botanical Garden 83: 67-86.[CrossRef][ISI]
Manning J. C. Goldblatt P.. 1997. The Moegistorhynchus longirostris (Diptera: Nemestrinidae) pollination guild: long-tubed flowers and a specialized long-proboscid fly pollination system in southern Africa. Plant Systematics and Evolution 206: 51-69.[CrossRef][ISI]
Manning J. Goldblatt P.. 2002. The pollination of Tritoniopsis parviflora (Iridaceae) by the oil-collecting bee Rediviva gigas (Hymenoptera: Melittidae): the first record of oil-secretion in African Iridaceae. South African Journal of Botany 68: 171-176.
Michener C. D.. 1944. Comparative external morphology, phylogeny, and classification of the bees (Hymenoptera). Bulletin of the American Museum of Natural History 82: 158-326.
Ollerton J.. 1998. Pollination: sunbird surprise for syndromes. Nature 394: 726-727.[CrossRef][ISI]
Pauw A.. 1998. Pollen transfer on birds' tongues. Nature 394: 731-732.[CrossRef]
Pauw A.. 2004. Variation in pollination across a fragmented landscape at the Cape of Africa. Ph.D dissertation, University of Cape Town, Cape Town, South Africa.
Pauw A.. 2005. Inversostyly: a new stylar polymorphism in an oil-secreting plant, Hemimeris racemosa (Scrophulariaceae). American Journal of Botany 92: 1878-1886.
Proctor M. Yeo P. Lack A.. 1996. The natural history of pollination Timber Press, Portland, Oregon, USA.
Rebelo A. G.. 1987. Bird pollination in the Cape flora. In A. G. Rebelo [ed.] A preliminary synthesis of pollination biology in the Cape flora 83-108 South African Council for Scientific and Industrial Research, Pretoria, South Africa.
Steiner K. E.. 1989. The pollination of Disperis (Orchidaceae) by oil-collecting bees in southern Africa. Lindleyana 4: 164-183.
Steiner K. E. Whitehead V. B.. 1991. Oil flowers and oil bees: further evidence for pollinator adaptation. Evolution 45: 1493-1501.[CrossRef][ISI]
Steiner K. E. Whitehead V. B.. 1996. The consequences of specialization for pollination in a rare South African shrub, Ixianthes retzioides (Scrophulariaceae). Plant Systematics and Evolution 201: 131-138.[CrossRef][ISI]
van der Pijl L.. 1961. Ecological aspects of flower evolution. II. Zoophilous flower classes. Evolution 15: 44-59.[CrossRef][ISI]
Vinson S. B. Williams H. J. Frankie G. W. Shrum G.. 1997. Floral lipid chemistry of Byrsonima crassifolia (Malpighiaceae) and a use of floral lipids by Centris bees (Hymenoptera: Apidae). Biotropica 29: 76-83.[CrossRef][ISI]
Vogel S.. 1954. Blütenbiologische Typen als Elemente der Sippengliederung: dargestellt anhand der Flora Südafrikas. Botanische Studien 1: 1-338.
Vogel S.. 1974. Ölblumen und ölsammelnde Bienen. Tropische und subtropische Pflanzenwelt 7: 285-547.
Vogel S.. 1984. The Diascia flower and its bee-and-oil-based symbiosis in southern Africa. Acta Botanica Neerlandica 33: 509-518.[ISI]
Vogel S. Michener C. D.. 1985. Long bee legs and oil-producing floral spurs, and a new Rediviva (Hymenoptera, Melittidae; Scrophulariaceae). Journal of Kansas Entomological Society 58: 359-364.
Waser N. M. Chittka L. Price M. V. Williams N. M. Ollerton J.. 1996. Generalization in pollination systems, and why it matters. Ecology 77: 1043-1060.[CrossRef]
Whitehead V. B. Steiner K. E.. 2001. Oil-collecting bees of the winter rainfall area of South Africa (Melittidae, Rediviva). Annals of the South African Museum 108: 143-277.
This article has been cited by other articles:
|
|
 |
|
  A. V. Etcheverry, M. M. Aleman, and T. F. Fleming Flower Morphology, Pollination Biology and Mating System of the Complex Flower of Vigna caracalla (Fabaceae: Papilionoideae) Ann. Bot., June 28, 2008; (2008) mcn106v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Waterman and M. I. Bidartondo Deception above, deception below: linking pollination and mycorrhizal biology of orchids J. Exp. Bot., March 2, 2008; (2008) erm366v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Stpiczynska and K. L. Davies Elaiophore Structure and Oil Secretion in Flowers of Oncidium trulliferum Lindl. and Ornithophora radicans (Rchb.f.) Garay & Pabst (Oncidiinae: Orchidaceae) Ann. Bot., February 1, 2008; 101(3): 375 - 384. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Fig. 3. Rediviva peringueyi pollination mechanism. (a) Female R. peringueyi collecting floral oil from the apex of the lip appendage of Pterygodium alatum with a rapid rubbing motion of the front tarsi. The bee hangs onto the lip appendage with the middle tarsi, onto which the pollinaria (visible) become attached. Bar = 3 mm. (b) Several pollinaria of P. catholicum attached precisely to the basitarsi of the middle legs of R. peringueyi via the sticky viscidia. Bar = 1 mm. (c) Pollinaria of Pterygodium volucris attached to the ventral surface of the last abdominal segment of R. peringueyi. Bar = 3 mm.
