| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2Center for Ecological Research, Kyoto University, Sakyo, Kyoto 606-8502, Japan; and 4Graduate School of Human and Environmental Studies, Kyoto University, Sakyo, Kyoto, Japan 606-8501, Japan
Received for publication February 5, 1998. Accepted for publication September 14, 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: bee • coevolution • Costaceae • floral morphology • pollination guild • Sarawak • spiderhunter • Zingiberaceae
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Janzen, 1980
). With the exception of a few special cases, a single plant species is pollinated by more than one species of visitors and each pollinator species utilizes many plant species as a floral resource. Therefore, a single plant species is usually influenced by a group of pollinator species, and vice versa (Feinsinger, 1983
).
Floral morphology is one of the most important aspects of plant–pollinator interactions. It determines pollinator accessibility to nectar, efficiency of pollen deposition on the pollinator body, and efficiency of pollen acquisition by the stigma from the pollen vectors. Variation in the site of pollen placement on the pollinator body can promote coexistence of plant species sharing common pollinators (e.g., Armbruster, Edwards, and Debevec, 1994
). Corollas also function as protection of floral resources against nonpollinating flower visitors. Recent research has demonstrated the importance of inherited morphological characters and their effect on fitness (Campbell, 1989
, 1996
; Galen, 1989
; Campbell et al., 1991
).
Although correlations between flower morphology and pollination systems have been discussed repeatedly, floral morphology has usually been described qualitatively, such as "bowl," "dish," or "tube," and morphological similarity of the flowers in the same or different pollination guilds has rarely been further quantified. At most, quantification of floral morphology is usually restricted to flower depth, which is thought to be related to pollinator tongue length. This is partly due to the functional shift of floral parts among species, which makes it difficult to evaluate with a single method various floral morphologies in different taxonomic groups. However, the importance of morphological characteristics other than flower depth was demonstrated by Campbell (1989
, 1996)
. Harder (1985)
and Inoue and Kato (1992)
suggested that more than one morphological characteristic causes differentiation in foraging niches among flower visitors. Instead of comparing a single floral characteristic among plant species belonging to diverse taxa, accurate evaluation and multivariate comparison of floral morphology among related species with basically the same construction will be a fruitful approach.
The Zingiberaceae are a family of the order Zingiberales with 
50 genera and 1000 species and is indigenous in the tropics, especially Indomalesia (Dahlgren, Clifford, and Yeo, 1985
). More than 150 species of the family have been identified in Borneo (Smith, 1985
, 1986
, 1987
, 1988
, 1989
; Sakai and Nagamasu, 1998
). The family shows the highest species diversity among herbaceous families in a particular area in Bornean lowland forests (Poulsen, 1997
). Considering their diverse and conspicuous flowers, there are relatively few studies on the pollination ecology of the Zingiberaceae (Endress, 1994
; but see Classen, 1987
; Ippolito and Armstrong, 1993
; Kato, Itino, and Nagamitsu, 1993
; Kato, 1996
). The Costaceae are also a member of the Zingiberales and sometimes included in Zingiberaceae. The family is pantropical but its distribution is clearly centered in Central America (Dahlgren, Clifford, and Yeo, 1985
), and only three species of Costus are known from Borneo (Maas, 1979
). These two families were treated together as gingers, because species in both the families are similar in their floral groundplan and habitat.
In this paper, the relationship between the floral morphology of gingers and their pollination systems in a Bornean tropical forest was studied. The ginger flowers of 44 species found in the forest had similar floral groundplans but differed in the shape and size of their floral parts. First, we analyze morphological variation among pollination guilds. The analysis shows that the morphological aspects of the pollination syndromes can be quantified. The paper then discusses the variations within pollination guilds and their functions in the plant–pollinator interactions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), which included a lowland mixed dipterocarp forest located in Lambir Hills National Park, as well as secondary forests and swamps in the surrounding area of the park. The forest is dominated by Dipterocarpaceae in the emergent layer, which sometimes reach a height of 70 m, while the secondary forests are dominated by Euphorbiaceae and Moraceae, which reach 20 m in height.
Plant collections and observation of flower visitors
Plant collections and observations were made from July 1994 to June 1995 and from April to June 1996. Flowers of 44 ginger species (Zingiberaceae and Costaceae) were collected and pressed for herbarium specimens (Table 1). Their inflorescences and flowers were preserved in 50% alcohol and their photos were taken to record color. These plant specimens were identified in E (Edinburgh), K (Kew), and SAR (Sarawak Herbarium, Sarawak Forest Department) herbaria. They were sent to some herbaria, among which SAR and KYO (Herbarium, Kyoto University) have a complete set of the collection (Plants of Sarawak, Canopy Biology Program). All the ginger species produce hermaphroditic flowers with the exception of Amomum polycarpum, which has male and hermaphroditic flowers on the same or different inflorescences within an individual (i.e., andromonoecy; Sakai and Nagamasu, 1998). All flowers are diurnal, opening in the morning with the exception of Zingiber longipedunculatum, whose flowers open around noon and wither within 6–24 h. Daily flower number per inflorescence is generally less than four. Flowers of all species secrete nectar (Appendix 1). Many ginger species flower more than once a year, and some of them show little synchronized flowering (Sakai, unpublished data).
|
; Kato, Itino, and Nagamitsu, 1993
; Kato, 1996
). If >75% of the total visits to a ginger species were made by members of one of the pollinator groups, the ginger species was categorized into a pollination guild corresponding to the pollinator group. No single pollinator group accounted for >75% of visits to Amomum angustipetalum, and this species was therefore not classified as a guild. There was great variation in sample size of pollinator visits, and pollinators of a few species were identified based on small numbers of visits. It was partly because some species had extremely low population density, long flowering intervals, and/or very short flowering periods. However, considering that most species are rare in tropical forests with overwhelming diversity, it may be misleading to conduct analyses only for the species with high population densities and/or frequent reproductions, especially when discussing diversity and variation. Therefore, we performed analyses for all species whose pollinator was identified. In addition, to avoid mistakes caused by pollinator misidentification, we also conducted analyses excluding each of the ginger species whose pollinator was identified based on a small number of visits and confirmed that the exclusion of the species did not cause large differences.
Nectar production of each ginger species was determined by bagging some flowers before anthesis to exclude flower visitors. Volume and sugar concentration of floral nectar for bagged and open flowers were measured, respectively, by microcapillary tubes and pocket reflectometers between 1000 and 1300. Measurement for Zingiber longipedunculatum was made at 1500, as its flowers opened after 1200. Mean daily sugar production per inflorescence was then calculated in terms of sucrose mass.
Other floral characteristics, such as color and position of the inflorescences, were also described (Appendix 1). Color of the corolla was recorded for exposed flowers. If the corolla was not exposed while flowering, the color of the exposed organs (such as the bracts) was recorded. Inflorescence position was categorized into four classes according to height above the ground: subterranean inflorescences, 0–0.3 m, 0.3–1 m, and >1 m. Total and daily flower numbers on an inflorescence were also recorded. Sugar concentration, nectar production rate, and total flower number were compared among the pollination guilds using Tukey's studentized range test. The test was devised for multiple pairwise comparisons of means between groups and can be applied when sample sizes of the groups are not equal (SAS, 1988
).
Flora morphology
Flowers of Costaceae and Zingiberaceae are monosymmetric and have similar groundplans. Flowers have only a single fertile stamen, while the others are replaced by petaloid staminodes. Three petals are congenitally fused with the androecium at the base. The flowers usually have a prominent lip, which is formed by the union of 3–5 staminodes. The only stamen of the flower is rather thick and hangs over the lip; the lip and stamen are fused into a tube at the base together with three petals; in contrast to the stamen, the pistil is slender and supported by the anthers, which embrace the style from above so that the stigma is exposed above the anthers (Fig. 1). To describe floral morphology, seven floral morphological characteristics were selected and measured: floral tube length, lip length and width, anther length, filament length and width (length and width of a free filament part), and stigma width (Fig. 1). In cases where the lower part of the anthers was within the floral tube without a free filament part, a negative value was assigned to filament length. The seven morphological characteristics were compared among pollination guilds using Tukey's studentized range test.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
). The distribution range of each pollination guild rarely overlapped with other guilds. Sixty-three percent of variation among the guilds was represented by CAN1 and 37% by CAN2. In CDA, the contribution of each characteristic to a reduced variable is expressed by its standardized canonical coefficient. In CAN1, the largest absolute value of a canonical coefficient was for lip width, indicating that this characteristic contributes most to CAN1 (Fig. 3). Its contribution was in a negative direction. The second largest was stigma width followed by anther length and floral tube length, all of which contributed in a positive direction. This means that flowers with larger CAN1 have narrower lips, wider stigmas, and longer anthers and floral tubes. On the other hand, positive contributions in CAN2 were by the size of the pistils and stamens (anther length, filament length, stigma width and filament width) and negative contributions by the size of the corolla (lip width and floral tube length) (Fig. 3). Values of CAN2 indicate robustness of the sexual parts of the flower compared to corolla size.
|
|
Spiderhunter-pollinated guild
Eight ginger species (Amomum roseisquamosum, Etlingera aff. brevilabris, Et. punicea, Hornstedtia reticulata, H. leonurus, H. aff. minor, Et. aff. metriocheilos and Plagiostachys strobilifera) were found to be spiderhunter pollinated. All birds observed visiting these gingers were little spiderhunters (Arachnothera longirostra) with the exception of a single visit by a copper-throated sunbird (Nectarinia calcostetha) on P. strobilifera and by a long-billed spiderhunter (Ar. robusta) on Et. aff. brevilabris (Appendix 2). Spiderhunters are resident birds with long bills and are seen throughout the year. They generally approached the inflorescence carefully and remained for only a short time (10–25 s), but sometimes lingered and sang loudly. Occasionally they used warning calls to discourage other conspecifics from approaching the inflorescence. The flowers of these spiderhunter-pollinated species had long floral tubes (>31 mm with the exception of P. strobilifera) and large anthers (6–16 mm) but short free filament parts (-1.7–3.0 mm with the exception of H. reticulata and H. leonurus). Pollen was generally deposited on the bill or forehead of the pollinator when the spiderhunters inserted their long bills (36 mm for little spiderhunters) into the floral tube and sucked nectar from the flower.
Etlingera punicea (E3, Fig. 4) was located at the center of the spiderhunter-pollinated range on the CDA plane (Fig. 2). This species produced showy flowers with a long yellow lip fringed with red on a subterranean inflorescence. The flowers secreted considerable amounts of nectar (474 mg sugar per day per inflorescence). In addition to the spiderhunters, Amegilla bees were also observed attracted to and flying around the flowers but could not insert their proboscides into the floral tube, which was blocked by the tightly rolled sidelobes of the lip.
|
Amegilla-pollinated guild
All medium-sized bees visiting ginger flowers were Amegilla insularis and A. pendleburyi. Amegilla bees are solitary bees, which make nests underground. They were also observed all year-round in the study site. These long-tongued shade-loving bees flew swiftly near the ground and traplined these rare ginger flowers. Most collected specimens of Amegilla bees were A. pendleburyi (Appendix 2). Eleven ginger species (Costus speciosus, C. globosus, Alpinia glabra, Globba brachyanthera, Amomum calyptratum, A. gyrolophos, A. oliganthum, Elettariopsis sp. 1, Plagiostachys crocydocalyx, P. sp. 1, and Zingiber longipedunculatum) were determined to be Amegilla pollinated. Some of these species had flowers with a large stamen that blocked the entrance of the floral tube and kept nectar thieves away. Amegilla bees visiting these flowers inserted their proboscides under the stamen to lift it and then crawled underneath (Fig. 7). After the bees had consumed the nectar, they pushed up the stamen and came out of the flower with pollen loads on their heads and dorsal thoraxes.
One of the species in the Amegilla-pollinated range on the CDA plane was Plagiostachys crocydocalyx (P1, Fig. 7), which has large white flowers (Fig. 2). The thick petaloid stamen of P. crocydocalyx hangs down over the lip. After crawling under the stamen to consume nectar, the dorsal abdomen of the Amegilla bee was covered with pollen. Although video records revealed that P. crocydocalyx was visited more frequently by little spiderhunters (6.0 flowers/h) than Amegilla bees (1.5 flowers/h), it is unlikely that spiderhunters are important pollinators of the species because spiderhunters inserted their bills into the short floral tubes of P. crocydocalyx (9.5 mm) from the side at the base of the lip without touching the anther and stigma. Thus, visits by spiderhunters were not counted as effective visits for pollination.
Other Amegilla-pollinated species, Plagiostachys sp. 1 (P2), were located in the halictid bee range on the CDA plane (Fig. 2). Small flowers of the species deposited their pollen on the pollinator head with the short filaments and small anthers. Although it seemed possible for halictid bees to visit the flowers, no such visits were observed. Instead, spiderhunters were observed infrequently visiting the flowers and touching the stigmas (Fig. 8).
Zingiber longipedunculatum (Z1, Fig. 9) was located at the border of the three pollination guilds on the plane (Fig. 2). The anther of its flower was almost sessile with a short filament, but the stigma protruded far beyond the anther. The bees observed and collected on the flowers of the species were exclusively female Amegilla collecting pollen. The bees came to the inflorescences before anthesis around 1230 and inserted their proboscides to open the flowers. After anthesis they rubbed their heads on the anther thecae several times and hovered above the flower to collect pollen onto their corbiculae on the hind legs. When the bees landed on the same or next flower, pollen on their corbiculae was acquired by the stigma.
Halictid-pollinated guild
Four species of halictid bees (Nomia sp. 1, N. sp. 2, N. sp. 3, and Thrinchostoma afaciatum) were observed visiting ginger flowers (Appendix 2). Halictids bees are also solitary bees making nests underground and are observed all the year-round, but they are much smaller than Amegilla bees. These shade-loving bees except N. sp. 3 had tongues longer than most other halictid bees, flew swiftly near the ground, and traplined these rare ginger flowers. Ten ginger species (Boesenbergia gracilipes, B. aff. variegata, B. grandifolia, Elettaria longituba, Elettariopsis aff. kerbyi, Elettariopsis sp. 2, Amomum coriaceum, A. durum, A. polycarpum, and A. somniculosum) were pollinated by halictid bees. Halictid-pollinated flowers usually had small anthers (length 2.8–8.5 mm) and short free filament parts (-0.4–3.6 mm). The entrances of their floral tubes were not tightly closed. When visiting the flowers, they landed on the lips and inserted their heads into the floral tubes (Fig. 10). Some of halictid-pollinated flowers had much longer floral tubes (up to 76 mm) than the proboscides of halictid bees (8.2–9.0 mm except N. sp. 3), though they cannot go into slender tubes of the flowers. Pollen was attached to the dorsal thoraxes, abdomens, or heads of the bees visiting the flowers.
Halictid-pollinated Amomum durum (M3, Fig. 11), A. coriaceum (M2), and A. somniculosum (M10) were located near the border of the Amegilla-pollinated and halictid-pollinated ranges. Amegilla bees were observed to try to enter a flower of Amomum somniculosum, whose lip was rolled up to form a tube that blocked Amegilla bees but not halictid bees. Amomum coriaceum and A. durum were observed being visited and also pollinated by Amegilla bees. No mechanism to block Amegilla bees was found in these two ginger species.
Floral nectar and other floral characteristics
More than 80% of the ginger species had inflorescences embedded in or just above the ground (Fig. 12). Some spiderhunter-pollinated or Amegilla-pollinated species produced inflorescences higher than 0.3 m. No halictid-pollinated species had inflorescences higher than 0.3 m above the ground. Total flower number on an inflorescence varied from four to 100 and did not differ significantly among the pollination guilds (Fig. 12).
|
Sugar concentration of floral nectar was slightly lower in spiderhunter-pollinated plants (26 ± 4%) than in bee-pollinated ones (Amegilla-pollinated: 29 ± 4%, halictid-pollinated: 27 ± 6%), but the difference was not significant. Daily sugar production per inflorescence was highly elevated in spiderhunter-pollinated plants (Fig. 12). Sugar mass in nectar produced per day by the spiderhunter-pollinated guild (149 ± 167 mg per inflorescence) was significantly higher than those of both the Amegilla-pollinated (6.1 ± 8.0 mg) and halictid-pollinated (2.5 ± 2.1 mg) guilds (P < 0.05). The difference between the latter two was not significant.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Such diffuse plant–pollinator relationships were also found in Neotropical species in the Zingiberales. Diverse species of hummingbirds and euglossine bees in the Neotropics were replaced by a few species of spiderhunters and Amegilla bees, respectively, while long-tongued small bee-pollinated species were rarely reported as pollinators of Neotropical Zingiberales (Stiles, 1975
, 1978
; Kennedy, 1978
; Schemske, 1981
, 1983
). Generally, hummingbirds and euglossine bees have higher local diversity than spiderhunters and Amegilla bees, and plants pollinated by hummingbirds and euglossine bees can often be divided into subguilds. The different species of euglossine bees have a wide range of body size and tongue length (Dressler, 1982
), and species diversity at reported sites is usually much higher (43–50 species collected with chemical attractants in Costa Rica and Panama; Janzen et al., 1982
; Roubik and Ackerman, 1987
) than those of Amegilla in Southeast Asia (1–4 species; Lieftinck, 1956
). Small euglossine bees in the Neotropics may be in similar position with halictid bees in Lambir. Kennedy (1978)
reported that Calanthea (Marantaceae) flowers with longer tubes were pollinated by long-tongued euglossine bees and those with shorter tubes by short-tongued bees. Stiles (1975,
1978)
, found two pollination guilds in Heliconia pollinated by hummingbirds. One was the guild pollinated by four species of nonterritorial, traplining foragers with long, curved bills and the other by five territorial species with shorter bills. In spite of lower species diversity of pollinators in Lambir, there is not much difference in the number of species of the Zingiberales between Lambir and a wet lowland forest in the Neotropics (54 species in Lambir vs. 54 species in La Selva, Costa Rica).
The syndrome of hermaphroditic flowers in the understory with a low daily output of flowers per plant and pollination by long-distance foragers was found in diverse plant taxa both in the Neotropics and in the Paleotropics, and this suite of reproductive traits has apparently evolved multiple times in tropical angiosperms (Kress and Beach, 1994
). Plants pollinated by spiderhunters, Amegilla, and halictid bees, common pollinators with gingers, were found among other families on the floor or understory of the forest (Kato, 1996
; Momose et al., 1998
). These plants share some similar characteristics with gingers, i.e., rather constant flower production, elongated flowering periods with a low daily output, and specialization for their pollinators. Characteristics of the environment on the forest floor partly explain similarities in plant–pollinator interactions in the two tropics. Flowers on the forest floor with low solar radiation should only provide rewards to reliable pollinators. They constantly produce small amounts of floral resource and protect against other flower visitors while forcing the pollinators to move among plants. They inevitably share the pollinator groups because a plant species alone could not feed a pollinator group. Within such plant–pollinator relationships, plants are thought to be much more influenced by their pollinators than the pollinators are by the plants (Feinsinger, 1983
).
Floral morphology and pollinators
Among the seven morphological characteristics measured, significant differences between the pollination guilds were found in four. Furthermore, CDA showed that three pollination guilds could be mostly distinguished on the plane of CAN1 and CAN2, by two linear functions of the seven characteristics. These results indicate that floral morphological characteristics are related to specific pollinator types.
The morphological characteristics of flowers in each pollination guild can be identified by examining the standardized canonical coefficients of CAN1 and CAN2. CAN1 is most influenced by lip width in a negative direction, followed by stigma width, anther length, and floral tube length in a positive direction. Amegilla-pollinated flowers have smaller CAN1 than the other two. This may be due to the lips often acting as a platform for pollinating bees and the platforms for larger bees are wider than for smaller bees. The lips of halictid-pollinated flowers also act as a platform but are smaller than those of Amegilla-pollinated flowers. The averaged CAN1 of spiderhunter-pollinated flowers was the largest. Species in the spiderhunter-pollinated guild had almost the same range of lip length with halictid-pollinated, but their stigmas and anthers were more robust than those of halictid-pollinated species. It is interesting that the two most important characteristics for CAN1, lip width and stigma width, did not show any significant difference in a pairwise comparison between pollination guilds. This disparity may have arisen because CDA treated variation among the three pollination guilds, while pairwise comparison examined differences between two of the three guilds; CDA also combined multiple variables and could evaluate relative size within a flower.
The fourth most important characteristic was floral tube length. Depth of the floral tube is usually found to have a close relationship with the proboscis or bill length of pollinators (Stiles, 1975
; Kennedy, 1978
; Kato, 1996
; Yumoto, Itino, and Nagamasu, 1997
). Spiderhunters had the longest mouth parts among the three pollinator groups, and spiderhunter-pollinated flowers had the longest floral tubes. However, it is not true of the two bee-pollinated guilds. In spite of the longer tongues of Amegilla bees, Amegilla-pollinated flowers had shorter floral tubes than halictid-pollinated flowers on average, though the difference was not significant. In addition, floral tubes of many halictid-pollinated species were much longer than the tongue length of their pollinators. The long floral tubes may function to defend the ovaries from herbivores by keeping them deep inside the bracts.
The halictid-pollinated guild and the other two were divided along the axis of CAN2, while the halictid-pollinated guild was between the spiderhunter-pollinated and Amegilla-pollinated guilds as for CAN1. Characteristics related to size of the pistils and stamens, anther length, filament length, filament width, and stigma width, contributed to CAN2 in a positive direction, and floral tube length and lip width contributed in a negative direction. CAN2 is considered to indicate robustness of the sexual parts of the hermaphroditic flower compared to size of the whole flower. Halictid-pollinated flowers had relatively delicate sexual parts compared to those of the other two. This may reflect the smaller body size and weaker power of the pollinators to handle and a lower protection requirement for the smaller amount of nectar.
Although the three pollination guilds could be distinguished on the CDA plane, some species were located on the boundary of the guilds or in the range of other pollination guilds. Thus, their floral morphology deviated from that of the typical flowers of their guild. To what extent these atypical flowers have adapted to their pollinators poses an interesting question.
Spiderhunter-pollinated Amomum roseisquamosum (M9) had a shorter floral tube than average for its guild, and dissected flowers bore similarities to other bee-pollinated species. However, under natural conditions, the lower half of its lip is rolled up tightly and functions as a tube. This shows a flexible floral morphology. Plagiostachys strobilifera (P3) had the shortest floral tube of the spiderhunter-pollinated guild. The floral nectar of the species was sometimes accessible to Amegilla or other bees, though they neither touched the anthers nor deposited pollen on the stigma protruding from the floral tube. In this case, floral morphology of the species did not seem to perfectly match pollinator morphology.
Amegilla-pollinated Plagiostachys sp. 1 (P2) and Zingiber longipedunclatum (Z1) are located outside the range of the Amegilla-pollinated guild on the CDA plane. Zingiber longipedunculatum had a different pollination mode from the other species in the guild. When Amegilla bees visited, pollen was deposited on their dorsal surface but pollen was acquired by the stigma not from their dorsal surface but from their corbiculae. This difference explains the successful pollination of Zingiber longipedunclatum despite its location outside the Amegilla-pollinated range. Plagiostachys sp. 1 was located at the center of the halictid-pollinated range on the CDA plane. Although visits by spiderhunters and Amegilla bees to its flowers were observed, no visits by halictid bees were observed.
Amomum durum (M3), A. coriaceum (M2), and A. somniculosum (M10) in the halictid-pollinated guild were located close to the range of the Amegilla-pollinated guild. Amomum somniculosum with its rolled up lip did not allow Amegilla bees to take floral nectar. The other two species, however, were visited and possibly pollinated by Amegilla bees. It can be said that the floral morphology of A. durum and A. coriaceum falls between that of the typical floral morphology of the two bee-pollinated guilds.
The function of variation in floral morphology within a pollination guild is one of the interesting problems that remain to be examined. The species with atypical flowers mentioned above were visited by their pollinators as frequently as the other members of the guild (Table 3). Visit frequency seems to reflect not floral morphology but resource (nectar for spiderhunters, and nectar and pollen for bees) availability of the plant species. However, pollination success is possibly effected by their morphology through other mechanisms. Differences in floral morphology are expected to promote pollinator sharing by differentiating pollen placement position on pollinators (Feinsinger, 1983
; Armbruster, Edwards, and Debevec, 1994
; but see Murcia and Feinsinger, 1996
). Flowering periods of ginger species in the forest are reported to considerably overlap among species within a pollination guild (Sakai, unpublished data), and it is possible that there is competition for pollinators.
Other floral characteristics
Spiderhunter-pollinated species secreted significantly more nectar than the others. Amegilla-pollinated species produced more nectar than the halictid-pollinated species, although the difference was not significant. This reflects a difference in the energy budgets of the pollinators. In general, larger flower visitors consume more energy during foraging (Heinrich, 1975
). The richer the floral reward is, the more tightly the flower is protected. No flower visitors other than pollinators were observed on spiderhunter-pollinated flowers, while Amegilla-pollinated Plagiostachys crocydocalyx and halictid-pollinated A. polycarpum were visited by spiderhunters, which did not contribute to pollination at all. It is probable that the small amount of floral resource is not worth the cost of protection and that less rewarding flowers are less attractive to nectar robbers.
Flower color is important to attract pollinators. Correlations between red flowers and bird pollination, and orange, yellow or white flowers and bee pollination, which have been reported for many plants (Endress, 1994
), were also detected in gingers. Conversely, many ginger species produced subterranean inflorescences, and there seemed to be little vertical difference among pollination systems. Total flower number on an inflorescence appeared to be related to the length of the flowering period of an inflorescence and the cost of a flower. Highly rewarding spiderhunter-pollinated species tended to have a smaller number of flowers, but the differences among the pollination guilds were not significant.
Structure of the plant–pollinator relationship
Floral morphology and the amount of floral nectar were correlated with the pollination guilds. However, there were considerable variations in morphology within the guilds, and the distributions of morphology of the pollination guilds were more or less continuous. From CDA analyses and field observations of flower visitors, it was observed that the floral morphology of some species did not perfectly match their pollinators.
Whether or not a pollinator visits the flower may partly depend on the amount of available nectar. This in turn is determined partly by the amount of nectar the plant produces and partly by the visit frequency of consumers. If the density or species compositions of the plants or pollinators changes, the amount of nectar may also change and the flower visitors may shift to another plant species. When the floral morphology is clearly different among pollination guilds, it is difficult for pollinators to shift to another plant species and for plants to overcome changes in fauna of flower visitors. In contrast, when there are species with flowers whose morphology falls between that of the typical flowers of two pollination guilds, they may be pollinated by either pollinator depending on the conditions. At different places, population densities of plants and pollinators, or pollinator fauna or flora are different and the plants and pollinators presumably have different relationships. Kato, Itino, and Nagamitsu (1993)
have reported that Sumatran Etlingera punicea was pollinated by Amegilla bees with longer tongues than those in Borneo, although the species was visited only by spiderhunters in Lambir. Though most birds visiting ginger flowers belonged to a single species, if there were two morphs in visiting birds in bill length, subguilds with different floral tube length might be formed within the bird-pollinated guild, since there seemed to be enough intraguild variation in floral tube length. Plant–pollinator interactions may be more flexible than previously thought.
Deviations of some species from the range of the guild on the CDA plane were partly explained by different pollination modes of the species or different functions of floral parts from the other species in the same pollination guild. Deviating species were usually found in Amomum and Plagiostachys, which included members of two or more pollination guilds. In genera where all members shared common pollinators, such as halictid-pollinated Boesenbergia and spiderhunter-pollinated Etlingera, the flowers tended to have the typical morphology of the pollination guild. The phylogenetic background may be related to these deviations in floral morphology.
|
|
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Campbell, D. R. 1989 Measurements of selection in a hermaphroditic plant: variation in male and female pollination success. Evolution 43: 318–334.[CrossRef][ISI]
———. 1996 Evolution of floral traits in a hermaphroditic plant: field measurements of heritabilities and genetic correlations. Evolution 50: 1442–1453.[CrossRef][ISI]
———, N. M. Waser, M. V. Price, E. A. Lynch, and R. J. Mitchell. 1991 Components of phenotypic selection: pollen export and flower corolla width in Ipomopsis aggregata. Evolution 45: 1458–1467.
Classen, R. 1987 Morphological adaptations for bird pollination in Nicolaia elatior (Jack) Horan (Zingiberaceae). Garden's Bulletin Singapore 40: 37–43.
Dahlgren, R. M. T., H. T. Clifford, and P. F. Yeo. 1985 The families of the monocotyledons. Springer Verlag, Berlin.
Dressler, R. L. 1982 Biology of the orchid bees (Euglossini). Annual Review of Ecology and Systematics 13: 373–94.
Endress, P. K. 1994 Diversity and evolutionary biology of tropical flowers. Cambridge University Press, New York, NY.
Feinsinger, P. 1983 Coevolution and pollination. In D. J. Futuyma, and M. Slatkin [eds.], Coevolution, 283–310. Sinauer, Sunderland, MA.
Galen, C. 1989 Measuring pollinator-mediated selection on morphometric floral traits: bumblebees and the alpine sky pilot, Polemonium viscosum. Evolution 43: 882–890.[CrossRef][ISI]
Gilbert, L. E., and R. H. Raven [eds.]. 1975 Coevolution of animals and plants. University of Texas Press, Austin, TX.
Harder, L. D. 1985 Morphology as a predictor of flower choice by bumble bees. Ecology 66: 198–210.[CrossRef][ISI]
Heinrich, B. 1975 Energetics of pollination. Annual Review of Ecology and Systematics 6: 139–170.
Inoue, T., and M. Kato. 1992 Inter and intraspecific morphological variation in bumble bee species, and competition in flower utilization. In M. D. Hunter, T. Ohgushi, and P. W. Price [eds.], Effects of resource distribution on animal-plant interactions, 393–427. Academic Press, San Diego, CA.
———, T. Yumoto, A. A. Hamid, H. S. Lee, and K. Ogino. 1995 Construction of a canopy observation system in a tropical rainforest of Sarawak. Selbyana 16: 100–111.
Ippolito, A., and J. E. Armstrong. 1993 Floral biology of Hornstedtia scottiana (Zingiberaceae) in a lowland rain forest of Australia. Biotropica 25: 281–289.[CrossRef][ISI]
Janzen, D. H. 1971 Euglossine bees as long distance pollinators of tropical plants. Science 171: 203–205.
———. 1980 When is it coevolution? Evolution 34: 611–612.[CrossRef][ISI]
———, P. J. Devries, M. L. Higgins, and L. S. Kimsey. 1982 Seasonal and site variation in Costa Rican Euglossine bees at chemical baits in lowland deciduous and evergreen forests. Ecology 63: 66–74.[CrossRef][ISI]
Kato, M. 1996 Plant-pollinator interactions in the understory of a lowland mixed dipterocarp forest in Sarawak. American Journal of Botany 83: 732–743.[CrossRef][ISI]
———, T. Itino, and T. Nagamitsu. 1993 Melittophily and ornithophily of long-tubed flowers in Zingiberaceae and Gesneriaceae in West Sumatra. Tropics 2: 129–142.
Kennedy, H. 1978 Systematics and pollination of the ‘closed flowered’ species of Calanthea (Marantaceae). University of California Publications of Botany 71: 1–90.
Kress, W. J., and J. H. Beach. 1994 Flowering plant reproductive systems. In L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn [eds.], La Selva, ecology and natural history of a Neotropical rain forest, 161–182. University of Chicago Press, Chicago, IL.
Lieftinck, M. A. 1956 Revision of some oriental Anthophorine bees on the genus Amegilla Friese (Hymenoptera, Apoidea). Zoologische Verhandelingen 30: 1–44.
Maas, P. J. M. 1979 Notes on Asiatic and Australian Costoideae (Zingiberaceae). Blumea 25: 543–549.
Momose, K., T. Yumoto, T. Nagamitsu, M. Kato, H. Nagamasu, S. Sakai, R. D. Harrison, T. Itioka, A. A. Hamid, and T. Inoue. 1998 Pollination biology in a lowland dipterocarp forest in Sarawak, Malaysia. I. Characteristics of the plant-pollinator community in a lowland dipterocarp forest. American Journal of Botany 85: 1477–1501.
Murcia, C., and P. Feinsinger. 1996 Interspecific pollen loss by hummingbird visiting flower mixtures: effects of floral architecture. Ecology 77: 550–560.[CrossRef][ISI]
Poulsen, A. D. 1997 The herbaceous ground flora of the Batu Apoi Forest Reserve, Brunei Daruesalam. In S. Edwards, W. E. Booth, and S. C. Choy [eds.], Tropical rainforest research — current issues. Kluwer Academic Publishers, Dordrecht.
Roubik, D. W., and J. D. Ackerman. 1987 Long-term ecology of euglossine orchid-bees (Apidae: Euglossini) in Panama. Oecologia 73: 321–333.[CrossRef][ISI]
Sakai, S., and H. Nagamasu. 1998 Systematic studies of Bornean Zingiberaceae I. Amomum in Lambir Hills. Edinburgh Journal of Botany 55: 45–64.
SAS. 1988 SAS user's guide. Statistics, version 6. SAS Institute, Cary, NC.
Schemske, D. W. 1981 Floral convergence and pollinator sharing in two bee-pollinated tropical herbs. Ecology 62: 946–954.[CrossRef][ISI]
———. 1983 Breeding system and habitat effects on fitness components in three Neotropical Costus (Zingiberaceae). Evolution 37: 523–539.[CrossRef][ISI]
Smith, R. M. 1985 A review of Bornean Zingiberaceae I: (Alpinieae). Notes from Royal Botanical Garden Edinburgh 42: 261–314.
———. 1986 A review of Bornean Zingiberaceae II: (Alpinieae, concluded). Notes from Royal Botanical Garden Edinburgh 43: 439–466.
———. 1987 A review of Bornean Zingiberaceae: III (Hedychieae). Notes from Royal Botanical Garden Edinburgh 44: 203–232.
———. 1988 A review of Bornean Zingiberaceae: IV (Globbeae). Notes from Royal Botanical Garden Edinburgh 45: 1–19.
———. 1989 A review of Bornean Zingiberaceae V: (Zingiberaceae). Notes from Royal Botanical Garden Edinburgh 45: 409–423.
Stiles, F. G. 1975 Ecology, flowering phenology, and hummingbird pollination of some Cost Rican Heliconia species. Ecology 56: 285–301.[CrossRef][ISI]
———. 1978 Temporal organization of flowering among the hummingbird foodplants of a tropical wet forest. Biotropica 10: 194–210.[CrossRef][ISI]
Yumoto, T., T. Itino, and H. Nagamasu. 1997 Pollination of hemiparasites (Loranthaceae) by spiderhunters (Nectariniidae) in the canopy of a Bornean tropical rainforest. Selbyana 18: 51–60.
This article has been cited by other articles:
|
|
 |
|
  Y.-Q. WANG, D.-X. ZHANG, and Z.-Y. CHEN Pollen Histochemistry and Pollen : Ovule Ratios in Zingiberaceae Ann. Bot., October 1, 2004; 94(4): 583 - 591. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-Y. Gao, L. Zhang, X.-B. Deng, P.-Y. Ren, J.-J. Kong, and Q.-J. Li The floral biology of Curcumorpha longiflora (Zingiberaceae): a ginger with two-day flowers Am. J. Botany, February 1, 2004; 91(2): 289 - 293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. OLLERTON, S. D. JOHNSON, L. CRANMER, and S. KELLIE The Pollination Ecology of an Assemblage of Grassland Asclepiads in South Africa Ann. Bot., December 1, 2003; 92(6): 807 - 834. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. L. BORBA, G. J. SHEPHERD, C. V. D. BERG, and J. SEMIR Floral and Vegetative Morphometrics of Five Pleurothallis (Orchidaceae) Species: Correlation with Taxonomy, Phylogeny, Genetic Variability and Pollination Systems Ann. Bot., August 1, 2002; 90(2): 219 - 230. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Yumoto Bird-pollination of three Durio species (Bombacaceae) in a tropical rainforest in Sarawak, Malaysia Am. J. Botany, August 1, 2000; 87(8): 1181 - 1188. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||
: spiderhunter-pollinated; 
: Amegilla-pollinated; 
: halictid-pollinated; +: not identified). Mean CAN1 and CAN2 values of the guilds are indicated by S (spiderhunter-pollinated), A (Amegilla-pollinated), and H (halictid-pollinated). For the plant species codes, see 
