| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Population Biology |
2Botany Department, University of Cape Town, P Bag Rondebosch, South Africa 7701; 3Centre of Excellence in Natural Reserve Management, University of Western Australia, 444 Albany Highway, Albany WA 6330, Australia
Received for publication September 10, 2002. Accepted for publication February 20, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: autogamy • facilitated selfing • mutualism • pollination • reproductive assurance • Roridula • seed set • self-fertilization
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Baker, 1955
; Jain, 1976
; Lloyd, 1980
; Holsinger, 1996
), even if inbreeding depression is strong (Schoen and Brown, 1991
; Lloyd, 1992
; Schoen et al., 1996
). However, the evolution of selfing as a mechanism for reproductive assurance depends on selfed progeny having enough fitness to reproduce and on the assumption that selfing does not occur at the expense of cross-pollination (Lloyd, 1979
). Reproductive assurance is when plants self-pollinate to ensure seed production when pollinators and/or potential mates are scarce (Herlihy and Eckert, 2002
). It may be promoted through autonomous autogamy—within-flower selfing (Lloyd and Schoen, 1992
). Alternatively it may be promoted by facilitated autogamy—within-flower selfing caused by animal pollen vectors (Lloyd and Schoen, 1992
). Although facilitated selfing is a common mechanism leading to mixed mating systems (Richards, 1986
), Lloyd (1992)
postulated that facilitated selfing is never of great benefit to plants because it provides very little, if any, reproductive assurance. One of his primary reasons for this is that facilitated selfing is only of value when pollinators are very scarce. But because facilitated selfing requires pollinators, it does not provide assurance when it is most needed (Lloyd, 1992
). However, this assumes that the conditions limiting outcrossing pollinators also limit the pollinators that facilitate selfing. Reproductive assurance can be tested by emasculating flowers. Emasculated flowers should set fewer seeds than intact flowers if seed set is limited by pollen from other flowers (Schoen and Lloyd, 1992
). Reproductive assurance can only operate when pollen from other flowers (geitonogamous and xenogamous) limit seed set. However, the benefits of reproductive assurance will be negated if selfed seeds, which are inferior to outcrossed seeds from inbreeding depression, preclude the development of outcrossed seed (Herlihy and Eckert, 2002
). This event is commonly called "seed discounting," the absence of which is important for reproductive assurance to be effective (Herlihy and Eckert, 2002
).
Although reproductive assurance is thought to be a primary factor driving the evolution of selfing, there have been few experimental field studies on whether self-fertilization actually does provide reproductive assurance. Of these, few examine seed discounting (Schoen and Lloyd, 1992
). The importance of reproductive assurance is controversial because the majority of experimental studies have shown that xenogamous and geitonogamous pollen does not limit seed production and that self-fertilization provides no reproductive assurance (Bernhardt, 1976
; Cruden and Lyon, 1989
; Leclerc-Potvin and Ritland, 1994
; Klips and Snow, 1997
; Eckert and Schaefer, 1998
). Only two studies (Motten, 1982
; Piper et al., 1986
) indicate that xenogamous and geitonogamous pollen may limit seed production and that self-fertilization can potentially provide reproductive assurance. Only Herlihy and Eckert (2002)
have combined experimental measures of reproductive assurance, seed discounting, selfing rates, and inbreeding depression. They showed in Aquilegia canadensis that autonomous selfing increases seed production. However, this benefit is outweighed by the loss of high quality seed as a result of seed discounting and inbreeding depression. Their results challenge the widely held view that reproductive assurance is a major factor driving the evolution of selfing.
The other general hypothesis for the evolution of self-fertilization is the automatic selection hypothesis (Schoen et al., 1996
). Under this hypothesis, genes promoting selfing spread more rapidly through a population than those promoting outcrossing, because selfers potentially sire seed on the same plant as well as other plants. In contrast, outcrossers can only sire seeds on other plants (Fisher, 1941
). However, the spread of genes promoting selfing is opposed by inbreeding depression, which only allows selfing to evolve if the fitness of progeny from selfing is greater than half that of outcrossed progeny (Fisher, 1941
).
In this study, we examine the pollination biology of the two species that comprise the family Roridulaceae, experimentally testing whether facilitated autogamy can provide reproductive assurance and thus be a strong selective pressure for the evolution of self-compatibility and self-fertilization. Roridula is a genus found in small, isolated populations in the fynbos biome of South Africa (Obermeyer, 1970
). The fynbos biome is characterized by frequent fires that are thought to maintain species richness and diversity (Cowling et al., 1992
). A wide variety of plant responses have evolved to cope with these disturbances. Although many plant species from a range of families resprout after fire (le Maitre and Midgley, 1992
), other plants, including Roridula do not resprout after fire. Instead, seed germination is stimulated by fire, and a single cohort of seedlings recruits after each fire event. Thus, reliable seed set between fires is crucial.
The family Roridulaceae consists of two geographically separated species, R. gorgonias Linnaeus and R. dentata Planchon (Obermeyer, 1970
; Carlquist, 1976
). Roridula dentata has pendulous, pink flowers with five ovate petals (c. 12 mm) and five stamens. Plants may have from one to 50 flowers depending on the age of the plant. Anthers are irritable, deflexed in bud, becoming erect soon after flower opening. The anthers can become erect with irritation as a result of insect activity or on their own (several hours to one day after flower opening). The tips of erect anthers (where pollen is released through a pore) lie close to the stigmas. Stigmas are receptive on flower opening although anthers normally release pollen on the second day of flower opening. Fruit capsules are about 10 mm long and take several months to mature. Each capsule has three locules, each with a single, large seed (c. 5 mm long). Roridula gorgonias flowers are also pink and petals about 15 mm long. Anthers are also irritable, and anthesis takes place in a similar way to R. dentata. However anthers of R. gorgonias dehisce down the sides and contain greater amounts of pollen than R. dentata. Capsules are of similar size to R. dentata although each locule can contain up to four seeds (c. 2.5 mm long). Plants of both species have sticky leaves that catch large numbers of insect prey. Carnivorous hemipterans (Miridae) Pameridea roridulae Reuter and Pameridea marlothi Poppius live permanently and obligately on R. gorgonias and R. dentata, respectively (Dolling and Palmer, 1991
). They roam the plants, apparently unaffected by the sticky secretions, consuming snared prey (Dolling and Palmer, 1991
; Ellis and Midgley, 1996
). Pameridea defecates on Roridula leaves, and nitrogen is then absorbed through the leaf cuticle (Ellis and Midgley, 1996
; Anderson and Midgley, 2002
).
Marloth (1903
, 1925
) observed that juvenile Pameridea apparently play a role in the pollination of Roridula. However, Lloyd (1934)
expressed doubts regarding Marloth's anecdotal accounts of Roridula pollination. Johnson (1992)
and Vogel (1978)
also noted that the flower and anther structure of R. dentata corresponds to a buzz pollination syndrome, and Vogel (1978)
postulated that Roridula is adapted to pollination by Apidae. But the flowering time of R. dentata extends from late June to November and that of R. gorgonias is from late June to late August. These flowering times span a large proportion of the Cape winter (June through August) when pollinators are scarce, contrary to the bee pollination theory. Moreover, the sticky traps of R. dentata are easily capable of catching large Apidae (Marloth, 1903
). In addition, the peduncles of R. dentata are very short and flowers are set amongst sticky traps, making for a difficult approach by bees (Marloth, 1925
). Marloth (1925)
postulated that Pameridea pollinators are Roridula's solution to the sticky trap problem. However, the large, bright showy flowers of Roridula seem "unnecessary" for the purpose of attracting pollinators that are already on the plant in large numbers (Lloyd, 1934
).
In this study, we used a combination of field observations and floral manipulations to elucidate the pollen vectors in the system. By excluding all possible pollinators and then all flying pollinators (but not sessile hemipterans), we determined the contribution of hemipterans to seed and fruit set. Using allozyme electrophoresis, we calculated the outcrossing rates of seed in unmanipulated flowers from several populations. Finally, we use both floral manipulations and allozyme electrophoresis to show the contribution of facilitated autogamy to seed set in Roridula.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Pollination
Visits to flowers by potential pollinators (other than Pameridea) were noted from an observation point where approximately 400–600 flowers could be seen. Observations took place on sunny days when insect pollinators are most active and Roridula flowers were open. A total of 69 h of observation over three flowering seasons (May–November, 17 d) were made for R. dentata at four different populations (Pop7, Pop8, Pop15, Pop16). Fifty-three hours of observations over three flowering seasons (June–August, 14 d) were made for R. gorgonias at five different populations (Pop1–5). Visits were recorded if insects alighted on the petals of the flowers. Small visiting insects were captured, killed, and immediately placed onto a slide with molten glycerol jelly and fuschin stain. The insects were later examined for pollen using a microscope. Pollen on the insects was compared to a microscope slide of Roridula pollen taken directly from the flowers. Larger visiting insects were also caught and later examined for pollen using a microscope.
Line transects were walked through one R. dentata population (Pop7) and random flowers examined every meter for the presence of potential pollinators (n = 275). Flower-visiting fauna (including Pameridea) were noted, and if possible, specimens were caught. Insects were preserved and examined for pollen grains as described. Flowers from a R. gorgonias population (Pop1) were also examined using the same methods. Because there were only 50 plants in this population, more than one flower was sampled per plant (n = 100).
Floral manipulations were performed in one R. dentata population (Pop7) and one R. gorgonias population (Pop1) to examine seed set due to autogamy and hemipterans. Eight different floral manipulations were carried out. (1) Flying insects were excluded from the flowers by bagging a rosette of leaves with an unopened bud on it (R. dentata, n = 30; R. gorgonias, n = 36). Bags were made from very light, transparent nylon pantyhose material, supported by a thin wire frame of florist wire to keep the bag away from the flower. At harvest time, bagged rosettes still contained live Pameridea. (2) Anthers were removed from 30 R. dentata flower buds on 30 plants to estimate the contribution of geitonogamy and xenogamy to fruit/seed set. If reproductive assurance is occuring in this system, we expect that emasculation will cause a significant decline in seed or fruit set. However, in this system pollen is the only reward offered by Roridula, and emasculation can potentially make flowers less attractive to pollinators. Thus it is impossible to determine if decreased fruit/seed set is due to a limitation of xenogamous/geitonoganous pollen or due to flowers being less attractive to pollinators. Consequently the emasculation of Roridula can lead to the rejection of the reproduction assurance hypothesis if emasculated flowers reach full seed/fruit set. However, a decline in seed/fruit set of emasculated flowers cannot be taken as clear support for the hypothesis. (3) To test the effect petals have on pollinator attraction, the petals of 30 R. dentata flower buds were removed. (4) To control for the effect of physical damage by petal removal, the petals of another 30 R. dentata flower buds were removed and the flowers cross-pollinated. Experimental cross-pollination in this treatment (and other hand-pollination treatments) were carried out as follows: Flower buds were emasculated and bagged before reproductive structures were mature. Pollen was shaken from several flowers into a small container and then transferred onto the receptive stigmas. Flowers were rebagged. (5) All pollinators were removed by bagging a flower and ensuring no hemipterans were in the bag (R. dentata, n = 50; R. gorgonias, n = 22). One-and-a-half months later, the seed capsules from all treatments were collected, dissected, and their seeds counted. At harvest, (6) natural seed set was examined along transects, and a seed capsule was picked every meter, one arm's length away (R. dentata, n = 226; R. gorgonias, n = 96). Seeds in the capsules were counted. Exclusion of hemipterans but not flying pollinators was impossible because hemipterans traversed sticky barriers between them and the flowers. Because only 50 plants existed at the R. gorgonias study site, more than one capsule was taken per plant (approximately 3000 capsules existed in the population at the time of sampling).
In addition to those flower manipulations, two more treatments were completed to determine the plants' breeding system: (7) hand-crossing flowers (R. dentata, n = 30; R. gorgonias, n = 16) and (8) self-pollinating (R. dentata, n = 30; R. gorgonias, n = 16). Pollen for the selfed treatment was obtained from other flowers on the same plant. After pollination, flowers were rebagged, ensuring that no hemipterans were in the bags. Floral manipulations 1, 5, 6, 7, and 8 were performed on R. gorgonias and treatments 1–8 were performed on R. dentata. Treatments 2–4 were not applied to R. gorgonias because population size was small, and these extra treatments may have had detrimental effects on the seed set of the population. The data analyzed were fruit set (i.e., the numbers of flowers from each treatment that set seed vs. the number that did not). Mean seed set of each treatment was also calculated by counting the number of seeds with embryos in each fruit capsule (including those that set no seed). Both of these measurements estimate pollinator visitation rates. The effects of the treatments (fruit set) were compared using Fisher's Exact Test and the sequential Bonferroni method was used for significance values to correct for type I errors (Rice, 1989
).
Allozyme electrophoresis
We collected seed material from 11 R. dentata localities (Pop7–Pop17) and six R. gorgonias localities (Pop1–Pop6). If plant populations were smaller 20 individuals, then one capsule was collected from each plant in the population. In larger populations, between 20 and 24 capsules were collected. In large populations, capsules were haphazardly picked along two perpendicular, bisecting transects from within half a meter of the transect line. The distance between samples depended on the population size. Seeds were placed on ice and later frozen at –80°C. Endosperm from Roridula seeds was excised, then homogenized in vegetative extraction buffer I of Cheliak and Pitel (1984)
with a glass rod attached to a variable-speed electric motor. Samples were used within 4 wk of collection and centrifuged for 5 min prior to use. Filter paper wicks were dipped into the supernant of centrifuged samples and inserted into 13% horizontal starch gels. The amount of sample tissue from each seed was very small, and a maximum of two gels were run per sample. Thus, we were unable to obtain a complete genotype of all loci for each individual.
The enzymes MDH (Enzyme Commission [E.C.] 1.1.1.37), ADH (E.C. 1.1.1.1), GPI (E.C. 5.3.1.9), and DIA (E.C. 1.6.2.2) were resolved on a continuous histidine-citrate buffer system, pH 6.0 (Stuber et al., 1977
). Pep LGG (E.C. 3.4), IDH (E.C. 1.1.1.42), ME (E.C. 1.1.1.40) and PGM (E.C. 5.4.2.2) were resolved using a discontinuous Tris-citrate-borate-lithium hydroxide buffer, with a gel buffer pH 8.7 and an electrode buffer pH 8.0 (Ridgeway et al., 1970
). G6PDH (E.C. 1.1.1.49) and MPI (E.C. 5.3.1.8) were resolved using a continuous Tris-borate-EDTA buffer system at pH 8.6 (Markert and Faulhaber, 1965
). The enzymes PGM (one locus), MPI, and IDH (both loci) could only be resolved clearly for R. dentata and not R. gorgonias.
Inbreeding was roughly estimated using the inbreeding coefficient (Fis). Fis and significance values were calculated using the program Genepop (Raymond and Rousset, 1995
) and the selfing rate (s) was approximated using the equation Fis = s/(2 – s) (Barrett and Kohn 1991
). Fis estimates gene flow within populations, which gives us important information about the biology and movement of the pollinators that contribute to plant gene flow. If Fis = 0, there is random mating within a population, whereas a value of one would indicate total inbreeding within the population.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Roridula gorgonias
During the observations on R. gorgonias, several species of insect visited the flowers. The most frequent visitors were Pameridea, which were present in 60% of all open flowers examined. On average, each flower contained 1.6 ± 1.8 hemipterans at any time. Of these hemipterans, 68% carried Roridula pollen. Of all the hemipterans in flowers, only 0.5% were adults, even though adults comprise approximately 7.5% of the total hemipteran population (B. Anderson, personal observation). Hemipteran juveniles were observed probing pollen on dehiscent anthers. This pollen was probably eaten as Miridae are known to be pollen feeders (Proctor et al., 1996
). Thrips were also occasionally found in flowers, although in comparison to Pameridea they were rare, occurring on 0.6% of all flowers examined. Six visits were recorded from the small bee Allodape punctata (Anthophoridae) and 50% were carrying pollen. Two visits from the carpenter bee, Xylocopa albifrons (Anthophoridae), were recorded (both bees became briefly entangled in the leaves). One was captured and it was carrying pollen. During our observation X. albifrons were frequently seen flying past Roridula populations but few landed. A visit by a blister beetle, Ceroctis capensis (Meloidae), was recorded once. This species seems to be opportunistic and destructive (consumes petals) and was found feeding on a wide variety of flower species. The blister beetle was observed flying between flowers and consuming floral parts. It was covered in Roridula pollen and was most likely cross-pollinating flowers, although potentially damaging the flowers.
Pollination
Trends in fruit set and seed set follow each other closely (Figs. 1 and 2), and they both measure pollinator visitation rates in this study. Hence, we only refer to fruit set in the results and discussion. The removal of all potential pollinators other than Pameridea from R. dentata did not significantly reduce fruiting success compared to open-pollinated control flowers (treatment 1 vs. 6, 
2 = 2.74, P > 0.05); thus, seeds produced in this treatment are either the product of autogamy or the result of Pameridea pollinators. But the removal of all potential pollinators (autonomous seed production) including Pameridea resulted in a 68% reduction in fruiting success (treatment 5 vs. 6, 2 = 75.38, P < 0.001). The difference between treatments 1 and 5 is the contribution of the hemipterans to fruit set (Fig. 1). The removal of R. dentata petals also significantly reduced fruit set (treatment 3 vs. 4, 2 = 8.29, P < 0.05), suggesting that petals help attract pollinators. Fruiting success in the petal control treatment was not significantly lower than in hand-crossed treatments, indicating that the removal of petals caused no damage to stigmas or pollen (treatment 4 vs. 7, 2 = 19.72, P > 0.05). Removal of petals produced the same seed set as when all pollinators were excluded (treatment 3 vs. 5, 2 = 0.29, P > 0.05), suggesting a visitation rate of close to zero. Emasculation of anthers ensured that no autogamy took place although pollinators still had access to flowers. This caused a large drop in fruit set (treatment 2 vs. 6, 2 = 91.66, P < 0.001), suggesting that most fertilization events involved pollen from the same flower.
|
|
2 = 2.44, P > 0.05). Almost all flowers in these treatments set fruit, whereas the removal of all pollinators from the system significantly lowered fruit set (treatment 5 vs. 6, 2 = 39.7, P < 0.001). This 25% reduction in seed set is attributable to pollinators, primarily Pameridea.
Crossed and selfed R. dentata flowers had equal success in fruiting (treatment 7 vs. 8, Fig. 1, 2 = 4.25, P > 0.05), and the mean seed set of these two treatments was also very similar (Fig. 1), suggesting that selfing is possible. Crossed and selfed R. gorgonias flowers were also equally successful at fruiting (treatment 7 vs. 8, Fig. 2, 2 = 100, P > 0.05). Differences in fruiting success of hand-pollinated flowers vs. unmanipulated flowers were nonsignificant (treatment 6 vs. 7, Fig. 2, 2 = 2.15, P > 0.05).
Allozyme electrophoresis
In populations in which diallelic loci were scored, all differed significantly from the null hypothesis of random mating (Table 1). Selfing rates in all R. dentata populations were very high, ranging from 0.918 to 1 (Table 1).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
recognized that specialized (i.e., relatively species-specific) pollination systems often have "compensation" measures that ensure their persistence in the event of pollinator failure. Self-compatibility and autogamy (facilitated and autonomous) are likely to perform this role in Roridula. Although Roridula flowers are seemingly adapted for bee pollination, bees were rarely seen pollinating flowers. The exclusion of only flying pollinators from flowers (treatment 1) did not decrease fruit set. Instead, fruit set was maintained in this treatment at values close to natural fruit set because hemipterans were still able to forage in flowers and they facilitated pollination. Because the hemipterans were constrained by nylon bags to a single flower, fruit set was autogamous. In contrast, the removal of all pollinators, including Pameridea, resulted in a very large drop in the fruiting success of both Roridula species. This drop suggests that self-pollination in the absence of Pameridea is not very efficient and that Pameridea is an important pollinator and can account for 68% of fruit set in R. dentata and 25% in R. gorgonias. The majority of hemipterans found in flowers were juveniles that were small (<3 mm), wingless, and not very motile (B. Anderson, personal observation). Because of their wingless state, juvenile Pameridea are unlikely to move pollen between plants and are therefore only likely to contribute geitonogamous or selfed pollen. Using allozymes, we found that when Fis and selfing rates could be calculated, very high degrees of selfing were always suggested in populations of both Roridula species. Thus, Roridula is predominantly a selfer, and outcrossing events are extremely rare in all populations tested. This is consistent with field observations and floral manipulations that show that the dominant pollinators are sessile hemipterans. Direct evidence for the absence of bee pollination is only available from their rarity observed in the field. However the scarcity of bee pollination is also supported by the rarity of outcrossed seeds in this system (bees are likely to have a large outcross component to their pollination).
In this system, outcrossed pollination events cannot account for the high fruit and seed set observed in Roridula. Without the aid of hemipterans, autogamy does not lead to the maximum seed set in R. dentata. Although autonomous autogamy is more effective in R. gorgonias than in R. dentata, it does not account for the full fruit set in either species of Roridula. Levels of natural fruit set can only be attained in the populations analyzed by a combination of autonomous selfing and pollination by juvenile hemipterans (almost entirely selfed or geitonogamous). Thus autonomous selfing and facilitated selfing are likely to play important reproductive assurance roles in Roridula.
The adaptive significance of facilitated selfing
One of the perceived obstacles to facilitated selfing offering reproductive assurance is that it is still reliant on pollinators at times when pollinators are supposedly scarce (Lloyd, 1992
). By stating that facilitated selfing provides no reproductive assurance when it is most needed, Lloyd (1992)
assumes that the abundance/effectiveness of all pollinators fluctuate together. However we show that although one guild of outcrossing pollinators may be rare, another guild of selfing pollinators may be common. We expect that facilitated selfing may be selected for when plants are pollinated by two guilds of insects with very different behavioral/life history traits. The first guild should be motile, contributing to outcrossed seed set. If the presence of this guild is unpredictable in time and space, a second guild may also be selected for. The primary characteristic of the second guild should be predictability in time and space (despite the fact that they may contribute mainly selfed seed). In Roridula, bees represent the first guild and hemipterans the second guild. A similar system was studied by Baker and Cruden (1991)
in which thrips reliably pollinated Potentilla in the absence of more motile pollinators. Zamora (1999)
found a similar situation in which thrips were important pollinators of Pinguicula in shady areas but not sunny areas when larger pollinators were more common. Baker and Cruden (1991)
hypothesized that pollination by nonmotile pollinators such as thrips may be an important and widespread phenomenon. However, Lloyd (1992)
refers to these small pollinators as squatters because they spend large amounts of time in flowers. He also suggests that pollination by squatters is generally not selected for because they may have detrimental genetic and physical effects on plants and their movements are often unpredictable. However, in the Roridula system, large numbers of hemipterans are found on every plant and are extremely prevalent year round in all life stages (B. Anderson, personal observation). This prevalence is important for both Roridula species because much of their flowering seasons stretch through winter when other pollinators are dormant or inactive (e.g., Hepburn and Jacot-Guillarmod, 1991
). Pameridea are also species-specific because they only occur on Roridula. As a result, Roridula loses no pollen to the stigmas of other plant species and pollen from other plants does not clog Roridula's stigmas. Another drawback of facilitated selfing is that it may often lead to seed discounting where self-pollination of ovules may preempt ovules that would otherwise have been outcrossed (Lloyd, 1992
). However, this disadvantage is minimized when cross-pollination events are extremely rare or absent. In addition, Roridula has weakly delayed self-pollination in which the stigma becomes receptive approximately 1 d before the anthers (B. Anderson, personal observation). This delay allows geitonogamous and xenogamous pollen a temporal advantage over autogamous pollen.
One of Lloyd's (1934)
arguments against hemipterans being pollinators of Roridula was the presence of showy petals. The need for advertising is another factor thought to limit the advantages of facilitated selfing because it is more expensive than autonomous modes of selfing (Lloyd, 1992
). This study shows that petals still have a vital function because their removal causes a large drop in fruit set. The large reduction in seed set caused by removing petals suggests that petals are also attractive to resident Pameridea. Flowers of R. gorgonias close at night, whereas those of R. dentata do not (B. Anderson, personal observation). Large numbers of juvenile hemipterans were observed sheltering in R. gorgonias flowers at night (personal observation), so flowers may also act as a refuge for Pameridea. Roridula dentata occurs in a much drier and hotter part of the country than R. gorgonias and its flowers are pendulous, whereas the flowers of R. gorgonias face the sun (personal observation). Flowers of R. dentata can be used by hemipterans in a thermoregulatory way, and the umbrella of petals may be used by hemipterans as shade. These results suggest that increased seed production alone (i.e., without lower advertising costs) is enough of a benefit for the evolution of selfing. Petals may play a role in attracting bees for rare cross-pollination events as well as hemipterans to ensure maximum seed set.
Emasculating flowers caused a large drop in the fruit set of R. dentata, which is consistent with the reproductive assurance hypothesis. Note that this reduction is not conclusive proof of the reproductive assurance hypothesis (see Materials and Methods) because emasculation may make flowers less attractive to pollinators.
Although the long-term effects of inbreeding are not addressed in this study, we nevertheless show that selfing, in particular facilitated selfing, is necessary for flowers to reach their full seed set. Facilitated selfing in this system most likely acts as a reproductive assurance mechanism that guarantees seed production in the likely event of pollination failure by bees. These data are among the first to show that facilitated selfing may be an important reproductive assurance mechanism in some systems, although this mode of reproductive assurance may be rare in other systems (Lloyd, 1992
). This study suggests that the extremely species-specific mutualism between Pameridea and Roridula may extend beyond simple nutritional benefits.
|
FOOTNOTES |
|---|
4 Author for reprint requests (banderso{at}botzoo.uct.ac.za)
.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Baker H. G. 1955 Self-compatibility and establishment after "long distance" dispersal. Evolution 21: 853-856[CrossRef]
Baker J. D. R. W. Cruden 1991 Thrips-mediated self-pollination of two facultatively xenogamous wetland species. American Journal of Botany 78: 959-963[CrossRef][ISI]
Barrett S. C. H. J. R. Kohn 1991 Genetic and evolutionary consequences of small population size in plants: implications for conservation. In D. A. Falk and K. E. Holsinger [eds.], Genetics and conservaiton of rare plants, 283. Oxford University Press, Nw York, New York, USA .
Bernhardt P. 1976 The pollination ecology of Hepatica acutiloba DC. (Ranunculaceae). Bulletin of the Torrey Botanical Club 103: 155-158
Bond W. J. 1994 Do mutualisms matter? Assessing the impact of pollinator and disperser disruption on plant extinction. Philosophical Transactions of the Royal Society of London B 344: 83-90[CrossRef]
Carlquist S. 1976 Wood anatomy of Roridulaceae: ecological and phylogenetic implications. American Journal of Botany 63: 1003-1008[CrossRef][ISI]
Cheliak W. M. I. A. Pitel 1984 Techniques for starch gel electrophoresis of enzymes from forest tree species. Information report PI-X-42, Petawawa National Forestry Institute, Canadian Forestry Service, Agriculture Canada, Toronto, Ontario, Canada
Cowling R. M. P. M. Holmes A. G. Rebelo 1992 Plant diversity and endemism. In R. M. Cowling [ed.], The ecology of fynbos, nutrients, fire and diversity, 62–112. Oxford University Press, Cape Town, South Africa
Cruden R. W. D. L. Lyon 1989 Facultative xenogamy: examination of a mixed mating system. In J. H. Bock and Y. B. Linhart [eds.], The evolutionary ecology of plants, 171–208. Westview Press, Boulder, Colorado, USA
Darwin C. 1876 The effects of cross and self-fertilization in the vegetable kingdom. John Murray, London, UK
Dolling W. R. J. M. Palmer 1991 Pameridea (Hemiptera: Miridae): predaceous bugs specific to highly viscid plant genus Roridula. Systematic Entomology 16: 319-328[ISI]
Eckert C. G. A. Schaefer 1998 Does self-pollination provide reproductive assurance in Aquilegia canadensis (Ranunculaceae)?. American Journal of Botany 85: 919-924[Abstract]
Ellis A. G. J. J. Midgley 1996 A new plant–animal mutualism involving a plant with sticky leaves and a resident hemipteran. Oecologia 106: 478-481[CrossRef][ISI]
Fisher R. A. 1941 Average excess and average effect of a gene substitution. Annals of Eugenics 11: 53-63[ISI]
Hepburn H. R. A. Jacot-Guillarmod 1991 The Cape honeybee and the fynbos biome. South African Journal of Science 87: 70-73[ISI]
Herlihy C. R. G. E. Eckert 2002 Genetic cost of reproductive assurance in a self-fertilizing plant. Nature 416: 320-323[CrossRef][Medline]
Holsinger K. E. 1996 Pollination biology and the evolution of mating systems in flowering plants. Evolutionary Biology 29: 107-149[ISI]
Jain S. K. 1976 The evolution of inbreeding in plants. Annual Review of Ecology and Systematics 7: 69-95
Johnson S. D. 1992 Buzz pollination of Orpheum frutescence. Veld and Flora 78: 36-37
Klips R. A. A. A. Snow 1997 Delayed autonomous self-pollination in Hibiscus laevis (Malvaceae). American Journal of Botany 84: 48-53[Abstract]
Leclerc-Potvin C. K. Ritland 1994 Modes of self-fertilization in Mimulus guttatus (Scrophulariaceae): a field experiment. American Journal of Botany 81: 199-205[CrossRef][ISI]
le Maitre D. C. J. J. Midgley 1992 Plant reproductive ecology. In R. Cowling [ed.], The ecology of fynbos, nutrients, fire and diversity, 135–174. Oxford University Press, Cape Town, South Africa
Lloyd D. G. 1979 Some reproductive factors affecting the selection of self-fertilization in plants. American Naturalist 113: 67-69[CrossRef][ISI]
Lloyd D. G. 1980 Demographic factors and mating patterns in angiosperms. In O. T. Solbrig [ed.], Demography and evolution in plant populations. Blackwell, Oxford, UK
Lloyd D. G. 1992 Self- and cross-fertilization in plants. II. The selection of self-fertilization. International Journal of Plant Sciences 153: 370-380[CrossRef]
Lloyd D. G. D. J. Schoen 1992 Self- and cross-fertilization in plants. I. Functional dimensions. International Journal of Plant Sciences 153: 358-369[CrossRef]
Lloyd F. E. 1934 Is Roridula a carnivorous plant?. Canadian Journal of Research 10: 780-786
Markert C. L. I. Faulhaber 1965 Lactate dehydrogenase isozyme patterns of fish. Journal of Experimental Zoology 159: 319-332
Marloth R. 1903 Some recent observations on the biology of Roridula. Annals of Botany 17: 151-157
Marloth R. 1925 The flora of South Africa. Vol. 2. Darter Brothers, Cape Town, South Africa
Motten A. F. 1982 Autogamy and competition for pollinators in Hepatica americana (Ranunculaceae). American Journal of Botany 69: 1296-1305[CrossRef][ISI]
Obermeyer A. A. 1970 Roridulaceae. In L. E. Codd, B. DeWinter, D. J. B. Killick, and H. B. Rycroft [eds.], Flora of Southern Africa 13. Kirstenbosch, 201–204, Darter Brothers, Cape Town, South Africa
Piper J. G. B. Charlesworth D. Charlesworth 1986 Breeding system evolution in Primula vulgaris and the role of reproductive assurance. Heredity 56: 207-217[ISI]
Proctor M. P. Yeo A. Lack 1996 The natural history of pollination. Harper Collins, London, UK
Raymond M. F. Rousset 1995 GENEPOP. Versions 1, 2. A population genetics software for exact tests and eucumenism. Journal of Heredity 86: 248-249
Rice W. R. 1989 Analysing tables of statistical tests. Evolution 43: 223-225[CrossRef][ISI]
Richards A. J. 1986 Plant breeding systems. George Allan & Unwin, London, UK
Ridgeway G. J. S. W. Sherburne R. D. Lewis 1970 Polymorphism in the esterase of Atlantic herring. Transactions of the American Fisheries Society 99: 147-151[CrossRef][ISI]
Schoen D. J. A. D. H. Brown 1991 Whole- and part-flower self-pollination in Glycine clandestine and G. argyrea and the evolution of autogamy. Evolution 45: 1665-1674[CrossRef][ISI]
Schoen D. J. D. G. Lloyd 1992 Self-and cross-fertilization in plants. III. Methods for studying modes and functional aspects of self-fertilization. International Journal of Plant Sciences 153: 381-393[CrossRef][ISI]
Schoen D. J. M. T. Morgan T. M. Bataillon 1996 How does self-pollination evolve? Inferences from floral ecology and molecular genetic variation. Philosophical Transactions of the Royal Society of London B 351: 1281-1290[CrossRef]
Stuber C. W. M. M. Goodman F. M. Johnson 1977 Genetic control and racial variation of ß-glucosidase isozymes in Maise (Zea mays L. ). Biochemical Genetics 15: 383-394[CrossRef][ISI][Medline]
Vogel S. 1978 Evolutionary shifts from reward to deception in pollen flowers. In A. J. Richards [ed.], The pollination of flowers by insects, 213. Academic Press London, UK
Zamora R. 1999 Conditional outcomes of interactions: the pollinator–prey conflict of an insectivorous plant. Ecology 80: 786-795[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
