| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
Department of Biology, Queen's University, Kingston, Ontario, Canada K7L 3N6
Received for publication January 12, 2001. Accepted for publication June 7, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1 d in each of male and female phases with an intervening 1-d neuter phase during which there is no available pollen in anthers and stigmas are not yet exposed to receive pollen. Morphological criteria used to delimit the beginning and end of each of these three sex phases were validated by quantifying the temporal schedule of pollen removal from anthers and pollen deposition on stigmas. Experimental pollinations showed that the morphological changes marking the end of female phase are hastened by pollen deposition. At the umbel level, synchronous development within sequential cohorts of flowers reduced overlap of male and female sexual phases between flowers. On average (±1 SE), 72 ± 3% of flowers completed their female phase while no other flowers on the same umbel were in male phase. Computer simulations of umbel development showed that this value is significantly higher than expected if the timing of flower development within umbels was random (30 ± 1%). Surveys of floral sex ratios in three populations revealed that 87% of umbels were either unisexual male or female at any given time. Pollinators usually visited more than one flower in sequence when foraging on umbels, suggesting that synchronous dichogamy may be an adaptation to avoid geitonogamy. The adaptiveness of both flower- and umbel-level dichogamy is also suggested because both traits are expressed to a lesser extent in obligately clonal, triploid populations, where flowers do not make seeds and hence floral adaptations are not maintained by natural selection.
Key Words: Butomaceae • Butomus umbellatus • dichogamy • floral longevity • flowering rush • pollen deposition • pollen removal • pollen tubes • pollination • pollinator visitation • temporal dioecism
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). However, relatively little is known about how dichogamy functions in natural populations. Dichogamy is usually defined morphologically in terms of the timing and duration of anther dehiscence and stigma receptivity. The functioning of dichogamy, however, depends upon the schedules of pollen removal and deposition (Lloyd and Webb, 1986
). For instance, a flower's female phase can be defined, nominally, by the initiation and duration of stigma receptivity, whereas the effective or functional female phase begins when stigmas become receptive but ends when they have received enough pollen to fertilize all ovules. Similarly, the nominal male phase begins with anther dehiscence and ends when pollen is no longer viable, but the effective male phase ends when pollen has been completely removed from anthers by pollinators (Lloyd and Webb, 1986
). Most of the information on dichogamy in natural populations is based on rough estimates of the degree of overlap between nominal sexual phases, usually defined by morphological criteria (Lloyd and Webb, 1986
; Bertin and Newman, 1993
). Very few studies have defined dichogamy in relation to the overlap between the effective sexual phases under natural pollination (e.g., Jonsson, Rosquist, and Widén, 1991
; Wells and Lloyd, 1991
; Robertson and Lloyd, 1993
; Bell and Cresswell, 1998
; Griffin, Mavraganis, and Eckert, 2000
).
Dichogamy is thought to have evolved to reduce interference between the two principal functions of a hermaphroditic flower: pollen export and pollen import (Wyatt, 1983
; Lloyd and Webb, 1986
; Bertin and Newman, 1993
). Self-pollination is a potentially important form of interference that may reduce female fitness by reducing the fertilization success of incoming outcross pollen and thus the number and/or quality of offspring produced and reduce male fitness by reducing the export of pollen to other individuals in the population (pollen discounting; Harder and Wilson, 1998
). There may also be mechanical interference between male and female sex organs if they are at the same position in a flower at the same time (Lloyd and Webb, 1986
). However, plants are highly hierarchical in structure, and individual dichogamous flowers are usually borne on multiflowered inflorescences. As a result, the adaptive significance of dichogamy, like many other aspects of floral morphology, is strongly influenced by how it functions in the context of the whole inflorescence (Harder and Barrett, 1996
). For instance, dichogamy may reduce autogamous self-pollination but will not, by itself, prevent geitonogamous selfing between flowers on the same inflorescence. However, dichogamy combined with some form of coordinated flower development may have a substantial effect on geitonogamy.
Geitonogamy may be reduced if the dichogamous flowers on an inflorescence develop so that all open flowers are in either male or female phase at any given time (Stout, 1928
). Lloyd and Webb (1986)
refer to this as "synchronous" dichogamy, and Cruden and Herman-Parker (1977)
consider this a form of "temporal dioecism." Cruden (1988)
reviews many examples of highly synchronous dichogamy that result in whole ramets passing through sequential phases of being purely male or purely female (e.g., Thomson and Barrett, 1981
; McDade, 1986
; Schlessman, Lloyd, and Lowry, 1990
). Although this phenomenon appears to have evolved repeatedly in diverse groups, it has not been studied extensively and is probably more common than currently appreciated (Cruden, 1988
; Schlessman, Lloyd, and Lowry, 1990
). Less pronounced synchrony may also be very common (Lloyd and Webb, 1986
) and quite effective at reducing geitonogamous selfing (Webb and Bawa, 1983
), but is liable to go unnoticed without careful observation.
Estimating the degree to which synchronous floral development combined with dichogamy reduces overlap between female and male sexual phases among flowers requires that the functional sexual phases of individual flowers be accurately defined. Very few studies have combined an examination of dichogamy at the inflorescence level with a functional analysis of dichogamy within individual flowers (e.g., Aizen and Basilio, 1995
). In cases where the temporal segregation of male and female sexual phases among flowers is not complete, it is also important to have some kind of null model of inflorescence development with which to compare empirical observations. This approach has not previously been used.
The adaptive significance of synchronous dichogamy depends on whether the foraging activities of pollinators causes geitonogamy (Snow et al., 1996
). If pollinators usually visit more than one flower in sequence on an inflorescence, then temporal segregation of sexual phases may be interpreted as a mechanism to reduce geitonogamy.
In this study, we investigate dichogamy at both the flower and inflorescence levels in flowering rush Butomus umbellatus (Butomaceae). Specifically, we describe dichogamy by defining male and female sexual phases morphologically, and then determine how well these morphological criteria reflect the amount of overlap between sexual phases based on the temporal pattern of pollen removal and receipt throughout floral development. We then combine detailed observations of umbel development and population surveys of floral sex ratios to quantify the extent to which dichogamy is coordinated among flowers within inflorescences. We compare our observations of umbel development to a null model generated by computer simulation to determine whether the sex expression of flowers in an umbel is more synchronous than expected by chance. Finally, we describe how insect pollinators forage on flowers and inflorescences to determine the opportunity for geitonogamy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
100 yr ago and has since spread across the continent (White, Haber, and Keddy, 1993
). Plants combine sexual reproduction with clonal reproduction via rhizome fragmentation and vegetative bulbils borne on both rhizomes and inflorescences (Eckert, Massonnet, and Thomas, 2000
). An individual plant consists of a monopodial rhizome that bears thin, linear, upright leaves in a distichous arrangement. Axillary meristems on the rhizome can develop into rhizome branches, bulbils, and inflorescences (Wilder, 1974
; Lieu, 1979
). Inflorescences are cymose umbels borne on thin, cylindrical, upright stalks and consist of 15–50 light pink flowers (Wilder, 1974
). Each umbel is subtended by three bracts, and a branching system of floral meristems is produced from the axil of each bract; hence umbels consist of three independent meristem lineages (Wilder, 1974
). Flowers consist of three pink sepals, three slightly larger pink petals, nine stamens (in an outer whorl of six and an inner whorl of three) and six conduplicate carpels, each of which contains 200 ovules (Singh and Sattler, 1974
; Kaul, 1976
; Lieu, 1979
). In introduced North American populations, flowers typically produce 200 seeds per fruit (Eckert, Massonnet, and Thomas, 2000
). The flowers of B. umbellatus were first described as dichogamous by Pohl (1935
; see also Faegri and van der Pijl, 1979
). Abundant nectar is secreted from nectaries located at the base of the carpels and, judging from observations of plants flowering under greenhouse conditions, appears to be produced throughout floral development (C. G. Eckert, personal observation). We studied three populations of B. umbellatus located within a 20-km radius of Kingston, in eastern Ontario, Canada. We used the largest population, located at Glenburnie (ONGL: 44°34'30.2'' N, 76°25'24.6'' W), for most of our observations. Supplementary observations were made at populations located near Gananoque (ONMA: 44°18'52.8'' N, 76°15'10.6'' W) and Bath (ONBP: 44°11'3.0'' N, 76°46'29.1'' W). All three populations were located in shallow water habitat typical of B. umbellatus in North America.
Flower and umbel development
To determine the prevalence of within-flower dichogamy, we recorded the development of all flowers on a haphazard sample of 35 umbels in population ONGL during June and July 1999. We visited each umbel every day, usually between 0900 and 1200, and recorded the sex expression of every flower, specifically the number of anthers dehisced, whether there was any bright orange pollen on the anthers, the extent to which stigma surfaces were exposed, the relative amount of pollen deposited on stigma surfaces, the angle between the petals and sepals and the floral axis (90° = petals fully open and perpendicular to axis; 0° = petals loosely enclosing the gynoecium; hereafter petals and sepals will be referred to collectively as petals), and the color of the petals (pink–brown). We observed that flowers appear to go through three distinct sexual phases: male phase (fresh pollen in dehisced anthers), female phase (stigma surfaces exposed and stigmatic papillae turgid and fresh-looking), and an intervening neuter phase (no pollen visible on anthers, stigma surfaces not yet exposed). The three sex phases are described in detail in the results and Table 1.
|
40 h.
Survey of floral sex ratios in natural populations
We conducted extensive observations on the sex expression of flowers in three populations to better estimate the prevalence of dichogamy within flowers as well as the relative duration of the three sexual phases. The relative duration of each sex phase is indicated by the proportion of flowers in each sex phase at any given time. We scored the number of male-, neuter-, and female-phase flowers for haphazard samples of 63–100 umbels (mean = 84) collected on each of 10 d in population ONGL, 3 d in ONBP, and 2 d in ONMA (total = 15 samples, 1260 umbels, 6955 flowers).
Estimating the functional duration of male phase
Male phase appears to be followed by neuter phase during which there is no visible pollen on anthers and stigma surfaces are not yet exposed. We confirmed that neuter phase, defined morphologically (Table 1), marked the end of the functional male phase by determining whether the removal of pollen from anthers was complete by neuter phase. We haphazardly sampled ten buds just before anthesis, ten neuter-phase flowers, and ten flowers that appeared to have just completed female phase on each of 10 d in ONGL, 3 d in ONBP, and 3 d in ONMA (16 d, 48 d x phase samples, 480 flowers). From each sample of mature buds, we took three subsamples each containing one anther from the outer whorl from each bud. We also took three subsamples, each containing one anther from the inner whorl from each bud. This yielded six subsamples of ten anthers each for each sample. Each subsample was ground in 300 µL of 3 : 1 glycerol : lactic acid in a 1.5-mL microcentrifuge tube using a tissue-grinding pestle. For each sample of neuter-phase or post-female-phase flowers, we took two subsamples containing all the anthers from five flowers per sample. This yielded two subsamples of 45 anthers each for each sample of each floral stage. Each subsample was ground in 600 µL of glycerol : lactic acid as above. The number of pollen grains in each subsample was estimated using a haemocytometer, with eight replicate counts per subsample. We averaged the replicate counts and divided by the number of anthers sampled to calculate the number of pollen grains per anther for each subsample. Subsample estimates were then averaged to yield one estimate of pollen per anther for each of the three floral stages for each of the 16 d.
Estimating the functional duration of female phase
Functionally, female phase begins when receptive stigma surfaces are exposed and start to collect pollen and ends when enough pollen has been deposited on stigmas to ensure the fertilization of all the ovules that will eventually develop into seeds. Morphological observations suggested that female phase begins when the stigma surfaces are exposed and ends when the stigmas appear covered in pollen, stigmatic papillae are wilted and brown, and petals begin to turn brown and move to loosely enclose the gynoecium (Table 1). We evaluated whether these morphological criteria accurately reflect the functional duration of female phase by quantifying pollen deposition on stigmas throughout floral development. Because pollen is very difficult to count accurately on stigmas of B. umbellatus, we counted pollen tubes in the upper portion of the style, using fluorescence microscopy following Kearns and Inouye (1993
, pp. 124–129; see also Eckert and Allen, 1997
). There is expected to be some lag time between the deposition of a pollen grain and the subsequent development of a pollen tube long enough to penetrate the style. Consequently, data from pollen tubes will tend to slightly underestimate the beginning of female phase and slightly overestimate the end of female phase.
We subdivided female phase into three sequential subphases based on the extent to which stigma surfaces were exposed (F1, F2, and F3), and subdivided the period after the end of female phase (defined morphologically) but before the onset of fruit development (i.e., swelling of the carpels) into three sequential subphases based on the degree to which the petals had turned brown and begun to wilt (P1, P2, and P3). On five different days, we collected ten flowers of each female and post-female subphase as well as ten neuter phase flowers from ONGL and counted the number of pollen tubes in the style for one carpel from each flower.
We evaluated variation in pollen tube number across developmental stages using a two-way, mixed-model ANOVA with stage as a fixed effect and date as a random effect. We used Tukey-Kramer tests to perform post-hoc multiple comparisons (Sokal and Rohlf, 1995
, pp. 251–252). However, the residuals from the analysis were not normally distributed and correlated positively with predicted values. Accordingly, the data were power transformed (Y' = Y
) to meet ANOVA assumptions using the Box-Cox method with = 0.2 (Neter, Wasserman, and Kutner, 1990
, pp. 149–150).
Effect of pollination on the morphological cues used to delimit female phase
We investigated whether changes in petal angle and color, the morphological cues that appear to mark the end of female phase (Table 1), are hastened by pollen deposition by hand-pollinating flowers at different times after the onset of stigma receptivity in a pollinator-free greenhouse environment and observing changes in floral morphology. We haphazardly collected ten plants from throughout ONGL before inflorescences were initiated in early June 1999, transplanted them in soilless medium (Sunshine Mix #3; Sungro Horticulture, Montreal, Quebec, Canada) within 15.25-cm pots, and grew them to flowering in random positions on a single greenhouse bench under ambient light and temperatures of 20°–30°C.
The ten plants produced inflorescences more or less simultaneously, and we pollinated two sets of five flowers on the first umbel produced by each plant. One set included the first flowers to open on the umbel (the "early" set). The other set included flowers that opened later in the development of the umbel ("late" set). Within each set, one flower was allocated to be pollinated at 0, 12, 24, 36, or 48 h after presentation of the stigma lobes. In all, 99 flowers were pollinated (the 48-h flower in the late set on one of the plants was not pollinated) and all flowers set abundant seed, indicating that female phase can last for at least 48 h.
Each flower was observed every 12 h after the exposure of the stigma lobes and we scored the extent to which the petals had closed around the flower by measuring the petal angle, between 90° (flower fully open with petals perpendicular to floral axis) and 0° (petals closed around the gynoecium and parallel to floral axis). For 80% of the flowers pollinated, petals were fully open (90°) at the beginning of stigma receptivity. We also recorded the time when the petals on each flower started to turn brown. We determined whether pollination hastened these floral changes using a two-way repeated-measures ANOVA with both time of pollination (0, 12, 24, 38, or 48 h after the onset of female phase) and flower set (early vs. late) as fixed, within-subject effects (Neter, Wasserman, and Kutner, 1990
, pp. 1049–1057).
Coordination of dichogamy at the umbel level
For each of the flowers on the 35 umbels followed through development at ONGL, we scored whether any other flowers on the same umbel were in male phase (defined morphologically as in Table 1) during the duration of its female phase (also defined morphologically). For each umbel, we then calculated the percentage of flowers experiencing no overlap between their female phase and the male phase of other flowers (PNO). We compared these PNO values to null expectations generated by computer simulation. We simulated umbels with the same number of flowers, the same duration of the three sexual phases, and the same total duration of flowering as the umbels sampled at ONGL (see results below), but the particular day that each flower entered male phase was chosen at random. For each set of umbel-specific parameter values, we repeated the simulation 1000 times, calculated the mean PNO across simulations, and tallied the number of times the simulated PNO was equal to or greater than the observed PNO. An observed value of PNO was considered significantly different from random if it was greater than at least 95% of the simulated values.
We also scored whether the umbels sampled during our survey of floral sex ratios in ONBP, ONGL, and ONMA were unisexual (i.e., bearing flowers of just male phase or just female phase; neuter phase flowers were not considered) vs. bisexual (bearing flowers of both male and female sexual phases). Of the 1260 umbels surveyed in the 15 population samples, 92 possessed only a single open flower and were, therefore, excluded from the analysis below because they could not possibly exhibit bisexuality.
Pollinator–umbel interactions
We examined how insect visitors interacted with umbels to determine if there is opportunity for geitonogamous self-pollination among flowers within umbels. During 9 d spanning an 18-d period during peak flowering, we observed 276 interactions between flowering umbels and insect visitors at ONGL and recorded for each the number of male-, neuter- and female-phase flowers open on each umbel as well as the number of flowers of each sex phase visited. All observations were conducted in fair weather between 0930 and 1310.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Individual flowers of B. umbellatus appear to pass through three distinct sexual phases based on the following morphological criteria (Table 1).
In the first phase, the male phase, as a bud develops, the subtending pedicel elongates and the six petals open to reveal nine large, bright orange, undehisced stamens in an outer whorl of six and an inner whorl of three. Shortly after the petals open, anthers begin to dehisce. Anthers in the outer whorl almost always dehisce before those in the inner whorl. The bright orange pollen grains are conspicuous, so that fresh pollen is clearly visible in dehisced anthers or on stigmas. Pollen remaining in anthers or on stigmas fades to white a couple of days after it is shed from the anthers. During this phase, carpels are small and light pink, and stigmas appear undeveloped and tightly closed. Petals are turgid, pink, and held at 90° to the floral axis. Any flower with at least one freshly dehisced anther plus at least one undehisced anther or with all anthers dehisced but orange pollen obviously remaining in the anthers was classified as in male phase.
In the second phase, the neuter phase, the anthers have dehisced and the filaments of the stamens bend away from the floral axis to position the shriveled anthers against the petals. During this process, there is no pollen visible in the anthers. The carpels swell and turn a dark burgundy color, but the stigmatic surfaces remain closed. Petals are turgid, pink, and held at 90° to the floral axis. Any flower without visible pollen in anthers and without visible stigma surface was classified as in neuter phase.
In the third phase, the female phase, the stigma surfaces are gradually revealed as the distal edges of the ovary wall curl outward. Stigmatic papillae are translucent and readily visible with the naked eye. Conspicuous orange pollen usually rapidly accumulates on opening stigmas. Petals are usually turgid, pink, and held at 90° to the floral axis. Any flower with exposed stigmas and fresh-looking, translucent stigmatic papillae was classified as in the female phase. Female phase appears to end when the stigmatic papillae turn brownish and are covered with a thick layer of faded pollen. Petals have begun to turn brown at this point and eventually wilt. The angle between the petals and the floral axis is greatly reduced as petals move to cover the gynoecium. Carpels turn dark burgundy–black and become markedly swollen. Eventually the developing fruit expands beyond the enclosing petals, and mature seed are released when carpels split down their inner seams (which were never fully fused). None of the 987 flowers we observed at ONGL possessed undehisced anthers or appeared to have pollen left in dehisced anthers while stigmas were exposed.
Based on means for the 35 umbels, flowers spent an average (±1 SD) of 1.11 ± 0.18 d in male phase, 0.61 ± 0.25 d in neuter phase, and 1.13 ± 0.13 d in female phase. Overall floral lifespan was 2.85 ± 0.25 d. The number of days an umbel flowered correlated positively with the number of flowers produced (r = 0.68, P < 0.00001), although the average floral lifespan correlated negatively with both flower number (r = –0.54, P = 0.0007) and days flowering (r = –0.36, P = 0.033).
Schedule of anther dehiscence within flowers
At ONGL, the period of anther dehiscence appears to last for fewer than 12 h after the first anther dehisces (Fig. 1). On 12 July, 27 of 30 flowers had dehisced all anthers by dusk. The remaining three flowers dehisced 7–8 of 9 anthers by dusk and the remaining anthers early the following morning. No anthers dehisced between dusk and dawn. On 28 July, all 30 flowers had completed anther dehiscence by dusk. The average duration of anther dehiscence did not differ between days, either when the three laggard flowers were included (12 July: mean ± 1 SE = 8.22 ± 0.98 h; 28 July: 6.37 ± 0.35 h; t = 1.77, df = 58, P = 0.082) or excluded (12 July: 6.54 ± 0.33 h; t = 0.35, df = 58, P = 0.73). Of the 60 flowers observed, 16.7% appeared to stay in neuter phase for the following 15–19 h, 53.3% for 20–29 h, 13.3% for 30–40 h, and 16.7% for 40 h.
|
Survey of floral sex ratios in natural populations
Surveys of floral sex ratios at ONBP, ONGL, and ONMA also suggested that individual flowers are strictly protandrous. None of the 6955 flowers scored on 1260 umbels in 15 population samples had undehisced anthers or appeared to have pollen left in dehisced anthers while stigmas were exposed. Even floral sex ratios suggest that the duration of the three sex phases was about equal. Based on the means from each survey, 32.2 ± 1.6% (mean ± 1 SE) of flowers were in male phase (range = 20.2–42.2%), 34.3 ± 1.3% were in neuter phase (range = 24.1–41.3%) and 33.5 ± 1.0% were in female phase (range = 29.0–43.5%).
Functional duration of male phase
The temporal pattern of pollen removal confirms that pollen dispersal does not occur during neuter phase, as flowers do not appear to disperse additional pollen after the end of male phase as defined in Table 1. Averaging across ten samples from ONGL, there was a very large difference in the number of pollen grains per anther between flowers before anther dehiscence (mean ± 1 SE = 31 570 ± 1239 pollen grains/anther) and flowers after anther dehiscence (neuter and post-female flowers: 609 ± 46 pollen grains/anther; t = 35.9, df = 28, P < 0.0001), but no difference between neuter-phase flowers (585 ± 56 pollen grains/anther) and post-female-phase flowers (634 ± 76 pollen grains/anther; t = 0.5, df = 18, P = 0.61; Fig. 2). On average, 98.1% of pollen was removed from anthers by neuter phase, after which no more was dispersed. We found a similar pattern at ONBP and ONMA, even though a lower proportion of pollen was removed from anthers. Of the 37 150 ± 2351 grains in anthers at the onset of male phase, only 7% remained by neuter phase (2597 ± 643 pollen grains/anther), and no further pollen was removed by the end of female phase (2978 ± 1100 pollen grains/anther; t = 0.3, df = 10, P = 0.77).
|
), hence there appear to be enough pollen tubes in styles by the end of the morphologically defined female phase to ensure full seed set.
|
|
|
The PNO varied significantly among the umbels sampled (2 x 35 contingency table 
2 = 140.5, df = 34, P < 0.0001) and generally increased with the extent to which the development of flowers in sequential cohorts was synchronized. This is illustrated in Fig. 5. For umbel 51 (PNO = 100%), the schedule of flower opening was very clumped, with five distinct synchronized cohorts. In contrast, the temporal pattern of flower opening was hardly clumped at all for umbel 3 (PNO = 44%). The PNO of each umbel also correlated positively with the average duration of neuter phase in individual flowers (Fig. 6; r = 0.53, P = 0.0010), but not the average duration of male phase (r = –0.01, P = 0.94) or female phase (r = –0.15, P = 0.40). The average duration of neuter phase varied much more among umbels (CV = 40.4%) than the duration of either male phase (16.0%) or female phase (11.5%). The PNO did not correlate with other umbel-level attributes such as flowers/umbel (r = –0.28, P = 0.10), number of days flowering (r = –0.26, P = 0.12) or average date of flower opening (r = –0.04, P = 0.81).
|
|
2 = 0.76, df = 2, P = 0.68) or among the ten samples from ONGL (range = 79.8–93.6%; 2 = 12.49, df = 9, P = 0.19). Whether an umbel was unisexual or bisexual was not related to the number of flowers open (unisexual umbels: mean ± 1 SE = 5.89 ± 0.21 flowers; bisexual umbels: 6.17 ± 0.30 flowers; paired t = 0.90, df = 14, P = 0.38).
Pollinator–umbel interactions
During 9 d of observation at ONGL, we observed 276 pollinator foraging bouts: 89.7% were made by honey bees (Apis mellifera), 7.6% by flies, 1.4% by bumble bees (Bombus spp.), and 0.7% by wasps (mostly Vespa spp.). More than one flower was visited in sequence during 55.1% of foraging bouts. Although the number of flowers visited increased significantly with the number open (r = 0.29, P < 0.0001), the increase was weak (Fig. 7) so that the proportion of flowers visited correlated negatively with the number open (r = –0.45, P < 0.0001). The number or proportion of flowers visited in sequence by a pollinator did not differ between pollen-bearing umbels (i.e., those with at least one male-phase flower) and umbels with only neuter and/or female-phase flowers (Table 3A). However, during foraging bouts on umbels with both male- and female-phase flowers, visitors were twice as likely to visit male- than female-phase flowers (Table 3B).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1 d (Table 1). Most anthers in a flower dehisce within the first day of opening (Fig. 1), and pollen is promptly removed by insects. The second day of floral life appears to be a neuter phase; pollen removal is complete (Fig. 2) but stigmas remain closed and receive little if any pollen (Fig. 3). Because the neuter phase begins when pollen removal is complete, its duration may vary strongly with the level of pollinator visitation. Neuter phases have been recorded in other dichogamous taxa, although it is not clear whether this is a common phenomenon (e.g., Aizen and Basilio, 1995
; Snow and Grove, 1995
). Female phase occurs on the third day after anthesis, during which stigmas become increasingly exposed and rapidly accumulate pollen (Fig. 3). At ONGL, we detected substantial variation among umbels in the average duration of the three phases, especially neuter phase (Fig. 6), but it is not clear whether this variation was caused by variation in the rate of pollen removal and deposition or variation in the inherent rate of floral development. Despite this variation, we never observed flowers that possessed undehisced anthers or appeared to have pollen left in dehisced anthers when stigmas were exposed.
The duration of sexual phases in dichogamous species and the degree of overlap between them are usually evaluated by morphological criteria such as the appearance of the anthers, stigmas, and other floral organs (reviewed in Lloyd and Webb, 1986
; Bertin and Newman, 1993
; Robertson and Lloyd, 1993
). In many cases, these observations may provide little information on the effective duration of sexual phases and the functional degree of dichogamy (e.g., Preston, 1991
; Robertson and Lloyd, 1993
; Griffin, Mavraganis, and Eckert, 2000
). However, our results suggest that morphological characteristics (Table 1) reliably delimit effective sexual phases in B. umbellatus. Using controlled pollinations under greenhouse conditions, we also showed that changes in petal angle and color, two of the morphological cues used to judge the end of female phase, are hastened by pollination (Fig. 4; Table 2). These results add to a growing body of evidence that the rate of floral development and senescence is influenced by pollen deposition (reviewed by O'Neill, 1997
; van Doorn, 1997
). It is also possible that the onset of female phase may be influenced by the rate at which pollen is removed from anthers (e.g., Bell and Cresswell, 1998
) so that the timing of stigma receptivity in relation to anther dehiscence is responsive to the local pollination environment. Experiments in which pollen is manually removed from anthers at different rates are required to explore this possibility.
Dichogamy at the umbel level and beyond
Strong protandry combined with a nonrandom schedule of flower opening results in synchronous dichogamy at the inflorescence level in B. umbellatus. Among the 35 umbels studied at ONGL, 72% of flowers showed no overlap between their female phase and the male phase of other flowers on the same umbel (PNO = 72%). In addition, our survey of floral sex ratios in three populations showed that 80–94% of umbels had either male-phase or female-phase flowers but not both. The data from 35 umbels at ONGL expressed in the same way yields a percentage of unisexual umbels of 80%. The segregation of the two sexual functions among flowers is promoted by the synchronous development of flowers within multiple cohorts (Fig. 5), so that individual umbels go through multiple cycles of being purely male and then purely female. The number of flowers in each cohort is usually a multiple of three, which probably reflects the fact that each umbel arises from three independent lineages of floral meristems (Wilder, 1974
).
Computer simulation of random umbel development provided a null model with which we could evaluate the significance of our observations. Although almost all of the umbels sampled at ONGL exhibited much higher PNO values than expected by chance alone, some degree of synchronous dichogamy at the inflorescence level may occur without coordinated flower development. In the average simulated umbel, 30% of flowers completed female phase when no other flowers were in male phase. Supplementary simulations (not shown) indicated that, as one might intuitively expect, the PNO of randomly developing umbels correlates positively with the total duration of flowering and negatively with the number of flowers.
We did not investigate the extent to which dichogamy might be coordinated beyond the umbel. Ramets of B. umbellatus usually produce multiple umbels separated by intervals of vegetative growth along a monopodial rhizome (Lieu, 1979
), so it is possible for female and male sexual phases of flowers on different umbels of the same ramet to overlap. However, data from an ongoing greenhouse growth experiment suggest that sufficient time elapses between the development of successive inflorescences on an individual rhizome to minimize or entirely eliminate overlap in flowering between umbels within ramets (F. L. Thompson and C. G. Eckert, unpublished data). Butomus umbellatus also possesses several mechanisms of clonal propagation (Eckert, Massonnet, and Thomas, 2000
), and pollination between flowers on different ramets of the same genet may be a major component of geitonogamy in clonal plants (Handel, 1985
; Eckert, 2000
). Genets of B. umbellatus can spread clonally by fragmentation of rhizome branches and by numerous small vegetative bulbils produced on rhizomes and inflorescences. Because related ramets are not physiologically connected, it seems unlikely that umbel development could be synchronized enough among ramets to result in dichogamy at this level of organization (Aizen and Basilio, 1995
).
These considerations suggest that in B. umbellatus, overlap between male and female sexual phases is minimal among umbels within ramets but probably not among ramets within genets. This is supported by a rudimentary nearest-neighbor analysis we conducted during surveys of floral sex ratios at ONBP, ONGL, and ONMA. For 9 of 15 samples, we located the umbel closest to each focal umbel, recorded the distance between the two, noted whether the neighboring umbel was flowering, pre-flowering, or post-flowering and, if flowering, classified it as unisexual male, unisexual female, or bisexual (total = 610 pairs of umbels). On average, flowering neighbors were located further away (mean ± 1 SE = 16.7 ± 0.8 cm) than nonflowering neighbors (14.1 ± 0.5 cm; t = 3.0, df = 608, P = 0.027). This is consistent with observations from greenhouse-grown plants indicating that sequential umbels on a rhizome rarely overlap flowering times. When both focal and neighboring umbels were in flower (n = 210 pairs), there was no tendency towards synchrony or asynchrony of sexual phases. For instance, when the focal umbel bore female-phase flowers, the neighbor bore male-phase flowers in 52.5% of pairs and only female-phase flowers in 47.4% of pairs (2 = 0.4, df = 1, P = 0.55). These data suggest that, as expected, there is no coordination of sexual phases among coflowering umbels, even at small spatial scales (see also Aizen and Basilio, 1995
). However, this analysis is highly inferential and should be confirmed by studies of genet-level flowering phenology in which ramets can be assigned to genets using genetic markers.
The adaptive significance of dichogamy
Within-flower dichogamy is generally viewed as an adaptive mechanism to reduce interference between pollen import and pollen export (Lloyd and Webb, 1986
; Bertin, 1993
; Bertin and Newman, 1993
). Self-pollination is considered to be a particularly common form of interference because it can reduce both female and male fitness (Lloyd, 1992
), however, physical interference between stamens and pistils may also occur (Lloyd and Webb, 1986
). In the case of B. umbellatus, it is likely that flower-level dichogamy reduces both forms of interference. The species is fully self-compatible (Eckert, Massonnet, and Thomas, 2000
) and thus self-pollination may result in self-fertilization. In addition, dichogamy involves the repositioning of anthers away from the gynoecia before the onset of female phase, possibly suggesting that physical interference would occur if flowers were not dichogamous. One of the main problems in determining the adaptive significance of within-flower dichogamy is that the trait is often difficult or impossible to manipulate experimentally (but see Griffin, Mavraganis, and Eckert, 2000
). Such is the case with B. umbellatus.
The evolutionary interpretation of dichogamy at the inflorescence level is more straightforward. Synchronous dichogamy is almost always interpreted as an adaptation to reduce geitonogamous self-pollination (Cruden and Hermann-Parker, 1977
; Lloyd and Webb, 1986
; Cruden, 1988
; Bertin and Newman, 1993
). Geitonogamy is expected to be common in multiflowered inflorescences (de Jong, Waser, and Klinkhamer, 1993
; Snow et al., 1996
), and all experimental studies to date have shown that it is an important mode of self-fertilization in self-compatible plants (Schoen and Lloyd, 1992
; Leclerc-Potvin and Ritland, 1994
; Eckert, 2000
) and may have negative effects on both male and female reproductive success (de Jong et al., 1992
; Harder and Barrett, 1995
; Hodges, 1995
). However, there are alternative interpretations. For instance, it could be that synchronous dichogamy in B. umbellatus is simply a nonadaptive by-product of umbels being composed of three independent lineages of meristems that develop at more or less the same rate (Wilder, 1974
). At this point, the available evidence for the functional significance of synchronous dichogamy in B. umbellatus is circumstantial, but generally supports the hypothesis that it is adaptive.
The results of this study indicate that insect visitors typically visit more than one flower in sequence when foraging on umbels (Fig. 7), suggesting that there would be opportunities for geitonogamy in B. umbellatus if dichogamy were not synchronized between flowers. However, geitonogamy in bisexual umbels may be partially reduced by the tendency of pollinators to preferentially visit male- over female-phase flowers (Table 3B; see also Aizen and Basilio, 1998
). Further investigation of the adaptive significance of synchronized dichogamy will involve determining whether sequential within-umbel visits by pollinators cause geitonogamy and whether geitonogamy reduces fitness. This will require estimating the extent to which selfing reduces offspring fitness via inbreeding depression. In addition, the consequences of increased overlap between sexual phases among flowers within umbels could be quantified by comparing the level of selfing and outcross siring success of synthetic bisexual umbels (constructed by tying together unisexual male and female umbels) and synthetic unisexual umbels in genetically marked experimental populations.
We have recently obtained additional indirect evidence that synchronous dichogamy is adaptive by comparing the degree of dichogamy at both the flower and umbel levels between populations in which sexual reproduction is potentially common (like the ones studied here) and populations in which seed production is prevented by triploidy. Obligately clonal, triploid populations of B. umbellatus are common in both the native and adventive geographical ranges (Hroudová and Zákravsky, 1993
; Krahulcová and Jarolímová, 1993
; K. Lui, F. L. Thompson, and C. G. Eckert, unpublished data). In these asexual populations, flowers do not make seeds and hence floral traits that may be adaptive in sexual populations no longer function to increase fitness and are expected to degenerate via the accumulation of mutations (see Eckert, Dorken, and Mitchell, 1999
). Surveys of floral sex ratios in large samples of sexual and asexual populations reveal that plants in asexual populations produce substantial frequencies of flowers in which pollen presentation and stigma receptivity overlap as well as umbels consisting of both male- and female-phase flowers (F. L. Thompson and C. G. Eckert, unpublished data). The apparent degeneration of flower- and umbel-level dichogamy in these asexual populations strongly suggests that both forms of dichogamy are actively maintained by selection in sexual populations.
|
FOOTNOTES |
|---|
2 Author for reprint requests (phone: 613-533-6158; FAX: 613-533-6617; eckertc{at}biology.queensu.ca
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Aizen M. A. A. Basilio 1998 Sex differential nectar secretion in protandrous Alstroemeria aurea (Alstroemeriaceae): is production altered by pollen removal and receipt?. American Journal of Botany 85: 245-252[Abstract]
Bell S. A. J. E. Cresswell 1998 The phenology of gender in homogamous flowers: temporal change in the residual sex function of flowers of oil-seed rape (Brassica napus). Functional Ecology 12: 298-306[CrossRef][ISI]
Bertin R. I. 1993 Incidence of monoecy and dichogamy in relation to self-fertilization in angiosperms. American Journal of Botany 80: 557-560[CrossRef][ISI]
Bertin R. I. C. M. Newman 1993 Dichogamy in the angiosperms. Botanical Review 59: 112-152[CrossRef][ISI]
Cruden R. W. 1988 Temporal dioecism: systematic breadth, associated traits, and temporal patterns. Botanical Gazette 149: 1-15[CrossRef]
Cruden R. W. S. M. Hermann-Parker 1977 Temporal dioecism: an alternative to dioecism?. Evolution 31: 863-866[CrossRef][ISI]
de Jong T. J. N. M. Waser P. G. L. Klinkhamer 1993 Geitonogamy: the neglected side of selfing. Trends in Ecology and Evolution 8: 321-325
de Jong T. J. N. M. Waser M. V. Price R. M. Ring 1992 Plant size, geitonogamy and seed set in Ipomopsis aggregata. Oecologia 89: 310-315[ISI]
Eckert C. G. 2000 Contributions of autogamy and geitonogamy to self-fertilization in a mass-flowering, clonal plant. Ecology 82: 532-542
Eckert C. G. M. Allen 1997 Cryptic self-incompatibility in tristylous Decodon verticillatus (Lythraceae). American Journal of Botany 84: 1391-1397[Abstract]
Eckert C. G. M. E. Dorken S. A. Mitchell 1999 Loss of sex in clonal populations of a flowering plant, Decodon verticillatus (Lythraceae). Evolution 53: 1079-1092[CrossRef][ISI]
Eckert C. G. B. Massonnet J. J. Thomas 2000 Variation in sexual and clonal reproduction among introduced populations of flowering rush, Butomus umbellatus (Butomaceae). Canadian Journal of Botany 78: 437-446
Faegri K. L. van der Pijl 1979 The principles of pollination ecology, 3rd ed. Pergamon, New York, New York, USA
Griffin S. R. K. Mavraganis C. G. Eckert 2000 Experimental analysis of protogyny in Aquilegia canadensis (Ranunculaceae). American Journal of Botany 87: 1246-1256
Handel S. N. 1985 The intrusion of clonal growth patterns on plant breeding systems. American Naturalist 125: 367-384[CrossRef][ISI]
Harder L. D. S. C. H. Barrett 1995 Mating cost of large floral displays in hermaphrodite plants. Nature 373: 512-515[CrossRef]
Harder L. D. S. C. H. Barrett 1996 Pollen dispersal and mating patterns in animal pollinated plants. In D. G. Lloyd and S. C. H. Barrett [eds.], Floral biology: studies on floral evolution in animal-pollinated plants, 140–190. Chapman and Hall, New York, New York, USA
Harder L. D. W. G. Wilson 1998 A clarification of pollen discounting and its joint effects with inbreeding depression on mating-system evolution. American Naturalist 152: 684-695[CrossRef][ISI]
Hodges S. S. 1995 The influence of nectar production on hawkmoth behavior, self-pollination and seed production in Mirabilis multiflora (Nyctaginaceae). American Journal of Botany 82: 197-204[CrossRef][ISI]
Hroudová Z. P. Zákravsky 1993 Ecology of two cytotypes of Butomus umbellatus. III. Distribution and habitat differentiation in the Czech and Slovak Republics. Folia Geobotanica et Phytotaxonomica 28: 425-435
Jonsson O. G. Rosquist N. Widén 1991 Operation of dichogamy and herkogamy in five taxa of Pulsatilla. Holarctic Ecology 14: 206-271
Kaul R. B. 1976 Conduplicate and specialized carpels in the Alismatales. American Journal of Botany 63: 175-182[CrossRef][ISI]
Kearns C. A. D. W. Inouye 1993 Techniques for pollination biologists. University Press of Colorado, Niwot, Colorado, USA
Krahulcová A. V. JarolÍmová 1993 Ecology of two cytotypes of Butomus umbellatus. I. Karyology and breeding. Folia Geobotanica et Phytotaxonomica 28: 385-411
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]
Lieu S. M. 1979 Growth forms in the Alismatales. II. Two rhizomatous species: Sagittaria lancifolia and Butomus umbellatus. Canadian Journal of Botany 63: 175-182
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. C. J. Webb 1986 The avoidance of interference between the presentation of pollen and stigmas in angiosperms. I. Dichogamy. New Zealand Journal of Botany 24: 135-162[ISI]
McDade L. A. 1986 Protandry, synchronized flowering and sequential phenotypic unisexuality in neotropical Pentagonia macrophylla (Rubiaceae). Oecologia 68: 218-223[CrossRef][ISI]
Neter J. W. Wasserman M. H. Kutner 1990 Applied linear statistical models, 3rd ed. Richard D. Irwin, Homewood, New Jersey, USA
O'Neill S. D. 1997 Pollination regulation of flower development. Annual Review of Plant Physiology and Plant Molecular Biology 48: 547-574[CrossRef][ISI]
Pohl F. 1935 Ein Fall von zweckloser Proterandrie (Butomus umbellatus). Berichte der Deutschen Botanischen Gesellschaft 53: 779-782
Preston R. E. 1991 The intrafloral phenology of Streptanthus tortuosus (Brassicaceae). American Journal of Botany 78: 1044-1053[CrossRef][ISI]
Robertson A. W. D. G. Lloyd 1993 Rates of pollen deposition and removal in Myosotis colensoi. Functional Ecology 7: 549-559[CrossRef][ISI]
SAS. 1997 JMP user's guide, version 3.2.1. SAS Institute, Cary, North Carolina, USA
Schlessman M. A. D. G. Lloyd P. P. Lowry III 1990 Evolution of sexual systems in New Caledonian Araliaceae. Memoirs of the New York Botanical Garden 55: 105-117
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]
Singh V. R. Sattler 1974 Floral development of Butomus umbellatus. Canadian Journal of Botany 52: 223-230
Snow A. A. K. F. Grove 1995 Protandry, a neuter phase, and unisexual umbels in a hermaphroditic, neotropical vine (Bomarea acutifolia, Alstroemeriaceae). American Journal of Botany 82: 741-744[CrossRef][ISI]
Snow A. A. T. P. Spira R. Simpson R. A. Klips 1996 The ecology of geitonogamous pollination. In D. G. Lloyd and S. C. H. Barrett [eds.], Floral biology: studies on floral evolution in animal-pollinated plants, 191–216. Chapman and Hall, New York, New York, USA
Sokal R. R. F. J. Rohlf 1995 Biometry, 3rd ed. W. H. Freeman and Company, New York, New York, USA
Stout A. B. 1928 Dichogamy in flowering plants. Bulletin of the Torrey Botanical Club 55: 141-153[CrossRef]
Thomson J. D. S. C. H. Barrett 1981 Temporal variation of gender in Aralia hispida vent. (Araliaceae). Evolution 35: 1094-1107[CrossRef][ISI]
van Doorn W. G. 1997 Effects of pollination on floral attraction and longevity. Journal of Experimental Biology 48: 1615-1622
Webb C. J. K. S. Bawa 1983 Pollen dispersal by hummingbirds and butterflies: a comparative study of two lowland plants. Evolution 37: 1258-1270[CrossRef][ISI]
Wells M. S. D. G. Lloyd 1991 Dichogamy, gender variation and bet-hedging in Pseudowintera colorata. Evolutionary Ecology 5: 310-326[CrossRef][ISI]
White D. J. E. Haber C. Keddy 1993 Invasive plants of natural habitats in Canada. Canadian Wildlife Service, Environment Canada, Ottawa, Ontario, Canada
Wilder G. J. 1974 Symmetry and development of Butomus umbellatus (Butomaceae) and Limnocharis flava (Limnocharitaceae). American Journal of Botany 61: 379-394[CrossRef][ISI]
Wyatt R. 1983 Pollinator–plant interactions and the evolution of breeding systems. In L. Real [ed.], Pollination biology, 51–95. Academic Press, Orlando, Florida, USA
