Small Pollen Loads Research Articles

Pollination of the strongly scented Sarcoglottis acaulis (Orchidaceae) by male orchid bees: nectar as resource instead ofperfume.

The Neotropical orchid genus Sarcoglottis comprises ~40 species, which emit strong floral scents, presumably involved in pollinator attraction. Information on basic aspects of its natural history is scant, with the few studies available so far pointing to nectar-seeking orchid bees as pollinators. Here, we investigated the reproductive biology of Sarcoglottis acaulis, addressing the ecological meaning of its floral scent. In Atlantic Forest fragments of NE Brazil, we described the floral biology, determined the breeding system and recorded the pollinators of S. acaulis. Additionally, we chemically characterized its floral scent and assessed its role on pollinator attraction. Although self-compatible, S. acaulis depends on pollinators to boost fruit set. Male orchid bees of Eulaema atleticana and E. niveofasciata were the only recorded pollinators. They foraged for nectar only, in spite of the strong scent emitted by S. acaulis flowers. The floral scent is composed of six compounds, of which geraniol and nerol elicited electroantennographic responses in Eulaema bees. A synthetic mixture of these compounds attracted Eulaema bees in field assays but did not trigger the stereotyped scent-gathering behaviour. The floral scent of S. acaulis acts in signalling and nectar is the sole reward for pollinators. Despite the low pollinator frequency, S. acaulis shows a high fruit set (77%), particularly when compared to other orchids. Attributes such as pollinia that release small pollen loads, allowing a single pollinia-carrying bee to pollinate several flowers, low amount of nectar, steady-state flowering and traplining behaviour of pollinators, might act together to assure this outstanding fruit set.

Nocturnal Bees Feed on Diurnal Leftovers and Pay the Price of Day – Night Lifestyle Transition

1. Bees exemplify flights under bright sunlight. A few species across bee families have evolved nocturnality, displaying remarkable adaptations to overcome limitations of their daylight-suited apposition eyes. Phase inversion to nocturnality in a minority of bees that co-exist with diurnal bees provide a unique opportunity to study ecological benefits that mediate total temporal niche shifts. While floral traits and sensory modalities associated with the evolution of classical nocturnal pollination syndromes, e.g. by bats and moths, are well-studied, nocturnality in bees represent a poorly understood, recently invaded, extreme niche. 2. To test the competitive release hypothesis, we examine how nocturnality shapes foraging by comparing pollen loads, nest pollen and flower visitation of sympatric nocturnal and diurnal carpenter bees. We predicted that nocturnal bees primarily use night-blooming flowers, show little/no resource overlap with diurnal species and competitive release favours night-time pollen collection for provisioning. 3. Contrarily, we found substantial resource overlap between nocturnal and diurnal bees. Flower opening times, floral longevity and plant abundance did not define nocturnal flower use. Smaller pollen loads on nocturnal foragers suggests subsistence on resource leftovers largely from diurnal flowers. Greater pollen types/diversity on nocturnal foragers indicates lower floral constancy compared to diurnal congenerics. Reduced activity during new moon compared to full moon suggests constraints to nocturnal foraging. 4. Invasion and sustenance within the nocturnal niche is characterised by: i)opportunistic foraging on residual resources as indicated by smaller pollen loads, extensive utilisation of day-blooming flowers and substantial overlap with diurnal bees, ii)generalisation at two levels – between and within foraging trips as indicated by lower floral constancy, iii)reduced foraging on darker nights, indicating visual constraints despite sensitive optics. This together with smaller populations and univoltine breeding in nocturnal compared to multivoltine diurnal counterparts, suggest that nocturnality imposes substantial fitness costs. 5. In conclusion, the evolution of nocturnality in bees is accompanied by resource generalisation instead of specialisation. Reduced floral constancy suggests differences in foraging strategies of nocturnal and diurnal bees which merits further investigation. The relative roles of competition, floral rewards and predators should be examined to fully understand the evolution and maintenance of nocturnality in bees.

Adjustment of honey load by honeybee pollen foragers departing from the hive: the effect of pollen load size

Honeybee foragers load a small amount of honey into the crop when they leave the hive. This “honey at departure” is used as a material to build pollen loads (glue honey) as well as fuel during flight in pollen foragers. We investigated the relationship between the size of pollen loads that a forager collected and the amount of honey at departure in the Western honeybee, Apis mellifera. Dancing pollen foragers increased honey at departure with the size of collected pollen loads. Analysis of the waggle-run duration revealed that the size of pollen load affected the rate of increase of honey at departure with food source distance in dancers. The increase rate was significantly higher in dancers that had returned to the hive with large pollen loads versus nectar-collecting dancers, but not in those with smaller pollen loads. The higher increase rate may reflect additional fuel carried for the return trip in pollen-foraging specialists that do not use collected nectar as fuel. Although honey carried by departing dance followers also increased significantly with the size of pollen load carried by dancers, the adjustment was different from that of dancers. These results suggest that recruited bees adjust the amount of honey at departure, including glue honey, based on communicated information and modify the amount of honey load subsequently based on their own experience.

Population size, pollination and phenotypic trait selection in Phyteuma spicatum

Plants in small populations may receive fewer visits, smaller pollen loads or pollen of poorer quality and suffer from reduced reproductive success compared to plants in larger populations. Consequently, pollen limitation of plants in small populations has been suggested to result in the evolution of reduced reliance on pollinators or the enhancement of traits that attract pollinators. The main aim of this study was to experimentally quantify the strength of pollinator-mediated selection on floral display size and flowering phenology in populations of varying size, using the self-incompatible, perennial herb Phyteuma spicatum as study species. We conducted supplementary hand pollinations in six populations (ranging in size between ca. 20–3000 flowering individuals) over two consecutive years and assessed selection gradients (i.e., trait–fitness relationships) in open- and hand-pollinated plants. Our results show that some populations are pollen limited in some years, but, contrary to our expectation, the degree of pollen limitation was not significantly related to population size. We found phenotypic selection for increased inflorescence size (in most populations and in both years), but we obtained no or no strong evidence that selection was pollinator-mediated or that the strength of selection was related to population size. This may have been the result of low statistical power, an inherent problem of studies examining effects of population size that require the inclusion of populations with only few individuals. In addition, given that selection appeared to be spatially and temporally variable, abiotic or biotic factors other than pollinators may have contributed to selection on inflorescence size.

Reloading the revolver - male fitness as a simple explanation for complex reward partitioning in Nasa macrothyrsa (Loasaceae, Cornales)

Reward partitioning and replenishment and specific mechanisms for pollen presentation are all geared towards the maximization of the number of effective pollinator visits to individual flowers. An extreme case of an apparently highly specialized plant‐pollinator interaction with thigmonastic pollen presentation has been described for the morphologically complex tilt-revolver flowers of Caiophora arechavaletae (Loasaceae) pollinated by oligolectic Bicolletes pampeana (Colletidae, Hymenoptera). We studied the floral biology of Nasa macrothyrsa (Loasaceae) in the field and in the glasshouse, which has very similar floral morphology, but is pollinated by polylectic Neoxylocopa bees (Apidae, Hymenoptera). We investigated the presence of thigmonastic anther presentation, visitor behaviour (pollinators and nectar robbers), co-ordination of pollinator visits with flower behaviour and the presence of nectar replenishment. The aim of this study was to understand whether complex flower morphology and behaviour can be explained by a specialized pollination syndrome, or whether alternative explanations can be offered. The results showed that Nasa macrothyrsa has thigmonastic pollen presentation, i.e. new pollen is rapidly (<< 10 min) presented after a pollinator visit. Nectar secretion is independent of removal and averages 7‐14 m Lh ‐1 . The complex flowers, however, fail to exclude either native (hummingbirds) or introduced (honeybees) nectar robbers, nor does polylectic Neoxylocopa actively collect the pollen presented. The findings do not support a causal link between complex flower morphology and functionality in Loasaceae and a highly specialized pollination. Rapid pollen presentation is best explained by the pollen presentation theory: the large proportion of pollinators coming shortly after a previous visit find little nectar and are more likely to move on to a different plant. The rapid presentation of pollen ensures that all these valuable ‘hungry pollinators’ are dusted with small pollen loads, thus increasing the male fitness of the plant by increasing the likelihood of siring outcrossed offspring. © 2010 The Linnean Society of London, Biological Journal of the Linnean Society, 2010, 100, 124‐131.

How robust is nonrandom mating in wild radish: do small pollen loads coupled with more competing pollen donors lead to random mating?

In previous studies of the weedy annual Raphanus sativus we have demonstrated that mating is nonrandom in greenhouse plants, suggesting that sexual selection is possible. To investigate how these greenhouse results might translate to conditions more similar to the field, we manipulated both pollen load size and the number of competing pollen donors on stigmas. While the smallest pollen loads (22 grains per stigma) were small enough to reduce fruit and seed set, seed siring success was unaffected by pollen load size. When the number of competing donors in a mixed pollination was increased to four, the proportion of seeds sired by the pollen donors was the furthest from expectation, suggesting that nonrandom mating increases as the number of donors per pollination increases. There was no significant interaction between pollen load size and number of competitors per pollination. Overall, mating remained nonrandom across all treatments. Thus differential seed paternity is likely to occur in the field as well as in the greenhouse.

A Group-1 Grass Pollen Allergen Influences the Outcome of Pollen Competition in Maize

Worldwide, 400 million people suffer from hay fever and seasonal asthma. The major causative agents of these allergies are pollen specific proteins called the group-1 grass pollen allergens. Although details of their antigenicity have been studied for 40 years with an eye towards immunotherapy, their function in the plant has drawn scant attention. Zea m 1 constitutes a class of abundant grass pollen allergens coded for by several genes that loosen the walls of grass cells, including the maize stigma and style. We have examined the impact of a transposon insertion into one of these genes (EXPB1, the most abundant isoform of Zea m 1) on the production of Zea m 1 protein, pollen viability, and pollen tube growth, both in vitro and in vivo. We also examined the effect of the insertional mutation on the competitive ability of the pollen by experimentally varying the sizes of the pollen load deposited onto stigmas using pollen from heterozygous plants and then screening the progeny for the presence of the transposon using PCR. We found that the insertional mutation reduced the levels of Zea m 1 in maize pollen, but had no effect on pollen viability, in vitro pollen tube growth or the proportion of progeny sired when small pollen loads are deposited onto stigmas. However, when large pollen loads are deposited onto the stigmas, the transposon mutation is vastly underrepresented in the progeny, indicating that this major pollen allergen has a large effect on pollen tube growth rates in vivo, and plays an important role in determining the outcome of the pollen-pollen competition for access to the ovules. We propose that the extraordinary abundance (4% of the extractable protein in maize pollen) of this major pollen allergen is the result of selection for a trait that functions primarily in providing differential access to ovules.

Incidence of geitonogamy differs between two populations in the hawkmoth-pollinated Platanthera bifolia (Orchidaceae)

We estimated geitonogamy in individuals with different inflorescence sizes in a small (100–200 flowering individuals) and a large population (&gt;700 flowering individuals) of the self-compatible, moth-pollinated orchid Platanthera bifolia (L.) L. C. Rich. (Orchidaceae). Geitonogamy was estimated as the percent reduction in pollen receipt by emasculated compared with control plants over seven nights. Geitonogamy in the small population was 23% and 38% during 2 years, respectively. In contrast, no geitonogamy was detected during a single flowering season in the large population. Geitonogamy did not vary with inflorescence size and emasculation had no impact on fruit set. The difference in geitonogamy between the populations in the present study may be related to pollinator abundance and behaviour. We suggest that incidence of geitonogamy will be higher if the pollinator carries smaller pollen loads when arriving at a plant because there will be a smaller fraction of cross-pollen carried after visiting one flower. Geitonogamy may be influenced by available number of mates, pollen load size, pollinator behaviour, and pollen carryover.Key words: geitonogamy, population size, inflorescence size, pollen-limitation, pollen carryover, self-pollination.

Under how wide a set of conditions will nonrandom mating occur in Raphanus sativus (Brassicaceae)?

Studies of the weedy annual Raphanus sativus have demonstrated that nonrandom mating, a prerequisite for sexual selection, can occur in greenhouse plants. To determine whether this nonrandom mating pattern can occur under a wide range of conditions, including conditions that might occur in the field, we considered variation in both maternal condition and pollen load size. Maternal condition was varied by altering the watering regime. Pollen load size was varied from approximately 26 to 343 pollen grains per stigma. At the smallest pollen load size, patterns of seed paternity were altered in two of the three pollen donor pairs; seed paternity became more equal among donors. For one of three pollen donor pairs, seed paternity was more divergent among donors on stressed maternal plants. Finally, for one pollen donor pair, rank order of pollen donor performance changed from the medium to the small pollen loads on stressed vs. control maternal plants. Thus, some field conditions may alter patterns of nonrandom mating in wild radish.

Reproductive biology of Colletia spinosissima (Rhamnaceae) in Argentina

We studied the reproductive behavior of Colletia spinosissima Gmel. at six sites in two regions of Argentina during the 1998–2000 flowering seasons. Flowering takes place in winter. The flowers are homogamous, fragrant, and entomophilous. There was evidence of self-incompatibility and cross-incompatibility between nearby partners. Fruit set of open-pollinated flowers (13.3%) seemed pollen-limited, since low fecundity was consistently associated with small stigmatic pollen loads and with low numbers of successful pollen tubes, and fecundity increased with abundance of efficient pollinators. Reproductive success (percent ovules becoming viable seeds) was 3.6. It is hypothesized that homogamy was established in Colletia because in an impoverished pollination environment – such as the one likely prevailing when this lineage evolved – achieving pollen export and receipt in a single pollinator visit has selective advantages. A large stigma and an extragynoecial compitum would compensate for the homogamy-driven, increased pollen-stigma interference.

Condition-dependent response to changes in pollen stores by honey bee ( Apis mellifera ) colonies with different parasitic loads

The impact of a parasitic infestation may be influenced by nutritional state, in both individuals and colonies. This study examined the interaction between pollen storage and the effects of an infestation by the mite, Varroa jacobsoni Oudemans, in colonies of the honey bee, Apis mellifera L. We manipulated the pollen storage and mite infestation levels of colonies, and measured pollen foraging and brood rearing. Increased pollen stores decreased both the number of pollen foragers and pollen load size, while initially at least foragers from colonies with moderate infestations carried smaller pollen loads than those from lightly infested colonies. Over the course of the experiment, all colonies significantly increased pollen-foraging rates and pollen consumption, which was presumably a seasonal effect. Lightly infested colonies exhibited a larger increase in pollen forager number than moderately infested colonies, suggesting that more intense mite infestations compromised forager recruitment. Brood production was not affected by the addition of pollen, but moderately infested colonies were rearing significantly less brood by the end of the experiment than lightly infested colonies. Furthermore, the efficiency with which colonies converted pollen to brood decreased as the pollen storage level decreased and the infestation level increased. The results of this study may indicate that honey bee colonies adaptively alter brood-production efficiency in response to parasitic infestations and seasonal changes.

The Effect of Population Size on Stigma Pollen Load, Fruit Set, and Seed Set in Allium stellatum Ker. (Liliaceae)

In Allium stellatum, a xenogamous self‐compatible prairie species, we examined stigma pollen load, percentage of fruit set, and percentage of seed set in large and small populations over 2 yr. We tested for significant effects of population size on these characteristics, for relationships between stigma pollen loads and fruit and seed set, and for significant variation between years in reproductive characteristics. In 1994 stigma pollen loads were inversely related to population size and were significantly lower than in 1993. The lower stigma pollen loads in 1994 were associated with adverse weather conditions (high rainfall), which are known to affect pollinator activity and may have led to competition among plants for pollinators in large populations. However, population size was not correlated with percentage of fruit set or percentage of seed set in either year. The lack of fruit and seed set differences between large and small populations, particularly in 1994, likely resulted from the small amount of pollen needed to trigger fruit and seed production. Our results indicate that for some self‐compatible species that similarly require small pollen loads, the effects of population size on reproduction are minimal, even if differences in pollen‐movement exist between large and small populations.

Foraging behaviour of bee pollinators on the tropical weed Triumfetta semitriloba : departure rules from flower patches

We studied the departure rules from flower patches used by bee pollinators of the tropical shrub weed Triumfetta semitriloba Jacq. (Tiliaceae). Flowering plants were distributed in well delimited clumps, in each of two pasture areas (A1 and A2) and one area of forest gap (A3), in Vicosa, southeastern Brazil. Five solitary bee species, Augochlorella michaelis, Augochloropsis cupreola, Pseudocentron paulistana, Ceratinula sp., Melissodes sexcincta, and one social bee, Plebeia droryana, were observed. Three departure rules were examined:(1) residence (time spent per patch); (2) giving-up (time spent in the patch between the last successful resource encounter and departure); (3) Three log-linear models were delineated to analyze the rule of departure: (1); DP*;TLF*;TPF*;TTP; (2) DP*;TLF*;TPF*;ALF; (3) DP*;TLF*;TPF*;NFV, where DP is departure from the patch; TLF is time spent in the last visited flower; TPF is time spent in penultimate visited flower; TTP is total time spent in the current patch; ALF is activity in the last visited flower; NVF is total number of visited flowers on the current patch. Flowering plants were considered as flower patches. The results showed that the departure from patches of flowers was influenced by intrinsic factors, such as body size, energetic requirements and pollen load size, and by extrinsic factors, such as resource availability and distribution, and density of resource patches. Small bodied bees, with lower energetic requirements and smaller pollen loads (Ceratinula sp. and P. droryana) or bees which occurred in A3, the area with fewer flowers per patch and lower density of flower patches ( Ceratinula sp. and A. michaelis), used a residence departure rule. Species with larger body size or which occurred on areas of higher density of flower patches and with more flowers per patch, A1 and A2 (P. paulistana, M. sexcincta, A. cupreola), used a probabilistic role of departure.

Effects of pollination intensity on the vigor of the sporophytic and gametophytic generation of Cucurbita texana

We examined the effects of pollen competition (pollen load size) on sporophytic vigor and gametophytic performance in Cucurbita texana, a wild gourd, while controlling for alternative interpretations of the data. Under field conditions we compared the vigor of progeny produced from large and small pollen loads and examined the in vitro performance of the pollen produced by the progeny. We found that the progeny from large pollen loads germinated faster and had a greater reproductive output (male flowers and fruits) than progeny produced from small pollen loads. In addition, we found that the pollen produced on plants derived from large pollen loads grew faster in vitro than the pollen produced on plants derived from small pollen loads. These findings indicate that pollen competition affects the performance of the resulting sporophytic generation and the microgametophytes they produce.

Effects of pollen competition on the reproductive performance in cucurbit hybrids (Cucurbitaceae): F1 and backcross generations

We examined the effect of pollen competition on the reproductive performance of Cucurbita F1 and backcross progeny while attempting to control for nongenetic maternal effects and nonrandom seed abortion as possible explanations for the data. We compared the vigor of F1 and backcross progeny produced from large and small pollen loads under field conditions. The results from both field studies indicated that the progeny produced from large pollen loads outperformed the progeny produced from small pollen loads for several measures of vigor. The F1 progeny produced from large pollen loads germinated faster, and produced 12% more staminate flowers and 9% more pistillate flowers than the progeny from the small pollen loads. Similarly, the backcross progeny produced from large pollen loads germinated faster, and produced 8% more pistillate flowers and 20% more fruits than the progeny from the small pollen loads. This study demonstrates a positive correlation between the size of the pollen load and the vigor of the resulting progeny under field conditions. Keywords: pollen competition, pollen performance, nonrandom fertilization, plant fitness, reproductive output.

Effects of pollen selection on progeny vigor in a Cucurbita pepo x C. texana hybrid

We examined the effects of pollen selection for rapid pollen-tube growth on progeny vigor. First, we crossed a wild gourd (Cucurbita texana) to a cultivated zucchini (Cucurbita pepo cv 'Black Beauty') to produce an F1 and then an F2 generation. Half of the F1 seeds were produced by depositing small loads of C. texana pollen onto the stigmas of C. pepo. These small pollen loads were insufficient to produce a full complement of seeds and, consequently, both the fast- and the slow-growing pollen tubes were permitted to achieve fertilization. An F2 generation was then produced by depositing small loads of F1 pollen onto stigmas of F1 plants. The F2 seeds resulting from two generations of small pollen loads are termed the non-selected line because there was little or no selection for pollen-tube growth rate on these plants. The other half of the F1 and F2 seeds were produced by depositing large pollen loads (>10 000 pollen grains) onto stigmas and then allowing only the first 1% or so of the pollen tubes that entered the ovary to fertilize the ovules. We did this by excising the styles at the ovary at 12-15 h after pollination. The resulting F2 seeds are termed 'the selected line' because they were produced by two generations of selection for only the fastest growing pollen tubes. Small pollen loads from the F2plants, both the selected and the non-selected lines, were then deposited onto stigmas of different C. pepo flowers, and the vigor of the resulting seeds was compared under greenhouse and field conditions. The results showed that the seeds fertilized by pollen from the selected line had greater vegetative vigor as seedlings and greater flower and fruit production as mature plants than the seeds fertilized by pollen from the non-selected line. This study demonstrates that selection for fast pollen-tube growth (selection on the microgametophyte) leads to a correlated increase in sporophyte (progeny) vigor.

Effects of Pollen Competition on Progeny Performance in a Heterozygous Cucurbit

We used an F1 hybrid of zucchini and its wild progenitor to examine the effects of pollen competition on progeny performance. We experimentally varied the intensity of pollen competition by depositing large or small pollen loads onto stigmas. To separate the effects of pollen competition from maternal effects, we excised the styles of flowers receiving the large pollen loads after only the fastest pollen tubes had entered the ovary. The styles from flowers receiving small pollen loads were excised after both fast- and slow-growing tubes had entered the ovary. Consequently, the mature fruits from the two treatments were similar in seed number and weight. Because our previous studies had revealed that fast- and slow-growing pollen tubes fertilize ovules in different regions (locations) within the ovary (fast into region I and slow into region 3), we examined the vigor of the progeny produced in regions I and 3 from both large and small pollen loads. The results revealed that the progeny from large pollen loads outperform progeny from small pollen loads. We also found that for small pollen loads, the progeny from region I outperform the progeny from region 3, which indicates that the progeny produced by the fastest pollen tubes outperform the progeny produced from the slowest.

Effects of Pollen-load Size on Sporophyte Competitive Ability in Two Epilobium Species

-Competition among pollen-tubes for ovules can lead to increased sporophytic vigor. This phenomenon has been well studied in a few cultivated species, but little is known about whether it occurs in natural populations. I tested for effects of pollen-tube competition on progeny performance in Epilobium septentrionale and E. canum ssp. canum. Three maternal plants of each species received small vs. excessive pollen-loads, using mixtures of pollen from three paternal plants. Progeny from these treatments were grown in competition with wild oats and harvested after nine weeks in the greenhouse. Intense pollen competition led to significantly larger progeny in only one of the six maternal families. In another maternal family, offspring from excessive pollen-loads were actually smaller than those from small pollen-loads. Pollination treatment had inconsistent effects on seed size, which did not appear to influence plant size. This study does not support the hypothesis that pollen competition fosters selection for faster-growing progeny.

EFFECTS OF POLLINATION INTENSITY IN CAMPSIS RADICANS.

I examined effects of pollination intensity on fruit, seed and seedling characteristics in Campsis radicans, and joint effects of pollen donor and pollination intensity on fruit production. Large pollen loads were more likely to initiate fruit production than small pollen loads, and the former fruits contained more seeds and a greater total seed mass. No further increases in seed number or mass occurred for pollen loads above 4,000 grains. The weight of individual seeds was unaffected by pollen load. Effects of pollen donor were generally larger than effects of pollen load, and fruit production from small loads of pollen from one donor were sometimes equal to fruit production from larger pollen loads from another donor. The ratio of pollen grains deposited to resultant seeds increased with pollen load, and several explanations are proposed. Seeds from heavy pollinations emerged better than seeds from light pollinations, but did not differ in speed of germination or in the performance of seedlings up to 126 days. The emergence differences are probably due to differing intensities of pollen tube competition.

Effects of Pollination Intensity in Campsis radicans

I examined effects of pollination intensity on fruit, seed and seedling characteristics in Campsis radicans, and joint effects of pollen donor and pollination intensity on fruit production. Large pollen loads were more likely to initiate fruit production than small pollen loads, and the former fruits contained more seeds and a greater total seed mass. No further increases in seed number or mass occurred for pollen loads above 4,000 grains. The weight of individual seeds was unaffected by pollen load. Effects of pollen donor were generally larger than effects of pollen load, and fruit production from small loads of pollen from one donor were sometimes equal to fruit production from larger pollen loads from another donor. The ratio of pollen grains deposited to resultant seeds increased with pollen load, and several explanations are proposed. Seeds from heavy pollinations emerged better than seeds from light pollinations, but did not differ in speed of germination or in the performance of seedlings up to 126 days. The emergence differences are probably due to differing intensities of pollen tube competition.