Flexible stigma presentation assists context-dependent pollination in a wild columbine


Author for correspondence:Shuang-Quan Huang Tel: +86 27 68753553 Fax: +86 27 68752560 Email: sqhuang@whu.edu.cn


  • • Stigma presentation has received much less attention than pollen presentation. Investigation of the various forms of stigma presentation may reveal previously undescribed reproductive mechanisms in plants.
  • • Here we investigated stigma presentation in the spring-flowering herb Aquilegia yabeana, whose linear stigma gradually develops over the lifetime of the flower. The degree of stigma development and recurvature are influenced by the timing of pollen deposition, apparently to favor cross-pollination.
  • • In this species, autogamous self-pollination was found to occur in the middle of the receptive period of the stigma if prior cross-pollination did not occur. However, because later-arriving cross-pollen was deposited closer to the ovary, it reached the base of the style before self-pollen which had been deposited earlier.
  • • Flexible stigma presentation helps the flower to maximize pollen resources under unreliable pollination conditions, taking advantage of both selfing and outcrossing. This is the first description of stigma development regulated by pollen deposition in a wild plant.


Temporal patterns in the presentation of pollen and stigma may influence pollen removal and receipt and consequently mating opportunities in flowering plants (Lloyd & Webb, 1986). It is clear that most plants spread the presentation of pollen in time instead of presenting simultaneously (Thomson & Thomson, 1992). The temporal staggering of pollen presentation is selected to increase the success of a pollen parent when pollinators transport pollen with diminishing returns (Lloyd & Yates, 1982; Harder & Thomson, 1989; Yeo, 1993; Harder & Wilson, 1994). Moreover, in addition to primary presentation on anthers, in several species pollen can be presented secondarily or tertiarily on stigmas and styles. This may in some way promote outcrossing (de Block & Igersheim, 2001). Yeo (1993) summarized, in 25 families, diverse forms of secondary pollen presentation that primarily function to reduce pollen losses, improving pollination efficiency.

In contrast, studies of stigma presentation are few. This topic has received little attention because of the belief that a flower needs to expose its stigmas only long enough to receive adequate pollen to fertilize its ovules (Lloyd & Yates, 1982; Wilson, 1995). However, forms of stigma presentation are also diverse and stigmas may evolve various adaptive mechanisms contributing to reproductive success. In Dalechampia (Euphorbiaceae), expanded stigmatic surfaces are presented laterally to the style. Pollen tubes start from the lateral stigma surfaces, grow up to the lip of the style, turn through 180°, and then grow through the style to the ovules. This circuitous growth is postulated to increase pollen competition after pollination (Armbruster et al., 1995). A comparative study of the stigmas of the African genera Rutidea and Nichallea, dissecting the role of the style/stigma as a secondary pollen presenter, showed rich variation in the location of stigma surfaces on styles (de Block & Igersheim, 2001). In many species in the order Scrophulariales, the two lobes of the stigma close in response to touch by pollinators (Newcombe, 1922), and the lobes reopen if the stigma receives too few pollen grains to fertilize its ovules (Fetscher & Kohn, 1999). This stigmatic sensitivity is considered to be a mechanism that minimizes intersexual interference and prevents intrafloral selfing (Lloyd & Webb, 1986; Fetscher & Kohn, 1999; Fetscher, 2001; Barrett, 2002). The flexistyle in Alpinia moves the stigma out of the way of self-pollen and back to a position where pollen from a different flower type can be received at a later time; this promotes assortative mating and restricts selfing (Li et al., 2002).

In some species there is one stigma per flower, while in others multiple stigmas are in service, related to syncarpy or apocarpy. Syncarpy with associated stigmas may permit pollen tubes to cross between carpels and increase pollen competition during pollen tube growth in the pistils (Carr & Carr, 1961; Mulcahy, 1979; Endress, 1982; Armbruster et al., 2002a; Huang, 2003). However, the multiple stigmas of apocarpous species can be presented over an extended receptive period. A recent study in Pyrus communis indicated that stigmas in asynchronous development coexist within a flower. Although the duration of stigma receptivity is brief for each stigma, the receptivity of the flower lasts longer (Sanzol et al., 2003). A staggered schedule for stigma presentation might be favored in unreliable pollination environments to optimize the use of pollen resources and enhance the maternal fitness of the flowers (Sanzol et al., 2003; Yu et al., 2005).

In a flower of Aquilegia caerulea James, the stigmas are presented gradually in succession during anthesis (Montalvo, 1992). This is also true in three other Aquilegia species that we have studied recently: Aquilegia ecalcarata Maximowicz, Aquilegia incurvata P. K. Hsiao and Aquilegia yabeana Kitagawa (Yu et al., 2005). The linear stigma on the ventral side of the style matures gradually from the apex towards the base over a period of days and becomes exposed to pollinators through a recurving movement of the style. Using experimental manipulations to examine the behaviour of stigmas in A. yabeana, we have found that pollination occurring at different phases of anthesis can influence stigma presentation in this species, and the potential adaptive significance of the flexible stigma presentation is discussed.

Materials and Methods

Study species and populations

Aquilegia yabeana Kitagawa (Ranunculaceae) is a perennial herb with a range extending from south-west to north China. It typically grows in the shaded understory of rich, mesic forests, along ditches and beside streams at elevations of 1000–2200 m above sea level. Plants generally flower in early May. This species is self-compatible and protogynous, with the female phase lasting 2–3 d followed by a male phase lasting 5–7 d. Hermaphrodite flowers have five purple sepals, and five purple petals each with a hooked nectar spur c. 2 cm long. The gynoecium consists of five unfused carpels surrounded by numerous stamens. A flower on average produces 65.7 ovules (Huang et al., 2004). Individuals produce several (one to eight) dichasial cymes (or a dichasium), each with an average of three to four flowers that bloom sequentially on the inflorescence, with distal flowers blooming first (Huang et al., 2004). Bumblebees are the most frequent pollinators of this species (L.-L. Tang, Q. Yu & S.-Q. Huang, unpublished data). Our study was performed on a natural population at the Huoditang Forestry Centre (33°27′ N, 108°28′ E), Qinling Mountain, Ningshan County, Shanxi Province, China, and in an experimental population at the Botanical Garden of Wuhan University, Wuhan, Hubei Province, China, composed of individuals transplanted from the natural population.

Measurements of autonomous self-pollination

To examine seed production resulting from autonomous self-pollination, 15 intact flowers from nine individual plants of the experimental population were caged. They were harvested at the end of anthesis and fixed in FAA (formalin : acetic acid : 70% ethanol at a ratio of 5 : 6 : 89 by volume). Seeds were counted in each immature fruit. As a control, we examined the seed set of 65 open-pollinated flowers from 30 individuals of the natural population.

To examine the timing and position of autonomous self-pollination, we observed the behavior of sex organs daily throughout anthesis. Twenty pistils were randomly picked from the caged flowers and were examined using a dissecting microscope to determine where self-pollen was deposited on the stigmas.

Artificial pollination treatments

To examine the effects of pollination on stigma development, we hand-pollinated bagged flowers in the experimental population at different stages of anthesis. Fifteen individual plants each with at least three consecutive buds were caged in fine mesh. All flowers were emasculated before anther dehiscence. The first flower was hand-pollinated as soon as it opened. At this time, styles were straight and stigmas were appressed together. The second flower was left until its styles just began recurving, and then we hand-pollinated one stigma per day until all the stigmas were pollinated. Styles normally began recurving on the second or third day of anthesis, after which time it became possible to pollinate stigmas of a given flower individually. Fresh pollen grains were collected from another individual and around 200 pollen grains were applied to each stigma along its length. The third flower was not pollinated. All flowers were harvested at the end of anthesis and fixed in FAA.

Measurement of stigma development status

We measured the stigma and style with a micrometer under a microscope. Surfaces with papillae were deemed to be stigmatic, and the smooth part of the pistil from the apex to where glandular hair appeared was deemed to be stylar tissue. The ratio of stigma length to style length was used as the indicator of stigma development status, to minimize variation from the effects of flower size. Data on stigma development status in different treatments were compared using a repeated-measures analysis of variance (ANOVA) followed by Tukey tests (SAS Institute, 1998).

Delayed pollination treatment

To determine whether late cross-pollination at the base of the stigma might give the cross-pollen prior access to the ovary, we compared pollen tube growth under sequential hand-pollinations. Nine plants were caged and one flower on each plant was hand-pollinated twice. First, pollen grains from another flower on the same plant (self-pollination) were added to the apex of the stigmas at the time when the self anthers were about to dehisce. The flowers were then emasculated. One day later, after more of the stigma was exposed, pollen from another plant (cross-pollination) was added to the basal portion. One flower was harvested 1 d after the second pollination, and the remainder were harvested at the end of anthesis. All flowers were fixed in FAA solution for later observation of pollen tubes.


In A. yabeana, both style movement and stigma development were influenced by pollen deposition. In the absence of pollination, the styles recurved gradually into a coil and stigma papillae covered the entire ventral side of the style (n = 75). The style was 7.32 ± 1.06 mm [mean ± standard deviation (SD)] long (n = 15) and the stigma : style length ratio was 1.00. However, this did not happen if pollen was deposited. In response to early pollination, all styles remained straight throughout anthesis (n = 71) and the stigma : style length ratio was only 0.25–0.55 (mean ± SD = 0.41 ± 0.09; n = 71) on the 6.1 ± 0.4 mm (n = 15) style by the end of anthesis, showing the developmental plasticity of stigmas.

Stigmas pollinated midway through anthesis showed intermediate development. In Fig. 1, the mean stigma : style length ratios of early-pollinated flowers are juxtaposed with data from later-pollinated pistils in the same flower to show the trend of the effect of pollination timing on the development status of the stigmas. The role of pollination in stopping the extension of stigma tissue was significant only at the first two time-points and its effect was negligible at mid-anthesis.

Figure 1.

The effect of pollination timing on stigma development in Aquilegia yabeana. The mean (± standard deviation) of the stigma : style length ratios of the pistils in an early-pollinated flower is juxtaposed with data from differentially pollinated pistils in the same flower to illustrate the trend. Points with the same subscript letters indicate that differences between stages are not significant (P > 0.05) using a Tukey post-hoc comparison. The sample size is shown above each treatment.

In the natural population, 72.8 ± 19.5% (mean ± SD) of ovules of open-pollinated flowers (n = 65) developed into seeds. In the garden, flowers set 37.2 ± 17.6% (n = 15) of seeds when pollinators were excluded. These seeds presumably resulted from autonomous self-pollination following incomplete protogyny and dehiscence of self-anthers in close proximity to stigmas. However, 90% of autonomously deposited pollen grains were on the apical 1.07 ± 0.40 mm of the stigma, leaving 79.9 ± 9.2% (n = 20) of the stigma length vacant.

Under a fluorescence microscope, pollen germination could be observed on any portion of the linear stigma. Self-pollen grains germinated near the apex and cross-pollen grains germinated in the basal portion of the stigma (Fig. 2). All pollen tubes grew downwards along the style. In all pistils accommodating both self-pollination and delayed cross-pollination, cross-pollen tubes reached the ovary but self-pollen tubes did not grow through the middle of styles (n = 9).

Figure 2.

Pollen tubes of cross-pollen deposited on the basal portion of the stigma (arrowhead) in Aquilegia yabeana reached the ovary before tubes (arrow) from self-pollen deposited 1 d earlier on the apical portion. (a) A style harvested 1 d after the second pollination. (b) A style harvested at the end of anthesis. Bar, 200 µm.


Our observations demonstrate that the stigma presentation of A. yabeana is flexible in receptivity scheduling and style movement, responding to pollen receipt. Maturation of columbine stigmas begins at the tip and spreads gradually towards the base of the style (Montalvo, 1992; Yu et al., 2005). If cross-pollen arrives on the day the flower opens, before self-anthers dehisce, there is less stigma development and the stigma/style remains straight. Without early pollination, the style recurves over time, exposing successively more basal portions of the stigma to later pollinations.

Self-pollen germinated and bagged flowers produced seeds, indicating self-compatibility in A. yabeana. Because it is protogynous, pollinator-mediated pollination has priority in this species. Ample cross-pollination at an early stage would keep the style straight and stop stigma extension. Hence, the flower would cease to be receptive by the time self-pollen was presented; this would limit potential self-pollination and encourage outcrossing. Meanwhile, the avoidance of interference between sexes might be a selective force on stigma behavior (Webb & Lloyd, 1986; Fetscher, 2001; Barrett, 2002). Stigmas that remain straight may be less likely to interfere with pollen dispersal. The architecture of styles within a flower also makes it unnecessary for stigmas to develop basally if they remain appressed to one another throughout anthesis, and thus inaccessible to pollinators.

In the absence of pollinators, flowers of A. yabeana can self-pollinate in the middle of anthesis while the stigmas continue to expose more and more receptive surface. Because of structural constraints, self-pollination is restricted to the tip of the stigma, allowing later pollinator-mediated pollinations to occur nearer the basal region of the extensive stigma. Although the self-pollen tubes have a head start in the race towards the ovary, they need to traverse more stylar tissue than the cross-pollen tubes. In this species it generally takes c. 60 h for pollen tubes to grow from the stigma tip to the ovary, leaving several days for the stigma to wait for pollinators. The positional advantage provides opportunities for delayed outcrossing to win the race for the ovules. Similarly, in some Collinsieae species, cross-pollen may also arrive later than self-pollen and win the race for reproductive success (Armbruster et al., 2002b).

A flowering plant is thought to have the ‘best of both worlds’ if it maximizes outcrossing while achieving back-up selfing to assure seed production in variable pollination environments (Becerra & Lloyd, 1992). Although many Aquilegia species invest heavily in attraction and reward for pollinators, scarcity of mates or pollinators may threaten their outcrossing as a result of vernal flowering and patchy distribution in marginal areas (Macior, 1978; Strand et al., 1996; Herlihy & Eckert, 2002; L.-L. Tang, Q. Yu & S.-Q. Huang, unpublished observations). Progressive stigma presentation seems to be a common feature in this genus. A fluctuating pollinator service might result in the interplay of selfing and outcrossing, and autonomous self-fertilization can be context-dependent, yielding reproductive assurance (Eckert & Herlihy, 2004; Kalisz et al., 2004). A. yabeana is the first wild species for which stigma development has been shown to be regulated by pollination conditions. Phenotypic plasticity in plants is considered to be a property of individual modular subunits triggered by local environmental conditions (de Kroon et al., 2005). It is hoped that further dynamic, in-depth studies of the reproductive behavior of this species will help to reveal more details of its reproductive strategies.


The authors wish to thank Shi-Guo Sun and Lu-Lu Tang for their help in the field; Charles Fenster, Chun-Feng Yang and You-Hao Guo for discussions; Scott Armbruster, Paul Wilson, Mark Rausher and an anonymous reviewer for their valuable comments; and Sarah Corbet for correcting the English and reviewing an earlier draft of the manuscript. This work was funded by the National Science Foundation of China (Grant no. 30400025 to S-QH).