Nectar replenishment and pollen receipt interact in their effects on seed production of Penstemon roseus

Resource supply and pollen delivery are often thought to equally limit seed production in animal-pollinated plants. At equilibrium, plants should show no response to experimental pollen supplementation because resources limit seed set above the current level of pollen attraction, while experimental reduction in pollen deposition below the equilibrium level would reduce seed set. The predicted equilibrium may be disrupted, however, if plants expend additional energy to replenish removed nectar. We investigated the combined effects of nectar removal and pollen delivery on female reproductive success of Penstemon roseus (Plantaginaceae), a hummingbird-pollinated plant that replenishes removed nectar. We first documented that the frequency of experimental nectar removal was correlated with total nectar secretion; and increased frequency of nectar removal resulted in increased female reproductive costs to the plant. Trade-offs between investing resources in nectar and investing resources in seeds were then investigated in two contrasting natural populations by removing nectar from flowers at increasing frequencies while simultaneously hand-pollinating flowers with increasing amounts of pollen. Seed set was lowest at low levels of pollen deposition, highest at medium-sized pollen loads, and intermediate when pollen loads were highest. At both sites, the frequency of nectar removal and pollen deposition had an interactive effect on seed production, in that intermediate levels of nectar removal result in the absolute highest seed set, but only at intermediate pollen loads. At high pollen loads, seed set was higher following little to no nectar removal, and at low pollen loads, all rates of nectar removal affected fecundity equally. Seed mass responded to nectar removal and pollination differently than did seed set. High levels of nectar removal and pollen delivery both lowered seed mass, with little interaction between main effects. Our findings are among the first to demonstrate that nectar replenishment costs and pollination intensity jointly affect seed production. This conflict between nectar replenishment costs and pollen-limiting factors results in trade-offs between pollinator attraction and seed provisioning. Thus, resource allocation towards nectar production should more often be considered in future studies of pollen limitation.     

Mobility of Impatiens capensis flowers: effect on pollen deposition and hummingbird foraging

Flexible pedicels are characteristic of birdpollinated plants, yet have received little attention in studies of hummingbird-flower interactions. A major implication of flexible pedicels is that flowers may move during pollination. We examined whether such motion affected interactions between ruby-throated hummingbirds (Archilochus colubris) and jewelweed (Impatiens capensis) by increasing pollen deposition and by altering the effectiveness of nectar removal. For I. capensis, flower mobility enhanced pollen deposition: birds had significantly longer contact with anthers and more pollen deposited on their bills and crowns when foraging at mobile flowers than at flowers that had been experimentally immobilized. In contrast, flower mobility imposed a cost on hummingbirds by significantly increasing their handling times and reducing their extraction rates relative to their interactions with immobile flowers. Field observations indicated that the motion observed during hummingbird visits did not occur when bees (Bombus spp., Apis mellifera) visited I. capensis flowers, which suggests that the mobility of I. capensis flowers is an adaptation for hummingbird pollination.     

Pollination ecology of Penstemon roseus (Plantaginaceae), an endemic perennial shifted toward hummingbird specialization?

Summary Pink-flowered tubular Penstemon roseus (Plantaginaceae), which has shifted partially to hummingbird pollination, blooms on high-elevation slopes in the mountains in Tlaxcala, Mexico. We studied the interactions between pollinator visitation rates to flowers, pollen removal and deposition, flower size, and nectar removal frequency on seed production in P. roseus. We combine observational and experimental studies in two contrasting natural populations. Our manual pollinations revealed that P. roseus is fully self-compatible. Autonomous self- and manual self-pollinated flowers matured as many seeds as when outcrossed, but outcrossing seems to become better than selfing as the flowering season progressed. Early in the season flowers that were bagged and hand-selfed, hand-outcrossed, or autonomously selfed, or unbagged and naturally pollinated had equal seed set in all four treatments. But later in the season, outcross pollen gave approximately twice as much seed set as the two self-treatments. Low levels of pollen receipt and pollen removal were consistent with the long time elapsed for a given plant to be visited by hummingbirds, which suggests pollen shortage in both sites. Despite differences in pollinator visitation rates to flowers, probability of flower visitation, removal and deposition of pollen, and nectar production rates between populations, we found that total nectar production had no effect on seed production at either site. The daily nectar secretion rate of 0.3–0.65 mg sugar per flower per 1–3 days was low relative to other hummingbird-adapted Penstemon species (typical range: 1.5–5 mg sugar per flower), and it might be intermediate between hummingbird- and bee-adapted Penstemon flowers. Our results support the hypothesis about a shift toward hummingbird pollination, and provide an example of a ‘despecialized’ Penstemon species, which attracts high-energy pollinators (hummingbirds) and profits from outcrossing, but retains bee-syndrome floral traits and low sugar production rates.     

Dioecy inHonkenya peploides var.major (Caryophyllaceae)

Sexual difference and pollination system inHonkenya peploides var.major were investigated. The present analyses indicate that sex expression of this plant is functionally dioecious, rather than androdioecious, as described previously. Its flowers are largely pollinated by small bees, hover-flies, flies, and ants. Female flowers were morphologically characterized by conspicuously longer styles and smaller petals, and had no pollen grains, while male flowers had shorter styles, larger petals, and longer filaments, and produced viable pollen grains. Although the male flowers possess many ovules in an ovary, they set fruits at a very low frequency in natural conditions and did not produce fruits by pollination experiments. We suggest that the dioecy of this species has been derived from hermaphrodite through gynodioecy, and not directly from hermaphrodite.     

Notes on the floral morphology in Vivianiaceae (Geraniales)

The functional floral morphology of the three genera of Vivianiaceae (= Ledocarpaceae, Geraniales), Rhynchotheca, Viviania and Balbisia, is compared. Likely pollination mechanisms are inferred from morphology and field observations. The flowers of Viviania are nectariferous and apparently zoophilous with nectar as the (primary) pollinator reward. Balbisia has pollen flowers without nectaries, its showy corolla indicates that it is also zoophilous with pollen as sole pollinator reward; bees were observed as flower visitors. One taxon (B. gracilis) may be anemophilous. Rhynchotheca has flowers without petals, with large, pendulous anthers and lacks nectaries. It shows synchronous mass flowering in its natural populations and is evidently anemophilous. A comparison with other Geraniales shows that nectar flowers with small anthers are likely the ancestral condition in Vivianiaceae. This suggests that the pollen flowers with larger anthers of Balbisia and Rhynchotheca may represent an apomorphic condition. The documentation of pollen flowers and anemophily in Vivianiaceae expands the range of known floral and pollination syndromes in Geraniales.     

Sites of ethylene production in the pollinated and unpollinated senescing carnation (Dianthus caryophyllus) inflorescence

Production of endogenous ethylene from the styles, ovary and petals of pollinated and unpollinated flowers of Dianthus caryophyllus L. was measured. The rate of ethylene production of cut, unpollinated flowers aged in water at 18°C was low until the onset of petal wilting, when a rapid surge of ethylene occurred in all tissues. The flower ethylene production was evolved mostly from the styles and petals. The bases of petals from unpollinated, senescing flowers evolved ethylene faster and sometimes earlier than the upper parts. Treatment of cut flowers with propylene, an ethylene analogue, accelerated wilting of flower petals and promoted endogenous ethylene production in all flower tissues. Pollination of intact flowers also promoted endogenous ethylene production and caused accelerated petal wilting within 2–3 days from pollination. Although the data are consistent with the hypothesis that ethylene forms a link between pollination of the style and petal wilting, in the unpollinated flower the style and petals can evolve a surge of ethylene independently of each other, about the time when the petals irreversibly wilt. The results are discussed in relation to the role of ethylene in flower senescence.     

Selection for pollination success and the mechanical fit of Imfiatiens flowers around bumblebee bodies

Selection on flowers has often been viewed as being particularly strict, constant, and responsible for species differences. Impatiens pallida and I . capensis flowers fit snugly around bees, leading one to expect a close relationship between floral morphology and pollination success. My studies on the amount of pollen removed from androecia and deposited on stigmas in single visits by bumblebees did not confirm this supposition. Trimming off parts of the floral vestibule with scissors and gluing in pleats had very little effect on die amount of pollen that bees moved. In reciprocal transfer experiments, flowers from different populations sometimes differed in the amount of pollen moved, but when the two species were compared in sympatry, pollen removal and deposition differed hardly at all. A comparison of the relationship between pollen movement and floral morphology among 15 populations showed that, although there was great heterogeneity in die amount of pollen moved, the observed differences were independent of floral morphology. None of this supports a belief in strong selection that fine-tunes the mechanical fit between bee and flower; selection for visitation success based on pollinator behaviour may have a much stronger influence on floral characters.     

Effectiveness of buzz pollination in Pedicularis chamissonis: significance of multiple visits by bumblebees

Pollination success of plants is highly susceptible to the frequency of visits and foraging behavior of pollinators. Pollination of the nectarless flowers of Pedicularis species depends on bumblebee workers collecting pollen by vibrating the anthers (buzz pollination). However, little is known about the efficiency of the pollination system. Foraging behavior, pollen removal from anthers and pollen deposition on stigmas of P. chamissonis were studied to assess the effectiveness of buzz pollination in an alpine snowbed population of northern Japan. Although bumblebees tended to visit most of the flowers open at a given time within inflorescences during a single visit, pollen removal rate at the first visit was about 20%, and buzzing period decreased with increasing number of previous visits, resulting in a decreasing proportion of pollen removed per visit as the number of visits increased. These trends enable plants to provide pollen for more pollinators. The number of pollen grains deposited on stigmas was not saturated during the first visit and increased with additional visits. Irrespective of weak self-compatibility, evidence of interference between self and outcross pollen was lacking for seed production. Therefore, buzz pollination in P. chamissonis acts as a mechanism that improves the chance of cross-pollination upon multiple visits if pollinator visitation is frequent.     

The dynamics of pollen removal and deposition,and its effects on sexual phases in a protandrous plant: Glechoma longituba (Lamiaceae)

The duration of sexual phases in dichogamous plants are affected by many factors. Using both experimental and observational studies, we investigated natural patterns of pollen removal and deposition, visiting frequency of pollinators, patterns of nectar secretion, and effects of pollen removal and stigmatic pollen deposition on the duration of sexual phases in a protandrous plant, Glechoma longituba. We found that visiting frequency of pollinators correlated with the nectar secretion pattern. The nectar volume during the male phase was higher than during the female phase. In the morning, the main pollinator, the bee Anthophora plumipes, mainly foraged for nectar and showed no preference for flowers in male or female phase, despite male phase flowers producing higher amounts of nectar. However, in the afternoon, they changed their behavior and foraged mainly for pollen, and then showed a preference for flowers in male phase. Furthermore, the rates of pollen removal and stigmatic pollen deposition can affect the starting time and the duration of the female phase. When pollen removal and pollination rates are low due to scarcity of pollinator services, the sexual phase can be prolonged, leading to an overlap, and thereby enhance the chance for sexual reproduction through pollinator‐facilitated self‐pollination. We consider the variation of sexual phases in Glechoma longituba an adaptive mechanism prepared for both cross‐pollination enhancement and reproductive assurance depending on the available pollination services.     

Ovule pre-emption and pollen limitation in a self-fertile perennial herb (Blandfordia grandiflora,Liliaceae)

The extent, frequency and causes of pollenlimited seed production were examined in partially selffertile populations of Blandfordia grandiflora for 2 years. Percentage seed set of open-pollinated plants (50–57%) did not differ within or between years, and was about 19% less than experimentally cross-pollinated plants (70–75%). Floral visits by honeybees did not differ through the flowering season and the number of pollen grains deposited on stigmas within 12 h of flowers opening exceeded the number of ovules per flower, indicating that the quality rather than the quantity of pollination limited seed set. Pollen limitation was caused by concurrent self- and cross-pollination and the subsequent abortion of some selfed ovules due to inbreeding depression. Natural seed set (55%) was intermediate between selfed (43%) and crossed (75%) flowers and was not increased when flowers that had been available to pollinators for 24 h were hand cross-pollinated, suggesting that ovules were already fertilized. Similarly, experimental pollination with both cross and self pollen within 24 h of flowers opening did not increase seed set relative to natural seed set, indicating that both cross- and self-fertilizations had occurred. In contrast, when selfing followed crossing by 48 h, or vice versa, seed set did not differ from crossed-only or selfed-only flowers, respectively, indicating that ovules were pre-empted by the first pollination. Collectively, these results indicate that under natural conditions self pollen pre-empts ovules, rendering them unavailable for cross-fertilization. This selfing reduces fecundity by 50%, as estimated from the natural production of cross seeds when selfing was prevented. Consequently, selection should favour floral traits, such as increased stigma-anther separation or protandry, that reduce interference between male and female functions that leads to selfing.     

The effect of petal size manipulation on pollinator/seed-predator mediated female reproductive success of Hibiscus moscheutos

The effects of petal-size manipulations on the behavior of pollinators and pollen/seed predators, and on pollen removal and deposition, were studied in Hibiscus moscheutos (Malvaceae) populations. The ultimate effects on the female reproductive success of flowers, such as fruit set, seed predation rate, and final seed set were also measured. We applied three levels of petal removal (100%, 50%, and 0% size reduction in radius) to flowers in natural populations. Two pollinators (Bombus pennsylvanicus and Ptilothrix bombiformis) ignored flowers without petals, suggesting that pollinators use petals as a visual cue to locate flowers. Consequently, 100% petal removal reduced female reproductive success considerably, mainly through a higher rate of fruit abortion due to failure of pollen deposition on stigmas. No significant differences between the 50% petal removal treatment and uncut control were detected in any components of female success examined. The results, therefore, suggest that differences in petal size have little influence on female reproductive success of Hibiscus flowers at our study site. Final seed set varied considerably depending on the larval densities of two coleopteran seed predators (Althaeus hibisci and Conotrachelus fissunguis). A. hibisci responded to petal size, and a higher density of adults was found in flowers in which petal size had not been reduced. Because Althaeus feed on pollen as adults and no effect of petal size on seed predation was detected, the preference of Althaeus for larger flowers may represent a foraging strategy for adult beetles and may exert counteracting selection pressure on petal size through male reproductive success of flowers. Received: 30 November 1997 / Accepted:12 June 1998     

FLORAL BIOLOGY OF MAGNOLIA

The floral biology of eight species of Magnolia native to the United States is described. The flowers are protogynous. They are pollinated by several species of beetles that enter buds as well as closed and open flowers to feed on nectar, stigmas, pollen, and secretions of the petals. Individual flowers persist from two to four days and undergo a series of petal, stigma, and stamen movements that assure pollination by beetles. It is suggested that the flowers of Magnolia are highly specialized for exclusive pollination by beetles. These specialization mechanisms produce large quantities of food for the beetles and deny other types of insects (bees, moths, etc.) access to the flowers at critical stages in the pollination process, i.e., when stigmas are mature and pollen is shed.     

Effect of nectar supplementation on male and female components of pollination success in the deceptive orchid Dactylorhiza sambucina

Many orchids lack floral nectar rewards and therefore rely on deception to attract pollinators. To determine the effect that a mutation for nectar production would have on overall pollination success of the deceptive orchid Dactylorhiza sambucina, we recorded pollen deposition and removal in flowers of plants that had either been supplemented with an artificial nectar solution or left unmanipulated as controls. Nectar supplementation resulted in significant increases in the proportion of flowers pollinated, regardless of morph colour and the density of plants supplemented in the population. However, nectar supplementation had a significant positive effect on pollinaria removal only for the yellow morph in one experiment in which a low proportion of plants were supplemented. Thus a mutation for nectar production would have a positive effect on overall pollination success in D. sambucina, particularly the female component. The observed patterns are discussed in relation to other factors, such as cross-pollination and the reallocation of nectar resources for other plant functions, which are traditionally considered to shape the rewardless strategies of orchids.     

The influence of nectar production on hawkmoth behavior,self pollination,and seed production in Mirabilis multiflora (Nyctaginaceae)

Nectar is the most common floral reward that plants produce to attract pollinators. To determine the effect of nectar production on hawkmoth behavior, pollen movement, and reproductive success in Mirabilis multiflora, I manipulated nectar volumes and observed the subsequent foraging behavior of the hawkmoth Hyles lineata and the resulting pollen movement patterns. Individual hawkmoths visited significantly more flowers on plants with more nectar. The increase in flower visits significantly increased pollen deposition on stigmas and pollen removal from anthers when nectar volume was raised to twice the highest level found in nature. As hawkmoths visited flowers consecutively on a plant, the proportion of self pollen deposited on stigmas increased significantly and rapidly. Based on simulated hawkmoth visits, seed set was significantly reduced for flowers later in a visit sequence. A simple model combining these results predicts that the form of selection on nectar production varies depending on pollinator abundance. Using a multiple regression analysis a nearly significant (P < 0.08) effect of stabilizing selection was detected during a single season as predicted by the model for the prevailing hawkmoth abundance. Although increased nectar production may indirectly affect plant fitness by reducing resources available for other plant functions, the direct effect of high nectar production on pollinator behavior and self pollination may generally limit floral nectar production.     

ANTHER SMUT DISEASE IN DIANTHUS SILVESTER (CARYOPHYLLACEAE): NATURAL SELECTION ON FLORAL TRAITS

Mating opportunities, pollination intensity, and pollen dispersal ability may vary with variation in floral traits such as color, size, and shape. Where these traits are selected by pollinators for enhanced elaboration, they should evolve toward the equilibrium between selection for further elaboration and selection against this through reduced fecundity or vitality. Here we show that pollinator-borne fungal diseases of plants may be a factor influencing the position of this equilibrium. Populations of the rock pink, Dianthus silvester often contain individuals infected with the anther smut fungus Microbotryum violaceum (= Ustilago violacea). In a naturally infected population in the Alps of eastern Switzerland we investigated how intrapopulation variation in flower size and nectar rewards influenced spore deposition and how floral traits varied with disease status. We found that spore deposition increased with increasing petal size, suggesting that large-flowered plants were at a greater risk of disease. Spore deposition was also higher for plants growing in patches with many or a high proportion of diseased neighbors. Multiple regression analyses showed that petal size or nectar reward influenced spore deposition when the effects of neighborhood disease abundance were controlled statistically. In sequential analyses, after removing the effects of disease density or frequency and plant gender, petal length explained significant variation in spore deposition. Diseased plants had reduced female reproductive organs, but calyx size was intermediate between that of healthy perfect and female flowers of this gynodioecious-gynomonoecious species, and diseased plants bore flowers with the largest petals. This may reflect a symptom of this disease or the cause, if larger-flowered plants are more likely to become infected. We conclude that investment to pollinator attraction may bring an enhanced risk of contracting this sterilizing pollinator-borne disease, so natural selection by the fungus M. violaceum acts to lower attractiveness to pollinators.     

POLLINATION DYNAMICS IN EPILOBIUM CANUM (ONAGRACEAE): CONSEQUENCES FOR GAMETOPHYTIC SELECTION

Several hand-pollination experiments have shown that pollen tube competition leads to gametophytic selection, and thus affects genotypic and phenotypic characteristics of the next generation. This study is one of the first to quantify natural levels of pollen tube competition. In a population of Epilobium canum, I measured both the amount of pollen deposited on stigmas and the timing of deposition. Approximately 20 tetrads were required for full seed set within fruits. Hummingbirds deposited >20 tetrads at 50–70% of the flowers examined, often in a single load. When pollen arrived in 2 loads, a portion of the ovules within each ovary was probably sired by competing pollen from the second load. Competition may be relatively weak unless at least 30 tetrads compete for ovules. About 20% of the flowers received >30 tetrads in the first load, and 13% acquired > 30 tetrads in 2 loads. The frequency and intensity of pollen tube competition varied among plants. In some styles, 80% of the pollen tubes were excluded from access to ovules, but in others no competition occurred. Further studies of pollination rates and progeny fitness are needed before we can assess the role of pollen tube competition in natural populations. Potential effects of gametophytic selection are discussed.     

Pollen limitation and variation in floral longevity in gynodioecious Potentilla tanacetifolia

Pollination limitation is common in flowering plants and is thought to be a factor driving the evolution of floral traits.The plasticity of floral longevity to pollination may be an adaptation of plants to pollen limitation.However,this adaptation is less critical in short-lived flowers.To evaluate pollen limitation and the plasticity of floral longevity to pollination in Potentilla tanacetifolia,a gynodioecious herb with short-lived flowers,we analyzed its breeding system,tested sex-differential pollen limitation,and compared variations in floral display size in natural populations in Duolun County,Inner Mongolia,China.Hand pollination experiments and pollinator exclusion treatments revealed that P tanacetifolia is self-compatible and non-autonomously apomictic and shows sex-differential pollen limitation.The plasticity of floral longevity to pollination was observed; the floral duration of female plants was prolonged by approximately 3-4 hours with pollination exclusion treatment.Moreover,the percentage of flowers displayed on female plants during pollination exclusion treatment was significantly higher than that during natural pollination.Under natural pollination conditions,the percentage of flowers displayed on female plants was significantly higher than on hermaphrodite plants.Furthermore,approximately 50％ of the pollen grains spread out of the anthers of hermaphrodite flowers within 2 h of anthesis; the number of pollen grains adhering to the stigmas of hermaphrodite flowers was significantly higher than that adhering to female flowers when flowers shed their petals.These results indicate that variation in floral longevity may be an adaptive strategy to pollination conditions for gynodioecious P tanacetifolia.     

宁夏枸杞的花粉呈现式样及交配系统研究北大核心CSCD

被子植物的花回馈、雌雄蕊时空分离特征和花粉呈现式样等花部特征及传粉者效率会影响雄性适合度和有性繁殖过程。宁夏枸杞(Lycium barbarum)是中国宁夏、新疆、内蒙等干旱和半干旱地区分布的特有种。该研究对新疆喀什地区宁夏枸杞自然种群的花部综合征、花粉呈现式样、花回馈、传粉者行为与交配方式的关系进行观察及统计分析,以探讨其花粉逐步呈现的适应性及其在提高雌雄繁殖过程中的意义。结果表明:(1)宁夏枸杞的单花寿命为(4.07±0.15) d,而雄性持续时间(0.07±0.01 d)比雌性持续时间短(4±0.01 d);花寿命内雌蕊长度比雄蕊长,属柱头探出式异位类型;花粉呈现式样为不完全逐步呈现。(2)花寿命不同阶段花回馈间存在显著差异。(3)意大利蜜蜂(Apis mellifera)、熊蜂(Bombus sp.)、食蚜蝇(Syrphidae sp.)是宁夏枸杞在自然居群的主要访花者,其中熊蜂和食蚜蝇是主要传粉者,但传粉者效率低,属于高移出低沉积类型。(4)宁夏枸杞的花粉胚珠比(2 448.11±448.32)及授粉实验结果均表现出兼性异交特征;自发自交及人工自花授粉花的座果率及结籽率很低,属于自交不亲和类型;自然传粉花的座果率及结籽率比人工去雄异花低,存在其较高的花粉限制(40.71%)。研究发现,宁夏枸杞花部综合征表现出雌雄异位和花粉不完全逐步呈现式样,这是避免雌雄功能及雄蕊各花药间的干扰、减少花粉同步移出及保障其雄性适合度的有效途径,但自然居群的传粉者种类限制、传粉效率低以及自交不亲性是导致宁夏枸杞花粉限制及降低雌性繁殖成功的主要原因。     

Both biotic and abiotic factors influence floral longevity in three species of Epidendrum (Orchidaceae)

Extended flower longevity has been widely defended to occur as a consequence of low pollination activity. To test this hypothesis in Neotropical rainforests, we selected three species of Epidendrum that flower during rainy conditions and commonly exhibit low levels of insect visits. Experimental studies carried out in 15 populations distributed along a wide spatial gradient detected the presence of pollinator limitation in all populations. Pollinia removal without pollen deposition increased longevity, and thus extended longevity could apparently be considered as an ecological advantage for these orchids. Nevertheless, flower longevity was severely shortened by abiotic factors like precipitation and humidity. Comparison with other tropical orchids showed that flower longevity was higher even in species that flower during periods of more pollinator activity than the ones studied here, which is contrary to the hypothesis that pollinators are the only driving force for long‐lived flowers in Epidendrum. Thus, although long floral longevities are apparently favorable for the reproductive success of these species, the high maintenance costs experienced during rainy conditions are also an important factor constraining the selection of long‐lived flowers in Epidendrum.     

Optimal pollinator attraction strategies in Trollius ranunculoides Hemsl. (Ranunculaceae) at different altitudes: increased floral display or promotion of nectar output?

For alpine plant species, patterns of resource allocation to functional floral traits for pollinator attraction can be highly significant in adaptation to low pollinator abundance and consequent pollen limitation. Increased pollination can be achieved either through a larger floral display or production of more pollen rewards. In this study, variation in resource allocation to different components for pollinator attraction was studied along an altitudinal gradient in Trollius ranunculoides, an obligate self‐incompatible out‐crosser of the Qinghai–Tibet Plateau. We compared resource allocation to conspicuous yellow sepals (which mainly provide visual attraction) and degenerate petals (which provide the major nectar reward) between populations at four altitudes. Furthermore, we investigated the contribution of sepals and petals to pollinator attraction and female reproductive success in an experiment with sepal or petal removal at sites at different altitudes. At the level of single flowers, resource allocation increased to sepals but decreased to petals with increasing altitude. Consistent with these results, sepals contributed much more to visitation rate and seed set than petals, as confirmed in the sepal or petal removal experiment. Sepals and petals contributed to female reproductive success by ensuring visitation rate rather than visitation duration. To alleviate increasing pollen limitation with increasing altitude, resource allocation patterns of T. ranunculoides altered to favour development of sepals rather than petals. This strategy may improve pollination and reproductive success through visual attraction (sepal) rather than nectar reward (petal) over a gradient of decreasing pollinator abundance.