Printer Friendly

Mutualism denied? Nectar-robbing bumble bees do not reduce female or male success of bluebells.


Many ecologists have decried the scanty attention paid to mutualistic interactions (Risch and Boucher 1976, Boucher et al. 1982, Vandermeer 1984, Keddy 1990, Bronstein 1991). Past history seems to support this concern. Of the 9453 papers published in Ecology, Ecological Monographs, and Ecological Applications between 1945 and 1993, only 55 used the word "mutualism" in the title or abstract, as opposed to 1087 for "competition," 767 for "predation," and 159 for "parasitism" or "disease." Although more papers have focused on mutualist species in recent years, many of these do not view mutualism per se as the object of study (Bronstein 1991, 1994). This lack of attention may reflect an attitude among ecologists that mutualisms are relatively uncommon because they are quickly eroded by "cheating" behavior, in which one species obtains a reward without providing a benefit in return. Although there is little doubt that a conflict of interest between the players is an integral part of many plant-pollinator interactions (Waser and Price 1983, Howe 1984), few studies have attempted to determine the balance point between benefit and detriment in mutualistic systems, even those in which apparently non-mutualistic behavior is common (but see Janzen 1979, Addicott 1986, Pellmyr 1989, Bronstein 1992, Tyre and Addicott 1993, Addicott and Tyre 1995). In this paper, I examine whether or not apparent "cheating" on the part of nectar-robbing bumble bees actually has negative impacts on the reproductive success of the plants they visit.

By definition, nectar robbers obtain a reward by piercing floral tissues without contacting the anthers and stigma, thereby failing to effect pollen transfer (Inouye 1979). Here, I follow common usage by describing floral visits in which the pollinator does contact anthers or stigma as "legitimate." As a result of nectar robbery, plants may suffer reduced reproductive success, due either to damage to floral tissues inflicted by robbers (McDade and Kinsman 1980, Galen 1983) or to reduced attractiveness of robbed flowers to legitimate pollinators (Heinrich 1976, Roubik 1982). However, some have argued that robbers might actually provide benefits to the plant in several ways. Putting aside cases in which purported "robbers" actually contact the reproductive parts of the flower while piercing the corolla (Macior 1966, Koeman-Kwak 1973, Waser 1979, Higashi et al. 1988, Scott et al. 1993), some have suggested that robbers could increase plant fitness, either by making nectar accessible to secondary robbers that also visit legitimately (Hawkins 1961) or by causing legitimate pollinators to visit more plants to collect an adequate reward, thereby increasing the outcrossing rate (Heinrich and Raven 1972, Gentry 1978, Zimmerman and Cook 1985). In addition, nectar robbers may pay some legitimate visits to the plants they otherwise rob. Taken together, these studies suggest that nectar robbers can have effects on plant fitness that range from negative to neutral to positive, and that it is dangerous to conclude, in the absence of experiments, that the "cheating" behavior of nectar robbers has detrimental effects on the plant.

Hermaphroditic plants gain fitness through both female and male function. Consequently, both the benefits and detriments of nectar robbers may have male and female components. Previous studies in other systems (e.g., Rust 1979, McDade and Kinsman 1980, Fritz and Morse 1981, Roubik 1982, Galen 1983, Roubik et al. 1985) have found a range of effects of nectar robbery on seed production (a component of female fitness). However, almost no studies have measured the effect of nectar robbery on correlates of male fitness. The only exceptions are studies of milkweeds (Wyatt 1980, Fritz and Morse 1981), in which removal of pollinaria is easily quantified. Yet, it is quite feasible that nectar robbers could influence a plant's male fitness even while having no effect on female fitness. Numerous studies have suggested that attractiveness to pollinators has a stronger effect on male fitness than on female fitness (Willson and Price 1977, Bell 1985, Queller 1985, Stanton et al. 1986, 1991, Cruzan et al. 1988, Young and Stanton 1990), although this is not universal (Schemske 1980, Schmid-Hempel and Speiser 1988, Campbell 1989, Broyles and Wyatt 1990). Hence, a reduction in legitimate visitation caused by nectar robbery, as long as some visits occur, might still allow full seed-set while reducing pollen export. To take more complete account of the net effect of nectar robbery, I measured correlates of both male and female fitness.


Natural history of the study system

Mertensia paniculata L. (Boraginaceae) is a clonal, but nonspreading, herbaceous perennial found across boreal North America (Hulten 1974). In Alaska, bluebells occur in communities ranging from alpine heaths to shaded sites beneath alder (Alnus incana) and white spruce (Picea glauca), but flowering occurs primarily in sunny locations (W. F. Morris, personal observation). Individual plants produce multiple stems from a fibrous caudex. Each stem produces one to several scorpioid inflorescences in which the pendulous flowers open sequentially over the course of the flowering period ([approximately equal to]6 wk in southcentral Alaska). Flowers contain four uniovulate mericarps; hence, each flower can produce 0-4 single-seeded nutlets. The corolla of bluebell flowers has a tubular base and a broad bell-shaped limb [ILLUSTRATION FOR FIGURE 1 OMITTED]. The corolla of unopened flowers is pink, and flower maturation occurs in the following sequence: the stigma is exserted from the unopened corolla [ILLUSTRATION FOR FIGURE 1A OMITTED], the anthers dehisce, and the corolla undergoes a color change from pink to blue as it opens. This transition is usually complete by 6-12 h after stigma exsertion, and flowers retain the blue corolla for 3-5 d. Due to sequential flower maturation, pink and blue flowers are always present on each inflorescence, except at the very beginning and end of the flowering period.

The bumble bees Bombus mixtus and B. frigidus are the primary visitors to bluebell flowers at the study sites in the Wrangell Mountains of southcentral Alaska. Of the 1974 bumble bee visits to bluebell flowers I observed in 1993 and 1994, 85.8% were made by individuals of these two species, which are difficult to distinguish in the field. Less common visitors include the bumble bees B. melanopygus (13.1% of all bumble bee visits) and B. occidentalis (1.1%), and a small fly that appears to use bluebell flowers primarily as mate-finding locations (W. F. Morris, personal observation). B. mixtus and B. frigidus foragers are commonly observed landing on the outside of opened (i.e., blue) flowers, chewing paired holes in the corolla tube [ILLUSTRATION FOR FIGURE 1B OMITTED], and inserting their tongues to remove nectar produced by nectaries below the ovaries. Because bees do not contact the anthers or stigma while robbing, such visits contribute to neither male nor female success of the plant. I have also observed individuals of the other two bumble bee species robbing flowers, but have never seen them making their own holes; this suggests that B. melanopygus acts only as a secondary robber of flowers that have been robbed previously by B. mixtus or B. frigidus. B. occidentalis is known to be a primary robber in other parts of its range (Inouye 1983); the fact that I have never seen it making holes in bluebell flowers may simply reflect its rarity at my study sites.

I quantified nectar robbing by bumble bees at three locations along a 5-km elevational gradient (Lower and Upper Bonanza Sites at 670 and 830 m, respectively, and the National Creek Pass Site at 1415 m) near Kennicott, Alaska in the summers of 1993 and 1994. Daily censuses of marked flowers, as well as a point census of marked flowers at the Pass Site in 1993, indicated that nectar robbery is frequent [ILLUSTRATION FOR FIGURE 2 OMITTED]. At the lowest and highest elevation sites, few flowers escaped robbing and most were robbed at or before anthesis. This intense robbery is comparable to the highest rates observed in other studies (e.g., McDade and Kinsman 1980, Roubik 1982, Young 1983, Kodric-Brown et al. 1984, Roubik et al. 1985). Nevertheless, robbery was relatively uncommon at the Upper Bonanza site in July 1994, indicating that its intensity varies over relatively short distances.

Consequences of nectar robbery for bluebell reproductive success

Per-flower female success. - To experimentally evaluate the effect of nectar robbery on bluebell reproductive success, I placed "antitheft devices" made from plastic tubing onto the bases of bluebell flowers. At the Lower Bonanza and National Creek Pass sites in 1993 and 1994, I randomly assigned each of a pair of unopened flowers on each of [greater than]100 stems per site to "robbed" or "unrobbed" treatments. In 1993, unrobbed flowers were fitted with collars (8 mm long) constructed from bands of transparent aquarium tubing slit down the sides and placed over the corolla tubes. "Robbed" flowers were left unmanipulated. In 1994, I used lighter weight dialysis tubing (1 cm wide) instead of aquarium tubing, and fitted 4 mm long half collars to the bases of flowers assigned to the "robbed" treatment. Half collars allowed bumble bees to rob flowers through the exposed portion of the corolla tube distal to the collar, but provided a partial control for effects of the collar by covering the portion of the corolla surrounding the ovaries and nectaries [ILLUSTRATION FOR FIGURE 1B OMITTED]. Observations suggest that dialysis tubing does not prevent bees from seeing the corolla within; when I completely enclosed flowers in tubing (see Male reproductive success), I found small droplets of dried nectar on the tubing where bumble bees had attempted to probe the enclosed flowers with their tongues. Flowers were monitored daily for evidence of robbery, and stems on which the uncollared "robbed" flowers remained unrobbed or the flowers with full collars were nevertheless robbed were omitted from the analysis.

As bluebell fruits develop, one or more of the four nutlets expands in size; when the enclosed seeds have filled, the expanded nutlets darken and fall separately from the fruit. Prior to fruit abscission, I collected all fruits, counted the number of expanded nutlets per fruit, and weighed all nutlets that contained mature seeds. Unfortunately, I was not always able to gauge accurately whether or not nutlets were fully mature before I collected them, and it was impractical to bag hundreds of tiny fruits. Hence, for most experiments, I used the number of nutlets initiated per flower at the time of collection as a measure of female reproductive success. Two pieces of evidence support this decision: (1) bagged flowers initiated very few nutlets (see Table 2 and [ILLUSTRATION FOR FIGURE 7 OMITTED]), indicating that nutlets usually do not expand unless the flower has been visited by a pollinator; and (2) data from the experiment conducted at the Lower Bonanza Site in 1994 (when I was most successful at collecting fruits after the seeds had matured) indicate that final seed production is linearly related to the number of nutlets initiated per fruit (linear regression equation for robbed and unrobbed treatments combined: mean number of seeds matured = 0.235 x number of nutlets initiated; [R.sup.2] = 0.991, P [less than] 0.001).

I found no evidence that nectar robbery reduced any of the correlates of female success measured at the scale of single flowers [ILLUSTRATION FOR FIGURE 3 OMITTED]. There were no significant differences at either site in either year in the percent of flowers initiating fruits, the number of nutlets initiated per successful flower, or the mass of seeds produced by robbed and unrobbed flowers (Table 1).

As a test of whether or not collars themselves influenced [TABULAR DATA FOR TABLE 1 OMITTED] female success, I performed the following "collar effects" experiment in 1994. One stem from each of 10 clones was enclosed in a mesh bag to prevent pollinator visitation, and three unopened flowers on each stem were randomly assigned to receive a full, half, or no collar. Once opened, all flowers were hand-pollinated using outcross pollen from a different, non-experimental clone each day. Hence, all flowers received an equivalent amount and quality of pollen, and differed only in the collar treatment. Three additional flowers without collars were left unpollinated in each bag to determine whether bluebell flowers can self-pollinate or require pollinator visitation to initiate fruits. The bags were removed once the corollas of all experimental flowers had abscised. The number of nutlets initiated per flower did not differ significantly among bagged hand-pollinated flowers with no, half, or full collars (Table 2; Kruskal-Wallis: [[Chi].sup.2] = 0.72, df = 2, P = 0.70). However, only 10.3% of bagged, unpollinated flowers initiated fruits (as compared with 86.2, 75.0, and 85.7% of flowers with no, half, and full collars, respectively), and the confidence interval for nutlets initiated per unpollinated flower overlapped zero (Table 2). Thus, typical rates of nutlet initiation appear to require pollinator visitation.

Whole-stern female success. - In the "collar effects" experiment, I protected single flowers on stems experiencing naturally high rates of nectar robbery. It is conceivable that an unrobbed flower on a stem whose flowers are mostly robbed would not show increased seed production, if legitimate pollinators quickly leave stems with depleted nectar. To test whether or not an effect of robbing would emerge at the whole-stem scale, I prevented or allowed robbing on entire stems from 12 bluebell clones at the Lower Bonanza Site in 1994. Two stems of similar height and diameter were chosen from each clone at the start of the flowering period. On one randomly chosen stem, half collars were placed on every flower immediately prior to anthesis; all flowers on the second stem received full collars. Stems were checked every 1-2 d throughout the flowering period, and all new flowers received collars as they prepared to open. Despite the pairing of similar-sized stems prior to the initiation of flowering, the two experimental stems from a clone often produced widely divergent numbers of flowers; hence it was not legitimate to compare the total number of fruits on paired stems. Instead, I looked for effects of robbery by comparing the average number of nutlets initiated per flower on the paired stems. There was no significant correlation between average nutlets per flower and number of flowers per stem in either treatment separately, or in both treatments combined.
TABLE 2. Number of nutlets initiated per bluebell flower in the
collar effects experiment.

           No collar    Half collar    Full collar    Not pollinated

Mean         2.31          2.04           2.04             0.21
95% CL   (1.77, 2.85)   (1.47, 2.61)   (1.53, 2.55)   (-0.04, 0.48)
n             29            28             28              29

The average number of nutlets initiated per flower in the whole-stem experiment varied widely, both among clones and within treatments [ILLUSTRATION FOR FIGURE 4 OMITTED]. On average, flowers on robbed stems initiated 1.56 nutlets per flower (95% CI = 1.21-1.92 nutlets) compared with 1.28 on unrobbed stems (95% CI = 0.99-1.58 nutlets), a difference that was not significant (Wilcoxon signed-ranks test, P = 0.17). Hence, the whole-stem experiment corroborated the lack of an effect of robbery on female success observed in the single-flower experiments.

Male reproductive success. - The experiments uncovered no evidence that bluebells suffer negative consequences, as measured by nutlet, production, due to nectar robbery. However, seed production was generally low in both experimental treatments [ILLUSTRATION FOR FIGURE 3 OMITTED], and even flowers that were hand-pollinated several times with outcrossed pollen initiated only about half of the four nutlets they are maximally capable of producing (Table 2). These data suggest that female success is limited more by resource availability than by pollen receipt; nevertheless, a reduction in legitimate visits could still reduce overall plant fitness by reducing the amount of pollen exported to other plants. To quantify the effect of robbery on male success, I measured pollen removal rates as a rough index of male fitness. Although the amount of pollen removed does not necessarily equate perfectly with paternal success, it provides an easily measurable component of male success in plants (for discussions of the advantages and limitations of measuring pollen removal, see Wilson and Thomson 1991, Snow and Lewis 1993).

At the Lower Bonanza site in 1994, four unopened flowers of similar age on each of three stems within each of 10 bluebell clones were marked and assigned haphazardly to one of four treatments. One flower was collected just before anthesis to provide an estimate of the initial amount of pollen per flower. Two experimental flowers received either a full collar ("unrobbed" flowers) or a half collar ("robbed" flowers), as in the female fitness experiment. A fourth ("unvisited") flower on each stem was completely enclosed in 2.5 cm wide dialysis tubing to prevent pollinator visitation. The rate of decline of pollen in unvisited flowers provided an estimate of the amount of pollen that may have simply fallen out of experimental flowers. All stems were censused each morning, and the experimental and unvisited flowers were collected at 1, 2, or 3 d after anthesis. As before, experimental pairs in which the appropriate collar failed to either prevent or allow robbing were discarded. All anthers from each flower were carefully removed, placed in open 1.5-mL microcentrifuge tubes for 24 h to allow the remaining pollen to dry, and preserved in 70% ethanol. The number of pollen grains remaining in each flower was estimated using a hemacytometer (each sample was homogenized on a vortex mixer for 60 s just before a drop of the solution was transferred to the hemacytometer; counts were averaged for two slides, on each of which pollen counts in eight subsamples (5 [[micro]liter] each) were combined). Pollen counts were performed with a digital image analysis system (Image 1).

The amount of pollen remaining in experimental flowers is illustrated in Fig. 5A. A preliminary ANOVA on In-transformed counts of pollen grains remaining in the three unopened flowers per clone indicated that clones did not differ in the initial amount of pollen per flower ([F.sub.9,20] = 1.48, P = 0.22); hence, clones were lumped across treatments in subsequent analyses. A regression of ln-transformed pollen against days since anthesis for unvisited flowers (i.e., unopened flowers, day 0, combined with flowers completely enclosed in dialysis tubing) indicated that pollen declined even in the absence of visitation, presumably because it fell from the dehisced anthers (regression equation: In(pollen) = 15.381 - 0.313 x day; [F.sub.1,58] = 47.2, P [less than] 0.001). Nevertheless, separate tests for days 1, 2, and 3 after anthesis showed that significantly less pollen remained in visited flowers (i.e., "robbed" and "unrobbed" flowers combined) than in unvisited ones (respective F statistics: [F.sub.1,28] = 16.50, P [less than] 0.001; [F.sub.1,26] = 54.23, P [less than] 0.001; [F.sub.1,28] = 69.20, P [less than] 0.001). I concluded that the decline in pollen in robbed and unrobbed flowers is due, at least in part, to removal by visitors.

I then compared pollen remaining in robbed and unrobbed flowers. Although one would expect more pollen grains to remain in robbed flowers if nectar robbery reduces visitation by legitimate pollinators, robbed flowers actually contained significantly fewer pollen grains 1 d after anthesis than did unrobbed flowers ([ILLUSTRATION FOR FIGURE 5A OMITTED]; [F.sub.1,18] = 5.53, P = 0.03). By days 2 and 3, the amount of remaining pollen did not differ ([F.sub.1,16] = 1.16, P = 0.30, and [F.sub.1,18] = 0.099, P = 0.76, respectively). Hence, these data lend no support to the hypothesis that nectar robbery results in either a delay in the timing or a reduction in the final magnitude of pollen removal.

Why doesn't nectar robbing reduce bluebell reproductive success?

In the experiments described, I found no evidence that nectar robbery reduced either the male or female success of bluebells. To determine why, I examined the behavior of individual bumble bees of the principal robbing species (B. mixtus and B. frigidus). I followed individuals for multiple visits, recording the type of flower visited (young pink flowers vs. older blue flowers) and the bee's behavior. Three types of behavior were recognized: (1) robbing (the bee never entered the mouth of the flower, but instead landed on the outside of the corolla and either chewed a new robbery hole or probed an existing hole); (2) legitimate visitation (the bee either entered the opened mouth of an older flower or hung upside-down from the tip of a newly opening, younger flower, usually while buzzing the flower to remove pollen (Buchmann 1983); or (3) both (after entering the mouth, the bee climbed to the outside of the flower to remove nectar through the pierced corolla tube). The striking result is that foragers adopted distinct behaviors when visiting pink vs. blue flowers [ILLUSTRATION FOR FIGURE 6 OMITTED]. At pink flowers, bees overwhelmingly paid legitimate visits to the mouth of the flower; in 98.7% of those visits, bees were observed vibrating the thorax to shake pollen from the dehiscing anthers onto their sterna, where it could then be transferred to the pollen baskets on the legs. Once flowers opened fully and turned blue, bees switched to robbing nectar (G test for independence of flower stage and bee behavior: [[Chi].sup.2] = 736.4, df = 2, P [less than] [10.sup.-5]). Because bees are likely to contact the exserted stigma while buzzing younger flowers for pollen [ILLUSTRATION FOR FIGURE 1A OMITTED], most flowers probably receive sufficient pollen to fertilize the four ovules before they experience nectar robbing. Similarly, the rapid removal of pollen [ILLUSTRATION FOR FIGURE 5A OMITTED] suggests that most of a flower's male success is secured before it is robbed.

The change in bumble bee foraging behavior is driven by the presentation of different rewards at different times in .the life of a flower. Unopened flowers are replete with pollen, but little pollen remains in the flower after 1 d [ILLUSTRATION FOR FIGURE 5A OMITTED]. Bluebell pollen grains do not stick together and, thus, drop freely from the anthers once they dehisce (W. F. Morris, personal observation). By buzzing pink flowers, bumble bees cause the anther sacs to break open, thus ensuring a large pollen reward. In contrast, pink flowers contain relatively little nectar. Calibrated capillary tubes were used to measure the nectar content of flowers from the male fitness experiment immediately after they were collected [ILLUSTRATION FOR FIGURE 5B OMITTED]. When protected from robbery, flowers contained substantially more nectar 1 d after opening than did unopened flowers (compare day 0 flowers to unvisited flowers on day 1 in [ILLUSTRATION FOR FIGURE 5B OMITTED]). A more detailed look at another collection of flowers showed nectar contents of 0.20 [+ or -] 0.15 [[micro]liter] before the stigma was exserted, 0.35 [+ or -] 0.18 [[micro]liter] after exsertion but before the corolla began to open, and 0.88 [+ or -] 0.17 [[micro]liter] in pink flowers that had begun to open (mean [+ or -] 1 SE, n = 9, 11, and 4, respectively), indicating that the onset of nectar production occurs at about the time of stigma exsertion. Bumble bees reduced nectar in robbed flowers to levels comparable to those in unopened flowers [ILLUSTRATION FOR FIGURE 5B OMITTED]. On average, flowers with full collars (accessible to legitimate pollinators) contained less nectar than did unvisited flowers [ILLUSTRATION FOR FIGURE 5B OMITTED], but this difference was not significant (P [greater than] 0.15 for main effects of day and treatment and for the interaction effect in a two-way ANOVA). Elevated nectar levels in unvisited flowers could be due to reduced evaporation of nectar from flowers completely enclosed in dialysis tubing, and not to the prevention of removal by legitimate visitors. Nevertheless, it is clear that most of the nectar removed from bluebell flowers is obtained by robbing.

Why are bluebells blue?

The absence of negative consequences of nectar robbery appears to be due to the fact that both male and female fitness are determined before flowers are robbed, at least at the study sites in 1993 and 1994. However, this raises an intriguing paradox: if flowers contribute to fitness primarily when they have not fully opened, what advantage does a plant gain by maintaining blue flowers for several days after pollination and provisioning them with a nectar reward that bumble bees obtain by larceny? Why doesn't the plant simply drop the corolla as soon as a flower is pollinated and save the energy and materials it invests in nectar production? I collected additional data to begin addressing two hypotheses that might explain the flowering pattern of bluebells.

First, blue flowers might contribute to plant fitness by acting as a form of insurance when pollinators are rare. Because summer weather in the Wrangell Mountains is variable, bumble bee visitation to bluebells may be lower in cold or rainy years than it was in 1993 and 1994, which were unusually warm and dry. When pollinators are rare, many or all bluebell flowers may not be visited during the brief pink phase, and the prolonged blue phase may provide extended opportunities for pollen receipt and removal. This hypothesis requires that blue flowers not receiving pollen until late in their flowering lifetimes actually be able to produce seeds. To test this, I bagged flowers early or late and compared their success to that of unbagged flowers. I marked two flowers that were about to open on each of four stems from 14 clones at the Pass Site in 1994. Each stem was randomly assigned to one of the following treatments. "Early" stems were bagged prior to anthesis and the bags were removed 2 d after the flowers opened; hence, flowers were only exposed to pollinators during the second half of their lifetimes. "Late" stems were bagged from 2 d after anthesis until the corollas had abscised. A third "control" stem on each clone was not bagged. The fourth stem was bagged both early and late as an additional test of whether or not bluebell flowers are capable of self-fertilization in the absence of pollinators.

As in previous experiments (see Table 2), flowers that were bagged for their entire lifetimes initiated significantly fewer nutlets than did unbagged flowers ([ILLUSTRATION FOR FIGURE 7 OMITTED]; Mann-Whitney U = 229.5, P = 0.001). Flowers that were bagged early initiated fewer nutlets, on average, than did unbagged flowers, but this difference was not significant (U = 355.0, P = 0.51). There was no difference between flowers bagged late and unbagged flowers (U = 404.4, P = 0.64). Although there may be a slight decline in seed production in flowers that receive pollen only late in their lifetime, such flowers are still capable of contributing to the plant's female fitness. This capability is a necessary requirement if blue flowers are to ensure seed-set during years of low pollinator visitation, but it is not sufficient to prove that blue flowers actually function in this manner. Experimental prevention of visits to blue flowers during years of low pollinator abundance would shed further light on this hypothesis. In addition, further experiments are needed to test whether or not pollen in blue flowers remains viable and, hence, capable of contributing to male fitness.

A second hypothesis posits that blue flowers might make an indirect contribution to fitness by attracting pollinators to pink flowers on the same plant. Sequential flower maturation assures that blue and pink flowers are usually present on the same stem. To measure the advertisement value of blue flowers, I compared visitation rates to pink flowers on stems with and without blue flowers. I chose paired stems with equal numbers of inflorescences on 25 bluebell clones. On one randomly chosen stem, I removed all blue flowers, leaving only the pinks; the other stem served as a control. The number of pink flowers on treatment and control stems did not differ (4.0 [+ or -] 0.47 vs. 4.4 [+ or -] 0.52, respectively, mean [+ or -] 1 SE; Wilcoxon signed-rank test: P = 0.43). I then recorded the number of legitimate visits to the pink flowers on each stem for a 15-min period. Pink flowers on stems that also displayed blue flowers received nearly double the visitation rate (0.048 [+ or -] 0.01 visits/min, mean [+ or -] 1 SE) of pink flowers on treatment stems (0.025 [+ or -] 0.009 visits/min; n = 25 trials; Wilcoxon: P = 0.006). Two non-exclusive processes could account for the enhanced visitation to pink flowers. First, the larger and showier blue flowers might simply act as a visual cue to the presence of nearby pink flowers; this effect would not require blue flowers to provide a nectar reward in order to contribute to plant fitness. Second, blue flowers might, because of the nectar reward they contain, attract robbers that then switch to buzz-pollinating pink flowers on the same clone. In fact, most individuals of B. mixtus and B. frigidus do alternate between visits to pink and blue flowers over the course of a single foraging bout [ILLUSTRATION FOR FIGURE 8 OMITTED]; only one of 62 bees I followed for [greater than] 10 visits failed to visit both pink and blue flowers. Bees visited 1.39 [+ or -] 0.12 flowers per stem (mean [+ or -] 1 SE), and most flights carried them to a neighboring stem. Because bluebell clones can produce dozens (or sometimes hundreds) of stems, robbers are highly likely to pollinate some pink flowers before leaving a clone.


Despite the expectation that nectar larceny should be detrimental to plant fitness, I found no evidence that robbing by bumble bees reduced the male or female success of bluebells, at least in the sites and years of my study. This result stems directly from the two-part behavior of bumble bees, which rob flowers that they have previously visited legitimately, thereby assuring both the fertilization of ovules and the export of pollen. Apparently, robbery does not cause reproductive tissue damage, which can reduce seed-set in other plants (Galen 1983). Although robbery exacted no measurable toll on plant reproduction, it is clear that part-time robbers (which account for [greater than]80% of bumble bee visits) provide a key benefit to bluebells; bagged flowers from which bumble bees were excluded initiated few fruits. Hence, from the plant's point of view, the bluebell-nectar robber mutualism appears to be strongly skewed toward benefit and away from detriment.

Because pollen removal is only a crude measure of male fitness, the absence of an effect of robbery on male success must be viewed as a preliminary conclusion. There are at least two reasons why male success of robbed and unrobbed flowers might differ even if their rates of pollen removal do not. First, an equivalent rate of pollen removal from robbed flowers could involve a decrease in pollen delivery to other plants and an increase in the amount of pollen dislodged from the flower. Second, robbery might alter the spatial distribution of those pollen grains that do arrive at the stigmas of other plants, causing concomitant changes in male success. For example, pollinators such as bumble bees that engage in area-restricted search (for example, see Pyke 1978) may move farther after visiting flowers that have been depleted by robbers. This might increase male (and female) fitness through the avoidance of local inbreeding effects (Waser and Price 1983, Zimmerman and Cook 1985). More detailed experiments that use genetic markers to determine the fate of removed pollen grains are currently underway.

Although observations and experiments suggest that successful pollination usually occurs when flowers are in the younger pink stage, older blue flowers do appear to contribute to reproductive success. The presence of blue corollas increases pollinator visitation to pink flowers on the same stem. Because bumble bees switch between robbing and buzz-pollinating over a short time span, the presentation of nectar rewards in blue flowers (perhaps in addition to their visual attractiveness) may be a key enticement to pollinators. Hence, nectar-producing blue flowers may act indirectly to enhance the reproductive success of the plant as a whole by increasing pollen receipt and export from pink flowers on the same stem or on nearby ramets of the same clone. Blue flowers also have the capacity to contribute directly to female reproductive success in years when pollinator services are scarce [ILLUSTRATION FOR FIGURE 7 OMITTED].

The change in corolla color as bluebell flowers age may function to signal a change in reward status, although it is not yet known if these bumble bees use flower color per se as a behavioral cue. Interestingly, a color change with age occurs in other species of Mertensia and in at least 11 other genera in the borage family (Weiss 1995). Visual cues (color or shape change) in bluebells may achieve the same functions that have been hypothesized for other plants. The retention of postchange flowers may enhance the long-distance attractiveness of a plant (Gori 1983, 1989, Cruzan et al. 1988, but see Casper and LaPine 1984, Delph and Lively 1989), but short-range pollinator discrimination against unrewarding flowers may increase pollen delivery to previously unvisited flowers (Gori 1983, Casper and LaPine 1984, Cruzan et al. 1988, Weiss 1991). In bluebells, mature flowers elicit robbing behavior from the principal pollinators; because bumble bees do not contact the stigmas of the flowers they rob, pollen deposited on their venters while buzz-pollinating other plants may still be available to effect fertilization of pink flowers (cf. Podolsky 1992).

Although nectar robbers have been shown to reduce plant reproductive success in other systems (Roubik 1982, Galen 1983, Roubik et al. 1985), this study shows that the existence of nectar robbery should not be used to argue in general that mutualisms are inherently prone to cheating. It is unclear whether or not the provision by bluebells of a robbed reward represents an evolved strategy to garner indirect benefits from robbers that legitimately visit other flowers on the plant. The nectar reward may have evolved in the past to attract legitimate pollinators, but may currently be maintained by selection because of its indirect contribution to reproductive success. In either case, both nectar production by bluebells and nectar robbery by bumble bees appear to be essential ingredients in their present-day mutualistic interaction.


Thanks to C. Christian, P. Chase, A. Denzer, G. Farley, and B. Hudgens for help in the field and lab, and to D. Doak, M. Rausher, N. Waser, and two anonymous reviewers for helpful suggestions. R. Thorp kindly identified the bumble bees. Research was supported by NSF grant BSR-9396119 and by a Duke University Arts and Sciences Research Council Faculty Research Award.


Addicott, J. F. 1986. Variation in the costs and benefits of mutualism: the interaction between yuccas and yucca moths. Oecologia 70:486-494.

Addicott, J. F., and A. J. Tyre. 1995. Cheating in an obligate mutualism: how often do yucca moths benefit yuccas? Oikos 72:382-394.

Bell, G. 1985. On the function of flowers. Proceedings of the Royal Society of London B 224:223-265.

Boucher, D. H., S. James, and K. H. Keeler. 1982. The ecology of mutualism. Annual Review of Ecology and Systematics 13:15-47.

Bronstein, J. L. 1991. Mutualism studies and studies of mutualism. Bulletin of the Ecological Society of America 72: 6-8.

-----. 1992. Mutualism, antagonism, and the fig-pollinator interaction. Ecology 69:1298-1302.

-----. 1994. Our current understanding of mutualism. Quarterly Review of Biology 69:31-51.

Broyles, S. B., and R. Wyatt. 1990. Paternity analysis in a natural population of Asclepias exaltata: multiple paternity, functional gender, and the "pollen-donation hypothesis." Evolution 44:1454-1468.

Buchmann, S. L. 1983. Buzz pollination in angiosperms. Pages 73-113 in C. E. Jones and R. J. Little, editors. Handbook of experimental pollination biology. Van Nostrand-Reinhold, New York, New York, USA.

Campbell, D. R. 1989. Inflorescence size: test of the male function hypothesis. American Journal of Botany 76:730-738.

Casper, B. B., and T. R. La Pine. 1984. Changes in corolla color and other floral characteristics in Cryptantha humilis (Boraginaceae): cues to discourage pollinators? Evolution 38:128-141.

Cruzan, M. G., P. R. Neal, and M. F. Willson. 1988. Floral display in Phyla incisa: consequences for male and female reproductive success. Evolution 42:505-515.

Delph, L. F., and C. M. Lively. 1989. The evolution of flora color change: pollinator attraction versus physiological constraints in Fuchsia excorticata. Evolution 43:1252-1262.

Fritz, R. S., and D. H. Morse. 1981. Nectar parasitism of Asclepias syriaca by ants: effect on nectar levels, pollinia insertion, pollinaria removal, and pod production. Oecologia 50:316-319.

Galen, C. 1983. The effects of nectar-thieving ants on seedset in floral scent morphs of Polemonium viscosum. Oikos 41:245-249.

Gentry, A. H. 1978. Anti-pollinators for mass-flowering plants? Biotropica 10:68-69.

Gori, D. F. 1983. Post-pollination phenomena and adaptive floral changes. Pages 31-49 in C. E. and R. J. Little, editors. Handbook of experimental pollination biology. Van Nostrand-Reinhold, New York, New York, USA.

-----. 1989. Floral color change in Lupinus argenteus (Fabaceae): why should plants advertise the location of unrewarding flowers to pollinators? Evolution 43:870-881.

Hawkins, R. P. 1961. Observations on the pollination of red clover by bees. Annals of Applied Biology 49:55-65.

Heinrich, B. 1976. Resource partitioning among some eu-social insects: bumblebees. Ecology 57:874-889.

Heinrich, B., and P. H. Raven. 1972. Energetics and pollination ecology. Science 176:597-602.

Higashi, S., M. Ohara, H. Arai, and K. Matsuo. 1988. Robber-like pollinators: overwintered queen bumblebees foraging on Corydalis ambigua. Ecological Entomology 13: 411-418.

Howe, H. F. 1984. Constraints on the evolution of mutualisms. American Naturalist 123:764-777.

Hulten, E. 1974. Flora of Alaska and neighboring territories. Stanford University Press, Stanford, California, USA.

Inouye, D. W. 1979. The terminology of floral larceny. Ecology 61:1251-1253.

-----. 1983. The ecology of nectar robbing. Pages 153173 in B. Bentley and T. Elias, editors. The biology of nectaries. Columbia University Press, New York, New York, USA.

Janzen, D. H. 1979. How many babies do figs pay for babies? Biotropica 11:48-50.

Keddy, P. 1990. Is mutualism really irrelevant to ecology? Bulletin of the Ecological Society of America 71:101-102.

Kodric-Brown, A., J. H. Brown, G. S. Byers, and D. F. Gori. 1984. Organization of a tropical island community of hummingbirds and flowers. Ecology 65:1358-1368.

Koeman-Kwak, M. 1973. The pollination of Pedicularis palustris by nectar thieves (short-tongued bumblebees). Acta Botanica Neerlandica 22:608-615.

Macior, L. W. 1966. Foraging behavior of Bombus (Hymenoptera: Apidae) in relation to Aquilegia pollination. American Journal of Botany 53:302-309.

McDade, L. A., and S. Kinsman. 1980. The impact of floral parasitism in two neotropical hummingbird-pollinated plant species. Evolution 34:944-958.

Pellmyr, O. 1989. The cost of mutualism: interactions between Trollius europaeus and its pollinating parasites. Oecologia 78:53-59.

Podolsky, R. D. 1992. Strange floral attractors: pollinator attraction and the evolution of plant sexual systems. Science 258:791-793.

Pyke, G. 1978. Optimal foraging: movement patterns of bumblebees between inflorescences. Journal of Theoretical Biology 13:72-98.

Queller, D.C. 1985. Proximate and ultimate causes of low fruit production in Asclepias exaltata. Oikos 44:373-381.

Risch, S. J., and D. Boucher. 1976. What ecologists look for. Bulletin of the Ecological Society of America 57:8-9.

Roubik, D. W. 1982. The ecological impact of nectar-robbing bees and pollinating hummingbirds on a tropical shrub. Ecology 63:354-360.

Roubik, D. W., N. M. Holbrook, and G. Parra V. 1985. Roles of nectar robbers in reproduction of the tropical treelet Quassia amara (Simaroubaceae). Oecologia 66:161-167.

Rust, R. W. 1979. Pollination of Impatiens capensis: pollinators and nectar robbers. Journal of the Kansas Entomological Society 52:297-308.

Schemske, D. W. 1980. Evolution of floral display in the orchid Brassavola nodosa. Evolution 34:489-493.

Schmid-Hempel, P., and B. Speiser. 1988. Effects of inflorescence size on pollination in Epilobium angustifolium. Oikos 53:98-104.

Scott, P. E., S. L. Buchmann, and M. K. O'Rourke. 1993. Evidence for mutualism between a flower-piercing carpenter bee and ocotillo: use of pollen and nectar by nesting bees. Ecological Entomology 18:234-240.

Snow, A. A., and P. O. Lewis. 1993. Reproductive traits and male fertility in plants: empirical approaches. Annual Review of Ecology and Systematics 24:331-351.

Stanton, M. L., A. A. Snow, and S. N. Handel. 1986. Floral evolution: attractiveness to pollinators increases male fitness. Science 233:1625-1626.

Stanton, M., H. J. Young, N. C. Ellstrand, and J. M. Clegg. 1991. Consequences of floral variation for male and female reproduction in experimental populations of wild radish, Raphanus sativus L. Evolution 45:268-280.

Tyre, A. J., and J. F. Addicott. 1993. Facultative non-mutualistic behaviour by an "obligate" mutualist: "cheating" by yucca moths. Oecologia 94:173-175.

Vandermeer, J. 1984. The evolution of mutualism. Pages 221-232 in B. Shorrocks, editor. Evolutionary ecology. Blackwell, Oxford, United Kingdom.

Waser, N. M. 1979. Pollinator availability as a determinant of flowering time in ocotillo (Fouquieria splendens). Oecologia 39:107-121.

Waser, N. M., and M. V. Price. 1983. Optimal and actual outcrossing in plants, and the nature of plant-pollinator interaction. Pages 342-359 in C. E. Jones and R. J. Little, editors. Handbook of experimental pollination biology. Van Nostrand-Reinhold, New York, New York, USA.

Weiss, M. R. 1991. Floral colour changes as cues for pollinators. Nature 354:227-229.

-----. 1995. Floral color change: a widespread functional convergence. American Journal of Botany 82:167-185.

Willson, M. L., and P. W. Price. 1977. The evolution of inflorescence size in Asclepias syriaca (Asclepiadaceae). Evolution 31:495-511.

Wilson, P., and J. D. Thomson. 1991. Heterogeneity among floral visitors leads to discordance between removal and deposition of pollen. Ecology 72:1503-1506.

Wyatt, R. 1980. The impact of nectar-robbing ants on the pollination system of Asclepias curassavica. Bulletin of the Torrey Botanical Club 107:313-321.

Young, A. M. 1983. Nectar and pollen robbing of Thunbergia grandiflora by Trigona bees in Costa Rica. Biotropica 15: 78-80.

Young, H. J., and M. L. Stanton. 1990. Influences of floral variation on pollen removal and seed production in wild radish. Ecology 71:536-547.

Zimmerman, M., and S. Cook. 1985. Pollinator foraging, experimental nectar-robbing, and plant fitness in Impatiens capensis. American Midland Naturalist 113: 84-91.
COPYRIGHT 1996 Ecological Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1996 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Morris, William F.
Date:Jul 1, 1996
Previous Article:Fire severity and vegetation response in the boreal Swedish forest.
Next Article:Mechanisms of hummingbird-mediated selection for flower width in Ipomopsis aggregata.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters