Nectar-robbing bumble bees reduce the fitness of Ipomopsis Aggregata (Polemoniaceae).
Floral larceny, or nectar robbing, is common among animal-pollinated plants (Inouye 1983). However, in comparison to the hundreds, if not thousands, of studies on the interactions between plants and their pollinators (for reviews see Faegri and van der Pijl 1979, Jones and Little 1983, Real 1983, Proctor et al. 1996), the interactions between plants and nectar robbers have received relatively little attention from pollination ecologists. Nevertheless, the interactions between plants and nectar robbers may be important to both the ecology and evolution of flowering plants. Moreover, nectar robbers may affect not only their host plants but also the legitimate pollinators of those hosts by depriving them of nectar. Due to a general paucity of information and lack of experimental studies, the importance of nectar robbing to plants and their mutualistic pollinators remains a mystery.
Most studies of nectar robbing have focused on how robbing affects female plant fitness (e.g., Macior 1966, Koeman-Kwak 1973, Rust 1979, Roubik 1982, Galen 1983, Newton and Hill 1983, Roubik et al. 1985, Reddy et al. 1992, Morris 1996, Olesen 1996). Nectar robbers may decrease fruit or seed set by damaging plant reproductive structures (McDade and Kinsman 1980), by deterring legitimate pollinators indirectly through their robbing behavior (Reddy et al. 1992), or by deterring pollinators directly through physical contact (Roubik 1982). On the other hand, robbers may enhance female fitness components if they effect some pollination themselves or if they lower geitonogamy by causing pollinators to visit fewer flowers on the same plant (Heinrich and Raven 1972, Zimmerman and Cook 1985, de Jong et al. 1992). Therefore, robbers can have both negative and positive effects on female fitness components.
Although the effects of nectar robbing on female fitness components have received some attention, the effects of robbing on male fitness are virtually unknown. To our knowledge, male fitness as a function of robbing has been examined in only three studies (Wyatt 1980, Fritz and Morse 1981, Morris 1996). Morris (1996) found that significantly more pollen was removed from robbed flowers one day after anthesis in the bumble bee-robbed Mertensia paniculata as compared to unrobbed flowers. Wyatt (1980) found equivocal results of the effects of pollinia removal from ant-robbed Asclepias curassavica. In one site, pollinia removal was higher in robbed plants; however, in a second site pollinia removal was higher in unrobbed plants. Fritz and Morse (1981) found no effect of nectar-robbing ants on pollinia removal in flowers of Asclepias syriaca. Each of these studies estimated male fitness as the removal of pollen from focal plants. In Asclepias, pollinia removal is related to the number of seeds sired (Broyles and Wyatt 1990). However, a positive relationship between pollen grains removed from anthers, viable pollen grains deposited on stigmas, and male success in fertilizing ovules, may not hold true for all plants (Stanton et al. 1992, Klinkhamer et al. 1994). Therefore, pollen removal may not be an accurate measure of male reproductive success.
In many of the systems studied to date, nectar robbers also act as pollinators and enhance the fitness of their host plants (e.g., Meidell 1944, Macior 1966, Schremmer 1972, Koeman-Kwak 1973, Waser 1979, Higashi et al. 1988, Morris 1996). Nectar robbers may increase plant fitness in one of two ways. Some robbers, such as bumble bees, need both pollen and nectar to meet their energetic and nutritional requirements (Heinrich 1979). Therefore, depending on the reward they are seeking, robbers will either bite a hole through the corolla to obtain nectar or will function as legitimate pollinators while collecting pollen (e.g., Meidell 1944, Macior 1966, Koeman-Kwak 1973, Morris 1996). Nectar robbers may also act as pollinators if their body parts brush up against the reproductive organs of the flowers they rob (e.g., Higashi et al. 1988). In the system we studied, however, nectar-robbing bumble bees do not pollinate the flowers they rob (R. E. Irwin and A. K. Brody, unpublished data).
Here we describe experimental and observational studies in which we examined the effects of nectar robbing by Bombus occidentalis on both male- and female-fitness components of the hummingbird-pollinated plant, Ipomopsis aggregata. Specifically, we asked: (1) Does nectar robbing affect the male fitness of I. aggregata on a per-plant or per-flower basis and (2) does nectar robbing affect the female fitness of I. aggregata on a per-plant or per-flower basis? We measured female fitness as (1) pollen receipt, (2) percent fruit set, (3) seed set per fruit, and (4) seed mass. Using florescent dyes as pollen analogs, we estimated pollen donation, rather than pollen removal, as a measure of male function in nectar robbed vs. unrobbed plants. We examined the effects of nectar robbing on both a per-flower and a per-plant basis because the effects of robbing might be seen at either of these hierarchical levels, and a knowledge of the effects at both levels provides strength to their interpretation.
We studied the montane perennial herb, Ipomopsis aggregata subspecies aggregata (Polemoniaceae; Grant and Wilken 1986), in meadows near the Rocky Mountain Biological Laboratory (RMBL), Gothic, Gunnison County, Colorado, USA Ipomopsis aggregata is self-incompatible and does not reproduce vegetatively (Waser and Price 1983). Therefore, all seeds are a result of inter-plant pollen transfer. In addition, I. aggregata is monocarpic so estimates of lifetime fitness can be obtained in a single flowering season (Waser and Price 1989). Plants usually begin flowering in mid-June and continue through mid-August. Each plant flowers for [approximately]6 wk, producing a single inflorescence with numerous, red, trumpet-shaped flowers (Waser 1978). The hermaphroditic flowers are protandrous with no disposition for flowers in different phases (staminate or pistillate) to occur on different parts of the plant or in different parts of the season (Pyke 1978a). Nectaries are located at the base of the flowers, and nectar is produced at a rate of 1-5 [[micro]liter] of nectar[center dot][flower.sup.-1][center dot][d.sup.-1], with a concentration of 20-25% sucrose equivalents (Pleasants 1983). The corolla usually abscises after 3-5 d.
The red, tubular flowers of 1. aggregata are pollinated primarily by Broad-tailed, Selasphorus platycercus, and Rufous, Selasphorus rufus, Hummingbirds near the RMBL (Waser 1978). Ipomopsis aggregata is robbed by the bumble bee, Bombus occidentalis. Bombus occidentalis uses its sharp, toothed mandibles to chew a hole through the side of the corolla. It inserts its proboscis into the hole and removes all available nectar without damaging the plant's reproductive or nectar-producing structures (Pleasants 1983, Irwin and Brody 1998) but does not pollinate the flowers (R. E. Irwin and A. K. Brody, unpublished data). Therefore, a nectar-robber visit to a flower does not result in pollination but reduces the standing crop of nectar to zero. On a per-plant basis over the entire season, robbing rates range from [approximately]0% to 100% of available flowers robbed, with a seasonal mean ([+ or -] 1 SD) of 51.3% [+ or -] 38.3% (Irwin and Brody 1998).
We examined the effects of natural and artificial nectar robbing on both male- and female-fitness components of I. aggregata during the summers of 1995 and 1996. To assess the effects of nectar robbing on male fitness, we tracked dye particles (used as pollen analogs) from artificially and naturally robbed plants and estimated pollen (dye) donation to other plants in local populations (Waser and Price 1982, Waser 1988, Campbell 1989a, b). To assess the effects of nectar robbing on female fitness, we quantified pollen receipt using natural pollen loads, subsequent fruit and seed production, and seed mass from both naturally and artificially robbed plants. In addition, we verified that artificial robbing adequately mimicked natural robbing in its effects on both male and female function by comparing results from naturally and experimentally robbed plants. We examined estimates of both male and female fitness per plant for each week of flowering to account for variation in hummingbird foraging behavior and its effects on plant fitness over the season.
Effects of nectar robbing on measures of whole-plant fitness
We assessed the effects of nectar robbing on dye donation and pollen receipt in six populations (three each in 1995 and 1996) and the effects of nectar robbing on fruit and seed production and seed mass in an additional 14 populations (six in 1995 and eight in 1996). Different populations of I. aggregata were used in each of the two years of study. Each population consisted of [approximately]200 flowering I. aggregata within a 10[m.sup.2] area. Populations were separated from each other by distances [greater than]1 km. Ninety-five percent of pollen is expected to stay within populations of these areas (Waser and Price 1983, Campbell 1991a).
For each population in 1995, we chose five focal plants that were morphologically similar in corolla length, corolla width, stalk diameter, plant height, and stigma and anther length. In 1995, two of the focal plants were randomly assigned to artificial robbing treatments: one received a low level of nectar robbing (10% of available flowers robbed) and one a high level of nectar robbing (80% of available flowers robbed). These levels simulate observed rates of average low and high natural robbing of 15 randomly sampled I. aggregata that we found in the 1995 field season around the RMBL town site (Irwin and Brody 1998), and they simulate reported rates of low and high robbing in I. aggregata studied in other years (Pleasants 1983). The remaining three plants in each population were left unmanipulated and received natural nectar robbing. In 1995, there were no significant differences in male and female fitness components among plants that were robbed naturally vs. artificially (see Results: Comparison of natural and artificial nectar robbing below). Therefore, in 1996 we used artificially robbed plants only. This allowed us to accurately control the levels of robbing each plant experienced and to increase replicates of low and high robbing treatments within sites. In 1996, we chose six focal plants and randomly assigned three to the low robbing treatment and three to the high robbing treatment. All other aspects of the design remained the same as in 1995.
We artificially robbed flowers by making a small hole in the side of the corolla near the base of the flower with a pair of dissecting scissors. We then removed nectar by inserting a micro-capillary tube into the hole and drawing out all available nectar. All newly robbed flowers were either in elongated bud or male phase. If a flower had been robbed previously, it was robbed again by extracting all available nectar through the existing hole each day that the robbed flower was in bloom. Flowers not artificially robbed were protected from natural robbing by placing a piece of translucent plastic drinking straw over the entire tubular corolla. We examined the effect of this "collaring" manipulation in a separate population and found no effect of collaring on pollen (dye) donation, pollen receipt, or fruit or seed production (dye donation: [F.sub.1,60] = 0.11, P = 0.75; pollen receipt: [F.sub.1,24] = 2.42, P = 0.13; percentage fruit set: [F.sub.1,38] = 0.19, P = 0.66; seeds/fruit: [F.sub.1,23] = 0.02, P = 0.97). By artificially robbing and protecting flowers within plants, we maintained an average rate of robbing of 10% and 80% for the low and high robbing treatments, respectively, over the entire season.
We estimated pollen movement within three populations in each year by applying powdered fluorescent dyes (Series JST-300, Radiant Color, Richmond, California) to the dehiscing anthers of flowers on focal plants. The dyes act as pollen analogs, and dye and pollen are transferred at similar rates by hummingbirds foraging at I. aggregata flowers. Mean dye donation over multiple flowers gives an accurate estimate of mean pollen donation (Waser and Price 1982) and therefore of male function. We could reliably distinguish among six colors of dye without using a UV lamp (blue, chartreuse, deep green, deep red, sunset orange, and light pink). In 1995, we used five focal plants per population, three assigned to receive natural levels of robbing and one each assigned to the low and high robbing treatments. In 1996, we used a maximum of six focal plants per population, three assigned to the low robbing treatment and three to the high treatment. The dyes were used to track pollen donation from the focal plants near the center of each population to other plants (hereafter referred to as "outer plants"). In each population, we estimated pollen donation to outer plants (estimate of male fitness) and natural pollen receipt (estimate of female fitness) for each focal plant. Because focal plants donated as well as received dye, we also measured the effects of nectar robbing on fruit and seed production and seed mass in an additional 14 populations where no dye was applied because dye can interfere with the fertilization of ovules (for a review see Kearns and Inouye 1993).
We began dye application in each population as soon as all focal plants came into bloom and continued until all focal plants had ceased blooming. We randomly assigned the focal plants in each population to one of the six different dye colors at the beginning of the flowering season. Each week, we applied dye every day for four consecutive days. We applied dye particles with a fiat-head toothpick to all of the anthers of the flowers on focal plants in staminate phase. On each of the four days, we artificially robbed flowers according to their assigned treatments and recorded: (1) the number of flowers in bloom on each focal plant, (2) each flower's phase, staminate or pistillate, and (3) the percentage of flowers naturally robbed for those plants in the natural nectar-robbing treatment in 1995. We robbed flowers before applying dye to anthers to prevent dislodging dye from the flowers. In addition, we marked the calyx of all open flowers with a dot of nontoxic paint to distinguish both the week those flowers were in bloom (a different color of paint was used for each week) as well as their robbing status (robbed vs. unrobbed).
Effects of nectar robbing on male fitness
Each afternoon after dye application, we randomly selected 15 outer plants and collected stigmas from one-half of their pistillate-phase flowers. To estimate male fitness, we quantified the amount of dye donated to outer plants by counting the number of dye particles of each color on each stigma. To standardize for daily differences in the number of flowers dyed per focal plant, we divided the average dye particle donation per plant by the number of flowers dyed. Dye donation was normalized using the square-root transformation. We employed a mixed-model repeated-measures ANCOVA using nectar-robbing treatment as a fixed effect, year, site, and plant nested within treatment as random effects, week as the repeated measure, and the number of open flowers per plant as a covariate. The treatment effect was tested with respect to the plant nested within treatment term. All other factors were tested over the residual. Treatment means were compared using a posteriori Tukey-Kramer HD means tests. Dye donation per plant was examined as the response variable. For naturally robbed plants in 1995, we performed a regression of average dye donation per plant on average robbing per plant to determine the relationship between natural levels of nectar robbing and dye donation.
Effects of nectar robbing on female fitness
Pollen receipt. - We collected all of the stigmas from focal plants after the corollas abscised. Collecting stigmas at this stage has no effect on pollen tube germination or fruit set (Waser and Fugate 1986). Using a dissecting microscope, we counted the number of pollen grains on each stigma. We estimated pollen receipt as the average number of pollen grains received per focal plant per week.
Fruit and seed production. - We collected all fruits from the focal plants in each study population. We recorded the number of fruits that did and did not set seed. Seeds of each fruit were counted and weighed together. We then calculated three components of female function: percentage fruit set per week, average seeds per fruit per week, and average seed mass per fruit per week. Percentage fruit set was calculated as the number of expanded fruits divided by the total number of flowers produced per week. We computed seeds per fruit as the average number of seeds produced per successful fruit per week. Fruits were collected in individual packets in 1995 so seed mass was determined on a per-fruit basis in 1995 only.
Analysis. - Pollen receipt was normalized with a square-root transformation. We normalized percentage fruit set and seeds per fruit using the arcsine square-root and natural-log transformations, respectively. We again employed a mixed-model repeated-measures ANOVA to examine the effects of robbing on pollen receipt per plant as well as fruit and seed production and seed mass per plant. Nectar-robbing treatment was included as a fixed effect; year, site, and plant nested within treatment as random effects, and week as the repeated measure. We used the number of open flowers as a covariate in the model where pollen receipt was the response variable. In all models, the treatment effect was tested with respect to the plant nested within treatment term. All other factors were tested over the residual. We measured seed mass in 1995 only; therefore, year was not included in the analysis for this variable. Treatment means were compared using a posteriori Tukey-Kramer HD means tests. For naturally robbed plants in 1995, we performed separate regressions of the average of each response variable per plant on average robbing per plant to determine the relationship between natural levels of nectar robbing and pollen receipt, percentage fruit set, seed set per fruit, and seed mass.
Comparison of artificial and natural nectar-robbing treatments
To examine if plants respond in a similar way to natural and artificial nectar robbing, we compared plants with similar natural and artificial robbing levels for both male and female fitness components in 1995. Because nectar robbing is variable among I. aggregata plants (Irwin and Brody 1998), and because we had no a priori knowledge of the amount of natural robbing plants would receive, we measured the percentage of natural robbing in three plants per site, suspecting that at least some of those plants would receive similar levels of natural robbing as our artificially robbed plants. To determine if artificial high robbing accurately mimicked natural high robbing, we compared artificially robbed plants to naturally robbed plants that had a seasonal mean [greater than or equal to]75% of their flowers naturally robbed. For low robbing, we compared artificially robbed plants to naturally robbed plants that had a seasonal mean of [less than or equal to]15% of their available flowers naturally robbed. We used a mixed-model repeated-measures ANOVA with treatment (artificial and natural robbing) as a fixed effect, plant nested within treatment and site as random effects, and week as the repeated factor. The treatment effect was tested with respect to the plant nested within treatment term. All other factors were tested over the residual. We analyzed dye donation, pollen receipt, percentage fruit set, seeds per fruit, and seed mass as response variables.
Effects of nectar robbing on measures of individual-flower success
Nectar robbing might have no effect on whole-plant measures of male or female function if pollinators do not discriminate between robbed and unrobbed flowers on the same plant. However, nectar robbing could still affect per-flower success, especially if unrobbed flowers within plants compensate for those that are robbed. Therefore, both of these nested hierarchical levels, per-flower and per-plant, may be important for interpreting the effects of nectar robbing on I. aggregata. Thus, we also examined the effects of nectar robbing on dye donation and pollen receipt on a per-flower basis in 1995.
To do so, we chose four morphologically similar focal plants at the center of one population. On each day that the plants were in bloom, we robbed 50% of the staminate-phase flowers using the techniques described above and collared the other 50% of the flowers to prevent robbing by B. occidentalis. We dyed the anthers of all robbed flowers one dye color and the anthers of all unrobbed flowers a second dye color. The same two colors were used on all four plants to distinguish dye donation by robbed vs. unrobbed flowers.
Each afternoon after dye application, we randomly selected 15 outer plants and collected stigmas from one-half of their pistillate-phase flowers. We counted the number of dye particles of each color on each stigma and quantified male fitness as the average number of dye particles donated from robbed and unrobbed flowers to outer plants on a per-flower basis. To standardize for daily differences in the number of robbed and unrobbed flowers dyed, we divided the average dye particle donation per nectar-robbing treatment by the total number of flowers dyed per treatment. We also sampled, in a similar fashion, all of the stigmas from focal plants 24 h after they became receptive. From these samples, we determined the average number of pollen grains received per flower per nectar-robbing treatment in a 24-h period. We examined dye donation by robbed and unrobbed flowers with an ANOVA with treatment (robbed and unrobbed flowers) as a fixed effect and date as a random effect. We examined pollen receipt by robbed and unrobbed flowers using a mixed-model repeated-measures ANOVA with robbing treatment (robbed and unrobbed flowers) as a fixed effect, plant nested within floral treatment as a random effect, and date as the repeated measure. Again, the treatment effect was tested with respect to the nested term.
In the overall models for the repeated-measures ANOVAs and ANCOVAs, we examined each component of plant reproductive success on a per-plant basis for variation among years and sites. For all response variables except pollen receipt (which we discuss separately below), year was insignificant in the overall model (P [greater than] 0.05 in all cases), and for all variables (except percentage fruit set and seed set per fruit), site was insignificant in the overall model. Therefore, in nonsignificant cases, we pooled data across years and/or sites. In the repeated-measures ANOVA and ANCOVA models with the appropriate pooled data, the effect of plant identity was insignificant for percentage fruit set and seed set per fruit (P [greater than] 0.05 in both cases). However, plant identity was significant for dye donation ([F.sub.22,67] = 1.95, P = 0.02) and seed mass ([F.sub.16,50] = 2.67, P = 0.004). Neither week nor the covariate (number of open flowers) had a significant effect on dye donation; however, estimates of female fitness - percentage fruit set, seeds per fruit, and average seed mass - did vary among weeks. For both the low and high robbing treatments, these estimates of female fitness decreased as the season progressed.
In 1995, the average percentage of natural nectar robbing per plant in plants in the natural robbing treatment ranged from 0% to 97.2% of available flowers robbed, with a mean ([+ or -] 1 SD) of 32.9% [+ or -] 37.2% (N = 27 plants).
Comparison of artificial and natural nectar-robbing treatments
For sites in which we measured dye donation and pollen receipt in 1995, two of the three sites contained plants with high levels of natural robbing ([greater than]75% of available flowers robbed; N = 3 plants), and two of the three sites contained plants with low levels of natural robbing ([less than]15% of available flowers robbed; N = three plants). And for sites in which we measured percentage fruit set, seed set per fruit, and seed mass, four of the nine sites contained plants with high levels of natural robbing (N = 7 plants), and five of the nine sites contained plants with low levels of natural robbing (N = 9 plants). based on the range of variation in natural nectar robbing among plants, we did not expect that all naturally robbed plants from all nine sites in 1995 would conform to the low (0-15%) or high (75-100%) robbing rates throughout the entire season. We found no differences between artificial and natural nectar robbing for any of the response variables measured at low and high robbing rates (Table 1). Therefore, our low and high robbing treatments accurately represent natural low and high robbing in terms of plant response.
Effects of nectar robbing on male fitness
Plants in the high robbing treatment donated significantly fewer dye particles than plants in the low robbing treatment ([F.sub.1,22] = 26.08, P [less than] 0.0001; [ILLUSTRATION FOR FIGURE 1 OMITTED]). In addition, there was a significant negative relationship between average percentage robbing and average dye donation for naturally robbed plants in 1995 (r = 0.64, N = 9 plants, P = 0.05). On a per-flower basis, robbed flowers donated significantly fewer dye particles than unrobbed flowers over the 1995 flowering season ([F.sub.1,13] = 12.88, P = 0.003). Mean dye donation (per flower dyed) by unrobbed flowers was over twice that of robbed flowers (mean [+ or -] 1 SE, unrobbed flowers: 0.09 [+ or -] 0.01 pollen grains donated; robbed flowers: 0.04 [+ or -] 0.01 pollen grains donated).
Effects of nectar robbing on female fitness
Pollen receipt. - Plants in the high robbing treatment received significantly fewer pollen grains than plants in the low robbing treatment in 1996 but not in 1995 (Table 2; [ILLUSTRATION FOR FIGURE 2 OMITTED]). Pollen receipt was significantly lower in 1995 (mean [+ or -] 1 SE = 65.61 [+ or -] 2.46 pollen grains) than in 1996 (mean [+ or -] 1 SE = 100.51 [+ or -] 2.49 pollen grains; [F.sub.1,79] = 8.44, P = 0.02). In the overall models for 1995 and 1996, site was insignificant (1995: [F.sub.2,17] = 1.60, P = 0.23; 1996: [F.sub.2,59] = 2.82, P = 0.07), so [ILLUSTRATION FOR FIGURE 1 OMITTED] data were combined from all sites within years. The effects of plant nested within treatment and week on pollen receipt varied between years (Table 2). On a per-flower basis in 1995, artificially unrobbed and robbed flowers received similar numbers of pollen grains (mean [+ or -] 1 SE, unrobbed flowers = 87.83 [+ or -] 5.88 pollen grains, robbed flowers = 77.32 [+ or -] 5.01 pollen grains; [F.sub.2,4] = 0.68, P = 0.45). For naturally robbed plants in 1995, there was a significant negative relationship between average percentage robbing and average pollen receipt (r = 0.71, N = 9 plants, P = 0.03).
Fruit set. - Percentage fruit set per week was significantly lower in sites that received fluorescent dyes (mean [+ or -] 1 SE, dyed sites = 46.34% [+ or -] 2.31%, undyed sites = 52.57% [+ or -] 1.42%; [F.sub.1,611] = 5.29, P = 0.04). Therefore, we only analyzed data from sites that did not receive dye treatments. We found no significant difference in percentage fruit set in the low vs. high artificial robbing treatments (mean [+ or -] 1 SE, low robbing = 54.77% [+ or -] 1.73%, high robbing = 47.16% [+ or -] 1.68%; [F.sub.1,4] = 0.18, P = 0.69). However, for naturally robbed plants in 1995 we did find a significant negative relationship between average percentage fruit set and average percentage robbing per plant (r = 0.55, N = 18 plants, P = 0.02).
Seed set per fruit. - Seed set per fruit did not differ significantly between sites in which fluorescent dyes were and were not applied ([F.sub.1,326] = 0.41, P = 0.52); therefore, we analyzed data from all sites. Seed set per fruit was significantly lower in plants that received the high robbing treatment ([F.sub.1,4] = 8.14, P = 0.04; [ILLUSTRATION FOR FIGURE 3 OMITTED]). In addition, plants in the high robbing treatment produced significantly fewer seeds per plant than plants in the low robbing treatment ([F.sub.1,4] = 16.08, P = 0.01; [ILLUSTRATION FOR FIGURE 4 OMITTED]). However, for naturally robbed plants in 1995, the negative relationship between average percentage nectar robbing and average seed set per fruit was not significant (r = 0.04, N = 27 plants, P = 0.82).
Seed mass. - Seed mass did not differ among sites in which fluorescent dyes were and were not applied ([F.sub.1,71] = 1.11, P = 0.29) so we analyzed our data across all sites. Seed mass did not differ significantly between the nectar-robbing treatments (mean [+ or -] 1 SE, low robbing = 0.83 [+ or -] 0.06 mg, high robbing = 1.01 [+ or -] 0.07 mg; [F.sub.1,16] = 1.70, P = 0.21). In addition, for the plants that received natural levels of robbing in 1995, the negative relationship between average percentage nectar robbing and average seed mass per plant was not significant (r = 0.14, N = 27 plants, P = 0.49).
TABLE 2. Repeated-measures ANCOVAs for the effects of treatment (low vs. high robbing), plant nested within robbing treatment, week, and the covariate (number of open flowers) on pollen receipt in 1995 and 1996. 1995 1996 Source of variation df F P df F P Robbing treatment 1 0.57 0.48 1 13.87 0.001 Plant (treatment) 4 1.15 0.37 16 2.05 0.03 Week 3 0.91 0.46 3 3.01 0.04 No. open flowers 1 0.09 0.77 1 0.74 0.40 Error 14 48 Notes: The treatment effect was tested with respect to the plant nested within treatment term. All other factors were tested over the residual.
Nectar robbing decreased both male and female fitness components of I. aggregata. Donation of dye to neighboring plants, a measure of male fitness, was significantly reduced in plants with high robbing. Assuming no differences in pollen tube growth or fertilization of ovules, lower levels of pollen donation will result in lower numbers of offspring sired. Pollen receipt was also significantly reduced in plants with high nectar robbing in 1996. Because I. aggregata is often pollen limited, decreased numbers of pollen grains deposited on stigmas will likely result in decreased seeds per fruit (Hainsworth et al. 1985, Campbell 1991b, Campbell and Halama 1993). Plants with high robbing also produced fewer seeds per fruit as well as fewer total seeds per plant.
How does nectar robbing reduce the male and female reproductive success of I. aggregata? In other systems, three major mechanisms have been identified or proposed: (1) Nectar robbers may damage plant reproductive structures (McDade and Kinsman 1980). (2) Robbers may aggressively interact with pollinators and deter them from visiting plants (Roubik 1982). (3) Nectar robbers may reduce the attractiveness of robbed plants and flowers to pollinators and, thus, indirectly cause a reduction in pollinator per flower visitation rate (Reddy et al. 1992). A fourth possibility exists as well. Nectar robbers may decrease the effectiveness (rather than the visitation rate) of pollinators by decreasing the frequency with which pollinators contact the anthers and stigmas of robbed flowers.
There is no evidence that nectar robbing by B. occidentalis damages the reproductive structures of I. aggregata. We found no macroscopic or microscopic damage to the stigmas, styles, or ovaries in nectar-robbed flowers (Irwin and Brody 1998). Nor did we find any evidence that B. occidentalis directly interferes with the foraging behavior of hummingbird pollinators. Furthermore, it is unlikely that robbing affects pollination success by reducing pollinator effectiveness. Mitchell and Waser (1992) varied nectar standing crop in I. aggregata flowers and found no detectable effect on single-visit pollination effectiveness estimated through pollen removal (anther contact) and pollen receipt (stigma contact). Because the immediate effect of robbing is a reduction in nectar standing crop, we expect that similar results would hold for our study as well.
The most plausible mechanism underlying the reduction in male and female fitness components by nectar robbers is that robbing decreases the attractiveness and profitability of robbed I. aggregata to pollinators. Hummingbirds avoid plants with high robbing and visit fewer flowers on those plants (Irwin and Brody 1998). Other studies have also highlighted hummingbird sensitivity to resource levels among plants (i.e., Pyke 1978b, Gass and Montgomerie 1981, Mitchell 1993). Ipomopsis aggregata is self-incompatible, and reproduction is often pollinator-limited (Hainsworth et al. 1985, Campbell 1991b, Campbell and Halama 1993). Therefore, reductions in visitation rate due to hummingbird avoidance of nectar-robbed plants are likely to decrease both male and female fitness components.
Examining the results of nectar robbing at different hierarchical levels (i.e., whole-plant vs. per-flower) would have been critically important had we found no whole-plant difference in high vs. low nectar-robbing treatments. Nectar robbing could have affected per-flower success but could have been masked if unrobbed flowers within plants compensated for those that were robbed. However, in 1995 we found that robbing decreased dye donation at the plant and flower levels and had no significant effect on pollen receipt at either the plant or flower levels. Thus, the patterns at the flower level did not elucidate further patterns or mechanisms unapparent at the whole-plant level.
Despite the strength of the relationship between high robbing and reduced plant fitness components, we found two apparently anomalous results, each of which can be explained in the context of the known natural history of I. aggregata. First, the high nectar-robbing treatment decreased seed set per fruit and seed set per plant but not percentage fruit set. Other studies of I. aggregata have also shown that reductions in pollinator visitation reduce seed set but not fruit set (Brody 1992, Campbell and Halama 1993). Some fruit set is effected by pollinators other than hummingbirds (Pleasants and Waser 1985, Brody 1992), but a reduction in visitation by hummingbirds can have a dramatic effect on seed set per fruit and whole-plant seed production even when fruit set is unaffected. Second, the importance of year, site, and week effects were not consistent across all response variables. For example, percentage fruit set and seed set per fruit differed significantly among sites and weeks. And the importance of robbing on pollen receipt varied among years. Plant reproductive success may vary among years, sites, and weeks for many reasons including variation in pollinator abundance as well as variation in the degree to which plants are pollinator- vs. resource-limited (e.g., Waser and Real 1979, Brody and Mitchell 1997). If plants are pollinator-limited, variation in pollinator population densities and foraging behaviors due to migration patterns, breeding activity by females, territorial aggression by males, and weather (e.g., Cody 1968, Stiles 1971, Pimm 1978, Gass and Montgomerie 1981) may have a significant impact on their reproductive success. Furthermore, one might expect that not all plant fitness components respond in the same degree to the same selective pressures (e.g., Stanton et. al. 1986, Stanton and Preston 1987, Cruzan et al. 1988, Mitchell 1993).
Despite these vagaries, the strength of our study lies in having taken an experimental approach and thus being able to separate the effects of robbing per se from other factors with which it may have been correlated. Few studies have manipulated robbing and examined subsequent plant fitness (but see Wyatt 1980, Fritz and Morse 1981, Rust 1979, Zimmerman and Cook 1985, Morris 1996). Studies relying on natural variation in nectar robbing rather than experimental manipulations cannot determine whether robbing was a direct determinant of plant fitness or merely correlated with it (e.g., Macior 1966, Koeman-Kwak 1973, Kendall and Smith 1976). From our study, we can conclude that nectar robbing affects reproductive success in I. aggregata, and does so through its indirect effects on pollinators (Irwin and Brody 1998).
Studies examining the effects of nectar robbing on both male and female reproductive success in a single plant species are rare (Wyatt 1980, Fritz and Morse 1981, Morris 1996). In hermaphroditic plants, nectar robbing may affect both male and female fitness. Therefore, both estimates should be taken into account when measuring ultimate reproductive consequences. Although we did not measure male reproductive success directly (number of seeds sired), powdered dyes provided an accurate measure of relative success in pollen delivery. There is a significant and positive correlation between dye and pollen receipt to stigmas (Waser and Price 1982; R. E. Irwin and A. K. Brody, unpublished data). Nonetheless, pollen donation is itself only an estimate of male fitness. An increase in delivery of pollen to stigmas may or may not increase seeds sired. We are currently conducting a paternity analysis using genetic markers to determine the effect of nectar robbing on the actual number of seeds sired by plants.
Nectar robbers can have any number of effects on plant fitness from negative to positive. Our study gives the first piece of conclusive evidence that nectar robbers can negatively effect male as well as female fitness components. We found no evidence that particular traits (i.e., toxic nectar or tough corolla tubes) of I. aggregata prevent nectar robbing per se, but such is not the case in other nectar-robbed plant species (e.g., Barrows 1976, Stephenson 1982, Wootton and Sun 1989, Proctor et al. 1996). However, robbing does appear to be nonrandom, and robbers may impose selection on floral phenology and morphology. Therefore, nectar robbing may still be an important evolutionary force for I. aggregata. Plants that produce flowers later in the season and flowers with longer corollas are more likely to be heavily robbed (R. E. Irwin and A. K. Brody, unpublished data). Our next step is to examine the degree to which nectar robbing affects the evolution of floral traits as well as the population ecology of nectar-robbed I. aggregata.
We thank S. Armbruster, A. Courant, N. Gotelli, G. Lowenberg, W. Morris, J. Olesen, A. Seidl, N. Waser, and an anonymous reviewer for helpful comments and suggestions that improved the manuscript. We are also grateful for the facilities of the Rocky Mountain Biological Laboratory. This work was supported by NSF grants DEB-9596178 and DEB9806501.
Barrows, E. M. 1976. Nectar robbing and pollination of Lantana camara (Verbenaceae). Biotropica 8:132-135.
Brody, A. K. 1992. Oviposition choices by a pre-dispersal seed predator (Hylemya sp.): II. A positive association between female choice and fruit set. Oecologia 91:63-67.
Brody, A. K., and R. J. Mitchell. 1997. Effects of experimental manipulation of inflorescence size on pollination and pre-dispersal seed predation in the hummingbird-pollinated plant Ipomopsis aggregata. Oecologia 110:86-93.
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.
Campbell, D. R. 1989a. Inflorescence size: test of the male function hypothesis. American Journal of Botany 76:730-738.
-----. 1989b. Measurements of selection in a hermaphroditic plant: variation in male and female pollination success. Evolution 43:318-334.
-----. 1991a. Comparing pollen dispersal and gene flow in a natural population. Evolution 45:1965-1968.
-----. 1991b. Effects of floral traits on sequential components of fitness in Ipomopsis aggregata. American Naturalist 137:713-737.
Campbell, D. R., and K. J. Halama. 1993. Resource and pollen limitations to lifetime seed production in a natural plant population. Ecology 74:1043-1051.
Cody, M. L. 1968. Interspecific territoriality among hummingbird species. Condor 70:270-271.
Cruzan, M. B., P. R. Neal, and M. F. Willson. 1988. Floral display in Phyla incisa: Consequences for male and female reproductive success. Evolution 42:505-515.
de Jong, T. J., N. M. Waser, M. V. Price, and R. M. Ring. 1992. Plant size, geitonogamy and seed set in Ipomopsis aggregata. Oecologia 89:310-315.
Faegri, K., and L. van der Pijl. 1979. The principles of pollination ecology. Third edition. Pergamon, Oxford, UK.
Fritz, R. S., and D. H. Morse. 1981. Nectar parasitism of Asclepias syriaca by ants: effects on nectar levels, pollinia insertion, pollinaria removal and pod production. Oecologia 50:316-319.
Galen, C. 1983. The effect of nectar-thieving ants on seedset in floral scent morphs of Polemonium viscosum. Oikos 41: 245-249.
Gass, C. L., and R. D. Montgomerie. 1981. Hummingbird foraging behavior: decision-making and energy regulation. Pages 159-194 in A. C. Kamil and T Sargent, editors. Foraging behavior: ecological, ethological, and psychological approaches. Garland Press, New York, New York, USA.
Grant, V., and D. H. Wilken. 1986. Taxonomy of the Ipomopsis aggregata group (Polemoniaceae). Botanical Gazette 147:359-371.
Hainsworth, F. R., L. L. Wolf, and T. Mercier. 1985. Pollen limitation in a monocarpic species, Ipomopsis aggregata. Journal of Ecology 73:263-270.
Heinrich, B. 1979. Bumble bee economics. Harvard University Press, Cambridge, Massachusetts, USA.
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 bumble bees foraging on Corydalis ambigua. Ecological Entomology 13:411-418.
Inouye, D. W. 1983. The ecology of nectar robbing. Pages 153-173 in B. Bentley and T. Elias, editors. The biology of nectaries. Columbia University Press, New York, New York, USA.
Irwin, R. E., and A. K. Brody. 1998. Nectar robbing in Ipomopsis aggregata: effects on pollinator behavior and plant fitness. Oecologia, in press.
Jones, C. E., and R. J. Little. 1983. Handbook of experimental pollination biology. Van Nostrand Reinhold, New York, New York, USA.
Kearns, C. A., and D. W. Inouye. 1993. Techniques for pollination biologists. University Press of Colorado, Niwot, Colorado, USA.
Kendall, D. A., and B. D. Smith. 1976. The pollinating efficiency of honeybee and bumble bee visits to flowers of the runner bean Phaseolus coccineus L. Journal of Applied Ecology 13:749-752.
Klinkhamer, P. G. L., T. J. de Jong, and J. A. J. Metz. 1994. Why plants can be too attractive - a discussion of measures to estimate male fitness. Journal of Ecology 82:191-194.
Koeman-Kwak, M. 1973. Pollination of Pedicularis palustris by nectar thieves (short-tongued bumble bees). 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.
Meidell, O. 1944. Notes on the pollination of Melampyrum pratense and the "honeystealing" of humble-bees and bees. Bergens Museums Arbok 11:5-11.
Mitchell, R. J. 1993. Adaptive significance of Ipomopsis aggregata nectar production: observation and experimentation in the field. Evolution 47:25-35.
Mitchell, R. J., and N. M. Waser. 1992. Adaptive significance of Ipomopsis aggregata nectar production: pollination success of single flowers. Ecology 73:633-638.
Morris, W. F. 1996. Mutualism denied? Nectar-robbing bumble bees do not reduce female or male fitness of bluebells. Ecology 77:1451-1462.
Newton, S. D., and G. D. Hill. 1983. Robbing of field bean flowers by the short-tongued bumble bee Bombus terrestris L. Journal of Agricultural Research 22:124-129.
Olesen, J. M. 1996. From naivete to experience: bumble bee queens (Bombus terrestris) foraging on Corydalis cava (Fumariaceae). Journal of the Kansas Entomological Society 69:274-286.
Pimm, S. L. 1978. An experimental approach to the effects of predictability on community structure. American Zoologist 18:797-808.
Pleasants, J. M. 1983. Nectar production in Ipomopsis aggregata (Polemoniaceae). American Journal of Botany 70:1468-1475.
Pleasants, J. M., and N. M. Waser. 1985. Bumble bee foraging at a "hummingbird" flower: reward economics and floral choice. American Midland Naturalist 114:283-291.
Proctor, M., P. Yeo, and A. Lack. 1996. The natural history of pollination. Timber Press, Portland, Oregon, USA.
Pyke, G. H. 1978a. Optimal foraging in bumble bees and coevolution with their plants. Oecologia 36:627-640.
-----. 1978b. Optimal foraging in hummingbirds: testing the marginal value theorem. American Zoologist 18:739-752.
Real, L. 1983. Pollination biology. Academic Press, London, UK.
Reddy, T. B., K. Rangaiah, E. U. B. Reddi, and C. S. Reddi. 1992. Consequences of nectar robbing in the pollination ecology of Vitex-negundo (Verbenaceae). Current Science 62:690-691.
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. V. Parra. 1985. Roles of nectar robbers in the 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.
Schremmer, F. 1972. Der Stechsaugrussel, der Nektarraub, das Pollensammlung, und der Blutenbesuch der Holzbienen (Xylocopa) (Hymenoptera, Apidae). Zeitschrift fur Morphologie der Tiere 72:263-294.
Stanton, M. L., T. Ashman, L. E Galloway, and H. J. Young. 1992. Estimating male fitness of plants in natural populations. Pages 62-90 in R. Wyatt, editor. Ecology and evolution of plant reproduction: new approaches. Chapman and Hall, New York, New York, USA.
Stanton, M. L., and R. E. Preston. 1987. The reproductive biology of petal color variants in wild populations of Raphanus sativus L.: I. Pollinator response to color morphs. American Journal of Botany 74:178-187.
Stanton, M. L., A. A. Snow, and S. N. Handel. 1986. Floral evolution: attractiveness to pollinators increases male fitness. Science 232:1625-1627.
Stephenson, A. G. 1982. Iridoid glycosides in the nectar of Catalpa speciosa are unpalatable to nectar thieves. Journal of Chemical Ecology 8:1025-1034.
Stiles, F. G. 1971. Time, energy, and territoriality of the Anna hummingbird (Calypte anna). Science 173:818-821.
Waser, N. M. 1978. Competition for hummingbird pollination and sequential flowering in two co-occurring wildflowers. Ecology 59:934-944.
-----. 1979. Pollinator availability as a determinant of flowering time in Ocotillo (Fouquieria splendens). Oecologia 39:107-121.
-----. 1988. Comparative pollen and dye transfer by natural pollinators of Delphinium nelsonii. Functional Ecology 2:41-48.
Waser, N. M., and M. L. Fugate. 1986. Pollen precedence and stigma closure: a mechanism of competition for pollination between Delphinium nelsonii and Ipomopsis aggregata. Oecologia 70:573-577.
Waser, N. M., and M. V. Price. 1982. A comparison of pollen and fluorescent dye carry-over by natural pollinators of Ipomopsis aggregata (Polemoniaceae). Ecology 63:1168-1172.
Waser, N. M., and M. V. Price. 1983. Optimal and actual outcrossing, and the nature of plant-animal interactions. Pages 341-359 in C. E. Jones and R. J. Little, editors. Handbook of experimental pollination biology. Van Nostrand Reinhold, New York, New York, USA.
Waser, N. M., and M. V. Price. 1989. Optimal outcrossing in Ipomopsis aggregata: seed set and offspring fitness. Evolution 43:1097-1109.
Waser, N. M., and L. A. Real. 1979. Effective mutualism between sequentially flowering plant species. Nature 281:670-672.
Wootton, T. J., and I-F Sun. 1989. Bract liquid as a herbivore defense mechanism for Heliconia wagneriana inflorescences. Biotropica 22:155-159.
Wyatt, R. 1980. The impact of nectar-robbing ants on the pollination system of Aselepias curassavica. Bulletin of the Torrey Botanical Club 107:313-321.
Zimmerman, M., and S. Cook. 1985. Pollinator foraging, experimental nectar-robbing and plant fitness in Impatiens capensis. American Midland Naturalist 113:81-84.
|Printer friendly Cite/link Email Feedback|
|Author:||Irwin, Rebecca E.; Brody, Alison K.|
|Date:||Jul 1, 1999|
|Previous Article:||Modeling the spread of pine wilt disease caused by nematodes with pine sawyers as vector.|
|Next Article:||Induced responses to herbivory in wild radish: effects on several herbivores and plant fitness.|