Oecologia (2001) 128:263–273 DOI 10.1007/s004420100640 Carlos Lara · Juan Francisco Ornelas Preferential nectar robbing of flowers with long corollas: experimental studies of two hummingbird species visiting three plant species Received: 29 September 2000 / Accepted: 28 December 2000 / Published online: 28 February 2001 © Springer-Verlag 2001 Abstract Long flower tubes have been traditionally viewed as the result of coevolution between plants and specialized, legitimate, long billed-pollinators. However, nectar robbers may have played a role in selection acting on corolla length. This study evaluated whether hummingbirds are more likely to rob flowers with longer corollas from which they cannot efficiently extract nectar with legitimate visits. We compared two hummingbird species with similar bill lengths (Lampornis amethystinus and Colibri thalassinus) visiting floral arrays of artificial flowers with exaggerated corolla lengths, and also evaluated how the birds extract nectar rewards from medium to long corollas of three hummingbird-pollinated plants (Salvia mexicana, S. iodantha and Ipomoea hederifolia). The consequences of foraging for plant fitness were evaluated in terms of seed production per flower. Variation in seed production after legitimate visits of hummingbird-pollinated plants was mostly explained by differences in pollinator effectiveness. Seed production did not increase with the number of legitimate visits to a flower, except in I. hederifolia. We found that birds were more likely to rob both artificial and natural flowers with long corolla tubes. Nectar robbing was not observed on short-corolla flowers of Salvia spp., but robbing negatively affected seed production of long-tubed flowers of I. hederifolia. Significant differences between hummingbird species in the use of this behavior were observed, but males and females behaved alike. We suggest that short-billed hummingbirds with enlarged bill serrations (the edge of both tomia finely toothed) may have an advantage in illegitimately feeding at long-corolla flowers. This raises the possibility of counter-selection on increasing corolla length by nectar robbers. C. Lara · J.F. Ornelas (✉) Departamento de Ecología y Comportamiento Animal, Instituto de Ecología, A.C., Apdo. Postal 63, Xalapa, 91000 Veracruz, México e-mail: [email protected] Tel.: +52-28-421800 ext. 4125, Fax: +52-28-187809 Keywords Colibri thalassinus · Lampornis amethystinus · Nectar robbing · Corolla length · Seed production Introduction Long floral tubes (corollas) have been traditionally viewed as a floral adaptation for pollination by longtongued and long-billed pollinators (Darwin 1859, 1877; Feinsinger and Colwell 1978; Stiles 1981; Feinsinger 1983). In this view, longer corollas are favored because pollinators remove more pollen grains from them (Darwin 1862; Wolf et al. 1976; Nilsson 1988; Fenster 1991). However, this situation may be complicated by visitors that extract nectar by piercing floral tissues without contacting the anther and stigma (nectar robbing; Feinsinger et al. 1987; Ornelas 1994; Navarro 1999; Maloof and Inouye 2000). Short-billed hummingbirds often obtain the nectar of flowers by making perforations at the base of the corolla tube or using a hole already made by insects and birds (e.g., Beal 1880; Ridgway 1890; Grant 1952; Skutch 1954, 1973; Kodric-Brown and Brown 1979; Inouye 1983; Kodric-Brown et al. 1984; Feinsinger et al. 1987; Ornelas 1994, 1998; Navarro 1999; but see Colwell et al. 1974). Insofar as birds preferentially rob flowers with longer corollas, from which they cannot efficiently extract nectar through the corolla mouth, they may have played a role as a selective force countering corolla elongation. Other floral traits such as chemical deterrents (Guerrant and Fiedler 1981), dilute nectar (Bolten and Feinsinger 1978), protective floral bracts (Inouye 1983; Ornelas 1996), thickening of floral tissue (Inouye 1983; McDade 1984; Ornelas 1998), shifts from diurnal to nocturnal nectar production (Haber and Frankie 1982), and alliances with protective ants (Becerra and Venable 1989; Oliveira et al. 1999) may also be the result of selection by nectarrobbing species. Nectar robbing by insects and passerines occurs commonly on flowers adapted for hummingbird pollination 264 (Barrows 1976; McDade and Kinsman 1980; Willmer and Corbet 1981; Inouye 1983; Arizmendi et al. 1996; Bittrich and Amaral 1996; Irwin and Brody 1998, 1999; Navarro 1999; Maloof and Inouye 2000). In particular, long-tubed hummingbird flowers are exploited regularly by various robbers that significantly reduce the floral nectar available to pollinators (e.g., McDade and Kinsman 1980; Snow 1981; Pleasants 1983; Arizmendi et al. 1996; Irwin and Brody 1998; Traveset et al. 1998; Navarro 1999; Lara and Ornelas 2001). Skutch (1954) stated that nectar robbing is quite a rare behavior among tropical hummingbirds, but species of Aglaiocercus, Anthracothorax, Chalybura, Chlorostilbon, Chrysolampis, Colibri, Eulampis, Eupherusa, Heliothryx, Thalurania, and Trochilus are consistently reported to act as nectar robbers (ffrench 1973; Skutch 1973; Feinsinger et al. 1979; McDade and Kinsman 1980; Snow and Snow 1980; Snow 1981; Berry 1982; Ruschi 1982; Kodric-Brown et al. 1984; McDade 1984; Stiles 1985; Feinsinger et al. 1987; Gill 1987; Ornelas 1994, 1995; Navarro 1999). Yet, the behavior, ecology and evolution of nectar robbing by hummingbirds has rarely been studied (Colwell 1973; McDade and Kinsman 1980; Feinsinger et al. 1987; Ornelas 1994; Navarro 1999). Hummingbirds may extract nectar illegitimately by (1) actively piercing the base of typically long corollas and leaving a slit in the floral tissue (primary nectar robbers); (2) using the hole left in flowers already pierced by other species (secondary nectar robbers); (3) visiting flowers morphologically adapted for pollination by insects or bats (nectar thieves); and (4) by reaching between the petals of polypetalous flowers (nectar thieves) (Ornelas 1994, 1995). Ornelas (1994) found that 28 genera (23% of hummingbird genera) have species with enlarged tomial serrations (a variable number of serrations in the terminal portion of the maxillary and mandibular tomia). He hypothesized that enlarged bill serrations may enhance the exploitation of resources such as long-tubed corollas and tough-tissue corollas by facilitating access to protected nectaries, the grasping of tough and waxy blooms, and the cutting of flower tissue. Serrate, shortbilled hummingbirds pierce the corolla tubes near the base to extract nectar, leaving a conspicuous slit in the floral tissues (Ornelas 1994; Navarro 1999). These robbing behaviors have been categorized based on how hummingbirds handle flowers to extract nectar (Inouye 1983), however, the dichotomy between legitimate and illegitimate visitation is more useful when trying to document facultative nectar robbing. Of particular interest is the observation that a hummingbird can behave as either a legitimate pollinator or a robber of the same plant species (e.g., McDade and Kinsman 1980; Soberón and Martínez del Rio 1985; Ornelas 1994), or a legitimate pollinator of one plant species and a robber of another (McDade and Kinsman 1980; Soberón and Martínez del Rio 1981; Inouye 1983). The outcomes of presumably mutualistic interactions between a plant and its diverse array of floral visitors are conditioned by the degree of morphological matching between the flower and visitor, but also by the individual behavior of the pollinator, in particular, the mode, sequence and frequency of visitation to single flowers. The aim of our study was to determine how individuals of two hummingbird species with relatively short bills behave on flowers with contrasting corolla lengths. We asked to what extent natural and experimental variation in corolla length affects (1) the frequency of robbing visits to artificial and natural floral arrays and (2) the behavior of individual hummingbirds during each visit. We also evaluated seed production of experimental plants. We predicted that robbing should be more common on longer corollas and would negatively affect seed production. To our knowledge, this is the first experimental study of facultative nectar robbing and its causes and consequences. Materials and methods Study site Fieldwork was conducted from November to March 1992 and from July to October 1993 at the Laboratorio Natural Las Joyas field station (19°35′–19°37′N, 103°15′–104°37′W; at 1592 m a.s.l.; Arizmendi et al. 1996). This 1245-ha preserve is located in the Sierra de Manantlán Biosphere Reserve, in the Mexican states of Jalisco and Colima. Mean annual precipitation is 1610 mm (Jardel 1991), most of it falling between June and October when hurricanes occur. A short dry season extends from March to May. Mean annual temperature is 14.6°C, with freezing temperatures occurring only for a few days during the winter (November to February; Arizmendi et al. 1996). The reserve has a heterogeneous topography (Jardel 1991), and the vegetation is a mosaic of wet coniferous, pine-oak, and fragments of cloud forest along ravines, and secondary vegetation (Vázquez et al. 1995). Hummingbird species Our study focused on 2 of the 21 hummingbird species recorded at Las Joyas (Ornelas and Arizmendi 1995), the amethyst hummingbird (Lampornis amethystinus) and green violet-ear (Colibri thalassinus), a resident and an altitudinal migrant in the study area, respectively. Amethyst hummingbirds have straight, smooth, medium-sized bills (mean±SD=20.72±0.10 mm, n=48 males; mean±SD=22.30±0.10 mm, n=41 females; Ornelas 1995). Green violet-ears have slightly curved, medium-sized bills with enlarged bill serrations (see Ornelas 1994) (mean±SD=21.02±0.11 mm, n=17 males; mean±SD=20.90±0.12 mm, n=19 females; Ornelas 1995). Green violet-ears nest at the field station at the end of the rainy season and during this time feed mostly on the introduced bee-pollinated runner bean, Phaseolus coccineus (Leguminoseae). Amethyst hummingbirds breed early in the rainy season; during this time, they feed mostly from flowers of Ipomoea hederifolia (Convolvulaceae) and Crusea coccinea (Rubiaceae) (Ornelas 1995; Ornelas and Arizmendi 1995). Both species feed from flowers that range in corolla length from 10 to 33 mm at our study site (Arizmendi 1994; J.F. Ornelas, unpublished work). Field procedures A total of 120 adult hummingbirds were captured in the field for this study. Before trials, hummingbirds were housed individually for 1–2 days in field-collapsible cages (61×61×61 cm). These cages were placed in a room with ambient light and temperature. 265 shown in Fig. 1. Original vegetation was not removed, to give a more natural setting for the bird being tested. A perch was placed inside the aviary 3 m away from the observer. Hummingbirds were released one by one into the aviary to visit the artificial flowers for 1 h and then removed. We noted for each hummingbird (1) the type of visit (legitimate or illegitimate), and (2) the number of times it probed each type of flower. Because plastic flowers cannot be pierced, probing illegitimately at the base of artificial flowers was considered as a robbing attempt (hereafter robbing visits). Observations were conducted from 0800 to 1200 hours and we used hummingbirds as we netted them. Nectar robbing on natural flowers Fig. 1 Diagram showing the arrangement of artificial flowers. Four groups of long (58 mm) and four groups of short (33 mm) artificial flowers were placed alternately in groups of three flowers as illustrated, with the corolla opening either facing up, down, or horizontally Hummingbirds had free access to 20% (by mass) sugar solution, and ≈20 live Drosophila flies introduced to the cage two or three times a day. Most hummingbirds acclimated to captivity within 1 day (Ornelas 1995; Lara and Ornelas 1998). We found no effect of pre-trial housing on the hummingbird's foraging performance. Nectar robbing on artificial flowers We first looked for effects of corolla length and flower angle on nectar robbing behavior using arrays of artificial flowers. We hypothesized that robbing should be more likely on longer corollas when legitimate visitation takes a greater effort for the rewards that can be obtained. For this part of our study 15 individuals of each hummingbird species (8 males and 7 females) were used. We manipulated corolla length using conical plastic micropipet tips of 33 and 58 mm length and simulated “petals” with red plastic material (Fig. 1). The lengths of corollas correspond to the range of long-tubed, hummingbird-pollinated flowers normally encountered at Manantlán (Arizmendi 1994; J.F. Ornelas, unpublished work). Each flower was filled with c. 20 µl of 20% (by mass) sugar solution (Ornelas 1995) based on the 2–18 µl nectar production range known for the species studied (Arizmendi et al. 1996; J.F. Ornelas, unpublished work). Artificial flowers were placed inside a portable, outdoor aviary (4×8×2 m). The flowers were hung along a wire across the aviary at 1.7 m height and at a distance of 5 m from the observer. The whole floral array consisted of eight groups with three artificial flowers each (4 groups with 33-mm flowers and 4 groups with 58-mm flowers) 20 cm apart from each other. To manipulate flower angle within each size group, one flower was attached with the corolla opening facing up (0°), one flower facing down (180°), and the third facing horizontally (90°). To minimize the effect of treatment position, we placed the floral arrays so as to alternate groups of small and large flowers as We chose the most common hummingbird plant species during the study period with contrasting corolla lengths that range from 18 to 30 mm (Lara 1995) to test whether nectar robbing is facultative on corolla length. Preliminary observations in the study area showed that flowers of Salvia mexicana L., S. iodantha Fernald (Labiatae), and Ipomoea hederifolia L. (Convolvulaceae) are visited by C. thalassinus and L. amethystinus (Arizmendi 1994; J.F. Ornelas, unpublished work). S. mexicana and S. iodantha are commonly found in successional areas of burned pine forest and less frequently in patches of cloud forest (Arizmendi 1994). Their flowers are visited by 12 hummingbird species (C. thalassinus, Hylocharis leucotis, Amazilia beryllina, A. rutila, A. violiceps, L. amethystinus, Eugenes fulgens, Tilmatura dupontii, Atthis heloisa, Stellula calliope, Selasphorus platycercus, and S. rufus; Arizmendi 1994; J.F. Ornelas, personal observations) and commonly pierced by passerine Diglossa baritula (Arizmendi et al. 1996). Arizmendi (1994) reported that more than 50% of Salvia flowers are pierced by D. baritula under natural conditions. While it is not known which hummingbird species are the most effective pollinators of these Salvia species, short-billed hummingbirds (e.g., Selasphorus spp.) are likely candidates, since they are very abundant during winter flowering and pollen is deposited abundantly on their foreheads during a flower visit (Arizmendi 1994). Flowers of S. mexicana, with relatively short purple corolla tubes (mean±SD=18.7±0.02 mm, n=40; Lara 1995), are hermaphroditic, self-compatible, and last 4 days (Arizmendi et al. 1996). Mean numbers of seeds produced by flowers range from 0.1±0.4 (SD) seeds by automatic self-pollination, 1.4±1.3 (SD) after a single visit by the long-billed, least efficient pollinator (Eugenes fulgens), to 2.5±1.2 (SD) by the short-billed, most efficient pollinator (Hylocharis leucotis) (Arizmendi et al. 1996). Flowers of S. iodantha, with medium-sized maroon corolla tubes (mean±SD=21.3±0.23 mm, n =30; Lara 1995), are also hermaphroditic, self-compatible, and also last 4 days (M.C. Arizmendi, personal communication). Both species bloom from December to March (J.F. Ornelas, unpublished work). I. hederifolia is a woody vine commonly growing in pine and pine-oak forests (Arizmendi 1994). Its solitary, 1-day flowers (Chemas-Jaramillo 1995) are visited by several hummingbird species but mostly by Eugenes fulgens and L. amethystinus, and several species of butterflies (Chemas-Jaramillo 1995; J.F. Ornelas, personal observations); Trigona bees and D. baritula frequently rob its flowers (Arizmendi 1994; Chemas-Jaramillo 1995). Arizmendi (1994) reported that less than 30% of its flowers are pierced by D. baritula under natural conditions. While it is not known which hummingbird species are the most effective pollinators of this species, long-billed E. fulgens are likely candidates because pollen is deposited abundantly on their foreheads during a flower visit (Arizmendi 1994). Flowers are hermaphroditic and self-incompatible (Chemas-Jaramillo 1995) with long orangeyellow corolla tubes (mean±SD=30.2±0.32 mm, n=30; Lara 1995). Blooming occurs asynchronously year-round, but peaks from July to September (J.F. Ornelas, unpublished work). The remaining 90 captured hummingbirds were used in this part of the experiment. We presented 15 hummingbirds of each species (8 males and 7 females) with each of the three selected 266 plant species. The same aviary was used as an enclosure and placed over naturally growing clumps of flowering plants to ensure hummingbird visitation. Clumps were formed by about eight plants of one of the plant species displaying numerous developed buds and open flowers. Salvia flowers last >1 day, so it was necessary to have newly opened flowers and flowers >1 day old in our experimental sample to ensure outcrossed pollination. Inside the aviary, flowers on plants were bagged and unbagged with bridal netting, as needed, to control hummingbird visitation. Vegetation surrounding plants was removed to facilitate observations. To begin behavioral observations, we unbagged 30 open flowers at anthesis and then introduced a hummingbird. Only one hummingbird at a time was released into the aviary. All observations were conducted from 0800 to 1200 hours when hummingbirds are more active foraging and nectar production is high in these plants (Arizmendi et al. 1996). For each individual hummingbird we observed (1) whether its floral visits on each flower were legitimate (flower visited by its entrance), illegitimate (flower probed at the base), or robbed (flower actively pierced by any way), and (2) how often it visited each flower within a foraging bout. Hummingbirds typically visited all 30 flowers at least once during a session of behavioral observations, however, some flowers were revisited and fewer not visited at all. Behavioral observations ended before anthesis of the 30 open flowers was finished. Then we released the hummingbird, inspected the flowers for any damage, and then bagged them until seed production. Lastly, a group of 30 non-manipulated flowers (unbagged and outside the aviary) of each plant species was monitored until seed set to estimate seed production under open conditions of pollination. All statistical analyses were run using General Linear Modeling with StatView and SuperANOVA (Abacus Concepts 1989, 1996). Results Foraging behavior on artificial flowers Both hummingbird species made the same number of legitimate visits to artificial flowers, independent of corolla length and flower angle, except that individuals of L. amethystinus made more legitimate visits on average than C. thalassinus to short flowers facing up and to long flowers facing horizontally (Table 1, Fig. 2). As predicted, the number of robbing visits was statistically higher among long corolla tubes (58 mm; Table 1) and C. thalassinus made significantly more robbing visits than L. amethystinus, independent of flower angle (Fig. 2). Statistical analyses Artificial flowers We used two-way ANOVAs to evaluate variation in the number and type of visits to artificial flowers by individuals of the two hummingbird species (Zar 1984). In the model, hummingbird species and sex, and their interaction were fixed main factors, and the number of legitimate and robbing visits were dependent variables. Next, a three-way ANOVA (Zar 1984) with fixed factors of hummingbird species, corolla length and flower angle, and all interactions, was used to explore responses to features of the artificial flowers. Post hoc mean comparisons (Games-Howell procedure) were conducted to examine differences between hummingbird species in visiting (legitimately and robbing attempts) short and long artificial flowers. Natural flowers The same two-way ANOVA model was used to evaluate variation in number of visits to flowers of S. mexicana, S. iodantha, and I. hederifolia, and the effects of hummingbird species and number of visits on seed production variation of flowers of these species receiving legitimate visits. Because some flowers of all plant species ended up being revisited by hummingbirds, it was necessary to control for these differences in the analysis, as they might otherwise obscure the effect that floral visitors have on seed production. We analyzed the relationship between seed production and hummingbird visitation with ANCOVA (Zar 1984). In the model, hummingbird species was a fixed factor, number of visits to a flower was a covariate, and seed production was the dependent variable. The same analysis was used for legitimate and legitimate with robbing or illegitimate probes. Robbing visits were not analyzed this way because of small sample sizes. Post hoc mean comparisons (GamesHowell procedure) were conducted to examine the data for differences in seed production among flowers visited legitimately, legitimately visited and then robbed or illegitimately probed, and only robbed by the two hummingbird species. Fig. 2 Variation in number of a legitimate and b robbing visits to short and long artificial flowers offered at various angles to individuals of Colibri thalassinus and Lampornis amethystinus. Asterisks indicate significance after post hoc mean comparisons (**P=0.001, ***P=0.0001). The two hummingbird species differed significantly in number of visits to artificial flowers (legitimate visits F1,176=13.45, P<0.001, robbing visits F1,176=13.88, P<0.001), but males and females did not differ from each other (legitimate visits F1,176=0.85, P>0.05, robbing visits F1,176=0.80, P>0.05). In all cases the species×sex interaction was not significant (legitimate visits F1,176=1.33, P>0.05; robbing visits F1,176=0.25, P>0.05) 267 Table 1 Results of three-way ANOVAs of number of visits to artificial flowers as a function of hummingbird species, corolla length, and flower angle for legitimate visits and robbing visits Source Legitimate visits Hummingbird species Corolla length Flower angle Hummingbird species×Corolla length Hummingbird species×Flower angle Corolla length×Flower angle Hummingbird species×Corolla length×Flower angle Residual Robbing visits Hummingbird species Corolla length Flower angle Hummingbird species×Corolla length Hummingbird species×Flower angle Corolla length×Flower angle Hummingbird species×Corolla length×Flower angle Residual Fig. 3a–c Variation in number of visits to flowers of Salvia mexicana, S. iodantha, and Ipomoea hederifolia by individuals of C. thalassinus and L. amethystinus. Asterisks indicate significance after one-way ANOVAs (*P=0.01, **P=0.001, ***P=0.0001). No differences in the number of visits between males and fe-males were observed for S. mexicana (Table 2). For S. iodantha, males of L. amethystinus made significantly more visits (mean±SE 1.52±0.05) to individual flowers than females (mean±SE 1.34±0.04) (post hoc mean contrasts, F=3.97, P<0.05). Females of C. thalassinus made more visits (mean±SE 1.63±0.07) to individual flowers than males (mean±SE 1.48±0.60), but this difference was not significant (post hoc mean contrasts, F=2.77, P>0.05). In all cases, no differences between males and females were observed in I. hederifolia, and the species×sex interaction was not significant (Table 2) Sum of squares df Mean square F P 14.975 7.561 13.736 0.699 0.350 1.974 1.647 <0.001 <0.01 <0.001 NS NS NS NS 22.130 60.380 5.297 30.151 0.443 1.176 0.128 <0.001 <0.001 <0.01 <0.001 NS NS NS 108.422 54.742 198.902 5.058 5.071 28.591 23.847 1 1 2 1 2 2 2 108.422 54.742 99.451 5.058 2.536 14.295 11.924 1216.360 168 7.240 45.371 123.791 21.719 61.816 1.817 4.822 0.527 1 1 2 1 2 2 2 45.371 123.791 10.860 61.816 0.908 2.411 0.263 344.435 168 2.050 268 Table 2 Results of two-way ANOVAs of number of visits by males and females of two hummingbird species to flowers of Salvia mexicana, S. iodantha and Ipomoea hederifolia for number of legitimate visits, legitimate with robbing visits, and robbing visits only Source Sum of squares df Mean square Legitimate visits Salvia mexicana Hummingbird species Sex Hummingbird species×Sex Residual 10.454 0.100 0.253 1078.858 1 1 1 896 10.454 0.100 0.253 1.204 8.757 0.083 0.210 <0.01 NS NS Salvia iodantha Hummingbird species Sex Hummingbird species×Sex Residual 3.552 0.056 5.758 768.740 1 1 1 896 3.552 0.056 5.758 0.858 4.140 0.065 6.711 0.042 NS <0.01 Ipomoea hederifolia Hummingbird species Sex Hummingbird species×Sex Residual 28.315 0.223 1.208 1127.582 1 1 1 671 28.315 0.233 1.028 1.680 16.849 0.133 0.719 <0.001 NS NS Legitimate with robbing visits Ipomoea hederifolia Hummingbird species Sex Hummingbird species×Sex Residual 0.121 0.553 0.050 44.633 1 1 1 178 0.121 0.533 0.050 0.251 0.483 2.205 0.199 NS NS NS Robbing visits only Ipomoea hederifolia Hummingbird species Sex Hummingbird species×Sex Residual 1.263 0.211 0.211 22.386 1 1 1 35 1.263 0.211 0.211 0.640 1.975 0.330 0.330 NS NS NS Foraging behavior on natural flowers No nectar robbing was observed in S. mexicana and S. iodantha. All flowers were legitimately visited at least once by C. thalassinus and L. amethystinus. In S. mexicana, individuals of C. thalassinus made significantly more legitimate visits on average than L. amethystinus (Fig. 3). In S. iodantha, we observed a marginal difference between hummingbird species in the number of legitimate visits (Fig. 3, Table 2). Nectar robbing was observed in I. hederifolia. Individuals of C. thalassinus actively pierced the base of the flowers (21% of the flowers), with no apparent damage to the ovaries but leaving a slit at the base of the flowers. In contrast, individuals of L. amethystinus did not pierce but probed at the base of the flowers. Although no apparent damage was left at the base of the flower or to the ovaries, we interpreted this behavior as illegitimate, facultative nectar robbing. Nine flowers of Ipomoea hederifolia were not visited (6 out of 450 presented to C. thalassinus and 3 out of 450 presented to L. amethystinus). Most flowers were legitimately visited at least once by C. thalassinus and L. amethystinus (64.4 and 83.7%, respectively), and the remaining were legitimately visited but then robbed or only robbed by C. thalassinus (25.1 and 9.5%, respectively), and legitimately visited but then probed at the F P base or only probed at the base by L. amethystinus (14 and 2.3%, respectively). We then compared the number of visits of both hummingbird species according to these behavioral categories, separately. Individuals of L. amethystinus made more legitimate visits and legitimate visits with robbing than C. thalassinus (Fig. 3), but the number of robbing visits to flowers of I. hederifolia was not statistically different between species (Fig. 3). Pollinator effectiveness In all cases, open pollination (unbagged flowers) resulted in a higher seed production than flowers legitimate visits by either hummingbird species (Fig. 4). C. thalassinus was significantly more effective in pollinating S. mexicana than L. amethystinus (F1,886=9.73, P<0.01). Both hummingbird species were equally effective in pollinating S. iodantha (F1,887=0.61, P>0.05) and I. hederifolia (F1,662=0.753, P>0.05). The number of seeds was not affected by the number of legitimate visits in the Salvia species (S. mexicana F7, 886=0.56, P>0.05; S. iodantha F6,887=0.381, P>0.05), but in I. hederifolia the effectiveness of both hummingbird species significantly increased up to full seed set after two visits to a flower (F1,662=101.24, P<0.001; Fig. 5). In Salvia spp, the species×number of visits interactions were not signif- 269 Fig. 4 Variation in number of seeds produced by S. mexicana, S. iodantha, and I. hederifolia under open conditions of pollination and after legitimate and illegitimate visitation by C. thalassinus and L. amethystinus. Asterisks indicate significance after post hoc mean comparisons (***P=0.0001) Fig. 5 Relationship between number of legitimate visits and number of seeds produced after hummingbird visitation. Numbers above bars indicate flowers visited a given number of times. Asterisks indicate significance after post-hoc mean comparisons (*P=0.01, **P=0.001, ***P=0.0001) icant (S. mexicana F5,886=1.87; P>0.05; S. iodantha F1,887=0.18, P>0.05), but it was significant in I. hederifolia (F5,662=4.43, P<0.001). significant differences between hummingbird species (F1, 897 =11.22, P<0.001). That is, the number of legitimate visits affects seed production. In I. hederifolia, the number of legitimate visits also affected seed production (Table 3). Because the F-value for the hummingbird species×number of visits interaction was not significant, it was removed from the model. After removing such an effect, the F-value for hummingbird species was significant (F1, 673 =4.22, P<0.05). Effect of number of legitimate visits on seed production Variation in S. mexicana seed production was not affected by the number of legitimate visits (Table 3). There were no significant differences between species, and the hummingbird species×number of visits interaction was significant (Table 3). Similar results were observed in S. iodantha, except that hummingbird species×number of visits interaction was not significant (Table 3). We removed it from the model and found Effect of number of robbing visits on seed production As expected, nectar robbing was not observed in short-tubed flowers (S. mexicana and S. iodantha), so 270 Table 3 Results of ANCOVAs for the regression of the mean number of seeds against the number of visits by the two hummingbird species in flowers of S. mexicana, S. iodantha and I. hederifo- lia. A low P-value for hummingbird species×number of visits interaction, indicates that the number of visits (covariate) was useful in predicting seed production for legitimate visits and legitimate with robbing visits Source Sum of squares df Legitimate visits Salvia mexicana Hummingbird species Number of visits Hummingbird species ×Number of visits Residual 2.049 2.223 8.050 1443.305 1 1 1 896 2.049 2.233 8.050 1.611 1.272 1.386 4.997 NS NS <0.05 Salvia iodantha Hummingbird species Number of visits Hummingbird species ×Number of visits Residual 2.610 0.134 0.001 743.204 1 1 1 896 2.610 0.134 0.001 0.829 3.147 0.161 0.001 NS NS NS Ipomoea hederifolia Hummingbird species Number of visits Hummingbird species ×Number of visits Residual 0.247 51.218 0.003 109.743 1 1 1 672 0.247 51.218 0.003 0.163 1.514 313.627 0.016 NS <0.0001 NS Legitimate with robbing visits Ipomoea hederifolia Hummingbird species Number of visits Hummingbird species ×Number of visits Residual 0.000 6.648 0.435 29.586 1 1 1 174 0.000 6.648 0.435 0.170 0.000 39.096 2.556 NS <0.0001 NS they are not analyzed further. In contrast, some longtubed flowers of I. hederifolia were robbed after a number of legitimate visits, and other flowers were only robbed. We analyzed these two categories separately. Flowers legitimately visited and then robbed by C. thalassinus produced significantly fewer seeds than those legitimately visited and then illegitimately probed by L. amethystinus (Fig. 4). Results of ANCOVA showed that seed production in I. hederifolia is also affected by number of legitimate visits with robbing (Table 3). After removing the non-significant hummingbird species×number of visits interaction, the F-value for hummingbird species was significant (F1,175 =6.58, P<0.05). Number of visits was useful in predicting seed production (P<0.001). Flowers that were only robbed by C. thalassinus produced almost no seeds and those only probed at the base by L. amethystinus produced no seeds (Fig. 4). There was a statistical difference in seed production between flowers legitimately visited by C. thalassinus, and those legitimately visited and then robbed by the same hummingbird species (one-way ANOVA, F1,414=14.03, P<0.001), suggesting that robbing negatively affected seed set in this species (Fig. 4). In contrast, no differences in seed set were observed between flowers legitimately visited by L. amethystinus and those legitimately visited and then probed at the base by the same hummingbird species (P>0.05; Fig. 4). Mean square F P Discussion Pollinator effectiveness Within hummingbird species, individuals visited flowers legitimately and robbed, sometimes simultaneously, and in other cases as separate activities. Nectar robbers are generally assumed not to pollinate (Inouye 1983); however, we have shown that pollination effectiveness depends on corolla length, frequency, and mode of nectar robbing. Reduced seed set in flowers of all three species compared to those under open conditions of pollination may be because flowers were excluded from other more effective pollinators (Arizmendi 1994) than the ones we studied, and/or received fewer visits. There are three possible explanations for the increase in seed production of I. hederifolia with the number of visits: (1) hummingbirds used in this comparative study transfer and deposit pollen loads inefficiently, (2) the plant is resource and/or pollen limited, and/or (3) the plant is obligately outcrossed. Although our study was not designed to distinguish among these non-mutually exclusive explanations, the among-plant variation we found in the number of seeds produced by each flower depending on the number of visits and the pollinator (Fig. 5), suggests that the number and origin of pollen grains deposited on the stigma may play an important role in determining seed set. The differences in the number of times a hummingbird 271 visits a given flower means that bill and floral morphological attributes can not solely account for the variation we observed on I. hederifolia reproductive success, but that flowers need more than one visit to reach full seed set. On the other hand, differences between Salvia species in seed set can be solely interpreted in terms of differences in pollinator effectiveness (morphological mismatch between hummingbird beaks and flowers). Although we cannot distinguish between those seeds sired as a result of hummingbird visitation from those that resulted by selfing (Salvia flowers visited only once produced the same number of seeds as those visited more times), it seems that individuals of C. thalassinus are more effective pollinators of S. mexicana flowers and those of L. amethystinus are more effective pollinators of S. iodantha and I. hederifolia flowers. Nevertheless, pollinator effectiveness was diminished by the intensity and the mode of nectar robbing by C. thalassinus because I. hederifolia flowers produced fewer seeds when legitimately visited and then robbed or only robbed, than those that were only legitimately visited (Fig. 4). Robbing of Ipomoea by Colibri yields some seeds, but robbing by Lampornis did not. Also, the fact that Lampornis robbing yields zero seed set whereas legitimate visits yields seed set is evidence for a cost of robbing in this species. It is possible that Colibri does some pollination during what looks like an illegitimate visit, that is, by piercing the corolla some pollen grains dropped onto the stigma. Studies looking at the negative effect of nectar robbers on seed set have shown contradictory results (see Maloof and Inouye 2000 for a review). The assumption that nectar robbing carries both male and female fitness costs to plants is based on the correlative evidence that robbed flowers have fewer seeds than flowers visited legitimately (Roubik et al. 1985; Reddy et al. 1992; Irwin and Brody 1998, 1999; Traveset et al. 1998; Navarro 1999; this study). In nature, however, robbed flowers might receive further visits that were legitimate, even when flowers were emptied of nectar by robbers, and therefore one legitimate visit might suffice to produce the full seed set. In contrast, robbed flowers would set fewer or no seeds if long-billed hummingbirds were less efficient in pollinating or avoided them because they are empty of nectar (less attractive) than those non-robbed flowers (see also Irwin and Brody 1998, 1999). There is some evidence that hummingbirds discriminate against robbed flowers (Gass and Montgomerie 1981; Irwin and Brody 1998), but whether legitimate pollinators are less efficient with robbed flowers remains to be investigated. Flowers that receive further legitimate visits after robbing might have a real fitness cost when nectar taken by the robbers is replaced by additional secretion. New nectar production may impose a high cost on the plant (Southwick 1984; Pyke 1991). Navarro (1999) showed that flowers of Macleania bullata (Ericaceae) produced more nectar as a consequence of nectar robbing by shortbilled hummingbirds. This additional nectar secretion entailed an energetic cost that reduced fruit set. The effects of corolla length Legitimate pollinators may select for long corollas and colors (and other floral traits), but the role of nectar robbers as agents of selection on flower morphology has mostly been ignored (Maloof and Inouye 2000). Robbing may represent a counter selective force that counter balance selection for longer corollas by specialist pollinators. Some floral traits, including variation in corolla length among populations (Waser 1979; Roubik et al. 1985), flower location (Colwell et al. 1974; Traveset et al. 1998; Arizmendi et al. 1996), and high rates of nectar production (Pyke 1981; Navarro 1999), may be the consequence of nectar robbing. Floral characters such as corolla length, curvature, and flower angle vary considerably among hummingbird-pollinated plants, and this affects foraging efficiency of the birds. It has been shown that hummingbirds take more time to extract nectar from long-tubed flowers than from short-tubed flowers (e.g., Hainsworth 1973; Hainsworth and Wolf 1976; Montgomerie 1984; Hainsworth et al. 1983). Earlier studies have compared foraging behavior of hummingbird species that differed in bill lengths (Wolf et al. 1972; Hainsworth 1973; Hainsworth and Wolf 1976; Ewald and Williams 1982; Montgomerie 1984; Temeles and Roberts 1993; Temeles 1996), and have shown that birds with longer bills have greater maximum extraction depths and faster handling times at relatively long flowers than birds with shorter bills. We were unable to quantify handling times and nectar volume extracted by the birds; however, our experimental approach showed that short-billed hummingbirds used nectar robbing as an alternative to forage from long-tubed flowers. We found that the longer of the three species (Ipomoea) (and the longer of the artificial flowers) receive illegitimate probes, as has been commonly observed in nature (e.g., Grant 1952; Skutch 1954; Colwell et al. 1974; Kodric-Brown and Brown 1979; Stiles 1981; Inouye 1983; Kodric-Brown et al. 1984; Feinsinger et al. 1987; Ornelas 1994; Navarro 1999). Although flower angle may also play an important role in explaining flower-handling variation (Montgomerie 1984), our results for nectar robbing on long-tubed flowers were not confounded by this factor by design. In short, we have shown that facultative robbing behavior in both hummingbird species is most likely if corollas are longer. This suggests that selection will not unilaterally favor longer corollas – facultative (and obligate) robbers may select in the opposite direction from non-robbers, that is, robbers may select for shorter corollas. However, a formal demonstration of such selection would require establishment of a relationship between corolla length and seed set within a plant species and show that this is related to robbing. Nectar robbing among short-billed hummingbirds We found significant variation between hummingbird species in use of robbing behavior. It is possible that the 272 variation among individuals in nectar robbing is a result of their previous foraging experience. A higher incidence of illegitimate feeding by males has been observed as a result of their having shorter bills (Bertin 1982; Temeles and Roberts 1993). However, we found that males and females for either hummingbird species in I. hederifolia used legitimate and robbing behaviors similarly, although both species are sexually dimorphic in bill length (Ornelas 1995). Here, we have documented that individuals of both species of hummingbirds are indeed facultative in robbing behavior, and that this presumably opportunistic behavior of C. thalassinus, which has a serrated bill, negatively affected seed production of long corollas. Additional research is needed to determine whether nectar robbing is most likely to evolve among shortbilled hummingbirds (Ornelas 1994). Because we found no visible evidence of physical damage to the ovaries and corolla tubes of the examined flowers, we think that the decreased fertility of flowers robbed by both hummingbird species may be due to effects on pollinator visitation, rather than to destruction of plant reproductive parts (see also Irwin and Brody 1999). However, individuals of C. thalassinus with enlarged tomial serrations had the highest negative effect on I. hederifolia seed production. Further studies should explore the role bill serrations play in perforating and handling long-tubed flowers. If the function of bill serrations were to hold the flowers in some fashion and then rob them, one would expect the act of legimate visitation and flower piercing to be somewhat slower than that of nectar robbery. A high-speed video analysis would show whether the flowers are stabbed, grasped, and/or sawed, and one could compare floral handling (e.g., foraging time) between nectar robbers with bill serrations and nectar thieves with smooth bills. In conclusion, we have shown that both of the two species of hummingbirds are more likely to rob artificial flowers with long rather than short corollas. Furthermore, the hummingbird with the shorter bill was more likely to rob than the species with the longer bill. In the field, only the plant species with a long corolla experienced nectar robbing, and this had a negative impact on seed set. Our experimental results suggest that nectar robbing may favor short corollas. Acknowledgements Comments by María del Coro Arizmendi, William A. Calder, Phyllis D. Coley, Barbara Ditsch, Astrid Eben, Theodore H. Fleming, Myriam Mermoz, Jorge López Portillo, Joan E. Maloof, Carlos Martínez del Rio, Victor Rico-Gray, Nick Waser and three anonymous reviewers greatly improved an earlier draft of this paper. We also thank R. Iorente Adame, C. Ortíz Arrona, G. Zavala, J. Sandoval, and L.I. Iñiguez for their help during fieldwork, and the staff of Las Joyas field station for the facilities to conduct our research. This research was supported by a doctoral scholarship (No. 56254) from the Consejo Nacional de Ciencia y Tecnología (CONACyT) and the James R. Silliman Memorial Research Grant to J.F.O. References Abacus Concepts (1989) SuperANOVA. Abacus Concepts, Berkeley Abacus Concepts (1996) StatView reference. Abacus Concepts, Berkeley Arizmendi MC (1994) Interacciones ecológicas múltiples: el caso del sistema mutualista colibríes-plantas y el ladrón de néctar Diglossa baritula (Passeriformes: Aves). PhD thesis, CCHCentro de Ecología, UNAM, México, DF Arizmendi MC, Dominguez CA, Dirzo R (1996) The role of an avian nectar robber and of hummingbird pollinators on the reproduction of two plant species. Funct Ecol 10:119–127 Barrows EM (1976) Nectar robbing and pollination of Lantana camara (Verbenaceae). Biotropica 8:132–135 Beal WJ (1880) Fertilization of flowers by hummingbirds. Am Nat 14:126–127 Becerra JXI, Venable DL (1989) Extrafloral nectaries: a defense against ant-Homoptera mutualisms? Oikos 55:276–280 Berry PE (1982) The systematics and evolution of Fuchsia Sect. Fuchsia (Onagraceae). Ann Mo Bot Gard 63:1–98 Bertin RI (1982) Floral biology, hummingbird pollination and fruit production of trumpet creeper (Campsis radicans, Bignoniaceae). Am J Bot 69:122–134 Bittrich V, Amaral MCE (1996) Pollination biology of Symphonia globulifera (Clusiaceae). Plant Syst Evol 200:101–110 Bolten AB, Feinsinger P (1978) Why do hummingbird flowers secrete dilute nectar? Biotropica 10:307–309 Chemas-Jaramillo A (1995) Biología floral y polinización de doce especies de Ipomoea (Convolvulaceae) en un bosque tropical caducifolio de Jalisco. MSc thesis, Facultad de Ciencias, UNAM, México, DF Colwell RK (1973) Competition and coexistence in a single tropical community. Am Nat 107:737–760 Colwell RK, Betts BJ, Bunnell P, Carpenter FL, Feinsinger P (1974) Competition for the nectar of Centropogon valerii by the hummingbird Colibri thalassinus and the flower-piercer Diglossa plumbea, and its evolutionary implications. Condor 76:447–452 Darwin C (1859) On the origin of species. Murray, London Darwin C (1862) On the various contrivances by which British and foreign orchids are fertilized by insects. Murray, London Darwin C (1877) The different forms of flowers on plants of the same species. Murray, London Ewald PW, Williams WA (1982) Function of the bill and tongue in nectar uptake by hummingbirds. Auk 99:573–576 Feinsinger P (1983) Coevolution and pollination. In: Futuyma D, Slatkin M (eds) Coevolution. Sinauer, Sunderland, pp 282– 310 Feinsinger P, Colwell RK (1978) Community organization among neotropical nectar-feeding birds Am Zool 18:779–795 Feinsinger P, Linhart YB, Swarm LA, Wolfe JA (1979) Aspects of the pollination biology of three Erythrina species on Trinidad and Tobago. Ann Mo Bot Gard 66:451–471 Feinsinger P, Beach JH, Linhart YB, Busby WH, Murray G (1987) Disturbance, pollinator predictability, and pollination success among Costa Rican cloud forest plants. Ecology 68:1294– 1305 Fenster CB (1991) Selection on floral morphology by hummingbirds. Biotropica 23:98–101 ffrench R (1973) A guide to the birds of Trinidad and Tobago. Livingston, Wynnewood Gass CL, Montgomerie RD (1981) Hummingbird foraging behavior: decision making and energy regulation. In: Kamil AC, Sargent TD (eds) Foraging behavior: ecological, ethological, and psychological approach. Garland, New York, pp 159–164 Gill FB (1987) Ecological fitting: use of floral nectar in Heliconia stilesii Daniels by three species of hermit hummingbirds. Condor 89:779–787 Grant V (1952) Isolation and hybridization between Aquilegia formosa and A. pubescens. Aliso 2:341–360 273 Guerrant EO, Fiedler PL (1981) Flower defenses against nectarpilferage by ants. Biotropica 13:25–33 Haber WA, Frankie GW (1982) Pollination of Luehea (Tiliaceae) in Costa Rican deciduous forest. Ecology 63:1740–1750 Hainsworth FR (1973) On the tongue of a hummingbird: its role in the rate and energetics of feeding. Comp Biochem Physiol 46:65–78 Hainsworth FR, Wolf LL (1976) Nectar characteristics and food selection by hummingbirds. Oecologia 25:101–114 Hainsworth FR, Mercier T, Wolf LL (1983) Floral arrangements and hummingbird feeding. Oecologia 58:225–229 Inouye DW (1983) The ecology of nectar robbing. In: Bentley B, Elias T (eds) The biology of nectaries. Columbia University Press, New York, pp 153–173 Irwin RE, Brody AK (1998) Nectar robbing in Ipomopsis aggregata: effects on pollinator behavior and plant fitness. Oecologia 116:519–527 Irwin RE, Brody AK (1999) Nectar-robbing bumble bees reduce the fitness of Ipomopsis aggregata (Polemoniaceae). Ecology 80:1703–1712 Jardel EJ (1991) Perturbaciones naturales y antropogénicas y su influencia en la dinámica sucesional de los bosques de Las Joyas, Sierra de Manantlán, Jalisco. Tiempos Cienc 22: 9–26 Kodric-Brown A, Brown JH (1979) Competition between distantly related taxa in the evolution of plants and pollinators. Am Zool 19:115–127 Kodric-Brown A, Brown JH, Byers GS, Gori DF (1984) Organization of a tropical island community of hummingbirds and flowers. Ecology 65:1358–1368 Lara C (1995) Ecología de la conducta de forrajeo del colibrí oreji violeta-verde (Colibri thalassinus): implicaciones en el robo de néctar. Tesis de Licenciatura, Benemérita Universidad Autónoma de Puebla, México Lara C, Ornelas JF (1998) Forrajeo de artrópodos por dos colibríes mexicanos en condiciones de aviario. Ornitol Neotrop 9:41–50 Lara C, Ornelas JF (2001) Nectar “theft” by hummingbird flower mites and its consequences for seed set in Moussonia deppeana. Funct Ecol 15:78–84 Maloof JE, Inouye DW (2000) Are nectar robbers cheaters or mutualists? Ecology 81:2651–2661 McDade LA (1984) Systematics and reproductive biology of Central American species of the Aphelandra pulcherrima complex (Acanthaceae). Ann Mo Bot Gard 71:104–165 McDade LA, Kinsman S (1980) The impact of floral parasitism in two Neotropical hummingbird-pollinated plant species. Evolution 34:944–958 Montgomerie RD (1984) Nectar extraction by hummingbirds: response to different floral characters. Oecologia 63:229–236 Navarro L (1999) Pollination ecology and effect of nectar removal in Macleania bullata (Ericaceae). Biotropica 31:618–625 Nilsson LA (1988) The evolution of flowers with deep corolla tubes. Nature 334:147–149 Oliveira PS, Rico Gray V, Díaz Castelazo C, Castillo Guevara C (1999) Interaction between ants, extrafloral nectaries and insect herbivores in Neotropical coastal sand dunes: herbivore deterrence by visiting ants increases fruit set in Opuntia stricta (Cactaceae) Funct Ecol 13:623–631 Ornelas JF (1994) Serrate tomia: an adaptation for nectar robbing in hummingbirds? Auk 111:703–710 Ornelas JF (1995) Radiation in the genus Amazilia: a comparative approach to understanding the diversification of hummingbirds. PhD Dissertation. University of Arizona, Tucson Ornelas JF (1996) El origen y evolución de los colibríes. Rev Cienc 42:38–47 Ornelas JF (1998) Filogenias y método comparativo: identificación y evaluación de hipótesis evolutivas y establecimiento de criterios para conservar especies en riesgo. Acta Zool Mex 74: 5–42 Ornelas JF, Arizmendi MC (1995) Altitudinal migration: implications for conservation of avian neotropical migrants in western Mexico. In: Wilson MH, Sade SA (eds) Conservation of Neo- tropical migratory birds in Mexico (Miscellaneous Publication 727). Maine Agricultural and Forest Experiment Station, Orono, pp 98–112 Pleasants JM (1983) Nectar production in Ipomopsis aggregata (Polemoniaceae). Am J Bot 70:1468–1475 Pyke GH (1981) Optimal nectar production in a hummingbird plant. Theor Popul Biol 20:326–343 Pyke GH (1991) How much does floral nectar cost? Nature 350:58–59 Reddy TB, Rangaiah K, Reddi EUB, Reddi CS (1992) Consequences of nectar robbing in the pollination ecology of Vitex negundo (Verbenaceae). Curr Sci 62:690–691 Ridgway R (1890) The hummingbirds (report). American Museum of Natural History, New York Roubik DW, Holbrook NM, Parra G (1985) Roles of nectar robbers in reproduction of the tropical treelet Quassia amara (Simarubaceae). Oecologia 66:161–167 Ruschi A (1982) Aves do Brasil. Expressão e Cultura, EXPED, Rio de Janeiro Skutch AF (1954) Life histories of Central American birds, vol 1 (Pacific Coast Avifauna 31). Cooper Ornithological Society, Berkeley Skutch AF (1973) The life of the hummingbird. Crown Press, New York Snow BK (1981) Relationships between hermit hummingbirds and their food plants in eastern Ecuador. Bull Br Ornithol Club 101:387–396 Snow DW, Snow BK (1980) Relationships between hummingbirds and flowers in the Andes of Colombia. Bull Br Mus Nat Hist Zool 38:105–139 Soberón J, Martínez del Rio C (1981) The dynamics of a plantpollinator interaction. J Theor Biol 91:363–378 Soberón J, Martínez del Rio C (1985) Cheating and taking advantage in mutualistic associations. In: Boucher DH (ed) The biology of mutualism. Croom Helm, London, pp 192–216 Southwick EE (1984) Photosynthate allocation to floral nectar: a neglected energy investment. Ecology 65:1765–1769 Stiles FG (1981) Geographical aspects of bird-flower coevolution with particular reference to Central America. Ann Mo Bot Gard 68:323–352 Stiles FG (1985) Seasonal patterns and coevolution in the hummingbird-flower community of a Costa Rican subtropical forest. In: Buckley PA, Foster MS, Morton ES, Ridgely RS, Buckley FG (eds) Neotropical Ornithology. Ornithol Monogr 36, pp 757–787 Temeles EJ (1996) A new dimension to hummingbird-flower relationships. Oecologia 105:517–523 Temeles EJ, Roberts WM (1993) Effect of sexual dimorphism in bill length on foraging behavior: an experimental analysis of hummingbirds. Oecologia 94:87–94 Traveset A, Willson MF, Sabag C (1998) Effect of nectar-robbing birds on fruit set of Fuchsia magellanica in Tierra del Fuego: a disrupted mutualism. Funct Ecol 12:459–464 Vázquez JA, Cuevas R, Cochrane TS, Iltis HH, Santana FJ, Guzmán L (1995) Flora de Manantlán. Universidad de Guadalajara, University of Wisconsin-Madison, Botanical Research Institute of Texas, Fort Worth Waser NM (1979) Pollinator availability as a determinant of flowering time in ocotillo (Fouquieria splendens). Oecologia 39:107–121 Willmer PG, Corbet SA (1981) Temporal and microclimatic partitioning of the floral resources of Justicia aurea amongst a concourse of pollen vectors and nectar robbers. Oecologia 51:67– 78 Wolf LL, Hainsworth FR, Stiles FG (1972) Energetics of foraging: rate and efficiency of nectar extraction by hummingbirds. Science 176:1351–1352 Wolf LL, Stiles FG, Hainsworth FR (1976) Ecological organization of a tropical highland community. J Anim Ecol 45:349– 379 Zar JH (1984) Biostatistical analysis. Prentice Hall, Englewood Cliffs
© Copyright 2026 Paperzz