Flowering plants involve mutualistic (e.g., pollinators, seed dispersers) and antagonistic interactions (e.g., herbivores, seed predators, nectar robbers) simultaneously (Bronstein, 2001; Karban & Baldwin, 1997). Nectar robbery is a common antagonistic interaction in species with tubular corollas (Irwin & Maloof, 2002; Milet-Pinheiro & Schlindwein, 2009), which occurs when robbers consume nectar through holes made at the base of the corolla, without transferring pollen among flowers (Irwin, Adler, & Brody, 2004). Impacts of nectar robbers affect, directly or indirectly, the plant reproduction performance (Milet-Pinheiro & Schlindwein, 2009). Direct impacts are functional damage of reproductive female flowers (e.g., ovaries and style; Maloof & Inouye, 2000; Roubik, 1982) and male (e.g., less pollen availability; Irwin, Bronstein, Manson, & Richardson, 2010) structures. Indirect impacts include interference with legitimate pollinators by changing their foraging behavior or causing flower avoidance (Irwin, 2003; Irwin et al., 2010). Nectar robbers can have negative effects on endemic species fitness. In Polygala vayredae flowers that were robbed produced fewer seeds and had less pollen transfer than intact flowers, spotlighting its relevance for conservation and management (Castro, Silveira, & Navarro, 2008).
The aim of this study was to document the intensity of nectar robbers and potential pollinators in the Brazilian endemic Collaea cipoensis. We addressed the following questions: (1) Which are the floral visitors and potential pollinators of C. cipoensis? (2) Is nectar robbing a frequent phenomenon in C. cipoensis flowers? (3) Does nectar robbery affect floral abortion?
Our study was conducted in an area under a recovery process belonging to Vellozia Private Reserve (19°16′54″ S, 43°35′45″ W) at Serra do Cipó, Minas Gerais, Brazil. The region has an average annual temperature of 21.3 °C and an average annual rainfall of 1247 mm (Inmet, 2015). Typically there is a cold, dry season from May to September and a hot rainy season from November to March (Madeira & Fernandes, 1999). The rupestrian grasslands region comprises 7000 km2 characterized by rocky outcrops and sandy fields with soils poor in nutrients (Furley & Ratter, 1988; Oliveira et al., 2015). The flora comprises ca. 4000 plant species with the highest levels of plant endemism of Brazil (Giulietti, Pirani, & Harley, 1997). Despite the high biodiversity and conservation importance of this region, many species with restricted distributions are endangered due to strong human disturbance (Menezes & Giulietti, 2000).
C. cipoensis (Fabaceae) is a restricted endemic species limited geographically to Serra do Cipó, Minas Gerais, Brazil (Rapini, Ribeiro, Lambert, & Pirani, 2008). It is known from only a few populations, so it is considered extremely rare (Fortunato, 1995). C. cipoensis is a perennial shrub with 2-3 m in height, with red, tubular flowers disposed in inflorescences with 2-6 flowers and is hummingbird pollinated, flowering and fruiting continuously throughout the year with peaks from May to January (Fortunato, 1995). It is distributed along watercourses, associated with acidic, poor in nutrients and high aluminum saturation soils (Negreiros, Borges-Moraes, & Fernandes, 2008).
The study was conducted in October and November of 2014 in plants located in a regeneration plantation inside to the Vellozia Reserve during peak flowering: (1) to evaluate the effects of nectar robbers on flowers with and without damage of C. cipoensis. We selected 6 plots of 5 m × 5 m (25 m2) each separated by 10 m throughout the study area and sampled all reproductive individuals; (2) to identify nectar robbers and flower visitors we observed 17 plants and 350 flowers; and (3) we performed a paired design considering robber and intact flowers to quantify visitation events (separated no more 50 cm), and we conducted focal observations (by 6 persons) during 07:00 to 18:00 h, totalizing 132 h in 2 days.
In total we found 44 flowering plants, and quantified the number of flowers and developing fruits with and without damage by nectar robbers. We consider nectar robbing as holes at the base of the corolla, easily distinguished in the calyx of flowers and in newly formed fruits (Fig. 1A). We assumed a floral abortion occurred when a viable flower fell from the plant when touched lightly. We calculated flower abortion ratio as the number of aborted flowers divided by the total number of flowers produced by each plant. We tested for differences in flower abortion ratio per plant between damaged and intact flowers using a Generalized Linear Mixed Model using SAS v. 9.2 (SAS Institute, 2008). A binomial distribution was specified for abortion ratio. Since damaged and non-damaged flowers were found in the same individual, the plant was included as a random factor in the analysis.
Our results showed that the main floral visitors of C. cipoensis were bees and hummingbirds. The most common bee visitor was Apis mellifera (N = 99 flowers visited), followed by Trigona spinipes (N = 22), Bombus sp. (N = 13) and wasps (N = 4). The main nectar robber was T. spinipes, which chewed a hole through the calyx and corolla to access the nectar chamber. A. mellifera was second in robbery visits, and less frequently consumed pollen and made legitimate visits. We recorded two hummingbird species visiting the flowers: Colibri serrirostris (Fig. 1B) and Eupetomena macroura (Fig. 1C). Based on the frequency of visits and the observation that both made contact with the reproductive organs of the flower, we consider these species as potential pollinators. Our results show that hummingbirds preferentially visit flowers without damage (mean visits flower 10 ± s.e. = 17.11 ± 2.7, N = 237) than flowers with damage (3.44 ± 1.05, N = 44) (Wilcoxon test Z = 3.97, p < 0.0001).
To assess the damage by nectar robbers we sampled 1286 flowers in 44 individuals of C. cipoensis; of these flowers, 716 (56%) were intact and 570 (44%) showed robbery. Flower abortion ratio did not statistically differ between damaged and intact flowers within plants (F 1,80 = 2.6, p = 0.11); however, there was a trend indicating mean flower abortion ratios were higher for robbed flowers (0.38 ± 0.061) than for intact flowers (0.22 ± 0.057) (mean ± s.e.). Additionally, sampling of 544 unripe and mature fruits of 44 plants showed that only 14% of fruits had signs of robbery as indicated by damage in the remnant calyx.
Our results show that the primary robber is T. spinipes, which agrees with findings reported for other plant species where major nectar robbers correspond to bees (Irwin et al., 2010; Navarro, 2000). As in the case of bumblebees, T. spinipes uses its proboscis to make holes at the base of the flowers (Irwin et al., 2010; Macior, 1966). This interaction may decrease fruit or seed set by damaging plant reproductive structures (McDade & Kinsman, 1980). Our results suggest that damage caused by robbers to C. cipoensis flowers could lead to flower abortion, which may reduce overall fruit production.
Nectar robbing leads to negative effects on plant reproductive success when it exists, for example, destruction of ovaries or ovules (Navarro & Medel, 2009), selective seed abortion (Zhang et al., 2011), or by inducing behavioral changes in pollinators with the ability to detect robbed flowers (Irwin & Brody, 1999). In C. cipoensis we observed T. spinipes making a large hole at the base of the corolla and this damage could lead to ovary injuries triggering flower abortion. Nonetheless, other variables, like pollen availability, must be considered. A similar scenario, although we did not record any other bird visiting C. cipoensis flowers, was found in Puya coerulea where robbing by the blackbird Curaeus curaeus on hummingbird pollinated flowers; negatively affected the flowers by causing intense damage androecium (González-Gómez & Valdivia, 2005). This study documented that almost 50% of the flowers in the population showed evidence of nectar robbing; hence the phenomenon is likely a significant determinant of plant reproductive success. A small proportion of fruits (14%) have evidence of damage, however, it could indicate that if injury does not affect the ovary the fruits can develop. Additionally, this suggests that nectar robbing by T. spinipes could have adverse effects on reproductive success of this restricted endemic Brazilian species. Further studies should address the intensity and cost of nectar robbing, interactions between robbers and pollinators, and consequences of floral abortion to the plant present and future reproductive success.
The authors thank Gonçalves-da-Silva, G. for field assistance, the Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre of the Universidade Federal de Minas Gerais, Programa Estudante-Convenio Pós-graduação-PEC-PG, CAPES, CNPq, INECOL and Reserva Vellozia.