Temporal variation in extrafloral nectar secretion in different ontogenic stages of the fruits of Alibertia verrucosa S. Moore (Rubiaceae) in a Neotropical savanna

Abstract

In this study, we evaluate the temporal variation in extrafloral nectaries (EFNs) secretion in different ontogenic stages of Alibertia verrucosa (Rubiaceae) fruits in a Neotropical savanna. We observe greater nectar secretion rate in fruits of intermediate size compared with young or ripe fruit, indicating that they are possibly more protected by ants. In addition, the nectar secretion was higher at night, a pattern that could be associated with an increase of herbivore pressure and water stress during the daylight hours. In fact, due to the high temperature and low humidity during the day in savannas, most herbivores display strong nocturnal activity, and plants can avoid nectar secretion in this period. Our results indicate that A. verrucosa can change the ants' attraction according to EFN secretion during the ontogenic stages of the fruits, probably secreting more nectar when the biotic defense are more necessary for the protection of the fruits and the plant as a whole.

Keywords:

• nectar secretion
• temporal variation
• multitrophic interactions
• biotic defense
• optimal defense theory

1. Introduction

Over evolutionary time, the strong pressure exerted by herbivores has led plants to develop different chemical, structural, and biological strategies that can reduce plant damage by herbivores (Marquis 1984; Coley & Kursar 1996). The main and best-known defense mechanism against herbivory is direct defense, in which the plants have secondary compounds in several structures that make the herbivores’ ability to exploit these plants more difficult (Coley & Kursar 1996; Cortesero et al. 2000). However, plants also have indirect defenses by the third trophic level, particularly predator insect and parasitic wasps and flies (parasitoids), which are known to protect plants against herbivory (Rico-Gray & Oliveira 2007; Romero et al. 2008). This indirect and biotic defense is usually induced, where the plant releases volatile and/or nutritive substances that attract predators and herbivore parasitoids (Kessler & Baldwin 2001; Sabelis et al. 2001; Heil 2008).

In tropical ecosystems, ants have an important role in the biotic defense of plants (Rico-Gray & Oliveira 2007; Rosumek et al. 2009; Romero & Koricheva 2011). In this region, thousands of angiosperms and pteridophytes have extrafloral nectaries (EFNs) that secrete an energy-rich liquid that attracts ants (Rico-Gray & Oliveira 2007). Several authors have reported not only the increase in the rate of leaf herbivory, but also the reduction of fruit production in plants isolated from ants (Oliveira et al. 1999; Nascimento & Del-Claro 2010), which is reflected in the increase of EFN-plants' fitness visited by ants (Rico-Gray & Thien 1989; Oliveira et al. 1999; Chamberlain & Holland 2009). In fact, most studies involving the interaction between ants and EFN plants have focused only on the decrease of leaf herbivory (see Rico-Gray & Oliveira 2007 and references therein). In such cases, EFNs are extremely active in young leaves, which are more vulnerable to damage caused by herbivores (Heil et al. 2000; Korndörfer & Del-Claro 2006). During leaf ontogeny, EFNs decrease nectar secretions when the leaf begins to invest in physical defenses, such as cellulose, lignin, and silica (Dyer et al. 2001; Korndörfer & Del-Claro 2006).

However, it has been found that some plant species have EFN in their fruits (Rico-Gray 1989; Díaz-Castelazo et al. 2005; Holland et al. 2010). Since the function of EFN is traditionally related to the attraction of ants (Bentley 1977), it is expected that the presence of these structures on fruits may be related to the protection of fruits and seeds in formation. In addition, several studies have shown that plants invest in defense especially in reproductive (flowers) and dispersion (fruits) structures (Herrera 1982; Cipollini & Levey 1997; Dicke 1999; Thornburg et al. 2003; Carter et al. 2007), mainly by the fact that these structures are directly related to their reproductive success (Cipollini & Levey 1997; Holland et al. 2009). In fact, during fruit growth and maturation more energy is allocated in these structures and the more valuable it becomes (Cousens et al. 2008). Therefore, it is expected that plants invest more in defense in the later stages along ontogeny of the fruits, once ripe fruits are more apt to dispersion (Cipollini & Levey 1997). Thus, if the strategy of a plant species against herbivory is biotic defense, an increase in nectar secretion is expected during ontogenic stages of the fruits, increasing the visitation of protective ants.

Nevertheless, once nectar is costly to the plant, its secretion may not be continuous (Southwick 1984; Pyke 1991; Heil et al. 2000). In regions with high temperature and low humidity during the day (e.g. Neotropical savanna), most herbivores display strong nocturnal activity due to their ecophysiological limitations (Fagundes et al. 1996; Strauss et al. 2009; Byk & Del-Claro 2010). Therefore, a change in the nectar secretion during periods with high herbivory and/or high levels of water stress can reduce energy loss in ‘superfluous’ EFN secretion, and reduce the risk of EFN fungal infections (Heil et al. 2000; Holland et al. 2010; Yamawo et al. 2012a).

In this study, we evaluate the temporal variation in EFN secretion in different ontogenic stages of Alibertia verrucosa S. Moore (Rubiaceae) fruits in a Brazilian Neotropical savanna. Since the reproductive success of A. verrucosa is directly related to its fruit dispersal performed by birds and mammals (J.C.F.F. and T.J.I. unpublished observations), it is expected that bigger and riper fruits are more often dispersed. In this context, we hypothesized that the bigger and riper the fruits of A. verrucosa, the greater the EFN secretion and, indirectly, the biotic defenses provided by ants. Moreover, we also hypothesized that due to nocturnal foraging of most herbivores and water stress during the daylight hours in the Neotropical savanna, the EFN secretion of A. verrucosa is higher at night, which attracts more ants.

2. Materials and methods

2.1. Study area

We conducted this study at Estação Ecológica Serra das Araras (EESA) (15°38′S and 57°12′W, elev. 217 m), located between the municipalities of Porto Estrela, Cáceres, and Barra do Bugres, west of the state of Mato Grosso, Brazil. The reserve is a Conservation Unit of Integral Protection under the administration of the Brazilian Institute for the Environment and Natural Resources (IBAMA). According to the Köppen classification, the climate is Tropical savanna (Aw) with an average annual temperature of 28°C, humidity 70%, and 1.400 mm of precipitation. It has two well-defined seasons, a rainy season between November and April and dry season between May and October. The reserve area covers 28.700 ha of continuous forest with different physiognomies. However, the study area is characterized as a Brazilian Neotropical savanna, a xeromorphic forest also known as cerrado. The terrain is undulating with altitudinal variation of 500 m between the plateaus, riverside terrains, and mountains (Zardo et al. 2010). Field observations were conducted in October 2011, at the end of the dry season and at the end of the A. verrucosa fruiting period.

2.2. Species studied

A. verrucosa S. Moore is a treelet species of the Rubiaceae family, which reaches up to 8 m tall and exhibits a dense, broad, and highly branched canopy. This species is commonly found in gallery forests, seasonal forests, and cerrado in the central region of the Brazilian Neotropical savanna. Fruiting period of A. verrucosa is between July and November. Fruits have fleshy mesocarp and rough epicarp, and nearly 30 mm in diameter. Interestingly, this species has only one EFN at the superior extremity of each fruit (Silva-Jr & Pereira 2009). In an analogous manner to Tocoyena formosa (Rubiaceae), after pollination and fall of the corolla, the floral nectaries of A. verrucosa continue secreting nectar, acting functionally as EFN (Santos & Del-Claro 2001).

2.3. Data collection

In order to evaluate the temporal variation in nectar secretion in fruits of A. verrucosa, we randomly selected 41 fruits in 35 individuals. All selected plants were at least 35 m apart. To quantify the amount of nectar secreted, we excluded the possible nectar consumers (winged and nonwinged) covering each fruit with paper bags and putting nontoxic grease in the fruit peduncle. Paper bags have the additional function of preventing desiccation of nectar secreted. Fruits were isolated from their visitors for 24 hours, and we measured the volume of nectar (by graduates microcapillary) every 4 hours (n=7 measurements). We quantified the nectar volume using 1-µL microcapillaries (graduated by 0.1 µL divisions) (Heil et al. 2000; Dáttilo et al. 2012). To evaluate the variation in nectar secretion during fruit ontogeny, we collected the fruits used previously and measured the diameter of each fruit. Later, we also evaluated the investment in carbon during ontogeny fruits of A. verrucosa by the dry weight of each fruit method. For this, the fruits were kept in an oven at 72°C for 24 hours (or until constant weight) and afterward, we measured their dry weight with a precision balance. Here, we used the fruit size as a measure to classify the different ontogenic stages of the A. verrucosa fruits. This classification was confirmed by the color and smell of fruits during their development.

2.4. Data analysis

From the fruit diameter, we estimated the fruit volume by approaching their shape to a sphere. We related the fruit volume with the dry weight of each fruit to determine if the carbon allocation is proportional to growth using simple linear regression (SLR). Initially, in order to test our hypothesis that the greater the fruit, the greater the EFN secretion, we investigated the relationship between dry weight of the fruits (predictor variables) and EFN secretion (response variable) using SLR. However, as the SLR did not fit to the data, we used a posteriori, a model of quadratic regression. This model was the best fit for the data, using the ‘curve fitting’ in the BioStat 5.3 software (Ayres et al. 2007). Finally, to determine whether the amount of nectar secreted varied throughout the day, we used a repeated measures analysis of variance (ANOVA). We performed ANOVA analysis in the SYSTAT software version 13.0 (Wilkinson 1998).

3. Results

In this population of A. verrucosa in the Neotropical savanna, the fruit diameter was 16.26±5.65 mm (mean±SD) and dry weight was 0.716±0.509 mg. As expected, the fruit volume was positively correlated with its dry weight (weight= 0.994×volume), and therefore, weight gain was proportional to growth. In addition, we observed that during fruit ontogeny of A. verrucosa, there are different patterns of nectar secretion. In fact, we did not observe a linear relationship between dry weight of fruits and volume of secreted nectar (r ² =0.02; P=0.351). However, a posteriori, the quadratic regression was the best fit for the data [nectar secreted=0.1+(0.6890×weight) – (0.3568×weight²); r ² =0.27; P<0.001] (Figure 1). This indicates that the fruits of intermediate size produce greater amount of nectar secreted, with little nectar secreted in young fruits, and the absence of nectar secretion in ripe fruits (Figure 1).

Figure 1. Quadratic regression between the mean amount of nectar secreted (mL) over 24 h and the dry weight (mg) of the A. verrucosa fruits (Rubiaceae) in a Neotropical savanna (P<0.001) (n=41 fruits).

We also observed a large variation in the amount of nectar secreted by the A. verrucosa fruits (mean±SD: 0.421±0.687 mL), with EFN producing nectar from 0 mL to 3.91 mL at the highest peaks of secretion. In fact, we observed a large variation in the amount of nectar secreted over 24-hour period (repeated measures ANOVA: F=7547; GL=6; P=0.001). Nectar production was significantly higher during the night (between 18 and 6 hours), decreasing drastically in the morning (10 hours), and remaining constant throughout the day (Figure 2).

Figure 2. Mean±standard error of nectar secreted per (mL) of A. verrucosa S. Moore fruit (Rubiaceae) over 24 h in a Neotropical savanna (repeated measures ANOVA: F=7.547; GL=6; P=0.001) (n=41 fruits).

4. Discussion

Several authors have demonstrated that the damage caused by predators is greater in young fruits, mainly due to their physical properties, which make them softer (Reuveni & Reuveni 1995; Alford 2007; Fernández et al. 2008; Wise & Hébert 2010). However, sometimes the predators do not have a special preference for young fruit (Clarke 1992; Lenzi et al. 2006; Silva et al. 2007), mainly because in some plant species the young fruits contain high levels of tannins compared with ripe fruits (Dement & Mooney 1974). Particularly in our study, we consistently observed an unidentified species of Coreidae (Hemiptera) with the rostrum inserted only in young fruits not tended by ants, probably feeding on the seed. Some plant species may abort their fruit damaged by herbivores or seed-sucking, avoiding subsequent spending of energy (Alford 2007; Boieiro et al. 2012). As we demonstrated, only a small amount of carbon (energy) was invested in young fruits. Thus, this smaller amount of energy allocated in young fruits can also justify the lower energetic investment in defenses (e.g. nectar) at this stage of fruit development.

Some studies show that in certain cases, the EFN secretion can be induced by jasmonic acid (Heil et al. 2001), a phytohormone which often increases during fruit development (Lopez et al. 1987; Czapski & Saniewski 1992; Creelman & Mullet 1997). In this context, such studies could support our initial hypothesis in which the EFN secretion can be higher in bigger and riper fruits of A. verrucosa. However, we observed that during fruit ontogeny of A. verrucosa, there are different patterns of nectar secretion, in which bigger and riper A. verrucosa fruits produced little or no nectar. The increase of nectar secretion in fruits of intermediate size demonstrates that the energy allocated in the fruit defense is proportional to the energy investment of plant in the fruit until some point in the fruit ontogeny. So why not protect fruit which invested much energy? The optimal defense hypothesis can be used to explain this pattern. Firstly, bigger fruits of A. verrucosa could be sturdier to specific herbivores of young A. verrucosa fruits, which could not access the seeds inside the bigger fruits. In fact, we did not observe Coreidae attacking bigger and riper fruits, even in the absence of ant foraging. It is possible that chemical modifications in the seed render them less attractive, or the thickness of the epicarp and mesocarp represents a physical barrier for herbivorous inaccessibility to the seed. Thus, with lower predation pressure by seed-sucking in bigger and riper fruits, reduction of investment in defense of these fruits would also be expected. Furthermore, there could be an exchange of defenses; when the fruit reaches an almost ripe stage, where more carbon is allocated in physical defenses causing hardening of fruit epicarp. This pattern could be similar to that observed in leaves that invest more in physical defense (e.g. leaf toughness) when they become fully expanded (Dyer et al. 2001; Korndörfer & Del-Claro 2006). However, the allocation of fruit mass in A. verrucosa was proportional to the fruit volume. Quantitative defenses, which imply an increase in the hardness of the fruit to the allocation of lignin or cellulose (Coley et al. 1985), are inevitably associated with a disproportionate increase in the investment carbon in relationship to the fruit volume, detectable by measuring of dry weight, which was not observed in this study.

However, the hypothesis of exchange defenses cannot be completely ruled out, since physical defenses or other chemical defenses could be involved in the defense of ripe fruits. In this case, the presence of such chemicals would be detectable only with the use of laboratory chemical analysis. Finally, the bigger fruits may be subject to predators and vertebrate dispersers (e.g. birds and mammals). In such cases, the presence of ants foraging on fruits may be nonfunctional, as ants may be inefficient in defending the host plant against some vertebrates (Hölldobler & Wilson 1990; Stapley 1998; Rico-Gray & Oliveira 2007). Moreover, in some cases, the presence of ants may even be undesirable, as they may frighten some possible disperses (Madden & Young 1992; Stapley 1998; Thomas 1988). A recent study performed by Yamawo et al. (2012b) showed that the plants can gradually shift from direct to indirect defense against herbivory during leaf aging. Here, we show that the likely change in defense tactics during the ontogenic stages of the fruits of A. verrucosa supports the ‘optimal defense theory’ (McKey 1974, 1979; Rhoades 1979), in which, plant tissues most closely linked to fitness, such as fruits, should be most defended at high and different levels. Similarly, working in the Sonoran Desert of North America, Holland et al. (2009) showed that the EFN production in Pachycereus schottii (Cactaceae) is greater in fruits than in buds. In fact, the framework of the optimal defense hypothesis has been corroborated by different studies involving ontogenetic or diurnal patterns in the secretion of nectar or other indirect defenses (Heil et al. 2000; Radhika et al. 2008; Rostás & Eggert 2008).

We observed higher nectar secretion in A. verrucosa fruits at nocturnal period, possibly mediated by both biotic and abiotic factors. Optimal defense theory can also help to explain this pattern, one most herbivores in regions with high temperature (e.g. Neotropical savanna) tend to be strictly nocturnal (Fagundes et al. 1996; Strauss et al. 2009; Byk & Del-Claro 2010). Moreover, due to the plants reduce greatly the nectar secretion when this resource is removed in low frequency (Koptur 1983; Pyke 1991; Heil et al. 2000), the ants' visitation on EFN of A. verrucosa may also have night foraging in arid and semi-arid regions (Briese & Macauley 1980; Oliveira et al. 1995). However, abiotic factors could also influence the nectar secretion of A. verrucosa. In fact, in regions with high temperatures throughout the day, most plants suffer water stress and can avoid nectar production during this period (Holland et al. 2010; Kulloli et al. 2011; Yamawo et al. 2012a).

In summary, our results indicate that A. verrucosa can change its energy allocation to defense against predation during ontogeny of its fruits. While in small fruits, where little energy was allocated, there is a low nectar production, fruits of intermediate size secreted higher amounts of nectar. This can increase the ants' visitation and consequently, the biotic defense offered by ants. However, in ripe and more energetically expensive fruits there is a strong decrease in nectar secretion, possibly in order to allow access to dispersers. Lastly, A. verrucosa may be altering the ants' visitation through increase in nectar secretion at night. This increase in ants' visitation may be crucial to reduce the herbivores foraging on A. verrucosa fruits in the Neotropical savanna. We suggest future studies to evaluate how ants' behavior, herbivores, and nectar seasonality influence this tritrophic interaction involving plant–ant–herbivore.

Acknowledgements

We are grateful to Víctor Rico-Gray, Ryan Smith, Kléber Del-Claro, Anselmo Nogueira, Haydée Yáñez, and two anonymous reviewers for their helpful comments and suggestions on earlier versions of this manuscript. J.C.F.F. was granted a master's degree fellowship from the CNPq. WD is grateful for financial support by the CNPq grant No. 237339/2012-9 and CONACYT.