Are the germination and vigor of Ficus and Cecropia seeds affected by the digestion process of a frugivorous bat?

Taking into account the importance of the bats as seed dispersers, the aim of this study was to evaluate the seed consumption of two arboreal species (Cecropia pachystachya and Ficus gomelleira) by a species of frugivorous bat (Sturnira lilium). As hypothesis, seeds of Cecropia and Ficus ingested by Sturnira are favored in the germination process under laboratory conditions. Bats were captured with mist nets and their fecal excretion were collected and submitted to the following treatments: (1) digestive system Cecropia Group, (2) digestive system Ficus Group, (3) Cecropia control Group, and (4) Ficus control Group. The passage of seeds through Sturnira lilum digestive system reduced the percentage of seed germination (19% – C. pachystachya and 16% – F. gomelleira) and vigor in both plant species. This result indicates that loss of seed viability occurs during the process of ingestion and digestion of seeds, reducing the percentage of germination. However, most seeds cross the digestive system without damage, resulting in good germination rates, which may allow the dispersion and establishment of Cecropia and Ficus propagules in other areas.


Introduction
Seed dispersal in tropical forests depends heavily on mammals and birds, a dynamic process that has great influence on the distribution patterns of plant individuals in the community. This process of transport and distribution of the seeds can provide an appropriate location for germination and seedling establishment (Fenner, 1985). Thus, seed dispersal is important for the maintenance of forests, as well as for recovering areas that have undergone human action (Galindo-González et al., 2000;Garcia et al., 2000). It is estimated that 50 to 90% of the species of fruiting trees found in tropical forests have their seeds dispersed by animals (Fleming, 1979;Howe and Smallwood, 1982).
Among the dispersion mechanisms, endozoochory, that is the dispersion after passing through the digestive system of vertebrates, is one of the most important, including the interaction known as dispersive mutualism, where the animals obtain nutrients from fruits that they consume dispersing the diaspores away from the parent plant, undergoing less intraspecific competition than when remaining close to the progenitor plant (Fenner, 2000;Galetti et al., 2006), besides decreasing the predation (Janzen, 1970). This process can also colonize recently open habitats or deposit the seeds specifically in microhabitats more favorable to their germination and establishment (Howe and Smallwood, 1982).
Among mammals, fruit-eating bats (Chiroptera) have been considered as efficient seed dispersers due their foraging behavior (Neuweiler, 2000;Preciado-Benítez et al., 2015), such as frequent exchange of feeding sites (Morrison, 1980), flights of several kilometers per night (Williams and Williams, 1970), consumption and rapid digestion of food during the pursuit trajectory (Fleming, 1982). In addition, the emptying of the intestines during the flight certainly increases the chance of adequate dispersion and the establishment of diaspores (Neuweiler, 2000), making them excellent dispersers of many species of plants, including pioneer and secondary plants (Galindo-González et al., 2000;Galletti et al., 2006;Munin et al., 2012). On the other hand, some bat species can also act as seed predators chewing and killing the embryo to obtain nutrients of seeds, as bat genus Chiroderma (Nogueira and Peracchi, 2003). However, little information is available regarding the negative effects of frugivorous bats on seeds of consumed plants taking into account the diversity of species in the tropics and their interactions.
The fruit-eating bat species Sturnira lilium (É. Geoffroy, 1810) (Phyllostomidae) occurs from the Antilles and Mexico to the northeast Argentina, Paraguay and Uruguay and throughout the Brazil. It is a medium-sized species (forearm: 42.0 mm; body mass: 21 g), with gray to orange coloring. It is considered an excellent seed disperser of several species of plants, especially pioneers, in the Neotropics (Simmons, 2005;Reis et al., 2007).
Sturnira lilium has been considered as a specialist in fruits of various species of Solanaceae (Marinho-Filho, 1991;Passos et al., 2003;Mello et al., 2008). Andrade et al. (2013) corroborate the hypothesis that Sturnira bats do not forage opportunistically, separating their diet hierarchically and selected fruits of their preferred plant genera first, even when other fruits are offered in abundance. This preference seems to be related to the fact that solanaceous fruits present a longer fruiting time (8 to 10 months), in relation to other species (Marinho-Filho, 1991). Fruits from other families, such as Cecropiaceae, Melastomataceae, Moraceae, Myrtaceae and Piperaceae, and nectar and pollen are also consumed by this bat species (Garcia et al., 2000;Passos et al., 2003;Munin et al., 2012). Among others, Sturnira bats visit Cecropia pachystachya Trécul (Cecropiaceae) and Ficus gomelleira Kuth & C. D. Bouché (Moraceae). Cecropia pachystachya is a pioneer and evergreen species, which is characteristic of moist soil on the edge of woods and glades (Lorenzi, 2008). Ficus gomelleira is a large evergreen species that belongs to the early secondary ecological group (Rolim et al., 1999).
Thus, this study was carried out taking into account the importance of Sturnira lilium as a seed disperser and the lack of information on whether there are some negative effects on seed dispersal and germination processes by this species. It aimed to evaluate the effect of digestion of the seeds of Cecropia pachystachya and Ficus gomelleira by S. lilium in the Rio Negro region of the Pantanal in Mato Grosso do Sul, Brazil.

Study area
The study was carried out in the Pantanal Research Institute (IPPAN) that has an area of 2,618 ha (19º33'20.9" south latitude; 55º36'47.19" west longitude) and is located in the Farm Santa Emília, Rio Negro Pantanal, municipality of Aquidauana, state of Mato Grosso do Sul, Brazil. It took place along the edges of vegetation areas, which presented Ficus and Cecropia trees, with the presence of ripe infructescenses at the time. The site is located in Pantanal wetlands, Aquidauana sub-region, which has approximately 70% of its surface occupied by the Cerrado biome, characterized by the predominant presence of woody vegetation (Pott et al., 2011).
According to Alvares et al. (2014), the region is localized in the Tropical zone and characterized as Awwith dry winter, presenting average annual temperature between 24 and 26°C and rainfall that ranges from 1300 to 1600 mm per year. From September to November, the maximum temperature exceeds 40°C and, in May, June and July, when there are cold fronts, the minimum average may drop below 20°C. The winter is dry (May to Septem-ber) with the driest months in July and August; the summer is rainy, with 80% of rainfall concentrated between November and March (Bononi et al., 2008).

Bat sampling
In June 2011, the capture activities were carried out with the capture effort corresponding to two sampling nights, each one with the aid of two nets of 7 x 3 m close to the ground (capture effort was calculated in accordance with Straube and Bianconi, 2002). Nets were set at 6h pm and remained open during six hours. The nets were nearby the trees with infructescenses and were inspected every hour when necessary. The specimens captured were removed from the nets and transported to the laboratory of the IPPAN, where they rested for about 30 minutes until the excretion of feces. This time is sufficient for processing the recently ingested food and evacuation of the intestines (Mikich, 2002). Feces were then collected and placed in paper bags.
At the end of the sampling, the specimens were released in the same sampling area, following the guidelines of the American Society of Mammalogists (Sikes and Gannon, 2011). Identification of captured specimens of S. lilium was confirmed based on Gannon et al. (1989). This species was already captured in the study region by Fischer et al. (2015). Analysis of the Ethics Committee for Animal was not necessary because the research was inserted inside the Laboratory of Chiroptera as standard procedures. Bat captures were allowed by license of capture -IBAMA License n o . 14/2005.

Seed collection and germination test
Initially, F. gomelleira and C. pachystachya trees were identified through direct observation, evaluating the family and genera level, confirmed through identification keys of Carauta (1989) for Ficus and Romaniuc Neto et al. (2009), for Cecropia. The ripe infructescenses were taken from different trees on the site and stored in paper bags. In relation to captured bats, 10 fecal units were obtained and inspected in order to find seeds of C. pachystachya and F. gomelleira, which were removed and identified. The collected seeds were transported to the Interdisciplinary Laboratory for Research in Environmental and Biodiversity Systems, University Anhanguera-Uniderp, where the experiment was conducted.
The seeds were placed in Petri dishes lined with two sheets of moistened germitest paper and an aqueous solution of 0.1% Ranol fungicide was used to control the infestation of pathogenic fungi in the seeds. The volume of the fungicide solution was equal to 2.5 times the weight of the substrate. These dishes were maintained in a growth chamber at 27°C, with periods of twelve hours of artificial lighting (fluorescent lamps). Four treatments were assembled, each one consisting of four replicates of 50 seeds each: (1) Cecropia Digestive System Group, with Cecropia seeds that passed through the bat digestive system; (2) Ficus Digestive System Group, with Ficus seeds that passed through the bat digestive system; (3) Cecropia Control Group, with Cecropia seeds extracted directly from plants; and, (4) Ficus Control Group, with Ficus seeds extracted directly from plants.
The counting of germinated seeds was daily (experimental time of 36 days), with seeds being considered as germinated when there was a protrusion of 2.0 mm primary root. At the end, the non-germinated seeds were subjected to a tetrazolium viability test (1% in aqueous solution) for assessing the viability of seeds after the test (Brasil, 2009).

Statistical treatment
Experimental design was completely randomized with four replications for each treatment. The percentage of germination was estimated according to the formula %G = (∑ni.N -1 ).100, where ∑ni corresponds to the total number of seeds germinated in relation to the number of seeds placed to germinate (N -1 ) (Borghetti and Ferreira, 2004). The relative frequency of germination was calculated by the formula Fr = 100.(ni/Σni), where ni is the number of germinated seeds per day and Σni, the sum of the total number of germinated seeds (Labouriau, 1983).
The seed vigor was also estimated, being indirectly measured by the germination mean time (GMT) estimated in days. The germination was quantified from the kinetic point of view, t = Σni.ti/Σni, where t indicates mean germination time, Σni = number of germinated seeds in a time interval and Σti = time interval in days (Laboriau and Agudo, 1987). The germination speed index (GSI) was calculated by the formula GSI = G1/N1+ G2/N2 + ... Gn/Nn), where G1 is equal to the number of seeds that had germinated at the first count, N1 is equal to the number of days elapsed for the first count, G2 equals the number of seeds germinated at the second count, N2 equals the number of days elapsed for the second count, and n, the final count (Maguire, 1962).
To know whether there is difference in the values of the seed germination parameters between Digestive System Group and Control Group, the non-parametric Mann-Whitney tests (U tests) (p < 0.05) were used. U tests were conducted using the Assistat statistical program, with independent tests for each plant species.

Results
Ten Sturnira lilium bats (four males and six females) were captured by a capture effort of 504 m 2 .h. Cecropia Control Group began the process of primary root protrusion on the first day of testing and Digestive System Ce cropia Group on the third day ( Figure 1) and presented significant differences in percentage of germinated seeds (U = 0, p = 0.014) and vigor (GMT and GSI: U = 0, p = 0.014) when compared to Cecropia Digestive System Group. Fi cus Control Group started the process on the seventh day, while Digestive System Ficus Group began on the 11 th day ( Figure 2) and similarly, presented significant differences in percentage of germinated seeds (U = 0, p = 0.014) and vigor (GSI: U = 0, p = 0.014 and GMT: U = 1, p = 0.029) when compared to Ficus Digestive System Group.
There was a lower percentage of germination of the seeds after their passage through the digestive system and slower germination over a longer period of time (decreased vigor) (Table 1). Both plant species had a polymodal distribution of germination type with multiple germination peaks (Figure 2). After the end of the test, the tetrazolium test indicated that the embryos were dead.

Discussion
The presented data suggest that the seed passage through the digestive system of Sturnira lilium leads to a reduction in germination percentage and a decrease in the seed vigor for both Cecropia pachystachya and Fi cus gomelleira. Moreover, the dead embryos observed in the end of the germination tests indicate a deleterious effect in germination of F. gomelleira and C. pachys tachya through their seed consumption by S. lilium. Similar results were obtained by Oliveira et al. (2013), with ingested seeds of Cecropia and Ficus by Platyrrhi nus lineatus (É. Geoffroy, 1810) in Rio Negro Panta- Table 1. Germination (%), germination speed index (GSI) and germination mean time (GMT) (in days) of Cecropia pachystachya and Ficus gomelleira before and after passage through the digestive system of Sturnira lilium. Means followed by the same letter in the column do not differ significantly by the Tukey test (p < 0.05). nal, demonstrating that this bat can alter the seed vigor in a negative way. The decrease in seed vigor, however, is not always considered a negative factor in the germination process because the reduction in vigor may mean that germination is being distributed in space and time, situation related to the characteristics of the disperser animals (Fenner, 1985). For this reason, the increase in GMT and the decrease in GSI obtained in this study and also by Sato et al. (2008) may mean a change in the distribution of germination over time. This is not necessarily negative, since the passage of Ce cropia and Ficus seeds through the digestive system could provide more favorable environmental conditions for their establishment than those found at the time they were dispersed (Lobova et al., 2003). This hypothesis is confirmed by the study of Brancalion and Marcos Filho (2008), who observed that the distribution of the germination over long periods favors an increased seedling survival rate.

C. pachystachya F. gomelleira
On the other hand, the decrease in the percentage of germination can be related to the possible damages that occur during the process of ingestion of the fruits. According to Barnea et al. (1992), damage may occur due to a great number of factors, such as the mastication of the seed and the pH of the digestive juice, characteristics which may influence the integrity of the ingested seeds. However, Tang et al. (2007), Oliveira and Lemes (2010) and Oliveira et al. (2013), working with consumed seeds of genus Cecropia and Ficus, demonstrated that most seeds normally pass through the bat digestive tract without damage and often did not lead to a change in the percentage of germination when compared with germination obtained from samples of fresh fruit seeds. Rossaneis et al. (2015) also demonstrated that the passage through the gastrointestinal tract of S. lilium produced no significant change in the rates of seed germination (Cecropia glaziovii and Ficus guaranitica).
Germination rates of the genus Cecropia are usually high, not needing for specialized dispersers so that their seeds can germinate in great quantity (Holthuijzen and Boerboom, 1982). However, Lobova et al. (2003) noticed that the passage of Cecropia seeds through the bat digestive system can increase the germination process because some part of fruit, as mesocarp and mucilage of seed, are removed during the passage. If there is no removal of these parts, the seed could fester in soil together with the fruit precluding the germination. Moreover, when the seeds are separated from the pulp by passing through the intestine of bats, this can increase their longevity by reducing the risk of predation, e.g., by ants or others insects. Tang et al. (2008) mention this situation with infructescenses of Morus macroura Miq. (Moraceae) consumed by fruiteating bats. Furthermore, the mature fruits that fall to the ground may deteriorate, and their seeds did not germinate (Tang et al., 2008). The loss of moisture, thereby reducing the attack by fungi, also improves germination (Heer et al., 2010).
Usually, the seeds of pioneer or early secondary plant species germinate quickly after being dispersed by the parent tree. However, some species of genus Cecropia do not present this pattern, because the soil seed bank, in certain regions, is largely composed of its seeds (Charles-Dominique, 1986;Vázquez-Yanes et al., 1996). According to Salomão et al. (2003), Cecropia seeds have tolerance to desiccation and cold, which could allow their maintenance in the seed bank over a period of time in order to wait for favorable period to germinate and establish the seedlings. These seeds can germinate up to nine years after dispersion (Lobova et al., 2003).
Regarding the genus Ficus, the germination rates of several species are usually low (Lorenzi, 2009), although the opposite has been observed in the present study, a fac- tor probably related to the region of origin of the seeds (Baskin and Baskin, 2001). In species with explosive fruiting, such as Ficus, some of the propagules have very little prospect of being effectively dispersed by animals because many seeds will not be removed from the parent plant (Fleming, 1982). In these cases, the disperser may contribute to the maintenance of a high percentage of diaspores germination, and thus may increase the chances of survival and establishment of the seedlings. According to Janzen (1970) and Connell (1971), the dispersion can promote the reduction of intraspecific competition by modifying the amount of dispersed seeds in a given location or if recruitment sites are far from the parent plant. This increase the probability of survival of seedlings (Janzen-Connell hypothesis). Even if the germination rate is decreased almost by half, as in the case of Platyrrhinus lineatus and Cecropia pachystachya (Sato et al., 2008), the bat can be considered an efficient disperser moving the seeds to large distance from their parents. For these reasons and according to the results obtained in this study, Sturnira lilium seems to present an appropriate behavior for an efficient disperser of Ficus gomelleira, transporting their seeds away from the parent plant.
Although frugivorous medium-sized bats, such as Sturnira lilium, tend to have a smaller home range than the large frugivorous bats (e.g., Artibeus litura tus) (Fleming et al., 1972), they probably do not always follow the same path in searching for food. Instead, they may increase their flight time searching for new sources of food, especially when the availability is spatially unpredictable, as is the case of Ficus (Fleming, 1982). Considering that bats that find a fig tree with ripe infructescenses may return for several consecutive nights for foraging (e.g., Artibeus ja maicensis) (Morrison, 1978), the mobility of the diaspores also increases because the bats usually return to their shelters every day (Fleming et al., 1972;Morrison, 1978;Fleming, 1982).
Plants can respond in markedly different ways to the process of ingestion of their fruits and the ingestion and digestion strategies used by dispersers (Fenner, 2000;Robertson et al., 2006). However, the potential for dispersal by bats is confirmed in many studies (Galindo-Gonzáles et al., 2000;Garcia et al., 2000;Lobova et al., 2003;Lopez and Vaughan, 2004;Henry and Jou ard, 2008;Heer et al., 2010;Oliveira and Lemes, 2010;Oliveira et al., 2013). These studies have also demonstrated that ingestion by fruit bats may be an important factor in regeneration processes and succession of forest areas.
Results indicate that loss of seed viability occurs during the process of ingestion and digestion of the seeds by Sturnira lilium, reducing the percentage of germination. However, most seeds cross the digestive system without damage, resulting in goods germination rates, which may allow the dispersion and establishment of Cecropia and Ficus propagules in other areas. Thus, this bat would be acting as a predator, but taking into consideration their ability to disperse the seeds, the bat could also be considered an efficient seed disperser. As S. lilium is a specialist in consuming fruits of the Solanum genus, it is possible that it acts as a predator of other plant species that are not Solanaceae -beyond those studied here.