Zoochory is one of the most important dispersal strategies of plants, with about 50 to 90% of trees dispersing their seeds by animals in tropical forests and savannas (^{Fleming 1987}; ^{Vieira et al. 2002}; ^{Jordano et al. 2006}). Seed dispersal that occurs after passing through the digestive tract of an animal (en dozoochory) is an important mutualistic relationship, where animals remove energy from fruits and the plant’s propagules are dispersed (^{Van Der Pijl 1982}; ^{Galetti et al. 2006}).

Seeds that pass through the digestive tract of frugivorous may improve their germination rates, due to the removal of germination inhibitors and the rehydration of tissues by chemical scarification (^{Traveset & Verdú 2002}; ^{Robertson et al. 2006}).

Between the frugivorous vertebrates, bats are recognized as one of the most important in the Neotropics (^{Galetti et al. 2006}; ^{Jordano et al. 2006}; ^{Muscarella & Fleming 2007}; ^{Fleming et al. 2009}). They are excellent seed disperser of pioneer and early secondary plant species, making them important for the secondary succession process, allowing the seeds flow from conserved to modified areas (^{Fleming 1987}; ^{Jordano et al. 2006}; ^{Galetti et al. 2006}; ^{Fleming et al. 2009}).

The family Phyllostomidae is widely distributed in the Neotropics with 160 species; and 120 species recorded as frugivorous and/or nectivorous (^{Simmons 2005}). Frugivorous phyllostomids are recognized as efficient seed disperser in tropical forests (^{Fleming 1987}). Between them, Carollia perspicillata (Linnaeus 1758) is one of the most widely distributed bats in the neotropics. It is a medium-sized bat (forearm ± 42.0 mm; weight 10-23 g) (^{Peracchi et al. 2006}) and an important frugivorous generalist, feeding mainly of Cecropia, Ficus, Piper and Solanum species (^{Garcia et al. 2000}; ^{Passos et al. 2003}; ^{Carvalho-Ricardo et al. 2014}; ^{Parolin et al. 2016}).

Bats can maintain mutualistic relationships with these species (^{Fleming et al. 2009}; ^{Parolin et al. 2016}), inducing seed germination rate (^{Heer et al. 2010}) or increasing speed of seed germination (^{Bocchese et al. 2007}). But in other cases, there is no effect or even a decrease over seed germination (^{Tang et al. 2008}; ^{Oliveira et al. 2013}).

Considering the importance of frugivorous bats for the dispersal of pioneer species and their role in the regeneration of degraded environments, here we investigate if the Neotropical frugivorous bat C. perspicillata can act as a germination inducer of a pioneer tree species, Cecropia pachystachya Trec. (Urticaceae). The species C. pachystachya is a pioneer tree with fast growth, heliophite and selectively hygrophilous. It occurs frequently in disturbed areas or in initial stages of ecological succession (^{Pott & Pott 1994}). We sought to answer the following questions: Are the seed germination induced or reduced after bat digestion? Are the seed viability maintained after storage?

The study was carried out in the Quinta do Sol Privately Owned Nature Reserve (PONR) (Latitude 19°46’34.81”; Longitude 55°14’46.28”), at Taboco region, municipality of Corguinho, Mato Grosso do Sul, Brazil (Fig. 1). The typical native vegetation of the region consists of savanna physiognomies, where the woodlands, with crown cover of 50% to 90%, made up of trees with 8 to12 m tall, predominate, and with gallery forest occurring along streams (^{Oliveira-Filho & Ratter 2002}). However, most of the landscape region is composed by planted pastures with Brachiaria humidicola (Rendlle) Schweick, used for beef cattle raising.

In October 2010, two nets (7x3 m) were set up near the ground in a riparian area. The nets were deployed two consecutive nights at 6:00 pm, and stayed open during six hours long, being inspected every hour when necessary. The catching effort was calculated as ^{Straube & Bianconi (2002)}, where E = 504 m².h. Seven bats of Carollia perspicillata species were captured. The identification of the captured bats was made based on ^{Vizotto & Taddei (1973)} and ^{Díaz et al. (2016)}; for the region under study only C. perspicillata was cited (^{Fischer et al. 2015}), among the four Carollia species existing in Brazil (^{Simmons 2005}).

In the same occasion of bat captures, we collected mature infructescences of Cecropia pachystachya from 10 individuals in different sites. The captured bat specimens were left for two hours in cotton bags for their feces disposal and then transported to the field laboratory and placed in cages (41x34x16 cm polypropylene) covered with dark cloth, containing the infructescence of C. pachystachya and a cotton ball moistened with water. Each bat was subjected to the experiment only once in individual cages. The animals were trapped until the following day and released at dusk at the same location where they had been captured. Feces samples on the cage floor were collected and packed in paper envelopes. Seeds from excreta as well as fresh seeds from infructescence (not digested by bat) were separated with the aid of tweezers and stereoscopic microscopes and stored in sterile petri dishes.

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 no. 14/2005.

To verify if C. perspicillata are a germination inducer of C. pachystachya seeds, we compare the germination of different treatments: (1) with fresh seeds (control); (2) seeds treated with pH 2 acid solution (pH 2); (3) seeds treated with pH 3 acid solution (pH 3) and (4) seeds collected from feces (digestive tract). For pH 2 and pH 3 groups, fresh seeds were subjected to chemical scarification through the seed immersion in hydrochloric acid (HCl) during 1 minute in pH 2 and pH 3 solutions respectively (^{Bocchese et al. 2007}).

Fig. 1 Location of Environmental Estancia Quinta do Sol, Taboco region, municipality of Corguinho, Mato Grosso do Sul, Brazil. 

All treatments consisted of four replicates with 25 seeds each, which were placed in Petri dishes (seven centimeters in diameter) on two sheets of filter paper and moistened with a 0.2% aqueous solution of Rovral fungida (volume of solution equivalent to 2.5 times the weight of the substrate). The dishes were kept in a germination chamber at a constant temperature of 30°C, with periods of 12 hours of artificial lighting (fluorescent lamps). Germinated seeds, those with a minimum protrusion of two mm of primary root, were counted daily for six consecutive days (^{Oliveira et al. 2013}).

The remained seeds not used in germination tests (fresh seeds and feces seeds) were placed in individual Petri dishes and stored under laboratory conditions (temperature of 28.5°C ± 1.0°C and relative humidity between 60 and 70%). To verify the seed viability (for both fresh seeds and feces seeds) along time, we carried out germination tests every 30 days after seed collection, from November 2010 to February 2011, following the same procedures described above.

For each replicate, in all experiments, we calculated final germination percentage G% = 100^. Σ ni · N^-1 and mean germination time MGT = Σ n_i^.t_i^.Σni ^-1 whereΣn_i is the amount of germinated seeds in relation to the number of seeds (N) placed to germinate and n_iis the number of germinated seeds within the time interval t_i -1 and t_i (^{Ranal & Santana 2006}). The germination percentage and mean germination time data were analyzed with one-way ANOVA and the germination percentage of storage essay was analyzed with two- way ANOVA. The Tukey test (p < 0.05) was used to compare means.

There were significant differences between treatments in germination percentage (F =99.6; p<0.001), and MGT (F =173.3; p<0.001) (Table 1). Seed germination from C. perspicillata feces was 24% lower than seeds from infructescence. There was no significant difference in germination among seeds from feces and pH 2 treatment. Seed germination from pH 3 treatment differed from feces’ seeds but was statistically equal to pH 2 treatment. MGT increased for all treatments in relation to control seeds (Table 1).

Table 1 Means (± standard error) of germination percentage (%) and mean germination time (MGT ) of Cecropia pachystachya seeds for fresh seeds (Control), seeds from Carollia perspicillata feces (Feces) and chemically scarified seeds (pH 2 and pH 3). 

There was significant interaction between treatment and storage time (F =144.0; p<0.001). Storage affected seed germination only in seeds from C. pachystachya infructescence, with germination percentage remaining smaller in seeds from feces in relation to seeds from infructescence even after four months of storage (Fig. 2).

Our results indicate that the bat C. perspicillata was not a germination inducer of C. pachystachya seeds. However, seed germination remained high after bat ingestion (76%). A similar study reported that seed germination percentage of C. pachystachya not differed between seeds from infructescence (48%) and from C. perspicillata feces (57.5%). However, seed germination of another Cecropia species (C. glaziovii) was increased after ingestion by C. perspicillata (from 51.5 to 98%) (^{Rossaneis et al. 2015}).

Regarding to other bat species, seeds of C. pachystachya ingested by Artibeus lituratus, Platyrrhinus lineatus, and Sturnira lilium, had not their germination percentage changed in relation to seeds from infructescence (^{Bocchese et al. 2007}; ^{Oliveira et al. 2013}; ^{Rossaneis et al. 2015}).

The Piper genus is indicated as the main food item of C. perspicillata (^{Parolin et al. 2016}). Seeds of this genus that have been ingested by C. perspicillata not changed or increased their germination for Piper aduncum, P. hispidinervum, and P. amalago (^{Garcia et al. 2000}; ^{Rossaneis et al. 2015}). Moreover, the germination percentage of Solanum americanum, S. mauritianum, S. granuloso-leprosum, and Ficuseximia seeds that are ingested by C. perspicillata,also did not differ from seeds extracted from fruits (^{Rossaneis et al. 2015}).

Although MGT was in general around four days (Table 1), the slight increase in MGT for seeds from feces and pH treatments may indicate some physiological change in the germination process. Similar results were notice in other studies where C. pachystachya seeds have MGT increased after P. lineatus and S. lilium ingestion (^{Oliveira et al. 2013}; ²⁰¹⁸).

The increase of MGT of seeds after bat ingestion has been interpreted for some studies as being a good aspect. Because a longer MGT may represent a germination distribution in time (^{Tang et al. 2008}), it may allow seedlings to reach suitable conditions for their establishment in different periods, which increases survival in environments that suffer temporal variations in their abiotic factors (^{Brancalion & Marcos Filho 2008}).

The distinct results of seed germination responses after bat ingestion found here and in other studies may be result of several factors. Such as differences in the pH of the bat species’ digestive system, digestion time, the amount of food consumption, and fruit consumption behavior (e.g. chewing), among others. The statistic similarity of pH 2 treatment and the feces treatment may indicate that the reduction in germination percentage of seeds from C. perspicillata feces is a result of stomach acids from digestive tract of bats. In general, the stomach of most mammals has a pH around 2, because the enzyme pepsin, which together with hydrochloric acid-releasing cells, promotes acidification (^{Schmidt-Nielsen 2002}).

Seed storage reduced germination of seeds from C. pachystachya fruits, while seeds from C. perspicillata feces were not affected by storage period tested. It may indicate that the passage through bat digestive tract affected only seeds that which had low quality. Moreover, the removal of the pulp by passage through digestive tract may also reduce predation risk by insects and/or decrease fungal infection (^{Tang et al. 2008}; ^{Heer et al. 2010}).

Cecropia pachystachya, a tree characteristic of the early stages of succession is one of the most abundant tree species in soil seed bank (^{Holthuijzen & Boerboom 1982}; ^{Grombone-Guaratini & Rodrigues 2002}). Therefore, the large proportion of viable seeds from C. perspicillata feces after four months of storage (68%) may indicate that these seeds can maintain their viability in soil seed bank even after bat dispersion during a period of time.

Fig. 2 Germination percentage means (± standard error) of Cecropia pachystachya seeds from fruits and Carollia perspcillata feces, just after collection (October) and after four months storage (February). Different letters indicate significant difference among means according to Tukey test (p<0.05). 

A recent study reported that C. perspicillata consumed 40.7% of plants recorded in that study, and presented the highest value of the dispersion index (DII = 1.536) and the highest abundance of seeds consumed (^{Casallas-Pabón et al. 2017}). Therefore, C. perspicillata may be an important seed disperser of Cecropia species and another plant species in several areas of the Neotropics. In fact, our results confirm the evidences of ^{Lobova et al. (2003)} that the bat dispersal is not necessary for germination of Cecropia seeds, but it may increase seed survival. Like other frugivorous bats (^{Lobova et al. 2003}), we conclude that C. perspicillata is an effective disperser of C. pachystachya seeds.