Reproductive pattern of the large fruit-eating bat, Artibeus amplus (Chiroptera: Phyllostomidae) in the Venezuelan Andes

Bimodal polyestry is the most common reproductive pattern in tropical bats, and it consists in producing one offspring per female twice a year. Reproductive patterns are closely related to rainfall regimes, frequently occurring twice a year in tropical regions. The goal of our study was to determine the reproductive pattern of the large fruit-eating bat, Artibeus amplus Handley, 1987 in a cave in the Venezuelan Andes inhabited by a large, stable colony. Thus, in this study we describe for the first time this important biological aspect of this unknown Neotropical bat species through the examination of external reproductive characteristics of males (inguinal or scrotal testes) and females (pregnant, lactating, post-lactating), based on 211 individuals (120 males and 91 females) captured between September 2008 and August 2009, in Cueva del Parque Las Escaleras, Estado Táchira, Venezuela. During this period of monthly sampling for a full year, most males displayed large scrotal testes, averaging 10 mm maximum length. The examination of females indicated that although pregnancy was first observed in November 2008, it reached a maximum during January and February 2009. Although adult males with scrotal testes throughout the year could imply that females have more than one pregnancy, our results suggested a seasonally monoestrous reproductive pattern for A. amplus. This study represents the first report of reproductive pattern for this poorly-known Neotropical frugivorous species. The observed monoestrous reproductive pattern supports the existence of synchronization between precipitation and reproduction. This synchronization has been frequently observed in most species of fruit bats. In this study, Artibeus amplus presumably adjust the parturition to anticipate the rainy season, as a strategy that allows maintenance of offspring during high availability of fruits. Rev. Biol. Trop. 65 (1): 335-344. Epub 2017 March 01.

The most common reproductive pattern in tropical bats is bimodal polyestry (Wilson, 1979;Racey, 1982;Racey & Entwistle, 2000), which consists in producing one offspring per female twice a year. Although the control of reproductive patterns is driven by hormonal changes in females (Bronson, 1985;Brown, 1999;Martin & Bernard, 2000), a common strategy in males is to remain fertile throughout the year once they reach adulthood (Racey, 1982). In frugivorous bats, it has been determined that reproduction is closely related to food availability, which in turn depends on climatic conditions such as temperature and precipitation (Wilson, 1979;Racey, 1982;Heideman, 2000). Thus, bimodal polyestry is closely related to a rainfall regime that occurs twice a year (Wilson, 1979;Racey, 1982;Racey & Entwistle, 2000).
In Venezuela, A. amplus has been found in a few roost sites and we were aware of a large, stable colony in one of them. We hypothesize that the reproductive pattern of A. amplus will reflect a clear relationship with the rainfall regime of the study site (a single maximum yearly), and this may result in a different pattern from most Artibeus species (i.e., bimodal polyestry). Thus, the goal of our study was to determine the reproductive pattern of the large fruit-eating bat, in a cave in the Venezuelan Andes inhabited by a large, stable colony.
Reproductive data presented here constitute the first study to describe this primordial aspect of the biology of this unknown species of Neotropical bat.

Study site:
The studied population of A. amplus was found in Cueva del Parque Las Escaleras (8°00'18.8" N -71°43'41.8" W, 1 320 masl), approximately 4 km SE of Pregonero town, Municipio Uribante, Estado Táchira, Venezuela (Fig. 1). The cave is frequently visited by people because it is located within the recreational park. The vegetation of the area mainly consist of human-disturbed habitats, as well as open savanna of oligotrophic origin, and small patches of secondary forest surrounding rivers and creeks (Ramoni-Perazzi et al., 2014). Monthly mean temperature of the study site is 20°C (Molina-Chacon, 1983). Annual precipitation averages 1 636.9 mm (n= 49 years, from the National Institute of Meteorology and Hydrology; INAMEH, http:// www.inameh.gob.ve), occurring in a unimodal pattern, with a period of drought from January to March, and a period of water availability the rest of the year, with maximum in July and August (Molina-Chacon, 1983). The cave is ca. 38 m long and it is divided into three consecutive chambers. Individuals of A. amplus occupied the second and third chambers (ca. 21 m from cave entrance), both about 2.5 m in height.

Sampling of individuals:
All sampling protocols were performed following guidelines of the American Society of Mammalogists for capture, handling, and care of mammals (Sikes, Gannon,& the Animal Care and Use Committee of the American Society of Mammalogists, 2011). Bats were captured monthly from September 2008 to August 2009, inside the cave between 13:00 h and 17:00 h, using hand nets and mist nets (Kunz, Hodgkison, & Weise, 2009). Individuals were examined to determine sex, body mass (Ohaus Cs 200.0 gx 0.1 g), and standard body measurements (Spi 2000 Plastic Dial Caliper ± 0.1 mm). Age (juvenile, immature, or adult) was estimated following common criteria (Brunet-Rossinni & Wilkinson, 2009), as well as reproductive condition (Racey, 2009). Males with scrotal testes and pregnant, lactating and post-lactating females were classified as reproductively active individuals (Racey, 2009). Post-lactating females presented large, hairless, flaccid, and dark nipples. Females lacking these characteristics were considered as non-reproductive (inactive). We are aware of the limitations of observation of external characteristics that may not be able to detect early pregnancy or actual reproductive conditions (Vela-Vargas, 2013). Maximum length of testis was also measured (Spi 2000 Plastic Dial Caliper ± 0.1 mm) to assess changes throughout the study. Individuals were marked ) on the right (males) or left (females) forearm using aluminum wing bands (Gey Band & Tag Co., Norristown, PA, USA, size 4, Style 374). Finally, all individuals were released within the cave. Statistical analysis: Generalized linear model analyses (GLM) were used to determine whether variation in maximum length of testes was significant, controlling for body mass, and using R (R Core Team, 2015). All statistical tests considered α = 0.05. Variation in reproductive condition of males and females during the study was established by monthly percentages of bats with each reproductive condition.

Male reproductive condition:
The proportion of males with scrotal testes was greater throughout the study than that of males with non-scrotal testes (Fig. 2). The highest proportion of males with non-scrotal testes (38.5 %) was recorded in November 2008. All examined males had scrotal testes during September and December 2008, and from April to August 2009 excluding June (Fig. 2). Scrotal testes had an average length of 10.0 ± 1.7 mm (n= 106) (Fig. 3A). The maximum length of testes observed was 13.3 mm, while the minimum was 4.9 mm (Fig. 3A). The variation in maximum length of testes showed significant differences during the study, regarding time (GLM, F= 10.527; P <0.001), and individuals (GLM, F= 2.673; P<0.001). These differences are shown by four frequently recaptured males, which exhibited changes in testes length during the study period (Fig. 3B).

Female
reproductive condition: Although the variation in female reproductive condition during the study indicated first signs of pregnancy in November 2008, maximum number of pregnant individuals occurred during January and February 2009, in which all examined females were pregnant (Fig. 4). Pregnant females were observed from November 2008 to May 2009, and the lowest observed proportion was 9 % (April 2009). After the period of maximum pregnancy (January-February 2009), lactating (Fig. 5) and post-lactating females were briefly observed (Fig. 4). The highest observed proportion of post-lactating females was 27.8 % (October 2008). After the period of maximum pregnancy (January to March 2009), the capture of pregnant females decreased abruptly. All captured females were considered reproductively inactive in June and August 2009 (Fig. 4).

DISCUSSION
Most males of A. amplus at Cueva del Parque Las Escaleras had scrotal testes. This permanent testicular descent has frequently been observed in other Artibeus such as A. jamaicensis and A. lituratus in an arid region of the Venezuelan Andes (Sosa & Ramoni-Perazzi, 1995), and in A. lituratus in Brazil (Duarte & Talamoni, 2010). These last authors also examined testicular activity through spermatogenic analysis and indicated that males were reproductively active and fertile throughout the annual cycle. Maintaining scrotal testes throughout the year ensures availability of sexually mature males to reproduce anytime during the reproductive cycle (Bronson, 1985;Krutzsch, 2000;McCracken & Wilkinson, 2000), although the cost of this condition requires high energy (Krutzsch, 2000;Speakman, 2008). The presence of male A. amplus able to reproduce anytime during the annual cycle in our study area is probably an strategy to ensure pregnancy in this species at any time that females may come into oestrus.
The values of maximum length of testes seem to increase from February 2009 in three of the four frequently-recaptured individuals, although this trend needs further research based on extensive recaptures. Variation in maximum length of testes in the frequently-recaptured males reinforces the generalized linear model analyses with regards to the existence of significant monthly variation of testes lengths, with different trajectories for each individual, which might suggest differences between males, but this trend needs further research.
We are aware of the limitations inherent to the examination of external reproductive characteristics, as these might not be sufficiently informative to detect early stages of females' reproductive cycle (Vela-Vargas, 2013). The only reproductive event observed (i.e., maximum pregnancy during the driest period in January-February) indicated a monoestrous reproductive pattern for A. amplus at Cueva del Parque Las Escaleras. Reproductive events are frequently synchronized to climate (Opler, Frankie, & Baker, 1980;Sobral & Machado, 2001). This synchronization of reproduction with maximum rainfall, frequently reported in tropical regions of South America, has been observed in most species of fruit bats (Willig, 1985;Dinerstein, 1986;Duarte & Talamoni, Fig. 4. Reproductive condition of adult females of Artibeus amplus (n=91), based on external reproductive characteristics, captured in Cueva del Parque Las Escaleras during the study. Inactive (striped lines, ascend), pregnant (white), lactating (grey), and post-lactating (striped lines, descend). Solid line indicates precipitation (mm). 2010), with most species exhibiting a polyestrous reproductive pattern (Wilson, 1979;Willig, 1985;Molinari & Soriano, 2014). However, monoestrous patterns, such as that observed in this study, have also been reported for Neotropical frugivorous species such as A. lituratus (Duarte & Talamoni, 2010) and Anoura geoffroyi (Baumgarten & Vieira, 1994;Zortéa, 2003) in populations subject to unimodal rainfall regimes in Brazilian localities.
The Venezuelan Andes are an orographic barrier between two regimes of precipitation (Monasterio & Reyes, 1980): the Llanos versant has the unimodal rainfall pattern typical of the North of South America, as described in Hastenrath (1984), while the Maracaibo Lake versant has the bimodal pattern, typical of the Caribbean region as described in Taylor and Alfaro (2005). Here, our results highlighted the existence of synchronization between precipitation and reproduction at least among large-size Artibeus species, on both sides of this Cordillera. The monoestrous pattern observed here for A. amplus on the Llanos versant contrasts with those polyestrous patterns reported by Sosa and Ramoni-Perazzi (1995) in a locality on the Maracaibo Lake versant, where rainfall has two peaks in April-May and September-October, A. jamaicensis showed reproductive maxima in May and in October, and A. lituratus showed reproductive maxima in March and July (Sosa & Ramoni-Perazzi, 1995).
In this study, Artibeus amplus presumably adjust the parturition to anticipate the rainy season, as a strategy that allows maintenance of offspring during high availability of fruits (Fleming, Hooper, & Wilson, 1972). In the study area, A. amplus could be favoring the greatest survival rate for the offspring, since the critical lactation period would then occur at a time of maximum food availability.
This study supported not only a clear relationship of reproduction with rainfall regimes, but also provides the first report of reproductive pattern of a Neotropical bat species that has remained unknown since its discovery, back in 1987. However, several aspects of its biology still remain to be determined. Some of these could be the detailed study of reproductive conditions of males and females based on the examination of hormonal profiles and histological conditions of their reproductive tracts, and the determination of whether giving birth to a single offspring/year in A. amplus in the study site is related to delayed development, which was reported in other large Artibeus (Fleming, 1971). Finally, does seasonal polyestry could occur in A. amplus under different rainfall regimes, as in other large Artibeus?