Human disturbances have reduced the three basic niche dimensions for the survival of some mammalian species (space, time and resources). Many studies consider that anthropic activities cause the extinction of ecological functionalities faster than the species extinctions. The effects of human disturbances on the niche structure and ecological relationships of bat communities are fairly unknown.  Based on management use, we identified 4 types of habitats in the Ecuadorian Andean Chocó: primary forest (PF), secondary forest in natural regeneration (SF1), secondary forest in assisted regeneration (SF2) and pastures for cattle (P). We used these habitats to determine the responses in the trophic niche and the ecological relationships of understory fruit bats. Our results show that niche overlap is higher in disturbed habitats (SF1, SF2, and P), while ecological relationships are weaker. These results mean that bat communities and the ecosystem services that these animals provide are highly endangered due to anthropogenic disturbances.

Sampling was organized according to the lunar calendar periods, avoiding full moon periods because many species reduce their activity patterns to lower predation risk. There were 12 sampling nights in February, 11 in March, and 12 in May. During these months, we captured bats in the primary forest of Mashpi Lodge (PF), the secondary forest in natural regeneration (SF1) in Pambiliño and Chontaloma, and the secondary forest in assisted regeneration (SF2) in Chontaloma. Additionally, between July and August 2021, a 9-night sampling trip was carried out between the Mashpi Shungo and Chontaloma reserves. In Mashpi Shungo, we sampled in primary forest and SF1, whereas in Chontaloma, we sampled in pastures for cattle (P).

For the capture sessions, we used 4-7 mist nets of variable sizes (6, 9 and 10 meters), opened between 19:00 pm to 00:30 am (Arias and Pacheco, 2019; Novoa et al. 2011). We placed capture stations on closed trails, mountain ridges, near water sources, and potential flight sites where bats could be intercepted (Kelm et al. 2008). Nets were checked in 15 to 20 minute intervals. Bat species were identified with the help of field guides (López-Baucells et al. 2016; Tirira, 2017). To obtain seeds, the animals were kept for an hour in cloth bags. Fecal samples obtained were preserved in 2 ml eppendorf tubes with 70% alcohol (Arias and Pacheco, 2019). We collected voucher specimens according to the guidelines of the American Society of Mammalogist (Sikes and Gannon, 2016). The specimens were deposited with their respective reference data in the Instituto Nacional de Biodiversidad (INABIO) under the research license with code MAE-ARSFC-2020-0.512 granted by the Ministerio del Ambiente Agua y Transición Ecológica del Ecuador (MAATE) (Appendix 1).

Counting seeds

The seeds collected were separated from the fecal remains in the laboratory with entomological tweezers. Once separated, they were dried on paper envelopes and then photographed with a stereomicroscope (Arias and Pacheco, 2019). To count very numerous seeds, such as those of the genus Piper, we used the app CounTThings from Photos (Dinamic Ventures INC, 2017), and for cases in which this application could not differentiate seeds from other fecal residues, we used ImageJ (Wayne Rasband (NIH), 2019), which allows counting the seeds based on the size of their pixels. Seed identification was performed to the lowest taxonomic level possible using field guides (Kirkbride et al. 2006; Lobova and Mori, 2007). Counts were made for each bat, then grouped by genus or morphospecies and type of habitat (PF, SF1, SF2, P). The seed collection was deposited in the MECN to act as a reference guide for other studies.

Counting pollen

Samples obtained from nectarivorous bats were spread on microscope slides, stained with methylene blue, and finally photographed (Arias and Pacheco, 2019). We used the same applications described before to count each grain of pollen. In a similar way, counts were made by individual, then grouped by pollen (in general) and type of habitat.

Counting insects

We counted how many insects existed in each fecal sample. We identified each insect order by comparing the morphology and anatomy of their wings, mouthparts, heads, eyes, antennae, and other structures with field guides (Borror and White, 1970; Pokhrel and Budha, 2015). Insect remains that could not be identified (undetermined insects) were counted as a single food item based on their presence, adding 1 to the count if this type of remains were present. However, all the counts were summarized as insects and for each type of habitat.

Counting vertebrates

Some omnivorous bats usually feed on small vertebrates (birds, frogs, rodents and other bats), fecal remains of these bats usually present fur, bones, or feathers (Gardner, 1977). We compared these remains with biological material available at INABIO. We counted the different types of preys identified in fecal samples for each individual, then counts were summarized by vertebrates and type of habitat.