Wildlife Acoustics | Bioacoustic monitoring systems for research, science, industry and governments.

Effectiveness of artificial roosts for Neotropical bat species

Dr. Michael Schöner
University of Costa Rica

Effectiveness of artificial roosts for Neotropical bat species Effectiveness of artificial roosts for Neotropical bat species Effectiveness of artificial roosts for Neotropical bat species Effectiveness of artificial roosts for Neotropical bat species

Update 1

The aim of our project, effectiveness of artificial roosts for Neotropical bat species is to efficiently attract especially vulnerable bat species to artificial roosts via sensory cues. Currently, we are running pre-experiments to find the most promising sensory cues before we start opening the artificial roosts. As a model species we chose wrinkled-lipped bats (Trachops cirrhosus). Using Wildlife Acoustic SM4s we recorded individuals in their roosts during the early morning hours when the individuals start coming back to their roosts. As groups of T. cirrhosis usually roost in relatively stable groups, we assume that conspecifics attract each other via calls. We then caught individuals in their roosts and brought them to a flight arena where we had set up two artificial roosts. From one we played back purely social calls (contact calls), from the other a mix of contact and echolocation calls was broadcasted. So far we found that although the contact calls are more attractive to the individuals (as indicated by the number of approaches to the referring artificial roosts), in the end the bats preferred to stay inside the roosts where the mixture of calls had been broadcasted. Moreover, we tested if olfactory cues also might be effective in attracting the individuals. As T. cirrhosis is usually roosting in mixed species groups, e.g. with Micronycteris spp., we tested pure T. cirrhosis feces against a feces mix of different bat species that we had freshly collected inside the roosts. It seems that the T. cirrhosis individuals prefer feces of their own species. We will continue testing further sensory cues until we have a large enough sample size to ensure that we collected the best working ones. After that we will start opening the artificial roosts in the wild to experimentally test which sensory stimulus (acoustic, olfactory or a mixture) is most reliable in attracting wild bats to novel roosts.

Update 2

With our project, Effectiveness of artificial roosts for Neotropical bat species" we want to attract bats to artificial roosts via sensory cues. For this aim, we use fringe-lipped bats (Trachops cirrhosus) as study species. Currently, we are still running pre-experiments to find the most attractive acoustic and olfactory cues. This has turned out to be more challenging than previously thought. To find optimal cues, we are running pre-experiments in a flight cage where individual T. cirrhosus can select between roosts with different acoustic cues (playbacks of T. cirrhosus social calls vs. a mix of social and echolocation calls). However, we needed to change the acoustic set-up several times and are still searching for the best possible acoustic design. Because finding an attractive sensory lure is so critical to the success of the project, we are investing considerable time and care in these pre-experiments to ensure a well-working study design once we open the artificial roosts. Additionally, we have started to conduct a second type of pre-experiments in old bunkers that are frequently used as roosts by wild-ranging T. cirrhosus. During the last months, we rarely recorded any individuals of T. cirrhosus with the SM4s inside the bunkers but our impression is that the individuals nevertheless preferred those bunkers where we played back a mix of social and echolocation calls of conspecifics compared to control bunkers without any playbacks. We will need more time and more recordings of individuals inside the bunkers to evaluate this statistically.

Update 3

During the last months we were very successful and could record many videos from the pre-experiments, which aim to evaluate the most promising sensory cues to be tested in the field.

I am currently analysing the videos, which takes a lot of time as one video lasts 8 to 12 hours. In a flight cage we have set up two artificial roosts from where different sensory cues are displayed (e.g., pure Trachops social calls vs. a mix of social and echolocation calls). For the video analysis I evaluate which roost/sensory cue is visited most often and which eventually serves as roost, i.e. where does the bat stay in the end. I will then run the according statistics.

Update 4

Effectiveness of artificial roosts for Neotropical bat species

We are about to finish the pre-experiments within the next one or two weeks. My colleagues could catch a lot of Trachops individuals for the pre-experiments and most of them were cooperative. When broadcasting bat calls with a BatLure from out of artificial roosts inside a flight cage, it seems that the Trachops prefer the mixed calls (social + echolocation calls) over the pure social calls. This would be close to what happens in nature, where there are also no pure calls of a certain type (social vs. echolocation call). We will need a few more of these pre-experiments to get significant values.

Situation is different when we are testing olfaction cues in further pre-experiments. Here, the bats randomly choose one of the presented olfactory cues (pure feces from Trachops vs. a mix of feces from Trachops and other bat species that can be found roosting together). It is difficult to say if olfaction does not play a role for roost selection or if the Trachops simply do not care. I hope that the field-experiment will tell us more on this question. For the field experiment, we are as well going to use a feces mix (Trachops + other bat species), as this resembles the most natural situation.

Finally, as we are so close to start with the field experiment, we are testing the set-up within the artificial roosts in our study sites and how to arrange the devices inside the roosts (e.g., we need to avoid that the SM4s constantly record the bat calls emitted by the BatLures. I have attached a short graph on how we are planning the set-up). This way, we will be able to directly start with the field-experiment once the last pre-experiment has been conducted.

I am very confident that we can finally start with the field-experiment mid of November.

Final Report

Effectiveness of artificial roosts for Neotropical bat species

Fig. 1. Artificial bat roost.

Habitat loss, conversion and fragmentation have posed major threats to bat species worldwide. This is especially true for bats in tropical rainforests. Many tropical bat species rely on intact forests not only for foraging but also, critically, for roosting. Several studies from Europe and North America have tested the effectiveness of artificial bat roosts (ABRs) as surrogates for the tree cavities found in older growth forests, critical roosting resources that are often lacking in forests suffering from anthropogenic impact. In contrast, in the tropics ABRs have rarely been offered to bats; knowledge on their effectiveness is lacking. Our project, “Effectiveness of artificial roosts for Neotropical bat species,” aims to investigate natural colonization of ABRs (Fig. 1), and to test whether the addition of bat-specific sensory stimuli enhance the roost colonization process. In addition to daily acoustic and visual monitoring of ABRs, we conducted field experiments with sensitive gleaning phyllostomids. We focus on Neotropical fringe-lipped bats (Trachops cirrhosus) as a model species to investigate which stimuli effectively attract individuals to ABRs. Knowledge gained from these investigations will offer valuable insight for conservation strategies that can be widely applied.

We hypothesized that 1) certain bat species would colonize ABRs more quickly than others and 2) these colonization processes might reflect succession patterns, with certain species depending on others for roost discovery: highly exploratory species colonizing ABRs first, less exploratory species following only once others had become established within a roost. 3) We further hypothesized that species which are less likely to use ABRs are at more risk from habitat loss, because they are more likely to rely on specific characteristics for roost finding, and are less flexible in adapting to changing environmental features. We predicted that by introducing sensory lures we could help close this gap, making new, artificial roosts attractive to sensitive species as well. We specifically predicted that multi-modal sensory lures (acoustic + olfaction cues combined) would be more effective in attracting bats than uni-modal lures (acoustic or olfactory cues alone), which in turn should be more effective than controls with no added sensory cues. Finally, we predicted that ABRs would be more effective when situated in disturbed forests where natural roosts are rare.


The project is located in the secondary tropical forests of Soberanía National Park near Gamboa, Panama (28 ABRs; Fig. 2), and in Osa, southwestern Costa Rica (40 ABRs). We are currently conducting experiments in the Panama study site; we focus on these results for this report. Moderate to more extensive land use characterizes the chosen study areas. For a comparative approach between disturbed and intact forests, the same experiments will be conducted on Barro Colorado Island, which is the site of more pristine, undisturbed forests. Afterwards, we will repeat the experiments in Costa Rica to investigate if the results transfer to other regions.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 2. The red markings indicate locations with artificial bat roosts (ABRs) in Panama. ABRs are located in pristine rainforests on Barro Colorado Island and in disturbed secondary forests around Gamboa.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 3 Trachops cirrhosus. Photo credit: Alexander Baugh.

Over 70 species of bats can be found in the forests around Gamboa, including the fringe-lipped bat, Trachops cirrhosus (Fig. 3), which is known for eavesdropping on the mating calls of its frog and insect prey. It is particularly known for feeding on the túngara frog, Engystomops pustulosus.

Selection of sensory stimuli

To select the sensory stimuli most effective at attracting our focal bat, T. cirrhosus, we tested different acoustic and olfactory cues. From earlier studies we knew that individuals of T. cirrhosus can be attracted with acoustic lures broadcasting the mating songs of túngara frogs. Such playbacks have been used for behavioral experiments in the context of foraging, but did not make sense in the context of roost finding. For roost attraction we instead used the calls of conspecifics. Using mist nets, we captured T. cirrhosus from their natural roosts and foraging sites around Gamboa, and brought them to our lab. After collecting standard measurements (e.g. forearm, weight, reproductive status, etc.) and marking them with PIT-tags for individual recognition, we fed them fish and offered water ad libitum. At dusk we put the bats in a flight cage in which we had set up an artificial roost. As soon as the bats had started to occupy the artificial roost, we recorded their calls (Avisoft Bioacoustics, Germany: USGH with condenser microphone CM16; Wildlife Acoustics, US: SM4BAT FS with SMM-U1 external ultrasonic microphones). We then released the bats into their natural habitat again.

To test which types of calls attracted bats to the roost, we conducted a series of behavioral choice experiments with new individuals (Fig. 4). We set up two artificial curtain roosts inside the flight cage and placed one ultrasound loudspeaker (Avisoft Bioacoustics, Germany: UltraSoundGate Player 416 H with Ultrasonic Dynamic Speaker Vifa; Apodemus, Netherlands: BatLure) behind each curtain. Each loudspeaker played back a different type of bat call. The test stimuli included social calls, echolocation calls, and a mixture of echolocation and social calls.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 4. To determine the most effective sensory stimuli to use in the artificial roosts in the forest, we first ran behavioral choice tests with Trachops cirrhosus in the lab.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 5. In a Y-maze we tested bat attraction to different olfactory stimuli.

Similarly, we tested bat responses to different olfactory stimuli. We first collected feces from T. cirrhosus individuals in the lab and from a natural roost exclusively used by T. cirrhosus. These pure T. cirrhosus feces were tested against a mix of feces from species that commonly roost together (e.g., T. cirrhosus, Micronycteris spp. and Saccopteryx spp.), freshly collected from natural roosts. We placed each of the two olfactory stimuli in one of the two arms of a Y-maze and placed a T. cirrhosus individual in the starting arm (Fig. 5). Beginning at approximately midnight each bat had 7 hours to decide if it stays in the starting arm, in the arm that contained pure Trachops feces, or in the arm with the mixed species feces.

In all experiments the different stimuli were randomly allocated to one of the two sides. All experiments were filmed with camcorders and infrared lights to see for which stimulus the bats finally decided as indicated by their final choice in the early morning hours.  

Field experiments

ABRs are located in sets of four in the same area (plots of 100 m²) to minimize the effects of environmental variation on colonization rates. Within a set we randomly assigned each ABR to one of four treatments to determine what sensory stimuli best attract bats. ABRs were fitted with (1) an acoustic lure (Apodemus, Netherlands: BatLure) broadcasting the bat calls found to be most attractive in our lab experiments, (2) an olfactory lure, using the fecal cues found to be most attractive in our lab experiments, (3) a multi-modal lure (both acoustic + olfactory lures), and (4) no sensory lures, as a control. ABRs are monitored daily by visual inspection and acoustically using ultrasonic recorders (Wildlife Acoustics, US: SM4BAT FS with SMM-U1 external ultrasonic microphones) to assess the colonization process. We opened the first set of ABRs November 2018. We plan to continue opening the remaining ABRs in groups of four as described above, and monitor each set for a minimum of 8 weeks.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 6. Artificial Bat Roost with equipment to acoustically monitor and attract bats.

Preliminary Results and Discussion

We recorded echolocation (Fig. 7) and social calls (Fig. 8) from T. cirrhosus individuals. By presenting these calls to new individuals (n = 10), we were able to select attracting (and not repelling) social calls. When testing these attracting social calls against echolocation calls, we found that the bats are generally more attracted by the social calls than by echolocation calls as indicated by the number of approaches to one of the two artificial roosts inside the flight cage. However, after initial approach bats were reluctant to stay inside artificial roosts. We then tested the same social calls against a combination of echolocation and social calls. Still, the bats approached more often to roosts where the social calls were presented. However, when it came to roosting, the bats preferred to stay in the roost with the combination of echolocation and social calls. While these results are preliminary and we are continuing to run tests with more individuals, this is a strong hint that the bats prefer the most natural situation where individuals inside a roost display a mixture of echolocation and social calls and not only one call type.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 7. Examples of Trachops cirrhosus echolocation calls.

Effectiveness of artificial roosts for Neotropical bat species

Fig. 8. Examples of Trachops cirrhosus social calls.

Regarding the olfaction stimuli, the results from the preliminary experiments are less clear as the tested individuals (n = 6) randomly selected pure T. cirrhosus feces or the feces mix. Because a mixture of feces from different species most closely resembles natural conditions in many of the roosts, we decided to use the feces mix as olfactory stimulus in the field-experiments.

We currently only have one set of opened ABRs, so it is too early to draw conclusions from the field experiments. We do see some interesting preliminary findings, however. As predicted, acoustic monitoring with SM4s shows the highest bat activity at the ABR where acoustic and olfactory stimuli are presented together, followed by the ABR with acoustic stimuli alone, which is in turn followed by the ABR with olfactory stimuli alone. Lowest bat activity was measured at the control ABR which had no sensory lures.

We were excited to see that not only did bats inspect the ABRs as evidenced by the acoustic recordings, they also very quickly moved in. We found our first bat residents 5 days after opening the roosts. During our daily visual checks of the roosts, we have found a total 12 Micronycteris (gleaning insectivorous bat), 2 Glossophaga (nectarivorous bat), and 1 Carollia (frugivorous bat) roosting in the ABRs since roost opening. These species are known to be exploratory and rapid roost colonizers, so it confirms our predictions that they would find our roosts first. We hope that with more time, and with these pioneer species residing in the ABRs and thus also acting as attractant lures, the more sensitive bat species such as T. cirrhosus will also find these new roosting resources. Opening the other ABRs and monitoring colonization of these roosts by different species will enable us to better implement ABRs in tropical forests, thereby protecting those vulnerable bat species which are affected most by roost destruction. By doing so, we will be able to make concrete recommendations for applied conservation projects, helping bats recover from anthropogenic impacts.


For all their help in the field and/or ideas and suggestions we are deeply grateful to Detlev Kelm, Nikolai Meyer, Ram Mohan, Michelle Nowak and Adriana Tapia. We cordially thank Wildlife Acoustics for providing us with four SM4BAT FS and associated acoustic equipment, especially Alexandra Donargo for her help throughout this project.

Additional Info

  • Products: Song Meter SM4BAT
  • Research Type: Detect and Identify Bats
  • Species: Bats
  • Field Story Type: Grant Recipient Field Stories