Data from: Poison frogs rely on vision for homing in natural environments
Data files
Oct 08, 2025 version files 24.76 KB
-
DataAnalysis_VisualNavigation.Rmd
12.40 KB
-
Exp1_final.csv
1.95 KB
-
Exp2_final.csv
5.58 KB
-
README.md
4.83 KB
Abstract
Neotropical poison frogs (Dendrobatoidea) can navigate back to their territories after large passive displacements. However, the underlying sensory mechanisms remain largely unknown. Dendrobatoids use simple visual stimuli for navigation in confined laboratory experiments, but larger movements under natural conditions are likely to be more challenging. Tropical forests have dense vegetation that creates cluttered and confusing visual environments. Here, we investigated visual cues and navigation of territory-holding male strawberry poison frogs (Oophaga pumilio) in tropical forests. First, we displaced frogs and blocked different aspects of their visual surroundings at the release site. To orient successfully, frogs required information from the lower half of the visual field, but not from the sky or canopy. Second, we displaced males in forested or open areas and characterized their homing trajectories according to visual information using 360° photos. Homing success was higher and faster in open areas, where frogs also showed better initial orientations and lower meandering. We show that low-level visual panoramas in open areas contain more directional information than in closed forests, which could explain these navigational differences. Our results shed light on the mechanisms by which frogs navigate natural environments.
Dataset DOI: 10.5061/dryad.msbcc2gb7
Description of the data and file structure
This study investigates the use of visual cues for navigation of territory-holding male strawberry poison frogs (Oophaga pumilio) in tropical forests and quantifies the orientation behavior and homing trajectories according to visual information.
Experiment 1: To determine whether male O. pumilio use certain components of the visual, we displaced males 8 m away from their territory and experimentally manipulated visual information. Each frog experienced one of four possible experimental conditions that differed in the visual information available at the release site (covered sky and canopy, covered panorama, all-covered and control .
Experiment 2: We investigated whether homing performance differs between sites with different visual complexity (Forest or Open areas). We displaced males 8 m away from their territory and recorded the homing trajectory of the frog, an quantified homing performance. We quantified the visual information available to O. pumilio at the release sites using 360° photos.
Files and variables
File: DataAnalysis_VisualNavigation.Rmd: Rmarkdown with all the necessary code to analyze the data in the csv files attached
File: Exp1_final.csv: Data for experiment 1 (E1)
File: Exp2_final.csv: Data for experiment 2 (E1)
Variables
Some variable names and descriptions are shared between both data files (Exp1_final.csv and Exp2_final.csv). Each variable indicates in parentheses in which experiment it was used (E1, E2, or both).
- ind (E1, E2): ID of individual used in the experiment. Different frogs were used for each experiment.
- site (E1, E2): Site where the experiment was performed (Open Areas/Forest)
- latency (seconds) (E1, E2): Latency to departure: the time between the beginning of the trial and the moment when the frog exited the goniometer
- initial_orientation (degrees) (E1, E2): The angle section on the goniometer within which the frog exited the wooden disk
- ang_error (degrees) (E1, E2): Angular error of the initial orientation: the angular difference between the frog’s initial orientation and the true territory direction. A value ranging between 0 and 180 degrees. For this calculation, we first subtracted 180 from the actual angle representing the actual territory direction. Then we took the absolute value of the previous result, subtracted 180 again and finally took the absolute value as the angular error value.
- condition (E1): The experimental conditions in experiment 1 that differed in the visual information available at the release site (Sky-Canopy covered/covered panorama/all covered/control)
- date (E2): Date of the trial (M/D/Y)
- desp (E2): Direction in which the frog was displaced from its territory
- succes_task (E2): Indicates whether the frog successfully returned to its territory after displacement. We defined a successful homing event as a 75% progression along the total length of the displacement route (i.e., arrival within 1.5 m of the original calling perch). (Y/N)
- group (E2): Combination of success task and site (successful forest/unsuccessful forest/successful open areas)
- rot_info (E2): Rotational image difference function (rIDF): the average difference (absolute pixel difference) between the 360° image and itself across all possible orientations of the image along the azimuth (360 increments of 1 degree). This results in a curve of image difference across azimuths characterized by a minimum when both images’ orientation match
- depth (E2): The depth of the valley around the minimum of the curve of the rIDF
- width (E2): The width of the valley around the minimum of the curve of the rIDF
- sky_ratio (E2): The ratio of sky/terrestrial pixels in the 360° visual field
- path_length (E2): The total distance traveled by the frog from the displacement site to its territory. Only for successful males (meters)
- meander (E2): Local meander index: measures the average directional angle between consecutive segments
- straightness (E2): Overall straightness index: corresponds to the length of the mean resultant vector of the distribution of segment direction
- total_time (seconds) (E2): The time taken for successful homing
Code/software
The data were analyzed using R. The Rmarkdown file "DataAnalysis_VisualNavigation.Rmd" contains all the necessary code to replicate the results of the study, including loading the tabular data sets (Exp1_final.csv and Exp2_final.csv), performing data manipulations, creating subsets for analysis and running all the models.