UV reflectance in poison frog
Data files
Aug 24, 2020 version files 191.59 MB
-
Lita_1_IMG_9136.CR2
-
Lita_1_UV_IMG_9138.CR2
-
Lita_2_IMG_9177.CR2
-
Lita_2_UV_IMG_9178.CR2
-
Lita_3_IMG_9179.CR2
-
Lita_3_UV_IMG_9181.CR2
-
Lita_4_IMG_9183.CR2
-
Lita_4_UV_IMG_9184.CR2
Abstract
We have identified strong UV reflectance in the white spots of only a single population of the polymorphic/polytypic poison frog species Oophaga sylvatica. As congeners (O. pumilio) are unable to view UV signals, it stands to reason that natural selection, or neutral processes are likely to be responsible for these signals. We employ visual modeling techniques to estimate the perception of these UV signals, and in a series of comparisons we show surprisingly minimal gains in frog contrast due to their addition.
Methods
Digital images were taken using a tripod mounted full spectrum quartz converted Canon EOS 7D combined with a metal body Nikkor EL 80mm lens (known for high UV transmission), and a series of pass filters. To capture the full range of wavelengths visible to UV-sensitive species we took two photographs of each frog: one in human-visible wavelengths (VIS) and one in the ultraviolet (UV). For human-visible spectra the lens was fitted with a Baader UV-IR blocking filter that allowed transmission from 420-680 nm. For the UV photographs we fitted a Baader UV pass filter that allowed transmission from 320-380 nm. To ensure our camera matched known sensitivities included in the Multispectral Image Calibration and Analysis (MICA) Toolbox, the quartz conversion was undertaken at the same facility as the toolbox authors (Troscianko and Stevens, 2015) (Advanced Camera Services Limited, Norfolk, UK). Images were taken in RAW format and each photograph included 10% and 77% reflectance standards. Downwelling illumination was provided by unfiltered full sun local lighting conditions within frog microhabitats at peak frog activity. Photograph histograms were manually checked for each photograph (UV and visible) to ensure proper exposure and prevent data loss due to overexposure.