Evidence for ecological tuning of anuran biofluorescent signals

Abstract

Although biologists have described biofluorescence in a diversity of taxa, there have been few systematic efforts to document the extent of biofluorescence within a taxonomic group or investigate its general significance. Through a field survey across South America, we discover and document patterns of biofluorescence in tropical amphibians. We more than triple the number of anuran species that have been tested for this trait. We find evidence for ecological tuning (i.e., the specific adaptation of a signal to the environment in which it is received) of the biofluorescent signals. For 56.58% of species tested, the fluorescence excitation peak matches the wavelengths most abundant at twilight, the light environment in which most frogs are active. Additionally, biofluorescence emission spans both wavelengths of low availability in twilight and the peak sensitivity of green-sensitive rods in the anuran eye, likely increasing contrast of this signal for a conspecific receiver. We propose an expanded key for testing the ecological significance of biofluorescence in future studies, providing potential explanations for the other half of fluorescent signals not originally meeting formerly proposed criteria. With evidence of tuning to the ecology and sensory systems of frogs, our results suggest frog biofluorescence is likely functioning in anuran communication.

Fig. 1. Biofluorescent emission by taxonomic family.

A summary of the percent biofluorescent emission by family and excitation source. The set of box plots for each family presents the percent biofluorescent emission under the corresponding excitation light source: UV – Ultraviolet (360–380 nm), VI – Violet (400–415 nm), RB – Royal blue (440–460 nm), CY – Cyan (490–515 nm), and GR – Green (510–540 nm). The centre of each boxplot represents the median of the data. The bounds of the box represent the “interquartile range” (IQR), and the whiskers represent the minimum and maximum values of the data set. Axis labels on the bottom panel hold for each set of box plots above. Each point on the plots represents one individual (the maximum percent biofluorescent emission recorded for that individual under that excitation light source). Each individual was measured under each light source. Created in BioRender. Whitcher, C. (2023) BioRender.com/v49t214. Supplementary Data Table 4 contains the respective numeric values for these measurements.

Fig. 2. Evidence for ecological significance of biofluorescence in Anurans.

The four criteria for demonstrating ecological significance of biofluorescence proposed by Marshall and Johnsen (2017), presented within the framework of a twilight environment and dim light photoreceptor visual sensitivities. The top panel presents the patterns expected when biofluorescence signals are tuned to the environment (criteria 1 and 2). Criterion 1: In a twilight environment when most frogs are active, the dominant wavelengths are ~450–460 nm (27; solid black line). The criterion predicts peak excitation (blue dotted line) of anuran fluorophores should match this wavelength range. Criterion 2: The least dominant wavelengths at twilight are ~580–610 nm (27; solid black line). The arrow represents the Stokes Shift of the biofluorescence from peak absorption wavelengths to peak re-emission wavelengths. The criterion predicts the peak biofluorescence re-emission will be centered around ~590 nm to provide the greatest contrasting background at twilight. The center panel presents the patterns expected if the biofluorescence is observable by a receiver (criteria 3 and 4). Criterion 3: Frogs have significantly more green-sensitive (peak absorption ~500 nm) than blue-sensitive rods (28; 39). Hence, this criterion predicts peak biofluorescence re-emission will be centered around ~500 nm to match the greatest spectral sensitivity of another anuran receiver. Criterion 4: The body locations displayed during frog intraspecific communication (29–34; % of species displaying location) should match the body locations that are biofluorescent (this study; % of species for which this location produced maximum biofluorescent recording when excited by blue light, 440–460 nm). The bottom panel presents the observed data for signal tuning and ecological significance from this study. When all fluorescent spectra recorded under blue excitation light (440–460 nm) are plotted (from all body locations), they follow the general shape presented by the dashed green line. This observed fluorescent emission pattern maximizes both sensitivity of the green-sensitive rod in the anuran eye and contrast with the background environment at twilight. The results from our study show that blue-light-induced green anuran biofluorescence meets all four criteria for ecological significance. Created in BioRender. Whitcher, C. (2023) BioRender.com/t37d584.

Fig. 3. Blue light produces the most intense biofluorescence emission in Anurans.

Emission intensity (percent of reflected light realized as biofluorescent emission) is shown for five excitation light sources: UV – Ultraviolet (360–380 nm), VI – Violet (400–415 nm), RB – Royal blue (440–460 nm), CY – Cyan (490–515 nm), and GR – Green (510–540 nm). Each point represents one individual (the maximum percent biofluorescent emission recorded for that individual under that excitation light source). The centre of each boxplot represents the median of the data. The bounds of the box represent the “interquartile range” (IQR), and the whiskers represent the minimum and maximum values of the data set. We utilized a non-parametric Kruskal-Wallis one-way analysis of variance test and Pairwise Dunn’s tests with Holm adjustment to determine if the wavelength of biofluorescent emission differed by excitation light source There is a significant difference in biofluorescent emission intensity by excitation source ( $\chi$2 = 446.88, p = 2.05e-95, n = 2380) with blue light excitation (RB, 440–460 nm) producing a significantly greater percent biofluorescent emission than any of the other excitation light sources. The blue light source also has the closest excitation wavelength to the dominant wavelengths of the twilight environment (see Supplementary Figure S3).

Fig. 4. Measures of biofluorescence emission in amphibians satisfies two of Marshall and Johnsen’s criteria for ecological relevance of biofluorescence.

A Criterion 2 (fluorescence will be viewed against a contrasting background): wavelength of peak emission in anurans (green circles) tends to be different than the most abundant wavelengths in background twilight (p < 0.0001; black line digitized with permission from Cronin et al., 2014). B Criterion 3 (organisms viewing the fluorescence will have spectral sensitivity in the fluorescent emission range): peak emission wavelengths (green circles) match peak sensitivity of anuran green-sensitive (GS) rod of the anuran visual system better than expected by chance (p < 0.0001; black line obtained from). In each panel, the observed wavelength of emission for each individual frog is presented as a colored circle (n = 194). The mean irradiance and sensitivity values of the environment and GS rod at each emission wavelength was not directly measured in this study but obtained from irradiance/sensitivity spectra. Randomization tests were used to generate null distributions and test for significance (see text for details).

Fig. 5. Biofluorescence increases visibility among anurans at twilight.

The black line represents the tradeoff between visual sensitivity and background contrast (green-sensitive rod sensitivity curve divided by twilight irradiance curve; i.e., the spectrum that maximizes both visual sensitivity and contrast with the background environment simultaneously). The observed average tradeoff value (the observed test statistic) is presented as a horizontal line for the observed green emission peaks produced by blue (440–460 nm) excitation light (n = 194 colored points on graph). The null distribution of 10,000 samples is presented on the panel to the right (blue distribution). The p-value in the bottom right-hand corner presents results of the comparison of the values in the randomization distribution to the observed test statistic. A randomization test was used to generate the null distribution and test for significance (see text for details). The green fluorescence emission peak wavelengths produced by blue excitation light match the tradeoff between the visual sensitivity of the green-sensitive anuran rod and the contrast with the twilight environment better than expected by chance (p < 0.0001).

Fig. 6. Maximum biofluorescence by body location.

Pie charts (left) present the body locations from which the maximum biofluorescent emission recording was taken from each individual. Body regions are summarized into the following nine groups: cloaca, dorsal surface, eye, facial pattern, flank, inguinal region, limb, throat, and ventral surface. Photographs (right) illustrate variation in the patterns of biofluorescence produced by blue (440-460 nm) excitation light. The species photographed, in order from left to right are: (top) (A) Boana atlantica,(B) Hamptophryne boliviana, (C) Scinax strigilatus, (middle) (D) Boana geographica, (E) Boana lanciformis, (F) Proceratophrys renalis, (bottom) (G) Chiasmocleis bassleri, (H) Scinax trapicheiroi, and (I) Boana calcarata. Each species panel includes a photograph taken under blue (440–460 nm) excitation light through a 500 nm longpass filter and a photograph of the same individual taken under a full spectrum light source (inset). Dorsal biofluorescence was exhibited via secretions from the frog’s skin (as in Boana atlantica) or located in specific positions on the skin (as in Hamptophryne boliviana and Scinax strigilatus). Ventral biofluorescence was documented as widespread, condensed to specific patterns, or scattered in a speckled pattern (as seen in each individual of the middle row respectively). Additionally, ventral biofluorescence often showed both green and orange emission (~527 nm and ~608 nm; as seen in Boana geographica and Proceratophrys renalis). Finally, distinct regions of the frog body, such as the arms, throat, or eyes sometimes produced the greatest biofluorescent emission recording from an individual (as seen in each individual of the bottom row respectively). Created in BioRender. Whitcher, C. (2023) BioRender.com/b34v891.

Fig. 7. Proposed expansion of Marshall and Johnsen’s criteria for ecological significance of biofluorescence.

We propose the presented dichotomous key to incorporate the complexity of the previously defined criteria. Marshall and Johnsen’s previous checklist for ecological significance of fluorescence is represented by the left most path of the key. Note the addition of different receiver contexts (conspecific vs. heterospecific; left vs. right panels) and the opportunity for not meeting certain criteria to be signals of cryptic coloration (via “NO” paths). Our newly proposed key now incorporates both cryptic and conspicuous signals as ecologically significant explanations for the evolution of biofluorescence. We also propose an expansion of the term “background environment” in Criterion 2 to include both the habitat environment and the individual’s skin environment as, depending on context and size/extent of biofluorescent patch, either could be the relevant background against which the signal is being viewed. Created in BioRender. Whitcher, C. (2024) BioRender.com/x53d269.