Diversity of Color in Pleasing Fungus Beetles (Coleoptera: Erotylidae: Erotylinae)

Abstract

Erotylidae, or pleasing fungus beetles, are a morphologically diverse lineage of Coleoptera notable for the variety of colors and patterns present on their dorsal surface. This study begins the characterization of this diversity and discusses patterns around coloration within Erotylinae. Using spectrophotometer data, we investigated the frequency of certain color motifs across erotyline tribes and discuss geographic patterns in these color motifs. The most frequently observed colors within Erotylinae are brown/black with orange/red maculations in the case of bicolored taxa. In terms of type of maculations, stripes were the most common pattern observed. When summarizing the diversity across major geographic areas, the Neotropical and Indomalay regions displayed the most color variation, followed by the Australasian region.

1. Introduction

Coloration is a central feature of visual communication for both intra- and inter-species signaling [1,2,3,4,5,6,7]. Color can shift, sometimes rapidly, within a lineage [2,8,9]. The evolution of color may be driven by ecological factors (e.g., habitat preference, climate change, and feeding behavior) [10,11,12,13], but may also be influenced by sexual selection [14,15] and/or predation [16,17,18]. Two of the most common evolutionary strategies involving color for interspecies communication are camouflage and aposematism. For example, there are several types of camouflage that are used to avoid predators, such as color change and blending into background environments (e.g., geckos [19], flounder [20], octopus [21]), disruptive coloration (e.g., cheetahs [22], praying mantis [22], reef fish [23]), countershading (e.g., crustations [24], sharks [22]), and mimicry (e.g., Viceroy butterflies [25]), to name a few. Unlike camouflage, where the goal is to blend in, aposematic coloration often consists of bold and bright contrasting colors in various patterns to warn predators that they are toxic or dangerous. Examples of aposematic coloration can be seen in many groups, and some common examples include poison dart frogs [26], black widow spiders [27], and coral snakes [28].

Insects are highly diverse in their coloration and demonstrate numerous examples of both camouflage and aposematism [5]. Walking sticks [29,30], grasshoppers [31,32], and butterflies [33] contain examples of camouflage. Specifically, katydids (Orthoptera) possess leaf-like wings, in both color and shape [32]. Alternately, other insects such as bees and wasps [34,35], as well as beetles [36,37,38,39,40], provide classic examples of aposematism. For instance, fireflies (Coleoptera) use both bioluminescent signals and coloration to warn predators of their distastefulness [38,39].

Coleoptera (beetles) form one of the most diverse animal lineages and unsurprisingly they express an astounding diversity of coloration. Within beetles, there are several different mechanisms of color including iridescence and pigmentation. Iridescence combines multilayer reflectors, diffraction gratings, and three-dimensional photonic crystals to bend and refract light, producing many bright and often metallic colors, but also making quantification very difficult [41,42,43,44]. Pigmentation is important in the formation of patterns that convey a variety of important ecological functions including mimicry, temperature response, mating, etc. [45,46,47,48,49]. Complex pigmentation is widespread across many beetle groups, especially the aptly named pleasing fungus beetles, or Erotylidae [50,51,52].

Erotylidae are the largest family within the Erotyloidea with approximately 3,670 species broken into six subfamilies, five tribes, and ~283 genera [53]. Several phylogenetic hypotheses have been proposed for erotylids, both morphological [54,55] and molecular [56], which consistently found a monophyletic family but numerous classification issues within [57,58,59]. Erotylids can be found globally, with the highest diversity in tropical regions in South America and Southeast Asia [58]. Most erotylids feed on fungi [56,58,60,61,62], but some groups, like the Langurini, are stem borers in green plants [63,64], or feed in cycad cones like the Pharaxonothinae [65,66,67], or are general detritivores [68]. The family displays an incredible amount of morphological diversity, with members ranging in body size from 2 to 30 mm, exhibiting varying body shapes from dorsoventrally flattened to extremely convex, as well as elongated to nearly perfectly round [54,56,62]. Coloration similarly varies from dull brown/black to bright colors and uniform coloration to complexly patterned and multicolored [46,52,56,69,70,71,72,73,74,75]. The complex patterns are often displayed as contrasting dark and bright colors through stripes or spots [56,70] (Figure 1). Erotylinae, the largest of the six currently described subfamilies, are broken into five tribes (Dacnini, Encaustini, Erotylini, Megalodacnini, and Tritomini); however, the relationships both within (e.g., Tritomini) and between tribes have yet to be solidified [54,59].

Color patterns have been examined for some lineages of Erotylidae [54,56,59,70,76,77] but have yet to be explored for most genera. Several hypotheses have been presented for the color variation observed in Erotylidae, most notably its function as an aposematic signal [52,56,78,79,80]. The often brightly colored and highly patterned body type is typical of aposematic signaling, and there is some evidence to suggest that erotylids are unpalatable and in some cases toxic [56,81]. Past studies [52,56] have attempted to correlate aposematism with coloration in erotylids while coding color and pattern using the human visible spectrum and human interpretation, which can be prone to errors, but provides a foundation for future studies such as this. Here, we survey and discuss the common color motifs across Erotylinae using spectral data. Specifically, we assessed whether more similar underlying wavelengths are seen in more closely related taxa (i.e., within tribes vs. between tribes). We attempt to describe the components of the observed color within each tribe and between major geographic areas.

2. Methods

2.1. Taxon Sampling

Representatives from all major clades of Erotylinae were selected to have dorsal body coloration quantified with a spectrophotometer. Recently collected specimens from Peru, Vietnam, and the USA, as well as museum specimens from the Florida State Collection of Arthropods (FSCA), were measured. In total, 123 taxa were included, representing 81 genera, currently classified into all five described tribes, with classification following [54]. Specimens were determined by the author P.E.S. with available pertinent literature and are all vouchered in the BYUC, FSCA, or NCSU with unique identifiers (Table S1).

2.2. Character Coding: Coloration and Pattern

Taxa were selected based on the availability of fresh material but were chosen to maximize higher-level clades sampled and to capture the color diversity observed in the group; metallic and/or iridescent species were removed due to limitations of recording accurate color under these conditions. The dorsal surface of specimens was categorized into bicolored or unicolored to determine how many spectral data points were necessary (Table S2). Bicolored was defined as a specimen with multiple observable colors (i.e., both background and marking). Unicolored was defined as a specimen with one observable color or a pattern that was too small (<2 mm) to be detected by the spectrophotometer probe [82].

For bicolored specimens, we attempted to summarize and group certain observations (i.e., background and marking) based on the following guidelines: (1) circular markings with a perimeter entirely or mostly visible is considered a spot, (2) longitudinal markings spanning the width or length of either the pronotum or elytra are considered stripes, (3) visual inspection of the surface area occupied by either the spots and stripes or the remaining dorsal surface was used to identify what we consider the background. While most of the markings are more brightly colored and the background is often dark brown or black, there are some representatives where, based on our coding, the markings are black and the background is the more brightly colored part of the beetle (e.g., Cyrtomorphus sp.). If the specimen had alternating stripes of relatively even size, then the background was considered black, and the brighter color was coded as the marking.

2.3. Measuring Spectral Data

An Ocean Optics spectrophotometer (model DH-2000-BAL) [Ocean Insight, Orlando, FL, USA] with an Ocean Optics Lightsource (model Flame-S-UV-VIS-ES) [Ocean Optics, Inc., Dunedin, FL, USA] and an Ocean Optics reflection probe (model QR400-7-SR) [Ocean Optics, Inc., Dunedin, FL, USA] was used to measure each specimen. The setup of the Oceanview 1.67 software settings on the computer was as follows: the Boxcar set to 10, the Integration time to 40, and scans to average to 25, as well as checking the box for correct linearity. The reflectance standard (LabSphere USRS-99-010 AS-01158-060 Spectraflex Uncalibrated Reflectance Standard) [Labsphere Inc., North Sutton, NH, USA] was used to calibrate the machine for accurate color measurements. These standards were remeasured after every 12 beetles to maintain proper calibration. Each specimen was placed approximately 2 mm from the probe at a 90 degree angle. Each visible color was measured in three locations, and each measurement was repeated three times at each location. Depending on the beetle, and whether it was unicolored or bicolored, 9 or 18 spectral data points were taken, respectively (Figure 2).

2.4. Analysis of Spectral Data

2.4.1. Preview Outliers

To preview spectral measurements for outliers, a visual inspection was conducted (i.e., all 9–18 spectral measurements) for each specimen in Rstudio v.4.2.1. The getspec function from the pavo package [83] in Rstudio, specifying the file format (“txt”) and x-axis limits (325 to 650), was used to read in the measurement, and then the explorespec function was used. This function generated graphs that displayed all three spectral measurements at once. The resulting graphs were manually inspected for any obvious outliers due to misnamed files or mismeasured readings due to human error. If no outliers were detected, the spectral readings were cleaned using the procspec function to smooth the curve using LOESS smoothing and convert any negative values to zeros (fixneg parameter). The smoothed spectral data were averaged together by the aggspec function to create one averaged line which was then plotted with ggplot2.

2.4.2. Peak Wavelength of Color

To find the wavelength at the highest reflectance percentage for the background and marking of each specimen, we wrote a function called “find_max”. The function first separated the spectral files by background or marking. Next, the files were smoothed and aggregated, again using the procspec and aggspec functions. The maximum value of the peaks was calculated by using the peakshape function, which outputs the corresponding x-values (wavelength nm).

2.4.3. Visualizing Average Spectral Data

The summarise function in the dplyr package [84] was used to calculate the spectral average (mean function) and standard deviation (sd function) for each tribe separated by background and marking and was plotted with ggplot2 (geom_ribbon). The summarise function was also used to calculate the spectral average (mean function) and standard deviation (sd function) for each genus in a tribe separated by background and marking and was plotted with ggplot2 (geom_ribbon). To visualize the spectral data in 3D colorspace, the aggregated spectral data were converted using the vismodel function (avg.uv parameter). Next, this visual model was converted to colorspace using the colspace function (tcs parameter). Then, the spec2rgb function with an alpha of one was used, and finally the col2rgb function was used. Lastly, scatterplot3d was used to plot the RGB values by tribe.

2.4.4. Visualizing Spectral Data by Biogeographical Region

The spectral data were first separated by biogeographical region, tribe, and background or marking. Next, they were smoothed and aggregated, as mentioned above. These data were then plotted using ggplot2 and colored by tribe.

2.4.5. FANOVA

A functional ANOVA was conducted between the tribes using the fanova.test function with default parameters in the fdANOVA package in Rstudio [85]; however, we note that the sampling scheme underrepresents certain tribes, which limits the interpretation of this test.

3. Results

The results of the dorsal pattern observations from our sampling showed that Encaustini (100%), Erotylini (60%), and Megalodacini (74%) have a much higher percentage of bicolored specimens compared to Dancini (7%) and Tritomini (17%). Of the three tribes with higher bicolored specimens, Encaustini (75%), Erotylini (80%), and Megalodacini (92%) have a higher percentage of specimens that display stripes compared to spots (Figure 1, Table S2). Of the bicolored specimens, ~30% had a stark difference in color, represented by a difference in maximum wavelength greater than 100 nm. Four specimens we measured (Episcapha semperi, Iphiclus odyneuroides, Pselaphacus rubricatus, and Scaphidomorphus bosci) from four genera and three tribes (Erotylini, Tritomini, and Megalodacnini) had the highest difference at 350 nm. The specimens with the highest difference were split evenly between stripes and spots. Nine specimens (Callischyrus melanogaster, Erotylus cassidoides, Episcapha sp. 1, Episcapha sp. 2, Ischyrus sp. 1, Iphiclus sp. 2, Iphiclus nr. musicalus, Tamboria coerulea, and Triplatoma sp.) from six genera and three tribes (Erotylini, Megalodacnini, Tritomini) had a difference between 200 and 300 nm. Four specimens (Aegithus cyanipennis, Cypherotylus sp. 1, Cypherotylus sp. 2, and Erotylus cassidoides 2,) from three genera and one tribe (Erotylini), had a difference between 100 and 200 nm.

The general background color for all tribes is dark brown/black, as expected, with Encaustini having the most variation in intensity (Figure 3C,D). The general marking colors for all tribes peak within the yellow to red range (575–675 nm) (Figure 3A,B). Tritomini, Erotylini, Megalodancini, and Encaustini are tightly grouped, with Encaustini again having the most variation in intensity; Dancini follows the trend but has a lower intensity. The FANOVA for the background wavelengths indicates that there is not a significant difference between the tribes (test stat = 2.1, p-value = 0.06); the FANOVA for the marking wavelengths was again not significant (test stat = 1.22, p-value= 0.32), indicating that there is no difference between the tribes.

However, within tribes, generic variation is displayed within the wavelengths and intensity (Figure 4). When comparing the background colors, Dacnini have some genera (e.g., Neothallis and Kuschelengis) with more of a blue (400–450 nm) tone compared to the other genera. Additionally, Thallis has the most variation within reflectance intensity ranging from 10 to 70%. Encaustini displays more variation in intensity than color. For example, all three genera (Aulacochilus, Micrencaustes, Encaustes) peak within the orange/red (600–675 nm) range, but Micrencaustes is much brighter with an intensity of ~60% compared to the other two genera who are a more subtle orange/red with an intensity of ~5%. Within Erotylini, most genera have a dark brown/black background, but Neopriotelus and Scaphidomorphus have a blue (400–450 nm) tone, while Strongylosomus, Oligocorynus, Plastococcus, Cyclomorphus, and Dyslexia have more of an orange/red (600–675 nm) tone. Within the background colors of Megalodancini, Tamboria has a blue (400–450 nm) tone and Coptengis has a yellow (550–600 nm) tone, while the remaining genera have a dark brown/black color. Within Tritomini, the genera are split between an orange/red and dark brown/black background. For example, Lybanodes, Laurenticola, Lybas, Mycolybas, Mycomystes, Mycophthorus, Pselaphacus, and Pseudolybas have a stark peak within the orange/red range compared to other genera such as Rhynchotritoma, Tritoma, and Haematochiton, which are also within the orange/red range. When comparing the marking colors, there is generic variation as well. Within Encaustini, all three genera peak in the orange/red range, but Micrencaustes again has a much greater intensity peaking at 100%. Within Erotylini, Scaphidomorphus has the strongest overall intensity and peaks within the yellow/orange range, whereas Aegithus, Barytopus, and Prepopharus peak within the orange/red range. Most genera within Megalodacnini peak within the orange/red range but vary in intensity. For example, Megalodacne, Nestitus, and Tamboria have higher reflectance intensities in the orange/red range compared to Endytus, Episcapha, Hybosoma, and Linodesmus. Additionally, Tamboria has a second peak within the violet/blue (325–375 nm) range. Within Tritomini, Megischyrus, Pselaphacus, and Spondotriplax have a dark brown/black marking, while Callischyrus, Ischyrus, and Mycotretus peak within the orange/red range but have a low reflectance intensity, with Callischyrus having the highest at ~60%.

While Erotylinae has a global distribution, tribal ranges vary. A tribal breakdown by biogeographical regions is as follows: (1) Tritomini is found in all regions except Australasia; (2) Erotylini is restricted to the Nearctic and Neotropics; (3) Dacnini is found in all regions, with only one species found in the Neotropics (Brazil), and there are no records in Central America; (4) Megalodacnini is abundant in the Palearctic, Indomalaya, and Australasia, while there are a few species in the Nearctic and Neotropics; (5) Encaustini is most abundant in the Palearctic, but there are a few species in the Afrotropics and one genus in the Nearctic (Mexico) (PES). We were able to include biogeographical tribal representatives as follows: (1) Tritomini in all regions except Australasia; (2) Erotylini from the Nearctic and Neotropics; (3) Dacnini from the Afrotropics, Palearctic, Indomalaya, and Australasia; (4) Megalodacnini from the Afrotropics, Indomalaya, Neotropics, and Australasia; (5) Encaustini from the Palearctic and Indomalaya. When the dorsal pattern observations are based on biogeographical regions with this sampling, unicolor has greater representation in the Palearctic (86%), Nearctic (100%), Afrotropics (79%), and Australasia (89%) (Figure 5, Table S2). Two continents have more of an even split between unicolor and bicolor specimens at 49% compared to 51% in the Neotropics and 44% compared to 55% in Indomalaya; within the bicolored specimens, spots are the most common pattern, at 93% in Indomalaya and 76% in the Neotropics. The general background color for all regions is dark brown/black and orange/red. The exceptions to this are as follows: Neotropics include peaks within the blue/green (450–525) and yellow (575–600) ranges, Afrotropics include peaks within the blue (425–475) range, Australasia has peaks within the blue (400–450) and green (~550) ranges, Indomalaya has peaks within throughout the blue to orange (400–600) range, and Palearctic includes peaks within the blue (~450) range. The general marking color for all the continents was within the orange/red range, with the exception of the Neotropics, which had peaks throughout the blue to orange (450–600) range, and the Nearctic, which did not have any bicolored specimens represented.

4. Discussion

The most common pattern across Erotylinae is spots, with a stark color contrast between the background being dark brown/black and the spot color being orange/red. The coloration and patterning within Erotylidae are most often due to pigmentation, with some exceptions (e.g., Cypherotylus). The spot pattern due to pigmentation is also very common within Coccinellidae (lady beetles), but with the coloration swapped. Coccinellidae often have a red, orange, or yellow background color with black spots, indicative of aposematism. Other incredibly color diverse lineages include Buprestidae (jewel beetles), Cetoniinae (flower chafers), and Cassidinae (tortoise beetles), but most of their coloration is structural, leading to high rates of iridescence and metallic coloration. For example, the most common colors within Buprestidae are a metallic green, iridescent blue, and gold, while the common colors within Cassidinae are metallic yellow and green, as well as spots that are red or orange with a black background. While Cetoniinae have high rates of iridescence, most commonly metallic green, there is also a higher rate of patterning between black and orange/red and black and yellow/green. While the FANOVA did not currently find significant results between the tribes, increasing the sample size as well as adjusting and solidifying tribal relationships would most likely lead to significant differences at the tribal level, and hence we explored the variation within each tribe and by biogeographical region in addition.

The biogeographical region analysis is reflective of most insect diversity patterns, with the highest richness and diversity occurring in the tropics (e.g., Neotropics, Indomalaya). Australasia contains a high color diversity, despite not having many representatives (9). Within Australasia, Megalodacnini is the most diverse clade we evaluated. The background colors peak in the blue (~425 nm) and yellow/green (500–600 nm) ranges compared to the Dacnini representatives. Dacnini tends to peak toward the red (600–650 nm) range or black, with the exception of a few peaks in the blue range but not as stark as the blue peak in Megalodacnini. Megalodacnini is also the only tribe that possesses a bicolored specimen that was evaluated.

In this analysis, the Nearctic (8), Afrotropics (14), and Palearctic (7) have a moderate amount of richness represented compared to the Neotropics (57) and Indomalaya (29). Within the Nearctic, Tritomini have the most representatives, but none of them are considered bicolored. The species in this tribe are split between black and the orange/red (600–650 nm) range. The one representative of Erotylini peaks within the orange/red range. Within the Afrotropics, Megalodacnini is the only tribe that has bicolored representatives, where the marking peaks in the orange/red range and the background tends to be black. The background of Tritomini is split between black and orange/red. Within the Palearctic, Encaustini is the only tribe that has a bicolored representative that has a peak within the orange/red range at a high reflectance intensity. The background colors are generally dark down/black but can be grouped by reflectance intensity. Dancini and Erotylini have the lowest reflectance at approximately 10–25%, Tritomini ranges from approximately 45 to 75%, and Encaustini is at approximately 75%. Neotropics and Indomalaya represent the highest richness. Indomalaya has the greatest tribal representation, only missing Erotylini. Within Indomalaya, Tritomini has the most diversity of background color with multiple peaks throughout the blue to red range, while Megalodacnini has the most diversity in marking color, peaking in the orange/red range at varying reflectance intensity. Within the Neotropics, Tritomini has the most diversity within the background colors with multiple peaks throughout the blue to red range, while Erotylini has the most variation within the marking colors, again with multiple peaks throughout the blue to red range.

The high diversity of coloration and pattern within Erotylinae most likely evolved for interspecies communication in biodiversity hotspots as an aposematic signal, as hypothesized in previous studies. Aposematism is often bright contrasting colors in striking patterns that act as a warning message to deter predators that they are dangerous or unpalatable, which is fairly common within beetles [86]. We detected numerous examples of stark contrasting colors and patterns within Erotylinae (i.e., Erotylus cassidoides, Barytopus lunulatus, Episcapha semperi, Tamboria coerulea) (Figure S1), indicating that aposematism could be an effective defense strategy within Erotylidae. The highest diversity and richness occur within the Neotropics and Indomalaya, where insect diversity is highest in general. This abundance of diversity could provide an alternative explanation for the observed coloration and patterns, in this case mimicry. Erotylids could be camouflaged by imitating other more defended insects in the area. However, mimicry does not explain the high color diversity within the less diverse regions such as Australasia and Afrotropics. The most common colors and patterns tested in other insects that indicate aposematic behavior are red, yellow, and orange with spots or stripes. Here, we began to explore the true underlying colors within Erotylinae, which showed that there is a high level of color and pattern variation within the group. The most common colors within Erotylinae are dark brown/black and orange/red. Of the bicolored specimens, stripes are the most common pattern (79%), and Megalodacnini is the most color-diverse tribe. However, specific drivers of the observed color variation across Erotylidae remains difficult to test until phylogenetic work can solidify more of the higher-level classification and specific tests are conducted exploring the chemical defense across this clade.