Molecular evidence for color discrimination in the Atlantic sand fiddler crab, Uca pugilator

Animal coloration has long been implicated in intraspecific signaling (Darwin, 1871), particularly in the context of sexual selection and mate choice. Not surprisingly, the ability to visually distinguish colors has been found to be a major sensory cue for these behavioral and ecological functions in many species [e.g. primates (Fernandez and Morris, 2007), spiders (Lim et al., 2008), cichlids (Seehausen et al., 2008), crabs (Baldwin and Johnsen, 2009), stomatopods (Cheroske et al., 2009) and frogs (Sztatecsny et al., 2010)]. Research using fiddler crabs (genus Uca) has provided several insights into the mechanisms of intraspecific visual signaling and communication, but this work has generally focused on form and motion cues or, if it involved the manipulation of color, could not rule out achromatic contrast as a confounding visual cue (for reviews, see Crane, 1975; Salmon and Hyatt, 1983; Christy and Salmon, 1984; Salmon, 1984; Salmon and Zucker, 1988; Christy and Salmon, 1991; Backwell and Passmore, 1996; Koga et al., 1998; Backwell et al., 1999; Backwell et al., 2000; Christy et al., 2001; Murai and Backwell, 2005; Christy and Backwell, 2006; How et al., 2008). To summarize a few salient findings, the males' claw-waving displays attract the attention of females, and because the waves' spatial and dynamic aspects are species specific, it probably also communicates species identity (Aspey, 1971; How et al., 2009). The conspicuousness of this display is aided by the enhanced visual contrast of the carapace and major cheliped against a mud background, provided by their particular spectral reflectance, which can even change to maintain high contrast after a change in background (Zeil and Zanker, 1997; Zeil and Hofmann, 2001; Zeil and Hemmi, 2006). When female U. mjoebergi and U. signata were given a choice between hetero- and conspecific males with painted chelae they preferred whichever one was painted to resemble a conspecific male (Detto et al., 2006).

Only a few behavioral experiments have been specifically designed to test for color vision in fiddler crabs, and these have been fairly successful. Female U. pugilator showed positive phototaxis towards monochromatic blue and red–orange light, but not towards white light (Hyatt, 1974). Recently, when U. mjoebergi females were subjected to a choice between males having a yellow claw and males whose claws had been painted one of several shades of grey, they preferred the yellow claw, suggesting that mere luminosity does not mediate the discrimination (Detto, 2007). Thus, for U. mjoebergi, coloration does not merely increase luminance contrast, but is a real color signal in both visible and ultraviolet wavelengths (Detto and Backwell, 2009).

Despite the evidence for color vision from behavioral studies, attempts to demonstrate color vision physiologically have yielded mixed results. In an early attempt, glass electrode electroretinograms (ERG) on U. pugilator, U. pugnax and U. minax produced broad spectral sensitivity curves with maxima between 450 and 500 nm, predicting at least two pigments with one sensitive in blue–violet and the other in red–orange (Hyatt, 1974). This contradicted results published in the same year from cotton wick ERGs from U. pugilator and U. pugnax, which showed an absorption curve with only a single peak at ∼508 nm (Scott and Mote, 1974). Both studies also included selective adaptation of photopigments with monochromatic light, and both studies reported results that supported their opposing conclusions from the ERG amplitudes, adding to the confusion over whether these species have one or more visual pigments. Much later, action potentials recorded from the optic nerve of U. thayeri showed broad spectral response curves that were best fit by two visual pigments having peak wavelength of spectral sensitivities (λ_max) at 430 nm and 510–530 nm (Horch et al., 2002). Finally, recent microspectrophotometric measurements from U. tangeri, U. vomeris, U. pugnax and U. pugilator showed only one visual pigment present in retinula cells 1–7 (R1–R7 out of a total of eight) with λ_max ranging from 500–540 nm depending on the species (Jordao et al., 2007). The possibility of a different visual pigment in the eighth retinula cell (R8) was not ruled out since the small size of these cells precluded any measurements.

Thus, although behavioral studies tend to indicate color perception, these have only occasionally been corroborated by physiological evidence. This disparity may exist because electrophysiological and spectrophotometric measurements tend to be of poor quality in Uca. The poor quality is most likely due to the enormous amount of screening pigments, which serve several functions, including restricting stray light, meta-rhodopsin conversion and light adaptation (Eguchi and Waterman, 1967; Ribi, 1978; Stavenga, 1989). Such studies therefore often produce results that only weakly support one side or the other.

The minimum requirement for color vision is the expression of at least two active opsins of different spectral absorbance in at least two different members of the eight photoreceptor cells comprising an ommatidium (also called retinula cells R1–R8). The present study aims to identify the physiological substrate for color vision in fiddler crabs, if it exists. We avoid the problems associated with the screening pigments by using molecular biological techniques. We identified three distinct opsin transcripts (UpRh1, UpRh2 and UpRh3) in the retinas of both male and female U. pugilator and found by in situ hybridization that all three opsins are expressed in all ommatidia. An ostensibly UV-sensitive opsin, UpRh3, was found in R8 and, of the remaining seven retinula cells, five were found to have UpRh1 and three to have UpRh2, meaning there is coexpression of two of the opsins in a single retinula cell in each ommatidium.

Uca pugilator (Bosc 1802) sand fiddler crabs were collected from Beaufort, NC, USA and maintained at room temperature in the laboratory (Cincinnati, OH, USA) in a tank containing soil from their native habitat. The enclosure was filled and emptied at six hour intervals with 20 parts per thousand salinity water (Instant Ocean, Spectrum Brands Inc., Alpharetta, GA, USA), thereby mimicking a 12 h circatidal cycle. Crabs were placed on a 12 h:12 h L:D cycle and fed pulverized fish food (Tetramin, Spectrum Brands Inc.).

Total RNA was extracted from the eyes of five U. pugilator male fiddler crabs using TRIzol reagent according to the manufacturer's protocol (Gibco BRL, Gaithersburg, MD, USA). Single-stranded complementary DNA (cDNA) was synthesized from ∼1 μg of total RNA using oligo(dT) primers provided with the Affinityscript QPCR cDNA synthesis kit (Stratagene, La Jolla, CA, USA). Synthesized cDNA was used in PCR reactions with degenerate primers F5 and R8 designed based on the conserved regions found in all arthropod opsins (Hariyama et al., 1993) (Table 1). Owing to the presence of excess amount of polysaccharides in the crustacean eye tissue, Plant RNA Isolation Aid (Applied Biosystems, Foster City, CA, USA) was used in every PCR reaction (1:10 dilution per reaction) to prevent PCR inhibition. PCR amplifications were typically performed in 25 μl reaction volumes consisting of: 2.5 μl PCR 10× buffer, 2.5 μl 25 mmol l^–1 MgCl₂, 4 μl 10 mmol l^–1 dNTPs, 1 μl 100 μmol l^–1 forward primer, 1 μl 100 μmol l^–1 reverse primer, 0.5 μl Taq DNA polymerase (Takara Bio Inc., Madison, WI, USA), 1.5 μl cDNA, 2.5 μl Plant RNA Isolation Aid and 9.5 μl H₂O. PCR conditions were as follows: 5 min of initial denaturation; 30 cycles of 1 min at 94°C, 1 min at 48–55°C (depending on the Tm of the primer pair), and 1 min at 72°C; and a final extension of 5 min at 72°C. The PCR products were separated on a 1% agarose gel, fragments of appropriate length were gel extracted (QIAquick Gel Extraction Kit; QIAGEN, Valencia, CA, USA), and these were cloned into pGEM-T Easy (Promega, Madison, WI, USA). Samples were bi-directionally sequenced with vector primers SP6 and T7. These partial opsin sequences were then used to design additional degenerate and opsin specific primers as listed in Table 1 for 5′ and 3′ RACE reactions.

To obtain full-length opsin sequences, the 3′ and 5′ ends of the opsin genes were amplified using RACE. For 3′ RACE, cDNA was synthesized from ∼1 μg of the total RNA as described above using the adapted oligo(dT) primer, Hybrid dT (Table 1). Adapter R primers (reverse complementary to the 5′ end of the Hybrid dT) were used as reverse primers in conjunction with opsin-specific forward primer. 5′ RACE reactions were performed from ∼10 μg of total RNA according to the manufacturer's protocol with the Firstchoice RLM-RACE kit (Applied Biosystems) using opsin-specific reverse primers (UpRh1 R1, UpRh2 R1 and UpRh3 R1) and kit-provided forward primers.

Full-length U. pugilator opsin cDNA sequences were translated ‘in silico’ and aligned using CLUSTALW [MEGA 4.0 software (Tamura et al., 2007)] with other arthropod opsin protein sequences obtained from GenBank (species and accession numbers listed in Table 2). A neighbor-joining algorithm (Saitou and Nei, 1987) with Poisson correction (Zuckercandl and Pauling, 1965) was used to construct an opsin tree, and bootstrapped by replicating 1000 times. Pairwise deletion algorithm was also used to eliminate any alignment gaps present in the sequence.

Riboprobes were generated from ∼200 bp present at unique 3′ untranslated regions (3′UTR) or 5′ ends of the coding region (Fig. 1). Appropriate fragments were PCR amplified, cloned into pGEM T-Easy (Promega, Madison, WI, USA), and digoxigenin- and fluorescein-labeled antisense and sense (control) riboprobes were synthesized using the SP6/T7 transcription kit (Roche Applied Science, Indianapolis, IN, USA). After riboprobe synthesis, the template DNA was degraded using 2 μl of DNase, provided in the kit, for 15 min at 37°C. Synthesized probes were stored in hybridization buffer [0.3 mmol l^–1 NaCl, 2.5 mmol l^–1 EDTA, 20 mmol l^–1 Tris-HCl (pH 8.0), 50% formamide, 10% dextran sulfate, 100 mg ml^–1 yeast tRNA and 1× Denhardt's medium (Sakamoto et al., 1996)] at –80°C until further use.

Fiddler crabs (U. pugilator) were immobilized on ice and their eyes were dissected directly into RNA Later Solution (Applied Biosystems, Foster City, CA, USA). The eyes were then fixed with 4% paraformaldehyde in 1× phosphate-buffered saline (PBS; pH 7.4) at room temperature (RT) for 2 h. After fixation, the tissue was washed twice in 1× PBS for 15 min and dehydrated through an ethanol series. After two 1 h washes in xylene, the tissue was embedded in paraffin wax (Thermo Scientific, Wilmington, DE, USA). The embedded tissue was frozen at –20°C overnight, and ∼10 μm paraffin sections were cut using a microtome (Leitz Wetzlar, Wetzlar, Hesse, Germany). Sections were then washed once in 1× PBS for 10 min, rehydrated in a graded ethanol series to remove the paraffin wax, and hybridized with 0.5 μg ml^–1 of digoxigenin-labeled probe at 55–60°C for 18 h in hydridization buffer. After hybridization, the slides were washed at 55–60°C with 5× SSC, followed by sequential 30-min washes of 2×, 1× and 0.2× SSC buffer. Sections were then blocked with 10% normal sheep serum (Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) in 1× PBX for 1 h at RT, incubated at 4°C overnight with a 1:250 dilution of sheep anti-digoxigenin–AP (Roche Applied Sciences, Indianapolis, IN, USA) in blocking solution, and then washed three times in 1× PBX at RT for 20 min. Alkaline phosphate (AP) staining was developed using either the 1-STEP NBT/BCIP plus suppressor (Thermo Scientific, Wilmington, DE, USA) or with Fast Red tablets (Roche Applied Sciences, Indianapolis, IN, USA) according to manufacturer's instructions. After staining, the sections were washed twice for 10 min in 1× PBX at RT, rinsed with distilled H₂O and mounted using Fluoromount-G (SouthernBiotech, Birmingham, AL, USA). Images were digitally captured using an Olympus BX51 microscope equipped with an Olympus 60806 camera (Olympus America Inc., Center Valley, PA, USA) and processed with the Adobe Photoshop image processing software (Adobe Systems Inc., San Jose, CA, USA). Fluorescent images were captured with a Texas Red filter channel.

Double in situ hybridizations were performed as described above with minor modifications. Two different 0.5 μg ml^–1 riboprobes labeled with digoxigenin and fluorescein, respectively, were used simultaneously in the hybridization buffer. Sections were blocked in a mixture of 10% normal sheep serum and 10% normal goat serum (Jackson ImmunoResearch) and tissue sections were simultaneously incubated at 4°C for 18 h with both sheep anti-digoxigenin–AP (1:250) and goat anti-fluorescein–POD (1:100; Roche Applied Sciences, Indianapolis, IN, USA) diluted in blocking solution. A fluorescein-labeled probe was detected using a DAB substrate kit (Vector Laboratories, Burlingame, CA, USA) according to manufacturer's instructions. Alkaline phosphatase activity was subsequently detected using Fast Red tablets (Roche Applied Sciences) in accordance with manufacturer instructions after a brief 10 min wash in 1× PBX at RT.

Three different opsin cDNAs of 1358, 1467 and 1459 bp were identified by RT-PCR using degenerate and gene-specific opsin primers (Table 1) and named UpRh1, UpRh2 and UpRh3, respectively (Fig. 1; GenBank accession numbers: UpRh1, HM765425; UpRh2, HM765426; UpRh3, HM765427). Opsin cDNAs differed in the length of their 5′ and 3′ untranslated regions and in nucleotide sequence conservation, with the 3′ end of each transcript in particular showing low (28%) sequence similarity.

Opsin amino acid sequences were deduced from the full-length cDNA sequences for each opsin. UpRh1 and UpRh2 each encode 377 amino acid opsins and UpRh3 encodes a 376 amino acid opsin. UpRh1 and UpRh2 are 76.72% identical and 89.12% similar in their amino acid sequences, and show 50.13% amino acid identity and 70.13% similarity to UpRh3. As is typical for opsins, all three U. pugilator opsins are predicted to be seven-transmembrane-domain proteins. In addition they show strict conservation of (1) the potential G-protein binding regions (QAKKMNV and DRY) (Gartner and Towner, 1995; Townson et al., 1998), (2) the lysine (K) amino acid responsible for covalent binding to the chromophore through a Schiff's base interaction (Gartner and Towner, 1995; Townson et al., 1998; Wang et al., 1980), (3) two conserved cysteine (C) residues (UpRh1, C128 and C205; UpRh2, C128 and C205; UpRh3, C132 and C209) responsible for disulfide linkage stabilization (Gartner and Towner, 1995; Townson et al., 1998) and (4) an aspargine (N) that is a potential glycosylation site (UpRh1 and UpRh2, N3; UpRh3, N9) (Gartner and Towner, 1995; Townson et al., 1998).

Several characteristic amino acids have been identified at specific positions that are functionally related to the spectral sensitivity of the opsins. For instance, a functional, conserved residue at position 131 in the third transmembrane domain of invertebrate opsins is thought to be one of the two negative counter ions stabilizing the protonated Schiff's base (Chang et al., 1995), and in all long-wavelength-sensitive opsins, this residue is a tyrosine (Y), whereas in all short-wavelength opsins, this residue is phenylalanine (F). UpRh1 and UpRh2 contain a tyrosine residue (Y131 and Y131) whereas UpRh3 has a phenylalanine (F134; Fig. 2) residue corresponding to that position. UpRh3 also contains a conserved lysine (K111) that is unique to all insect short-wavelength opsins. In contrast, UpRh1 and UpRh2 have a conserved serine (S142) found in all opsins except the short-wavelength opsins (Fig. 2). Together, these data predict that UpRh1 and UpRh2 are long-wavelength-sensitive opsins, and UpRh3 is a short-wavelength-sensitive opsin.

To further test these predictions, we determined the phylogenetic relationships of the U. pugilator opsins to other arthropod opsins of known sensitivity (Fig. 3). UpRh1 and UpRh2 are nested within the middle-wavelength opsins. In particular, UpRh1 is 91% identical to U. vomeris opsin1 and 84% identical to H. sanguineus BCRH1 opsin, the latter has an estimated maxima at 480 nm (Sakamoto et al., 1996). UpRh2 is 87% identical to U. vomeris opsin2 and 91% identical to H. sanguineus BCRH2 opsin, also sensitive to light at 480 nm (Sakamoto et al., 1996). UpRh3, by contrast, is only 50% identical to the opsins identified from U. vomeris and H. sanguineus. It is more closely related to short-wavelength opsins (SW), with 71% similarity to ultraviolet-sensitive opsins from B. impatiens (absorption maxima unknown) and A. mellifera (absorption max. 353 nm), respectively.

The retina of U. pugilator contains ∼8000 nearly identical subunits called ommatidia (Smolka and Hemmi, 2009), each containing eight photoreceptor cells whose microvilli make up a fused rhabdom. In brachyurans, one of these photoreceptors is small and is situated at the distal ends of the cell bodies of the remaining seven photoreceptors. The rhabdomere of this small retinula cell, R8, is found only at the distal tip of the rhabdom in every ommatidium (Marshall et al., 1999). The remaining seven photoreceptors in U. pugilator have different sized and shaped cell bodies (Fig. 4). To determine the expression patterns of the U. pugilator opsin genes, we performed in situ hybridizations on wax sections of randomly oriented eyes. Antisense riboprobes were synthesized from regions with high nucleotide sequence specificity for each opsin to identify transcript-specific expression. No staining was detectable with the sense riboprobes (data not shown).

The mRNAs for all three U. pugilator opsin genes are present in every ommatidium in the eye, with no apparent spatial differences between ommatidia from different eye regions. However, opsin-specific differences in expression were observed between photoreceptor cells within individual ommatidia. UpRh3, for instance, was restricted to the small R8 cell at the distal end of the photoreceptor portion of each ommatidium (Fig. 5), whereas the putative middle-wavelength opsins are present in the proximal cells R1–R7. UpRh1 mRNA was found in five of these seven cells (Fig. 6), and UpRh2 mRNA was found in three of these cells (Fig. 7).

Since seven retinula cells form the proximal ommatidia, our observation that UpRh1 and UpRh2 were expressed in five and three proximal photoreceptors respectively indicates that these genes must be coexpressed in one of the retinula cells. To better define the expression patterns of these related genes, we performed double in situ hybridization. The results confirmed those from single in situ hybridization, that these two genes share overlapping expression in one of the seven proximal photoreceptor cells (Fig. 8). Fig. 8A shows seven photoreceptor cells, five of which express UpRh1 and three express UpRh2. Expression patterns of mRNA for the latter three using a riboprobe specific to UpRh2 show that in a single ommatidium only these three photoreceptor cells express UpRh2 (Fig. 8B). Comparison of several such image pairs indicates coexpression of UpRh1 and UpRh2 opsins in a single photoreceptor cell.

Three sequentially distinct, full-length cDNAs of opsin genes were identified from U. pugilator retinas. By molecular analysis, UpRh1 and UpRh2 are predicted to be middle-wavelength sensitive, and UpRh3 is predicted to be short-wavelength sensitive. These data are consistent with previous behavioral tests and the physiological tests of Hyatt (Hyatt, 1974) and Horch et al. (Horch et al., 2002). The uniform expression pattern in ommatidia across the retina would seem to indicate a similarly uniform sensing capacity, but detailed studies on the physiological effect of screening pigments, and sensitizing pigments, if there are any, are required to determine whether there is any physiological heterogeneity between ommatidia.

The results from in situ hybridization show that all ommatidia of U. pugilator are composed of at least four types of photoreceptor cells that are unique in their spectral sensitivity, thereby satisfying the most fundamental requirement for color vision. This uniform expression pattern is somewhat surprising, since the dorsal and ventral visual fields of semi-terrestrial crabs are probably as consistently exposed to different light environments as any terrestrial animal, because of a highly efficient multi-sensory control system for maintaining eyestalk pitch and roll (Nalbach et al., 1989a; Nalbach et al., 1989b; Nalbach, 1990). However, with only three opsin genes, color vision in all directions probably requires such uniformity of expression. In addition to color discrimination, polarization sensitivity is common in several marine organisms, and its role is similar to color vision in enhancing contrast and conspicuousness, intraspecific signaling, and object identification (Cronin et al., 2003; Glantz and Schroeter, 2006; Glantz, 2008). In fiddler crabs, polarization sensitivity can apparently be used as a celestial direction cue for orienting to the crabs' home beach (Herrnkind, 1968; Chiussi and Diaz, 2002), although it appears to play no role in finer-scale navigation by path integration (Layne et al., 2003). Typically in insects, the dorsal rim area contains UV-sensitive photoreceptors which are also polarization sensitive, coinciding with the fact that natural light tends to be more polarized at shorter wavelengths. UpRh3, the opsin in the most distal cell in each ommatidium, is UV-sensitive and may likewise mediate polarization sensitivity, but this must be done in conjunction with at least one other cell, normally within the same ommatidium. Since the other cells in the ommatidium are not UV sensitive, there is potential for confounding the e-vector orientation and wavelength.

Based on their peak sensitivity to light, opsins are often classified into three groups: short-wavelength-sensitive opsins (SWS, ultraviolet–blue), middle-wavelength-sensitive (MWS, blue–green) and long-wavelength-sensitive (LWS, green–red). It is thought that the visual system of ancestral insects was trichromatic, with opsins sensitive to UV, blue and green (Briscoe and Chittka, 2001), and in some groups an expansion of the green opsin added red sensitivity, followed by the loss of the blue opsin (Jackowska et al., 2007). Although much less is known of the molecular evolution of crustaceans, crabs do not appear to follow this scheme. They are thought to descend from ancestors with two opsins, one sensitive to violet–ultraviolet, and the other sensitive to blue–green or green (Porter et al., 2007). Among the three opsins identified from the photoreceptor cells of U. pugilator, UpRh3 is predicted to be UV sensitive, and UpRh1 and UpRh2 to be MW sensitive. Although most other crustaceans possess an opsin sensitive to considerably longer wavelengths (∼500–530 nm), the only brachyuran crab whose opsin gene sequence and spectral sensitivity are both known, Hemigrapsus sanguineus, has two MWS opsins, both with peak sensitivity around 480 nm (Sakamoto et al., 1996). The highly similar fiddler crab opsins UpRh1 and UpRh2 can be predicted to have a similar peak, and are likely to be the result of an expansion event of the ancestral MWS opsin. The functional significance of this ostensible bifurcation of the MWS opsin as an adaptation in the color system of semi-terrestrial animals should be further examined in the context of their particular light environment.

The results from the double in situ hybridization reveal that the probes for the two opsins, UpRh1 and UpRh2, both label a single photoreceptor cell in each ommatidium. The colocalization of these two probes in a single photoreceptor cell demonstrates that the cell expresses two opsin visual pigments. The once commonly held notion that a single photoreceptor cell expresses only one functional visual pigment has been dispelled many times over, as evidence from both vertebrates and invertebrates has shown this to be an over-generalization. The expression of multiple opsins in a single photoreceptor cell has been reported in mice (Applebury et al., 2000; Lukats et al., 2002; Lyubarski et al., 1999), hamsters (Lukats et al., 2002), rabbits and guinea pigs (Röhlich et al., 1994), fish (Archer and Lythgoe, 1990; Wood and Partridge, 1993; Hope et al., 1998), salamanders (Makino and Dodd, 1996), butterflies (Kitamoto et al., 1998; Arikawa et al., 2003; Sison-Mangus et al., 2006), flies (Mazzoni et al., 2008), beetles (Jackowska et al., 2007), and several distantly related crustaceans (Sakamoto et al., 1996; Oakley and Huber, 2004; Frank et al., 2009). In fact, coexpression of opsins in single photoreceptors may even be unusually common in crustaceans (Cronin and Porter, 2008).

The coexpression of opsins now seems to be a fairly common phenomenon, but its functional significance is still a matter of speculation, although at least two reasonable theories exist. The first theory proposes a sensitivity enhancement function, and no role in color vision. Multiple opsin expression causes photoreceptor cells to have a broader spectral sensitivity and, presumably, increases the absolute sensitivity of the cell. These cells may therefore mediate achromatic vision in diurnal invertebrates under low light conditions, similar to rods in vertebrate scotopic vision, a function that has also been proposed for opsin coexpression in murine cones (Applebury et al., 2000). The second theory proposes a role in color vision. If the absorbance spectra of solo-expressed visual pigments are fairly similar, a ‘broad-band’ cell may be useful for spectral tuning. The two similar signals can be better discriminated by sharpening and amplifying their peaks, essentially by multiplication of the signal from the broad-band cell with the respective single peak signals, thereby providing the animal with finer color discrimination. A variation of this theory that does not strictly involve color perception is that the activity of a broad-spectrum photoreceptor is compared to that of a narrower UV photoreceptor to judge the ratio between UV and all longer wavelengths, which may be useful for orientation (Mazzoni et al., 2008). Although the specific absorbance curves to support this theory have yet to be demonstrated, we speculate that, of the two theories, the opsins expressed in U. pugilator conform better to the latter, color perception theory. Unlike, for instance, lycaenid butterflies which have a photoreceptor type coexpressing both a short and a long-wavelength opsin, probably conferring an extremely broad-sensitivity spectrum (Sison-Mangus et al., 2006), the two coexpressed opsins in fiddler crabs fall within the same spectral class. This means that the broadness of their combined spectral absorbance is narrower than it potentially could be, limiting its effectiveness for improving absolute sensitivity and potentially introducing an additional need for signal sharpening.

We would like to thank Dr Dan Rittschof for providing us with the experimental animals, Dr Elke Buschbeck for kindly allowing us to use her lab equipment and Dr Srdjan Maksimovic and Mark Charlton-Perkins for their technical assistance and helpful discussions.