Comparative study of spectral sensitivity, irradiance sensitivity, spatial resolution and temporal resolution in the visual systems of Ocypode quadrata and Aratus pisonii Free

The visual systems of most animals are highly evolved to efficiently extract visual signals from the background noise. Since the photons in visual signals are limited in dim-light environments, inhabitants face the classic trade-off between irradiance sensitivity and temporal resolution. Higher irradiance sensitivity usually requires a longer neural integration time, whereas better temporal resolution demands a shorter neural integration time in response to light. The best balance between sensitivity and resolution can be achieved by altering the composition of the visual pigments (Forward et al., 1988), the membrane properties of the photoreceptor cells (de Souza and Ventura, 1989; Laughlin and Weckström, 1993) and the structure of the eye (reviewed in Meyer-Rochow, 2001). Nocturnal animals and deep-sea animals require higher sensitivity (and the resultant lower resolution) to see in dim light (Frank, 2000; Johnson et al., 2002), whereas carnivorous (usually diurnal) predatory animals require higher temporal resolution at the expense of sensitivity to track their prey (Howard et al., 1984; de Souza and Ventura, 1989; Laughlin and Weckström, 1993; Frank, 2003). It has been hypothesized that differences in organisms' visual systems result mostly from differences in their ecology – primarily habitat and lifestyle, which includes prey preferences and activity cycles (Autrum and Stoecker, 1950; Autrum, 1958).

The animals in this study belong to the subphylum Crustacea, which consists of around 52,000 species worldwide (Cronin and Porter, 2008). Crustaceans occupy almost every conceivable niche within marine ecosystems (Cronin and Porter, 2008), and the visual systems of crustaceans are very diverse with respect to their eye morphology and physiology. This study sought to compare the visual physiology of the Atlantic ghost crab (Ocypode quadrata) and mangrove tree crab (Aratus pisonii), two decapod crustaceans occupying different niches in South Florida, specifically studying their spectral sensitivity, irradiance sensitivity, spatial resolution and temporal resolution.

The feeding ecologies of both species in this study were well described, but their visual ecologies were poorly understood. The Atlantic ghost crab Ocypode quadrata (Fig. 1A,B) is a fast-moving carnivorous crab of the family Ocypodidae inhabiting the sandy beaches from Block Island, Rhode Island, USA to Santa Catarina, Brazil (Williams, 1984). Most of the ghost crabs can reach a speed of over 2 m s⁻¹ (Burrows and Hoyle, 1973) and feed exclusively on other smaller crustaceans and mollusks at night (Wolcott, 1978). The mangrove tree crab Aratus pisonii (Fig. 1C,D) is an arboreal crab of the family Sesarmidae inhabiting the mangrove forests of the western Atlantic from central Florida to Brazil and the eastern Pacific from Sonora to Peru (Rathbun, 1918). Previous studies on the diet of the mangrove tree crab indicate that they mainly feed on mangrove leaves, arboreal algae (Beever et al., 1979) and occasionally on animal corpses (Erickson et al., 2008). These two species were chosen because of their vastly different feeding ecologies, providing an excellent test of Autrum's hypothesis that visual physiology is correlated with the life history of an animal.

The mangrove tree crabs [Aratus pisonii (H. Milne-Edwards, 1837)] were collected from the mangrove forests, and Atlantic ghost crabs [Ocypode quadrata (Fabricius, 1787)] were collected from the intertidal zones of exposed sandy beaches in Hollywood, Florida. Both species included both male and female specimens. All animals were transported to the laboratory in plastic containers containing seawater from the collection site and were kept in complete darkness until the experiments. The temperature of the laboratory remained constantly at 21.8°C. Fresh mangrove leaves were supplied in abundance in the container of the mangrove tree crab, while fish meat was provided to the Atlantic ghost crab once every 2 days. The seawater was changed daily, and all animals were used for experiments within 10 days of collection.

In this study, the spectral sensitivity, irradiance sensitivity, and temporal resolution of the crustaceans were quantified through the electroretinogram (ERG), which is an extracellular signal that represents the summed mass response of multiple photoreceptor cells to a light stimulus. The ERG is the method of choice because it directly records the responses in the receptor layer that takes into account any changes caused by pre-retinal filtering (Bryceson, 1986). All ERG experiments were conducted between 11:00 h and 17:00 h.

The electrophysiological recording method was based on methods used by Frank et al. (2012). All animals were dark adapted in a dark room for 7 h prior to the experiment. Animals were prepared for recordings under dim red light and remained alive during the experiments. They were suspended in a seawater bath inside a Faraday cage that was covered with a lightproof sheet. All the appendages of the animal were secured to its body using a thread, and the animal body was submerged in the seawater with only the tested eye slightly above the level of the water. A drop of cyanoacrylate glue was applied to the base of eye stalk to stabilize the eye. With the aid of a micromanipulator and a dissecting microscope (Olympus), a glass-insulated tungsten microelectrode (Frederick Haer) was inserted into the eye at the midpoint between the anterior and dorsal regions, resting its tip directly above the photoreceptor layer, to record electrical responses. A silver chloride electrode was placed in the water near the animal to ground out the background electrical noise in the water bath, and the entire apparatus was shielded with a Faraday cage to block extraneous electrical noise. The response was amplified with an X Cell-3 Microelectrode Amplifier (FHC, Inc.) with a high impedance probe (Kugel, 1977). Level of amplification was set to 2000×, and filters were set between 1 and 1000 Hz. Data were displayed on a laptop computer (PowerBook Titanium G4, Apple) and then digitized using a program written in LabView (National Instructions, Austin, TX, USA).

The pathway of a light stimulus from the light source to the eye was in the following order: light source, monochromator, shutter, neutral-density filter wheel, light guide and the eye. The full spectrum light (150 W quartz halogen source) was first adjusted to monochromatic light at the tested wavelength by a monochromator (CM110 monochromator, Spectral Products, Putnam, CT, USA). Flash duration was regulated by a computer-controlled shutter (Model VS14, Vincent Associates, Rochester, NY, USA), and light irradiance was controlled with a neutral-density filter wheel (Edmund Optics) driven by a computer-controlled stepper motor (all control programs were written in LabView). Test flashes of light were transmitted towards the eye using a bifurcated light guide (EXFO) composed of randomized silica fibers. The end of the light guide was positioned 1 cm above the insertion site and aligned at a right angle to the corneal surface, ensuring the entire anterior and dorsal regions of the eye were bathed in light. The irradiance of each flash was calibrated in units of photons cm⁻² s⁻¹.

At the onset of the experiment, a dim test flash with a wavelength of 490 nm was administered, starting at an irradiance of 10⁸ photons cm⁻² s⁻¹. The irradiance was then gradually increased until the elicited response ranged between 35–50 μV. Measurements began when the response to this initial test flash had not changed for 1 h. Subsequently, the eye was stimulated with stimulus flashes of monochromatic light, and flash irradiance was adjusted until the eye produced a criterion response of 50 μV. Duration of each flash was 0.1 s, with an interval of at least 1 min between flashes. The wavelengths of the flashes were in random order, ranging from 380 to 600 nm in increments of 10 nm. After testing each wavelength, a test flash identical to the initial one was given in order to assess the dark-adapted state of the eye. If indications of slight light adaptation were observed, the experiment was paused until the response to the test flash returned to its original value, within a tolerance of ±1 μV, ensuring that the eye had fully recovered to its dark-adapted state.

After obtaining the criterion response for each wavelength in the dark-adapted experiment, the study proceeded with chromatic adaptation experiments. An adapting light was channeled to the eye through the secondary path of the bifurcated light guide. This light, sourced from a halogen lamp (LS-1, Ocean Optics, Dunedin, FL, USA), passed through a 500 nm band-pass filter. The irradiance of the adapting light was adjusted with the neutral density filters, ensuring it exceeded the dark-adapted criterion response irradiance by at least one log unit. To confirm full chromatic adaptation, a dim test flash (irradiance differing from the dark-adapted test) was administered, and only when the elicited response stabilized for 10 min did the measurements commence. The eye was subsequently exposed to the identical series of wavelengths as those used in the dark-adapted protocol.

Data were plotted as the inverse of the irradiance required to evoke the criterion response at each wavelength and normalized to the wavelength of maximum sensitivity. Absorbance curves were fitted to the spectral sensitivity data in accordance with visual pigment templates (Stavenga et al., 1993).

Voltage versus log irradiance (V/logI) curves were generated from measurements made in dark-adapted eyes to compare the irradiance sensitivity of two species. The dark-adapted eyes were stimulated with 0.1 s stimulus flashes of increasing irradiances of 490 nm monochromatic light. The first stimulus flash was the dimmest flash and the irradiance started at around 10⁸ photons cm⁻² s⁻¹. The irradiance of each following flash was increased by a half log unit until the response was saturated. To ensure that the stimulus was given to a fully dark-adapted eye, a dim test flash was administered after each stimulus, and no further stimulus flashes were given until the response to the test flash had recovered to the dark-adapted level. The V/log I curves were plotted with the Zettler modification of the Naka–Rushton equation (Naka and Rushton, 1966a,b; Zettler, 1969): V/V_max=I^m/(I^m+k^m), where I is stimulus irradiance, V_max is maximum response amplitude the eye is capable of generating, m is the Hill coefficient which describes the steepness of the curve and k is the irradiance yielding a response that is 50% V_max. The dynamic range of 5–95% V_max was also labeled in the curves.

Temporal resolution was examined by determining the maximum critical flicker fusion frequency (CFF_max) and the response latency. CFF_max was tested under both dark-adapted and light adapted conditions, while the response latency was quantified in dark-adapted eyes. In the dark-adapted test, all animals remained fully dark-adapted, because light adaptation could affect the temporal resolution in certain species (Frank, 2003). A flickering stimulus of 490 nm monochromatic light at 15 Hz was presented to the dark-adapted eye for 2 s, and the flicker frequency was gradually increased until critical flicker fusion was achieved. The flickering light stimulus was generated by a computer-controlled electromagnetic shutter with a constant 50% duty cycle (50:50 light:dark ratio). To ensure that the eye remained dark-adapted, a dim test flash was given between each measurement, and subsequent flickering stimulus would not be given until the amplitude of the test flash had returned to its dark-adapted level. Irradiance was then increased by one log unit, and the flicker rate of the stimulus light was increased until critical flicker fusion is again achieved. CFF_max was determined as the point at which the eye could no longer respond to each individual flash of light. The response latency, equal to the elapsed time between the onset of stimulus and the onset of response, was measured under a test flash with an irradiance of k (measured in irradiance sensitivity test), which is the irradiance that produced a response that was 50% of the V_max.

It was not known whether the animals have apposition or superposition eyes prior to the experiment. The clear zone in superposition eyes would be heavily covered with screening pigments if the eyes were light adapted, which will make it extremely difficult to differentiate apposition from superposition eyes. Therefore, the collection of the eyes for histology was conducted under red light so that all specimens were fully dark adapted.

To make sure the fixative penetrated to the center of the eye, each eye was cut into two equal sagittal sections before fixation in a mixture of 2.5% glutaraldehyde and 3.7% formaldehyde in seawater for 2–4 h at room temperature (Alkaladi and Zeil, 2014). Samples were subsequently washed in three changes of seawater for 5 min each and then dehydrated through a sequence of 20%, 50%, 70%, 90%, 95%, and 100% ethanol for 15 min each. Samples were embedded in paraffin wax and sectioned at a thickness of 5 μm.

The tissue sections were then deparaffinized, hydrated with water, and then stained for 5 min using Mallory–Heidenhain stain (1 g phosphotungstic acid, 2 g Orange G, 1 g water-soluble aniline blue and 3 g acid fuchsin in 200 ml distilled water). Sections were viewed and photographed under a light microscope, and the digital images were analyzed using Image J. The following structures were examined histologically: screening pigments, the presence or absence of clear zone, rhabdoms and facets.

Since histological examination only offers a two-dimensional perspective of eye sections, CT scanning was employed to offer a comprehensive three-dimensional view of the eye's morphology to validate and corroborate the findings from histological analyses. This 3D visualization provides a detailed assessment of the eye surface, facilitating accurate measurements of facet diameters and elucidating the overall shape and curvature of the eye. Samples were stained with 3% phosphotungstic acid in 70% ethanol for 21 days and scanned by a nano-CT machine (phoenix v|tome|x m). The 3D visualization was generated using Dragonfly software (Object Research Systems, 2020).

The minimum interommatidial angle (ΔΦ), the angle of separation between adjacent ommatidial axes, was used to quantify spatial resolution and was calculated by dividing the effective distance between adjacent facets by the radius of the eye curvature. In an eye arranged with hexagonal facets, the effective facet distance, which occurs in the vertical direction in this study, is √3/2 of the facet diameter. The spatial resolution of eye regions on the vertical plane was assessed using localized data of these distances and eye curvature. The relative facet elevation, reflecting the vertical position of the facet, is determined by the vertical distance from the base of the eye (where it meets the eye stalk) in proportion to the overall eye length from base to tip. The local eye curvature, assessed vertically, was measured by fitting circles to the images of eyes, with the radius of the eye curvature being equal to the radius of the circle (following Baldwin Fergus et al., 2015). Cycles per degree (CPD) was used for studying visual acuity and was calculated as the reciprocal of 2ΔΦ.

In order to determine if the animal possessed an acute zone, an eye region with high visual acuity, the pseudopupils of both species were examined visually. The live animal was placed 20 cm in front of the eyes of the observer and rotated 360 deg vertically and then 360 deg horizontally. Any increase in the size of pseudopupil would indicate the presence of an acute zone. After an acute zone was identified, the spatial resolution was quantified from the ΔΦ in the acute zone, using the same methodology described above for calculating the ΔΦ in non-acute zone. Photos of pseudopupils in the anterior, dorsal, and posterior of the eye were taken with a camera (Canon I) connected to a stereo microscope (Meiji Techno).

The CFF_max, the response latency, the log k value, the minimum interommatidial angle and the cycles per degree in the acute zone, and the facet diameter were analyzed statistically to determine if there were significant differences between O. quadrata and A. pisonii. A Shapiro–Wilk test was used to test for normality, and a Bartlett's test was used to test the homogeneity of variances. Afterward, a two-sample t-test was used to analyze normally distributed data, while a two-sample Mann–Whitney Wilcoxon test was used to analyze non-normal data. All statistical analyses were conducted using the statistical software package R (https://www.r-project.org/) and null hypotheses (no difference between the two species for each of the factors mentioned above) were rejected when P≤0.05.

Each species showed only one peak in the dark-adapted spectral sensitivity curve, with peak sensitivity of 494 nm (O. quadrata) and 499 nm (A. pisonii). Chromatic adaptation at a wavelength of 500 nm was used to determine whether the animals have additional photopigments. The chromatic-adapted spectral sensitivity curves of both species were the same as their dark-adapted curves, which indicates that each species has only one blue sensitive photopigment. Any flashes at a wavelength below 400 nm or above 600 nm could not evoke a criterion response of 50 μV, and therefore the range of the sensitivity curves is between 400 and 600 nm (Fig. 2).

There was also no significance difference in response latency and dark-adapted CFF_max between the two species, and both species had a higher CFF_max when light adapted than when dark adapted (Table 1). The light-adapted CFF_max of A. pisonii was significantly higher than that of O. quadrata, even though the absolute difference was only 1.5 Hz.

The log k value, which represents the log irradiance required to elicit a response that is 50% of V_max, was significantly lower in O. quadrata compared with A. pisonii. Additionally, both species showed a similar dynamic range, which represents the log irradiance range between response limits of 5–95% V_max, with approximately 3 log units for both O. quadrata and A. pisonii (Fig. 3).

Ocypode quadrata has apposition eyes, as the end of the crystalline cone is directly connected to the rhabdom (Fig. 4C). The shape of the eyes is between a hemisphere and an ellipsoid (Fig. 4A). The facet size increased from the top and bottom part of the eye toward the middle (Fig. 4B). The radius of eye curvature and the size of pseudopupil increased substantially in the middle anterior region of the eye, indicating that the O. quadrata eye possesses an acute zone. The pseudopupil of O. quadrata only elongated vertically in the acute zone but not horizontally (Fig. 5A,B), and therefore the spatial resolution varies between their vertical and horizontal vision. In the acute zone, the vertical interommatidial angle (ΔΦ) has reached the minimum value of 0.40 deg with a CPD of 1.25.

A. pisonii also possess apposition eyes (Fig. 4D–F), with males having larger corneal facet size as a result of their larger body and eye size compared with females. This difference in size results in facet radii ranging from 25 to 43 μm. However, despite this variation in the facet radius, ΔΦ and spatial resolution visually appeared consistent across all tested individuals, as a larger facets radius is also associated with a proportionally greater radius of eye curvature in A. pisonii. It is imperative to note that this study does not offer a comprehensive analysis of eye size variations across distinct age groups or between sexes. The shape of the eyes in A. pisonii is a hemisphere (Fig. 4D), and an acute zone with a ΔΦ of 0.86 deg was found in the anterior region of the eye. Unlike O. quadrata, spatial resolution does not vary between the vertical and horizontal vision in A. pisonii, and the pseudopupil in the acute zone increased in size both vertically and horizontally (Fig. 5C,D).

The results of the dark-adapted sensitivity experiments showed that there is no difference in the spectral sensitivity between O. quadrata and A. pisonii – both peaked in the blue region of the spectrum (494 nm and 499 nm, respectively). This finding is consistent with earlier studies that demonstrated that most terrestrial and shallow-water decapods are maximally sensitive to wavelengths between 480 and 540 nm (Cronin et al., 1995; Forward et al., 1988; Johnson et al., 2002; Scott and Mote, 1974; reviewed in Marshall et al., 1999). Although the Atlantic ghost crab and mangrove tree crab have different habitats – a sandy beach for the former and a mangrove forest for the latter – both species live by the shoreline and have a planktonic larval stage in shallow-water areas where the light spectrum peaks in the blue–green region. The maximum sensitivity peak of both species in the blue region of the spectrum aligns with the wavelength of light that penetrates best in shallow water. This is consistent with the sensitivity hypothesis, which states that the photopigment matches the spectral composition of light in its habitat for maximum sensitivity to the available light (Clarke, 1936; Munz, 1958).

Whether an animal was diurnally active (mangrove tree crab) versus nocturnally active (Atlantic ghost crab) did not play an important role in their spectral sensitivity since the spectrum of moonlight is very similar to daylight, with only a slight shift toward the longer wavelengths (Ciocca and Wang, 2013). Atlantic ghost crabs and mangrove tree crabs are both terrestrial species, and the light that passes through their eyes has been scattered similarly by the atmosphere. As a result, the background light in their respective habitats has a similar spectral composition, which led to a similar evolution in their spectral sensitivity.

The spectral sensitivity curves did not change between dark adaptation and chromatic adaptation, which suggests that both the Atlantic ghost crab and the mangrove tree crab have only one photopigment. The absence of the sensitivity at UV wavelengths indicates that their R8 rhabdom cells, which are known to contain UV-sensitive visual pigments in decapod crustaceans (Cummins and Goldsmith, 1981; reviewed in Marshall et al., 2003), are either lost or reduced. However, numerous previous studies have demonstrated that several species of fiddler crab, which belong to the same family as the Atlantic ghost crab (Ocypodidae), display varied spectral sensitivities, notably including UV perception (Detto and Backwell, 2009; Horch et al., 2002; Jessop et al., 2020; Rajkumar et al., 2010). This difference may suggest a potential evolutionary distinction in the visual systems of phylogenetically proximate species. The flat sandy beach is a relatively monotone environment (Fig. 1A), and a monochromatic visual system should provide sufficient contrast detection for an Atlantic ghost crab to spot prey. On the other hand, the mangrove forest habitat of the mangrove tree crab is a complex three-dimensional environment (Fig. 1C). However, since the mangrove tree crab is an herbivore, a monochromatic visual system is adequate for it to recognize mangrove leaves in mangrove forests. Therefore, while color vision could be advantageous, it might not act as a substantial driving force in the evolution of these two species.

The dark-adapted CFF_max of the Atlantic ghost crab was the same as that of the mangrove crab (33.5 vs. 33.7 Hz), indicating that both species have the same temporal resolution. Although the light-adapted CFF_max of the ghost crab was significantly higher than that of the mangrove crab, the difference of only 1.5 Hz is too small a difference to have an effect on their ability to track moving objects, and this small difference, together with no difference in their dark-adapted CFF_max or response latency indicates that both species have the same temporal resolution. Considering the dark-adapted CFF_max of the mangrove tree crab and Atlantic ghost crab, their temporal resolution aligns closely with other shallow-water non-predatory decapods For instance, cleaner shrimps exhibit dark-adapted CFF_max values that range between 34 and 39 Hz (Caves et al., 2016), and fiddler crabs display values around 34.6 Hz (Brodrick et al., 2022) and 32 Hz (Layne et al., 1997). However, when compared with flying insects such as tsetse flies (85 Hz) and fruit flies (60 Hz) (Miall, 1978), and diurnal predatory crustaceans such as snapping shrimps (160 Hz) (Kingston et al., 2020), their temporal resolution appears lower.

Like most insects and other crustaceans (Laughlin and Weckström, 1993; Frank, 2003), both Atlantic ghost crab and mangrove tree crab have shown an improvement in the temporal resolution under light adaptation. This increase in temporal resolution in brighter environments is thought to reflect the trade-off between temporal resolution and irradiance sensitivity. A higher temporal resolution implies that the brain processes visual information more rapidly, which can enhance the capability of an animal to track fast-moving objects. However, this increased processing speed also means there is less time to collect photons for each visual response, which can compromise the quality of vision, especially in dim light conditions. This trade-off between temporal resolution and irradiance sensitivity is reduced in brighter environments because the abundance of light helps maintain an adequate signal-to-noise ratio even during rapid visual processing. Moreover, an individual's temporal resolution, as measured by its CFF_max, can also be influenced by other factors such as ontogeny and temperature, which have been shown to increase temporal resolution in many species (Frank, 2017). In a recent study on fiddler crabs, both light-adapted and dark-adapted CFF_max showed significantly higher values during the middle of the day compared to other times, indicating a strong correlation between temporal resolution and the circadian clock (Brodrick et al., 2022). Notably, the light-adapted CFF_max in this study was tested immediately after the dark-adapted test, which may have induced a transition from night to day for the crabs. Therefore, it is possible that these crabs could achieve an even higher light-adapted CFF_max when their circadian clock is fully entrained to the middle of the daytime.

Feeding ecology may have played a role in the evolution of temporal resolution in both the Atlantic ghost crab and the mangrove tree crab. Based on feeding habits alone, the Atlantic ghost crab, a predator that preys on mobile prey, would be expected to have a higher temporal resolution than the mangrove tree crab, which feeds on non-motile leaves. However, the day/night activity patterns of the two species may also have an impact on temporal resolution. For instance, as a nocturnal predator, the Atlantic ghost crab may require higher photosensitivity and longer integration time to detect prey in dim light conditions, which comes at the cost of lower temporal resolution. This is similar to how increasing exposure time in photography increases brightness but causes blurriness of any moving objects in the image. It is possible that the effects of predator/prey interaction and day/night activity patterns have counterbalanced each other, leading to similar temporal resolution in these two species with different ecological niches.

Irradiance sensitivity of apposition eyes is a function of integration time and facet size. The response latency of the Atlantic ghost crab is about the same as that of the mangrove tree crab, suggesting no difference in their integration time. However, the log k value of the Atlantic ghost crab eye is significantly smaller than that of the mangrove tree crab, meaning that it takes significantly less light to produce a response that is 50% of the maximum response that the eye is capable of generating (V_max), which indicates a significantly higher irradiance sensitivity in the Atlantic ghost crab. As there is no difference in temporal resolution, this difference in the irradiance sensitivity must result from the significant differences in their optics or eye sizes. Both species have apposition eyes, but the Atlantic ghost crab has a significantly larger corneal facet diameter (53.8 vs. 33.4 μm), meaning that each rhabdom has a larger aperture, collecting more light and thus providing a partial explanation for the significantly higher photosensitivity in the Atlantic ghost crab.

This is similar to what has been found in two other species of crabs with similar activity levels – one nocturnal and one diurnal. The facet diameter of Leptograpsus variegatus, a nocturnally active crab living by the shoreline (similarly to the Atlantic ghost crab), has a facet diameter of 45 μm (Stowe, 1980), which is considerably larger than the 25 μm facet diameter in Uca lacteal, a low-tide diurnal crab (Alkaladi and Zeil, 2014).

Enlarged corneal facets have also been found in many other nocturnal and deep-sea arthropods. Only 0.0001% of the surface light remains at a depth of 400 m in clearest ocean water (Jerlov, 1976) and some deep-sea crustaceans have evolved extraordinarily huge facets for increased photosensitivity (reviewed in Land and Nilsson, 1990). For example, both the isopod Cirolana (a shallow water species but in water so murky that the light intensity is equivalent to that at 600 m in clear ocean water) and the deep-sea amphipod Phronima have apposition eyes with facet diameters of 150 μm and 100–135 μm, respectively (Nilsson and Nilsson, 1981; Land, 1981). Warrant et al. (2004) found similar adaptations in several species of bees. Like all other bees, the nocturnal sweat bee, Megalopta genalis, has apposition eyes, but the photosensitivity of their eyes is almost 30 times greater than the eyes of diurnal honeybees. This is primarily due to the relatively large facets, which are 36 μm average diameter in the nocturnal species, whereas the diurnal species have an average facet diameter of 20 μm. All these studies including the current study support the hypothesis that in general, nocturnally active species would have larger corneal facets than diurnally active species.

Both the Atlantic ghost crab and mangrove tree crab have apposition eyes, which is indicated by the absence of a clear zone between crystalline cones and rhabdoms, and the lack of reflecting tapetum. A reflecting tapetum is a tissue layer located behind the photoreceptors that enhances photosensitivity by reflecting light back to the photoreceptors. This usually leads to the appearance of ‘eye glow’ or tapetal reflection when tested with a beam of light at night, which was not observed in either of these species. Moreover, apposition eye structure is consistent with what has been observed in other species of the families Ocypodidae and Sesarmidae (Arikawa et al., 1987; Alkaladi and Zeil, 2014; reviewed in Gaten, 1998). Although the apposition eye type of the mangrove tree crab is consistent with their diurnal lifestyle, where less light is needed to reach the retinula cells, the apposition eye type of the nocturnal Atlantic ghost crab is not consistent with most nocturnally active species. However, the results presented in this study show that the Atlantic ghost crab has a large facet diameter, which allows their eyes to be as sensitive as a superposition eye at night. Although superposition vision is generally more advantageous for nocturnally active species, there are also nocturnally active insects, such as crickets, locusts and cockroaches, with apposition optics (reviewed in Honkanen et al., 2017), indicating that superposition vision is not necessary for nocturnal activity.

In both species, the facet size is larger in the acute zone, a trait observed in numerous other animals possessing compound eyes (as discussed in Land and Nilsson, 2012). Diffraction of light has the potential to diminish spatial resolution, allowing light from unintended directions to enter the rhabdom through diffraction at the aperture. A more expansive facet in the acute zone can mitigate this effect and improve the signal-to-noise ratio.

Even though the Atlantic ghost crab has larger facets, which are often associated with animals living in dim light environments with lower spatial resolution, it has a larger eye than the mangrove tree crab (average diameter of 3.18 versus 1.87 mm) and possesses a larger number of ommatidia with longer rhabdoms and a flatter eye curvature. A flatter curvature increases the spatial resolution as it allows more ommatidia to view an area. Longer rhabdoms increase photosensitivity, as there is a greater chance that photons will be absorbed by visual pigments during their trip through a long rhabdom. The ΔΦ, the ratio of facet diameter to the radius of eye curvature, is significantly smaller in the Atlantic ghost crab than that in the mangrove tree crab, giving them the higher spatial resolution they would need to identify motile prey, while the larger eye diameter counteracts the usual reduction in sensitivity that is associated with a better spatial resolution.

The larger radius of eye curvature in the Atlantic ghost crab compared with the mangrove tree crab mainly originates from the difference in the shape of their eyes. The ellipsoid eye of the Atlantic ghost crab is a divergent trait in the Ocypodidae family, which typically has an acute zone in the middle anterior region of their eyes, resulting in high vertical spatial resolution compared to other arthropods (Zeil and Al-Mutairi, 1996; Zeil and Hemmi, 2006). The Atlantic ghost crab has evolved these traits to an extreme, with a huge acute zone and extremely high vertical resolution. This adaptation is crucial since Atlantic ghost crabs are not only voracious predators but also prey to many other animals, such as raccoons and shore birds. Their high vertical resolution enables them to distinguish predators from their prey easily, as the former typically are seen above the retinal equator while the latter usually appear at or below it (Layne, 1998; reviewed in Zeil and Hemmi, 2006). This high vertical resolution, as hypothesized by Zeil et al. (1986), may further allow them to monocularly gain depth and size information directly from retinal position and retinal size. Eyes on raised stalks further enhance the visual advantage of the Atlantic ghost crab by minimizing the portion of the scene obstructed by the carapace, offering a fully panoramic field of view (Zeil and Hemmi, 2006). However, since these crabs need to fold their long eyes into the optic sockets on their carapace when they burrow, their two eyes are very close to each other, leading to poor binocular vision for stereopsis (perception of 3D environments, reviewed in Schwind, 1989; Alkaladi and Zeil, 2014). As a result, most Ocypodidae crabs, including the Atlantic ghost crab, are ground dwellers living in a flat environment. In contrast, mangrove tree crabs have spherical eyes, which are common in the Sesarmidae family. Their eyes are widely spaced, providing them with excellent binocular vision that enhances their stereopsis (reviewed in Schwind, 1989), aiding their navigation in 3D mangrove forests.

When attempting to correlate visual adaptations with environmental characteristics, it is important to not only look at activity cycles (diurnal versus nocturnal), but also feeding ecology. The ghost crab is an active predator, needing higher spatial resolution, but is also nocturnal, needing greater sensitivity, and the two requirements seem to conflict with each other. This study demonstrated the importance of using both histological and electrophysiological methods to study visual adaptations, as the electrophysiological results demonstrated that the nocturnal ghost crab was significantly more photosensitive than the diurnal mangrove crab, as expected, but the lack of differences in temporal resolution would have made this a puzzling result. However, the histological studies demonstrated that this difference in photosensitivity originated in differences in the eye morphology of the two species. Larger eyes together with larger facet diameters can produce both an increase in spatial resolution and an increase in photosensitivity, which a nocturnal active predator would need, whereas the smaller facet diameters and smaller eyes are sufficient for a diurnally active species specializing on non-motile prey. The prey differences also explain why the ghost crab can afford to have a metabolically more expensive larger eye, while the less active mangrove crab can obtain sufficient nutrients from its herbivorous diet to support its smaller eyes.

Portions of the results and discussion in this paper are reproduced from the Masters thesis of Ruchao Qian at Nova Southeastern University (Qian, 2020).

All relevant data can be found within the article and its supplementary information.