Introduction

The assertion that photoreception and vision are probably the most important modes by which animals belonging to a multitude of different evolutionary lineages explore their environment, modulate their behaviour and shape their learning processes, is surely obvious and widely shared. Light is an important environmental signal for almost all cellular organisms, and photoreceptors are present in bacteria, fungi, plants, and animals (Karniol et al., 2005). Despite the wide variety of molecules involved in the processes of photoreception and image formation in animals, the basic biochemical mechanisms show remarkable similarities (Porter et al., 2016). The appearance of eyes can be traced back to a very early stage of metazoan evolution, leading to the formation of visual systems of varying complexity (Land, 2014). The oldest evidence of a visual system is a compound eye of trilobites from the Cambrian (Schoenemann et al., 2017). It is precisely amongst Arthropoda that there is an extraordinary diversity of photoreceptors, optical mechanisms, and eye types, which in turn influence their behaviour and ecological functions (Yilmaz et al., 2022). Within this phylum, a large range of different visual systems can be specifically found in crustaceans (Palecanda et al., 2022).

The available information on the structural and physiological characteristics of the photoreceptive apparatuses in Ostracoda, a class of Crustacea, is largely related to marine species, whilst knowledge of these topics for non-marine species is still rather limited. Ostracods are small, bivalved crustaceans found in all types of aquatic habitats and also in semi-terrestrial environments. They are the most abundantly preserved arthropod group in the fossil record ranging from the Ordovician through to the Holocene, where they are present with their calcitic carapaces and very rarely also with preserved soft parts (Matzke-Karasz & Smith, 2022). According to the World Ostracoda Database, the global number of extant and extinct species of marine, non-marine ostracods amounts to over 45,000 (Huang et al., 2022). Living representatives belong to the subclasses Myodocopa and Podocopa (Horne et al., 2005). Myodocopa are exclusively marine and comprise planktonic and nekto-benthic species; Podocopa occur in marine, brackish and freshwater environments and are mostly benthic. For living non-marine ostracods, > 2300 subjective species are reported in four superfamilies: Cypridoidea, Cytheroidea, Darwinuloidea, and Terrestricytheroidea (Meisch et al., 2019). Adaptations to habitats with different ecological requirements could have probably shaped the morphology and evolution of the eye in distinct phylogenetic lineages of this subclass, as it has been hypothesised for the Copepoda, another highly diverse class of Crustacea (Steck et al., 2023). A comprehensive overview on morphology, ecology and behaviour of non-marine ostracods can be found in Smith et al. (2015).

We here present a review of studies on visual systems in non-marine ostracods, including information on their anatomical, biochemical and physiological characteristics. Special attention is paid to the adaptive significance of light-gathering ability in relation to ecological interactions and learning processes, and areas of ignorance that still need to be adequately addressed are also highlighted.

The photoreceptive systems of ostracods

Ostracods use different sensory systems to explore their environment and interact with other organisms (see Appendix S1). For what concerns sight, Myodocopa and Podocopa are characterised by different light-gathering organs. The photoreceptive system of Podocopa is characterised by naupliar eyes, which are often called "median", "frontal", or even "simple" eyes, and have similar characteristics in different crustacean groups (Elofsson, 1992). Myodocopa may have simple apposition compound eyes, both compound and naupliar eyes, or, more rarely, only naupliar eyes (Land, 1980; Land & Nilsson, 1990; Elofsson, 1992). In both subclasses, the absence of any eye structure or at least of the eye pigment is possible (e.g. in the marine suborders Cladocopina and Platycopina and many hypogean non-marine species).

The evolution of the current ostracod eyes and their diversification into compound and naupliar types is still under debate (see e.g., Fryer, 1996; Oakley & Cunningham, 2002; Dingle, 2003; Oakley, 2003; Elofsson, 2006; Hunt, 2007; Horne, 2010; Syme & Oakley, 2012; Oakley et al., 2013; Schoenemann & Clarkson, 2023).

Evidence suggests that Myodocopa are monophyletic and nested in other groups with different eye types, based on comparative studies using molecular data of some genes coding for ribosomal RNA (Oakley & Cunningham, 2002; Oakley, 2003; Elofsson, 2006), although previous studies found very different results (see e.g., Fryer, 1996). The naupliar eyes’ homology is supported by phylogenetic data and morphological features (Paulus, 1972, 1979, 2000; Bitsch & Bitsch, 2005).

Even if the debate on the homology of the two types of ostracod eyes is still open and research is ongoing, providing a complete overview of the topic is beyond the scope of this review (see Appendix S2 for further information). It is, however, of interest to confront the ostracod naupliar eyes, namely in non-marine podocopids, with those potentially homologous, such as the eyes of Branchiura, Copepoda, and Cirripedia (Elofsson, 1965, 1966; Oakley & Cunningham, 2002; Regier et al., 2005).

General morphology of naupliar eyes in ostracods

Structurally, the naupliar eye of Podocopa can be observed in different subtypes with little or conspicuous "variation on the theme". The literature is constellated with detailed research papers and reviews on the topic (Müller, 1894; Kesling, 1951; Fahrenbach, 1964; Hartmann, 1966; Tanaka, 2005; Elofsson, 1966, 2006).

Since the beginning of ostracod research in the eighteenth century, naupliar eyes have been illustrated as a black dot, incidentially indicating the anterior body half and thus defining the orientation of the animal (Fig. 1a). Ostracod naupliar eyes are located between the brain and the anterodorsal part of the body, near the bases of the antennulae (Fig. 1b,c), yet more or less detached from the protocerebrum (Elofsson, 1966, 1992, 2006). The ostracod naupliar eye is usually divided into three parts, two lateral and one median, each with a photoreceptor organ housed within a cup-shaped shielding pigment. The three eyecups are either fused together (Fig. 2), as observed in most freshwater Cyprididae (e.g. in genera Heterocypris, Cypridopsis, Herpetocypris, Eucypris, and Cypris), or more or less separated (Fig. 2), as found in e.g. Notodromas monacha (O.F. Müller, 1776) and the brackish-marine species Xestoleberis aurantia (Baird, 1838) (Vávra, 1891; Claus, 1892; Müller, 1894; Nowikoff, 1908; Schreiber, 1922; Rome, 1947; Kesling, 1951; Elofsson, 1966). When the cups are separated, their position depends on the degree of separation between the eyes. The general disposition entails an anterior ventral cup, often underdeveloped, and two dorso-lateral paired ones (Elofsson, 1966, 1992, 2006). Such organization of what are often called “frontal eyes”, is common amongst crustaceans in general, and it is most likely used for body orientation in relation to light, although image-forming cannot be excluded (Brusca & Brusca, 2002; Srinivasan & Zhang, 2004; Elofsson, 2006; Nilsson, 2009, 2013). Lüders (1909) advanced the hypothesis that median eyes, which are iridescent when lit by light, such as those of podocopid ostracods, may serve the role of coloration by reflecting the light, making the animal “glowing”. This hypothesis is discussed in Hartmann (1967). The specific conformation of the median eyes with separated cups is typical in ostracods and in a few other taxa, whilst some specific cells composing the eye are peculiar to Podocopa, with only a few similarities with Copepoda. For example, for copepods, see Steck et al. (2023), Elofsson (1969), Ong (1970), Fahrenbach (1964) and Dudley (1969); for podocopid ostracods, see Andersson & Nilsson (1981) and Meyer-Rochow (1999); finally, for a comparison between copepods and ostracods, see Elofsson (1966) and Kaji & Tsukagoshi (2010b). The axons often reach the protocerebral region independently, with the ventral eye axon group running along with the two lateral ones (Nowikoff, 1908; Elofsson, 2006). In Cyprididae, especially in non-marine representatives, Turner (1896) observed a single optic nerve that branches into three separate courses when approaching the eyes. However, Rome (1947) states that Turner was mistaken, after finding in Heterocypris incongruens (Ramdohr, 1808) that three completely separate nerves connect the naupliar eye with the thick anterio-median area of the protocerebrum (Fig. 3).

Fig. 1
figure 1

A Before photographs became a common medium to illustrate living ostracods, the naupliar eyes were one of the features included in the figured set of characters, not at least to define the body orientation. The colourful pencil drawings by Walter Klie, a renowned ostracodologist of the first half of the twentieth century here serve as an example. (W. Klie, unpublished, without scale.) B Light microscopic image of the right body side of a Tanycypris alfonsi (Nagler et al., 2014) with naupliar eye visible at the base of the pair of antennulae (arrow). C Dorsal view of a Eucypris virens specimen, with valves removed. The black naupliar eye is located directly under the carapace, in between the bases of the antennulae. Scales: 100 µm

Full size image
Fig. 2
figure 2

Upper part: Lateral views of Eucypris virens (left) and Newnhamia fenestrata King, 1855 (subfamily Notodromadinae) (right), valves removed. Whilst E. virens possesses a fused naupliar eye, the one of N. fenestrata consists of three separate eye cups, forming a Y-shape. The iridescent spot (here visible only in the left lateral eye) results from the light reflection by the tapetal cells. Scales: 250 µm. Lower part: Sketches of a fused eye (left) as in E. virens and a separated eye (right) as in N. fenestrata with their detailed possible composition of different elements. CL = cuticular lens; LC = lens cell; RC = retina cells; TC tapetal cells; PC = pigment cells. Cuticular lenses are only potentially present when the naupliar eyes are separated

Full size image
Fig. 3
figure 3

Sketch of a non-marine ostracod. Left valve and left set of appendages removed to allow view on fused naupliar eye, three nerves (blue) connecting the eye cups with the protocerebrum, and the ventral nerve chord (red). After Kesling, 1951

Full size image

Each cup can be formed by several specialised cells: pigment cells, tapetal cells, retinal cells, which form two-celled microvillar rhabdoms, and lens cells (see below for a detailed description of these cell types). The number of cells for each type tends to vary amongst taxa. When the cups are not fused, specific types of lenses are formed by transparent regions of the carapace directly above the lateral cups. A schematic comparison of the anatomy of the two possible types of podocopid naupliar eyes (Fig. 4) illustrates that so-called cuticular lenses can occur additionally to the lenses within the eye cup; their characteristics will be described in more detail below.

Fig. 4
figure 4

Scheme of podocopid naupliar eyes. A fused eyecups; B separated eyecups. Each vertical rectangular box represents one eyecup, with the inside lines being the rhabdoms, underlain by the tapetum (striped horizontal rectangles) and pigments cells (dark horizontal rectangles). Lens cells are shown as elliptic blue shapes above each cup (L = lateral, An = anterior, V = ventral), whilst cuticular lenses are only present above the lateral eyecups in separated eyes

Full size image

An example of a non-marine ostracod whose median eyes are characterised by the presence of all the above-mentioned specialised cells, including highly refractive cuticular lenses of the lateral cups, is Notodromas monacha, a neustonic species (Andersson & Nilsson, 1981). In Notodromas trulla Smith & Kamiya, 2014, from the A-7 instar onwards, the dorsolateral eyecups are separated, resembling those of the adults (Smith & Kamiya, 2014; Smith et al., 2022). Species in the subfamily Notodromadinae, such as the ones mentioned above, usually swim upside-down when feeding from the water surface (Smith & Kamiya, 2014). The existence of separate eyecups could be related to this lifestyle, as the transition from one diametrically opposite body position to the other—that is, from normal to upside down position—is a more complex behavioural pattern than simply swimming and requires a higher level of postural control (Fischer & Laforsch, 2018). In addition, the eyes are well positioned to detect predators approaching from beneath (Zhang et al., 2024). The hypodermal cells sometimes connect the eye stalk to the caparace, and, as seen in a Vargula species, may even produce a protective secretion for the eye (Huvard, 1990; Tanaka, 2005). The pigment and tapetal cells with the lens cells (if all present) create a sort of sandwich around the retinal sensory cells, with the pigment cells covering the space between the cups and connecting them (Fig. 5) (Müller, 1894). The retinal cells face downward, and their axons emerge from the oculus by moving outward and then bending inward towards the cerebrum (Elofsson, 1992). The presence or absence of one or more specialised cells influences the actual function for which the eyes are ultimately used. Most likely, the evolution and the loss of specific cells are consequences of the adaptation of an organism to its particular habitat, consistent with its phylogenetic and structural limitations (Seilacher, 1970). According to Hartmann (1967), smaller taxa have less developed eyes. Given the light scarcity, it is reasonable to imagine that deposit-feeders, subterranean species, and scavengers will rely more on chemoreception than photoreception (Hartmann, 1967; Speiser et al., 2013; Cronin & Feller, 2014; Glantz et al., 2014; Thiel & Watling 2015). An example of a podocopid genus with probably blind species is Typhlocypris, as embryos show eyes, whilst adults have reduced and depigmented eyes (Turner, 1896; Namiotko et al., 2014). In the literature, the number of specialised cells that compose the naupliar eye in podocopid ostracods does not always coincide (Vávra, 1891; Claus, 1892; Müller, 1894; Nowikoff, 1908; Rome, 1947; Kesling, 1951).

Fig. 5
figure 5

Histological section of Eucypris virens. A Section in frontal body plane (dividing dorsal from ventral), near dorsum, cutting through naupliar eye (N) and other organs. H = hepatopancreas; I = intestine (with food ball); F = female genital area with spiral canal; C = decalcified carapace. B Frontal section through the naupliar eye below the anterior eye cup, cutting the axons (Ax) of nerve cells connected to the retinal cells (RC). Underneath the latter, each cup possesses a layer of tapetal cells (TC) and pigment cells (PC). The arrow points to the nucleus of a pigment cell. Pigments in the area between the two lateral lenses (LC) indicate that pigment cells in fact connect the three eye cups of this fused naupliar eye. (Paraplast® embedding, 5 µm-section, Domagk’s Azan staining (Mulisch & Welsch, 2010)). Scales: 100 µm for A, 5 µm for B

Full size image

Pigment cells

The approximate number of pigment cells depends on the general anatomy of the visual system. When the ocelli are fused in one organ, three to five thick cells hold the eye together by filling the space between the cups (Fig. 5) (Elofsson, 2006). Conversely, when the cups are separated, roughly four pigment cells hold the eye together, forming a "Y" shape with the eyecups (Andersson & Nilsson, 1981). The pigment cell organelles, such as mitochondria, glycogen granules and the endoplasmic reticulum are intermingled with dark-coloured electron-dense granules of different sizes, which are non-uniformly scattered but tend to mainly occupy the cell borders (Elofsson, 1969; Andersson & Nilsson, 1981). The colour and chemical properties of granules suggest they are based on ommochromes, and their pigmentation can strongly differ between species (Fox & Vevers, 1960; Needham, 1974; Andersson & Nilsson, 1981; Nilsson, 2013). The specific function of such cells has yet to be entirely understood. However, in compound eyes, pigment cells work as protective shields for photoreceptors and rhabdomeres against excessive light intensity, as well as determining the angle of light acceptance for retinal cells (Stavenga, 1979; Hallberg & Elofsson, 1989; Douglas & Marshall, 1999).

Tapetal cells

These cells have little cytoplasmic content and a highly reflective ultra-structure originated by layers of anisotropic crystals arranged neatly parallel to each other and the strained cytoplasm but tangential to the cup curvature (Andersson & Nilsson, 1981). Since the cells completely cover the eyecups, they create a highly reflective concave mirror-like stratum upon each cup (Fig. 6). The number of crystalline layers varies incredibly between taxa, but it is certainly more than ten. The specific composition of the crystals is rarely discussed in literature. Although the presence of guanine crystals in the tapetal cell body seems a reasonable explanation for the characteristics of tapetal cells, other hypotheses involve crystals of pteridines, analogues of guanine, and spheres of lipid or melanin, the latter found in fish and decapods (Land & Nilsson, 2002; Palmer et al., 2018). The function of the tapetal cells, fundamental components of the naupliar visual systems, is to collect and focus light on the sensory cells layer. The tapetal strata of the ventral cup and lateral cups can display significant differences in some taxa. For example, the tapetal cells of the ventral cup can contain more cytoplasm surrounding their spherical nuclei and a higher number of crystal layers (Andersson & Nilsson, 1981). Thus, the cells appear thicker. Conversely, the tapetal cells covering the lateral cups can present oval nuclei disposed towards the borderline of the pigment cell layer (Andersson & Nilsson, 1981). Similarities are found in the tapetum structure between ostracods and copepods (Fahrenbach, 1964; Dudley, 1969; Elofsson, 1969). In the ostracod naupliar eyes, two kinds of structural degeneration can be observed: one that lacks the reflective tapetum, which seems to disappear independently of the other eye cells, the second presenting degeneration in the pigment cells layer accompanied by underdevelopment of the retina, suggesting that they may form a developing ocular module (Kaji & Tsukagoshi, 2010). Tapetal cells show various degrees of degeneration in ostracods and may frequently disappear in species with reduced naupliar eyes, as those adapted to interstitial environments (Kaji & Tsukagoshi, 2010).

Fig. 6
figure 6

Naupliar eye of Australocypris robusta De Decker, 1974, in transmission light microscope, using (A) inbuilt lighting from above and (B) and external lateral light. Only in (B) the light hits the eye cups in a way that the tapetal cells can reflect it, creating the impression of a glowing organ. Anterior is to bottom left corner. Scale: 100 µm

Full size image

Retinal cells

Retinal cells are pear-shaped sensory cells (Andersson & Nilsson, 1981). They arrange their long axis perpendicularly to the tapetum (Andersson & Nilsson, 1981). The number of sensory cells varies dramatically amongst taxa. Hartmann (1967) provides a summary of the findings regarding the number of sensory cells—from six up to 40—observed in several species. It is worth mentioning that the genus Cypris has an exceptionally high number of retinal cells and a large brain (Elofsson, 2006). In some species of the genera Leptocythere and Loxoconcha the retinal cells are degenerated (Elofsson, 2006). The cytoplasm of the retinal cell is often described as the transparent area between rhabdoms (Tanaka, 2005). The retinal cell contents are mitochondria, endoplasmatic reticulum, Golgi bodies, vacuoles and granular material (Andersson & Nilsson, 1981; Huvard, 1990). Differences in electron density and size can be observed by comparing the cytoplasms of the ventral and lateral eyecups. Each cell is attached to a single axon, and then the axons of several cells group together. Cross-sections of the axon show mitochondria, smooth endoplasmatic reticulum and neurotubuli both in Podocopa (e.g., Notodromas) and Myodocopa (e.g., Vargula) (Fahrenbach, 1964; Andersson & Nilsson, 1981; Huvard, 1990). From different studies on the podocopids Notodromas monacha, Herpetocypris reptans (Baird, 1835) and Cypris pubera O.F. Müller, 1776, generally, a rhabdom consists of the apical portion of no more than two adjacent sensory cells. The rhabdom then consists of two horizontal microvillar rhabdomeres, one for each sensory cell (Nowikoff, 1908; Rome, 1947; Andersson & Nilsson, 1981). Thus, whilst the rhabdoms are arranged perpendicularly to the tapetal cells layer, the microvilli lie on a plane parallel to the reflective tapetum (Elofsson, 1966; Andersson & Nilsson, 1981). The microvilli are oriented perpendicularly to the incident light if the cells are close to the centre of the tapetum, or obliquely in its more peripheral parts. The alignment of microvilli does not have a consistent directional structure amongst all rhabdoms in the few literature reports for podocopid ostracods (see Eguchi & Waterman, 1976; Douglas & Marshall, 1999; Meyer-Rochow, 2001; Marshall & Cronin, 2011) and, therefore, polarised vision should be precluded (Andersson & Nilsson, 1981). Similar structures have been observed in copepods, suggesting that they might serve for physiological functions such as membrane cycling (Elofsson, 1969; Eguchi & Waterman, 1976).

Lens cells

In some species, e.g., Heterocypris incongruens, the cells are positioned nearly outside the ocelli (Elofsson, 1966). The cup may contain one or more twisted cells that serve as a single lens as in the genus Leptocythere (Elofsson, 1966). Hartmann (1967) refers to two cells, whilst Nowikoff (1908) mentions three. When only one cell is present, it tends to be large, and half of its body is overabundant compared to the eyecup containing it, with the nucleus displaced toward the outer region (Elofsson, 2006). The cytoplasm is transparent, but its density can vary between the interior and outer areas of the cell (Andersson & Nilsson, 1981; Elofsson, 2006). In some cases, ventral and lateral lens cells show significant differences in electron density, with the former having the highest value (Andersson & Nilsson, 1981). Lens cells, when present, have a surprisingly low refractive index (about 1.358–1.371), even lower in the ventral cup, at least in Notodromas monacha (Andersson & Nilsson, 1981). The values of the refractive indexes of the lens cells are close to the ones of the retinal cells (approx. 1.35), but more importantly, of water (approx. 1.33) (Andersson & Nilsson, 1981; Tanaka, 2006). This observation suggests that, in general, lens cells play only a minor role in focusing light (Andersson & Nilsson, 1981; Tanaka, 2005, 2006). Instead, this lens possibly can improve the light-gathering efficiency of the overall eye, acting as a diffuser. Examples of similar structures exists in copepods, such as in the marine copepods Labidocera and Pontella, and in other taxa such as in the scallop Pecten (Land, 1965; Andersson & Nilsson, 1981; Ali, 2013). However, in the taxa mentioned above, the lens cells may have a higher refractive index or a refractive index gradient.

Cuticular lenses

The calcified cuticular carapace areas dorso-lateral to the two upper eyecups are formed like lenses (Fig. 7) that diffract the light towards the inside of the eyecup, possibly achieved by the arrangement of closely placed crystals parallel to the plane of the shell (Bonaduce & Danielopol, 1988). In support of this hypothesis, it was observed in the study by Tanaka et al. (2009) that the ultrastructure of the ocular tubercle of the fossil ostracod Primitiopsis planifrons Jones, 1887 is characterised by a prismatic layer perpendicular to the tangent of the curvature of the valves, a structure not found elsewhere in the valves. Only those upper, more laterally positioned cups, possess such calcified cuticular lenses, whilst the ventral cup does not. When the eyes are fused, the ocelli may be arranged a little more ventrally than when they are separated, so that the eyes receive the light simply by opening the valves (Fig. 8) (Bonaduce & Danielopol, 1988). There is little information on the existence and disposition of the crystals inside the calcified cuticular lens (Yamada, 2019). Cuticular lenses possess an exceptionally high refractive index, which in Notodromas range between 1.62 and 1.64, almost that of pure calcite (Andersson & Nilsson, 1981; Tanaka, 2005). The carapace outside cuticular lenses still preserves its role as a protecting shield (Benson, 1974; Tanaka, 2005). Due to their calcification, cuticular lenses are well preserved in fossil records, providing important paleo-evolutionary evidence for phylogenetic considerations in ostracods (Van Morkhoven, 1963; Benson, 1975, 1984; McKenzie & Peypouquet, 1984; Bonaduce & Danielopol, 1988; Kontrovitz & Myers, 1988; Puckett, 1991; Tanaka, 2006). The environment in which the organism lives strongly influences the permanence of certain vision-related traits. This hypothesis is coherent with the finding that deep-water species in the cytheroidean genera Cytheropteron (Cytheruridae), Palmoconcha (Loxoconchidae) and Cluthia (Leptocytheridae) lack cuticular lenses (Tanaka, 2005). In addition, the thickness of the carapace and the size of the animal are important in limiting the development of the cuticular lens (Hartmann, 1967; Bonaduce & Danielopol, 1988). Indeed, intuitively, the thinner and smaller the carapace, the more unlikely it is that the cuticular lens will develop. For example, littoral species of Xestoleberididae do not display any cuticular lens even if many species of the same family have well-developed lenses. In a fundamental work, Bonaduce & Danielopol (1988) discussed the differences in cuticular lens morphology between the marine representatives of the family Xestoleberididae and the non-marine subfamily Notodromadinae. Surprisingly, these two groups, although different phylogenetically and in terms of habitat preferences, show great similarities in cuticular lens structure, if present (Bonaduce & Danielopol, 1988). However, unlike in Xestoleberididae, Notodromas monacha and Notodromas persica Gurney, 1921, display a pronounced sexual dimorphism in the shape of the cuticular lenses: Notodromas males have slightly bi-convex lenses, whilst females have rounded lenses (Bonaduce & Danielopol, 1988).

Fig. 7
figure 7

Newnhamia fenestrata. a. Right valve removed, with the (separate, right) lateral naupliar eye cup clearly visible by the light reflection of the tapetum. Light microscopical image. b. Right valve, with calcified cuticular lens located directly above the area of the right naupliar eye cup. Scanning electron microscopy image. Scale: 100 µm

Full size image
Fig. 8
figure 8

Anterior-dorsal view of Australocypris robusta. The opened carapace allows the fused naupliar eye to be directly exposed to light, as seen by the reflection of the tapetum of the frontal lens. Scale: 250 µm

Full size image

A broad categorisation of naupliar eyes considers the existence of three groups: one where the three cups are fused to create a unique organ without any lens system, another with fused eyecups and lens cells, and a last one with separated cups and with both lens cells and cuticular lenses (Rome, 1947; Elofsson, 1966; Andersson, 1979; Andersson & Nilsson, 1981; Huvard, 1990). Then, as highlighted by Tanaka (2005), at least six different subtypes can be distinguished by considering the presence or absence of specialised cells and their morphology (see Supplementary table S1 for a summary of the naupliar eyes subtypes in non-marine ostracods). Most species studied by Tanaka (2005) are marine; nonetheless, the paper cites some examples of non-marine species and families with non-marine representatives.

A second classification characterises the cuticular lenses based on specific morphological parameters, calculated according to a model on naupliar eye geometrical optics (Tanaka, 2006). Different types of naupliar eyes are classified by studying the relationship between the standardised curvature of the outer part of the cuticular lens, the relative lens thickness, and the light-gathering ability of the rhabdoms. Most families containing non-marine species, such as hemicytherids, loxoconchids, cytherids, cytherurids, and possibly leptocytherids, limnocytherids, cytherideids and eucytherids, display a medium degree of gathering ability and significant variability in the remaining two morphological parameters, although the species considered by Tanaka (2006) are all marine. It seems that genera, families, or species living in habitats with high irradiance tend to occupy the lower-ranked ability level.

The categories presented above do not explicitly consider the morphotypes of naupliar eyes in the case of species with degeneration of specialised cell tissues. Moreover, in the embryonic stage, the pigment and sensory cells appear almost simultaneously and before the appearance of tapetal cells. These observations suggest that in the hierarchical development of the structure of the eyes, the sensory and pigment cells might belong to the same module (Kaji & Tsukagoshi, 2010).

Physics of the podocopid naupliar eyes

Like any other optical system, eyes are subject to unavoidable flows originating from light diffraction and the defects of the optical elements composing them, such as lenses (for details see Appendix S3). Moreover, in general, animal eyes face a certain level of trade-off between spatial resolution and light sensitivity, with the latter developed preferentially in organisms living in low-light environments; naupliar eyes do not deviate from this rule (Land & Nilsson, 2002).

Spatial resolution can be quantified in terms of inter-receptor angle, defined as the ratio between the distance of the centres of two adjacent rhabdoms, and the distance from the rhabdoms to the nodal point, which is often the centre of the optical elements, and here it is considered as such for simplicity (see Appendix S4). The distance thus defined is approximately the focal length (Land & Nilsson, 2002). Since the distance between rhabdom centres has a lower physical limit, the most efficient way to increase resolution is to increase the focal length of the lenses, which is particularly unfeasible for small eyes such as those of the majority of ostracod species (Land & Nilsson, 2002). The focal length in non-marine ostracods may vary greatly amongst species. It is hypothesised that for visual systems with short focal lengths of about 100 μm diffractions have the main role in inducing image blurring. For Notodromas monacha, Andersson & Nilsson (1981) estimated a focal length for the lateral cups of approximately 103 μm. However, this observation does not help in determining the resolution capacities of the eyes of Notodromas monacha.

The same authors also hypothesised that in Notodromas monacha the diameter of blurred pattern formed in the retina and derived by light diffraction, a phenomenon called the Airy disc, is around 0.5–0.7 µm, which is much smaller than the rhabdoms, which, in general, have a cross-sectional area of 100 µm2 (Turner, 1896; Rome, 1947; Elofsson, 1966; Andersson & Nilsson, 1981).

The shape, size, and disposition of the rhabdoms play a fundamental role in determining the resolution and light sensitivity of the eye (see Appendix S5). For example, in podocopid naupliar eyes, the receptors may work as lightguides (Warrant & McIntyre, 1993; see Appendix S6). The major problem originating from lightguide receptors causing the diminishing of spatial resolution is that there is a threshold for the angle of incident light below which the receptor of interest is not able to collect the light by total internal reflection, and other non-targeted receptors will absorb it (Warrant & Mclntyre, 1990, 1993). This typically occurs in receptors of organisms displaying an exceptional light sensitivity that, unfortunately, leads to a so-called light spread in the retina (Warrant & McIntyre, 1990, 1991, 1993). As possible solutions to the absorption of the "escaped" light from receptors other than the one of interest, arthropods and thus also crustaceans evolved rhabdom shields made of light-absorbing granules of pigments or rhabdoms with non-cylindrical shapes (Land, 1984; Warrant & McIntyre, 1991, 1993). There are no clear mentions to this aspect in the literature with regard to the ostracod naupliar eyes, but there are references to black granules inside the retinal cells of some Leptocythere species and to ovoid-like shaped rhabdoms in Notodromas (Elofsson, 1966; Andersson & Nilsson, 1981). Therefore, it cannot be dismissed that other non-marine ostracods have developed ways to reduce the scattering of light information.

Again, Andersson & Nilsson (1981) measured the light-rays acceptance angle of the rhabdoms of Notodromas monacha and concluded that they displayed a rounded-up angle of 0,17 rad for the lateral cups and 0,35 rad for the ventral one, respectively. We can compare these results to known receptor acceptance angles in other animals, for example, that of deep-sea isopod crustacean Cirolana, which is around 0,78 rad, that of the stomatopod Odontodactylus, 0,017 rad, that of the copepod Labidocera, 0,06 rad, or that of humans, which is on the order of 10–4 rad (Land & Nilsson, 2002). After Nilsson (2009, 2013), simple eyes composed of optical elements, such as mirrors and lenses, may have evolved for serving either behaviours requiring low- or high-resolution vision. It is often assumed that naupliar eyes can only serve as simple photoreceptors. However, many aspects about the functions of naupliar eyes in podocopid ostracods are still unknown. It is thus uncertain in which category to place the podocopid naupliar eyes according to Nilsson’s classification. A certain level of imaging is expected in Notodromas monacha, for which a detection angle below the low-resolution threshold of 0,52 rad has been measured, in favour of a slightly higher resolution (see Andersson & Nilsson, 1981; Nilsson, 2013). Overall, the reported values seem to suggest a rather weak resolution power in Notodromas monacha and, consequently, the possibility for this ostracod to produce a poorly defined image on the retina (Land & Nilsson, 2002). Moreover, given the position of the retina compared to the tapetal cells layer, receptors first receive the unfocused light passing through the lens and then the focused rays collected by the mirror, forming an inevitably fuzzy image (Nilsson, 1990).

Conversely, such value of the receptor light-acceptance angle for Notodromas monacha implies a discretely high degree of light sensitivity. In general, the larger the acceptance angle of receptors and the size of the eye's aperture, the higher the sensitivity of the eyes (Land, 1981; Land & Nilsson, 2002; see Appendix S7).

The F-number is another useful measure of eye sensitivity, related to the inverse of the focal length and aperture of an optical system. The lower the F-number, the greater the performance in light gathering (see Appendix S7). Usually, in animal eyes, although not the rule, a low F-number increases the chance of severe aberrations and thus reduces the image quality. Notodromas monacha has a particularly low F-number of around 0.27 and thus should possess great light sensitivity (Andersson & Nilsson, 1981; Kontrovitz & Myers, 1988), compared to, for example, the eye of the scallop Pecten (Mollusca), which is around 0.6 (Land, 1965).

Nonetheless there is a maximum water depth below which even organisms with an exceptionally high degree of sensitivity cannot compensate for the scarcity of light. Kontrovitz & Myers (1988) proposed an expression for this limiting depth, considering as a model organism marine podocopid ostracods (Lythgoe, 1979; Kontrovitz & Myers, 1988). According to these authors, such expression has never been used to predict the visual limitations of non-marine ostracods. Also, as Kontrovitz & Myers (1988) state, below a measurable threshold, no further improvement in the eyes system can be made in order to gain light gathering ability (Benson, 1975; Lythgoe, 1979; Kontrovitz & Myers, 1988).

The visual system of podocopid ostracods comprises a combination of a mirror and one or two lenses. The fundamental optical element in such a system is the concave mirror formed by the tapetal layer. Its main role is directing the light toward the receptors. The reflecting properties of this optical element arise from the basic principle of refraction and light interference (Land, 1972; see Appendix S8).

For Notodromas monacha, the tapetal cell crystals appear optically anisotropic, particularly sensitive to light polarisation and thus potentially birefringent (Andersson & Nilsson, 1981), although birefringence does not relate to the perception of polarised light in terms of visual clue. The consequences of the birefringence of the tapetal cell crystals on the animal optics are not clear. Biological mirrors similar to the ones here described for ostracods are also observed in some other organisms, such as copepods (see Vaissière, 1961; Fahrenbach, 1964; Land, 1972, 1981).

The concave mirror may not be the only element to exploit internal periodic ultrastructures in order to induce transmission, refraction, and selective reflection of incident light. Carapace ultrastructure has been vastly studied for both, living myodocopid and podocopid ostracods, also in relation to the habitat characteristics (see, e.g., Yamada et al., 2005; Yamada, 2019). However, the ultrastructure of these transparent carapace windows in the vicinity of the eyes has only been described for Macrocypridina castanea (Brady, 1897), a planktonic deep-sea ostracod species, by Parker et al. (2019). The eye tubercles of some living podocopid ostracods have been vastly discussed, but the ultrastructure of the carapace has not been taken into consideration (Kontrovitz & Slack, 1995; Kontrovitz & Puckett, 1998).

Concave mirrors form the image in the direction of the retina and, if spherical-like, present an approximated focal length of half of their radius (see Appendix S9). To form a final image for the visual system, the initial image produced by the lens system works as an intermediate one, an object for the mirror element (Myers & Kontrovitz, 1988). The most straightforward approach to studying podocopid ostracod eyes is to consider the mirror-lens system as composed of thin lenses and simple concave hemispherical thin mirrors (see Fig. 9 for a schematic representation of lens-mirror optics in podocopid naupliar eye; see Appendix S9). In ventral cups, lens cells have a minor role in focusing the light into an image; thus, the convergence of light is affected by the spherical mirror alone. Without any lenses, the object is inevitably positioned in front of the eye, and the image has a distance approximately equal to, or greater than, half of the mirror radius (see Jenkins & White, 1957). Nevertheless, when considering both the thin lens and mirror approximation, the final image is positioned closer to the mirror concavity than in the lensless scenario. Therefore, the role of the lens is almost to compress the optical system, bringing the focal length of the whole lens-mirror apparatus closer to the mirror vertex.

Fig. 9
figure 9

Generalization of the naupliar eye optics based on ray tracing applied to the naupliar eye of the Silurian ostracod Primitiopsis planifrons (Jones, 1887) in Tanaka et al. (2009). The black dot on the lens shows the nodal point, whilst that on the path of the rays represents the focal point of the image when the mirror is not considered. The position of the cuticular lens (CL), lens cells (LC), retinal cells (RC), tapetal cells (TC), and pigment cells (PC) in this schematic representation are approximated. Primitiopsis planifrons lacks lens cells, but we represent a generalized version of the image where the lens cells are present

Full size image

In Myers & Kontrovitz (1988), another approximation to calculate the effective focal length of the lens-mirror system is applied (see Appendix S9). They consider a thick lens and mirror, forming a double lens system (Jenkins & White, 1957). As a result, they obtain an expression for the system focal length as a whole. With this approach, the authors show that the focal length of the system is both lower and upper bounded. Given the relationship between focal length and F-number, Myers & Kontrovitz (1988) then derive a realistic range of F-numbers for the podocopid naupliar eye. As a result, the F-number calculated by this approach is within 0.25–0.50, i.e., values which are surprisingly coherent with that actually measured for ostracods (Andersson & Nilsson, 1981; Myers & Kontrovitz, 1988). Tanaka (2006) made a further step forward by combining the thick-lens tapetum model of Kontrovitz and Myers (1988) with a computationally based investigation of the light-gathering abilities of the various podocopid eye types possessing cuticular lenses. This author then verified the results obtained through this approach with those of behavioural photostatic experiments in different microhabitats, demonstrating that the light-gathering ability of naupliar eyes is closely influenced by the shape of the cuticular lens, namely its thickness and external curvature.

Moreover, Andersson & Nilsson (1981) compared the results obtained by ray tracing of incidence 0 and 25 degrees from the optical axis with those obtained by applying the formula for refraction in spherical surfaces to the ventral and lateral lenses. Thus, the lens system is approximated as two media with different refractive indices, encountering the boundary throughout a spherical surface. An estimation of the focal length was then obtained as a function of the refractive indices and curvature of the surface. Andersson & Nilsson (1981) concluded, by deriving quantitative measurements, that the lens cells in the ventral cup do not contribute to the focus of the image, which is prerogative of the tapetum—the focal distance calculated resulted in at least a hundred times longer than the distance between the lens cell and the mirror. When applied to the lateral lenses, considering both cuticular and lens cells, the focal distance was twice as large as the distance from the tapetum, confirming that the lens system works as a pre-focus system.

Andersson & Nilsson (1981) also hypothesised that podocopid ostracods may possess stereoscopic vision—at least Notodromas monacha. The visual field of this species can be estimated to be around 100 degrees dorsoventrally for the lateral cups, and 150 degrees for the ventral one. Moreover, Notodromas monacha overlaps the visual fields of 35 degrees in the lateral cups, positioned at 25 degrees from the longitudinal axis. Finally, with the same approach, the angular resolution has been approximated at 25 degrees for the lateral cups and 40 for the ventral ones (Andersson & Nilsson, 1981).

Visual stimuli transduction in podocopid ostracods

The absorbance spectrum of a given visual pigment depends on the type of transmembrane opsin protein, its associated subgroup of guanine nucleotide-binding proteins, and the chromophore ligand molecule that composes the receptor (Rayer et al., 1990; Hunt et al., 2014).

It has been suggested that the ancestor of Pancrustacea had four to five groups of photoreceptors inherited from the arthropods' ancestor (Koyanagi et al., 2008; Kashiyama et al., 2009; Henze & Oakley, 2015). The groups of photoreceptors comprised one or two opsin clades sensitive to long wavelengths (LW1-LW2) two in the range of middle wavelengths (MW1-MW2), and one sensitive to short wavelengths (SW) (Henze & Oakley, 2015). The opsin clades then diversified into subclades or got lost with the evolution of the Pancrustacea (Henze et al., 2012; Henze & Oakley, 2015).

Several transcriptional, translational, and physiological studies have investigated colour vision in different representatives of Crustacea, suggesting that various taxa possess from monochromatic to true-colour vision (Goldsmith & Fernandez, 1968; Scott & Mote, 1974; Martin & Mote, 1982; Knight & Leggett, 1985; Hariyama et al., 1993; Smith et al., 1993; Marshall et al., 1996; Horch et al., 2002; Johnson et al., 2002; Dalal et al., 2003; Koyanagi et al., 2008; Kashiyama et al., 2009; Porter et al., 2009; Katti et al., 2010; Rajkumar et al., 2010; Colbourne et al., 2011; Henze et al., 2012; Derby & Thiel, 2014). However, information on the visual capability of ostracods is scarce even if they might represent an example in which extensive opsin diversification occurred within a class. The literature search shows that the only non-marine ostracod genus for which molecular studies on opsin expression are available is Heterocypris. Palecanda et al. (2022) analysed five RNA transcripts of Heterocypris incongruens and discovered that it expressed only long-wavelengths sensitive rhabdomeric opsins (r-opsins), here classified as belonging to the LW1 opsin clade. Moreover, Henze and Oakley (2015) inferred that the transcripts studied of a species in the genus Heterocypris expressed visual r-opsins sensitive to long wavelengths amongst those potentially perceived by crustaceans, belonging to the arthropod LW2 clade. The characteristic of r-opsin-based receptors is their transduction, typical of invertebrates (Rayer et al., 1990; Land & Nilsson, 2002; Nilsson, 2004, 2013; Hunt et al., 2014; see Appendix S10).

The r-type opsins in the arthropod long wavelength-sensitive clade seem to have a broad spectrum of absorbance, which is thought to be green-sensitive and extends to higher lengths (Porter et al., 2007; Hunt et al., 2014; Henze & Oakley, 2015; Yilmaz et al., 2022; Liénard et al., 2022).

The results of these studies concerning transcriptional investigations do not necessarily imply the translation of proteins into functional receptors (Henze & Oakley, 2015). It seems that Heterocypris has lost the capacity to express opsins in the medium-wavelength sensitivity clades (Henze & Oakley, 2015), but it is currently unknown whether this result is extendable to other Podocopa lineages. Furthermore, some species of Myodocopa have lost the LW2 opsin clade, leading to their blindness to long wavelength (Henze & Oakley, 2015; see Appendix S11). Another hypothesis is that Oligostraca lacks the MW1 clade of opsins (peak around 480 nm) and that ostracods, like most crustaceans, lack r-opsins sensitive to short wavelengths (Henze & Oakley, 2015). In the ancestral arthropod, such clade most likely created mainly UV-sensitive visual opsins (Henze & Oakley, 2015). No apparent reason has been provided as to why several crustaceans have lost SWs, but those that do still have them often show lysine residues similar to what should be bovine rhodopsin 90 glycine (Salcedo et al., 2003).

In crustacean groups with short-wavelength sensitivity, sensitivity tends to be absent in individuals without compound eyes. This is probably because SW-UV-sensitive opsins are often expressed in specialised receptors (R8 cells) only present in the ommatidia of some compound eyes (Marshall et al., 1999; Palecanda et al., 2022). Evidence supports the differential expression of opsins in the two visual systems, naupliar and compound eyes, most likely representing recent paralogous duplications (Henze & Oakley, 2015; see Appendix S11). Nonetheless, it must be taken into account that the expression of opsins does not trivially determine the spectral sensitivity of an organism (see Appendix S12).

In addition to visual opsins, there are several non-visual opsin clades whose function still needs to be wholly clarified in vertebrates and invertebrates (Velarde et al., 2005; Eriksson et al., 2013; Hunt et al., 2014; Battelle et al., 2016; Ni et al., 2017). The transcriptome of Heterocypris incongruens studied by Palecanda et al. (2022) presents three non-visual opsins, two RH7 opsins and one neuropsin. In arthropods, non-visual opsins are classified into the R-type Rh7 and arthropsins, the C-type (ciliary opsins) pteropsins, and finally, peropsins and neuropsins (Hunt et al., 2014; Palecanda et al., 2022). The latter is often expressed in podocopid ostracods, specifically in their neuronal tissue (Henze & Oakley, 2015; Brandon et al., 2017; Nagata et al., 2018; Palecanda et al., 2022).

Non-visual opsins are frequently expressed extraocularly (Palecanda et al., 2022). It is hypothesised that both, neuropsins and Rh7 opsins may play a role as photoreceptors maintaining a circadian rhythm, as observed in Drosophila melanogaster Meigen, 1830 (Kistenpfennig, 2012; Henze & Oakley, 2015; Battelle et al., 2016; Cronin & Bok, 2016; Senthilan & Helfrich-Förster, 2016; Ni et al., 2017; Sakai et al., 2017; Liénard et al., 2022; Palecanda et al., 2022).

Both, Rh7 opsins and neuropsins are considered bistable opsins that create particularly sensitive UV-SW photopigments and, after being excited, can return to their initial resting state (Koyanagi & Terakita, 2008; Yamashita et al., 2010; Kojima et al., 2011; Matsuyama et al., 2012; Koyanagi et al., 2013; Hunt et al., 2014; Cronin & Bok, 2016; Sakai et al., 2017; Liénard et al., 2022; Palecanda et al., 2022).

Podocopid ostracod brain

When the light stimulus causes excitation of receptors, signals reach the central nervous system, which in the Podocopa comprises a circum-oesophageal collar of fused ganglia, the latter subdivided into protocerebrum, deuterocerebrum and triterocerebrum, and a ventral chain of ganglia (Rome, 1947; Hartmann, 1966). For a more detailed description of the central nervous systems and nerve cells in non-marine ostracods, see e.g. Claus (1891, 1895), Turner (1896), Lüders (1909), Hanström (1928, 1931), Hartmann (1967). The protocerebrum is the cerebrum part that elaborates the visual stimuli through the optic nerves (Dahl, 1965; Elofsson, 1992). The axons reach from the sensory cells to a single glomerular neuropil slightly separated from the protocerebrum (Elofsson, 1966, 1992, 2006). The position of the protocerebrum is always perpendicular to the axis spanning the length of the ostracod dorsally (Andersson & Nilsson, 1981; Huvard, 1990). Overall, if the eyes are somewhat non-efficient or degraded, a specific neuronal compensation may occur once the information is processed in the brain (Andersson & Nilsson, 1981; Elofsson, 2006). Hanström (1928, 1931) hypothesised that the protocerebrum and deutocerebrum had developed from the same primordial brain, in analogy with what observed in polychaetes—see Hartmann (1967) for a discussion on this subject.

Weygoldt (1961) showed that five large cells are present in the protocerebrum of Cyprideis torosa (Jones, 1850) using chrome-hematoxylin to stain the cells. By studying the anatomy of Heterocypris incongruens, Aramant & Elofsson (1976) proposed that the brain of ostracods may use monoamine neurotransmitters to transmit and elaborate information. Specifically, Aramant & Elofsson (1976) used a fluorescence histochemical technique to create fluorophores from biogenic monoamines present in the organism and artificial chemical compounds (Falck & Owman, 1965). They found a fluorophore in the protocerebrum of Heterocypris incongruens. The fluorophore emitted weakly in the naupliar eyes and strongly in the neuropile. At the same time, complex fluorescent patterns were observed in the deutocerebrum and tritocerebrum. A second fluorophore was conspicuously present, scattered amongst the deutocerebrum and other brain areas. Fluorescence was also present along the ventral nerve cord, lip nerve and fibres of antennae. The presence of neurotransmitters of the family of indolamines or catecholamines has been hypothesised (Elofsson et al., 1968; Welsh, 1972; Aramant & Elofsson, 1976). Aramant & Elofsson (1976) suggested that the abundance of deutocerebral monoaminergic neurons results from the preferential use of the olfactory sense over the visual one. However, it must be taken into account that this hypothesis emerges from the study of only one species, namely Heterocypris incongruens. For this reason, future research should be extended to different species to draw general conclusions on neuronal processing in non-marine ostracods.

Photoreception and learning in non-marine ostracods

In podocopid ostracods, it is reasonable to hypothesise recognition mechanisms involving sensory capacities to optimise the energy required for mate searching and reduce the reproductive costs, the latter being exceptionally high in amphigonic non-marine species, given the production of giant spermatozoa (Matzke-Karasz, 2005; López‐Fernández et al., 2007; Schmit et al., 2013). However, in amphigonic species, copulation of males with other males, non-conspecific females or even with grains of sand has also been observed (Horne & Martens, 2000, pers. observation). Nonetheless, Eucypris virens (Jurine, 1820) recognise sexual partners precisely in mixed populations, where parthenogenetic and sexual females coexist (Horne et al., 1998; Schmit et al., 2013). Overall, several peculiar behaviours in approaching partners are well documented in non-marine ostracods (Danielopol, 1980; Danielopol & Horne, 1996; Smith & Kamiya, 2007; Mesquita-Joanes et al., 2012). Although mate recognition is most likely achieved by direct contact, the exact mechanism is unknown and may vary from species to species leaving some space for hypotheses (Danielopol, 1980; Horne et al., 1998; Martens, 2000; Schmit et al., 2013). Auditive signals per se have been discarded in Ostracoda (Martens, 2000). Other possible mate recognition mechanisms involve chemical clues, such as the release of pheromones (Horne et al., 1998). In the case of Eucypris virens and other freshwater species, communication via chemical stimuli is unlikely since the populations usually have high densities of individuals (Schmit et al., 2013); consequently, under these conditions, dissolved chemical signals might be efficient only in certain cases, for example as an alarm in risk situations, but hardly as a means of recognising potential mates. Possible visual mate recognition in Podocopa, contrarily to Myodocopa, is often overlooked since their visual abilities, especially of non-marine ostracods, are thought to be primitive (McGregor & Kesling, 1969; Cohen & Morin, 1990; Horne et al., 1998; Horne & Martens, 2000; Zhang et al., 2024), and so there is no definitive evidence against or in favour of the role of vision in mate recognition in non-marine ostracods. Similarly, there are no known generalisable differences in perception between taxa depending on the environment they inhabit (Oakley & Huber, 2004). For example, species in the genus Notodromas not only display the already mentioned sexual dimorphism in the visual apparatus, but they also have distinct morphological features between sexes, as larger or more pigmented carapaces in females (De Deckker, 1979; Horne et al., 1998). This genus of non-marine ostracods is currently the only one for which optical mate recognition has been hypothesised (Horne et al., 1998).

Other examples of evident sexual dimorphism are present in several limnocytherids and candonids, and generally are a common feature of sexually reproducing organisms (Meisch, 2000; Smith et al., 2015). Nonetheless, sexual dimorphism can originate for different reasons than mate recognition, such as sexes differences in daily activity and diet (Speiser et al., 2013; Arenz et al., 2018) or brood care (Smith et al., 2015). For example, despite sexual dimorphism, epigean candonids are considered nearly blind, with very reduced, red-pigmented eyes, thus making it unlikely that morphological differences between genders play a role in partner recognition (Turner, 1896).

Non-marine ostracods, related to their lifestyle and habitat, show remarkable differences in swimming ability and motility (Hartmann, 1966; Mbahinzireki et al., 1991), as well as in dietary alternatives, being generalists, either omnivorous, detritivorous or scavengers (Liperovskaya, 1948; Margalef, 1953; De Deckker, 1983; Stepien & Brusca, 1985; Fryer, 1997; Vannier et al., 1998; Meisch, 2000; Danielopol et al., 2002; Schmit et al., 2007; Smith & Delorme, 2010; Ottonello & Romano, 2011; Mesquita-Joanes et al., 2012). Coloured carapaces are typical of non-marine ostracods inhabiting littoral and sublittoral environments, whilst interstitial species display depigmented carapaces and reduced photoreceptive systems. Cypridopsis vidua (O.F. Müller, 1776), whose valves have a “zebra” pattern with dark brown bands, uses the macroalga Chara as a hiding place from predators to survive by quickly reacting to chemical cues indicating the predators' presence (Mbahinzireki et al., 1991). Integrating chemical, photoreceptive and other senses stimuli in order to detect predators and mates is considered quite a rule rather than an exception amongst animals (Acquistapace et al., 2005). Therefore, we cannot exclude that similar integration may also occur in non-marine ostracods with light-gathering apparatuses. Non-marine ostracods seem to rely on olfactory cues when harvesting food (Roca et al., 1993; Uiblein et al., 1994, 1996; Vannier et al., 1998; Mesquita-Joanes et al., 2012; Vandekerkhove et al., 2012; Miličić et al., 2015). Moreover, they show a preference for some food sources compared to others (Liperovskaya, 1948; Schmit et al., 2007). Throughout behavioural experiments, it has been observed that both Eucypris virens and Heterocypris incongruens feed preferentially on spinach (Schmit et al., 2007; Miličić et al., 2015). Cypridopsis vidua preferentially searches for Chara sp., whilst Notodromas monacha feeds from neustonic protists (Benzie, 1989; Roca et al., 1993; Uiblein et al., 1994, 1996; Kiss, 2004; Schmit et al., 2007; Karanovic, 2012). Overall, it is believed that in ostracods, vision does not participate in the feeding process (Land & Nilsson, 1990; Vannier et al., 1998; Miličić et al., 2015). Nonetheless, food preferences can also be determined by innate colour preferences. Colour preferences are generated when the visual signal immediately leads to motor control. So far, no studies have been conducted on these aspects for ostracods.

Non-marine ostracods seem to be capable of reacting to their preferential light components, as shown through phototaxis experiments. In Barcelo & Calkins (1979), a species of the freshwater genus Cypris, within three possible chambers, i.e. dark (non-illuminated), illuminated, and irradiated only with UV-B, show a statistically relevant preference towards the UV-B one, swimming horizontally towards it. Nonetheless, in a second experiment, no preferential vertical swimming heading to the most illuminated region of the column is observed for any of the irradiation types. Moreover, in both experiments, the survival rate of Cypris, irradiated by the UV-B light, is exceptionally high compared to the other taxa. The preferential choice of the UV-B chamber is surprising since ostracods are thought to lack the opsins that can make them sense UV (Henze & Oakley, 2015). However, the choice may not be related to colour vision. On the contrary, the relatively high tolerance of Cypris for UV can be explained by the characteristics of ostracod carapaces, which seem to protect the animal by absorbing up to 80% of the UV radiation and avoiding cellular damages – except for Heterocypris incongruens whose valves have a lower absorption rate (Van Den Broecke et al., 2012). The high tolerance for UV may be connected to the specific lifestyle, reproductive mode and carapace coloration of the different ostracod species (Van Den Broecke et al., 2012). A higher UV-tolerance is to be expected in species living in very shallow water bodies with high degrees of UV-light penetration. In water bodies of greater depth, active swimmers dwelling in the water column will have to be more UV-tolerant than fully benthic species.

The role of photoreception and vision in associative learning processes, confirmed by experiments in vertebrates, crustaceans and molluscs, is probably also important in other groups of animals characterised by simple neuronal systems (Bielecki et al., 2023). The degree of colour vision, light perception, and image formation an organism displays can be related to cognitive capabilities, associations, and learning (Marshall et al., 2015; Kelber, 2016; Kelber et al., 2003; Kelber & Osorio, 2010). Intuitively, colour learning and preference may have represented significant evolutionary assets (Kelber & Jacobs, 2016).

Previous studies on learning abilities of non-marine ostracods did not consider visual capabilities. An example of the latter behaviour is Cypridopsis vidua, which can undergo a decision-making process to balance harvesting and predation avoidance (Roca et al., 1993). Moreover, behavioural experiments on non-marine ostracods are generally very few, and are conducted on a limited number of species, such as Darwinula stevensoni (Brady & Robertson, 1870), Eucypris virens and Heterocypris incongruens (Kohler & McPeek, 1989; Roca et al., 1993; Rossi & Menozzi, 1993; Baltanás et al., 2000; Gandolfi et al., 2001; Van Doninck et al., 2003; Schmit et al., 2007; Miličić et al., 2015; Bieszke et al., 2020).

The first behavioural study concerning light stimuli perceptions in non-marine ostracods probably is that of Applewhite & Morowitz (1966). In the experiments, both sexes of Cyclocypris forbesi Sharpe, 1897 preferred dark environments and learned how to avoid light by solving a light-guided maze. Additionally, Cyclocypris forbesi displayed associative capabilities of classical pseudo-conditioning by closing its valves when illuminated by a light source passing through a blue ground glass filter associated with later electrical shocks. This type of learning has also been confirmed for other invertebrates (Perry et al., 2013; Hawkins & Byrne, 2015; Nargeot & Bedecarrats, 2017). Fischer & Laforsch (2018) showed that both Heterocypris incongruens and Notodromas monacha, when grown in microgravity, cannot adapt to use light sources to orientate in the new condition. Finally, the only ethological experiments combining colour perception and learning in non-marine ostracods are those conducted by Romano et al. (2022). They used miniaturised analytical systems to assess whether Heterocypris incongruens can display associative operant conditioning. Miniaturised devices had previously been employed to assess the phototactic abilities of other crustaceans, but never in terms of conditioning (Cartlidge et al., 2016). In the experiments, Heterocypris incongruens was first trained to associate a light colour, either green or yellow, with a source of food or stress (increased salinity). One individual at a time was placed in a miniaturised micro-arena and was forced to choose between two possible chambers differently organized in three different treatments. In each of the three treatments, Heterocypris incongruens could either enter in, respectively, i) a chamber with the light of the colour with which it was trained (correct colour) or in a dark chamber, ii) a chamber with a different light colour from that with which it was trained or in a dark chamber, iii) a chamber with a light of the correct colour or in a chamber with a light of different colour. Romano et al. (2022) showed that when trained for rewards, Heterocypris incongruens significantly chose the correct light when possible and otherwise chose the light of different colour rather than the dark chamber. Conversely, when trained to avoid stressors, Heterocypris incongruens chose the chamber with no light and did not display preferences when both lights are on. Since the light colours used in the experiments are always green or yellow, Heterocypris incongruens can distinguish at least one of these colours. However, ostracods have never been tested for object recognition and discrete and continuous quantity perception, especially in terms of learning.

Conclusive remarks and open questions

The morphology and ultrastructure of naupliar eyes and cerebra in ostracods is well documented. Nonetheless, the references mainly focus on marine podocopid species, and only limited information exists on non-marine ones. Most studies on non-marine podocopid naupliar eyes refer to species in the superfamily Cypridoidea, such as, for example, species in the genera Heterocypris, Cypridopsis, Herpetocypris, Eucypris, Cypris, Cypria, Cyclocypris and Notodromas. Regarding the superfamily Cytheroidea, there are some references to non-marine species in the genus Xestoleberis. Otherwise, works on the naupliar eyes in the superfamily Cytheroidea consider marine representatives of the families Leptocytheridae, Hemicytheridae, Cytheruridae and Loxoconchidae, which, in general, also contain non-marine species. As far as we know, no records on the morphology of naupliar eyes exist for the family Darwinulidae, the only living representatives of the superfamily Darwinuloidea.

Cuticular lenses in living non-marine podocopid ostracods have been described by in-depth morphological investigations only for genera in the family Notodromatidae. However, such structures are present in other ostracod taxa. For example, cuticular lenses are observed in different families of Cytheroidea (i.e., hemicytherids, loxoconchids, cytherids, cytherurids, xestoleberidids, paradoxostomatids, hemicytherids, leptocytherids, limnocytherids and cytherideids), containing mostly marine species but with several non-marine representatives as well. In podocopid ostracods with fused eyes and without cuticular lenses, a transparent region of the carapace above the eyes is often observed.

Podocopid naupliar eyes can be approximated to a mirror-lens visual system. The reflective properties of the concave biological mirror are derived from the basic principle of refraction and interference of light. Imperfections of optical systems lead to aberrations, whilst image blurring due to light diffraction is always present. Optical models describing naupliar eyes of non-marine ostracods and their efficiency often refer to Notodromas. In this genus, the role of aberrations and diffraction on image formation is known, and sensitivity and resolution of the eye have been quantified by calculating the F-number, i.e., image brightness index and the light ray’s acceptance angle of receptors. Podocopid ostracods seem to be particularly adapted for dim habitats, having a particularly low range of F-number, estimated theoretically. However, as already mentioned, careful investigations combining theoretical, computational, and experimental approaches show that, when the cuticular lens is present, its shape strongly influences the light-gathering ability of the podocopid naupliar eye. Moreover, despite the seemingly poor resolution power, Notodromas monacha can produce a modest image on the retina. In other podocopid species, the arrangement and structure of receptors are often ignored, although they are fundamental units in defining the resolution of an eye. Rhabdom shields or non-cylindrical rhabdoms can limit the degree to which light escapes from the receptor, which would otherwise lead to a loss of resolution. Both structures may be present in podocopid naupliar eyes, but they are rarely mentioned in relation to their role. Overall, simple eyes using mirror optics like those described for non-marine podocopid ostracods may serve behaviours requiring a moderately low resolution. A more comprehensive knowledge of which cells and lens types are present or absent in different non-marine species would lead to a better understanding of the degree of visual capacity to be expected, also depending on environmental and phylogenetic constraints.

Another area of research where only a small amount of information is available on non-marine podocopid ostracods is that of opsin expression and, therefore, the possibility of colour vision. In fact, the only evidence of opsin expression concerns Heterocypris incongruens, which expresses, amongst the known crustacean opsins, the r-opsins in the clade sensitive to long wavelengths in the visible spectrum. Generally, it is hypothesised that most crustaceans lack r-opsins sensitive to short wavelengths. Moreover, Heterocypris incongruens also possesses nonvisual opsins, whose roles are still debated. The expression of opsins must be considered of primary importance as much as the anatomy of the eyes, as knowledge of both is indispensable for obtaining well-founded information on the significance and use of vision in animal life. The expression of genes coding for multiple visual opsins is not merely linked to colour vision. Their actual transcription and translation are not guaranteed, and a particular distribution of opsin proteins between receptors may have purposes other than colour vision. It would certainly represent a major milestone to map the expression of opsins in species of non-marine ostracod from different phylogenetic lineages, and in relation to their habitats and behavioural adaptations. Opsin expression could be related to specific light perception requirements, intrinsically connected to phenotypic traits. For example, species in the genus Notodromas show not only sexual dimorphisms in the cuticular lenses of their visual apparatuses, but also distinct morphological differences in the carapace between sexes; however, no records of opsin expression in this genus exists. Vision and the ability to perceive colours can be tested through ethological experiments, which would intuitively complement, along with anatomical and molecular analyses, the study of the visual abilities of an organism through a multidisciplinary approach. For example, it has been shown that some non-marine species in the genus Cypris can react to wavelength in the UV range by positive phototaxis, which is unexpected since non-marine ostracods seem not to perceive radiation within this spectral range. Unfortunately, from the results of this review, there is no study in the literature on non-marine ostracod species that has combined structural and molecular aspects with behavioural descriptions.

Finally, non-marine ostracods show learning abilities, exhibiting associative learning of the classical type and operant conditioning when trained with light stimuli. The processes of learning and conditioning in non-marine ostracods have so far been described in two species belonging to the superfamily Cypridoidea, in two articles published fifty years apart. The first study showed that Cyclocypris forbesi learned to close its valves when illuminated by a light source associated with successive electric shocks. The second study demonstrated that Heterocypris incongruens, when trained to associate a light colour with a source of food or stress, chose the light that led towards the food or away from the stress. At present, it is impossible to determine whether these capabilities are present in other species of the same superfamily or in other non-marine podocopid superfamilies. Intuitively, the ability to learn mediated by visual stimuli is an outstanding adaptive advantage for survival and enhanced resource utilisation. Since some podocopid ostracods possess eyes, which can be described by simple optical models, possibly express different opsins, and learn to associate light stimuli to behaviours and even choose accordingly, it is surprising how little research has been devoted to investigating the role of these characteristics as a whole from an evolutionary and adaptive point of view. For these reasons, we do hope that future work will shed light on the importance of photoreception in non-marine ostracods and its potential benefits in the exploration of their habitat (e.g., search for food, use of shelters, detection of predators, recognition of mates, etc.) and for learning processes.