The Kingdom of the Blind: Disentangling Fundamental Drivers in the Evolution of Eye Loss

The Kingdom of the Blind: Disentangling Fundamental Drivers in the Evolution of Eye Loss Abstract Light is a fundamentally important biological cue used by almost every animal on earth, to maintain daily rhythms, navigate, forage, find mates, or avoid predators. But an enormous number of species live in darkness: in subterranean caves, deep oceans, underground burrows, and within parasitic host bodies, and the loss of eyes appears consistently across these ecosystems. However, the evolutionary mechanisms that lead to the reduction of the visual system remain the subject of great interest and debate more than 150 years after Darwin tackled the issue. Studies of model taxa have discovered significant roles for natural selection, neutral evolution, and pleiotropy, but the interplay between them remains unclear. To nail down unifying concepts surrounding the evolution of eye loss, we must embrace the enormous range of affected animals and habitats. The fine developmental details of model systems such as the Mexican cave tetra Astyanax mexicanus have transformed and enriched the field, but these should be complemented by wider studies to identify truly overarching patterns that apply throughout animals. Here, the major evolutionary drivers are placed within a conceptual cost–benefit framework that incorporates the fundamental constraints and forces that influence evolution in the dark. Major physiological, ecological, and environmental factors are considered within the context of this framework, which appears faithful to observed patterns in deep-sea and cavernicolous animals. To test evolutionary hypotheses, a comparative phylogenetic approach is recommended, with the goal of studying large groups exhibiting repeated reduction, and then comparing these across habitats, taxa, and lifestyles. Currently, developmental and physiological methods cannot feasibly be used on such large scales, but penetrative imaging techniques could provide detailed morphological data non-invasively and economically for large numbers of species. Comprehensive structural datasets can then be contextualized phylogenetically to examine recurrent trends and associations, and to reconstruct character histories through multiple independent transitions into darkness. By assessing these evolutionary trajectories within an energetic cost–benefit framework, the relationships between fundamental influences can be inferred and compared across different biological and physical parameters. However, substantial numbers of biological and environmental factors affect the evolutionary trajectory of loss, and it is critical that researchers make fair and reasonable comparisons between objectively similar groups. Introduction Biologists have long been fascinated with the pale, eerie, creatures that inhabit the world of darkness. Animals that live in true darkness, rather than the dim light associated with nocturnality and the mesopelagic zone, are found in a multitude of different habitats, including the bathypelagic (ocean depths below around 1000 m), subterranean, and submarine caves, fossorial burrows and, to some extent, endoparasitic host bodies. Many dark-living animals share striking morphological features, including albinism, increased investment in chemo- and mechanoreceptors, and the reduction or loss of eyes (Poulson 2001). Eye loss in particular has captured attention for more than a century, and has been reported across Metazoa from burrowing amphibians (Fig. 1; Mohun et al. 2010; Mohun and Wilkinson 2015) to deep-sea trilobites (Clarkson et al. 2006; Lerosey-Aubril 2006). Fig. 1 View largeDownload slide Eye loss affects a wide variety of animals living in many different dark habitats. Clockwise from top left, Bathymophila diadema, a marine gastropod found at depths of 650–1000 m; Microcaecilia dermatophaga, a burrowing caecilian (image courtesy of Mark Wilkinson); Metagonia jamaica, a cavernicolous spider (specimen SMNH 1408109); Astyanax mexicanus, cave population (image courtesy of Bill Jeffery). Fig. 1 View largeDownload slide Eye loss affects a wide variety of animals living in many different dark habitats. Clockwise from top left, Bathymophila diadema, a marine gastropod found at depths of 650–1000 m; Microcaecilia dermatophaga, a burrowing caecilian (image courtesy of Mark Wilkinson); Metagonia jamaica, a cavernicolous spider (specimen SMNH 1408109); Astyanax mexicanus, cave population (image courtesy of Bill Jeffery). We are now able to identify incremental molecular and developmental changes occurring in dark-living species, which have started to demystify this recurrent evolutionary pattern (e.g., Jeffery 2009; Stahl et al. 2015; Tierney et al. 2015). However, the overall evolutionary mechanisms behind eye loss remain controversial, more than 150 years after it was first explored by Darwin (1859; Rétaux and Casane 2013). Eye loss appears to be driven at least in part by direct selective pressure, but both pleiotropy and genetic drift are also key influencers (Rétaux and Casane 2013). Despite substantial recent progress, thanks to a combination of developmental and molecular techniques, it is not yet clear how these drivers interact, or, crucially, whether their relationships are similar across taxa and habitat types. Unifying concepts have been identified in the constructive evolution of eyes that apply across all eye types and animal groups (Nilsson 2013), but their regressive evolution remains a tangled matter. Here, a conceptual framework is proposed that could inform future studies of eye loss effectively across large and distant taxonomic groups, and several recommendations are made for such research directions. This paper will not comprehensively review the substantial body of existing work on individual taxa, particularly the Mexican cave tetra Astyanax mexicanus by Jeffery, Wilkens, and colleagues (e.g., Wilkens and Strecker 2003; Protas et al. 2007; Jeffery 2009; Borowsky 2013). This species is an exceptional model for studying the developmental mechanisms behind eyelessness in numerous instances across a cave network, and the resulting work has transformed our understanding of eye loss (see Jeffery [2009]; Wilkens and Strecker [2017] for reviews; and “The state of play” below). However, the current focus is on unifying concepts, cross-habitat comparisons, and non-model groups, and not individual case studies. The majority of eyeless species have not yet been kept or bred in captivity (and some have never been seen live), for reasons of accessibility, availability, or fragility; this paper mainly addresses how we can learn from these taxa in the absence of manipulative methods. The overarching aims are to explore the possibility of a generalized model for eye reduction, incorporating macroevolutionary trends that apply across taxonomic and ecological divides; and to promote a large-scale comparative morphological approach. This does not supplant the exceptionally important ongoing developmental, physiological, and transcriptomic work, but is intended to complement that work, and provide a taxonomic breadth and macroscopic evolutionary perspective that may be challenging to achieve using such methods. If evolutionary biologists aim to identify unifying concepts in the reductive evolution of eyes (or any other system), we must embrace the enormous range of affected animals and habitats (Fig. 1), and all the complications this entails. Broader taxonomic approaches have proved successful in other cases of repeated constructive evolution, including vision, which has been elegantly rationalized using trends that apparently apply to all animal groups (Nilsson 2013). Features including screening pigment, membrane folding, and focussing optics are causally linked to the increasing sophistication of visual tasks, and to the minimum light intensity mathematically required for these tasks. Such relationships are likely to be equally important in reductive evolution, as they affect the functionality and therefore the fitness value of visual structures in different light environments. By adopting this functional evolutionary approach to loss, we may more easily rationalize what we observe in nature. For example, high-resolution vision requires minimum light intensities roughly equivalent to those at ocean depths of 350–650 m (Nilsson 2013). Unsurprisingly, this is the point beyond which dramatic alterations to eye structure, including loss (Syme and Oakley 2012; Sumner-Rooney et al. 2016; Gonzalez et al. 2018), begin to appear in multiple marine groups (Warrant et al. 2003; Warrant and Locket 2004; Johnsen et al. 2012). There must be fundamental relationships between the physiological, ecological, and environmental characteristics of study systems and their evolutionary trajectory under darkness. However, unlike the constructive evolution of vision, a wealth of information about structure, function, development, and ecology is largely lacking for reduced-eyed animals, and some of this information is more readily obtained than others. The state of play: a (very) brief summary of eye loss research The consistent loss of eyes in dark-living animals has intrigued evolutionary biologists for centuries. In On the Origin of Species, Darwin wrestled with a selective basis for eye loss, uncertain as to what disadvantage eyes might confer. He eventually attributed it to “disuse” of eyes in the dark, in what we would now consider to be neutral evolution (Darwin 1859). Neutral evolution theory in its modern form developed from the 1930s and 1940s onwards, with its importance in regressive evolution being acknowledged decades before Kimura’s (1968) seminal paper. Although this view was not universally held, it dominated for another century (Culver and Pipan 2009). Following the new synthesis, and the accelerating discovery of more eyeless species, the evolution of loss became a subject of renewed interest in the second half of the 20th century. A new experimental approach, using cave fishes and crustaceans, produced evidence of convergent patterns in repeated cave invasions, and that natural selection favors the reduction and loss of eyes in energy-limited aphotic environments (Christiansen 1961; Poulson 1963; Sadoglu 1967; Barr 1968; Jones and Culver 1989). However, the dominant role of neutral evolution in eye loss was still championed by Wilkens and colleagues (e.g., Wilkens 1988; Wilkens and Strecker 2003, 2017: Chapter 7), who highlight the variation in regressive phenotypes and genotypes, pre-adaptation, and potential misconception of energy limitation in caves. The past two decades have seen enormous growth in molecular, developmental, and physiological studies, most notably using Astyanax (see Jeffery [2009]; Wilkens and Strecker [2017] for reviews). These have generated substantial evidence to demonstrate that direct selection favors energetic economy via eye reduction (Culver et al. 1995; Niven and Laughlin 2008; Yoshizawa et al. 2012; Moran et al. 2015; contraWilkens and Strecker 2017). Several also support a crucial role for pleiotropy in eye development through molecular players such as Shh (Yamamoto et al. 2009; Carlini et al. 2013). In the last 10 years, the capabilities of genetics and transcriptomics have improved dramatically, and transcriptome sequencing especially has been applied to a range of eyeless taxa (Friedrich et al. 2011; Carlini et al. 2013; Meng et al. 2013; Stahl et al. 2015; Tierney et al. 2015; Kim et al. 2017). Many of these found persistent expression of opsins and clock genes in (mainly cavernicolous) dark-living animals (Friedrich et al. 2011), often subject to downregulation or functional shifts (Gross et al. 2013; Hinaux et al. 2013; Tierney et al. 2015; Wong et al. 2015; Carlini and Fong 2017). It must also be noted that there are increasing examples of other roles for opsins beside photoreception (Leung and Montell 2017). Knowledge gaps Two key patterns have emerged that are potentially problematic to drawing broader conclusions about the reductive evolution of visual systems. Firstly, cave fauna has formed the focus of the majority of studies to date. As a crude measure, searches for “eyeless” or “blind” species on Zoobank return 65% cavernicolous, 24% fossorial, 5.8% ground-living (leaf litter), 4.8% marine and <1% parasitic taxa (as of May 15, 2018). In studies specifically addressing the evolution of eye loss, the skew is far greater, with 78% focusing on cavernicolous species, 8% on fossorial species, 7% on marine species, and 7% nocturnal. Secondly, certain taxonomic groups have received more attention than others. Fishes account for 51% of ZooBank search results returned for “eyeless” or “blind” species, yet there are fewer than 200 described teleosts of a predicted 21,000 global troglobionts (Juan et al. 2010; Soares and Niemiller 2013). Arthropods are, more understandably, the next largest group (36%), followed by snakes (12%), and others (including mammals, amphibians, and molluscs; <1%). Again, this trend is reflected in studies specifically examining eye loss, where target systems were 57% fish, 26% arthropods, 8% mammals, 5% amphibians, 3% molluscs, and <1% reptiles. Specialization within one or two of these ecosystems or taxa has created artificial divisions between researchers working on cavernicolous, deep-sea, fossorial, and parasitic species, and opportunities for synthesizing information have largely been underexploited, aside from a handful of research groups. Of course, there are fundamental differences between habitat types and the transitions undergone by different lineages. An integrative approach exploiting these differences will facilitate more objective and informative comparisons between them. For instance, cave taxa may be the subject of greater research interest because eye loss appears to occur more consistently than in deep-sea taxa, but this also represents an opportunity to investigate why they differ. What is “loss”? As an added obstacle, many studies and species descriptions dedicate just a handful of words to eye structure, even when investigating trait evolution (Raupach et al. 2009; Williams et al. 2013; Gonzalez et al. 2018). Absence of retinal pigmentation alone may be taken to signify eye loss, with the implication that loss is total, and pigmented (therefore superficially visible) eyes are presumed to be unaffected. Absolutist attitudes toward loss are deeply inhibitory to fully understanding the complex evolutionary processes behind it. Many taxa have been found to have non-pigmented eyes, vestigial or intact, and eyes with retained pigmentation may be reduced in size, sunken beneath skin, affected by retinal disruption or lens degradation, or impacted by altered gene sequences or expression (Jeffery 2009; Bloom et al. 2014; Sumner-Rooney et al. 2016). These are all vitally important characters that could shed further light on the evolution of loss but are going unrecorded. Similarly, degradation of structure does not necessarily equate to complete loss of function. While pupil obstruction, lens degradation and loss of pigmentation may impede spatial resolution, there is substantial evidence for persistent photoreceptor function and light-responsive behavior in species with dramatically reduced eyes (Friedrich et al. 2011; Fišer et al. 2016; Langille et al. 2018). This may include photoreceptor activity within the retina (David-Gray et al. 1998; Espinasa and Jeffery 2006; Zubidat et al. 2011) or elsewhere (Yoshizawa and Jeffery 2008), for phototaxis or circadian entrainment. Moving forwards: the future of the field The unifying questions in the evolution of eye loss remain; what are the evolutionary forces at play, what are their relative roles; and how do animals move toward eye loss? (See Culver and Pipan [2009, 109–30]; and Rétaux and Casane [2013] for reviews) As outlined above, significant progress has been made to address these questions in various taxa, but there are still substantial gaps in both our knowledge and, to an extent, within the research community working in the field of eye reduction and loss. To provide a unifying framework, we must identify fundamental factors that influence the trajectory toward eye loss across all habitat types and taxa. Here, such a framework is outlined for the first time. Several specific considerations for future work are outlined; many have been proposed by other authors, but none, as far as the author is aware, have been addressed directly in empirical studies. In some cases, it may be possible to incorporate these considerations into experimental designs, but otherwise they should be accounted for in interpretation of results and the drawing of wider conclusions. Fundamental drivers in the evolution of eye loss—a first attempt at conceptualization Eyes and corresponding brain activity are energetically expensive, which is fundamental to their loss; once their utility decreases, this energy could be reinvested elsewhere with greater results, and their high cost amplifies the potential benefits (Niven and Laughlin 2008; Moran et al. 2015). As with any costly trait, selection should favor the maximization of fitness benefits against these energetic costs, consistently drawing animals toward a cost–benefit equilibrium (Fig. 2, in green). The further an animal from equilibrium, the stronger the selective pressure, so we might imagine a fairly simple vector gradient of selective pressure toward equilibrium (Fig. 2, background gray arrows). Fig. 2 View largeDownload slide Schematic cost–benefit relationships for visual systems in changing photic environments, and the evolutionary drivers involved in loss. (A) Selection. Animal A inhabits a bright, visually complex and stable environment, and has evolved a visual system well-suited to its habitat and lifestyle. Overall energetic costs and fitness benefits from the visual system are at (green line) or near (pale green zone) equilibrium. If A moves to a darker habitat (A1) the return on these costs decreases (Δb). Natural selection (background, gray arrows) draws A1 back toward equilibrium, favoring the shortest route (black arrow) and with greater strength at greater distance from equilibrium. The fitness area beneath the line of equilibrium (in yellow) is inaccessible due to physical constraint; the eye cannot optically obtain more information from a stable environment without increased investment in size, complexity, innervation, etc. However, A1 can fully or partially recover its cost deficit (Δc) by reinvesting energy elsewhere (r) to maximize potential fitness gains (g) in other ways, such as an alternative sensory mode. Species with simpler visual systems and lower initial investment (B) undergo smaller displacements (B1) and are likely subjected to weaker selective pressure. (B) Drift and pleiotropy. Evolutionary trajectory is simultaneously affected by two additional drivers: drift (departure from the direction of selection) and pleiotropic effects (orange obstacles). Selection is strongest after greater displacement (1) and decreases as equilibrium is approached (2). The relative importance of drift, and therefore trajectory variability, increases with proximity to equilibrium (inset). Pleiotropic effects on genes involved in vision may obstruct certain evolutionary routes and cause additional diversions (p1 and p2). (C) Constraints. Owing to optical constraints, there is a threshold light intensity value below which vision is no longer physically viable (purple). Beyond this threshold (A2, A3), the selective pressure landscape changes, with the animal no longer being able to recoup benefits from the visual system and with no alternative but to economize by reducing it. A second potential threshold is a lower limit to cost (orange), with pleiotropic effects preventing extreme reduction if it would cause collateral damage elsewhere. This may prevent optimization and leave the animal with a deleterious vestigial structure. Fig. 2 View largeDownload slide Schematic cost–benefit relationships for visual systems in changing photic environments, and the evolutionary drivers involved in loss. (A) Selection. Animal A inhabits a bright, visually complex and stable environment, and has evolved a visual system well-suited to its habitat and lifestyle. Overall energetic costs and fitness benefits from the visual system are at (green line) or near (pale green zone) equilibrium. If A moves to a darker habitat (A1) the return on these costs decreases (Δb). Natural selection (background, gray arrows) draws A1 back toward equilibrium, favoring the shortest route (black arrow) and with greater strength at greater distance from equilibrium. The fitness area beneath the line of equilibrium (in yellow) is inaccessible due to physical constraint; the eye cannot optically obtain more information from a stable environment without increased investment in size, complexity, innervation, etc. However, A1 can fully or partially recover its cost deficit (Δc) by reinvesting energy elsewhere (r) to maximize potential fitness gains (g) in other ways, such as an alternative sensory mode. Species with simpler visual systems and lower initial investment (B) undergo smaller displacements (B1) and are likely subjected to weaker selective pressure. (B) Drift and pleiotropy. Evolutionary trajectory is simultaneously affected by two additional drivers: drift (departure from the direction of selection) and pleiotropic effects (orange obstacles). Selection is strongest after greater displacement (1) and decreases as equilibrium is approached (2). The relative importance of drift, and therefore trajectory variability, increases with proximity to equilibrium (inset). Pleiotropic effects on genes involved in vision may obstruct certain evolutionary routes and cause additional diversions (p1 and p2). (C) Constraints. Owing to optical constraints, there is a threshold light intensity value below which vision is no longer physically viable (purple). Beyond this threshold (A2, A3), the selective pressure landscape changes, with the animal no longer being able to recoup benefits from the visual system and with no alternative but to economize by reducing it. A second potential threshold is a lower limit to cost (orange), with pleiotropic effects preventing extreme reduction if it would cause collateral damage elsewhere. This may prevent optimization and leave the animal with a deleterious vestigial structure. If an animal, A, has evolved sophisticated eyes within a bright, stable, visually complex environment, we can assume that A approaches cost–benefit equilibrium; their energetic investment is adequately offset by the fitness benefits afforded by vision, through foraging, predator avoidance, or finding mates. If A moves to a darker habitat, their investment is the same, but the benefits may decrease (Δb), displacing A from equilibrium (Fig. 2A, A1). Under stable conditions, directional selective pressure will draw A1 along the shortest route back to equilibrium (Fig. 2A). If the available light is still sufficient to support vision, the shortest and most favorable route is often to decrease some cost and recover some benefit, perpendicular to the equilibrium gradient (Fig. 2A). The eye may adapt to exploit lower light levels by maximizing sensitivity at the cost of resolution, as in nocturnal and mesopelagic animals (Warrant 1999; Tierney et al. 2017). This can include simplification through loss of color vision, lower spatial resolution, neural summation, and slower photoreceptors, which may also reduce energetic cost (Warrant 1999; Johnson et al. 2000; Theobald et al. 2006; Stöckl et al. 2016b; Tierney et al. 2017; Emerling 2018; Valdez-Lopez et al. 2018). These savings (Δc) can be reinvested (r) to recoup potential fitness gains (g), in another sensory mode for example. A1 may not fully reduce investment to match Δc; many animals living in dim light actually increase eye size, while simplifying other aspects, to maximize photon capture and boost sensitivity. If A1 successfully survives this transition, it should eventually return to equilibrium (Fig. 2A). Note that another animal, B, may inhabit the same original light environment as A but have simpler eyes (Fig. 2A). These may confer lesser cost and return less information, but still satisfy this equilibrium, perhaps due to a simpler lifestyle, lower predation pressure, or preferential investment in another sensory system. If B undergoes the same habitat shift as A, it will experience lesser displacement from equilibrium and be subject to weaker selective pressure. The above model includes selection as the sole evolutionary driver, which is deeply biologically unrealistic. Neutral evolution also plays a crucial role in reduction and loss (Wilkens and Strecker 2017), so the path toward cost–benefit equilibrium is not usually direct (Fig. 2B). Their relative strengths depend on the magnitude of the deviation from equilibrium; as A approaches equilibrium, selective pressure decreases and the relative role of drift increases (Fig. 2B, inset). Neutral evolution also means that animals in stable environments will float within a tolerable zone surrounding equilibrium (Figs. 2–4, pale green), rather than remaining static and optimized. The rate of neutral evolution varies taxonomically and is influenced by mutation and metabolic rates, population size, DNA repair efficiency, and generation time (Brirten 1984; Ohta 1992). However, rates are generally low, so the effects of drift may take many generations to become apparent—possibly leading to the apparent dominance of selection in “young” dark-living groups (Fumey et al. 2018). Fig. 3 View largeDownload slide Physiological, ecological, and environmental factors in the evolution of eye loss. (A) Non-solar radiation (NSR) recovers the viability of visual tasks below threshold values for solar radiation, so animals exposed to bioluminescence (A1+) may respond as to dim light whereas those without bioluminescence (A1–) are more likely to exhibit reduction. Ecological niche may also introduce additional selective pressures. Active species and predators (e1) may be subject to stronger selective pressures to adapt to exploit dim light. Sedentary species (e2) may economize more on visual systems to reinvest in chemo- or mechanoreceptors. In the absence of NSR, e1 species, having higher overall energetic costs, are likely to be subject to stronger pressure reduce the visual system. (B) Energy limitation. Low-light and aphotic habitats are often energy-limited due to reduced photosynthetic capacity. This imposes an upper limit on energetic cost and enforces very strong directional selective pressure to reduce wastage in costly visual systems. (C) Gradual transitions from light to dark environments. (D) Different cost–benefit relationships will change the selective landscape and therefore the evolutionary dynamics of eye reduction and loss. For example, a sigmoidal curve may result in a shallower Δb/Δc during habitat transitions (A1), reducing the cost deficit and decreasing selective pressure. Fig. 3 View largeDownload slide Physiological, ecological, and environmental factors in the evolution of eye loss. (A) Non-solar radiation (NSR) recovers the viability of visual tasks below threshold values for solar radiation, so animals exposed to bioluminescence (A1+) may respond as to dim light whereas those without bioluminescence (A1–) are more likely to exhibit reduction. Ecological niche may also introduce additional selective pressures. Active species and predators (e1) may be subject to stronger selective pressures to adapt to exploit dim light. Sedentary species (e2) may economize more on visual systems to reinvest in chemo- or mechanoreceptors. In the absence of NSR, e1 species, having higher overall energetic costs, are likely to be subject to stronger pressure reduce the visual system. (B) Energy limitation. Low-light and aphotic habitats are often energy-limited due to reduced photosynthetic capacity. This imposes an upper limit on energetic cost and enforces very strong directional selective pressure to reduce wastage in costly visual systems. (C) Gradual transitions from light to dark environments. (D) Different cost–benefit relationships will change the selective landscape and therefore the evolutionary dynamics of eye reduction and loss. For example, a sigmoidal curve may result in a shallower Δb/Δc during habitat transitions (A1), reducing the cost deficit and decreasing selective pressure. Fig. 4 View largeDownload slide Using phylogenetic character reconstructions to chart evolutionary trajectories in the cost–benefit framework. Tracing character histories across phylogenetic trees (A) allows the “story-boarding” of multiple independent eye loss events, and how these may relate and differ to each other. Lineages sharing reduced features in the same order have likely moved through the adaptive landscape in the same way. By charting this within our cost–benefit framework (B), we can begin to visualize the relative importance of evolutionary drivers of selection, neutral mutation, and pleiotropy. Here, reduced eye characters (shape icons) have appeared in five out of nine species examined. Repeated evolution of the white square character in independent lineages h and d + e indicates that they occupied the same regions on the adaptive landscape, for example. Fig. 4 View largeDownload slide Using phylogenetic character reconstructions to chart evolutionary trajectories in the cost–benefit framework. Tracing character histories across phylogenetic trees (A) allows the “story-boarding” of multiple independent eye loss events, and how these may relate and differ to each other. Lineages sharing reduced features in the same order have likely moved through the adaptive landscape in the same way. By charting this within our cost–benefit framework (B), we can begin to visualize the relative importance of evolutionary drivers of selection, neutral mutation, and pleiotropy. Here, reduced eye characters (shape icons) have appeared in five out of nine species examined. Repeated evolution of the white square character in independent lineages h and d + e indicates that they occupied the same regions on the adaptive landscape, for example. The third key evolutionary driver is pleiotropy. Many genes involved in the visual system also fulfil other functions, such as non-visual phototransduction or multiple developmental roles (Yamamoto et al. 2009; Carlini et al. 2013). These may be protected from degradation by selective pressure, resulting in pleiotropic obstructions on the evolutionary route back to equilibrium, valleys in the adaptive landscape, where certain genes or characteristics cannot be altered without deleterious consequences elsewhere (Fig. 2B). Combined with the influence of drift, even closely related species or populations may therefore take different evolutionary trajectories when subjected to the same displacement (Fig. 2B;Niemiller et al. 2013; Sumner-Rooney et al. 2016). So far, our model does not include any constraints. Visual systems are physically constrained, in that they are subject to optical and biological laws that limit the amount and quality of information they can collect within certain means. Nilsson (2013) demonstrated that there are minimum light intensities for different visual tasks to be optically viable. This introduces a threshold that, once crossed, limits the potential to recover benefit via adaptation as seen in dim-light environments (Fig. 2C). This is the real crux of eye loss. If A moves into a truly dark habitat, beyond this threshold (Fig. 2C, A2), the selective landscape will differ from that under dim (A1) and bright (A) light. Instead of balancing pressures to reduce cost and improve sensitivity, the pressure to economize now dominates, and the visual system will be reduced. This may also increase the role of drift relative to animals in dimly lit environments, which remain optically constrained and subject to two directional selective pressures. If there is any benefit to basic photosensitivity, for circadian rhythms or phototaxis, this is likely to be retained in line with the cost–benefit equilibrium; otherwise, the eventual outcome may be complete loss. However, pleiotropy can also introduce a lower limit on energy expenditure in visual systems, where further reduction, such as genetic losses of function, would cause collateral damage in other areas. This may lead to the persistence of vestigial structures, which are degraded and have no functional benefit (Fig. 2C, A3). Several caveats must be mentioned: these schematics (Figs. 2 and 3) are highly simplified, with absolute threshold limits, a linear cost–benefit relationship, assumptions of no fundamental change in eye type, caeteris paribus in other physiological, ecological, and evolutionary drivers, and a simple vector gradient selective landscape. Although these assumptions may be flawed in their reflection of true biological systems, this could provide a starting framework to facilitate future developments. Biological and environmental factors Although these drivers are the three major influences on the general directions and trends we might expect to see in dark habitats, there are of course many additional factors at play. These may affect the nature and location of thresholds, the geometry of cost–benefit relationships, the extent of developmental constraints, and the nature of the selective pressure landscape both before and after transitions (Table 1). These will have significant impacts on the evolutionary trajectories of different taxa in different habitats (e.g., Fig. 3), and are likely responsible for the substantial differences we observe between instances of eye loss in nature, but to date, few studies have attempted to characterize the relative effects of these parameters (Gonzalez et al. 2018). Table 1 Biological and environmental factors affecting the evolution of eye loss Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Table 1 Biological and environmental factors affecting the evolution of eye loss Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Combined, these evolutionary, physiological, ecological, and environmental factors make disentangling eye loss a daunting task. However, they already go some way in explaining the existing variation we see in different taxa and habitats, and caution that extrapolations from model systems may not always be appropriately applied to alternative species or groups. Two examples are highlighted below, but the reader will appreciate that there are myriad combinations of similarities and differences in these factors that could apply to any pair of animals or habitats. Example 1. Non-solar radiation. The availability of non-solar radiation is critical and affects multiple habitats. Indirect solar radiation and starlight promote investment in highly sensitive vision in nocturnal taxa, but non-solar radiation accounts for deviations in habitats where we usually expect to observe regression (Fig. 3A). The most obvious example is marine bioluminescence, with signals potentially indicating prey, predators, or mates and therefore exerting strong selective pressure in favor of retaining vision. This represents a crucial difference between the evolution of cavefishes and deep sea fishes, despite them sharing similar lifestyles, energy-limited habitats, and developmental constraints (Poulson 2001). This is further supported by findings that in glow worm caves, the only cave habitats to host bioluminescence, animals do not lose their eyes and maximize sensitivity instead (Meyer-Rochow and Liddle 1988). Example 2. Lifestyle and activity. Although bioluminescence enables many deep sea fish, cephalopods, and even sea stars to sustain large, sensitive eyes (Johnsen et al. 2012; Birk et al. 2018), it does not “rescue” vision in many other groups. Scallops, gastropods, ostracods, scale worms, and isopods have all been shown to reduce and lose visual systems with increasing depth (Raupach et al. 2009; Syme and Oakley 2012; Malkowsky and Götze 2014; Sumner-Rooney et al. 2016; Gonzalez et al. 2018). The extent of investment in dim-light vision is often greatest in highly mobile predators, such as fish and cephalopods, which have the greatest potential to recover benefit from vision in low light, while the taxa that exhibit loss are often more sedentary. Interestingly, eye type and complexity may not be dominant factors, with camera (gastropods, fish) and compound (crustaceans, sea stars), simple (gastropods, sea stars) and complex (scallops, cephalopods) eyes being both lost and retained. Strategies and recommendations So, how can all these factors be recognized and evaluated on a large scale? First, comparisons between distant groups (geographically or phylogenetically) should incorporate variation within those groups, too. Despite superficially convergent eye losses, histological-level investigations consistently reveal more variation than anticipated within closely related groups (Malkowsky and Götze 2014; Sumner-Rooney et al. 2016). Phylogenetically contextualized comparative studies have been underused in the study of eye reduction, and have enormous potential to resolve evolutionary trajectories for comparison and hypothesis testing (Desuter-Grandcolas 1997; Klaus et al. 2013; Gonzalez et al. 2018). Second, if caves are “natural replicates” of evolutionary experiments, we can exploit the similarities and differences between dark ecosystems and taxa as experimental treatment groups. To incorporate the various factors outlined above, similar or identical methods should be deployed across different study groups (taxa, habitats, etc.) where possible to identify common trends. Molecular and physiological techniques are being used with great success to quantify the impacts of selection versus drift, and identify potential pleiotropic effects in agents such as Shh (Culver et al. 1995; Jeffery and Martasian 1998; Yamamoto et al. 2004, 2009; Espinasa and Jeffery 2006; Protas et al. 2008; Carlini et al. 2013). However, physiological techniques especially cannot currently be feasibly extended to multiple large groups of animals, yet such comparisons are critical to identifying truly unifying patterns in eye loss. Morphological studies can be deployed across large groups non-destructively and with relatively low costs. Coupled with continued transcriptomic, physiological, developmental, and behavioral research in model systems, there is considerable unexplored potential in comparative morphological studies. Comparative morphology and phylogenetic reconstruction The detailed study and comparison of visual and central nervous system anatomy can provide a wealth of information on evolutionary trajectories, as well as indicating to some extent the underlying mechanisms. Examining multiple lineages that have independently lost or reduced eyes provide “natural replicates” of the evolutionary process, and by applying this to multiple taxonomic groups in multiple habitat types, we can conduct a macroevolutionary experiment to evaluate the proposed conceptual framework, and test evolutionary hypotheses. It is clear that presence/absence data does not capture the extent of structural diversity in visual systems (Malkowsky and Götze 2014; Sumner-Rooney et al. 2016; Birk et al. 2018), but phylogenetic reconstruction of eye loss is a highly effective tool (Raupach et al. 2009; Syme and Oakley 2012; Gonzalez et al. 2018). By combining detailed morphological studies with phylogenetic methods, the evolutionary trajectories of individual eye characters can be reconstructed across large groups (Fig. 4A). This approach revealed variable evolutionary trajectories during multiple deep-sea invasions within a family of ecologically similar and even sympatric snails (Sumner-Rooney et al. 2016). Resolution was not sufficient to draw concrete conclusions about evolutionary processes, but supported strong roles for both selection and neutral evolution in sedentary, benthic detritivores with simple ancestral eye structures, in line with the expectations presented here. If expanded to include more taxa and more characters, including transcriptomic data where possible, this technique could facilitate both within- and between-group comparisons by charting the progression of individual cases of eye loss and reduction phylogenetically. Behavior and physiology: are there robust morphological proxies? Behavioral and physiological data are hugely desirable but not always obtainable, as animals may be too rare, inaccessible, or fragile. However, some basic indicators may be given by features such as neural investment (Banister 1984), numbers of ommatidia, focal length, aperture size, and the predicted functionality of photopigments if sequences are available. Blind mole rats, for example, are sensitive to circadian light regimes but cannot see, and this is elegantly reflected in their morphology: very reduced eyes and visual processing centers in the brain, but an intact suprachiasmatic nucleus (Zubidat et al. 2011). Similarly, differential investment in other areas of the central nervous system can indicate the importance of alternative sensory modalities, including olfaction (Cooper et al. 2001; Stöckl et al. 2016a), gustation, and mechanoreception (Franz-Odendaal and Hall 2006; Soares and Niemiller 2013). Testing evolutionary hypotheses Assembling large datasets of visual, central nervous system, and other sensory characters will enable us to identify regressive features that appear more or less frequently, recurrent associations between specific characters, habitat specificity, and the extent of reinvestment in, and trade-offs with, other sensory modes. Ancestral state reconstructions would allow researchers to story-board the evolution of eye reduction and loss in multiple replicates, multiple habitat types, and multiple ecological niches (Fig. 4). Armed with such a powerful dataset, we could test evolutionary hypotheses on an enormous scale. Differential reduction and reinvestment under varying physical and biological conditions may provide clues as to which factors dominate in different taxa and environments. Recurring features may indicate cost-effective routes to reduction; conversely, the repeated or unexpected survival of other characters may betray pleiotropic limitations, persistence of functionality, or repurposing of retained structures. These features can be used to plot trajectories across adaptive landscapes and help visualize the strength of the various evolutionary drivers (Fig. 4B). Multiple lineages convergently following the same progressive changes have likely moved through the adaptive landscape in the same way and would suggest significant constraints on evolutionary trajectory. These patterns may be consistent across biological settings, but more likely the differences between them will reveal more about the most influential factors and the more variable subjects. Beyond eye loss, this approach could help us to better understand the nature of reductive evolution overall, applied to any number of traits. Specific recommendations Study material Such large-scale comparative studies will require large amounts of specimen material and careful selection of focal groups (see below). While collecting dedicated fresh specimens is ideal, it is not always possible. Natural history collections offer an invaluable resource to supplement or substitute fresh material, covering a large range of taxa, geographical locations, and time periods (Sumner-Rooney and Sigwart 2017). As with any resource, there are limitations to using collection material; samples may be unsuitable for certain analyses due to fixation or preservation methods, age, or the need for destructive sampling. However, high-quality morphological data can sometimes be extracted from even very old ethanol-preserved samples, and spirit collections represent a major untapped resource for large comparative studies. Methods Rapid, non-destructive penetrative imaging tools, such as X-ray tomography (µ-CT and SXRT), are ideal for large-scale studies of precious specimens. Both have produced stunning results in very small visual and nervous systems (Sombke et al. 2015; Taylor et al. 2016), and improved scanners and visualization software now allow the study of eye structure in very small, unstained, wet material (L. Sumner-Rooney et al., in preparation). Taxon selection The selection of study groups requires careful consideration. For comparisons across habitats, researchers should ensure as much similarity between groups as possible in line with the considerations in Table 1. Criteria include: Multiple transitions into dark niches within a relatively low taxonomic rank. All empirical studies should incorporate multiple independent replicates, i.e., multiple lineages independently transitioning into dark habitats or niches (Tierney et al. 2017). Taxonomic rank is also important: the lower the rank, the less likely that confounding phyletic factors are introduced. In cases such as Cicurina and Astyanax, multiple cave invasions have occurred within genera or species (Wilkens and Strecker 2003; Hedin 2015), but in others, family-level analyses are required to capture sufficient replicates (Syme and Oakley 2012; Sumner-Rooney et al. 2016). Existence of, or ability to produce, a robust phylogeny. Resolved phylogenetic relationships are fundamental to any comparative study. If relationships between taxa are not clear, trait history reconstructions are ambiguous and character evolution cannot be reliably traced (Harvey and Pagel 1991). Inclusion of sighted taxa. Sighted relatives must be included in studies of loss, especially when examining large groups of eyeless species; without them, assigning character polarity and the resolution of fine evolutionary changes between reductive lineages are problematic. Multiple types of habitat transition. Comparisons between different habitat shifts are a major potential avenue for future research, but identifying suitable candidate study systems poses a major obstacle. Relatively few taxonomic groups have undergone transitions from, for example, both shallow to deep water and from surface to subterranean habitats in multiple lineages. Conclusions After more than 150 years of study, the biological conundrum of eye loss is firmly back in the research spotlight, and modern methodologies offer real possibilities of collecting large datasets across broad taxonomic ranges. Armed with these new tools, researchers can truly close in on the evolutionary patterns underlying loss. A basic evolutionary framework attempts to incorporate the fundamental evolutionary forces acting on species that transition into dark habitats, with the aim of facilitating future development and guiding future research. We urgently need to survey the morphological variety observed in dark-living animals on a broader scale, alongside continued high-resolution developmental and transcriptomic studies on specific species. Harnessing this variation using the comparative method across a large range of animals and habitats will help us identify the fundamental influences on the evolution of eye loss on a global scale. Acknowledgments I am very grateful to the Company of Biologists, Palaeontological Association (PA-GA201707), American Microscopical Society, Crustacean Society, and Society for Integrative and Comparative Biology (divisions DEDB, DNB, DPCB, DEE, and DIZ) for their support of this symposium. I would also like to thank my co-organizer Megan Porter, Mark Wilkinson and Bill Jeffery for the use of their images in Fig. 1, and two anonymous reviewers and the editors for their constructive feedback on the manuscript. Funding This work was supported by the Oxford University Museum of Natural History. References Banister KE. 1984 . A subterranean population of Garra barreimiae (Teleostei: Cyprinidae) from Oman, with comments on the concept of regressive evolution . J Nat Hist 18 : 927 – 38 . Google Scholar Crossref Search ADS Barr TC. 1968 . Cave ecology and the evolution of troglobites. In: Dobzhansky T , Hecht MK , Steere WC , editors. Evolutionary biology. New York (NY) : Plenum Press . p. 35 – 101 . Birk MH , Blicher ME , Garm A. 2018 . Deep-sea starfish from the Arctic have well-developed eyes in the dark . Proc R Soc B 285 : 20172743. Google Scholar Crossref Search ADS PubMed Bloom T , Binford GA , Esposito L , Garcia GA , Peterson I , Nishida A , Loubet-Senear K , Agnarsson I. 2014 . Discovery of two new species of eyeless spiders within a single Hispaniola cave . J Arachnol 42 : 148 – 54 . Google Scholar Crossref Search ADS Borowsky R. 2013 . Evolution of an adaptive behavior and its sensory receptors facilitates eye regression in blind cavefish . BMC Biol 11 : 81 – 4 . Google Scholar Crossref Search ADS PubMed Brirten RJ. 1984 . Sequence evolution differ between taxonomic groups of DNA . Science 39 : 1393 – 8 . Carlini DB , Fong DW. 2017 . The transcriptomes of cave and surface populations of Gammarus minus (Crustacea: Amphipoda) provide evidence for positive selection on cave downregulated transcripts . PLoS One 12 : e0186173. Google Scholar Crossref Search ADS PubMed Carlini DB , Satish S , Fong DW. 2013 . Parallel reduction in expression, but no loss of functional constraint, in two opsin paralogs within cave populations of Gammarus minus (Crustacea: Amphipoda) . BMC Evol Biol 13 : 89. Google Scholar Crossref Search ADS PubMed Childress JJ. 1995 . Are there physiological and biochemical adaptations of metabolism in deep-sea animals? Trends Ecol Evol 10 : 30 – 6 . Christiansen K. 1961 . Convergence and parallelism in cave Entomobryinae . Evolution 15 : 288. Google Scholar Crossref Search ADS Clarkson E , Levi-setti R , Horva G. 2006 . The eyes of trilobites: the oldest preserved visual system . Arthropod Struct Dev 35 : 247 – 59 . Google Scholar Crossref Search ADS PubMed Cooper RL , Li H , Long LY , Cole JL , Hopper HL. 2001 . Anatomical comparisons of neural systems in sighted epigean and troglobitic crayfish species . J Crustac Biol 21 : 360 – 74 . Google Scholar Crossref Search ADS Culver DC , Kane TC , Fong DW. 1995 . Adaptation and natural selection in caves: the evolution of Gammarus minus . Cambridge (MA) : Harvard University Press . Culver DC , Pipan T. 2009 . The biology of caves and other subterranean habitats . Oxford : Oxford University Press . Danielopol DL, Baltanás A, Bonaduce G. 1996. The darkness syndrome in subsurface-shallow and deep-sea dwelling Ostracoda (Crustacea). In: Uiblein F, Ott J, Stachowitsch M, editors. Deep-sea and extreme shallow-water habitats: affinities and adaptations. Vienna: Austrian Academy of Sciences. p. 123–43. Darwin C. 1859 . On the origin of species . London : John Murray . David-Gray ZK , Janssen JWH , Degrip WJ , Nevo E , Foster RG. 1998 . Light detection in a ‘blind’ mammal . Nat Neurosci 1 : 655 – 6 . Google Scholar Crossref Search ADS PubMed Derkarabetian S, Steinmann DB, Hedin M. 2010. Repeated and time-correlated morphological convergence in cave-dwelling harvestmen (Opiliones, Laniatores) from Montane Western North America. PLoS One 5:e10388. Desuter-Grandcolas L. 1997 . Studies in cave life evolution: a rationale for future theoretical developments using phylogenetic inference . J Zool Syst Evol Res 35 : 23 – 31 . Google Scholar Crossref Search ADS Emerling CA. 2018 . Regressed but not gone: patterns of vision gene loss and retention in subterranean mammals . Integr Comp Biol (doi: 10.1093/icb/icy004). Espinasa L , Jeffery WR. 2006 . Conservation of retinal circadian rhythms during cavefish eye degeneration . Evol Dev 8 : 16 – 22 . Google Scholar Crossref Search ADS PubMed Fišer Ž , Novak L , Luštrik R , Fišer C. 2016 . Light triggers habitat choice of eyeless subterranean but not of eyed surface amphipods . Naturwissenschaften 103 : 7. Google Scholar Crossref Search ADS PubMed Franz-Odendaal TA , Hall BK. 2006 . Modularity and sense organs in the blind cavefish, Astyanax mexicanus . Evol Dev 8 : 94 – 100 . Google Scholar Crossref Search ADS PubMed Friedrich M , Chen R , Daines B , Bao R , Caravas J , Rai PK , Zagmajster M , Peck SB. 2011 . Phototransduction and clock gene expression in the troglobiont beetle Ptomaphagus hirtus of Mammoth cave . J Exp Biol 214 : 3532 – 41 . Google Scholar Crossref Search ADS PubMed Fumey J , Hinaux H , Noirot C , Thermes C , Rétaux S , Casane D. 2018 . Evidence for late Pleistocene origin of Astyanax mexicanus cavefish . BMC Evol Biol 18 : 1 – 19 . Google Scholar Crossref Search ADS PubMed Gibert J, Deharveng L. 2002. Subterranean ecosystems: a truncated functional biodiversity. Bioscience 52:473–81. Gonzalez BC , Worsaae K , Fontaneto D , Martínez A. 2018 . Anophthalmia and elongation of body appendages in cave scale worms (Annelida: Aphroditiformia) . Zool Scr 47 : 106 – 21 . Google Scholar Crossref Search ADS Gross JB , Furterer A , Carlson BM , Stahl BA. 2013 . An integrated transcriptome wide analysis of cave and surface dwelling Astyanax mexicanus . PLoS One 8 : e55659. Google Scholar Crossref Search ADS PubMed Harvey PH , Pagel MD. 1991 . The comparative method in evolutionary biology. Oxford : Oxford University Press . Hedin M. 2015 . High-stakes species delimitation in eyeless cave spiders (Cicurina, Dictynidae, Araneae) from central Texas . Mol Ecol 24 : 346 – 61 . Google Scholar Crossref Search ADS PubMed Hinaux H , Poulain J , da Silva C , Noirot C , Jeffery WR , Casane D , Rétaux S. 2013 . De novo sequencing of Astyanax mexicanus surface fish and Pachón cavefish transcriptomes reveals enrichment of mutations in cavefish putative eye genes . PLoS One 8 : e53553. Google Scholar Crossref Search ADS PubMed Izutsu M, Toyoda A, Fujiyama A, Agata K, Fuse N. 2016. Dynamics of dark-fly genome under environmental selections. G3 Genes Genom Genet 6:365–76. Jeffery WR. 2009 . Regressive evolution in Astyanax cavefish . Annu Rev Genet 43 : 25 – 47 . Google Scholar Crossref Search ADS PubMed Jeffery WR , Martasian DP. 1998 . Evolution of eye regression in the cavefish Astyanax: apoptosis and the Pax-6 gene . Integr Comp Biol 38 : 685 – 96 . Johnsen S , Frank TM , Haddock SHD , Widder EA , Messing CG. 2012 . Light and vision in the deep-sea benthos: i. Bioluminescence at 500–1000 m depth in the Bahamian islands . J Exp Biol 215 : 3335 – 43 . Google Scholar Crossref Search ADS PubMed Johnson ML , Shelton PMJ , Gaten E. 2000 . Temporal resolution in the eyes of marine decapods from coastal and deep-sea habitats . Mar Biol 136 : 243 – 8 . Google Scholar Crossref Search ADS Jones R , Culver DC. 1989 . Evidence for selection on sensory structures in a cave population of Gammarus minus . Evolution 43 : 688 – 93 . Google Scholar Crossref Search ADS PubMed Juan C , Guzik MT , Jaume D , Cooper SJB. 2010 . Evolution in caves: Darwin’s ‘wrecks of ancient life’ in the molecular era . Mol Ecol 19 : 3865 – 80 . Google Scholar Crossref Search ADS PubMed Kim BM , Kang S , Ahn DH , Kim JH , Ahn I , Lee CW , Cho JL , Min GS , Park H. 2017 . First insights into the subterranean crustacean Bathynellacea transcriptome: transcriptionally reduced opsin repertoire and evidence of conserved homeostasis regulatory mechanisms . PLoS One 12 : e0170424 – 2 . Google Scholar Crossref Search ADS PubMed Kimura M. 1968 . Evolutionary rate at the molecular level . Nature 217 : 624 – 6 . Google Scholar Crossref Search ADS PubMed Klaus S , Mendoza JCE , Liew JH , Plath M , Meier R , Yeo DCJ. 2013 . Rapid evolution of troglomorphic characters suggests selection rather than neutral mutation as a driver of eye reduction in cave crabs . Biol Lett 9 : 20121098. Google Scholar Crossref Search ADS PubMed Land MF, Nilsson D–E. 2012. Animal eyes. 2nd ed. Oxford: Oxford University Press. Langille BL , Tierney SM , Austin AD , Humphreys WF , Cooper SJB. 2018 . How blind are they? Phototactic responses in stygobiont diving beetles (Coleoptera: Dytiscidae) from calcrete aquifers of Western Australia . Aust Entomol Soc published online (doi:10.1111/aen.12330). Lerosey-Aubril R. 2006 . Ontogeny of Drevermannia and the origin of blindness in Late Devonian proetoid trilobites . Geol Mag 143 : 89 – 104 . Google Scholar Crossref Search ADS Leung NY , Montell C. 2017 . Unconventional roles of opsins . Annu Rev Cell Dev Biol 33 : 241 – 64 . Google Scholar Crossref Search ADS PubMed Malkowsky Y , Götze M-C. 2014 . Impact of habitat and life trait on character evolution of pallial eyes in Pectinidae (Mollusca: Bivalvia) . Org Divers Evol 14 : 173 . Google Scholar Crossref Search ADS Meng F , Braasch I , Phillips JB , Lin X , Titus T , Zhang C , Postlethwait JH. 2013 . Evolution of the eye transcriptome under constant darkness in Sinocyclocheilus cavefish . Mol Biol Evol 30 : 1527 – 43 . Google Scholar Crossref Search ADS PubMed Meyer-Rochow VB , Liddle AR. 1988 . Structure and function of the eyes of two species of opilionid from New Zealand glow-worm caves (Megalopsalis tumida: Palpatores, and Hendea myersi cavernicola: Laniatores) . Proc R Soc B Biol Sci 233 : 293 – 319 . Google Scholar Crossref Search ADS Mohun SM , Davies WL , Bowmaker JK , Pisani D , Himstedt W , Gower DJ , Hunt DM , Wilkinson M. 2010 . Identification and characterization of visual pigments in caecilians (Amphibia: Gymnophiona), an order of limbless vertebrates with rudimentary eyes . J Exp Biol 213 : 3586 – 92 . Google Scholar Crossref Search ADS PubMed Mohun SM , Wilkinson M. 2015 . The eye of the caecilian Rhinatrema bivittatum (Amphibia: Gymnophiona: Rhinatrematidae) . Acta Zool 96 : 147 – 53 . Google Scholar Crossref Search ADS Moran D , Softley R , Warrant EJ. 2015 . The energetic cost of vision and evolution of eyeless Mexican cavefish . Sci Adv 1 : e1500363. Google Scholar Crossref Search ADS PubMed Niemiller ML , Fitzpatrick BM , Shah P , Schmitz L , Near TJ. 2013 . Evidence for repeated loss of selective constraint in rhodopsin of Amblyopsid cavefishes (Teleosti: Amblyopsidae) . Evolution 67 : 732 – 48 . Google Scholar Crossref Search ADS PubMed Nilsson D-E. 2013 . Eye evolution and its functional basis . Vis Neurosci 30 : 5 – 20 . Google Scholar Crossref Search ADS PubMed Niven JE , Laughlin SB. 2008 . Energy limitation as a selective pressure on the evolution of sensory systems . J Exp Biol 211 : 1792 – 804 . Google Scholar Crossref Search ADS PubMed Ohta T. 1972. Population size and rate of evolution. J Mol Evol 1:305–14 Ohta T. 1992 . The nearly neutral theory of molecular evolution . Annu Rev Ecol Syst 23 : 263 – 86 . Google Scholar Crossref Search ADS Poulson TL. 2001 . Adaptations of cave fishes with some comparisons to deep-sea fishes . Environ Biol Fish 62 : 345 – 64 . Google Scholar Crossref Search ADS Poulson TL. 1963 . Cave adaptation in amblyopsid fishes . Am Midl Nat 70 : 257 – 90 . Google Scholar Crossref Search ADS Protas M , Conrad M , Gross JB , Tabin C , Borowsky R. 2007 . Regressive evolution in the Mexican cave tetra, Astyanax mexicanus . Curr Biol 17 : 452 – 4 . Google Scholar Crossref Search ADS PubMed Protas M , Tabansky I , Conrad M , Gross JB , Vidal O , Tabin CJ , Borowsky R. 2008 . Multi-trait evolution in a cave fish, Astyanax mexicanus . Evol Dev 10 : 196 – 209 . Google Scholar Crossref Search ADS PubMed Raupach MJ , Mayer C , Malyutina M , Wagele J-W. 2009 . Multiple origins of deep-sea Asellota (Crustacea: Isopoda) from shallow waters revealed by molecular data . Proc R Soc B Biol Sci 276 : 799 – 808 . Google Scholar Crossref Search ADS Rétaux S , Casane D. 2013 . Evolution of eye development in the darkness of caves: adaptation, drift, or both? Evodevo 4 : 26. Google Scholar Crossref Search ADS PubMed Sadoglu P. 1967 . The selective value of eye and pigment loss in mexican cave fish . Evolution 21 : 541 – 9 . Google Scholar Crossref Search ADS PubMed Soares D , Niemiller ML. 2013 . Sensory adaptations of fishes to subterranean environments . Bioscience 63 : 274 – 83 . Google Scholar Crossref Search ADS Sombke A , Lipke E , Michalik P , Uhl G , Harzsch S. 2015 . Potential and limitations of X-ray micro-computed tomography in arthropod neuroanatomy: a methodological and comparative survey . J Comp Neurol 523 : 1281 – 95 . Google Scholar Crossref Search ADS PubMed Stahl BA , Gross JB , Speiser DI , Oakley TH , Patel NH , Gould DB , Protas ME. 2015 . A transcriptomic analysis of cave, surface, and hybrid isopod crustaceans of the species Asellus aquaticus . PLoS One 10 : e0140484. Google Scholar Crossref Search ADS PubMed Stöckl A , Heinze S , Charalabidis A , El Jundi B , Warrant E , Kelber A. 2016 . Differential investment in visual and olfactory brain areas reflects behavioural choices in hawk moths . Sci Rep 6 : 26041. Google Scholar Crossref Search ADS PubMed Stöckl AL , Ribi WA , Warrant EJ. 2016 . Adaptations for nocturnal and diurnal vision in the hawkmoth lamina . J Comp Neurol 524 : 160 – 75 . Google Scholar Crossref Search ADS PubMed Sumner-Rooney LH , Sigwart JD , McAfee J , Smith L , Williams ST. 2016 . Repeated eye reduction events reveal multiple pathways to degeneration in a family of marine snails . Evolution 70 : 2268 – 95 . Google Scholar Crossref Search ADS PubMed Sumner-Rooney L , Sigwart JD. 2017 . Lazarus in the museum: resurrecting historic specimens through new technology . Invertebr Zool 14 : 73 – 84 . Syme AE , Oakley TH. 2012 . Dispersal between shallow and abyssal seas and evolutionary loss and regain of compound eyes in cylindroleberidid ostracods: conflicting conclusions from different comparative methods . Syst Biol 61 : 314 – 36 . Google Scholar Crossref Search ADS PubMed Taylor GJ , Ribi W , Bech M , Bodey AJ , Rau C , Steuwer A , Warrant EJ , Baird E. 2016 . The dual function of orchid bee ocelli as revealed by X-ray microtomography . Curr Biol 26 : 1319 – 24 . Google Scholar Crossref Search ADS PubMed Theobald JC , Greiner B , Wcislo WT , Warrant EJ. 2006 . Visual summation in night-flying sweat bees: a theoretical study . Vision Res 46 : 2298 – 309 . Google Scholar Crossref Search ADS PubMed Tierney SM , Cooper SJB , Saint KM , Bertozzi T , Hyde J , Humphreys WF , Austin AD , Tierney SM. 2015 . Opsin transcripts of predatory diving beetles: a comparison of surface and subterranean photic niches . R Soc Open Sci 2 : 140386. Google Scholar Crossref Search ADS PubMed Tierney SM , Friedrich M , Humphreys WF , Jones TM , Warrant EJ , Wcislo WT. 2017 . Consequences of evolutionary transitions in changing photic environments . Austral Entomol 56 : 23 – 46 . Google Scholar Crossref Search ADS Valdez-Lopez JC , Donohue MW , Bok MJ , Wolf J , Cronin TW , Porter ML. 2018 . Sequence, structure, and expression of opsins in the monochromatic stomatopod Squilla empusa . Integr Comp Biol published online (doi:10.1093/icb/icy007/4985722). Warrant EJ. 1999 . Seeing better at night: life style, eye design and the optimum strategy of spatial and temporal summation . Vision Res 39 : 1611 – 30 . Google Scholar Crossref Search ADS PubMed Warrant EJ , Collin SP , Locket NA. 2003 . Eye design in deep-sea fishes. In: Collin SP , Marshall NJ , editors. Sensory processing in aquatic environments. New York (NY ): Springer . p. 303 – 22 . Warrant EJ , Locket NA. 2004 . Vision in the deep sea . Biol Rev Camb Philos Soc 79 : 671 – 712 . Google Scholar Crossref Search ADS PubMed Wilkens H. 1988 . Evolution and genetics of epigean and cave Astyanax fasciatus (Characidae, Pisces) support for the neutral mutation theory . Evol Biol 23 : 271 – 367 . Wilkens H , Strecker U. 2003 . Convergent evolution of the cavefish Astyanax (Characidae, Teleostei): genetic evidence from reduced eye-size and pigmentation . Biol J Linn Soc 80 : 545 – 54 . Google Scholar Crossref Search ADS Wilkens H , Strecker U. 2017 . Evolution in the dark: Darwin’s loss without selection . Berlin : Springer . Williams ST , Smith LM , Herbert DG , Marshall BA , Warén A , Kiel S , Dyal P , Linse K , Vilvens C , Kano Y. 2013 . Cenozoic climate change and diversification on the continental shelf and slope: evolution of gastropod diversity in the family Solariellidae (Trochoidea) . Ecol Evol 3 : 887 – 917 . Google Scholar Crossref Search ADS PubMed Wong JM , Pérez-Moreno JL , Chan TY , Frank TM , Bracken-Grissom HD. 2015 . Phylogenetic and transcriptomic analyses reveal the evolution of bioluminescence and light detection in marine deep-sea shrimps of the family Oplophoridae (Crustacea: Decapoda) . Mol Phylogenet Evol 83 : 278 – 92 . Google Scholar Crossref Search ADS PubMed Yamamoto Y , Byerly MS , Jackman WR , Jeffery WR. 2009 . Pleiotropic functions of embryonic sonic hedgehog expression link jaw and taste bud amplification with eye loss during cave fish evolution . Dev Biol 330 : 200 – 11 . Google Scholar Crossref Search ADS PubMed Yamamoto Y , Stock DW , Jeffery WR. 2004 . Hedgehog signalling controls eye degeneration in blind cavefish . Nature 431 : 844 – 7 . Google Scholar Crossref Search ADS PubMed Yoshizawa M , Jeffery WR. 2008 . Shadow response in the blind cavefish Astyanax reveals conservation of a functional pineal eye . J Exp Biol 211 : 292 – 9 . Google Scholar Crossref Search ADS PubMed Yoshizawa M , Yamamoto Y , O’Quin KE , Jeffery WR. 2012 . Evolution of an adaptive behavior and its sensory receptors promotes eye regression in blind cavefish . BMC Biol 10 : 108. Google Scholar Crossref Search ADS PubMed Zubidat AE , Nelson RJ , Haim A. 2011 . Spectral and duration sensitivity to light-at-night in “blind” and sighted rodent species . J Exp Biol 214 : 3206 – 17 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Integrative and Comparative Biology Oxford University Press

The Kingdom of the Blind: Disentangling Fundamental Drivers in the Evolution of Eye Loss

Loading next page...
 
/lp/ou_press/the-kingdom-of-the-blind-disentangling-fundamental-drivers-in-the-LGgutAZ5sW
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oup.com.
ISSN
1540-7063
eISSN
1557-7023
D.O.I.
10.1093/icb/icy047
Publisher site
See Article on Publisher Site

Abstract

Abstract Light is a fundamentally important biological cue used by almost every animal on earth, to maintain daily rhythms, navigate, forage, find mates, or avoid predators. But an enormous number of species live in darkness: in subterranean caves, deep oceans, underground burrows, and within parasitic host bodies, and the loss of eyes appears consistently across these ecosystems. However, the evolutionary mechanisms that lead to the reduction of the visual system remain the subject of great interest and debate more than 150 years after Darwin tackled the issue. Studies of model taxa have discovered significant roles for natural selection, neutral evolution, and pleiotropy, but the interplay between them remains unclear. To nail down unifying concepts surrounding the evolution of eye loss, we must embrace the enormous range of affected animals and habitats. The fine developmental details of model systems such as the Mexican cave tetra Astyanax mexicanus have transformed and enriched the field, but these should be complemented by wider studies to identify truly overarching patterns that apply throughout animals. Here, the major evolutionary drivers are placed within a conceptual cost–benefit framework that incorporates the fundamental constraints and forces that influence evolution in the dark. Major physiological, ecological, and environmental factors are considered within the context of this framework, which appears faithful to observed patterns in deep-sea and cavernicolous animals. To test evolutionary hypotheses, a comparative phylogenetic approach is recommended, with the goal of studying large groups exhibiting repeated reduction, and then comparing these across habitats, taxa, and lifestyles. Currently, developmental and physiological methods cannot feasibly be used on such large scales, but penetrative imaging techniques could provide detailed morphological data non-invasively and economically for large numbers of species. Comprehensive structural datasets can then be contextualized phylogenetically to examine recurrent trends and associations, and to reconstruct character histories through multiple independent transitions into darkness. By assessing these evolutionary trajectories within an energetic cost–benefit framework, the relationships between fundamental influences can be inferred and compared across different biological and physical parameters. However, substantial numbers of biological and environmental factors affect the evolutionary trajectory of loss, and it is critical that researchers make fair and reasonable comparisons between objectively similar groups. Introduction Biologists have long been fascinated with the pale, eerie, creatures that inhabit the world of darkness. Animals that live in true darkness, rather than the dim light associated with nocturnality and the mesopelagic zone, are found in a multitude of different habitats, including the bathypelagic (ocean depths below around 1000 m), subterranean, and submarine caves, fossorial burrows and, to some extent, endoparasitic host bodies. Many dark-living animals share striking morphological features, including albinism, increased investment in chemo- and mechanoreceptors, and the reduction or loss of eyes (Poulson 2001). Eye loss in particular has captured attention for more than a century, and has been reported across Metazoa from burrowing amphibians (Fig. 1; Mohun et al. 2010; Mohun and Wilkinson 2015) to deep-sea trilobites (Clarkson et al. 2006; Lerosey-Aubril 2006). Fig. 1 View largeDownload slide Eye loss affects a wide variety of animals living in many different dark habitats. Clockwise from top left, Bathymophila diadema, a marine gastropod found at depths of 650–1000 m; Microcaecilia dermatophaga, a burrowing caecilian (image courtesy of Mark Wilkinson); Metagonia jamaica, a cavernicolous spider (specimen SMNH 1408109); Astyanax mexicanus, cave population (image courtesy of Bill Jeffery). Fig. 1 View largeDownload slide Eye loss affects a wide variety of animals living in many different dark habitats. Clockwise from top left, Bathymophila diadema, a marine gastropod found at depths of 650–1000 m; Microcaecilia dermatophaga, a burrowing caecilian (image courtesy of Mark Wilkinson); Metagonia jamaica, a cavernicolous spider (specimen SMNH 1408109); Astyanax mexicanus, cave population (image courtesy of Bill Jeffery). We are now able to identify incremental molecular and developmental changes occurring in dark-living species, which have started to demystify this recurrent evolutionary pattern (e.g., Jeffery 2009; Stahl et al. 2015; Tierney et al. 2015). However, the overall evolutionary mechanisms behind eye loss remain controversial, more than 150 years after it was first explored by Darwin (1859; Rétaux and Casane 2013). Eye loss appears to be driven at least in part by direct selective pressure, but both pleiotropy and genetic drift are also key influencers (Rétaux and Casane 2013). Despite substantial recent progress, thanks to a combination of developmental and molecular techniques, it is not yet clear how these drivers interact, or, crucially, whether their relationships are similar across taxa and habitat types. Unifying concepts have been identified in the constructive evolution of eyes that apply across all eye types and animal groups (Nilsson 2013), but their regressive evolution remains a tangled matter. Here, a conceptual framework is proposed that could inform future studies of eye loss effectively across large and distant taxonomic groups, and several recommendations are made for such research directions. This paper will not comprehensively review the substantial body of existing work on individual taxa, particularly the Mexican cave tetra Astyanax mexicanus by Jeffery, Wilkens, and colleagues (e.g., Wilkens and Strecker 2003; Protas et al. 2007; Jeffery 2009; Borowsky 2013). This species is an exceptional model for studying the developmental mechanisms behind eyelessness in numerous instances across a cave network, and the resulting work has transformed our understanding of eye loss (see Jeffery [2009]; Wilkens and Strecker [2017] for reviews; and “The state of play” below). However, the current focus is on unifying concepts, cross-habitat comparisons, and non-model groups, and not individual case studies. The majority of eyeless species have not yet been kept or bred in captivity (and some have never been seen live), for reasons of accessibility, availability, or fragility; this paper mainly addresses how we can learn from these taxa in the absence of manipulative methods. The overarching aims are to explore the possibility of a generalized model for eye reduction, incorporating macroevolutionary trends that apply across taxonomic and ecological divides; and to promote a large-scale comparative morphological approach. This does not supplant the exceptionally important ongoing developmental, physiological, and transcriptomic work, but is intended to complement that work, and provide a taxonomic breadth and macroscopic evolutionary perspective that may be challenging to achieve using such methods. If evolutionary biologists aim to identify unifying concepts in the reductive evolution of eyes (or any other system), we must embrace the enormous range of affected animals and habitats (Fig. 1), and all the complications this entails. Broader taxonomic approaches have proved successful in other cases of repeated constructive evolution, including vision, which has been elegantly rationalized using trends that apparently apply to all animal groups (Nilsson 2013). Features including screening pigment, membrane folding, and focussing optics are causally linked to the increasing sophistication of visual tasks, and to the minimum light intensity mathematically required for these tasks. Such relationships are likely to be equally important in reductive evolution, as they affect the functionality and therefore the fitness value of visual structures in different light environments. By adopting this functional evolutionary approach to loss, we may more easily rationalize what we observe in nature. For example, high-resolution vision requires minimum light intensities roughly equivalent to those at ocean depths of 350–650 m (Nilsson 2013). Unsurprisingly, this is the point beyond which dramatic alterations to eye structure, including loss (Syme and Oakley 2012; Sumner-Rooney et al. 2016; Gonzalez et al. 2018), begin to appear in multiple marine groups (Warrant et al. 2003; Warrant and Locket 2004; Johnsen et al. 2012). There must be fundamental relationships between the physiological, ecological, and environmental characteristics of study systems and their evolutionary trajectory under darkness. However, unlike the constructive evolution of vision, a wealth of information about structure, function, development, and ecology is largely lacking for reduced-eyed animals, and some of this information is more readily obtained than others. The state of play: a (very) brief summary of eye loss research The consistent loss of eyes in dark-living animals has intrigued evolutionary biologists for centuries. In On the Origin of Species, Darwin wrestled with a selective basis for eye loss, uncertain as to what disadvantage eyes might confer. He eventually attributed it to “disuse” of eyes in the dark, in what we would now consider to be neutral evolution (Darwin 1859). Neutral evolution theory in its modern form developed from the 1930s and 1940s onwards, with its importance in regressive evolution being acknowledged decades before Kimura’s (1968) seminal paper. Although this view was not universally held, it dominated for another century (Culver and Pipan 2009). Following the new synthesis, and the accelerating discovery of more eyeless species, the evolution of loss became a subject of renewed interest in the second half of the 20th century. A new experimental approach, using cave fishes and crustaceans, produced evidence of convergent patterns in repeated cave invasions, and that natural selection favors the reduction and loss of eyes in energy-limited aphotic environments (Christiansen 1961; Poulson 1963; Sadoglu 1967; Barr 1968; Jones and Culver 1989). However, the dominant role of neutral evolution in eye loss was still championed by Wilkens and colleagues (e.g., Wilkens 1988; Wilkens and Strecker 2003, 2017: Chapter 7), who highlight the variation in regressive phenotypes and genotypes, pre-adaptation, and potential misconception of energy limitation in caves. The past two decades have seen enormous growth in molecular, developmental, and physiological studies, most notably using Astyanax (see Jeffery [2009]; Wilkens and Strecker [2017] for reviews). These have generated substantial evidence to demonstrate that direct selection favors energetic economy via eye reduction (Culver et al. 1995; Niven and Laughlin 2008; Yoshizawa et al. 2012; Moran et al. 2015; contraWilkens and Strecker 2017). Several also support a crucial role for pleiotropy in eye development through molecular players such as Shh (Yamamoto et al. 2009; Carlini et al. 2013). In the last 10 years, the capabilities of genetics and transcriptomics have improved dramatically, and transcriptome sequencing especially has been applied to a range of eyeless taxa (Friedrich et al. 2011; Carlini et al. 2013; Meng et al. 2013; Stahl et al. 2015; Tierney et al. 2015; Kim et al. 2017). Many of these found persistent expression of opsins and clock genes in (mainly cavernicolous) dark-living animals (Friedrich et al. 2011), often subject to downregulation or functional shifts (Gross et al. 2013; Hinaux et al. 2013; Tierney et al. 2015; Wong et al. 2015; Carlini and Fong 2017). It must also be noted that there are increasing examples of other roles for opsins beside photoreception (Leung and Montell 2017). Knowledge gaps Two key patterns have emerged that are potentially problematic to drawing broader conclusions about the reductive evolution of visual systems. Firstly, cave fauna has formed the focus of the majority of studies to date. As a crude measure, searches for “eyeless” or “blind” species on Zoobank return 65% cavernicolous, 24% fossorial, 5.8% ground-living (leaf litter), 4.8% marine and <1% parasitic taxa (as of May 15, 2018). In studies specifically addressing the evolution of eye loss, the skew is far greater, with 78% focusing on cavernicolous species, 8% on fossorial species, 7% on marine species, and 7% nocturnal. Secondly, certain taxonomic groups have received more attention than others. Fishes account for 51% of ZooBank search results returned for “eyeless” or “blind” species, yet there are fewer than 200 described teleosts of a predicted 21,000 global troglobionts (Juan et al. 2010; Soares and Niemiller 2013). Arthropods are, more understandably, the next largest group (36%), followed by snakes (12%), and others (including mammals, amphibians, and molluscs; <1%). Again, this trend is reflected in studies specifically examining eye loss, where target systems were 57% fish, 26% arthropods, 8% mammals, 5% amphibians, 3% molluscs, and <1% reptiles. Specialization within one or two of these ecosystems or taxa has created artificial divisions between researchers working on cavernicolous, deep-sea, fossorial, and parasitic species, and opportunities for synthesizing information have largely been underexploited, aside from a handful of research groups. Of course, there are fundamental differences between habitat types and the transitions undergone by different lineages. An integrative approach exploiting these differences will facilitate more objective and informative comparisons between them. For instance, cave taxa may be the subject of greater research interest because eye loss appears to occur more consistently than in deep-sea taxa, but this also represents an opportunity to investigate why they differ. What is “loss”? As an added obstacle, many studies and species descriptions dedicate just a handful of words to eye structure, even when investigating trait evolution (Raupach et al. 2009; Williams et al. 2013; Gonzalez et al. 2018). Absence of retinal pigmentation alone may be taken to signify eye loss, with the implication that loss is total, and pigmented (therefore superficially visible) eyes are presumed to be unaffected. Absolutist attitudes toward loss are deeply inhibitory to fully understanding the complex evolutionary processes behind it. Many taxa have been found to have non-pigmented eyes, vestigial or intact, and eyes with retained pigmentation may be reduced in size, sunken beneath skin, affected by retinal disruption or lens degradation, or impacted by altered gene sequences or expression (Jeffery 2009; Bloom et al. 2014; Sumner-Rooney et al. 2016). These are all vitally important characters that could shed further light on the evolution of loss but are going unrecorded. Similarly, degradation of structure does not necessarily equate to complete loss of function. While pupil obstruction, lens degradation and loss of pigmentation may impede spatial resolution, there is substantial evidence for persistent photoreceptor function and light-responsive behavior in species with dramatically reduced eyes (Friedrich et al. 2011; Fišer et al. 2016; Langille et al. 2018). This may include photoreceptor activity within the retina (David-Gray et al. 1998; Espinasa and Jeffery 2006; Zubidat et al. 2011) or elsewhere (Yoshizawa and Jeffery 2008), for phototaxis or circadian entrainment. Moving forwards: the future of the field The unifying questions in the evolution of eye loss remain; what are the evolutionary forces at play, what are their relative roles; and how do animals move toward eye loss? (See Culver and Pipan [2009, 109–30]; and Rétaux and Casane [2013] for reviews) As outlined above, significant progress has been made to address these questions in various taxa, but there are still substantial gaps in both our knowledge and, to an extent, within the research community working in the field of eye reduction and loss. To provide a unifying framework, we must identify fundamental factors that influence the trajectory toward eye loss across all habitat types and taxa. Here, such a framework is outlined for the first time. Several specific considerations for future work are outlined; many have been proposed by other authors, but none, as far as the author is aware, have been addressed directly in empirical studies. In some cases, it may be possible to incorporate these considerations into experimental designs, but otherwise they should be accounted for in interpretation of results and the drawing of wider conclusions. Fundamental drivers in the evolution of eye loss—a first attempt at conceptualization Eyes and corresponding brain activity are energetically expensive, which is fundamental to their loss; once their utility decreases, this energy could be reinvested elsewhere with greater results, and their high cost amplifies the potential benefits (Niven and Laughlin 2008; Moran et al. 2015). As with any costly trait, selection should favor the maximization of fitness benefits against these energetic costs, consistently drawing animals toward a cost–benefit equilibrium (Fig. 2, in green). The further an animal from equilibrium, the stronger the selective pressure, so we might imagine a fairly simple vector gradient of selective pressure toward equilibrium (Fig. 2, background gray arrows). Fig. 2 View largeDownload slide Schematic cost–benefit relationships for visual systems in changing photic environments, and the evolutionary drivers involved in loss. (A) Selection. Animal A inhabits a bright, visually complex and stable environment, and has evolved a visual system well-suited to its habitat and lifestyle. Overall energetic costs and fitness benefits from the visual system are at (green line) or near (pale green zone) equilibrium. If A moves to a darker habitat (A1) the return on these costs decreases (Δb). Natural selection (background, gray arrows) draws A1 back toward equilibrium, favoring the shortest route (black arrow) and with greater strength at greater distance from equilibrium. The fitness area beneath the line of equilibrium (in yellow) is inaccessible due to physical constraint; the eye cannot optically obtain more information from a stable environment without increased investment in size, complexity, innervation, etc. However, A1 can fully or partially recover its cost deficit (Δc) by reinvesting energy elsewhere (r) to maximize potential fitness gains (g) in other ways, such as an alternative sensory mode. Species with simpler visual systems and lower initial investment (B) undergo smaller displacements (B1) and are likely subjected to weaker selective pressure. (B) Drift and pleiotropy. Evolutionary trajectory is simultaneously affected by two additional drivers: drift (departure from the direction of selection) and pleiotropic effects (orange obstacles). Selection is strongest after greater displacement (1) and decreases as equilibrium is approached (2). The relative importance of drift, and therefore trajectory variability, increases with proximity to equilibrium (inset). Pleiotropic effects on genes involved in vision may obstruct certain evolutionary routes and cause additional diversions (p1 and p2). (C) Constraints. Owing to optical constraints, there is a threshold light intensity value below which vision is no longer physically viable (purple). Beyond this threshold (A2, A3), the selective pressure landscape changes, with the animal no longer being able to recoup benefits from the visual system and with no alternative but to economize by reducing it. A second potential threshold is a lower limit to cost (orange), with pleiotropic effects preventing extreme reduction if it would cause collateral damage elsewhere. This may prevent optimization and leave the animal with a deleterious vestigial structure. Fig. 2 View largeDownload slide Schematic cost–benefit relationships for visual systems in changing photic environments, and the evolutionary drivers involved in loss. (A) Selection. Animal A inhabits a bright, visually complex and stable environment, and has evolved a visual system well-suited to its habitat and lifestyle. Overall energetic costs and fitness benefits from the visual system are at (green line) or near (pale green zone) equilibrium. If A moves to a darker habitat (A1) the return on these costs decreases (Δb). Natural selection (background, gray arrows) draws A1 back toward equilibrium, favoring the shortest route (black arrow) and with greater strength at greater distance from equilibrium. The fitness area beneath the line of equilibrium (in yellow) is inaccessible due to physical constraint; the eye cannot optically obtain more information from a stable environment without increased investment in size, complexity, innervation, etc. However, A1 can fully or partially recover its cost deficit (Δc) by reinvesting energy elsewhere (r) to maximize potential fitness gains (g) in other ways, such as an alternative sensory mode. Species with simpler visual systems and lower initial investment (B) undergo smaller displacements (B1) and are likely subjected to weaker selective pressure. (B) Drift and pleiotropy. Evolutionary trajectory is simultaneously affected by two additional drivers: drift (departure from the direction of selection) and pleiotropic effects (orange obstacles). Selection is strongest after greater displacement (1) and decreases as equilibrium is approached (2). The relative importance of drift, and therefore trajectory variability, increases with proximity to equilibrium (inset). Pleiotropic effects on genes involved in vision may obstruct certain evolutionary routes and cause additional diversions (p1 and p2). (C) Constraints. Owing to optical constraints, there is a threshold light intensity value below which vision is no longer physically viable (purple). Beyond this threshold (A2, A3), the selective pressure landscape changes, with the animal no longer being able to recoup benefits from the visual system and with no alternative but to economize by reducing it. A second potential threshold is a lower limit to cost (orange), with pleiotropic effects preventing extreme reduction if it would cause collateral damage elsewhere. This may prevent optimization and leave the animal with a deleterious vestigial structure. If an animal, A, has evolved sophisticated eyes within a bright, stable, visually complex environment, we can assume that A approaches cost–benefit equilibrium; their energetic investment is adequately offset by the fitness benefits afforded by vision, through foraging, predator avoidance, or finding mates. If A moves to a darker habitat, their investment is the same, but the benefits may decrease (Δb), displacing A from equilibrium (Fig. 2A, A1). Under stable conditions, directional selective pressure will draw A1 along the shortest route back to equilibrium (Fig. 2A). If the available light is still sufficient to support vision, the shortest and most favorable route is often to decrease some cost and recover some benefit, perpendicular to the equilibrium gradient (Fig. 2A). The eye may adapt to exploit lower light levels by maximizing sensitivity at the cost of resolution, as in nocturnal and mesopelagic animals (Warrant 1999; Tierney et al. 2017). This can include simplification through loss of color vision, lower spatial resolution, neural summation, and slower photoreceptors, which may also reduce energetic cost (Warrant 1999; Johnson et al. 2000; Theobald et al. 2006; Stöckl et al. 2016b; Tierney et al. 2017; Emerling 2018; Valdez-Lopez et al. 2018). These savings (Δc) can be reinvested (r) to recoup potential fitness gains (g), in another sensory mode for example. A1 may not fully reduce investment to match Δc; many animals living in dim light actually increase eye size, while simplifying other aspects, to maximize photon capture and boost sensitivity. If A1 successfully survives this transition, it should eventually return to equilibrium (Fig. 2A). Note that another animal, B, may inhabit the same original light environment as A but have simpler eyes (Fig. 2A). These may confer lesser cost and return less information, but still satisfy this equilibrium, perhaps due to a simpler lifestyle, lower predation pressure, or preferential investment in another sensory system. If B undergoes the same habitat shift as A, it will experience lesser displacement from equilibrium and be subject to weaker selective pressure. The above model includes selection as the sole evolutionary driver, which is deeply biologically unrealistic. Neutral evolution also plays a crucial role in reduction and loss (Wilkens and Strecker 2017), so the path toward cost–benefit equilibrium is not usually direct (Fig. 2B). Their relative strengths depend on the magnitude of the deviation from equilibrium; as A approaches equilibrium, selective pressure decreases and the relative role of drift increases (Fig. 2B, inset). Neutral evolution also means that animals in stable environments will float within a tolerable zone surrounding equilibrium (Figs. 2–4, pale green), rather than remaining static and optimized. The rate of neutral evolution varies taxonomically and is influenced by mutation and metabolic rates, population size, DNA repair efficiency, and generation time (Brirten 1984; Ohta 1992). However, rates are generally low, so the effects of drift may take many generations to become apparent—possibly leading to the apparent dominance of selection in “young” dark-living groups (Fumey et al. 2018). Fig. 3 View largeDownload slide Physiological, ecological, and environmental factors in the evolution of eye loss. (A) Non-solar radiation (NSR) recovers the viability of visual tasks below threshold values for solar radiation, so animals exposed to bioluminescence (A1+) may respond as to dim light whereas those without bioluminescence (A1–) are more likely to exhibit reduction. Ecological niche may also introduce additional selective pressures. Active species and predators (e1) may be subject to stronger selective pressures to adapt to exploit dim light. Sedentary species (e2) may economize more on visual systems to reinvest in chemo- or mechanoreceptors. In the absence of NSR, e1 species, having higher overall energetic costs, are likely to be subject to stronger pressure reduce the visual system. (B) Energy limitation. Low-light and aphotic habitats are often energy-limited due to reduced photosynthetic capacity. This imposes an upper limit on energetic cost and enforces very strong directional selective pressure to reduce wastage in costly visual systems. (C) Gradual transitions from light to dark environments. (D) Different cost–benefit relationships will change the selective landscape and therefore the evolutionary dynamics of eye reduction and loss. For example, a sigmoidal curve may result in a shallower Δb/Δc during habitat transitions (A1), reducing the cost deficit and decreasing selective pressure. Fig. 3 View largeDownload slide Physiological, ecological, and environmental factors in the evolution of eye loss. (A) Non-solar radiation (NSR) recovers the viability of visual tasks below threshold values for solar radiation, so animals exposed to bioluminescence (A1+) may respond as to dim light whereas those without bioluminescence (A1–) are more likely to exhibit reduction. Ecological niche may also introduce additional selective pressures. Active species and predators (e1) may be subject to stronger selective pressures to adapt to exploit dim light. Sedentary species (e2) may economize more on visual systems to reinvest in chemo- or mechanoreceptors. In the absence of NSR, e1 species, having higher overall energetic costs, are likely to be subject to stronger pressure reduce the visual system. (B) Energy limitation. Low-light and aphotic habitats are often energy-limited due to reduced photosynthetic capacity. This imposes an upper limit on energetic cost and enforces very strong directional selective pressure to reduce wastage in costly visual systems. (C) Gradual transitions from light to dark environments. (D) Different cost–benefit relationships will change the selective landscape and therefore the evolutionary dynamics of eye reduction and loss. For example, a sigmoidal curve may result in a shallower Δb/Δc during habitat transitions (A1), reducing the cost deficit and decreasing selective pressure. Fig. 4 View largeDownload slide Using phylogenetic character reconstructions to chart evolutionary trajectories in the cost–benefit framework. Tracing character histories across phylogenetic trees (A) allows the “story-boarding” of multiple independent eye loss events, and how these may relate and differ to each other. Lineages sharing reduced features in the same order have likely moved through the adaptive landscape in the same way. By charting this within our cost–benefit framework (B), we can begin to visualize the relative importance of evolutionary drivers of selection, neutral mutation, and pleiotropy. Here, reduced eye characters (shape icons) have appeared in five out of nine species examined. Repeated evolution of the white square character in independent lineages h and d + e indicates that they occupied the same regions on the adaptive landscape, for example. Fig. 4 View largeDownload slide Using phylogenetic character reconstructions to chart evolutionary trajectories in the cost–benefit framework. Tracing character histories across phylogenetic trees (A) allows the “story-boarding” of multiple independent eye loss events, and how these may relate and differ to each other. Lineages sharing reduced features in the same order have likely moved through the adaptive landscape in the same way. By charting this within our cost–benefit framework (B), we can begin to visualize the relative importance of evolutionary drivers of selection, neutral mutation, and pleiotropy. Here, reduced eye characters (shape icons) have appeared in five out of nine species examined. Repeated evolution of the white square character in independent lineages h and d + e indicates that they occupied the same regions on the adaptive landscape, for example. The third key evolutionary driver is pleiotropy. Many genes involved in the visual system also fulfil other functions, such as non-visual phototransduction or multiple developmental roles (Yamamoto et al. 2009; Carlini et al. 2013). These may be protected from degradation by selective pressure, resulting in pleiotropic obstructions on the evolutionary route back to equilibrium, valleys in the adaptive landscape, where certain genes or characteristics cannot be altered without deleterious consequences elsewhere (Fig. 2B). Combined with the influence of drift, even closely related species or populations may therefore take different evolutionary trajectories when subjected to the same displacement (Fig. 2B;Niemiller et al. 2013; Sumner-Rooney et al. 2016). So far, our model does not include any constraints. Visual systems are physically constrained, in that they are subject to optical and biological laws that limit the amount and quality of information they can collect within certain means. Nilsson (2013) demonstrated that there are minimum light intensities for different visual tasks to be optically viable. This introduces a threshold that, once crossed, limits the potential to recover benefit via adaptation as seen in dim-light environments (Fig. 2C). This is the real crux of eye loss. If A moves into a truly dark habitat, beyond this threshold (Fig. 2C, A2), the selective landscape will differ from that under dim (A1) and bright (A) light. Instead of balancing pressures to reduce cost and improve sensitivity, the pressure to economize now dominates, and the visual system will be reduced. This may also increase the role of drift relative to animals in dimly lit environments, which remain optically constrained and subject to two directional selective pressures. If there is any benefit to basic photosensitivity, for circadian rhythms or phototaxis, this is likely to be retained in line with the cost–benefit equilibrium; otherwise, the eventual outcome may be complete loss. However, pleiotropy can also introduce a lower limit on energy expenditure in visual systems, where further reduction, such as genetic losses of function, would cause collateral damage in other areas. This may lead to the persistence of vestigial structures, which are degraded and have no functional benefit (Fig. 2C, A3). Several caveats must be mentioned: these schematics (Figs. 2 and 3) are highly simplified, with absolute threshold limits, a linear cost–benefit relationship, assumptions of no fundamental change in eye type, caeteris paribus in other physiological, ecological, and evolutionary drivers, and a simple vector gradient selective landscape. Although these assumptions may be flawed in their reflection of true biological systems, this could provide a starting framework to facilitate future developments. Biological and environmental factors Although these drivers are the three major influences on the general directions and trends we might expect to see in dark habitats, there are of course many additional factors at play. These may affect the nature and location of thresholds, the geometry of cost–benefit relationships, the extent of developmental constraints, and the nature of the selective pressure landscape both before and after transitions (Table 1). These will have significant impacts on the evolutionary trajectories of different taxa in different habitats (e.g., Fig. 3), and are likely responsible for the substantial differences we observe between instances of eye loss in nature, but to date, few studies have attempted to characterize the relative effects of these parameters (Gonzalez et al. 2018). Table 1 Biological and environmental factors affecting the evolution of eye loss Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Table 1 Biological and environmental factors affecting the evolution of eye loss Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Factor More affected Less affected Effect Additional references Physiological  Complexity of eye Complex Simple Higher Δb, stronger selective pressure to reduce complex eyes  Possession of extraocular photoreceptors (EOPs) EOPs No EOPs More potential benefit to maintaining reduced eyes, e.g., for circadian entrainment, if no other EOPs (Yoshizawa and Jeffery 2008)  Eye type Apposition eyes, small camera eyes, rapid photoreceptors Superposition eyes, large camera eyes, slower photoreceptors, eyespots Systems better suited for high sensitivity have lower visual thresholds and sustain vision in darker habitats, may be more likely to adapt than reduce (Land and Nilsson 2012; Nilsson 2013)  Life history Direct developers, no photic phase; longer generation times Larvae or juveniles in photic habitats; shorter generation times Pressure to retain larval eye and development more constrained where a photic phase exists. Shorter generation times allow faster degradation (Poulson 2001)  Rate of drift Smaller populations; higher metabolism Larger populations; lower metabolism Higher relative influence of drift vs. selection (Ohta 1992)  Pre-adaptation to darkness Not pre-adapted Pre-adapted (e.g., nocturnal, mesopelagic) Pre-adapted animals have lower Δb (Poulson 2001) Ecological  Pelagic/benthic Benthic Pelagic Benthic species have increased availability of other cues and likely to be less active (below) (Childress 1995)  Active predator/scavenger Scavenger Active predator Greater selective pressure to sustain visual hunting where possible in active predators (Gibert and Deharveng 2002)  Sedentary/active Sedentary Active Lower selective pressure to sustain vision in sedentary species (Warrant 1999; Warrant and Locket 2004)  Predation pressure Caves, fossorial Deep sea, parasites? Weaker selective pressure to maintain vision where few top predators (Gibert and Deharveng 2002)  Connectivity Isolated populations (e.g., caves, fossorial, parasites, deep sea trenches, and vents) Mixing populations (e.g., open deep sea, nocturnal) Restricted gene flow enables faster change (Ohta 1972)  Age of transition Older populations? e.g., deep sea Younger populations? e.g., caves Increased time may be needed for change to occur, particularly by neutral evolution (Culver and Pipan 2009; Derkarabetian et al. 2010) Environmental  Non-solar radiation Most caves, fossorial, parasites? Deep sea, glow worm caves, nocturnal Where present, greater functionality and improved fitness (foraging/mate finding) (Poulson 2001; Warrant and Locket 2004)  Stability in light level Caves, deep sea, parasites Nocturnal, fossorial Interrupted darkness may maintain benefits of vision (Culver and Pipan 2009)  Speed of transition Sudden (some caves) Gradual (deep sea, some caves, fossorial) Gradual transitions may enable constant adaptation and less departure from cost-benefit equilibrium (Danielopol et al. 1996)  Energy limitation Most caves, deep sea Some caves (e.g., guano caves), fossorial, parasites, nocturnal Energy limitation as an upper cost threshold, with greater pressure to economize (Poulson 2001; Niven and Laughlin 2008; Izutsu et al. 2016) Combined, these evolutionary, physiological, ecological, and environmental factors make disentangling eye loss a daunting task. However, they already go some way in explaining the existing variation we see in different taxa and habitats, and caution that extrapolations from model systems may not always be appropriately applied to alternative species or groups. Two examples are highlighted below, but the reader will appreciate that there are myriad combinations of similarities and differences in these factors that could apply to any pair of animals or habitats. Example 1. Non-solar radiation. The availability of non-solar radiation is critical and affects multiple habitats. Indirect solar radiation and starlight promote investment in highly sensitive vision in nocturnal taxa, but non-solar radiation accounts for deviations in habitats where we usually expect to observe regression (Fig. 3A). The most obvious example is marine bioluminescence, with signals potentially indicating prey, predators, or mates and therefore exerting strong selective pressure in favor of retaining vision. This represents a crucial difference between the evolution of cavefishes and deep sea fishes, despite them sharing similar lifestyles, energy-limited habitats, and developmental constraints (Poulson 2001). This is further supported by findings that in glow worm caves, the only cave habitats to host bioluminescence, animals do not lose their eyes and maximize sensitivity instead (Meyer-Rochow and Liddle 1988). Example 2. Lifestyle and activity. Although bioluminescence enables many deep sea fish, cephalopods, and even sea stars to sustain large, sensitive eyes (Johnsen et al. 2012; Birk et al. 2018), it does not “rescue” vision in many other groups. Scallops, gastropods, ostracods, scale worms, and isopods have all been shown to reduce and lose visual systems with increasing depth (Raupach et al. 2009; Syme and Oakley 2012; Malkowsky and Götze 2014; Sumner-Rooney et al. 2016; Gonzalez et al. 2018). The extent of investment in dim-light vision is often greatest in highly mobile predators, such as fish and cephalopods, which have the greatest potential to recover benefit from vision in low light, while the taxa that exhibit loss are often more sedentary. Interestingly, eye type and complexity may not be dominant factors, with camera (gastropods, fish) and compound (crustaceans, sea stars), simple (gastropods, sea stars) and complex (scallops, cephalopods) eyes being both lost and retained. Strategies and recommendations So, how can all these factors be recognized and evaluated on a large scale? First, comparisons between distant groups (geographically or phylogenetically) should incorporate variation within those groups, too. Despite superficially convergent eye losses, histological-level investigations consistently reveal more variation than anticipated within closely related groups (Malkowsky and Götze 2014; Sumner-Rooney et al. 2016). Phylogenetically contextualized comparative studies have been underused in the study of eye reduction, and have enormous potential to resolve evolutionary trajectories for comparison and hypothesis testing (Desuter-Grandcolas 1997; Klaus et al. 2013; Gonzalez et al. 2018). Second, if caves are “natural replicates” of evolutionary experiments, we can exploit the similarities and differences between dark ecosystems and taxa as experimental treatment groups. To incorporate the various factors outlined above, similar or identical methods should be deployed across different study groups (taxa, habitats, etc.) where possible to identify common trends. Molecular and physiological techniques are being used with great success to quantify the impacts of selection versus drift, and identify potential pleiotropic effects in agents such as Shh (Culver et al. 1995; Jeffery and Martasian 1998; Yamamoto et al. 2004, 2009; Espinasa and Jeffery 2006; Protas et al. 2008; Carlini et al. 2013). However, physiological techniques especially cannot currently be feasibly extended to multiple large groups of animals, yet such comparisons are critical to identifying truly unifying patterns in eye loss. Morphological studies can be deployed across large groups non-destructively and with relatively low costs. Coupled with continued transcriptomic, physiological, developmental, and behavioral research in model systems, there is considerable unexplored potential in comparative morphological studies. Comparative morphology and phylogenetic reconstruction The detailed study and comparison of visual and central nervous system anatomy can provide a wealth of information on evolutionary trajectories, as well as indicating to some extent the underlying mechanisms. Examining multiple lineages that have independently lost or reduced eyes provide “natural replicates” of the evolutionary process, and by applying this to multiple taxonomic groups in multiple habitat types, we can conduct a macroevolutionary experiment to evaluate the proposed conceptual framework, and test evolutionary hypotheses. It is clear that presence/absence data does not capture the extent of structural diversity in visual systems (Malkowsky and Götze 2014; Sumner-Rooney et al. 2016; Birk et al. 2018), but phylogenetic reconstruction of eye loss is a highly effective tool (Raupach et al. 2009; Syme and Oakley 2012; Gonzalez et al. 2018). By combining detailed morphological studies with phylogenetic methods, the evolutionary trajectories of individual eye characters can be reconstructed across large groups (Fig. 4A). This approach revealed variable evolutionary trajectories during multiple deep-sea invasions within a family of ecologically similar and even sympatric snails (Sumner-Rooney et al. 2016). Resolution was not sufficient to draw concrete conclusions about evolutionary processes, but supported strong roles for both selection and neutral evolution in sedentary, benthic detritivores with simple ancestral eye structures, in line with the expectations presented here. If expanded to include more taxa and more characters, including transcriptomic data where possible, this technique could facilitate both within- and between-group comparisons by charting the progression of individual cases of eye loss and reduction phylogenetically. Behavior and physiology: are there robust morphological proxies? Behavioral and physiological data are hugely desirable but not always obtainable, as animals may be too rare, inaccessible, or fragile. However, some basic indicators may be given by features such as neural investment (Banister 1984), numbers of ommatidia, focal length, aperture size, and the predicted functionality of photopigments if sequences are available. Blind mole rats, for example, are sensitive to circadian light regimes but cannot see, and this is elegantly reflected in their morphology: very reduced eyes and visual processing centers in the brain, but an intact suprachiasmatic nucleus (Zubidat et al. 2011). Similarly, differential investment in other areas of the central nervous system can indicate the importance of alternative sensory modalities, including olfaction (Cooper et al. 2001; Stöckl et al. 2016a), gustation, and mechanoreception (Franz-Odendaal and Hall 2006; Soares and Niemiller 2013). Testing evolutionary hypotheses Assembling large datasets of visual, central nervous system, and other sensory characters will enable us to identify regressive features that appear more or less frequently, recurrent associations between specific characters, habitat specificity, and the extent of reinvestment in, and trade-offs with, other sensory modes. Ancestral state reconstructions would allow researchers to story-board the evolution of eye reduction and loss in multiple replicates, multiple habitat types, and multiple ecological niches (Fig. 4). Armed with such a powerful dataset, we could test evolutionary hypotheses on an enormous scale. Differential reduction and reinvestment under varying physical and biological conditions may provide clues as to which factors dominate in different taxa and environments. Recurring features may indicate cost-effective routes to reduction; conversely, the repeated or unexpected survival of other characters may betray pleiotropic limitations, persistence of functionality, or repurposing of retained structures. These features can be used to plot trajectories across adaptive landscapes and help visualize the strength of the various evolutionary drivers (Fig. 4B). Multiple lineages convergently following the same progressive changes have likely moved through the adaptive landscape in the same way and would suggest significant constraints on evolutionary trajectory. These patterns may be consistent across biological settings, but more likely the differences between them will reveal more about the most influential factors and the more variable subjects. Beyond eye loss, this approach could help us to better understand the nature of reductive evolution overall, applied to any number of traits. Specific recommendations Study material Such large-scale comparative studies will require large amounts of specimen material and careful selection of focal groups (see below). While collecting dedicated fresh specimens is ideal, it is not always possible. Natural history collections offer an invaluable resource to supplement or substitute fresh material, covering a large range of taxa, geographical locations, and time periods (Sumner-Rooney and Sigwart 2017). As with any resource, there are limitations to using collection material; samples may be unsuitable for certain analyses due to fixation or preservation methods, age, or the need for destructive sampling. However, high-quality morphological data can sometimes be extracted from even very old ethanol-preserved samples, and spirit collections represent a major untapped resource for large comparative studies. Methods Rapid, non-destructive penetrative imaging tools, such as X-ray tomography (µ-CT and SXRT), are ideal for large-scale studies of precious specimens. Both have produced stunning results in very small visual and nervous systems (Sombke et al. 2015; Taylor et al. 2016), and improved scanners and visualization software now allow the study of eye structure in very small, unstained, wet material (L. Sumner-Rooney et al., in preparation). Taxon selection The selection of study groups requires careful consideration. For comparisons across habitats, researchers should ensure as much similarity between groups as possible in line with the considerations in Table 1. Criteria include: Multiple transitions into dark niches within a relatively low taxonomic rank. All empirical studies should incorporate multiple independent replicates, i.e., multiple lineages independently transitioning into dark habitats or niches (Tierney et al. 2017). Taxonomic rank is also important: the lower the rank, the less likely that confounding phyletic factors are introduced. In cases such as Cicurina and Astyanax, multiple cave invasions have occurred within genera or species (Wilkens and Strecker 2003; Hedin 2015), but in others, family-level analyses are required to capture sufficient replicates (Syme and Oakley 2012; Sumner-Rooney et al. 2016). Existence of, or ability to produce, a robust phylogeny. Resolved phylogenetic relationships are fundamental to any comparative study. If relationships between taxa are not clear, trait history reconstructions are ambiguous and character evolution cannot be reliably traced (Harvey and Pagel 1991). Inclusion of sighted taxa. Sighted relatives must be included in studies of loss, especially when examining large groups of eyeless species; without them, assigning character polarity and the resolution of fine evolutionary changes between reductive lineages are problematic. Multiple types of habitat transition. Comparisons between different habitat shifts are a major potential avenue for future research, but identifying suitable candidate study systems poses a major obstacle. Relatively few taxonomic groups have undergone transitions from, for example, both shallow to deep water and from surface to subterranean habitats in multiple lineages. Conclusions After more than 150 years of study, the biological conundrum of eye loss is firmly back in the research spotlight, and modern methodologies offer real possibilities of collecting large datasets across broad taxonomic ranges. Armed with these new tools, researchers can truly close in on the evolutionary patterns underlying loss. A basic evolutionary framework attempts to incorporate the fundamental evolutionary forces acting on species that transition into dark habitats, with the aim of facilitating future development and guiding future research. We urgently need to survey the morphological variety observed in dark-living animals on a broader scale, alongside continued high-resolution developmental and transcriptomic studies on specific species. Harnessing this variation using the comparative method across a large range of animals and habitats will help us identify the fundamental influences on the evolution of eye loss on a global scale. Acknowledgments I am very grateful to the Company of Biologists, Palaeontological Association (PA-GA201707), American Microscopical Society, Crustacean Society, and Society for Integrative and Comparative Biology (divisions DEDB, DNB, DPCB, DEE, and DIZ) for their support of this symposium. I would also like to thank my co-organizer Megan Porter, Mark Wilkinson and Bill Jeffery for the use of their images in Fig. 1, and two anonymous reviewers and the editors for their constructive feedback on the manuscript. Funding This work was supported by the Oxford University Museum of Natural History. References Banister KE. 1984 . A subterranean population of Garra barreimiae (Teleostei: Cyprinidae) from Oman, with comments on the concept of regressive evolution . J Nat Hist 18 : 927 – 38 . Google Scholar Crossref Search ADS Barr TC. 1968 . Cave ecology and the evolution of troglobites. In: Dobzhansky T , Hecht MK , Steere WC , editors. Evolutionary biology. New York (NY) : Plenum Press . p. 35 – 101 . Birk MH , Blicher ME , Garm A. 2018 . Deep-sea starfish from the Arctic have well-developed eyes in the dark . Proc R Soc B 285 : 20172743. Google Scholar Crossref Search ADS PubMed Bloom T , Binford GA , Esposito L , Garcia GA , Peterson I , Nishida A , Loubet-Senear K , Agnarsson I. 2014 . Discovery of two new species of eyeless spiders within a single Hispaniola cave . J Arachnol 42 : 148 – 54 . Google Scholar Crossref Search ADS Borowsky R. 2013 . Evolution of an adaptive behavior and its sensory receptors facilitates eye regression in blind cavefish . BMC Biol 11 : 81 – 4 . Google Scholar Crossref Search ADS PubMed Brirten RJ. 1984 . Sequence evolution differ between taxonomic groups of DNA . Science 39 : 1393 – 8 . Carlini DB , Fong DW. 2017 . The transcriptomes of cave and surface populations of Gammarus minus (Crustacea: Amphipoda) provide evidence for positive selection on cave downregulated transcripts . PLoS One 12 : e0186173. Google Scholar Crossref Search ADS PubMed Carlini DB , Satish S , Fong DW. 2013 . Parallel reduction in expression, but no loss of functional constraint, in two opsin paralogs within cave populations of Gammarus minus (Crustacea: Amphipoda) . BMC Evol Biol 13 : 89. Google Scholar Crossref Search ADS PubMed Childress JJ. 1995 . Are there physiological and biochemical adaptations of metabolism in deep-sea animals? Trends Ecol Evol 10 : 30 – 6 . Christiansen K. 1961 . Convergence and parallelism in cave Entomobryinae . Evolution 15 : 288. Google Scholar Crossref Search ADS Clarkson E , Levi-setti R , Horva G. 2006 . The eyes of trilobites: the oldest preserved visual system . Arthropod Struct Dev 35 : 247 – 59 . Google Scholar Crossref Search ADS PubMed Cooper RL , Li H , Long LY , Cole JL , Hopper HL. 2001 . Anatomical comparisons of neural systems in sighted epigean and troglobitic crayfish species . J Crustac Biol 21 : 360 – 74 . Google Scholar Crossref Search ADS Culver DC , Kane TC , Fong DW. 1995 . Adaptation and natural selection in caves: the evolution of Gammarus minus . Cambridge (MA) : Harvard University Press . Culver DC , Pipan T. 2009 . The biology of caves and other subterranean habitats . Oxford : Oxford University Press . Danielopol DL, Baltanás A, Bonaduce G. 1996. The darkness syndrome in subsurface-shallow and deep-sea dwelling Ostracoda (Crustacea). In: Uiblein F, Ott J, Stachowitsch M, editors. Deep-sea and extreme shallow-water habitats: affinities and adaptations. Vienna: Austrian Academy of Sciences. p. 123–43. Darwin C. 1859 . On the origin of species . London : John Murray . David-Gray ZK , Janssen JWH , Degrip WJ , Nevo E , Foster RG. 1998 . Light detection in a ‘blind’ mammal . Nat Neurosci 1 : 655 – 6 . Google Scholar Crossref Search ADS PubMed Derkarabetian S, Steinmann DB, Hedin M. 2010. Repeated and time-correlated morphological convergence in cave-dwelling harvestmen (Opiliones, Laniatores) from Montane Western North America. PLoS One 5:e10388. Desuter-Grandcolas L. 1997 . Studies in cave life evolution: a rationale for future theoretical developments using phylogenetic inference . J Zool Syst Evol Res 35 : 23 – 31 . Google Scholar Crossref Search ADS Emerling CA. 2018 . Regressed but not gone: patterns of vision gene loss and retention in subterranean mammals . Integr Comp Biol (doi: 10.1093/icb/icy004). Espinasa L , Jeffery WR. 2006 . Conservation of retinal circadian rhythms during cavefish eye degeneration . Evol Dev 8 : 16 – 22 . Google Scholar Crossref Search ADS PubMed Fišer Ž , Novak L , Luštrik R , Fišer C. 2016 . Light triggers habitat choice of eyeless subterranean but not of eyed surface amphipods . Naturwissenschaften 103 : 7. Google Scholar Crossref Search ADS PubMed Franz-Odendaal TA , Hall BK. 2006 . Modularity and sense organs in the blind cavefish, Astyanax mexicanus . Evol Dev 8 : 94 – 100 . Google Scholar Crossref Search ADS PubMed Friedrich M , Chen R , Daines B , Bao R , Caravas J , Rai PK , Zagmajster M , Peck SB. 2011 . Phototransduction and clock gene expression in the troglobiont beetle Ptomaphagus hirtus of Mammoth cave . J Exp Biol 214 : 3532 – 41 . Google Scholar Crossref Search ADS PubMed Fumey J , Hinaux H , Noirot C , Thermes C , Rétaux S , Casane D. 2018 . Evidence for late Pleistocene origin of Astyanax mexicanus cavefish . BMC Evol Biol 18 : 1 – 19 . Google Scholar Crossref Search ADS PubMed Gibert J, Deharveng L. 2002. Subterranean ecosystems: a truncated functional biodiversity. Bioscience 52:473–81. Gonzalez BC , Worsaae K , Fontaneto D , Martínez A. 2018 . Anophthalmia and elongation of body appendages in cave scale worms (Annelida: Aphroditiformia) . Zool Scr 47 : 106 – 21 . Google Scholar Crossref Search ADS Gross JB , Furterer A , Carlson BM , Stahl BA. 2013 . An integrated transcriptome wide analysis of cave and surface dwelling Astyanax mexicanus . PLoS One 8 : e55659. Google Scholar Crossref Search ADS PubMed Harvey PH , Pagel MD. 1991 . The comparative method in evolutionary biology. Oxford : Oxford University Press . Hedin M. 2015 . High-stakes species delimitation in eyeless cave spiders (Cicurina, Dictynidae, Araneae) from central Texas . Mol Ecol 24 : 346 – 61 . Google Scholar Crossref Search ADS PubMed Hinaux H , Poulain J , da Silva C , Noirot C , Jeffery WR , Casane D , Rétaux S. 2013 . De novo sequencing of Astyanax mexicanus surface fish and Pachón cavefish transcriptomes reveals enrichment of mutations in cavefish putative eye genes . PLoS One 8 : e53553. Google Scholar Crossref Search ADS PubMed Izutsu M, Toyoda A, Fujiyama A, Agata K, Fuse N. 2016. Dynamics of dark-fly genome under environmental selections. G3 Genes Genom Genet 6:365–76. Jeffery WR. 2009 . Regressive evolution in Astyanax cavefish . Annu Rev Genet 43 : 25 – 47 . Google Scholar Crossref Search ADS PubMed Jeffery WR , Martasian DP. 1998 . Evolution of eye regression in the cavefish Astyanax: apoptosis and the Pax-6 gene . Integr Comp Biol 38 : 685 – 96 . Johnsen S , Frank TM , Haddock SHD , Widder EA , Messing CG. 2012 . Light and vision in the deep-sea benthos: i. Bioluminescence at 500–1000 m depth in the Bahamian islands . J Exp Biol 215 : 3335 – 43 . Google Scholar Crossref Search ADS PubMed Johnson ML , Shelton PMJ , Gaten E. 2000 . Temporal resolution in the eyes of marine decapods from coastal and deep-sea habitats . Mar Biol 136 : 243 – 8 . Google Scholar Crossref Search ADS Jones R , Culver DC. 1989 . Evidence for selection on sensory structures in a cave population of Gammarus minus . Evolution 43 : 688 – 93 . Google Scholar Crossref Search ADS PubMed Juan C , Guzik MT , Jaume D , Cooper SJB. 2010 . Evolution in caves: Darwin’s ‘wrecks of ancient life’ in the molecular era . Mol Ecol 19 : 3865 – 80 . Google Scholar Crossref Search ADS PubMed Kim BM , Kang S , Ahn DH , Kim JH , Ahn I , Lee CW , Cho JL , Min GS , Park H. 2017 . First insights into the subterranean crustacean Bathynellacea transcriptome: transcriptionally reduced opsin repertoire and evidence of conserved homeostasis regulatory mechanisms . PLoS One 12 : e0170424 – 2 . Google Scholar Crossref Search ADS PubMed Kimura M. 1968 . Evolutionary rate at the molecular level . Nature 217 : 624 – 6 . Google Scholar Crossref Search ADS PubMed Klaus S , Mendoza JCE , Liew JH , Plath M , Meier R , Yeo DCJ. 2013 . Rapid evolution of troglomorphic characters suggests selection rather than neutral mutation as a driver of eye reduction in cave crabs . Biol Lett 9 : 20121098. Google Scholar Crossref Search ADS PubMed Land MF, Nilsson D–E. 2012. Animal eyes. 2nd ed. Oxford: Oxford University Press. Langille BL , Tierney SM , Austin AD , Humphreys WF , Cooper SJB. 2018 . How blind are they? Phototactic responses in stygobiont diving beetles (Coleoptera: Dytiscidae) from calcrete aquifers of Western Australia . Aust Entomol Soc published online (doi:10.1111/aen.12330). Lerosey-Aubril R. 2006 . Ontogeny of Drevermannia and the origin of blindness in Late Devonian proetoid trilobites . Geol Mag 143 : 89 – 104 . Google Scholar Crossref Search ADS Leung NY , Montell C. 2017 . Unconventional roles of opsins . Annu Rev Cell Dev Biol 33 : 241 – 64 . Google Scholar Crossref Search ADS PubMed Malkowsky Y , Götze M-C. 2014 . Impact of habitat and life trait on character evolution of pallial eyes in Pectinidae (Mollusca: Bivalvia) . Org Divers Evol 14 : 173 . Google Scholar Crossref Search ADS Meng F , Braasch I , Phillips JB , Lin X , Titus T , Zhang C , Postlethwait JH. 2013 . Evolution of the eye transcriptome under constant darkness in Sinocyclocheilus cavefish . Mol Biol Evol 30 : 1527 – 43 . Google Scholar Crossref Search ADS PubMed Meyer-Rochow VB , Liddle AR. 1988 . Structure and function of the eyes of two species of opilionid from New Zealand glow-worm caves (Megalopsalis tumida: Palpatores, and Hendea myersi cavernicola: Laniatores) . Proc R Soc B Biol Sci 233 : 293 – 319 . Google Scholar Crossref Search ADS Mohun SM , Davies WL , Bowmaker JK , Pisani D , Himstedt W , Gower DJ , Hunt DM , Wilkinson M. 2010 . Identification and characterization of visual pigments in caecilians (Amphibia: Gymnophiona), an order of limbless vertebrates with rudimentary eyes . J Exp Biol 213 : 3586 – 92 . Google Scholar Crossref Search ADS PubMed Mohun SM , Wilkinson M. 2015 . The eye of the caecilian Rhinatrema bivittatum (Amphibia: Gymnophiona: Rhinatrematidae) . Acta Zool 96 : 147 – 53 . Google Scholar Crossref Search ADS Moran D , Softley R , Warrant EJ. 2015 . The energetic cost of vision and evolution of eyeless Mexican cavefish . Sci Adv 1 : e1500363. Google Scholar Crossref Search ADS PubMed Niemiller ML , Fitzpatrick BM , Shah P , Schmitz L , Near TJ. 2013 . Evidence for repeated loss of selective constraint in rhodopsin of Amblyopsid cavefishes (Teleosti: Amblyopsidae) . Evolution 67 : 732 – 48 . Google Scholar Crossref Search ADS PubMed Nilsson D-E. 2013 . Eye evolution and its functional basis . Vis Neurosci 30 : 5 – 20 . Google Scholar Crossref Search ADS PubMed Niven JE , Laughlin SB. 2008 . Energy limitation as a selective pressure on the evolution of sensory systems . J Exp Biol 211 : 1792 – 804 . Google Scholar Crossref Search ADS PubMed Ohta T. 1972. Population size and rate of evolution. J Mol Evol 1:305–14 Ohta T. 1992 . The nearly neutral theory of molecular evolution . Annu Rev Ecol Syst 23 : 263 – 86 . Google Scholar Crossref Search ADS Poulson TL. 2001 . Adaptations of cave fishes with some comparisons to deep-sea fishes . Environ Biol Fish 62 : 345 – 64 . Google Scholar Crossref Search ADS Poulson TL. 1963 . Cave adaptation in amblyopsid fishes . Am Midl Nat 70 : 257 – 90 . Google Scholar Crossref Search ADS Protas M , Conrad M , Gross JB , Tabin C , Borowsky R. 2007 . Regressive evolution in the Mexican cave tetra, Astyanax mexicanus . Curr Biol 17 : 452 – 4 . Google Scholar Crossref Search ADS PubMed Protas M , Tabansky I , Conrad M , Gross JB , Vidal O , Tabin CJ , Borowsky R. 2008 . Multi-trait evolution in a cave fish, Astyanax mexicanus . Evol Dev 10 : 196 – 209 . Google Scholar Crossref Search ADS PubMed Raupach MJ , Mayer C , Malyutina M , Wagele J-W. 2009 . Multiple origins of deep-sea Asellota (Crustacea: Isopoda) from shallow waters revealed by molecular data . Proc R Soc B Biol Sci 276 : 799 – 808 . Google Scholar Crossref Search ADS Rétaux S , Casane D. 2013 . Evolution of eye development in the darkness of caves: adaptation, drift, or both? Evodevo 4 : 26. Google Scholar Crossref Search ADS PubMed Sadoglu P. 1967 . The selective value of eye and pigment loss in mexican cave fish . Evolution 21 : 541 – 9 . Google Scholar Crossref Search ADS PubMed Soares D , Niemiller ML. 2013 . Sensory adaptations of fishes to subterranean environments . Bioscience 63 : 274 – 83 . Google Scholar Crossref Search ADS Sombke A , Lipke E , Michalik P , Uhl G , Harzsch S. 2015 . Potential and limitations of X-ray micro-computed tomography in arthropod neuroanatomy: a methodological and comparative survey . J Comp Neurol 523 : 1281 – 95 . Google Scholar Crossref Search ADS PubMed Stahl BA , Gross JB , Speiser DI , Oakley TH , Patel NH , Gould DB , Protas ME. 2015 . A transcriptomic analysis of cave, surface, and hybrid isopod crustaceans of the species Asellus aquaticus . PLoS One 10 : e0140484. Google Scholar Crossref Search ADS PubMed Stöckl A , Heinze S , Charalabidis A , El Jundi B , Warrant E , Kelber A. 2016 . Differential investment in visual and olfactory brain areas reflects behavioural choices in hawk moths . Sci Rep 6 : 26041. Google Scholar Crossref Search ADS PubMed Stöckl AL , Ribi WA , Warrant EJ. 2016 . Adaptations for nocturnal and diurnal vision in the hawkmoth lamina . J Comp Neurol 524 : 160 – 75 . Google Scholar Crossref Search ADS PubMed Sumner-Rooney LH , Sigwart JD , McAfee J , Smith L , Williams ST. 2016 . Repeated eye reduction events reveal multiple pathways to degeneration in a family of marine snails . Evolution 70 : 2268 – 95 . Google Scholar Crossref Search ADS PubMed Sumner-Rooney L , Sigwart JD. 2017 . Lazarus in the museum: resurrecting historic specimens through new technology . Invertebr Zool 14 : 73 – 84 . Syme AE , Oakley TH. 2012 . Dispersal between shallow and abyssal seas and evolutionary loss and regain of compound eyes in cylindroleberidid ostracods: conflicting conclusions from different comparative methods . Syst Biol 61 : 314 – 36 . Google Scholar Crossref Search ADS PubMed Taylor GJ , Ribi W , Bech M , Bodey AJ , Rau C , Steuwer A , Warrant EJ , Baird E. 2016 . The dual function of orchid bee ocelli as revealed by X-ray microtomography . Curr Biol 26 : 1319 – 24 . Google Scholar Crossref Search ADS PubMed Theobald JC , Greiner B , Wcislo WT , Warrant EJ. 2006 . Visual summation in night-flying sweat bees: a theoretical study . Vision Res 46 : 2298 – 309 . Google Scholar Crossref Search ADS PubMed Tierney SM , Cooper SJB , Saint KM , Bertozzi T , Hyde J , Humphreys WF , Austin AD , Tierney SM. 2015 . Opsin transcripts of predatory diving beetles: a comparison of surface and subterranean photic niches . R Soc Open Sci 2 : 140386. Google Scholar Crossref Search ADS PubMed Tierney SM , Friedrich M , Humphreys WF , Jones TM , Warrant EJ , Wcislo WT. 2017 . Consequences of evolutionary transitions in changing photic environments . Austral Entomol 56 : 23 – 46 . Google Scholar Crossref Search ADS Valdez-Lopez JC , Donohue MW , Bok MJ , Wolf J , Cronin TW , Porter ML. 2018 . Sequence, structure, and expression of opsins in the monochromatic stomatopod Squilla empusa . Integr Comp Biol published online (doi:10.1093/icb/icy007/4985722). Warrant EJ. 1999 . Seeing better at night: life style, eye design and the optimum strategy of spatial and temporal summation . Vision Res 39 : 1611 – 30 . Google Scholar Crossref Search ADS PubMed Warrant EJ , Collin SP , Locket NA. 2003 . Eye design in deep-sea fishes. In: Collin SP , Marshall NJ , editors. Sensory processing in aquatic environments. New York (NY ): Springer . p. 303 – 22 . Warrant EJ , Locket NA. 2004 . Vision in the deep sea . Biol Rev Camb Philos Soc 79 : 671 – 712 . Google Scholar Crossref Search ADS PubMed Wilkens H. 1988 . Evolution and genetics of epigean and cave Astyanax fasciatus (Characidae, Pisces) support for the neutral mutation theory . Evol Biol 23 : 271 – 367 . Wilkens H , Strecker U. 2003 . Convergent evolution of the cavefish Astyanax (Characidae, Teleostei): genetic evidence from reduced eye-size and pigmentation . Biol J Linn Soc 80 : 545 – 54 . Google Scholar Crossref Search ADS Wilkens H , Strecker U. 2017 . Evolution in the dark: Darwin’s loss without selection . Berlin : Springer . Williams ST , Smith LM , Herbert DG , Marshall BA , Warén A , Kiel S , Dyal P , Linse K , Vilvens C , Kano Y. 2013 . Cenozoic climate change and diversification on the continental shelf and slope: evolution of gastropod diversity in the family Solariellidae (Trochoidea) . Ecol Evol 3 : 887 – 917 . Google Scholar Crossref Search ADS PubMed Wong JM , Pérez-Moreno JL , Chan TY , Frank TM , Bracken-Grissom HD. 2015 . Phylogenetic and transcriptomic analyses reveal the evolution of bioluminescence and light detection in marine deep-sea shrimps of the family Oplophoridae (Crustacea: Decapoda) . Mol Phylogenet Evol 83 : 278 – 92 . Google Scholar Crossref Search ADS PubMed Yamamoto Y , Byerly MS , Jackman WR , Jeffery WR. 2009 . Pleiotropic functions of embryonic sonic hedgehog expression link jaw and taste bud amplification with eye loss during cave fish evolution . Dev Biol 330 : 200 – 11 . Google Scholar Crossref Search ADS PubMed Yamamoto Y , Stock DW , Jeffery WR. 2004 . Hedgehog signalling controls eye degeneration in blind cavefish . Nature 431 : 844 – 7 . Google Scholar Crossref Search ADS PubMed Yoshizawa M , Jeffery WR. 2008 . Shadow response in the blind cavefish Astyanax reveals conservation of a functional pineal eye . J Exp Biol 211 : 292 – 9 . Google Scholar Crossref Search ADS PubMed Yoshizawa M , Yamamoto Y , O’Quin KE , Jeffery WR. 2012 . Evolution of an adaptive behavior and its sensory receptors promotes eye regression in blind cavefish . BMC Biol 10 : 108. Google Scholar Crossref Search ADS PubMed Zubidat AE , Nelson RJ , Haim A. 2011 . Spectral and duration sensitivity to light-at-night in “blind” and sighted rodent species . J Exp Biol 214 : 3206 – 17 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Integrative and Comparative BiologyOxford University Press

Published: Sep 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off