For a nicely-formatted PDF of this post, click here!
Most animals have receptors allowing them to detect light – photoreceptors. Depending on how advanced the photoreceptors are, this can allow them to sense just the presence of light, going through to the direction of light, to the highest levels allowing animals to see predators and other objects. We can distinguish between several types of photoreceptors. The most general of these distinctions is whether the photoreceptor is ciliary or rhabdomeric, the difference being whether the light-sensitive membrane derives from cilia (ciliary; e.g. in vertebrates) or whether many microvilli of apical cells form a rhabdom (e.g. flatworms, molluscs, arthropods). Keep in mind though that these types coexist in most organisms; this is why there is such a bewildering array of photoreceptor types, from isolated photoreceptor cells to simple eyespots to proper eyes with lenses. Some photoreceptors have no real morphological features: the retinal ganglion cells in vertebrates and photoreceptors in the neural ganglia of several invertebrates are examples of these. We do not know how (or if!) they work.
Physiologically, ciliary photoreceptors generally hyperpolarise light, while rhabomeric ones depolarise it; exceptions abound though. Unambiguous distinctions can be obtained by looking at the biochemical differences, namely what they use for phototransduction. Phototransduction is how light information is changed to electrical signals to be transferred by the neurons. Ciliary photoreceptors all share a cyclic-nucleotide motif, whereas rhabdomeric ones all use a phopholipase C (PLC) motif.
All photoreceptors share a common visual pigment: a mixture of a vitamin A-based chromophore (retinal) and a seven-transmembrane-helix apoprotein (opsin). Together, they form rhodopsin. The pigments belong to a special class of receptors called G protein-coupled receptors, characterised by signaling through heterotrimeric G proteins. In vertebrates, the rhodopsin is thermally unstable, breaking down with light into the opsin and chromophore; rhodopsins are maintained by an enzymatic process. Invertebrate rhodopsins are stable, although renewal also must take place there.
Different animals have different opsins (variations of the same theme; over 1000 identified so far) and it is largely believed that opsin was present in the last common ancestor of the animals. Bacteria also have opsins (bacteriorhodopsin and channelopsins), which are structurally similar. There are two major groups of animal opsins, corresponding to the two types of photoreceptors: c-opsins (ciliary) and r-opsins (rhabdomeric). There are also minor, lineage-specific opsin groups, such as the G-opsins (cephalochordates and molluscs), pteropsins (insects), encephalopsin/teleost multiple tissue opsin and photoisomerase (regenerate the chromophore).
What must be said before going through the animal kingdom is that eyes are polyphyletic (Salwini-Plawen & Mayr, 1977). In the name of fairness though, I must mention that some have postualted a monophyletic origin (i.e. that the eyes of annelids, arthropods and vertebrates trace their origin back to a common ancestor) (Vanfleteren, 1982), or diphyletic (Eakin, 1982). We will scratch the surface of the debate later on.
Before going through the animal kingdom (in no particular order), we still need to define what an ‘eye’ is. Note that I’ll be using the terms eye and photoreceptor interchangeably in the post (photoreceptor is just too long for my poor fingers). Generally, we classify eyes on a scale of complexity. The absolute simplest eye is what’s found in unicellular algae or turbellarian epidermal eyes: one single cell with both a photosensitive and a shading pigment (there is a simpler version that does not have shading pigments, such as in the deep brain photoreceptors I’ll mention later, but I won’t count them as eyes). The next step is two-celled eyes: one photoreceptor and one pigment cell (as found in trochophore larvae and planarians). The photoreceptor cell is the one containing rhodopsin: its membrane is folded to enhance the amount of areas in contact with light. The pigment cell contains pigments that aren’t photosensitive, but instead absorb light, such as melanins or pterins. Also, they are the ones that form the lens.
These two-celled eyes are the simplest kinds of ocelli. Ocelli are simply multicellular arrangements of photoreceptor and pigment cells. They can sometimes form what is known as a pigment-cup eye, which can detect the direction of light. This pigment cup then forms the basis for the next step. The eye can be inverse (i.e. the folding of the photorecptor points towards the cup) or everse (the foldings point towards the light); there are also transitional forms between these two types.
The eye considered to be the most complex is the complex eye of vertebrates and cephalopods, with a cornea, iris and retina. These complex eyes are classic examples of functional convergence – they are not homologous (that would imply reduction in every other animal group – not a parsimonious explanation).
In total, there are 8 types of eye, summarised in the pictures above (Fernald, 2006). A is a pit eye and uses shadows, as does B. The pit eye led to the lensed, refraction-dependent eyes in C and D. B led to the refraction-dependent apposition compound eye (also found in bees, crabs and fruit flies) and to the reflection-dependent superposition eye of shrimp and lobsters (H). G also works on the basis of reflection.
First, let’s look at the eyes of vertebrates (picture above, Sharma, 2007). Everyone knows from high school biology that there are only two photoreceptors in mammals: rods and cones. This is now outdated, as a third class has been discovered: photosensitive retinal ganglion cells. Rods are responsible for low-light vision. The outer segment of the rod cell is packed full of discs with the pigment rhodopsin. Here’s what happens in phototransduction in rods: photoismoerised rhodopsin (i.e. a structural variant caused by light) activates a G-protein in the membrane called transducin. Transducin then activates an enzyme of the phosphodiesterase class which hydrolises cGMP (the same cGMP you should recognise from Biochemistry 101). So how does this translate to light information? Unhydrolised cGMP forms a gate for a nonselective cation channel in the cell membrane – it is through it that cations pass and since they are nonselective, anything can go through; this is called the “dark current” and it keeps the cell depolarised, sustaining continuous glutamate release. As cGMP gets hydrolised, the gate closes (gradually, like a portcullis); the cell then becomes hyperpolarised and glutamate is not released anymore. And that’s basically it: once glutamate stops getting released, the signal gets transmitted to the optic nerve.
This is one example of complexity in biochemistry. Unlike those intelligent design morons (case in point: Behe 1996 Darwin’s Black Box), we understand that the components of the cellular machinery evolved and we have ways of elucidating this evolution. To outline it, let’s examine phototransduction in all the animals. The vertebrate model above is an example of a cyclic nucleotide gated (CNG)-based phototransduction. In non-deuterostomian animals, a different cascade is present (see picture above, Montell, 1999), employing transient receptor potential channels (TRPC), which allow depolarisation to happen (as opposed to hyperpolarisation as above) – both of these refer to passages or doors through which ions can or cannot flow (no magic here, just basic chemistry). Basically, when light is absorbed, the PLC pathway opens up this TRPC and allows calcium ions to flow through. The picture below (Tong et al., 2009) is a simpler summary of both phototransduction pathways.
Which one of these originated first though? Or are they both variants of a non-extinct phototransduction cascade? Are they too complex to have evolved? No. In fact, it’s quite clear: the TRPC is a variant of CNG phototransduction. The change can be elegantly followed even. It takes mutating the G protein to Gα-q, changing the physiology of the cell membrane and causing the new type of ion flow. The CNG-based system was already present before (light-sensing is ubiquitous in all life). There really is no mystery: while the mutations may have proved fatal, because the building blocks from the CNG-system were already there (because they evolved millions of years before), the new TRPC system actually worked.
Birds have photoreceptors in their hypothalamus, of all places. They are what respond to day length and play into circadian rhythms. This placement of photoreceptors in the brain may sound surprising, but has been known since 1911, when Karl von Frisch noticed that minnows (who usually change their skin colour in response to light) with no eyes and no pineal gland gland still get the skin colour change, meaning there must be some other photoreceptor (it turned out to be in the diencephalon). The reason it works is because the skull of birds is rather fragile and thin, allowing light to go through.
Such extraocular (i.e. outside the eye) photoreceptors are also known from other organisms, including many invertebrates. This is best seen in aphids, where one of the morphs (aphids are highly-polymorphic, depending on the generation and season) has no ocelli, but still responds to light cues (counting the length of the days is the only way to induce reproduction in them). In mammals, they are found in the neurons of the retina and play a role in transmitting non-image-forming visual informaion. They use melanin, not opsin, and their role is now being intensely studied, but is most likely to do with circadian rhythms. In fact, they are present and active from a much earlier time than rods and cones in mice, being active from birth, not at 10 days after birth (Sekaran et al., 2005).
It’s known that, in general, vision plays a large role in controlling and regulating circadian rhythms in all organisms, but my knowledge of the field is too broken to even summarise how. Sorry :(
Box jellyfish (Cubozoa) are renowned for their eyes, which they need for their lifesyle: unlike most jellyfish, they are active swimmers and actively seek out their prey. Their visual system consists of four sensory structures (rhopalia), each with six eyes (total: 24 eyes). There are four types of eye: there can be paired slit eyes, paired pit eyes, an upper lens and a lower lens; see the picture above (O’Connor et al., 2010). The photoreceptors, in contrast to most invertebrates, are cilliary. The reason they are famous is that it is supposedly unexpected for such ‘basal’ animals to have such complex structures: it’s a good counterpoint to the fallacious notion of “evolutionary trends”.
But that’s not all: of the five cnidarian classes, four have species with eyes: the afore-mentioned cubozoans, the hydrozoans, scyphozoans and staurozoans – all the classes that have a distinct medusan stage (the missing class is the Anthozoa, the corals). From a phylogenetic point of view (see thought experiment at the end of the post), Pax gene expression is interesting in the cnidarians. First, some background into Pax: it’s a gene family encoding for transcription factors important for development in all metazoans. There are five subfamilies: Pax-4/6, Pax-2/5/8, Pax-1/9, Pax-3/7 and pox neuro, all originating by gene duplication before the cnidarian-bilaterian split. Pax-6 is an integral component of the developmental cascade leading to eye development in many bilaterians, leading to the consideration that this is a conserved status present in the last common ancestor of the bilaterians, concluding that eyes in bilaterians are all homologous (i.e. that arthropod compound eyes and vertebrate camera eyes are derived from the same ancestral structure). Since this is still a point of debate, and I quite strongly disagree with it, I will not bias the post by stating an intellectually dishonest opinion. To get back on track, cnidarians also have Pax genes, but do not use Pax-6 for eye development, instead using members of other subfamilies. This definitely means that Pax genes are functionally flexible. But it does raise some points about the whole Pax-6 = homologous eyes scenario, since the flexibility means that the different bilaterian phyla converged on Pax-6 for the same function because it was ideally suited for forming eyes. Food for thought. The picture above (Kozmik et al., 2008) and its caption shows how both convergence and common ancestry fit the current data.
Now we can move to the arthropods, in no particular order, starting with the insects. Even though they’re small and their brains have relatively few neurons, insects have very sophisticated visual systems that allow them to recognise conspecifics, predators, food and prey; they can memorise landmarks and use them as navigation aids. Additionally, most can do this at night – since most insects are nocturnal animals. This night vision isn’t only achieved by specialisations in their visual system though: they move more slowly (allowing more light to go in), for example. But most of it is due to adaptation. Most nocturnal insects have superposition compound eyes, where the ommatidia are bunched together (except for some exceptions with apposition compound eyes, where the ommatidia are more or less isolated; see picture above (Warrant et al, 2011)). This eye type is specialised for a high sensitivity.
Flies have compound eyes with up to thousands of facet lenses. Each lens focuses light on the rhabdomeres of the eight photoreceptors (R1-R6 are peripheral and achromatic, R7-R8 are central and chromatic). This is when phototransduction takes place. Fly pigments have a special chromophore called 3-hydroxyretinal that connects to the opsin of the peripheral photoreceptors, absorbing mainly blue-green (and a bit of UV; 500 nm) light. Additionally, the peripheral rhodopsins can bind to 3-hydroxyretinol (the alcohol of 3-hydroxyretinal) to give flies UV vision (360 nm).
It’s relevant to bring up a study that made quite a splash late last year: Xiang et al., 2010. It was surprising mostly because Drosophila is one of the best-studied organisms, yet it was only now that they discovered a whole new class of photoreceptors in the larva, specifically in the body wall. Their function is to help in avoiding light: in the wild, the larvae are burrowing animals – now we know that when they sense light on their body, they know they’re on the surface and dig right back down. This just highlights that while we may think we know everything about certain organisms (certainly true for the model organisms), there will always be surprises lurking (right beneath the epidermis, in this case).
Similarly, a model organism for just about every field of biology, the nematode Caenorhabditis elegans, has no eyes, but uses photoreceptors along its body to detect light and stay underground.
Such extraocular photoreceptors aren’t always diffuse and spread throughout the body. The photoreceptive vesicles and nuchal organs of cephalopods are concentrations of extraocular photoreceptors at specific locations on the body. In squid, the vesicles are used for counterillumination, a camouflage technique in which the colour of the underbody is made to match the light from above: the vesicles on the dorsal side detect the light spectrum and thus the chromophores/bioluminescent bacteria on the ventral side know what colour to take on.
Cockroaches, as we all know, are nocturnal insects. They have apposition eyes, though, as well as several noteable differences in their photoreceptors (and behaviour and morphology) compared to other insects, that have made them ideally suited for being active in dim light. The ommatidia themselves are irregular, resulting in a not-so-stellar image. However, it’s known that cockroaches are dependent on vision for their behaviour and for guiding their antennae. What makes their visual system efficient is in fact its variability in response. The signals from the photoreceptors are seemingly pooled at random – instead of muddling up the image, this results in an increase in the total information that can be extracted: instead of having all similar ommatidia doing their own thing, all the types are forced to work together, so to speak, and this results in a large pool of slightly different images that are then combined.
The oriental hornet has a cuticle that could be termed photoreceptive. Specifically, the yellow segments contain pterins (the brown ones contain melanin), and they are actually functional for sensing light. The hornet is a social insect and builds nests on the ground in sunlight (true correlation: when in the sun, the hornet digs more) – they detect the insolation level with the pterins. Not only that, pterins absorb some light from the UV spectrum – as do photosynthetic plants. I’m not simply drawing weird analogies here: hornets actually do harvest solar energy – their cuticle is folded in such a way that light gets reflected stronger in the pterin-containing segments, concentrating the light there and actually producing an energy boost. It’s nowhere near as powerful as photosynthesis (itself extremely wasteful anyway), but every bit counts – this could also explain the correlation between digging activity and insolation.
Butterflies are something of a model taxon for studying sexual dimorphism and its evolution, and this is due to their dichromatism: males and females vary in wing colour – even when we don’t see it, as the variations extend to spectrum ranges not resolved by our eyes (UV). Additionally, many lepidopteran (butterfly + moth) eyes are also sexually dimorphic, and males and females see differently. Basically, the situation is more complex than ‘shiny male = good’ – the females evolved specifically to be able to see the shininess of the male, and this evolution has to also be investigated. As a basic ground plan, unlike flies, lepidopterans have 9 photoreceptor types in their compound eyes. They have at least three visual pigments absorbing UV (350 nm), blue (440 nm) and long wavelength (530 nm) (Briscoe & Chittka, 2001). However, many variations on the theme exist, for example the most primitive butterfly family, the Papillionidae, has undergone two duplications of its opsin genes, leading to more colours being absorbable.
The classical example for sexual dimorphism in photoreceptors comes from another popular insect, the regular housefly. Males have to chase females in order to mate, and this has led to their having a highly-tuned tracking system. They have a special region on the fronto-dorsal side of their compound eyes called the love spot – females don’t have this – which they use to track the female while pursuing them from underneath. The love spot is characterised by having enlarged ommatidial facets, meaning higher sensitivity. Additionally, R1-R7 contribute to the achromatic pathway, meaning that darker objects (i.e. females) are easier to detect. Finally, the love spot is innervated specially to maximise chasing behaviour. This is simply a testament to the power of sexual selection.
This brings up another aspect of insect vision: motion detection. Anything that’s flying and hovering like that has to have some sort of stabilising mechanisms so the image doesn’t jitter all the time. The classical experiment to investigate the most basic aspect is to have a tethered fly fly inside a rotating drum. The fly will fly in the same direction that the drum is rotating (if it’s rotating clockwise, it will generate some yaw in a clockwise direction) – transferring this to nature, replace the drum with wind, and what this means is that the fly will compensate for a gust of wind by turning in the direction the wind is blowing, thereby negating the destabilising effect. Flies, of course, don’t fly forever, they have to land at some point. Landing is also visually mediated: they measure how fast their landing spot is expanding as they approach it – once the rate reaches a specific threshold, they decelerate and extend their legs and land.
That brings up another problem, first realised by Exner (1891) and one that we as humans can’t really wrap our heads around. We are innately used to having binocular vision and estimating distances using our vision – it’s an instinct. Of the insects, only praying mantises have binocular vision. In the rest, the compound eyes are too close together (and immobile) to be able to judge distances properly except at close range. It wasn’t until a random observation and subsequent experimentation by Wallace (1959) that Exner’s 1891 hypothesis was confirmed. That hypothesis (based on the movements of crab eyestalks) basically said that they use motion to estimate distance. Wallace saw that desert locusts sway their heads from side to side before making a jump; he then confirmed with elegant experiments that the locusts are in fact measuring distance, by judging how the various obstacles in the landscape change in size with the head’s movement. This principle is used by all insects, simply because it works (for example, observe a bee flying through a gap: it always flies through the center).
Anyone who has dabbled in arthropod phylogeny will have encountered the lamentable claim that there is only one semi-convincing autapomorphy for the crustaceans proving that they are a monophyletic group: their naupliar eye (a.k.a. frontal eye; these are in addition to the compound eyes). Crustaceans have a bunch of frontal organs innervated from the same brain area, and some of these organs function as eyes, as was recognised over 200 years ago. I refer to them out of habit as the naupliar eye, but keep in mind that the correct term is frontal eyes – naupliar eye sensu stricto refers to an arrangement of three joined simple eyes (two lateral, one ventral). But there are other arrangements (dorsal, ventral, paired, unpaired, etc), so frontal eye is the general term. They come in four types (neatly distributed phylogenetically, but the use of naupliar eye diversity for phylogenetics is only now beginning to be realised): malacostracan, ostracodan-maxillopodan, anostracan, phyllopodan (cephalocarids and mysatocarids don’t have any). It’s overkill to go into the details of their differences (mostly in the arrangement and innervation). They are somewhat problematic from a functional sense though: what is their use when coexisting with the much more powerful compound eyes? It is only in copepods and ostracods that the frontal eyes are the only visual system present – in all other eyed crustaceans, we simply do not know why they are still there.
It’s relevant to do a small thought experiment on phylogenetics. As I said, the naupliar eye is widely assumed to be autapomorphic for crustaceans, with some authors going as far as homologising it for all arthropods. But it has been known for a long time (Salvini-Plawen & Mayr, 1977) that eyes evolved many times in evolution, a result subsequently confirmed by the recognition of the ubiquity of eye convergence in evolution (Land & Fernald, 1992). So it is not outrageous to propose that eyes in fact originated independently in each crustacean group. The way to solve the problem is first of all to identify criteria for homology – you don’t need to do this, as Remane already took care of that in 1952 and 1961 (Remane, 1952, 1961). For the non-German-reading, the criteria are as follows: the structures have to be similar in connection (nervature, musculature, etc) and position; the structures have to resemble each other; some form of continuity through intermediate species. Common descent is not a criterion, it is a result (saying it is a criterion for homology is engaging in circular reasoning).
There is no point in giving an answer for the crustacean frontal eyes (there isn’t one yet: some say they are homologous, others say no; I fluctuate between the two, though tending towards non-homologous, to the ire of many of my colleagues), I just thought it would be a neat way to introduce the basics of phylogenetic research.
The thought experiment can of course be taken further to consider all the bilaterians. The homology of eyes across all the bilaterians is naturally rejected by morphologists (me included; see e.g. Nilsson, 1996). But new studies have shown some very surprising similarities between insect and vertebrate eye development. Some have taken this as a sign of homology of the eyes, completely ignoring the ~30 other phyla. A more reasonable group has pointed to the use of the same ‘master gene’, Pax6 (and homologues), involved in eye formation in many different phyla (Gehring & Ikeo, 1999), hinting at a conserved ancestral function at the base of the bilaterian tree. Again, no conclusive answer exists. If you are interested in going through with any of these thought experiments, let me know and I’ll send you papers on the subject so you’re properly informed.
Anyway, to conclude the arthropods, we still need a chelicerate, so we’ll look at the horseshoe crab. During embryogenesis, there are three types of eyes: a lateral compound eye, median ocelli and ventral eyespots. They’re all associated with pigment cells called guanophores and are used during the larval stages. Through the subsequent moulting, the lateral compound eye and median ocelli develop further and are the ones used by the adult. They all use opsin, with the median ocelli also capable of sensing UV light.
To conclude, I want to highlight a pair of experiments done on the evolution of colour vision, simply to show just how much it is possible for us to do. The experiments were done by the same group over the past two decades and were about how vertebrate visual pigments evolved to absorb the specific wavelengths of light that they absorb. To do this, they used phylogenetics to infer how the ancestral pigments looked like and used directed mutagenesis to make organisms with those ancestral molecules. Basically, they got organisms that saw the same wavelengths as their ancestors. Awesome. Besides the specific knowledge gained on coelacanths and mammal colour vision evolution, it also gives us insight into just how individual mutations can result in new functions, even if they appear in sites that don’t appear very important. (If you want individual citations and papers, let me know – I didn’t put them here because there’s no review paper that I know of, just many individual studies).
Behe, M. 1996. Darwin’s Black Box.
Borst, A., Haag, J. & Reiff, D. F. 2010. Fly Motion Vision. Annual Review of Neuroscience 33, 49-70.
Briscoe, A. D. & Chittka, L. 2001. The evolution of color vision in insects. Annual Review of Entomology 46, 471-510.
Eakin, R. M. 1982. Continuity and diversity in photoreceptors. In: Westfall, J. A. (ed.). Visual Cells in Evolution.
Exner, S. 1891. The physiology of the compound eyes of insects and crustaceans.
Fernald, R. D. 2006. Casting a Genetic Light on the Evolution of Eyes. Science 313, 1914-1918.
Gehring, W. J. & Ikeo, K. 1999. Pax 6: mastering eye morphogenesis and eye evolution. Trends in Genetics 15, 371-377.
Kozmik, Z., Ruzickova, J., Matsumoto, Y., Vopalensky, P., Kozmikova, I., Strnad, H., Kawamura, S., Piatigorsky, J., Paces, V. & Vlcek, C. 2008. Assembly of the cnidarian camera-type eye from vertebrate-like components. PNAS 105, 8989-8993.
Montell, C. 1999. Visual Transduction in Drosophila. Annual Review of Cell and Developmental Biology 15, 231-268.
Land, M. F. & Fernald, R. D. 1992. The evolution of eyes. Annual Review of Neuroscience 15, 1-29.
Nilsson, D.-E. 1996. Eye ancestry: Old genes for new eyes. Current Biology 6, 39-42.
O’Connor, M., Garm, A., Marshall, J. N., Hart, N. S., Ekström, P., Skogh, C. & Nilsson, D.-E. 2010. Visual pigment in the lens eyes of the box jellyfish Chiropsella bronzie. Proceedings of the Royal Society B 277, 1843-1848.
Remane, A. 1952. Die Grundlagen des natürlichen Systems.
Remane, A. 1961. Gedanken zum Problem: Homologie und Analogie, Praeadaptation und Parallelität. Zoologischer Anzeiger 166, 447-465.
Salvini-Plawen, L. V. & Mayr, E. 1977. On the evolution of photoreceptors and eyes.
Sekaran, S., Lupi, D., Jones, S. L., Sheely, C. J., Hattar, S., Yau, K. W., Lucas, R. J., Foster, R. G. & Hankins, M. W. 2005. Melanopsin-dependent photoreception provides earliest light detection in the mammalian retina. Current Biology 15, 1099-1107.
Sharma, R. K. 2007. Molecular Neurobiology of Retinal Degradation. In: Lajtha & Johnson (eds.). Handbook of Neurochemistry and Molecular Neurobiology.
Tong, D., Rozas, N. S., Oakley, T. H., Mitchell, J., Colley, N. J. & McFall-Ngai, M. J. 2009. Evidence for light perception in a bioluminescent organ. PNAS 106, 9836-9841.
Vanfleteren, J. R. 1982. A monophyletic line of evolution? Ciliary induced photoreceptor membranes. In: Westfall, J. A. (ed.). Visual Cells in Evolution.
Wallace, G. K. 1959. Visual scanning in the desert locust Schistocera gregaria, Forskal. Journal of Experimental Biology 36, 512-525.
Warrant, E. & Dacke, M. 2011. Vision and Visual Navigation in Nocturnal Insects. Annual Review of Entomology 56, 239-254.
Xiang, Y., Yuan, Q., Vogt, N., Looger, L. L., Jan, L. Y. & Jan, Y. N. 2010. Light-avoidance-mediating photoreceptors tile the Drosophila larval body wall. Nature 468, 921-928.
Research Blogging necessities :)
Fernald, R. (2006). Casting a Genetic Light on the Evolution of Eyes Science, 313 (5795), 1914-1918 DOI: 10.1126/science.1127889
Montell, C. (1999). Visual Transduction in Drosophila Annual Review of Cell and Developmental Biology, 15 (1), 231-268 DOI: 10.1146/annurev.cellbio.15.1.231
Gehring, W., & Ikeo, K. (1999). Pax 6: mastering eye morphogenesis and eye evolution Trends in Genetics, 15 (9), 371-377 DOI: 10.1016/S0168-9525(99)01776-X