SEARCH

SEARCH BY CITATION

Keywords:

  • accommodation;
  • amphibious eyes;
  • graded index lens;
  • scanning eyes;
  • spider eyes

Abstract

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

Structures which bend light and so form images are present in all the major phyla. Lenses with a graded refractive index, and hence reduced spherical aberration, evolved in the vertebrates, arthropods, annelid worms, and several times in the molluscs. Even cubozoan jellyfish have lens eyes. In some vertebrate eyes, multiple focal lengths allow some correction for chromatic aberration. In land vertebrates the cornea took over the main ray-bending task, leaving accommodation as the main function of the lens. The spiders are the only other group to make use of a single cornea as the optical system in their main eyes, and some of these – the salticids – have evolved a remarkable system based on image scanning. Similar scanning arrangements are found in some crustaceans, sea-snails and insect larvae.


The origins of eyes

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

There are three ways that animals can produce images: using pinholes, concave mirrors and lenses. Pinhole eyes exist, in Nautilus and giant clams, and mirrors are found in the eyes of scallops, a few crustaceans and one deep-sea fish, but lenses provide by far the most common way of splitting up light according to its direction of origin. Although most of the world's described animal species have compound eyes, I will restrict this review to eyes with single optical systems. For a discussion of all animal eyes the reader is referred to Land and Nilsson.[1]

Early lens eyes did not readily fossilise, and information on the earliest chordate eyes is sparse: Ordovician conodonts such as Clydognathus had large eyes, and some Cambrian chordates such as Haikouichthys also had a pair of eyes, but there is no information on their internal structure. The existing relatives of the first fish, the agnathan lampreys, have eyes so similar to other fish that they too tell us little of their early evolution. Fortunately, the situation in the molluscs is different. Amongst existing forms one can trace a logical sequence (not necessarily quite the one followed by evolution) from the pit-like eyes of limpets, through simple lens eyes of various gastropod snails, to the large and sophisticated eyes of the cephalopods – Octopus, cuttlefish and squid (Figure 1). These have lenses with an inhomogeneous construction, a fine-grain retina, a variable pupil and external eye muscles, and provide vision that is in all ways comparable with that of fishes. This is perhaps the most famous example of convergent evolution in the animal kingdom.

image

Figure 1. Eyes of currently existing molluscs, showing a probable route for the evolution of the lens. Left: eyes of six gastropod snails, ranging from a simple pit to an eye with a good lens: a, Patella; b, Pleurotomaria; c, Haliotis, d, Turbo; e, Murex; f, Nucella. (after Salvini-Plawen & Mayr.[24] ep, epithelium; la, lacuna; li, lens; vm, vitreous mass. Right: Eye of Octopus. (From Young).[25]

Download figure to PowerPoint

The first stages in producing a lens eye involved creating a partially closed cavity and filling it with some substance with a higher refractive index than the surrounding sea-water. This did not produce a sharp image, but each stage would result in a more restricted acceptance angle for the receptors lining the cavity, and so provide a direction for improvement upon which evolution could operate. The final step in this process was to evolve a lens whose construction allowed the production of a sharp image, and I will consider this in the next section. It is of interest, though, to estimate how long all this might have taken. Starting from a simple patch of photoreceptive skin, and using typical values for variation in a population and conservative estimates of selection pressure, Nilsson & Pelger[2] estimated that it would take about 40 000 generations to evolve an eye that produced a sharp image. Thus a quite respectable eye could evolve in about half a million years. This feasibility study showed that eyes could evolve quite rapidly, and that the 50 million years of the Cambrian period provided more than enough time for this to happen.

Making a good lens

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

Eyes evolved in the sea, so initially the cornea, with fluid on both faces, could play no part in forming an image: the lens had to do all the work. Producing a lens that will perform well in water is not quite as easy as it may seem at first. It turns out that a lens made simply of a glass-like material (dry crystalline protein for example) will not produce an image of good enough quality, nor have a focal length short enough to be useful. The focal length needs to be kept short in relation to the size of the lens to keep the eye as a whole reasonably small. This means that the radii of curvature of the surfaces have to be small, which in turn makes a spherical shape for the lens more or less obligatory. However, spherical lenses have serious defects. The worst is known as spherical aberration, in which rays at a distance from the axis of the lens are bent through too great an angle to come to the same focus as the on-axis rays, and the result is a blur circle on the retina rather than a sharp image (Figure 2). With a spherical lens this would be wide, and the image would be very poor. The other problem is that the lens would have a rather long focal length. A single surface of radius r, separating two media of refractive indices n1 and n2, has a focal length given by rn2/(n2 − n1). A spherical lens, where light encounters two surfaces, both of radius r, has a focal length (f) equal to half this:

  • display math
image

Figure 2. Spherical aberration. (a) Paths of rays through a homogeneous lens of refractive index 1.66, showing how rays far from the axis are refracted too much (spherical aberration). (b) Lens with the same focal length as (a), but with a gradient of refractive index and a central refractive index of 1.52. (c) Form of the gradient in a fish lens, capable of producing a spherical aberration-free image. From Land & Nilsson.[1]

Download figure to PowerPoint

The refractive index (n2) of a dry protein such as the crystallin found in lenses is about 1.53, and with sea-water (n1 = 1.34) as the outside medium, the focal length of a lens made of such material would be four lens radii. In fact, this is much longer than the focal lengths of real lenses in fish and cephalopods. It has been known since the studies of Matthiessen in the 1880s that the lenses of fish as well as cephalopods and marine mammals nearly all have focal lengths of about 2.5 lens radii, a number that has become known as Matthiessen's ratio.

Clearly, a spherical lens made of homogeneous protein does not fit with what we know of fish and cephalopod lenses, namely that they are of excellent optical quality and short focal length. This apparent contradiction interested a number of 19th-century scientists including James Clerk Maxwell, who came up with the idea that such lenses must have a gradient of refractive index, highest in the centre and lowest near the periphery (Figure 2). Matthiessen had shown that there was such a gradient in fish lenses, and believed that its form was that of an inverted parabola, with the refractive index falling as the square of the distance from the lens centre. Matthiessen, it turns out, was not far wrong in his guess, although more recent theoretical studies have suggested that there are other functions that give a somewhat better performance in terms of the correction for spherical aberration (for a review see Jagger).[3]

What does the refractive index gradient achieve? In the first place it changes the pattern of refraction from a discrete bending of the rays at each interface to one in which rays are bent continuously within the body of the lens. The effect on spherical aberration is that the outermost rays, which travel shorter distances within the lens, are bent relatively less than they are at the interfaces of the homogeneous lens. Given the correct gradient, all rays can be brought to a focus at the same point, for light of a single wavelength. The shorter focal length is achieved because continuous refraction results in greater total ray-bending than does two-surface refraction. In fact, an f/r ratio of 2.5 can be achieved in a gradient index lens with a central refractive index of 1.52, whereas the same ratio would require a homogeneous lens to have an index of 1.66. The real value of the short focal length of fish lenses lies in the effect this has on light-gathering power. In photographic terms the F-number of the eye (focal length/diameter) is 1.25, which gives an image 2.6-times brighter than the image behind a homogeneous protein lens (n = 1.52) which has an F-number of about 2.

The other important defect of biological lenses in general is chromatic aberration, in which light of shorter wavelengths is brought to a focus closer to the lens than longer wavelength light. This means that a single retina at a fixed distance from the lens cannot be in focus for all wavelengths simultaneously. For animals that have only one visual pigment (deep-sea fish and most cephalopod molluscs, for example) this is not a problem. However, shallow water fish have excellent colour vision, and typically they possess four cone types whose wavelengths of maximum sensitivity cover a 250 nm range from ultraviolet to red. One way round the problem would be to place the different cone types at different distances from the lens, and there is some evidence for this. However, the distances involved are quite large (up to 10% of the average focal length, or 1 mm in a 10 mm focal length eye) and cone separations as great as this are not physically possible in a thin retinal sheet. It is now clear that some fish use another method. This is to produce lenses with multiple focal lengths, brought about by variations in the basic Matthiessen gradient.[4] The way this works is shown in Figure 3. For light of a single wavelength, the inner zones of the lens bring light to a closer focus than the outer zones (this effectively means that the lens is over-corrected for spherical aberration). However, for white light with a range of wavelengths the images from inner and outer zones have a spread of focal lengths, because of chromatic aberration. This means that the position of the image for short wavelengths formed by the outer zone can be made to coincide with the image for long wavelengths formed by the inner zone. This in turn allows cones with different wavelengths of maximum sensitivity to receive in-focus images in the same plane. This is not a perfect solution, because these in-focus images are contaminated by light from the other out-of-focus images, and so will have reduced contrast compared with a perfectly corrected monochromatic image. However, this is better than not having a sharp image at all, and it seems that a wide variety of teleost fish and many other vertebrates have opted for this solution. Figure 3 is somewhat over-simplified; in the species studied by Kröger et al.[4] there were three distinct images corresponding to the sensitivity maxima of the three cone types, rather than the two shown in the figure.

image

Figure 3. Method used by some fish and other vertebrates to overcome chromatic aberration. (a) By slightly varying the refractive index gradient (Figure 2c) the lens produces several sharp images at different distances (F1 & F2). Although each image has longitudinal chromatic aberration, their locations can be adjusted so that the images for different wavelengths coincide as shown. This means that cones with different spectral sensitivities can be located in a single layer. Based on Kröger et al.[4] (b) A circular pupil will cut off the outer zones of the lens, compromising chromatic correction. A slit pupil can include parts of all zones.

Download figure to PowerPoint

A problem with a multifocal lens of this kind is that when a circular pupil is used to restrict the light entering the eye this will interfere with the chromatic correction by progressively occluding the outer zones. One way round this is to use a slit pupil, which still allows all zones to be sampled (Figure 3). This appears to be one reason for the common occurrence of slit pupils among terrestrial vertebrates.[5] Slit pupils can also be closed down more effectively in bright light.

The impressive optical performance of the ‘Matthiessen lens’ has ensured its evolutionary popularity. Because an f/r ratio of around 2.5 immediately tells one that a spherical lens has a gradient structure, it is easy to survey the animal kingdom for occasions on which this type of eye has evolved. It evolved in the fish, presumably once, in the cephalopod molluscs, and possibly more than once in the gastropod molluscs. Matthiessen lenses also evolved independently in two unlikely places: among the annelid worms in the alciopids, a family of polychaetes that have become active carnivores in the marine plankton; and just once in the Crustacea, in the copepod Labidocera where the males have a pair of lenses that share a line-like retina of 10 receptors.

A particularly interesting variant of the spherical lens eye is found in the Cubozoans, the class of cnidarians that includes the notorious box jellyfish of the Great Barrier Reef. Most cnidarians are eyeless, but cubozoans have 24 eyes of various kinds arranged on four rhopalia – collections of sense organs suspended from the swimming bell (Figure 4). Two of the eyes on each rhopalium, although small (ca. 200 μm across), are remarkably similar in geometry to fish eyes, with a spherical lens and hemispheric retina. The lenses have a refractive index gradient, and a focal length of about three lens radii. Intriguingly, the retina is too close to the lens to be in focus, and the result is a blur circle. The reason for this appears to be straightforward: the blurred image will pass the low spatial frequencies needed to guide the animals’ behaviour, but not the higher frequencies, from ripples or debris in the water, which are not useful. Box jellies use their eyes to maintain their position relative to the bottom or shoreline when swimming in moving bodies of water, but they do not use them to hunt prey.[6] The phylum Cnidaria, to which the Cubozoa belong, split from the ancestors of most other animals (the Bilateria) at least 700 years ago, well before the Cambrian, so the potential to make really quite good eyes was present in this early metazoan lineage.

image

Figure 4. Cubomedusan eyes. (a) One of the four rhopalia from the bell of Tripedalia cystophora. The two large eyes (ULE and LLE) point upwards and obliquely downwards. There are also two slit eyes (SE) and pit eyes (PE) without lenses. The statolith ensures that the rhopalium stays vertical. (b) In the lower lens eye (LLE) the lens is too weak to focus inside the retina, but allows for wider aperture. From Land & Nilsson.[1]

Download figure to PowerPoint

Deep water variations on the Matthiessen theme

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

A spherical lens produces the same image quality across a hemispheric retina, and so is ideal for the wide-angle eyes of most fish and molluscs. Not all fish, however, make use of a 180° field. In the mid-waters of the ocean, at depths of 300 m or more, usable light from the surface becomes restricted to a downward–directed cone about 60° across. This is reflected in the structure of the eyes of many fish and cephalopods, which have become tubular and upward-directed (Figure 5). Catching prey in the mesopelagic environment essentially involves spotting the dark silhouette of some creature against the dim background of residual daylight. Of course, if you wish to avoid capture yourself, this means minimising your own silhouette, and there are various strategies for doing this. Many fish and crustaceans are transparent, but it is not possible to have transparent eyes or, usually, a transparent gut. Hatchet fish have become very thin, and their highly reflecting silvery sides help to make them appear invisible from the side. However, a very common technique employed by fish, cephalopods and crustaceans is counter-illumination (Figure 5). The undersides of these animals are studded with photophores, sometimes along the whole length, and in other cases just under the eyes or gut. This strategy will only be effective if the light emitted by the photophores matches the downwelling light in intensity, and angular distribution and so the construction of these devices is often quite intricate, involving lenses, mirrors and filters.[7] Impressively, some fish, squid and crustaceans can regulate the light output of their photophores over a 1000-fold range, and in some cases have a photophore that points into the eye, in parallel with the others, to allow a direct comparison with the light from above.

image

Figure 5. Vision in the deep sea. (a) Disguise techniques. A, Dorsal counter-shading. B, Reflecting sides: the light reflected has the same intensity as the light that would have been transmitted at the same angle relative to the sea surface. C, Counter-illumination of the silhouette with photophores. (b) Section of an eye of the deep-sea fish Scopelarchus, showing the tubular form, and the size it would have if the shape were conventional (dashes). (c) Double eye of Bathylychnops exilis. The lens of the downward-pointing secondary eye is formed from the sclera. Ret, retina; Cho, choroid; Scl, sclera. (a) Based on Denton,[26] (b and c) redrawn from Locket.[27]

Download figure to PowerPoint

There is another source of light in the deep sea, and this derives from the luminescence of other organisms, which may also be potential prey. Some mid-water fish also have secondary eyes of various designs which point obliquely downwards. In Bathylychnops (Figure 5) the secondary eye is similar to the main eye, but the lens is formed from the sclera. However, in Dolichopteryx the secondary optical system is not based on a lens at all, but on a concave mirror with a complex stepped construction. Detecting self-luminous objects against a dark background is a less demanding visual task than detecting small contrast differences against a dim background, so these secondary eyes are usually much smaller than the main eyes.

Non-spherical lenses in marine animals

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

There are remarkably few aquatic eyes of the single-chambered type that do not contain spherical lenses. There are one or two, however, where the required ray-bending is achieved by refraction at a number of surfaces, more in the manner of a multi-component camera lens. A particularly impressive system is found in the copepod Pontella, where a total of six surfaces in three lenses are used to produce the image.[8] This eye is the ventral component of the typical tripartite ‘nauplius’ eye and, as in other copepods, contains a retina with very few receptors, in this instance only six (Figure 6). This apparent simplicity is the more remarkable because of the amazing development of the eye's optics. In the male there are three lenses, one attached to the eye-cup itself, and another two in the animal's rostrum. In the female, curiously, the most anterior component is missing, making the lens a doublet rather than a triplet. Seen from below it is clear that while most of the surfaces are approximately spherical, this is not true of the first surface, which is distinctly parabolic. Ray tracing through the lenses, assuming them to have a uniform refractive index of 1.52, shows that on axis the image is good. With a spherical front surface the system as a whole gives a poor image, with obvious spherical aberration, but with the parabolic surface this disappears, giving a well-corrected point image. From an optical standpoint, it seems that Pontella has hit upon the alternative way of avoiding the perils of spherical refracting surfaces: it has made an aspheric lens, rather than a graded index one.

image

Figure 6. Copepods with unusual lenses. (a) Pontella. Female left and male right. Only the male has a triplet lens. (b) Drawing of the male eye from below showing the parabolic profile. The retina has only six receptors. (c) eyes of Sapphirina, showing the binocular construction. (d) One eye of Copilia quadrata, showing the extent of scanning (A–B) by the second lens. (a–c) from Land and Nilsson.[1] (d) after Exner.[28]

Download figure to PowerPoint

The eyes of another group of copepods – Sapphirina, Copilia and their relatives – have intrigued biologists for well over a century, and it is not hard to see why (Figure 6). Each eye has a pair of lenses. The larger anterior lens is part of the carapace, and it throws an image onto a second smaller lens attached to the front of a tiny retina containing 5–7 receptors. The design is thus somewhat like a pair of telescopes, each with objective and eyepiece lenses. The other reason why these eyes have merited so much attention concerns the way they move. The rear part of each eye, including the second lens, moves sideways in the body, through an angle of about 14°, as measured from the front lens. In Copilia the eyes move together, but in opposite directions, fast medially and slower laterally. The rate varies between 0.5 and 10 Hz.[9] Although the scanning movements of the eyes increase the effective field of view of the retina, which on its own is only about 3° across, they still only enable the animal to scan a tiny line in the surrounding space. Unfortunately, there are no direct observations of the behaviour or eye movements of Copilia in its natural marine environment.

It is worth mentioning in this context that one fish, the sandlance Limnichyes fasciatus, also splits its refraction between four surfaces by making use of a thickened corneal ‘lenticle’ with a relatively high refractive index (1.38). This tiny but remarkable fish, with its independently-moveable eyes, catches copepods and other plankton with a rapid, visually-guided lunge. The lenticle can change shape during accommodation and forms part of a very fast focusing mechanism. In conjunction with a rather weak lens, the lenticle brings the nodal point of the optical system towards the front of the eye, thus increasing the focal length and magnifying the image.[10]

Materials for constructing lenses

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

Biological lenses are constructed from a wide variety of materials, most of them having other functions in different parts of the body. The α-crystallins in vertebrate lenses are related to a heat-shock protein that protects tissue from high temperatures. A major constituent of the cornea is a tumour cytosolic enzyme. In Octopus lenses one protein is an aldehyde dehydrogenase, and another is glutathione S-transferase. And this pattern of borrowing molecules with other biochemical functions is repeated across the animal kingdom.[11] The requirements are that the molecules should be suitable for mass expression and dense packing, but unlikely to aggregate to form lumps. It seems that there is a range of suitable candidates.

Emergence onto land: a new role for the cornea

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

When they emerged from water, the early land vertebrates would have found that their eyes had a new optical arrangement. The cornea, which in water was simply a tough transparent membrane protecting the front surface of the eyeball, became an image-forming structure in its own right, rivalling the lens in its ability to bring rays of light to a focus. In water the cornea has little or no optical effect, because it has a fluid of the same refractive index on both sides. On land, however, the front surface is in air, so there is a now a large refractive index difference, across which rays are bent by refraction. It turns out that the ray-bending power of a fish lens and a cornea in air are quite similar. If the radius of curvature of a surface is r, and the refractive indices on the two sides are n1 and n2, then the focal length f of the surface is given by the formula f = n1r/(n− n1). This means that there is a focused image of distant objects at a distance f from the centre of curvature of the surface. For a cornea in air the outside refractive index n1 is 1, and inside the eye n2 is about 1.34, so that f becomes r/0.34, or about 3r. We saw earlier that a fish lens of radius r has a focal length of about 2.5r, so the focal lengths of corneas and lenses with the same radius are quite comparable.

An eye with both a cornea and a fish-type lens has too much focussing power, and if the first proto-amphibian to come on land had done nothing about this it would have been very myopic (Figure 7). The blurring would be comparable to what happens to our vision when we go swimming without goggles; in this case, however, we lose the power of the cornea (which now has fluid on both sides) and become hyperopic, which means that we do not have clear vision at any distance. To solve the problem of excessive optical power, land vertebrates could have done a number of things. They might have abandoned the lens altogether and adopted the cornea as the sole image-forming structure, or they could have kept the lens and flattened the cornea so that it had no power, or they could have retained both but shrunk the eye to fit the shorter focal length of the combined system. The last of these possibilities, or something like it, does occur in nocturnal mammals; but most of the reptiles, birds and mammals have opted for a compromise, in which the lens is retained, but with much less power than the ancestral fish lens. The lens and cornea then divide the optical power between them: in humans this ratio is about 1–2 (Figure 5). Even in man, however, the lens has retained some of the inhomogeneity of its fishy ancestor, with a core refractive index of about 1.5, falling to 1.37 at the cortical surface.

image

Figure 7. (a) Terrestrial eyes are hyperopic in water, and aquatic eyes are myopic in air. (b) In the human eye, and most terrestrial vertebrates, about 2/3 of the focussing is done by the cornea, with the lens mainly responsible for accommodation. Ic, image formed by cornea alone; Ir, final image on the retina.

Download figure to PowerPoint

Animals which are active day and night, such as horses, owls, and many mammalian carnivores, have eyes in which the lenses have a diameter of about 0.4–0.5 times the diameter of the eye itself. In truly diurnal animals, for example monkeys and parrots, the ratio is lower, between 0.3 and 0.4. However, in nocturnal animals that rarely emerge in daylight, such as the house mouse, opossum, and bush baby have lenses that are almost spherical, with lens diameters that are 0.6–0.8 times the eye diameter.[12] These differences are of relative not absolute lens size, and are concerned with getting as bright an image as possible for a given size of eye. A large almost spherical lens, combined with a strongly curved cornea, gives a very short focal length, and combined with a wide aperture this gives the eye a very high light gathering power (image brightness is proportional to (D/f)2, where D is aperture diameter and f focal length). In photographic terms a house mouse has an F-number (f/D) of about 0.9, compared with about 2.0 for a human with a wide-open pupil. The mouse's image is brighter by a factor of nearly 5. Generally speaking the power of the cornea is relatively more important in diurnal eyes, and the lens in nocturnal eyes.

Accommodation

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

Once an eye reaches a size of a millimetre or more the depth of focus shrinks, and some means of altering focus for near and far distances is needed (Figure 8). In fish, with a spherical lens of fixed dimensions, the only way to do this is to move the lens towards or away from the retina. In elasmobranchs such as sharks, and also in amphibians, the protractor lentis muscle moves the lens away from the retina to bring near objects to a focus, but in teleost fish the reverse happens: the retractor lentis pulls the lens closer for distant vision. In land vertebrates accommodation occurs by deformation of the elastic lens (at least it does until you reach your mid-forties!). In birds this is done by Brücke's muscle which pushes the ciliary body inwards, allowing the lens to bulge and so increasing its power. In mammals the ciliary muscle does the same thing. Birds have a second muscle, Crampton's muscle, which pulls on the periphery of the cornea, decreasing its radius of curvature and increasing its power.

image

Figure 8. Accommodation mechanisms in vertebrates. See text for description. From Land and Nilsson.[1]

Download figure to PowerPoint

Chamaeleons have remarkable lenses which actually have negative power at rest, implying a refractive index profile quite different from the usual ‘highest in the centre’ gradient.[13] Muscle action deforms the lens and gives it positive power, and chamaeleons use a readout of this action to estimate the distance of insect prey.[14] The combination of the cornea and negative lens gives the optical system a particularly long focal length, and hence a magnified image, which, together with a distinct high-resolution fovea, provides exceptional acuity. The high mobility of the turret-like eyes also aids the search for insects amongst foliage.

A slightly different telephoto mechanism is found in hawks. Here the negative lens takes the form of a concave foveal pit in front of the receptors. The difference in refractive index between the vitreous and the foveal matrix turns this interface into a negative lens which shifts the focus backwards and magnifies it a factor of 1.45.[15] Hawks have slightly narrower receptors than humans, and wider pupils (about 6 mm) improving the diffraction limit. These modifications give them the best acuity of any animal, about 160 cycles per degree, compared with 60 cycles per degree for humans with similar-sized eyes.

Some birds and flat fish employ a ‘static’ method of accommodation known as a ramp retina, in which the upper retina, which images the substrate, is further from the lens than the lower retina. In fruit bats the retina undulates, providing a range of focal distances.[12]

Amphibious lenses

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

Seals, like mice, have spherical lenses but for a different reason. Their problem is that, having returned to water, they no longer have an optically useful cornea. They therefore require a much more powerful lens, and that means a spherical lens, just as in fish. However, they are not wholly aquatic, and when they come onto land the re-appearance of a strong cornea would make them very myopic. One solution is a flattened cornea with little power in either medium. Seals have corneas that are much flatter than their terrestrial carnivore relatives, and do indeed have similar acuity in both air and in water, at least in daylight. Incidentally this is one reason why baby seals look so appealing. Their eyes are like limpid pools, not because of the purity of their souls, but because the corneas are almost flat.

Besides seals, there are some species in all vertebrate groups that need to see reasonably well in both air and water. Flying fish, mud skippers, most amphibians, turtles, diving birds, and otters all spend part of their lives in each medium, and have to cope with the sudden large changes in optical power when they dive or surface (Figure 9). Like seals, many diving birds such as penguins, and some rock-pool fish minimise the problem by having a much flatter cornea than their non-amphibious relatives. One problem with a flat cornea is that it tends to restrict the field of view, and also results in serious distortion in the periphery. The rock-pool fishes Dialommus fuscus and Mnierpes macrocephalus have solved this problem in a particularly interesting way. They have two flat goggle-like corneas in each eye, making an angle of about 135°. Presumably this both increases the field of view and decreases distortion, although at the price of having a distinct ‘join’ through the centre of the visual field. A somewhat similar arrangement occurs in the flying fish Cypselurus heterurus which has a tent-like cornea consisting of three almost flat triangular facets.

image

Figure 9. Amphibious eyes in vertebrates. (a) The rock-pool fish Dialommus fuscus with flat-faced ‘goggles’. (b) Accommodation in the merganser, a diving duck, is achieved by squeezing the lens through the iris to produce a high curvature. (c) The ‘four eyed fish’ Anableps achieves simultaneous vision in air and water by the use of an ovoid lens with different curvatures on different axes. (a, c) redrawn from Munk,[29] (b) from Sivak et al.[16]

Download figure to PowerPoint

An alternative is to have a focusing mechanism so strong that it can make up the shortfall in optical power. When we dive we lose 40 D (dioptres) and can accommodate by a maximum of 10 D, so we are still left with an unbridgeable 30 D. Certain diving birds, however, have a method of altering the curvature of the front surface of the lens that is much more effective than ours. Birds and reptiles have a muscular iris supported by a ring of bony ossicles, and they are able to squeeze the lens into the constricted pupil using the powerful ciliary (Brücke's) muscle, creating a very high curvature in the resulting blip (Figure 9). Using this technique, mergansers and goldeneyes, both diving ducks, are able to generate 80 and 67 dioptres of extra power respectively, whereas the non-diving wood duck and mallard produced only about six and three dioptres.[16] Other diving birds probably use this method of accommodation, as do aquatic turtles and water snakes.

The ‘four-eyed fish’ Anableps anableps from South America has solved the problem of seeing in air and water simultaneously. Anableps cruises with half its eye above the surface meniscus, and half below. It has two pupils, one looking into each medium, and a lens whose shape ‘combines an aquatic optical system harmoniously with an aerial one, in a perfectly static situation’.[12] The compromise is achieved by the ovoid shape of the lens, with its long axis in the direction that looks down into the water (Figure 9). Rays parallel to the axis meet the strongest curvature of the lens, and so are refracted relatively more than rays coming from air, which meet the weaker curvatures of the short axis. The latter rays, however, are also bent by the cornea, so that the total amount of refraction is much the same in the two cases. It seems that this wonderful design is unique.

The corneal lens eyes of spiders and insects

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

It comes as something of a surprise to find that the cornea-lens combination (our kind of eye) is not particularly popular in the animal kingdom. It is necessarily confined to terrestrial animals, and since insects have opted predominantly for compound eyes (which are beyond the scope of this article) that only leaves one major land-living invertebrate group, the spiders, which makes exclusive use of the simple corneal eye system. Some insects also have simple eyes when they are larvae, and adults may also use them as flight-stabilizing devices, in conjunction with the compound eyes.

Spiders and their terrestrial relatives all have eyes of the simple type with the cornea as the main refracting surface. Their distant chelicerate relatives the horseshoe crabs (Limulus) have compound eyes, and it may well be that the eyes of modern arachnids are derived from compound eyes by a process of simplification. True spiders usually have eight eyes (sometimes reduced to six) and these are of two kinds: the principal eyes which point forwards and the secondary eyes which cover more peripheral fields of view. The two kinds of eye have different embryological origins, and the layout of the receptors is also different. Optically the eyes are very varied. They are small, by vertebrate standards, mostly much less than a millimetre across, but this still makes their lenses larger than the facets of compound eyes by an order of magnitude. For most insects diffraction limits their resolution to about 1°, but in jumping spiders this reduces to a few minutes of arc.

The most impressive spider eyes belong to the families that catch their prey by hunting rather than by using webs. Of the hunters some stalk prey in daylight, whilst others ambush their prey at night. The jumping spider Portia fimbriata, which eats other spiders, has the highest resolution of all, with an inter-receptor angle of 2.4 arc-min: only five times greater than the human fovea. Others have modest resolution but enormous light gathering power. Some spiders of the genus Dinopis, which catch cockroaches in forests at night, have eyes up to 1.4 mm in diameter, and an f-number of 0.6. The majority of web-building spiders have rather poor eyesight. The principal eyes usually do form low-resolution images, but the secondary eyes have unfocused lens-mirror combinations concerned with direction finding using the sun and other celestial cues.

Of all the spiders, the jumping spiders (Salticidae) undoubtedly have the most acute vision, and the most sophisticated visual systems. They stalk their prey (usually insects) in much the same way that a cat stalks a bird. There is an interesting division of labour between the secondary and principal eyes (Figure 10). The secondary eyes induce turns towards moving objects, which are then picked up, by and identified, by the much larger, forward-pointing, principal eyes. These are high resolution eyes with long focal lengths, increased by a telephoto arrangement similar to that in hawks.[17] Oscar Drees studied these spiders in the 1950s, and found that the principal eyes were responsible for distinguishing between prey and potential mates, and that this judgement was made on the basis of the geometry of the leg pattern of the target animal.[18] The principal eyes are specialized in two other ways. First, they each have a very narrow field of view (about 5° horizontally by 20° vertically), but this is offset by the fact that they ‘scan’ targets, with a complex pattern of eye movements (Figure 10) involving lateral, vertical, and rotational movements of the retinae.[19] Unlike vertebrate eyes, the lens itself remains still: it is only the retina that moves. Second, the retina is arranged in four layers, one behind the other. These animals have dichromatic colour vision, based on green sensitive pigment in the proximal layers and ultra-violet in the distal layers. This arrangement compensates for the longitudinal chromatic aberration of the optics. It may also allow objects at different distances to be focused on different sub-layers, and so act in lieu of an active accommodation system. One of the attractive features of the salticid visual system is its compactness. By confining high resolution to one pair of narrow, long focal length eyes, whilst using much smaller eyes for peripheral vision which requires lower resolution, jumping spiders have saved a great deal of space. If the same eye performed both tasks (as in vertebrates), its volume would be at least ten times greater.

image

Figure 10. Eyes of jumping spiders. (a) Prosoma of a jumping spider, showing the fields of view of the six eyes. The antero-median (AM) or principal eyes have a very narrow horizontal field of view extended by movements of the retina to cover a 50° arc. The antero-lateral (AL) and postero-lateral (PL) eyes are fixed and cover most of the field around the animal. The postero-median eyes are tiny, and non-functional in most species. (b) Front view of a jumping spider (Phidippus sp.) showing the characteristic leg posture. (c) Scanning movements of the two vertically extended AM retinae when viewing a novel target. Each horizontal scan lasts 1–2 s and the torsional movements are slower, taking 5–8 s. Based on Land.[19]

Download figure to PowerPoint

The largest eyes of any spider, and probably the largest simple eyes of any land invertebrate, are found in the genus Dinopis. As mentioned above, they are nocturnal hunters. They ambush insects passing beneath them by pinning them to the substrate with a net of sticky silk – rather like a Roman retarius gladiator. The trigger for this action is visually detected movement. Interestingly, the enormous (for a spider) postero-median eyes have near-spherical inhomogeneous lenses reminiscent of those of fish.[20] They have a double structure – a low index outer layer surrounding a denser core – the core itself behaving as a graded-index lens. The receptors are also huge, with receptive segments 20 μm wide and 55 μm long during the day, lengthening to twice this in the dark. The net effect of these various heroic adaptations is that the sensitivity of these eyes is enormous. Compared with a jumping spider like Phidippus, the sensitivity (measured as the number of photons absorbed per receptor, for a given field luminance) is roughly 2000 times greater, although with an inter-receptor angle of about 1.5°, the resolution is about ten times coarser.

Insect simple eyes, or ocelli, fall into two main groups: larval eyes, and the dorsal ocelli present in most winged adult insects (Figure 11). In both, the curved air-tissue cornea interface is the main refracting surface, although as in vertebrate eyes, a lens of some kind often augments the optical power of the system and aids in the formation of the image.

image

Figure 11. Ocelli in insects. Left: eyes of tiger beetle larvae (Cicindelida) (a) Head showing the six eyes. (b) Larva in ambush position in its burrow. The eyes are used to spot insect prey on the surface. These are grasped and pulled down the burrow. (c) Section of the largest eye: the inter-receptor angle is about 1.8°. Based on Friederichs (1931). Right: dorsal ocelli of a locust. (d) The three ocelli are on the sides and front of the head. (e) Section of an ocellus showing the pigmented iris, receptors and neuropil, from which only a few large axons emerge. The eye is profoundly out of focus. Foci for the light (LA) and dark (DA) adapted eye are shown, about 400 μm below the receptor layer. Redrawn from Goodman.[22]

Download figure to PowerPoint

In insects with a distinct larval stage, the ocelli are the only eyes the larvae possess. They vary greatly in size and complexity. In many, for example, butterfly and moth caterpillars, the ocelli are small and have a structure resembling that of a single ommatidium from a compound eye. However, in a few predatory species the eyes are larger, with an extended retina. Particularly impressive ocelli are found in tiger beetles (Cicindela). These ambush insect prey as they pass their burrows (Figure 11a–c). There are again six ocelli on each side of the head, but two are much larger than the others. The largest has a diameter of 0.2 mm and a retina containing 6350 receptors. The inter-receptor angle is about 1.8°, comparable with or better than the resolution of the compound eyes of most adult insects. This raises the interesting evolutionary question as to why the insects did not retain eyes like this into adult life. The larvae of water beetles (Acilius, Thermonectus) have equally large and intriguing ocelli, which they use for catching mosquito larvae.[21] Like jumping spiders, they have narrow retinae, but in these the long dimension is horizontal rather than vertical. They also scan, but they do this not with eye muscles, but by performing a vertical movement of the whole head and thorax, resembling a press-up.

Adult insects that fly typically have three simple eyes on the top of their heads. These dorsal ocelli resemble larval ocelli in possessing a lens and an extended retina (Figure 11d–f), but they are not embryologically related to the larval eyes. Some dorsal ocelli have tapeta, and some a mobile iris. They each have a wide field of view of 150° or more, and may have as many as 10 000 receptors. So far all this suggests that these are ‘good’ eyes, like those of hunting spiders. However, there is a problem. Everyone who has tried to get to grips with the optics of these eyes agrees that they are profoundly out of focus, with the retina much too close to the lens.[22] For example in the blowfly Calliphora the receptors extend from 40 to 100 μm behind the lens, but the focus is at 120 μm. It appears that this is not a mistake; dorsal ocelli are deliberately defocused. This raises the same questions as the defocused eyes of box jellyfish, discussed earlier. Most recent studies support the idea that the ocelli are horizon detectors, involved in enabling an insect to make fast corrections for pitch and roll.[23] The defocus then makes sense; high spatial frequency clutter such as leaves and branches will be removed, allowing the receptors to respond to changes in the overall distribution of light in the sky. The idea that these ocelli contribute to flight equilibrium is confirmed by the fact that the receptors converge massively onto a relatively few second-order neurons, and that these project directly into the optomotor system.

Final thought on lens evolution

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References

One could ask: if, as Nilsson and Pelger showed in 1994,[2] it is relatively straightforward to evolve a well focussed lens, or a properly curved cornea, why haven't all animals done this? In our book on animal eyes[1] we discuss the different levels of light-controlled behaviour that animals with different life-styles require. At the simplest level there are behaviours that only require non-directional monitoring of light: circadian rhythm control, light-avoidance, shadow responses to avoid predation and the detection of when an animal is close to the surface of the substrate. The next level requires directional light sensitivity: for phototaxis and the control of body posture. It is only at the next level that some kind of optical system, other than simple shadowing, is needed, and even then it need not resolve particularly well. Such low spatial resolution tasks would include the monitoring of self-motion, collision avoidance, habitat selection and orientation to coarse landmarks or major celestial objects such as the sun or moon. It is only at the final level that visual tasks require high resolution. These include pursuit of prey, predator detection and evasion, mate detection, orientation to fine landmarks, visual communication and recognition of individuals. In some cases an animal achieves only one or two of these behaviours. The copepods in Figure 6 or the tiger beetle larva in Figure 7 have good optics for one purpose only, and their brains are appropriately small. However, in the vertebrates, the cephalopod molluscs and the higher arthropods (including the jumping spiders), all of which use their eyes for multiple purposes, the visual regions occupy at least half the volume of the brain. The full exploitation of high resolution images is a costly business, and not to be embarked on lightly. Evolution does not seek arbitrary perfection, once the result is good enough.

References

  1. Top of page
  2. Abstract
  3. The origins of eyes
  4. Making a good lens
  5. Deep water variations on the Matthiessen theme
  6. Non-spherical lenses in marine animals
  7. Materials for constructing lenses
  8. Emergence onto land: a new role for the cornea
  9. Accommodation
  10. Amphibious lenses
  11. The corneal lens eyes of spiders and insects
  12. Final thought on lens evolution
  13. References
  • 1
    Land MF & Nilsson D-E. Animal Eyes. Oxford University Press: Oxford, 2012.
  • 2
    Nilsson D-E & Pelger S. A pessimistic estimate of the time required for an eye to evolve. Proc R Soc Lond B Biol Sci 1994; 256: 5358.
  • 3
    Jagger WS. The optics of the spherical fish lens. Vision Res 1992; 32: 12711284.
  • 4
    Kröger RHH, Campbell MCW, Fernald RD & Wagner H-J. Multifocal lenses compensate for chromatic defocus in vertebrate eyes. J Comp Physiol A 1999; 184: 361369.
  • 5
    Malmström T & Kröger RHH. Pupil shapes and lens optics in the eyes of terrestrial vertebrates. J Exp Biol 2006; 209: 1825.
  • 6
    Garm A, Oskarsson M & Nilsson D-E. Box jellyfish use terrestrial visual cues for navigation. Curr Biol 2011; 21: 798803.
  • 7
    Herring PJ. The Biology of the Deep Ocean. Oxford University Press: Oxford, 2002.
  • 8
    Land MF. Crustacea. In: Photoreception and Vision in Invertebrates (Ali MA, editor), Plenum Press: New York, 1984; pp. 401438.
  • 9
    Gregory RL. Origins of eyes – with speculations on scanning eyes. In: Vision and Visual Dysfunction, Vol.2. (Cronly-Dillon JR & Gregory RL, editors), Macmillan: Basingstoke, 1991; pp. 5259.
  • 10
    Pettigrew JD, Collin SP & Ott M. Convergence of specialised behaviour, eye movements and visual optics in the sandlance (Teleostei) and the chameleon (Reptilia). Curr Biol 1999; 9: 421424.
  • 11
    Piatigorsky J. Gene Sharing and Evolution. Harvard University Press: Cambridge, 2007.
  • 12
    Walls GL. The Vertebrate Eye and its Adaptive Radiation. Cranbrook Institute of Science: Bloomington Hills, MI. Reprinted (1967) Haffner, New York, 1942.
  • 13
    Ott M & Schaeffel F. A negatively powered lens in the chamaeleon. Nature 1996; 373: 692694.
  • 14
    Harkness L. Chamaeleons use accommodation cues to judge distance. Nature 1977; 267: 346351.
  • 15
    Snyder AW & Miller WH. Telephoto lens system of falconiform eyes. Nature 1978; 275: 127129.
  • 16
    Sivak JG, Hildebrand T & Lebert C. Magnitude and rate of accommodation in diving and non-diving birds. Vision Res 1985; 25: 925933.
  • 17
    Williams DS & McIntyre P. The principal eyes of a jumping spider have a telephoto component. Nature 1980; 288: 578580.
  • 18
    Forster L. Target discrimination in jumping spiders. In: Neurobiology of Arachnids (Barth FG, editor), Springer: Berlin, 1985; pp. 249274.
  • 19
    Land MF. Movements of the retinae of jumping spiders in response to visual stimuli. J Exp Biol 1969; 51: 471493.
  • 20
    Blest AD & Land MF. The physiological optics of Dinopis subrufus L. Koch: a fish lens in a spider. Proc R Soc Lond B Biol Sci 1977; 196: 198222.
  • 21
    Buschbeck E, Sbita SJ & Morgan RC. Scanning behavior by larvae of the predaceous diving beetle, Thermonectes marmoratus (Coleoptera: Dytiscidae) enlarges visual field prior to prey capture. J Comp Physiol A 2007; 193: 973982.
  • 22
    Goodman LJ. Organization and physiology of the insect dorsal ocellar system. In: Handbook of Sensory Physiology. Vol. VII/6C (Autrum H, editor), Springer: Berlin, 1981; pp. 201286.
  • 23
    Stange G. The ocellar component of flight equilibrium control in dragonflies. J Comp Physiol A 1981; 141: 335347.
  • 24
    Salvini-Plawen L. von & Mayr E. On the evolution of photoreceptors and eyes. Evol Biol 1977; 10: 207263.
  • 25
    Young JZ. A Model of the Brain. Oxford University Press: Oxford, 1964.
  • 26
    Denton EJ. On the organization of reflecting surfaces in some marine animals. Philos Trans R Soc Lond B Biol Sci 1970; 258: 285313.
  • 27
    Locket NA. Adaptations to the deep-sea environment. In: Handbook of Sensory Physiology, Vol. VII/5 (Crescitelli F., editor), Springer: Berlin, 1977; pp. 67192.
  • 28
    Exner S. The physiology of the compound eyes of insects and crustaceans. Translated from the German by R.C.Hardie (1989). Republished by Springer-Verlag, 1891.
  • 29
    Munk O. Hvirveldyrøjet. Berlingske Forlag: København, 1980.
  • 30
    Friederichs HF. Beiträge zur Morphologie und Physiologie der Sehorgane der Cicindeliden (Col.). Zeitschrift für Morphologie und Ökologie der Tiere 1932; 21: 1172.
Thumbnail image of

Michael Land

Vision has been central to my scientific career, first in the study of animal eyes, and then human eye movements. After a zoology degree from Cambridge and a PhD from University College London, where I worked on the reflecting eyes of scallops, I spent four years in Berkeley, mainly studying the scanning eyes of jumping spiders. I came to the University of Sussex in 1971, and since then have worked on vision in flies and deep-sea animals - especially shrimp, which also have reflecting eyes. During a spell in Australia I worked on butterfly vision with Dan-Eric Nilsson. He was co-author of a book Animal Eyes published in 2002, where we attempt to summarise the optical mechanisms of all types of eye. After 1990 I turned to human vision, particularly the eye movements we make to obtain the information we need for actions. I am currently Emeritus Professor of Neurobiology at Sussex.