Файл: Pye D. Polarised light in science and nature (IOP)(133s).pdf

Figure 9.5. The eye of the honey bee showing the dorsal rim area (DRA in black) where about 150 upward-facing ommatidia contain polarisation-sensitive ultraviolet-sensitive cells, responsible for analysing the polarisation pattern of the sky.

dorsal rim units (figure 9.5) and they have no twist to their retinula cells. The cells are therefore highly dichroic as a whole and are thus ideally adapted to perform rapid assessment of the polarisation pattern of the sky (see chapter 6). It is this band of dorsal rim ommatidia that is responsible for the sky compass navigational abilities of the bee. Because they point upwards they are unlikely ever to be used for discriminating colours, of flowers for example, and the possible ambiguity described previously is therefore irrelevant. In some of these ommatidia the ninth cell is not short but runs up the bundle with the others, while the microvilli, with their dichroic pigment molecules, may be arranged not in just two orthogonal directions in different cells but in three or even more directions within the retinula of one ommatidium.

A different and rather more ‘extravagant’ solution to this problem of colour/polarisation ambiguity is found in the remarkable eyes of mantis shrimps, Squilla. Most crustacea, like insects, have compound eyes. Those of mantis shrimps contain as many as ten different sensitive pigments so their colour vision must be quite outstanding. Perhaps because of this variety, in the mid-band of the eye there are horizontal rows of ommatidia that ‘look at’ the same regions of the visual field, instead of all pointing in different directions. Two rows of these midband ommatidia are sensitive to polarisation in the broad range from blue to yellow light. So in these animals colour vision and polarisation analysis are served by different specialised ommatidia that face in the same direction. The function of polarisation sensitivity in mantis shrimps is as yet unknown.

Ants also show discrimination of sky polarisation. Some especially revealing studies have been made on desert ants, Cataglyphis, of North

Africa and the Middle East. These insects must forage across open spaces, often with no visible landmarks to guide them. Even if their outward path is quite tortuous they can still take a direct line back to their nest over as much as 200 m. Furthermore, unlike bees, they can be followed by experimenters who can manipulate their visual environment all the way by devices held over the walking ants. For instance the sun can be obscured by a shadowing card, leaving most of the sky clearly visible, or conversely the sky polarisation can be changed by a large sheet of polaroid while leaving the sun visible through it. It has been known since 1911 that some ants, such as the garden ant Lasius, rely predominantly on the sun—using so-called sun compass orientation.

Shielding them from the sun, and at the same time showing them a reflection of the sun in a mirror, makes them change or even reverse their direction of travel accordingly. But the desert ant relies more on the pattern of sky polarisation than on the sun itself. In this case each eye has about a thousand ommatidia of which about 80 dorsal rim units are specialised polarisation-sensitive, ultraviolet units. Reliable navigation back to the nest without delay is essential in the desert where long exposure to the hot sand can be lethal. Perhaps a small cloud over the sun could impose a dangerous delay if only the sun compass were used.

Polarisation sensitivity of this kind now seems to be very common among the insects and examples of a different kind of application by dragonflies and water boatman bugs have been given in chapter 7. As in the bees and ants, it is generally especially associated with the ultraviolet-sensitive cells. An exception is seen in crickets where the dorsal rim cells are all blue sensitive dichroic units although ultravioletand green-sensitive cells occur elsewhere in the eye. The probable explanation is that whereas bees and ants are active by day, crickets tend to be nocturnal, when levels of ultraviolet light are low. It has been suggested that this may be common in other nocturnal insects although some diurnal flies also seem to be most sensitive to polarisation at blue wavelengths. Strangely, a few insects, such as certain water beetles, show best polarisation analysis at the (to them) very long wavelengths of yellow–green light and this has not yet been very convincingly explained. In general, the main significance of polarisation analysis seems to be the ability it confers of using the sky pattern for orientation, not just for homing as in bees and ants but also for keeping to a straight track. Many different insects, including butterflies, flies and mosquitoes, appear to be reluctant to fly across open spaces when the sky is heavily overcast. Even on a sunny day, it is not possible to keep the sun in sight

when moving through a wood whereas patches of sky can nearly always be seen here and there through the canopy.

The presence of microvilli with aligned receptor molecules, and their orthogonal orientation in different cells of the ommatidium is common to insects, crustacea (including Daphnia) and many other arthropods in general. There are some variations, for instance bunches of microvilli of different cells may actually interdigitate in an orthogonal ‘dovetailed’ arrangement so that light passes through each in turn along the axis of the cells. But the basic arrangement is ubiquitous. In a number of cases it has been possible to observe dichroicity within individual receptor cells by measuring light absorption for different directions of polarisation. It is tempting, therefore, to suppose that all these animals are able to detect the direction of polarisation. But this may not always be so. Twisting of the bundle of receptor cells has already been mentioned as a reason why the basic dichroicity may be lost—and twisting is not easy to observe as it requires cell orientation to be followed through a lengthy series of electron microscope sections. In many compound eyes the retinula cells within an ommatidium may also be coupled together electrically so that their excitation is shared. This increases the overall sensitivity to light but it destroys any individual dichroicity. Any such coupling through the membranes of adjacent cells cannot be seen under the electron microscope and is only revealed by electrophysiological recording of the responses of cells to light.

Actual evidence for polarisation analysis can take a number of forms, either behavioural or physiological. Examples of spontaneous behaviour in response to polarised light, and of changes of behaviour when the direction of polarisation is manipulated experimentally, have been described already. A number of investigators have also trained animals to respond to the direction of polarisation in an artificial visual stimulus. For instance an animal may be rewarded for pressing a pedal when shown vertically polarised light but not when shown horizontally polarised light. Such experiments, however, may be misleading even when the responses appear to be completely reliable. The problem is that an animal working for a reward will identify any clue as to which is the ‘correct’ signal and it may not be the clue the experimenter has in mind. This is especially true of polarised light stimuli. For instance, the light is sure to be reflected from some part of the enclosure or from objects within it. But light polarised in one direction is often (indeed generally, to some extent) reflected more strongly than light polarised at right angles (chapter 7). So an animal that is completely insensitive

to polarisation might nevertheless spot differences in the brightness of a reflection and learn to respond quite consistently to achieve a highly ‘correct’ score.

A more direct way of demonstrating sensitivity to the direction of polarisation is to record the excitation of individual receptor cells or their messages to the brain in single nerve fibres. A very fine electrode is inserted into a receptor cell or nerve fibre and is used to record the electrical changes brought about by light stimuli. When polarised light is used, the direction of polarisation is varied to give the strongest response, and this is compared with the response to light of equal intensity polarised at right angles to the first. A large difference means that the response is greatly dependent on polarisation and therefore the animal as a whole could be affected; a small difference of response naturally means the unit is only weakly sensitive to the direction of polarisation. Then if two responses with orthogonal directions can be found in different units, the animal should be able to analyse the direction of polarisation. This approach has now been used successfully to study bees, ants, crustaceans and a range of other creatures .

Incidentally, it seems to be commonly assumed that two orthogonal sensitivities are necessary to analyse the direction of polarisation but of course one strongly sensitive direction will do provided that the animal is able to rotate its head, or at least its eye, and make successive observations. Conversely, even two orthogonal sensitivities can give ambiguous answers if no movement is made—vertically and horizontally sensitive cells will be equally stimulated by oblique polarisation at 45◦. The problems are exactly the same as with the polariscopes described in chapters 1 and 2. In the case of reflections from water, which is always horizontal, no such ambiguity need arise and two orthogonal sensitivities will do for the water boatman bug or the dragonfly (chapter 7). But the sky compass analyser of the dorsal rim of the bee’s eye has a more difficult task that justifies a greater complexity, with more than two sensitive directions.

Octopus, squid and cuttlefish are very highly developed molluscs. Their eyes are not multifaceted compound eyes but simple or ‘camera’ eyes, optically very like our own although they have evolved quite independently. The receptor cells of the retina contain their receptor molecules within microvilli and half the cells have their microvilli running horizontally while the other half have them vertically. Two cells of each kind form tetrads throughout the retina, suggesting the presence of polarisation sensitivity. Recently, behavioural tests have shown that an

octopus can indeed analyse the direction of polarisation. They have been trained to respond to lamps that have a polarisation contrast pattern, say a vertical polarisation in the centre with a horizontal polarisation in the surround, or vice versa. They can even respond when the difference of direction is as small as 20◦ instead of 90◦. The risk of ‘false’ cues being involved here can be made very small. Moreover, octopuses are also able to distinguish between a clear piece of Pyrex glass with no strains and an otherwise identical piece that has been subjected to heat stress to create internal strains. These produce birefringence that is only made visible by polarisation analysis (see chapter 2 and figure 2.10)—both glasses look identical to our eyes.

The significance of all this for vision in their environment is not yet clear but perhaps it enhances their ability to detect the shiny fishes on which they prey. Silvery scales reflect the colour of their surroundings, which is perfect camouflage underwater, but the reflections can be strongly polarised in ways that do not match the polarisation produced by scattering in the surrounding water. So polarisation analysis can be used to ‘break’ the camouflage. Quite apart from this, some cuttlefish apparently use polarisation to communicate with each other. As well as signalling by their well-known transient patterns of light, shade and colour, they can also produce patterns of light polarisation in their skin by means of iridophores: cells that are iridescent because they produce multiple internal reflections which interfere and polarise the light they reflect (see chapter 7). Like all iridophores, these cells are not themselves changeable but in cuttlefish they can be quickly concealed or exposed by tiny overlying sacs of black pigment controlled by radial muscle fibres. It has been suggested that these cuttlefish are thus able to communicate ‘secretly’ with each other without disturbing their general camouflage patterns of shade and colour that alone would be seen by any predator that lacks polarisation sensitivity.

Spiders are much more closely related to insects and crustacea, since all three belong to the Arthropoda, or joint-limbed animals with external skeletons. Yet their eyes are of the simple or ‘camera’ type (actually most insects also have three simple eyes called ocelli, but they are quite crude). The eyes of spiders vary considerably but most have four pairs of eyes looking in somewhat different directions. In the hunting or ‘wolf’ spiders, Arctosa (figure 9.6), the two ‘principal’, or anterior median eyes have large lenses and form good images. They are used for orientation and respond to the polarisation pattern of the ‘sky compass’. Both they and the upward-looking posterior median

Figure 9.6. Left: the eyes of a wolf spider, Arctosa species, as seen from above; the forward-facing anterior median or ‘principal’ eyes (am) and upward facing posterior median (pm) eyes have been shown to be sensitive to the direction of polarisation. Centre: the ‘squinty’ boat-shaped polarisation sensitive posterior median (pm) eyes of the spider Drassodes as seen from above. Right: a section across a pm eye of Drassodes, with two reflecting layers (shown black) that lie obliquely below the masses of light-sensitive cells (dotted). The arrows show how light that is not absorbed on its first passage through one group of receptor cells is reflected back through them, across the eye and then through the other group of cells. Polarisation by reflection (chapter 7) acts together with the dichroicity of the receptor cells to enhance their polarisation sensitivity. There is no lens and no image is formed in these eyes but the two set at right angles are well adapted to detect the polarisation of sky light.

eyes have microvilli which, in parts of the retina, are arranged in orthogonal groupings. Electrical recordings have confirmed the presence of polarisation sensitivity in both wolf spiders and jumping spiders.

In 1999 a new kind of eye was discovered in another spider called Drassodes. Here again the two posterior median eyes are on top of the head, looking vertically upwards (figure 9.6). They are boat shaped and orientated at right angles to each other in a way that is aptly described as ‘squinty’. In common with many other spider eyes, the retina is backed by a reflective layer which, as in the eyes of cats and many other nocturnal creatures, directs any unabsorbed light back through the retina. Such reflections from spiders’ eyes can often be seen at night if one holds a torch close to one’s face and looks around a garden lawn for instance. In Drassodes the reflection is bright blue and is polarised along the axis of the eye so that a rotating polariser extinguishes the reflection from each eye in turn. As shown in figure 9.6, the reflective

layer is folded into two flat plates forming a V-shaped trench along the sides of the ‘boat’. Light therefore reaches the retinal cells either directly or after reflection by one or both reflectors. Even an isolated piece of a reflector is found to polarise light and this clearly enhances the inherent dichroicity of the 60 or so main retinal cells whose microvilli run along the longer axis of the eye. The electrical responses of the cells peak in the ultraviolet and show an unusually high dependence on the direction of polarisation. These eyes have no lenses so they can form no image and they receive light from a wide angle, about 125 ◦. The spiders are active around dawn and dusk when the sky overhead is polarised in a rather simple north–south band (figure 6.3 and chapter 6). Behavioural tests support the suggestion that the two eyes act together as a polarising sky compass to enable the spiders to return to their lairs after hunting forays. The structure of the polarising reflector layers is not yet known but it probably involves multiple layers and a Brewster-type reflection (chapter 7).

In the retinas of vertebrate eyes, the receptor cells, rods and cones, are quite different in structure from any of the foregoing. Instead of having microvilli, they have a stack of transverse disclike membranes that lie across the light path and contain their receptor molecules in random orientation (figure 9.7) in their walls. Indeed within the rather fluid membranes the individual molecules are free to rotate around the optical axis so that they all point randomly in different directions and fluctuate continuously. This arrangement would not be expected to be dichroic and optical measurements have confirmed this. In fact, when light is shone experimentally across the cell, there is marked dichroicity since all the receptor molecules are at least parallel with the planes of their holding membranes and strongly absorb when the direction of polarisation is in this plane. But in life and a whole eye, light never shines across the cells, only down their length.

The freedom of receptor molecules to rotate within the membranes of vertebrate retinal cells has been demonstrated by some ingenious observations. First, the visual cells of a frog were treated with a chemical (glutaraldehyde) that killed the cells and ‘set’ the fluid membranes, so preventing further movement of their constituents. A bright flash of polarised light shining in the normal direction along the length of the cells then left them strongly dichroic for further beams of light along this axis. Any receptor molecules that had absorbed light with the direction of polarisation of the first flash were inactivated (when they are ‘bleached’ and no longer absorb light), so further absorption in a

Figure 9.7. The light-sensitive cells, both rods and cones, of vertebrate eyes contain disclike double membranes that lie transversely to the light path (arrows). The light-sensitive molecules (dashes) are contained within these membranes but they point in all directions and are therefore not sensitive to the polarisation of light coming from the lens. This arrangement contrasts with the regularly orientated molecules in the eyes of bees and other invertebrates (figure 9.4).

second flash was favoured for light polarised at right angles to the first, in the molecules that were still receptive. But in living cells, where these molecules are free to rotate rapidly and randomly, such dichroicity persists for only a very brief time, around 20 µs (millionths of a second). Restoration of sensitivity by the regeneration of bleached molecules is very slow by comparison, taking a matter of minutes, and so cannot explain the rapid loss of dichroicity in normal eyes after the first flash. In this way paired flashes of polarised light can be used to measure the rate of molecular rotation within the membranes of living visual cells.

From all this one would not expect vertebrates to be sensitive to the direction of polarisation of light, and until a few years ago this was believed to be so. One exception may be found in the anchovy fish, Anchoa species, where the cone cells of the retina are contorted so that the membranous discs run almost lengthways and so lie edge on to the light path (figure 9.8). The cones are arranged in the retina in vertical rows (with rods in between) and there are two kinds of cone that alternate in each row, with the two types having their discs arranged orthogonally, either vertically or horizontally. This arrangement is expected to make each cone dichroic and the two orthogonal directions could give the anchovy the ability to detect the direction of polarisation. The way in which one kind of cone fits neatly under the other may increase overall sensitivity since the horizontally polarised light that passes through the short cones may then be absorbed by the horizontally sensitive long cones and not be wasted. In this respect there is a functional resemblence