Portal de Periódicos da CAPES

Animal navigation

2004; Elsevier BV; Volume: 14; Issue: 6 Linguagem: Inglês

10.1016/j.cub.2004.02.049

1879-0445

James L. Gould,

Insect and Arachnid Ecology and Behavior

Nearly all animals move in an oriented way, but navigation is something more: the directed movement toward a goal, as opposed to steering toward or away from, say, light or gravity. Navigation involves the neural processing of sensory inputs to determine a direction and perhaps distance. For instance, if a honey bee were to seek food south of its hive, it would depart from home with the sun to its left in the morning, but to its right in the afternoon. Several trends reflecting favorably on natural selection and poorly on human imagination characterized early studies of navigation. One tendency was the assumption that animals sense at most the same cues as we do. Thus, being blind to our own blindness, it came as a total surprise when honey bees and many other species were found to be able to see UV light. As navigation depends on the processing of such cues, the number of 'new' senses uncovered in the past fifty years has greatly expanded our thinking about what may be going on in the minds of animals – and there is no reason to assume the list is complete. To UV must be added polarized light, infra-red light, special odors (pheromones), magnetic fields, electric fields, ultrasonic sounds and infrasonic sounds. The second crippling propensity is what navigation pioneer Donald Griffin called our innate 'simplicity filter': the desire to believe that animals do things in the least complex way possible. Experience, however, tells us that animals whose lives depend on accurate navigation are uniformly overengineered. Not only do they frequently wring more information out of the cues that surround them than we can, or use more exotic or weaker cues than we find conceivable, they usually come equipped with alternative strategies – a series of backups between which they switch depending on which is providing the most reliable information. A honey bee, for instance, may set off for a goal using its time-compensated sun compass. When a cloud covers the sun, it may change to inferring the sun's position from UV patterns in the sky and opt a minute later for a map-like strategy when it encounters a distinctive landmark. Lastly, it may ignore all of these cues as it gets close enough to its goal to detect the odors or visual cues provided by the flowers. This is not to say that animals do not often rely on approximations and neural shortcuts to simplify these daunting tasks. A third stumbling block has been our presumption that, because the earliest cases studied involved 'imprinting' (irreversible one-trial learning), animals must have simple navigation programs, which need merely to be calibrated to the local contingencies. This is just what at least some relatively short-lived animals do – like honey bees for instance, who rarely forage for more than three weeks. But as we will see, most animals live longer, and in consequence many need to recalibrate themselves. Finally, most researchers deliberately ignored or denigrated the evidence for cognitive processing in navigating animals. This legacy of behaviorism (the school of psychology that denied instinct) puts a ceiling on the maximum level of mental activity subject to legitimate investigation. There are many navigating animals whose behavior lacks any hint of cognitive intervention. However, the obvious abilities of hunting spiders and honey bees to plan novel routes make it equally clear that phylogenetic distance to humans is no sure guide to the sophistication of a species' orientation strategies. Some navigating animals move over relatively short distances, a home range, whereas others migrate to new home ranges. The home-range species include mice, honey bees, hunting spiders, non-migratory birds and territorial fish. How do these animals come to know about their range? One of the problems we inherited from behaviorism was the assumption that exploratory behavior must be rewarded. However, many species examine their surroundings voluntarily and, in fact, do so in detail. Consider mice, which not only gallop endlessly in running wheels, but actually prefer difficulty, such as square 'wheels', or wheels with barriers that must be jumped. Given a 430 meter long opaque three-dimensional maze of pipes, mice will work out the shortest path within three days, and without reward. At least some animals must be innately motivated to learn about their home range. They acquire information in advance of its need and use it during navigation. For nocturnal animals, like mice, this drive is very strong indeed. When they encounter suitably familiar and distinctive landmarks, they know where they are. But to navigate, they need to determine direction. This can be achieved in two ways, and mice use both: they can use another landmark from their mental map and triangulate the direction of the goal, or they can use a landmark-independent compass like the earth's magnetic field. These animals can also navigate perfectly well, even if the habitat fails to provide useful landmarks. They will remember the direction and length of each leg of their outward journey and integrate the result when they are ready to return and set off home. The inevitable systematic errors in measuring distances and angles make this a less desirable strategy; thus, landmarks are preferred if available. For diurnal animals as well as nocturnal ones with a view of the sky, celestial cues provide considerable information about direction. Most obviously, the sun moves across the sky from east to west, and once the relationship between azimuth and time of day is memorized, the animal has a highly accurate compass (Figure 1). The patterns of polarized light generated by sunlight scattering in the atmosphere indicate the sun's location with good accuracy, particularly in the UV range. Thus, they can be substituted if the sun is not directly visible, even before the sun rises or after it sets. There is a sizable school of thought that holds that many birds and insects actually prefer UV-polarized light to the sun itself. Sufficiently memorizing the pattern of stars can allow a nocturnal animal to determine the pole point – the spot, due north or south, about which the stars rotate. If an animal lives long enough, the relationship between the sun's azimuth and the time of day will change with the seasons, and some recalibration will be needed. Finally, the earth's magnetic field can serve as a cue for navigation, though less accurate it would seem, as animals tend to rely on it only in the absence of landmark and celestial cues. Of course, diurnal animals can also have landmarks available, and thus employ either a true cognitive map (a map-like representation of all the landmarks in the home range) or a memorized series of landmarks along a familiar route. These two strategies are not mutually exclusive: honey bees can do either, depending on which seems to work best under the circumstances. However, bees and other navigating insects have poor visual resolution; ours is a thousand times better. Thus, a clear and unambiguous landmark for a mouse or human will generally be totally useless to a bee. It is safe to say that for home-range animals, the key area of research today is the formation and use of cognitive maps: we know far too little about what may be the most common weapon in the navigational arsenal (Figure 2). Migrating animals face a greater set of challenges. For most waterfowl, the young follow their parents during the first autumnal migration, and thus might memorize the route to the target. But what do the young geese, swans, and ducks learn that allows them to return to their birthplace in the spring and then back to the wintering grounds in subsequent years? One possibility is that the birds recall every landmark in sequence along the way. However, this seems unlikely, as the return routes are often quite different. Perhaps they might integrate the directions and distances, and thus be able to set a return course. As we will see, however, other birds do not do it this way. Putting aside for a moment the question of how the birds learn where the wintering ground is, any mechanism is going to require a compass. The prime suspects for cues are the sun, polarized light, the stars and the earth's magnetic field. I have mentioned the problem of recalibration of the sun's path as the year wears on, and home-range landmarks are important only at the end of the return, but the magnetic field is another matter for these animals. The North Magnetic Pole lies 1200 km from the Geographical North Pole, among the Queen Elizabeth Islands in Canada. This produces a discrepancy, the so called 'declination', between true, geographic north and magnetic north. For a home-range creature, this inconsistency hardly matters as it is unvarying. But for a migrator, movement to or from high latitudes, which are the breeding areas of many birds, almost inevitably involves changing declination. To use a magnetic backup under overcast, the bird has to update (recalibrate) its declination correction. Only in the last decade has this problem been recognized and it subsequently has been solved for at least two species. In essence, once on the move, the birds recalibrate to celestial cues, such as polarized light or the pole point, as often as possible at high latitudes. A little thought will convince you that the 'parallax' between the two poles declines with decreasing latitude. Accordingly, recalibration seems to occur less frequently at lower latitudes, and may not occur at all in tropical animals. As most birds migrate at night to minimize heat stress, not much attention has been paid to the recalibration that is necessary to use a sun compass. The sun's pattern of movement changes with season, such that the steepest angle of climb is found at the summer solstice. This means slow rates of azimuth movement near dawn and dusk but rapid rates near noon. The angle also differs with latitude, being steeper with decreasing latitude. Thus, any long-lived long-distance migrant will need to adjust. The same argument applies to a lesser degree to nocturnal migrants. As the seasons change, some new stars become visible at night while others disappear. To infer the pole point through broken clouds, the animal's map of the sky must be updated. And as the migrants move south in the fall, new sets of stars in the southern sky appear, while northern stars slip below the horizon. Clearly, changes in both season and latitude make relearning the stars essential. Only fairly recently has this constant updating been demonstrated, Although there is no a priori reason that this ought to be so, nocturnal migrants calibrate their star pole to the magnetic pole. Instead of simply taking the pole point as the true guide, the birds constantly recalibrate the magnetic pole to the geographic pole, and then the geographic pole to the magnetic pole. This apparently unnecessary operation seems to occur almost on a daily basis while moving at high latitudes. Some migrants, perhaps most, know which latitude they are at. In essentially every case, e.g. birds, fish, turtles, lobsters, the cue is the dip angle of the lines of magnetic force. The lines are approximately horizontal at the equator and vertical at the magnetic pole and the angle varies systematically with latitude. In theory, animals could obtain the same information from the sun's noon elevation, but I know of no case in which this traditional human solution is used. Knowledge of latitude tells migrators when to make turns and when to stop. This strategy does not provide a measure of longitude. The 'problem of longitude' bedeviled human navigators until very recently. With only a compass and a latitude detector, a migrant cannot do much about drift due to crosswinds or currents. Some autumnal migrants do not seem to need much longitude information: they winter over large longitudinal bands. However, when they come back in spring, many seem to return accurately, even to the same nest box, like house wrens. As many species follow different routes, the odds seem vanishingly small that the same low-resolution strategy would put them in their small home range, thus enabling them to use a cognitive map to find the previous season's nest area. The apparent answer to this conundrum is provided by a map sense, which is strikingly evident in homing pigeons, but similarly applies to other species such as sharks, salmon, sea turtles and spiny lobsters. The argument for a map is simple: if you were kidnapped, deprived of cues during transport, and released in an unfamiliar area, to get home you would need a map to know where you are, and a compass to find the homeward direction. Numerous experiments demonstrate that: the map does not depend on celestial cues; almost no conceivable manipulation on the outward journey from the loft affects performance; pigeons know where they are, either at release or after circling a couple of times; they make systematic departure errors at specific sites; flight directions become more accurate as the bird flies; and departure error and scatter decline with increasing distance of the release site from the loft. Most pigeons appear to calibrate the map on the first day they take wing and circle above and around their loft. A pigeon moved to another loft after this date will return home even years later. There are two possible explanations for the pigeon map. One proposes that the birds associate the odor borne on the wind with the direction the breeze blows. There is an amazing body of literature on this hypothesis, but four facts argue against it: the rearing technique used in the studies does not permit the normal imprinting flight; the scatter of pigeons using olfactory cues should increase with the distance of the release site from the loft, but exactly the opposite is observed; these results have been very difficult to repeat, and pigeons have a notoriously poor olfactory sensitivity. The other hypothesis invokes magnetic fields. The earth has numerous magnetic-field gradients, and these gradients generally form a skewed grid from which, in theory, map location could be inferred. To use this grid for navigation, a bird would have to fly around its home area to determine the local directions and rates of change of the gradients. At the release site, it would have to measure the same values and extrapolate its location based on the home values. If the degree of accuracy of this measurement is independent of release distance, as it should be, release-bearing scatter should decrease with distance, as it does. Any difference between the gradients at the site versus the loft would cause the birds to misread their maps, leading to release site biases. Any error in the gradient strength or field values at the loft relative to the larger-scale pattern around it should generate a systematic misreading of map locations, producing a wide-scale pattern of release-site biases that are symmetrical to the gradient direction. This has been observed for several lofts. Releases on days of magnetic storms, when compass readings are unaffected, but map readings would be, shift site-biases and increase return times. Release at magnetic anomalies, where iron-rich rock increases local field strength enough to alter map readings but not compass direction, should leave the birds reoriented or disoriented. Indeed, at smooth gradients they are reoriented, whereas at irregular ones they are ludicrously disoriented (Figure 3). Pigeons, like most animals with maps, possess enormous numbers of magnetite grains in the ethymoid sinus. There are species in which magnetic sensitivity may be mediated in other ways, for example magnetic induction in elasmobranchs, or paramagnetic effects in the visual pigments of silvereye birds, but magnetite-based compasses seem to fit the behavioral data in nearly all cases. In species using a map for navigation, there are far too many of these grains for a mere compass, but enough to easily determine map location at least at an order of magnitude better than the 2 km estimated from returns of visually impaired pigeons. Pigeons subjected to gradually increasing and decreasing intense magnetic fields should have their magnetite grains remagnetized. If the map sense is directional, like a typical hand compass, the birds should be disoriented. On the other hand, if the map is axial, that is, if it cannot distinguish between north and south, the orientation should be better. Birds treated with magnetic fields are indeed better oriented. However, if the birds are subjected to a sharp intense pulse, either kind of map organ should be severely impaired, which is exactly what is observed. Whichever cues the map sense involves, it seems likely that it is this information which allows low-resolution first-time migrants to achieve high-resolution returns–and high-resolution southward trips in future years. Far and away the biggest question in long-distance navigation is the nature of the map sense. Second is the issue of recalibration, which is still puzzling. The interaction of these specific learning programs doubtlessly holds many magnificent secrets.The editors of Current Biology welcome correspondence on any article in the journal, but reserve the right to reduce the length of any letter to be published. All Correspondence containing data or scientific argument will be refereed. Items for publication should either be submitted typed, double-spaced to: The Editor, Current Biology, Elsevier Science London, 84 Theobald's Road, London, WC1X 8RR, UK, or sent by e-mail to [email protected] The editors of Current Biology welcome correspondence on any article in the journal, but reserve the right to reduce the length of any letter to be published. All Correspondence containing data or scientific argument will be refereed. Items for publication should either be submitted typed, double-spaced to: The Editor, Current Biology, Elsevier Science London, 84 Theobald's Road, London, WC1X 8RR, UK, or sent by e-mail to [email protected]