retroyou nostal(G) 2002 - 2003 : Illusions in Flight

Illusions in Flight

Visual IllusionsVestibular Illusions

An illusion is a false percept. An orientational illusion is a false percept of one's position or motion--either linear or angular--relative to the plane of the earth's surface. A great number of orientational illusions occur during flight: some named, others unnamed; some understood, others not understood. Those that are sufficiently impressive to cause pilots to report them, whether because of their repeatability or because of their emotional impact, have been described in the aeromedical literature and will be discussed here. The illusions in flight are categorized into those resulting from visual misperceptions and those involving vestibular errors.

Visual Illusions

Shape ConstancySize Constancy Aerial PerspectiveAbsent Focal CuesAbsent Ambient CuesAutokinesis Vection IllusionsFalse Horizons and Surface PlanesOther False Ambient Cues

We shall organize the visual illusions in flight according to whether they involve primarily the focal mode of visual processing or primarily the ambient mode. Although this categorization is somewhat arbitrary and may seem too coarse in some cases, it serves to emphasize the dichotomous nature of visual orientation information processing. We begin with illusions involving primarily focal vision. (Many of the visual illusions related in this section were first described by Pitts.23)

Shape Constancy

To appreciate how false shape constancy cueing can create orientational illusions in flight, consider the example provided by a runway that is constructed on other than level terrain. Figure 14a shows the pilot's view of the runway during an approach to landing and demonstrates the linear perspective and foreshortening of the runway that the pilot associates with a 3° approach slope. If the runway slopes upward 1° (a rise of only 35 m in a 2-km runway), the foreshortening of the runway for a pilot on a 3° approach slope is substantially less (the height of the retinal image of the runway is greater) than it would be if the runway were level. This can give the pilot the illusion of being too high on the approach. The pilot's natural response to such an illusion is to reshape the image of the runway by seeking a shallower approach slope (Fig. 14b). This response, of course, could be hazardous. The opposite situation results when the runway slopes downward. To perceive the accustomed runway shape under this condition, the pilot flies a steeper approach slope than usual (Fig. 14c).

Figure 14

Figure 14. Effect of runway slope on the pilot's image of runway during final approach (left) and potential effect on the approach slope angle flown (right). a. Flat runway -- normal approach. b. An up-sloping runway creates the illusion of being high on approach -- pilot flies the approach too low. c. A down-sloping runway has the opposite effect.

Size Constancy

Size constancy is very important in judging distance, and false cues are frequently responsible for aircraft mishaps due to illusions of focal visual origin. The runway width illusions are particularly instructive in this context. Figure 15a shows the accustomed runway width and a normal approach. A runway that is narrower than that to which a pilot is accustomed can create a hazardous illusion on the approach to landing. Size constancy causes the pilot to perceive the narrow runway to be farther away (i.e., the aircraft is higher) than is actually the case; hence, the pilot may flare too late and touch fly a steeper approach than if the approach terrain were level (Fig. 16a). If the approach terrain slopes up to the runway, on the other hand, the pilot tends to fly a less steep approach (Fig. 16b). Although the estimation of height above the approach terrain depends on both focal and ambient vision, the contribution of focal vision is particularly clear: consider the pilot who looks at a building below and, seeing it to be closer than such buildings usually are, seeks a higher approach slope. By the same token, focal vision and size constancy are responsible for poor height and distance judgments pilots sometimes make when flying over terrain having an unfamiliar composition (Fig. 17). A reported example of this is the tendency to misjudge the approach height when landing in the Aleutians, where the evergreen trees are much smaller than those to which most pilots are accustomed. Such height-estimation difficulties are by no means restricted to the approach and landing phases of flight. One fatal mishap occurred during air combat training over the Southwest desert when the pilot of a high-performance fighter aircraft presumably misjudged the aircraft's height over the desert floor because of the small, sparse vegetation and was unable to arrest a deliberate descent to a ground-hugging altitude.

Figure 15

Figure 15.Effect of runway width on the pilot's image of runway (left) and the potential effect on approach flown (right). a. Accustomed width--normal approach. b. A narrow runway makes the pilot feel the aircraft is higher than it actually is, which results in the approach being too low and the flare being too late. c. A wide runway gives the illusion of being closer than it actually is--the pilot tends to approach too high and flares too soon.


Figure 16

Figure 16. Potential effect of the slope of the terrain under the approach on the approach slope flown. a. The terrain slopes down to the runway; the pilot thinks the approach is too shallow and steepens it. b. Upsloping terrain makes the pilot think the aircraft is too high, which is corrected by making the approach too shallow.

Aerial Perspective

Aerial perspective also can play a role in deceiving the pilot, and the approach-to-landing regime again provides examples. In daytime, fog or haze can make a runway appear farther away as a result of the loss of visual discrimination. At night, runway and approach lights in fog or rain appear less bright than they do in clear weather and can create the illusion that they are farther away. It has even been reported that a pilot can have an illusion of being banked to the right, for example, if the runway lights are brighter on the right side of the runway than they are on the left. Another hazardous illusion of this type can occur during approach to landing in a shallow fog or haze, especially during a night approach. The vertical visibility under such conditions is much better than the horizontal visibility, so that descent into the fog causes the more distant approach or runway lights to diminish in intensity at the same time that the peripheral visual cues are suddenly occluded by the fog. The result is an illusion that the aircraft has pitched up, with the concomitant danger of a nose-down corrective action by the pilot.

Absent Focal Cues

A well-known pair of approach-to-landing situations that create illusions because of the absence of adequate focal visual orientation cues are the smooth-water (or glassy-water) and snow-covered approaches. In a seaplane, one's perception of height is degraded substantially when the water below is still. For that reason, a seaplane pilot routinely just sets up a safe descent rate and waits for the seaplane to touch down, rather than attempting to flare to a landing when the water is smooth. A blanket of fresh snow on the ground and runway also deprives the pilot of visual cues with which to estimate height above ground, thus making the approach extremely difficult. Again, approaches are not the only regime in which smooth water and fresh snow cause problems. A number of aircraft have crashed as a result of pilots maneuvering over smooth water or snow-covered ground and misjudging their height above the surface.

Absent Ambient Cues

Two runway approach conditions that create considerable difficulty for the pilot, by requiring focal vision alone to accomplish what is normally accomplished with both focal and ambient vision, are the black-hole and whiteout approaches. A black-hole approach is one that is made on a dark night over water or unlighted terrain to a runway beyond which the horizon is indiscernible, the worst case being when only the runway lights are visible (Fig. 18). Without peripheral visual cues to help orient the aircraft relative to the earth, the pilot tends to feel that the aircraft is stable and situated appropriately but that the runway itself moves about or remains malpositioned (is down-sloping, for example). Such illusions make the black-hole approach difficult and dangerous, and often result in a landing far short of the runway. A particularly hazardous type of black-hole approach is one made under conditions wherein the earth is totally dark except for the runway and the lights of a city situated on rising terrain beyond the runway. Under these conditions, the pilot may try to maintain a constant vertical visual angle for the distant city lights, thus causing the aircraft to arc far below the intended approach slope as it gets closer to the runway (Fig. 19).24 An alternative explanation is that the pilot falsely perceives through ambient vision that the rising terrain is flat, which leads to a lower-than-normal approach.

An approach made under whiteout conditions can be as difficult as a black-hole approach, and for essentially the same reason--lack of sufficient ambient visual orientation cues. There are actually two types of whiteout, the atmospheric whiteout and the blowing-snow whiteout. In the atmospheric whiteout, a snow-covered ground merges with a white overcast, creating a condition in which ground textural cues are absent and the horizon is indistinguishable. Although visibility may be unrestricted in the atmospheric whiteout, there is essentially nothing to see except the runway or runway markers; an approach made in this condition must therefore be accomplished with a close eye on the altitude and attitude instruments to prevent spatial disorientation and inadvertent ground contact. In the blowing-snow whiteout, visibility is restricted drastically by snowflakes, and often those snowflakes have been driven into the air by the propeller or rotor wash of the affected aircraft. Helicopter landings on snow-covered ground are particularly likely to create blowing-snow whiteouts. Typically, the helicopter pilot tries to maintain visual contact with the ground during the sudden rotor-induced whiteout, gets into an unrecognized drift to one side, and shortly thereafter contacts the ground with sufficient lateral motion to cause the craft to roll over. Pilots flying where whiteouts can occur must be made aware of the hazards of whiteout approaches, as the disorientation induced usually occurs unexpectedly under visual rather than instrument meteorological conditions.

Another condition in which a pilot is apt to make a serious misjudgment is in closing on another aircraft at high speed. When the pilot has numerous peripheral visual cues by which to establish both the aircraft's position and velocity relative to the earth and the target's position and velocity relative to the earth, the difficulty of tracking and closing is not much different from what it would be on the ground giving chase to a moving quarry. When relative position and closure rate cues must come from foveal vision alone, however--as is generally the case at altitude, at night, or under other conditions of reduced visibility--the tracking and closing problem is much more difficult.


Figure 18
Figure 18. Effect of loss of ambient orientation cues on the perception of runway orientation during a black-hole approach. a. When ambient orientation cues are absent, the pilot feels horizontal and (in this example) perceives the runway to be tilted left and upsloping. b. With the horizon visible, the pilot can orient correctly with peripheral vision and the runway appears horizontal in central vision.

An overshoot, or worse, a midair collision, can easily result from the perceptual difficulties inherent in such circumstances, especially when the pilot lacks experience in an environment devoid of peripheral visual cues.


Figure 19

  Figure 19. A common and particularly dangerous type of black-hole approach, in which the pilot falsely perceives the distant city to be on flat terrain and arcs below the desired approach slope

A related phenomenon that pilots need especially to be aware of is the dip illusion. It occurs during formation flying at night, when one aircraft is in trail behind another. To avoid wake turbulence and maintain sight of the lead aircraft, the pilot in trail flies at a small but constant angle below the lead aircraft by placing the image of the lead aircraft in a particular position on the windscreen and keeping it there. Now suppose the pilot is told to "take spacing" (separate) to 5 nautical miles (10 km). For every 1° below lead the trailing pilot flies, he or she is lower than lead by 1.7% (sin 1°) of the distance behind lead. Thus, if the trail pilot is 2° below lead and keeps the image of the lead aircraft at the same spot on the windscreen all the way back to 5 miles, the trail aircraft will descend to about 1100 ft (350 m) below the lead aircraft. To make matters worse, when the aircraft in trail slows down to establish separation, its pitch attitude increases by several degrees, and if the pilot does not compensate for this additional angle and tries to maintain the lead aircraft image in the same relative position, the altitude difference between the two aircraft can be doubled or even tripled. In the absence of ambient visual orientation cues, the pilot cannot detect the large loss of altitude without monitoring the flight instruments, and may inadvertently "dip" far below the intended flight path. Clearly this situation would be extremely hazardous if it were to occur at low altitude or during maneuvers in which altitude separation from other aircraft is critical.

One puzzling illusion that occurs when ambient visual orientation cues are minimal is visual autokinesis (Fig. 20). A small, dim light seen against a dark background is an ideal stimulus for producing autokinesis. After 6 to 12 seconds of visually fixating the light, one can observe it to move at up to 20° /sec in a particular direction or in several directions in succession, but there is little apparent displacement of the object fixated. In general, the larger and brighter the object, the less the autokinetic effect. The physiologic mechanism of visual autokinesis is not entirely understood. One suggested explanation for the autokinetic phenomenon is that the eyes tend to drift involuntarily, perhaps because of inadequate or inappropriate vestibular stabilization and that checking the drift requires efferent oculomotor activity having sensory correlates that create the illusion.

Whatever the mechanism, the effect of visual autokinesis on pilots is of some importance. Anecdotes abound of pilots who fixate a star or a stationary ground light at night, and seeing it move because of autokinesis, mistake it for another aircraft and try to intercept or join up with it. Another untoward effect of the illusion occurs when a pilot flying at night perceives a relatively stable tracked aircraft to be moving erratically when in fact it is not; the unnecessary and undesirable control inputs the pilot makes to compensate for the illusory movement of the target aircraft represent increased work and wasted motion at best and an operational hazard at worst.

To help avoid or reduce the autokinetic illusion, the pilot should try to maintain a well-structured visual environment in which spatial orientation is unambiguous. Because this is rarely possible in night flying, it has been suggested that (1) the pilot's gaze should be shifted frequently to avoid prolonged fixation of a target light; (2) the target should be viewed beside, through, or in some other reference to a relatively stationary structure such as a canopy bow; (3) the pilot should make eye, head, and body movements to try to destroy the illusion; and (4) as always, the pilot should monitor the flight instruments to help prevent or resolve any perceptual conflict. Equipping aircraft with more than one light or with luminescent strips to enhance recognition at night probably has helped reduce problems with autokinesis.


Figure 20

Figure 20. Visual autokinesis. A small, solitary light or small group of lights seen in the dark can appear to move, when in fact they are stationary.

So far, this section has dealt with visual illusions created by excessive orientation-processing demands being placed on focal vision when adequate orientation cues are not available through ambient vision or when strong but false orientation cues are received through focal vision. Ambient vision can itself be responsible for creating orientational illusions, however, whenever orientation cues received in the visual periphery are misleading or misinterpreted. Probably the most compelling of such illusions are the vection illusions. Vection is the visually induced perception of self-motion in the spatial environment and can be a sensation of linear self-motion (linear vection) or angular self-motion (angular vection).

Nearly everyone who drives an automobile has experiencedone very common linear vection illusion: when we are stopped at a stoplight and a large, presumably stationary vehicle in the adjacent lane creeps forward, a compelling illusion that our own car is creeping backward can result (prompting a swift but surprisingly ineffectual stomp on the brakes). Similarly, when one is sitting in a stationary train and the train on the adjacent track begins to move, a strong sensation that one's own train is moving in the opposite direction can be experienced (Fig. 21a). Linear vection is one of the things that make close formation flying so difficult, because the pilot can never be sure whether his or her own aircraft or that of the lead or wingman is responsible for the perceived relative motion.

Angular vection occurs when peripheral visual cues convey the information that one is rotating; the perceived rotation can be in pitch, roll, yaw, or other plane. Although angular vection illusions are not common in life, they can be generated readily in a laboratory by enclosing a subject in a rotating striped drum. Usually within 10 seconds after the motion begins, the subject perceives that he or she rather than the drum is rotating. A pilot can experience angular vection if the anticollision light on the aircraft is left on during flight through clouds or the revolving reflection provides a strong ambient visual stimulus, rotation in the yaw plane.

Another example of vection illusions is the so-called "Star Wars" named after the popular motion picture famous for its vection-inducing effects. This phenomenon involves linearly and angularly moving of ground lights off the curved inside surface of a fighter aircraft which create in the pilot disconcerting sensations of motion that conflict with the actual motion of the aircraft.

Fortunately, vection illusions are not all bad. The most advanced simulators depend on linear and angular vection to create the illusion of angular motion (Fig. 21b). When the visual flight environment is dynamically portrayed by flight simulators with a wide field-of-view and infinity-optics, the illusion of actual flight is so compelling that additional mechanical motion is often not even needed (although mechanically generated motion-onset cues do seem to improve the fidelity of the simulation).

False Horizons and Surface Planes

Sometimes the horizon perceived through ambient vision is not horizontal. Quite naturally, this misperception of the horizontal creates illusions in flight. A sloping cloud deck, for example, is very difficult to perceive as anything but horizontal if it extends for any great distance into the peripheral visual field (Fig. 22). Uniformly sloping terrain, particularly featureless terrain, can also create an illusion of horizontality with disastrous results, for the pilot thus deceived. Many aircraft have crashed as a result of pilot's entering a canyon with an apparently level floor, only to find that floor actually rose faster than the airplane could climb. At night, the lights of a city built on sloping terrain can create the false impression that the extended plane of the city lights is the horizontal plane of the earth's surface as already noted (Fig. 19). A distant rain shower can obscure the real horizon and create the impression of a horizon at the proximal edge (base) of the rainfall. If the shower is seen just beyond the runway during an approach to landing, the pilot can misjudge the pitch attitude of the aircraft and make inappropriate pitch corrections on the approach.

Figure 21

Figure 21. Vection illusions. a. Linear vection. In this example, the adjacent vehicle seen moving aft in the subject's peripheral vision causes the sensation of moving forward. b. Angular vection. Objects seen revolving around the subject in the flight simulator result in a perception of self-rotation in the opposite direction--in this case, a rolling motion to the right.

Pilots are especially susceptible to misperception of the horizontal while flying at night (Figs. 23a and 23b). Isolated ground lights can appear to the pilot as stars, creating the perception of a nose-high or one-wing-low attitude. Flying under such a false impression can, of course, be fatal. Frequently, no stars are visible because of overcast conditions. Unlighted areas of terrain can then blend with the dark overcast to create the illusion that the unlighted terrain is part of the sky. One extremely hazardous situation is that in which a takeoff is made over an ocean or other large body of water that cannot be distinguished visually from the night sky. Many pilots in this situation have perceived the shoreline receding beneath them to be the horizon, and some have responded to this false percept with disastrous nose-down control input.

Figure 22

Figure 22. A sloping cloud deck, which the pilot misperceives as a horizontal surface.

Pilots flying at high altitudes can sometimes experience difficulties in control of aircraft attitude, because at high altitudes the horizon is lower with respect to the plane of level flight than it is at the lower altitudes most pilots are accustomed to flying. As a reasonable approximation, angle of depression of the horizon in degrees equals the square root of altitude in kilometers. A pilot flying at an altitude of 49,000 ft (15 km) sees the horizon almost 4° below the extension of the horizontal plane of aircraft. By visually orienting to the view from the left cockpit window, a pilot might be inclined to fly with the left wing 4° down to level it with horizon. If the pilot does this and then looks out the right window, the wing would be seen 8° above the horizon, with half of that elevation due to the erroneous control input. The pilot also might experience problems with pitch control because the depressed horizon can cause a false perception of 4° nose-high pitch attitude.

Figure 23a

Figure 23b

Figure 23. Misperception of the horizontal at night. a. Ground lights appearing to be stars cause the earth and sky to blend and a false horizon to be perceived. b. Blending of overcast sky with unlighted terrain or water causes the horizon to appear lower than is actually the case.

Other False Ambient Cues

One very important aspect of ambient visual orientational cuing is in the stabilizing effect of the surrounding instrument panel, glare shield, canopy bow or windshield frame, and especially the reflections of panel and other cockpit structures off the windshield or canopy at night. The stable visual provided by these objects tends to cause the aircraft motion to appear not above the threshold for vestibular motion though it may be well above the usual threshold for vestibular motion. While flying at night or in instrument weather, a pilot may thus have a sense of security because no motion is felt, due to the apparently ambient visual environment. Of course, this falsely stabilizing effect occurs when the visual environment contains the usually valid ambient references (natural horizon, earth's surface, etc.).

Another result of false ambient visual orientational cuing is the sun illusion. On the ground, we are accustomed to seeing brighter surroundings above and darker below, regardless of the position of the sun. The direction of this gradient in light intensity thus helps us orient with respect to the surface of the earth. In clouds, however, such a gradient generally does not exist, and when it does, the lighter direction is generally toward the sun and the darker direction is away from it. But the sun is not always directly overhead; as a consequence, a pilot flying in a thin cloud layer may perceive falsely that the sun is directly overhead. This prompts the pilot to bank in the direction of the sun, hence the name of this illusion.

Finally, the disorienting effects of the northern lights and of aerial flares should be mentioned. Aerial refueling at night in high northern latitudes often is made quite difficult by the northern lights, which provide false cues of verticality to the pilot's peripheral vision. Similarly, when aerial flares are dropped, they may drift with the wind, creating false cues of verticality. Their motion also may create vection illusions. Another phenomenon associated with use of aerial flares at night is the "moth" effect. The size of the area on the ground illuminated by a dropped flare slowly decreases as the flare descends. Because of the size constancy mechanism of visual orientation discussed earlier, a pilot circling the illuminated area may tend to fly in a descending spiral with gradually decreasing radius. Another important factor is that the aerial flares can be so bright as to reduce the apparent intensity of the aircraft instrument displays and thereby minimize their orientational cuing strength.

Somatogyral Illusion, Oculogyral IllusionCoriolis IllusionSomatogravic Illusion, Inversion IllusionG-Excess EffectOculogravic IllusionsThe Leans

The vestibulocerebellar axis processes orientation information from the vestibular, visual, and other sensory systems. In the absence of adequate ambient visual orientation cues, the inadequacies of the vestibular and other orienting senses can result in orientational illusions. It is convenient and conventional to discuss the vestibular illusions in relation to the two functional components of the labyrinth that generate them--the semicircular ducts and the otolith organs.

Somatogyral Illusion

A somatogyral illusion is a false sensation of rotation (or absence of rotation) that results from misperceiving the magnitude or direction of an actual rotation. In essence, somatogyral illusions result from the inability of the semicircular ducts to register accurately a prolonged rotation, i.e., sustained angular velocity. When a person is subjected to an angular acceleration about the yaw axis, for example, the angular motion is at first perceived accurately because the dynamics of the cupula-endolymph system cause it to respond as an integrating angular accelerometer (i.e., as a rotation-rate sensor) at stimulus frequencies in the physiologic range (Fig. 24). If the acceleration is followed immediately by a deceleration, as usually happens in the terrestrial environment, the total sensation of turning one way and then stopping the turn is quite accurate (Fig. 25). If, however, the angular acceleration is not followed by a deceleration and a constant angular velocity results instead, the sensation of rotation becomes less and less and eventually disappears as the cupula gradually returns to its resting position in the absence of an angular acceleratory stimulus (Fig. 26). If we are subsequently subjected to an angular deceleration after a period of prolonged constant angular velocity, say after 10 seconds or so of constant-rate turning, our cupula-endolymph systems signal a turn in the direction opposite that of the prolonged constant angular velocity, even though we are really only turning less rapidly in the same direction. This is because the angular momentum of the rotating endolymph causes it to press against the cupula, forcing the cupula to deviate in the direction of endolymph flow, which is the same direction the cupula would deviate if we were to accelerate in the direction opposite the initial acceleration. Even after rotation actually ceases, the sensation of rotation in the direction opposite that of the sustained angular velocity persists for several seconds--half a minute or longer with a large decelerating rotational impulse. Another, more mechanistic, definition of the somatogyral illusion is "any discrepancy between actual and perceived rate of self-rotation that results from an abnormal angular acceleratory stimulus pattern." The term "abnormal" in this case implies the application of low-frequency stimuli outside the useful portion of the transfer characteristics of the semicircular duct system.

In flight under conditions of reduced visibility, somatogyral illusions can be deadly. The graveyard spin is the classic example of how somatogyral illusions can disorient a pilot with fatal results. This situation begins with the pilot intentionally or unintentionally entering a spin, let's say to the left (Fig. 27). At first, the pilot perceives the spin correctly because the angular acceleration associated with entering the spin deviates the cupulae the appropriate amount in the appropriate direction. The longer the spin persists, however, the more the sensation of spinning to the left diminishes as the cupulae return to their resting positions. On trying to stop the spin to the left by applying the right rudder, the angular deceleration causes the pilot to perceive a spin to the right, even though the only real result of this action was termination of the spin to the left. A pilot who is ignorant of the possibility of such an illusion is then likely to make counterproductive left-rudder inputs to negate the unwanted erroneous sensation of spinning to the right. These inputs keep the airplane spinning to the left, which gives the pilot the desired sensation of not spinning but does not bring the airplane under control. To extricate himself from this very hazardous situation, the pilot must read the aircraft flight instruments and apply control inputs to make the instruments give the desired readings (push right rudder to center the turn needle, in this example). Unfortunately, this may not be so easy to do. The angular accelerations created by both the multiple-turn spin and the pilot's spin-recovery attempts can elicit strong but inappropriate vestibulo-ocular reflexes, including nystagmus. In the usual terrestrial environment, these reflexes help stabilize the retinal image of the visual surround; in this situation, however, they only destabilize the retinal image because the visual surround (cockpit) is already fixed with respect to the pilot. Reading the flight instruments thus becomes difficult or impossible, and the pilot is left with only false sensations of rotation to rely on for spatial orientation and aircraft control.26

Figure 24
Figure 24. Transfer characteristics of the semicircular duct system as a function of sinusoidal stimulus frequency. Gain is the ratio of the magnitude of the peak perceived angular velocity to the peak delivered angular velocity; phase angle is a measure of the amount of advance or delay between the peak perceived and peak delivered angular velocities. Note that in the physiologic frequency range (roughly 0.05 to 1 Hz), perception is accurate; that is, gain is close to unity (0 dB) and phase angle is minimal. At lower stimulus frequencies, however, the gain drops off rapidly, and the phase shift approaches 90°, which means that angular velocity becomes difficult to detect and that angular acceleration is perceived as velocity. (Adapted from Peters.25)

Figure 25
Figure 25. Effect of the stimulus pattern on the perception of angular velocity. On the left, the high-frequency character of the applied angular acceleration results in a cupular deviation that is nearly identical to, and a perceived angular velocity that is nearly identical to, the angular velocity developed. On the right, the peak angular velocity developed is the same as that on the left, but the low-frequency character of the applied acceleration results in cupular deviation and perceived angular velocity that appear more like the applied acceleration than the resulting velocity. This causes one to perceive: (a) less than the full amount of the angular velocity, (b) absence of rotation while turning persists, (c) a turn in the opposite direction from that of the actual turn, and (d) that turning persists after it has actually stopped. These false percepts are somatogyral illusions.

Although the lore of early aviation provided the graveyard spin as an illustration of the hazardous nature of somatogyral illusions, a much more common example occurring all too often in modern aviation is the graveyard spiral (Fig. 28). In this situation, the pilot has intentionally or unintentionally entered into a prolonged turn with a moderate amount of bank. After a number of seconds in the turn, the pilot loses the sensation of turning because the cupula-endolymph system cannot respond to constant angular velocity. The percept of being in a bank as a result of the initial roll into the banked attitude also decays with time because the net gravitoinertial force vector points toward the floor of the aircraft during coordinated flight (whether the aircraft is in a banked turn or flying straight and level), and the otolith organs and other graviceptors normally signal that down is in the direction of the net sustained gravitoinertial force. As a result, when trying to stop the turn by rolling back to a wings-level attitude, the pilot feels not only a turning in the direction opposite to that of the original turn, but also a bank in the direction opposite to that of the original bank. Unwilling to accept this sensation of making the wrong control input, the hapless pilot rolls back into the original banked turn. Now the pilot's sensation is compatible with a desired mode of flight, but the flight instruments indicate a loss of altitude (because the banked turn is wasting lift) and a continuing turn. So the pilot pulls back on the stick and perhaps adds power to arrest the unwanted descent and regain the lost altitude. This action would be successful if the aircraft were flying wings-level, but with the aircraft in a steeply banked attitude it tightens the turn, serving only to make matters worse. Unless the pilot eventually recognizes what is occurring and rolls out of the unperceived banked turn, the aircraft will continue to descend in an ever-tightening spiral toward the ground, hence the name graveyard spiral.

Figure 26

Figure 26. Representation of the mechanical events occurring in a semicircular duct and resulting action potentials in the associated ampullary nerve during somatogyral illusions. The angular acceleration pattern applied is that shown in the right side of Figure 25.

Figure 27

Figure 27. The Graveyard Spin. After several turns of a spin the pilot begins to lose the sensation of spinning. When trying to stop the spin, the resulting somatogyral illusion of spinning in the opposite direction makes the pilot reenter the original spin. (The solid line indicates actual motion; the dotted line indicates perceived motion. illusion of spinning in the opposite direction makes the pilot reenter the original spin. (The solid line indicates actual motion; the dotted line indicates perceived motion.


Figure 28

Figure 28. The Graveyard Spiral. The pilot in a banked turn loses the sensation of being banked and turning. Upon trying to establish a wings-level attitude and stop the turn, the pilot perceives a bank and a turning in the opposite direction from the original banked turn. Unable to tolerate the sensation of making an inappropriate control input, the pilot banks back into the original turn.

Oculogyral Illusion

Whereas a somatogyral illusion is a false sensation, or lack of sensation, of self-rotation in a subject undergoing unusual angular motion, an oculogyral illusion is a false sensation of motion of an object viewed by such a subject.27 For example, if a vehicle with a subject inside is rotating about a vertical axis at a constant velocity and suddenly stops rotating, the subject experiences not only a somatogyral illusion of rotation in the opposite direction, but also an oculogyral illusion of an object in front moving in the opposite direction. Thus, a somewhat oversimplified definition of the oculogyral illusion is that it is the visual correlate of the somatogyral illusion; however, its low threshold and lack of total correspondence with presumed cupular deviation suggest a more complex mechanism. The attempt to maintain visual fixation during a vestibulo-ocular reflex elicited by angular acceleration is probably at least partially responsible for the oculogyral illusion. In an aircraft during flight at night or in weather, an oculogyral illusion generally confirms a somatogyral illusion: the pilot who falsely perceives a turning in a particular direction also observes the instrument panel to move in the same direction.

Coriolis Illusion

The vestibular Coriolis effect, also called the Coriolis cross-coupling effect, vestibular cross-coupling effect, or simply the Coriolis illusion, is another false percept that can result from unusual stimulation of the semicircular duct system. To illustrate this phenomenon, let us consider a subject who has been rotating in the plane of the horizontal semicircular ducts {roughly the yaw plane) long enough for the endolymph in those ducts to attain the same angular velocity as the head: the cupulae in the ampullae of the horizontal ducts have returned to their resting positions, and the sensation of rotation has ceased {Fig. 29a). If the subject then nods forward in the pitch plane, let's say a full 90° for the sake of simplicity, the horizontal semicircular ducts are removed from the plane of rotation and the two sets of vertical semicircular ducts are inserted into the plane of rotation {Fig. 29b). Although the angular momentum of the subject's rotating head is forcibly transferred at once out of the old plane of rotation relative to the head, the angular momentum of the endolymph in the horizontal duct is dissipated more gradually. The torque resulting from the continuing rotation of the endolymph causes the cupulae in the horizontal ducts to be deviated, and a sensation of angular motion occurs in the new plane of the horizontal ducts--now the roll plane relative to the subject's body. Simultaneously, the endolymph in the two sets of vertical semicircular ducts must acquire angular momentum because these ducts have been brought into the plane of constant rotation. The torque required to impart this change in momentum causes deflection of the cupulae in the ampullae of these ducts, and a sensation of angular motion in this plane--the yaw plane relative to the subject's body--results. The combined effect of the cupular deflection in all three sets of semicircular ducts is that of a suddenly imposed angular velocity in a plane in which no actual angular acceleration relative to the subject has occurred. In the example given, if the original constant-velocity yaw is to the right and the subject's head pitches forward, the resulting Coriolis illusion experienced is that of suddenly rolling and yawing to the right.

Figure 29

Figure 29. Mechanism of the Coriolis illusion. A subject rotating in the yaw plane long enough for the endolymph to stabilize in the horizontal semicircular duct (a) pitches the head forward (b). Angular momentum of the endolymph causes the cupula to deviate, and the subject perceives rotation in the new (i.e., roll) plane of the horizontal semicircular duct, even though no actual rotation occurred in that plane.

A particular perceptual phenomenon experienced occasionally by pilots of relatively high-performance aircraft during instrument flight has been attributed to the Coriolis illusion because it occurs in conjunction with large movements of the head under conditions of prolonged constant angular velocity. It consists of a sensation of rolling and/or pitching that appears suddenly after the pilot's attention has been diverted from the instruments in front and his or her head is moved to view some switches or displays elsewhere in the cockpit. This illusion is especially deadly because it is most likely to occur during an instrument approach, a phase of flight in which altitude is being lost rapidly and cockpit chores (e.g., radio frequency changes) repeatedly require the pilot to break up his instrument cross-check. The sustained angular velocities associated with instrument flying are insufficient to create Coriolis illusions of any great magnitude, however;28 and another mechanism (the G-excess effect) has been proposed to explain the illusory rotations experienced with head movements in flight.29 Even if not responsible for spatial disorientation in flight, the Coriolis illusion is useful as a tool to demonstrate the fallibility of our nonvisual orientation senses. Nearly every military pilot living today has experienced the Coriolis illusion in the Barany chair or some other rotating device as part of physiological training, and for most of these pilots it was then they first realized that their own orientation senses really cannot be trusted--the most important lesson of all for instrument flying.

Somatogravic Illusion

The otolith organs are responsible for a set of illusions known as somatogravic illusions. The mechanism of illusions of this type involves the displacement of otolithic membranes on their maculae by inertial forces so as to signal a false orientation when the resultant gravitoinertial force is perceived as gravitational (and therefore vertical). Thus, a somatogravic illusion can be defined as a false sensation of body tilt that results from perceiving as vertical the direction of a nonvertical gravitoinertial force. The illusion of pitching up after taking off into conditions of reduced visibility is perhaps the best illustration of this mechanism. Consider the pilot of a high-performance aircraft waiting at the end of the runway to take off. Here, the only force acting on the otolithic membranes is the force of gravity, and the positions of those membranes on their maculae signal accurately that down is toward the floor of the aircraft. Suppose the aircraft now accelerates on the runway, rotates, takes off, cleans up gear and flaps, and maintains a forward acceleration of 1 g until reaching the desired climb speed. The 1 G of inertial force resulting from the acceleration displaces the otolithic membranes toward the back of the pilot's head. In fact, the new positions of the otolithic membranes are nearly the same as they would be if the aircraft and pilot had pitched up 45°, because the new direction of the resultant gravitoinertial force vector, if one neglects the angle of attack and climb angle, is 45° aft relative to the gravitational vertical (Fig. 30). Naturally, the pilot's percept of pitch attitude based on the information from the otolith organs is one of having pitched up 45°; and the information from nonvestibular proprioceptive and cutaneous mechanoreceptive senses supports this false percept, because the sense organs subserving those modalities also respond to the direction and intensity of the resultant gravitoinertial force. Given the very strong sensation of a nose-high pitch attitude, one that is not challenged effectively by the focal visual orientation cues provided by the attitude indicator, the pilot is tempted to push the nose of the aircraft down to cancel the unwanted sensation of flying nose-high. Pilots succumbing to this temptation characteristically crash in a nose-low attitude a few miles beyond the end of the runway. Sometimes, however, they are seen to descend out of the overcast nose-low and try belatedly to pull up, as though they suddenly regained the correct orientation upon seeing the ground again. Pilots of carrier-launched aircraft need to be especially wary of the somatogravic illusion. These pilots experience pulse accelerations lasting 2 to 4 seconds and generating peak inertial forces of +3 to +5 Gx. Although the major acceleration is over quickly, the resulting illusion of nose-high pitch can persist for half a minute or more afterward, resulting in a particularly hazardous situation for the pilot who is unaware of this phenomenon. 30


Figure 30

Figure 30. A somatogravic illusion occurring on takeoff. The inertial force resulting from the forward acceleration combines with the force of gravity to create a resultant gravitoinertial force directed down and aft. The pilot, perceiving down to be in the direction of the resultant gravitoinertial force, feels the aircraft is in an excessively nose-high attitude. 

Do not be misled by the above example into believing that only pilots of high-performance aircraft suffer the somatogravic illusion of pitching up after takeoff. More than a dozen air transport aircraft are believed to have crashed as a result of the somatogravic illusion occurring on takeoff.31 A relatively slow aircraft, accelerating from 100 to 130 knots over a l0-second period just after takeoff, generates +0.16 Gx on the pilot. Although the resultant gravitoinertial force is only 1.01 G, barely perceptibly more than the force of gravity, it is directed 9° aft, signifying to the unwary pilot a 9° nose-up pitch attitude. Because many slower aircraft climb out at 6° or less, a 9° downward pitch correction would put such an aircraft into a descent of 3° or more--the same as a normal final-approach slope. In the absence of a distinct external visual horizon or, even worse, in the presence of a false visual horizon (e.g., a shoreline) receding under the aircraft and reinforcing the vestibular illusion, the pilot's temptation to push the nose down can be overwhelming. This type of mishap has happened at one particular civil airport so often that a notice has been placed on navigational charts cautioning pilots flying from this airport to be aware of the potential for loss of attitude reference.

Although the classic graveyard spiral was indicated earlier to be a consequence of the pilot's suffering a somatogyral illusion, it also can be said to result from a somatogravic illusion. A pilot who is flying "by the seat of the pants" applies the necessary control inputs to create a resultant G-force vector having the same magnitude and direction as that which the desired flight path would create. Unfortunately, any particular G vector is not unique to one particular condition of aircraft attitude and motion, and the likelihood that the G vector created by a pilot flying in this mode corresponds for more than a few seconds to the flight condition desired is remote indeed. Specifically, once an aircraft has departed a desired wings-level attitude because of an unperceived roll, and the pilot does not correct the resulting bank, the only way to create a G vector that matches the G vector of the straight and level condition is with a descending spiral. In this condition, as is always the case in a coordinated turn, the centrifugal force resulting from the turn provides a Gy force that cancels the Gy component of the force of gravity that exists when the aircraft is banked. In addition, the tangential linear acceleration associated with the increasing airspeed resulting from the dive provides a +Gx force that cancels the -Gx component of the gravity vector that exists when the nose of the aircraft is pointed downward. Although the vector analysis of the forces involved in the graveyard spiral is somewhat complicated, a skillful pilot can easily manipulate the stick and rudder pedals to cancel all vestibular and other nonvisual sensory indications that the aircraft is turning and diving. In one mishap involving a dark-night takeoff of a commercial airliner, the recorded flight data showed that the resultant G force which the pilot created by his control inputs allowed him to perceive his desired 10° to 12° climb angle and a net G force between 0.9 and 1.1 G for virtually the whole flight, even though he actually leveled off and then descended in an accelerating spiral until the aircraft crashed nearly inverted.

Inversion Illusion

The inversion illusion is a type of somatogravic illusion in which the resultant gravitoinertial force vector rotates backward so far as to be pointing away from rather than toward the earth's surface, thus giving the pilot the false sensation of being upside down. Figure 31 shows how this can happen. Typically, a steeply climbing high-performance aircraft levels off more or less abruptly at the desired altitude. This maneuver subjects the aircraft and pilot to a -Gz centrifugal force resulting from the arc flown just prior to level-off. Simultaneously, as the aircraft changes to a more level attitude, airspeed picks up rapidly, adding a +Gx tangential inertial force to the overall force environment. Adding the -Gz centrifugal force and the +Gx tangential force to the 1-G gravitational force results in a net gravitoinertial force vector that rotates backward and upward relative to the pilot. This stimulates the pilot's otolith organs in a manner similar to the way a pitch upward into an inverted position would. Even though the semicircular ducts should respond to the actual pitch downward, for some reason this conflict is resolved in favor of the otolith-organ information, perhaps because the semicircular-duct response is transient while the otolith-organ response persists, or perhaps because the information from the other mechanoreceptors reinforces the information from the otolith organs. The pilot who responds to the inversion illusion by pushing forward on the stick to counter the perceived pitching up and over backward only prolongs the illusion by creating more -Gz and +Gx forces, thus aggravating the situation. Turbulent weather usually contributes to the development of the illusion; certainly, downdrafts are a source of -Gz forces that can add to the net gravitoinertial force producing the inversion illusion. Again, do not assume one must be flying a jet fighter to experience this illusion. Several reports of the inversion illusion involve crews of large airliners who lost control of their aircraft because the pilot lowered the nose inappropriately after experiencing the illusion. Jet upset is the name for the sequence of events that includes instrument weather, turbulence, the inability of the pilot to read the instruments, the inversion illusion, a pitch-down control input, and difficulty recovering the aircraft because of resulting aerodynamic or mechanical forces.33


Figure 31

Figure 31. The inversion illusion. Centrifugal and tangential inertial forces during a level-off combine with the force of gravity to produce a resultant gravitoinertial force that rotates backward and upward with respect to the pilot, causing a false percept of suddenly being upside down. Turbulent weather can produce additional inertial forces that contribute to the illusion. (Adapted from Martin and Jones. 32)

G-Excess Effect

Whereas the somatogravic illusion results from a change in the direction of the net G force, the G-excess effect results from a change in G magnitude. The G-excess effect is a false or exaggerated sensation of body tilt that can occur when the G environment is sustained at greater than 1 G. For a simplistic illustration of this phenomenon, let us imagine a subject is sitting upright in a + 1 Gz environment and then tips the head forward 30° (Fig. 32). As a result of this change in head position, the subject's otolithic membranes slide forward the appropriate amount for a 30° tilt relative to vertical, say a distance of x JA-m. Now suppose that the same subject is sitting upright in a +2 Gz environment and again tips the head forward 30°. This time, the subject's otolithic membranes slide forward more than x JA-m because of the doubled gravitoinertial force acting on them. The displacement of the otolithic membranes, however, now corresponds not to a 30° forward tilt in the normal 1-G environment but to a much greater tilt, theoretically as much as 90° (2 sin 30° = sin 90°). The subject had initiated only a 30° head tilt, however, and expects to perceive no more than that. The unexpected additional perceived tilt is thus referred to the immediate environment, i.e., the subject perceives his or her aircraft to have tilted by the amount equal to the difference between the actual and expected percepts of tilt. The actual perceptual mechanism underlying the G-excess effect is more complicated than the illustration suggests: first, the plane of the utricular maculae is not really horizontal but slopes upward 20-30° from back to front; second, the saccular maculae contribute in an undetermined manner to the net percept of tilt; and third, as is usually the case with vestibular illusions, good visual orientational cues attenuate the illusory percept. But experimental evidence clearly demonstrates the existence of the G-excess effect. Perceptual errors of 10° to 20° are generated at 2 G, and at 1.5 G the errors are about half that amount.34,35

Figure 32

Figure 32. Mechanism of G-excess illusion. In this oversimplified illustration, the subject in a 1-G environment (upper half of figure) experiences the result of a 0.5-G pull on the utricular otolithic membranes when the head is tilted 30° off the vertical, and the result of a 1-G pull when the head is tilted a full 90°. The subject in a 2-G environment (lower half of figure) experiences the result of a 1-G pull when the head is tilted only 30°. The illusory excess tilt perceived by the subject is attributed to external forces (lower right). Note that the actual plane or the utricular macula slopes 20-30° upward.

In fast-moving aircraft, the G-excess illusion can occur as a result of the moderate amount of G force pulled in a turn--a penetration turn or procedure turn, for example. A pilot who has to look down and to the side to select a new radio frequency or to pick up a dropped pencil while in a turn should experience an uncommanded tilt in both the pitch and roll planes due to the G-excess illusion. As noted previously, the G-excess illusion may be responsible for the false sensation of pitch and/or roll generally attributed to the Coriolis illusion under such circumstances. The G-excess effect has recently become a suspect in a number of mishaps involving fighter/attack aircraft making 2- to 5.5-G turns at low altitudes in conditions of essentially good visibility. For some reason, the aircraft were overbanked while the pilots were looking out of the cockpit for an adversary, wingman, or some other object of visual attention, and as a result they descended into the terrain. In theory, the G-excess effect causes an illusion of underbank if the pilot's head is either facing the inside of the turn and elevated (Fig. 33) or facing the outside of the turn and depressed. If facing forward, the pilot would have an illusion of pitching up, i.e., clubing, during the turn. Thus, in any of these common circumstances, the pilot who fails to maintain a continuous visual reference to the earth's surface would likely cause the aircraft to descend in response to the illusory change of attitude caused by the G-excess effect. Perhaps in some of the mishaps mentioned, the pilot's view of the spatial environment was inadequate encompassing sky rather than ground, or perhaps G-induced tunnel vision was responsible for loss of ambient visual cues. In any case, it is apparent that the pilots failed to perceive correctly their aircraft attitude, vertical velocity, and height above the ground, i.e., they were spatially disoriented.


Figure 33

Figure 33. The G-excess illusion during a turn in flight. G-induced excessive movement of the pilot's otolithic membranes causes the pilot to feel an extra amount of head and body tilt, which is interpreted as an underbank of the aircraft when looking up to the inside of the turn. Correcting for the illusion, the pilot overbanks the aircraft and it descends.

The elevator illusion is a special kind of G-excess effect. Because of the way the utricular otolithic membranes are variably displaced with respect to their maculae by increases and decreases in +Gz force, false sensations of pitch and vertical velocity can result even when the head remains in the normal, upright position. When an upward acceleration (as occurs in an elevator) causes the net Gz force to increase, a sensation of climbing and tilting backward can occur. In flight, such an upward acceleration occurs when an aircraft levels off from a sustained descent. This temporary increase in +Gz loading can make pilots feel a pitch up and climb if their views of the outside world are restricted by night, weather, or head-down cockpit chores. Compensating for the illusory pitch up sensation, the pilot would likely put the aircraft back into a descent, all the while feeling that the aircraft is maintaining a constant altitude. In one inflight study of the elevator illusion, blindfolded pilots were told to maintain perceived level flight after a relatively brisk level-off from a sustained 2000-ft/min (10 m/sec) descent: the mean response of the six pilots was a 1300-ft/min (6.6 m/sec) descent.15 Clearly this tendency to reestablish a descent is especially dangerous during the final stage of a nonprecision instrument approach at night or in weather. Upon leveling off at the published minimum descent altitude, the pilot typically starts a visual search for the runway. In conjunction with failing to monitor the flight instruments during this critical time, the elevator illusion can cause the pilot to unwittingly put the aircraft into a descent, and thus squander the altitude buffer protecting the air

Oculogravic lllusion

The oculogravic illusion can be thought of as a visual correlate of the somatogravic illusion and occurs under the same stimulus conditions.36 A pilot who is subjected to the deceleration resulting from the application of speed brakes, for example, experiences a nose-down pitch because of the somatogravic illusion. Simultaneously, the pilot observes the instrument panel to move downward, confirming the sensation of tilting forward. The oculogravic illusion is thus the visually apparent movement of an object that is actually fixed relative to the subject during the changing direction of the net gravitoinertial force. Like the oculogyral illusion, the oculogravic illusion probably results from the attempt to maintain visual fixation during a vestibulo-ocular reflex elicited, in this case, by the change in direction of the applied G vector rather than by angular acceleration.

The elevator illusion was originally thought of as a visual phenomenon like the oculogravic illusion, except that the false percept was believed to result from a vestibulo-ocular reflex generated by a change in magnitude of the +Gz force instead of by a change in its direction. When an individual is accelerated upward, as in an elevator, the increase in +Gz force elicits a vestibulo-ocular reflex of otolith-organ origin (the elevator reflex) that drives the eyes downward. Attempting to stabilize visually the objects in a fixed position relative to the observer causes those objects to appear to shift upward when the G force is increased. The opposite effect occurs when the individual is accelerated downward; the reduction in the magnitude of the net gravitoinertial force to less than +1 Gz causes a reflex upward shift of the direction of gaze, and the immediate surroundings appear to shift downward. (The latter effect also has been called the oculoagravic illusion because of its occurrence during transient weightlessness.) Although the described visual effect undoubtedly contributes to the expression of the elevator illusion, it is not essential for its generation, since the illusion can occur even in the absence of vision, as just noted.

The Leans By far the most common vestibular illusion in flight is the leans. Virtually every instrument-rated pilot has had or will get the leans in one form or another at some time during his or her flying career. The leans consists of a false percept of angular displacement about the roll axis (i.e., an illusion of bank) and is frequently associated with a vestibulospinal reflex, appropriate to the false percept, that results in the pilot's actually leaning in the direction of the falsely perceived vertical (Fig. 34). The usual explanations of the leans invoke the known deficiencies of both otolith-organ and semicircular-duct sensory mechanisms. As indicated previously, the otolith organs are not reliable sources of information about the exact direction of the true vertical because they respond to the resultant gravitoinertial force, not to gravity alone. Furthermore, other sensory inputs can sometimes override otolith-organ cues and result in a false perception of the vertical, even when the gravitoinertial force experienced is truly vertical. The semicircular ducts provide false inputs in flight by responding accurately to some roll stimuli but not responding at all to others because they are below threshold. For example, a pilot who is subjected to an angular acceleration in roll so that the product of the acceleration and its time of application does not reach some threshold value, say 2°/sec, does not perceive the roll. Suppose that this pilot, who is trying to fly straight and level, is subjected to an unrecognized and uncorrected 2°/sec roll for 10 seconds: a 20° bank results. If the unwanted bank becomes suddenly apparent and is corrected by rolling the aircraft back upright with a suprathreshold rate, the pilot experiences only half of the actual roll motion that took place--the half resulting from the correcting roll. As the aircraft started from a perceived wings-Ievel position, the pilot upon returning to an actual wings-Ievel attitude is left with the illusion of having rolled into a 20° bank in the direction of the correcting roll and experiences the leans. Even though the pilot may be able to fly the aircraft properly by the deliberate and difficult process of forcing the attitude indicator to read correctly, the leans can last for many minutes, seriously degrading flying efficiency during that time.

Figure 34

Figure 34. The leans, the most common of all vestibular illusions in night. Falsely perceived to be in a right bank, but flying the aircraft straight and level by means of the flight instruments, this pilot is leaning to the left in an attempt to assume an upright posture compatible with the illusion of bank.

Interestingly, pilots frequently get the leans after prolonged turning maneuvers and not because of alternating subthreshold and suprathreshold angular motion stimuli. In a holding pattern, for example, the pilot rolls into a 3°/sec standard-rate turn, holds the turn for 1 minute, rolls out and flies straight for 1 minute, turns again for 1 minute, and so on until traffic conditions permit the continuation of the flight toward its destination. During the turning segments, the pilot initially feels the roll into the turn and accurately perceives the banked attitude. But as the turn continues, the percept of being in a banked turn dissipates and is replaced by a feeling of flying straight with wings level, both because the sensation of turning is lost when the endolymph comes up to speed in the semicircular ducts (somatogyral illusion) and because the net G force being directed toward the floor of the aircraft provides a false cue of verticality (somatogravic illusion). Upon rolling out of the turn, the pilot's perception is of a banked turn in the opposite direction. With experience, a pilot learns to suppress this false sensation quickly by paying strict attention to the attitude indicator. Sometimes, however, pilots cannot dispel the illusion of banking--usually when they are particularly busy, unfortunately. The leans also can be caused by misleading peripheral visual orientation cues, as mentioned in the section entitled "Visual Illusions." Roll vection is particularly effective in this regard, at least in the laboratory. One thing about the leans is apparent: there is no single explanation for this illusion. The deficiencies of several orientation-sensing systems in some cases reinforce each other to create the illusion; in other cases, the inaccurate information from one sensory modality for some reason is selected over the accurate information from others to create the illusion. Stories have surfaced of pilots suddenly experiencing the leans for no apparent reason at all, or even of experiencing it voluntarily by imagining the earth to be in a different direction from the aircraft. The point is that one must not think that the leans, or any other illusion for that matter, occurs as a totally predictable response to a physical stimulus; there is much more to perception than stimulation of the end-organs.