ALBERT

Notizen: Summary 1. Chases in which male flies (Fannia canicularis) pursue other flies were studied by filming such encounters from directly below. Males will start to chase whenever a second fly comes within 10–15 cm (Fig. 3). 2. Throughout these chases there was a continuous relationship between the angle (θ e ) made by the leading fly and the direction of flight of the chasing fly, and the angular velocity of the chasing fly (ω f ). This relation was approximately linear, with a slope of 20 ° s−1 per degree θ e (Figs. 4–7). 3. The maximum correlation between ω f and θite occurs after a lag of approximately 30 ms, which represents the total delay in the system (Fig. 8). 4. In the region close to the chasing fly's axis (θite less than about 35 °) a second mechanism exists in which the angular velocity of the chasing fly (ω f ) is controlled by the relative angular velocity of the leading fly (ω e ), rather than its relative position. The ratio of ω f to ω e in this region is approximately 0.7. 5. Using the results in 2–4 above, and an empirically determined relation between the angular and forward velocities of the chasing fly, it was possible to simulate the flight path of the chasing fly, given that of the leading fly (Fig. 11). Because these simulations predict correctly the manoeuvres and outcomes of quite complicated chases, it is concluded that the control system actually used by the fly is accurately described by conclusions 2–4. 6. The physiological implications of this behaviour, and the possible function of chasing, are discussed.

Notizen: Summary 1. The visually guided flight behaviour of groups of male and femaleSyritta pipiens was filmed at 50 f.p.s. and analysed frame by frame. Sometimes the flies cruise around ignoring each other. At other times males but not females track other flies closely, during which the body axis points accurately towards the leading fly. 2. The eyes of males but not females have a forward directed region of enlarged facets where the resolution is 2 to 3 times greater than elsewhere. The inter-ommatidial angle in this “fovea” is 0.6°. 3. Targets outside the fovea are fixated by accurately directed, intermittent, open-loop body saccades. Fixation of moving targets within the fovea is maintained by “continuous” tracking in which the angular position of the target on the retina (Θ e) is continuously translated into the angular velocity of the tracking fly ( $$\dot \Phi _p $$ ) with a latency of roughly 20 ms ( $$\dot \Phi _p = k \Theta _e $$ , wherek≏30 s−1). 4. The tracking fly maintains a roughly constant distance (in the range 5–15 cm) from the target. If the distance between the two flies is more than some set value the fly moves forwards, if it is less the fly moves backwards. The forward or backward velocity ( $$\dot F_p $$ ) increases with the difference (D-D 0) between the actual and desired distance ( $$\dot F_p = k^\prime (D - D_0 )$$ ), wherek′=10 to 20 s−1). It is argued that the fly computes distance by measuring the vertical substense of the target image on the retina. 5. Angular tracking is sometimes, at the tracking fly's choice,supplemented by changes in sideways velocity. The fly predicts a suitable sideways velocity probably on the basis of a running averageΘ e , but not its instantaneous value. Alternatively, when the target is almost stationary, angular tracking may bereplaced by sideways tracking. In this case the sideways velocity ( $$\dot S$$ ) is related toΘ e about 30 ms earlier ( $$\dot S_p = k\prime \prime \Theta _e $$ , wherek″=2.5 cm · s−1 · deg−1), and the angular tracking system is inoperative. 6. When the leading fly settles the tracking fly often moves rapidly sideways in an arc centred on the leading fly. During thesevoluntary sideways movements the male continues to point his head at the target. He does this not by correctingΘ e , which is usually zero, but by predicting the angular velocity needed to maintain fixation. This prediction requires knowledge of both the distance between the flies and the tracking fly's sideways velocity. It is shown that the fly tends to over-estimate distance by about 20%. 7. When two males meet head on during tracking the pursuit may be cut short as a result of vigorous sideways oscillations of both flies. These side-to-side movements are synchronised so that the males move in opposite directions, and the oscillations usually grow in size until the males separate. The angular tracking system is active during “wobbling” and it is shown that to synchronise the two flies the sideways tracking system must also be operative. The combined action of both systems in the two flies leads to instability and so provides a simple way of automatically separating two males. 8. Tracking is probably sexual in function and often culminates in a rapid dart towards the leading fly, after the latter has settled. During these “rapes” the male accelerates continuously at about 500 cm · s−2, turning just before it lands so that it is in the copulatory position. The male rapes flies of either sex indicating that successful copulation involves more trial and error than recognition. 9. During cruising flight the angular velocity of the fly is zero except for brief saccadic turns. There is often a sideways component to flight which means that the body axis is not necessarily in the direction of flight. Changes in flight direction are made either by means of saccades or by adjusting the ratio of sideways to forward velocity ( $$\dot S/\dot F$$ ). Changes in body axis are frequently made without any change in the direction of flight. On these occasions, when the fly makes an angular saccade, it simultaneously adjusts $$\dot S/\dot F$$ by an appropriate amount. 10. Flies change course when they approach flowers using the same variety of mechanisms: a series of saccades, adjustments to $$\dot S/\dot F$$ , or by a mixture of the two. 11. The optomotor response, which tends to prevent rotation except during saccades, is active both during cruising and tracking flight.

Notizen: Abstract Experiments by Fabre (1915), Thorpe (1950), Chmurzynski (1964), and most recently Gould (1986) suggest that insects have “maps” of their terrain which enable them to find their way directly to a goal when they are displaced several hundred metres from it. This paper discusses what might constitute an insect's map in terms of a two-part computational model. The first part describes how an insect reaches a goal when the insect is sufficiently close that it can see some of the landmarks which are visible from the goal. The second part considers the problem of navigating when there is no similarity between the view from the release-site and the view from the goal. We start from a model designed to explain how a bee might return to a goal using a two-dimensional “snapshot” of the landscape seen from the goal (Collett and Cartwright 1983). To guide its return, the model bee continuously compares its snapshot with its current retinal image and moves so as to reduce the discrepancy between the two. Bees can only be guided in the right direction by the difference between current retinal image and snapshot when there is some resemblance between the two. In a realistically cluttered world, snapshot and retinal image become very dis-similar only a short distance from the goal. To increase the distance from which a model bee can return, the bee takes two snapshots at the goal. The first snapshot excludes landmarks near to the goal and the second snapshot includes them. With close landmarks filtered from both snapshot and retinal image, the match between the two deteriorates gradually as the bee moves away from the goal. A model bee using a filtered snapshot and image finds its way back to the neighbourhood of the goal from a relatively long distance (Fig. 2). The bee then switches to the second snapshot and is guided to the precise spot by its memory of the close landmarks. For longer range guidance, the model bee is equipped with an album of snapshots, each taken at a different location within the terrain. Linked to each snapshot is a vector encoding the distance and direction from the place where the snapshot was taken to the hive. When the bee is displaced to a new position, it selects the snapshot which best matches its current image and follows the associated home-vector back to the hive (Fig. 3). Such a hive-centred map can also be used to devise novel routes to places other than the hive. For instance, a bee can reach a foraging site from anywhere in its terrain by adding the home-vector recalled at the starting position to a vector specifying the distance and direction of the foraging site from the hive. The sum of these two vectors defines a direct trajectory to the foraging site.

Notizen: [Auszug] Under some circumstances, Diptera and Hymenoptera learn visual shapes retinotopically, so that they only recognize the shape when it is viewed by the same region of retina that was exposed to it during learning,. One use of such retinotopically stored views is in guiding an insect's path to a ...

Notizen: [Auszug] Desert ants returning from a foraging trip to their nest navigate both by path integration and by visual landmarks. In path integration, ants compute their net distance and direction from the nest throughout their outward and return journeys, and so can always return directly home from their ...

Notizen: [Auszug] Cataglyphid ants travelling between their nest and feeding site follow familiar routes along which they are guided by views of the surrounding landscape,. On bare terrain, with no landmarks available, ants can still navigate using path integration. They continually monitor their net distance ...

Notizen: Summary 1. Males of many species of hoverfly hover in one spot ready to pursue passing objects, presumably in order to catch a mate. We have filmed two of the larger species as they begin their pursuit of an approaching projectile shot at them from a peashooter. Flies do not turn and fly towards the projectile as they would if they were tracking (Land and Collett, 1974). Instead they adopt an interception path, accelerating at a uniform rate (30–35 m · s−2) approximately in the direction in which the target is moving (Fig. 1). 2. If the projectile is made to reverse direction, the male does not respond to the change in the target's course for about 90 ms (Fig. 2). Thus, in contrast to the situation later (Fig. 3), the start of a pursuit is not under continuous sensory control. The fly selects its course when it first sees the target and maintains its interception path as an open-loop response. 3. Males only need to catch conspecifics and can thus assume that their quarry has a typical speed and, since it is of a uniform size, will be detected at a predetermined distance. These assumptions mean that the approach angle of the target at the moment of first sighting can be specified by the image velocity of the target across the retina ( $$\dot \theta _e $$ ). Since the male also “knows” its own acceleration, the direction of flight which will enable it to intercept the target can also be specified in terms of the initial position (θ e) and velocity ( $$\dot \theta _e $$ ) of the target image (Figs. 6 and 7). 4. We show that interception should occur if the male obeys the simple rule that the size of the turn it makes (Δφ) is given by $$\Delta \phi \simeq \theta _e - 0.1{\text{ }}\dot \theta _e \pm 180^\circ $$ . Our data indicate that the initial turns of the males do obey this rule (Fig. 8), and that males are “designed” to catch females travelling at about 8 m · s−1. 5. If males are calibrated to catch targets of a particular size and speed, they will be unable to intercept projectiles that are very different in these respects. When large, slowly moving projectiles are launched to pass by the fly, its attempted interception path is predictably inappropriate (Fig. 9).

Notizen: Summary Gerbils (Meriones unguiculatus) can specify the location of a goal by means of visual landmarks and will return to such a goal from different starting positions in the vicinity of the landmarks. To discover whether landmark-cues are used continuously during an approach to the goal, gerbils were trained to forage for sunflower seeds close to a single illuminated light-bulb on the floor of an arena. As they approached the bulb, it was switched off and another bulb in a variable position with respect to the first turned on. On 52 out of 71 trials the gerbils changed their trajectory (latency ca. 240 ms) to aim for the newly lit bulb (Fig. 1 A, B). On the remaining trials, gerbils maintained their original course towards the first bulb as though it were still lit and then paused after a longer delay before eventually changing direction (Fig. 1C). Thus, an approach to a beacon is usually under continuous visual control. This ensures that the gerbil will reach its goal correctly despite any inaccuracies in its initial computation of its approach. When switches were made between more complex arrays of landmarks, the gerbils' behaviour was less clear-cut. Possible reasons for this difference are suggested.

Notizen: Summary 1. Honeybees learn the location of a source of sucrose in relation to the positions of visual landmarks. It is shown here that bees pay most attention to those landmarks which are close to the sucrose and which will therefore guide a bee most accurately to its goal. 2. Individually marked bees were trained to collect sucrose from a reservoir whose location was specified by an array of landmarks, with individual landmarks placed at different distances from the reservoir. Once trained, the bees' flight path was recorded in tests with the sucrose removed. In these tests, the array of landmarks was transformed from the training configuration to provide two possible sites for the sucrose. One was defined by the landmarks close to the sucrose reservoir in training, the other by more distant landmarks. In all experiments (Figs.1, 2 and 3), bees spent more time flying in the site defined by the close landmarks. They did so even when the apparent sizes of the close and distant landmarks viewed from the reservoir were the same. We conclude that bees learn the distance of landmarks from the goal, and that, during their search for the goal, they weight nearer landmarks more heavily than distant ones. 3. Landmarks are also weighted according to their (apparent) size. Bees were trained to find sucrose at a reservoir placed equidistantly from different-sized landmarks. Tests were given with the training array stretched to provide two possible sites for the sucrose. One site was specified by the large, the other by the small landmarks. Bees spent more time flying in the site defined by the larger landmarks (Fig. 5). 4. We discuss, with the aid of computer simulation, different ways in which the distance between landmark and goal could influence a bee's searching behaviour.

Notizen: Summary This paper addresses two questions. 1. Does Schistocerca gregaria detect edges which are defined solely by velocity-contrast, that is by the difference in the image speeds generated by an object and its background when the locust moves? 2. Is the locust's ability to measure the distance of a target by motion parallax independent of the relative motion between target and back-ground? A locust walking on a circular platform was surrounded by a stationary cylinder which was lined with an irregular texture. Against this background, the insect viewed 3 stationary, equidistant targets. One target was black, one grey and the last was textured like the cylinder. Peers and jumps were aimed preferentially at the textured and black targets showing that targets can be detected by virtue of their velocity-contrast with the background. When textured targets were wide, jumps were seen to be aimed at the targets' edge. To assess whether velocity-contrast between target and background distorts distance-estimates, we used jump-velocity as a measure of apparent distance and examined how it varied with different arrangements of target and background. When a textured background is close to a target or the target is very wide, velocity contrast is small. The locust's jump-velocity is then 10% greater than when velocity-contrast is increased by making the background distant or the target narrow. This suggests that the locust is efficient at separating signals encoding absolute motion from those encoding relative motion.