Dung beetles ignore landmarks for straight-line orientation
- First Online:
- Cite this article as:
- Dacke, M., Byrne, M., Smolka, J. et al. J Comp Physiol A (2013) 199: 17. doi:10.1007/s00359-012-0764-8
- 573 Views
Upon locating a suitable dung pile, ball-rolling dung beetles shape a piece of dung into a ball and roll it away in a straight line. This guarantees that they will not return to the dung pile, where they risk having their ball stolen by other beetles. Dung beetles are known to use celestial compass cues such as the sun, the moon and the pattern of polarised light formed around these light sources to roll their balls of dung along straight paths. Here, we investigate whether terrestrial landmarks have any influence on straight-line orientation in dung beetles. We find that the removal or re-arrangement of landmarks has no effect on the beetle’s orientation precision. Celestial compass cues dominate straight-line orientation in dung beetles so strongly that, under heavily overcast conditions or when prevented from seeing the sky, the beetles can no longer orient along straight paths. To our knowledge, this is the only animal with a visual compass system that ignores the extra orientation precision that landmarks can offer.
Navigating insects use a number of compass cues to find their way to a food source and back home again. These include the sun, the pattern of polarised skylight and landmarks (for reviews, see Collett and Collett 2009; Wehner 1992, 2003). By monitoring and integrating the directions and distance travelled on the foraging journey, many insects continuously compute the most direct route back to their nest. This system of navigation is known as path integration (Collett and Collett 2000; Wehner and Srinivasan 2003; Wolf 2011). Due to the noise inherent in all sensory systems however, an animal’s perception of the direction in which it has travelled will invariably accumulate errors that will increase with path length (Cheung et al. 2007; Müller and Wehner 1988). To correct for these accumulated errors and to recalibrate their path integrator, many navigating insects rely on a combination of celestial compass cues and terrestrial landmark cues to follow a known route (Collett 1992; Graham and Cheng 2009; Kohler and Wehner 2005; Merkle and Wehner 2008; Narendra 2007a, b; von Frisch and Lindauer 1954). In bull ants, the importance of landmarks for navigation is so strong that if the view of familiar landmarks along the foraging path is blocked, some of the ants are unable to move past the obstruction (Reid et al. 2011). The shape of the distant skyline, rather than the position of local landmarks, can also provide important cues for foraging ants to navigate back to their nest (Fukushi 2001; Graham and Cheng 2009).
The importance of terrestrial landmarks for minimising errors in directional information is exemplified by the behaviour of male fiddler crabs (Uca lactea) that live in a featureless mudflat environment. During courtship, U. lactea create their own landmarks by building sandy structures near their burrow. The presence of a self-made landmark improves the homing accuracy of these crabs and allows them to travel further away from their burrow in search of a mate (Kim et al. 2010).
Unlike other animals, dung beetles appear not to rely heavily on landmark cues to orient. After locating a dung pat, a dung beetle quickly forms a ball and rolls it away. To avoid competition from other beetles congregating at the dung pile, the beetle must roll its dung ball away as swiftly and efficiently as possible, i.e. along a straight-line path (Dacke et al. 2003a, b, 2011; Byrne et al. 2003). To achieve this, dung beetles are known to rely primarily on the sun, the moon and the celestial polarisation pattern (Byrne et al. 2003; Dacke et al. 2003a, 2004). However, distant landmarks or landmarks that are held in line with the direction of movement could also aid dung beetles in maintaining a straight course. For example, a beetle would maintain a straight path by rolling directly away from a landmark.
If a beetle rolling away from the dung pile is robbed of its dung ball and then forced to make a new ball, it will depart from the dung pile in a new direction (Baird et al. 2010). This holds true even if the second ball is made at the same dung pat within minutes of the first one. During this time, the distribution of the surrounding landmarks remained constant, suggesting that dung beetles did not use these cues when selecting a roll bearing. Further evidence that dung beetles do not rely heavily on landmarks for orientation comes from experiments that place celestial compass cues and landmark cues in conflict. If the image of the sun is reflected by 180°, a ball-rolling beetle will also change its rolling direction by approximately 180° (Byrne et al. 2003). Similarly, if the pattern of polarised moonlight is artificially rotated by 90°, nocturnal ball-rolling dung beetles will change their path by approximately 90° (Dacke et al. 2003a, b). These findings indicate that, when presented with a conflict between terrestrial and celestial cues, terrestrial cues are ignored although the relative importance of different orientation cues for dung beetles has never been studied in detail. Here, we investigate the relative importance of landmarks and celestial cues for straight-line orientation in South African dung beetles.
Diurnal dung beetles Scarabaeus (Kheper) nigroaeneus were collected at the field site in South Africa using pit-fall traps. Beetles transported to Lund University in Sweden, were kept in large sand-filled boxes at an ambient temperature of 25 °C and a 12 h light (9:00–21:00)/12 h dark cycle. Field experiments in South Africa (experiments 1–3) were conducted within the game farm “Stonehenge”, 70 km north-west of Vryburg, North-West Province, South Africa (24.3° E, 26.4° S) in February 2009, 2010 and 2012. Field experiments in Sweden (experiment 4) were performed in the village Fjelie, 4 km west of Lund (13.1° E, 55.7° N) in March 2009.
Orientation relative to the dung pile
Orientation with and without surrounding landmarks
To test whether dung beetles use landmarks to minimise course errors after a disruption to their roll path, two circular arenas (1.6 m diameter) were created on level sandy ground 20 m apart. The view from the first arena ‘landmark arena’ was unobstructed, allowing a full view of the surrounding landscape, including high grass, bushes, trees and a water tower. The second arena ‘no-landmark arena’ was surrounded by a featureless beige wall (1 m height, 2 m diameter) that obstructed the view of all terrestrial landmarks. Two concentric circles were drawn on the ground in each arena, one of 40 cm radius, and the other of 80 cm radius (Fig 1b). Individual beetles were placed alongside their balls at the centre of one of the two arenas and allowed to roll along their chosen direction until they crossed both the 40 cm and 80 cm circles. 15 beetles started their first rolls in the landmark arena, and 15 beetles started their first rolls in the no-landmark arena, with the order alternating between beetles. A beetle’s rolling direction was defined as the bearing of the imaginary straight-line that connected the crossing points of each circle. After crossing the 80 cm circle, each beetle was picked up and carried over 20 m to the centre of the second arena (20 s walk away), where it started to roll again. The beetle’s rolling direction after this displacement was again measured as described above. As a control for disturbances due to the displacement between arenas, the same beetles were also carried half way to a second arena, but then returned to the centre of the original arena to roll for the second time.
Orientation performance without dorsal visual cues
To test whether landmarks can guide straight-line orientation in the absence of celestial cues, the dorsal visual field of ball-rolling beetles was occluded using small caps cut from thin black card. When taped to the dorsal thorax of a dung beetle with masking tape, these caps extended over the head and obscured the dorsal visual fields and a part of the lateral visual fields of the two dorsal eyes. In control experiments, beetles also wore identically shaped caps constructed from transparent plastic. Experiments were performed under three different conditions: beetles rolling under the clear sky with (1) no cap (n = 11) (2) a cardboard cap (n = 9) and (3) a transparent cap (n = 9). In each treatment, beetles were released, with their dung balls, from the centre of a circular, wall-free arena (diameter 3 m) of flattened and levelled sand. The beetles’ paths were recorded with a Sony HDR-HC5E Handycam fitted with a 0.42× wide-angle lens, suspended 3 m above the centre of the arena. Optical distortions were corrected by calibrating the camera (Bouguet 1999) with a black-and-white checkerboard (40 × 150 cm) that was placed at different locations on the arena and filmed from above. The roll paths were reconstructed from the video footage and their length determined over a radial distance of 120 cm extending from the perimeter of the 3 m diameter arena into the perimeter of a central circle with a diameter of 60 cm. This was done to exclude any initial re-orientations that the beetles may have made at the beginning of their roll path.
Orientation performance without celestial cues
In the previous experiment, it may have been possible that the caps, while blocking all celestial cues, also blocked some important terrestrial cues, such as the skyline. To account for this possibility, we recorded the paths of seven dung beetles rolling out from the centre of a circular, wall-free concrete arena (2 m diameter) under a fully overcast sky (devoid of celestial cues) and under a clear sky in Sweden. The experiments were performed on two successive days in March 2009 between 9:00 and 13:00 GMT, which corresponds to the normal activity period of the dung beetles at the field site. The ambient temperature during the experiments was 4 °C when the beetles were rolling under the overcast sky and 3 °C when they were rolling under the clear sky. During both days, the beetles had a full view of the surrounding landmarks. Beetle paths were analysed as in experiment 3, with the exception that their paths were evaluated from a centrally placed 40 cm circle to the perimeter; i.e. over a radial distance of 80 cm (due to the smaller arena).
Orientation relative to the dung pile
Dung beetles roll their balls along straight paths, facing backwards. The dung pile that marks the start of their roll path thus remains in the centre of their visual field as they move away from it, providing the beetles with a potential landmark cue. If the beetles use the dung pile as a landmark to aid them in maintaining a straight roll path, they should change their roll bearing by 45° counter-clockwise (CCW) when the dung pile is displaced 45° to the left. Conversely, they should change 45° degrees clockwise (CW) when the dung pile is displaced 45° to the right. This is not what we observed. The beetles changed their rolling direction by 0.70° ± 3.58° CCW (mean ± SE) after a leftward displacement of the dung pile and 1.00° ± 2.45° CW after a rightward displacement. These changes of bearings are not significantly different from each other (paired t test, t9 = 0.61, p = 0.55). This indicates that dung beetles do not use the dung pile as a dominant orientation cue as they roll away from it.
Orientation with and without surrounding landmarks
A ball-rolling dung beetle could use landmarks to maintain a straight line and/or to return to its original rolling direction after experiencing a disturbance to its roll path. To investigate these possibilities, 30 ball-rolling beetles were moved between two different arenas; one with a full view of the surroundings ‘landmark arena’ and one enclosed within a wall that blocked the view of surrounding landmarks ‘no-landmark arena’. The 15 beetles that started to roll their dung balls in the landmark arena changed their course by 16.3° ± 3.7° (mean ± SE) when displaced to the no-landmark arena. This change in direction is not significantly different from the course change the beetles exhibited when they were instead returned to the landmark arena after a partial displacement during the control experiment (22.7° ± 9.5°; paired t test, t14 = −0.66, p = 0.52). The 15 beetles that started to roll in the no-landmark arena changed their course by 19.0° ± 3.9° when transferred to the landmark arena. This is, again, not significantly different from the course change the beetles exhibited when returned to the no-landmark arena after a partial displacement during the control experiment (20.0 ± 4.3°; paired t test, t14 = −0.16, p = 0.87). These results indicate that, in the presence of reliable skylight information, the removal (or addition) of landmarks after a disturbance to the roll path does not affect the precision of the beetle’s re-orientation behaviour.
Orientation performance without dorsal visual cues
Orientation performance without celestial input
When orienting under a clear sky, dung beetles can potentially use the sun’s disk, the pattern of polarised skylight, the skyline or landmarks for their orientation. We find that the celestial compass dominates straight-line orientation in dung beetles so strongly that the visual removal of landmarks, such as bushes, trees and a high water tower, is of no importance for their orientation precision when the sun is visible (experiment 2). Even the dung pile, which is a centrally placed landmark that the beetles could use as they roll backwards away from it, is ignored when put in conflict with celestial cues (experiment 1). The importance of celestial cues for orientation is further demonstrated by the observation that beetles lose their ability to roll along straight paths when they are fitted with a cap that prevents them from seeing celestial compass cues (experiment 3). Similarly, beetles can no longer orient along straight lines under a heavily overcast sky (experiment 4) when the sun is no longer visible and the degree of skylight polarisation is greatly reduced (Pomozi et al. 2001; Hegedüs et al. 2007). Thus, even when celestial compass information is absent, the beetles do not use the skyline or other terrestrial landmarks to move along a straight path.
Interestingly, the paths of beetles rolling in overcast conditions appear less convoluted than those of beetles wearing caps. It is possible that a faint light intensity gradient (Ugolini et al. 2009), or a faint pattern of polarised light (Pomozi et al. 2001; Hegedüs et al. 2007) was still present in the overcast sky and that the beetles were able to use these cues to prevent rolling in loops. Another possibility is that beetles use some additional cue to keep a straight course while rolling (such as rotational optic flow cues in the lateral visual field) and that this system was obstructed by the caps. However, most beetles rolling under the overcast sky, or with their view of the sky obscured by a cap, still change their rolling direction by more than 180° over an effective distance of less than 80 cm. This strongly suggests that the beetles ignore landmarks, including the skyline, for their straight-line orientation.
Ego-centric systems of navigation are inherently susceptible to cumulative errors, while landmarks in most environments remain stable throughout the entire journey of an animal. To minimise the effect of these errors, it would thus be advantageous for the navigator to every now and then take a positional fix from landmark-based information (Wehner et al. 1996). Many animal navigators are also known to combine celestial compass information with landmark information to increase the precision of their navigational systems (see for example Collett 1992; Graham and Cheng 2009; Kohler and Wehner 2005; Merkle and Wehner 2008; Narendra 2007a, b; von Frisch and Lindauer 1954). Landmarks could also be used for maintaining a straight course if they are either very far away or held in line with the direction of movement. We find, however, that dung beetles do not use landmarks for orientation even when their celestial compass receives very weak information (experiment 4), or none at all (experiment 3). To our knowledge, this is the first demonstration of an animal with a celestial compass system that ignores the extra precision that landmarks can offer.
Why is it then that dung beetles are unique in ignoring landmarks? One possible reason is that ball-rolling dung beetles, unlike most insect navigators, do not need to locate a stationary nest at the end of their foraging journey. This makes the nature of their orientation task fundamentally different from that of homing insects. Despite the directional errors it accumulates while travelling, an animal using only celestial compass cues can theoretically travel arbitrarily far away from its starting point (Cheung et al. 2007). The animal will thus travel further away from its starting point—a dung pile, for example—with every step. In the case of the straight-line orientation of dung beetles, celestial compass cues appear to provide all the precision these insects need to ensure a safe exit from the dung pile.
The authors would like to thank Ted and Winnie Harvey for their assistance at the field site. We also thank Martin Kohler for help with the illustrations. The experiments were carried out according to the current laws on animal experimentation in Sweden.