Intricacies of running a route without success in night-active bull ants (Myrmecia midas).

Abstract

How do ants resolve conflicts between different sets of navigational cues during navigation? When two cue sets point to diametrically opposite directions, theories predict that animals should pick one set of cues or the other. Here we tested how nocturnal bull ants Myrmecia midas adjust their paths along established routes if route following does not lead to their entry into their nest. During testing, foragers were repeatedly placed back along their homeward route up to nine times, a procedure called rewinding. This procedure produced an accumulating path integrator, or vector, in diametric opposition to the learned landmark views of the route. Repeated rewinding made some individuals head initially in the nest-to-feeder vector direction, but all ants ended up using the visual scene for homing, demonstrating the importance of view-based homing in this species. Repeated rewinding, however, led to path deteriorations; with increased path meander and scanning, results also found in desert ants. After nine rewinding trips, ants were displaced off their route in further manipulations, to a site near the nest, an unfamiliar site, or with the terrestrial surround entirely covered. The results show that a change in visual conditions diminished the weight accorded to path integration: the off-route ants no longer headed off in the vector direction as they did on the immediately preceding trial. They relied on celestial compass cues in other ways for homing. Experiment 2 showed the effects of rewinding in the unaltered natural habitat were not view-specific in these bull ants.

Effect of repetition of vertical and horizontal routes on navigation performance in Australian bull ants

“…Increases in meandering, U-turns, and scanning behaviours are all common responses to rewinding (Deeti et al, 2023 ; Wystrach et al, 2019 ), and while we found an effect of being rewound on meandering when walking on a vertical surface, its effect on U-turns was a non-significant trend, and there was no change in scanning across any treatment. The meandering of these foragers reached an asymptote by the second trip, as further captures and rewinding did not lead to higher meandering.…”

“…The meandering of these foragers reached an asymptote by the second trip, as further captures and rewinding did not lead to higher meandering. Such an asymptote contrasts with M. bagoti foragers which showed increasing meandering, and U-turns over four repeated rewinding procedures (Wystrach et al, 2019 ) and in contrast to increasing meandering, U-turns and scans in the previous M. midas ’s horizontal rewinding experiment (Deeti et al, 2023 ). We also found weak evidence that U-turns increased due to vertical rewinding and full trip rewinding relative to captures on this horizontal portion of the trip.…”

“…While it may be that rewinding four times for a total length of 16 meters is insufficient for this species, it could also be that this species does not attend to their path integration vector when visual cues are available. Other studies have shown longer path integration vector lengths lead to stronger weighting and greater behavioural changes (Deeti et al, 2023 ; Freas et al, 2017a , b ; Wystrach et al, 2019 ). And so, while we would expect Rewind Trip foragers to therefore have the greatest behavioural response, paradoxically we instead found that foragers rewound only on the Vertical section (total of 6 meters) meandered more than those rewound on both Vertical and Horizontal sections.…”

“…We did not detect any effect of Rewind Terrain, or Rewind Trip on any measures of navigational uncertainty. Other studies rewound M. bagoti foragers four times for a total of 40 meters and detected an increase in meandering and U-turn behaviours (Wystrach et al, 2019 ), Deeti et al ( 2023 ) rewound M. midas foragers and first detected increases in meandering, U-turns, and scanning behaviours after six horizontal rewinding trips, a total distance of 42 meters. While it may be that rewinding four times for a total length of 16 meters is insufficient for this species, it could also be that this species does not attend to their path integration vector when visual cues are available.…”

“…Additionally, M. midas foragers use terrestrial cues to avoid barriers by taking alternative paths during their vertical ascent on foraging trips (Islam et al, 2023 ). Rewound over the horizontal foraging corridor, M. midas foragers relied on terrestrial cues even after nine rewinding procedures, although they exhibited an increase in meandering and scanning behaviours, and occasional brief U-turns shorter than 2% of the vector length (Deeti et al, 2023 ). Nothing is known, however, about how they integrate path integration and terrestrial cues during vertical navigation.…”