01whole.pdf (15.5 MB)
Learning, memory and cue choice in navigating ants
thesis
posted on 2022-03-28, 15:37 authored by Cody A. FreasA critical reoccurring challenge facing mobile animals is the need to reliably find goal locations (e.g. food or nest). To solve this problem, animal navigators acquire and use multiple cue sets within their environment, which designate direction and distance estimates of these locations. Navigational cues can conflict, which adds a further challenge, requiring the navigator to integrate these cues before moving. Foraging ants are expert visual navigators known to use both learned panorama cues and path integration to navigate. This thesis examines navigational memory, learning and cue choice in two Australian ant species, the diurnal ant Melophorus bagoti and the nocturnal bull ant Myrmecia midas. M. bagoti is a solitary foraging desert ant endemic to Central Australia. Foragers are shown to retain robust memories of both the nest skyline and multiple non-nest site skylines. Forager orientation performance shows evidence of retroactive interference after changes are made to the skyline at the same site. When presented with cue conflicts between the terrestrial and celestial cue sets, foragers appear to choose dynamically based on experience and cue reliability in accordance with the Temporal Weighting Rule. Foragers with directional conflicts in their inbound and outbound routes show evidence of rapid vector calibration to the inbound route.This calibration appears to have a directional limit of around 45˚. When foragers are restricted to the nest site, they are unable to extrapolate panorama cues from the nest to local sites but only require one experience of the homeward route to learn the correct nest direction. Furthermore, exposure to the outbound foraging path appears critical for efficient homeward route formation in this species. M. midas is a previously unstudied solitary foraging bull ant that forages nocturnally. Foragers show a strong reliance on terrestrial panorama cues in cue choice tests and only weak use of accumulated vectors. Foragers that are subjected to polarised light manipulations during route navigation dynamically weight the terrestrial and celestial cues during the foraging trip. M. midas foragers also navigate using the panorama while on the tree face, exhibiting similar solutions to navigational challenges on the ground and when navigating in three dimensions. Both species show heavy use of panorama learning and memory along with flexible use of navigation strategies.