Learning in the navigation of night-active bull ants

Chapter 2 examines navigation behaviour in familiar and unfamiliar environments. While Australian bull ants are known to rely heavily on familiar panoramic views to navigate, their behaviour in new unexperienced locations with unfamiliar panoramic views and with conflicting navigational cues is not well understood. Previous studies usually rely on a single technique to record their ants’ behaviour, but different studies use different techniques, and the comparability of these techniques is poorly understood. I investigated multiple common methods of recording navigational behaviour, in the hope of measuring their strengths and weaknesses. I compared: 1) videotaping of ants walking on a trackball, 2) videotaping of walking on the natural terrain, 3) tracking using a differential GPS device, and 4) recording initial trajectories on a goniometer. I found that ants headed towards the nest when displaced in a location with a familiar view, independently of whether the location was previously visited or not. An unfamiliar view, on the other hand, led the foragers to not orient towards the nest. In addition, I found that foragers exhibited a broader heading direction when captured and restrained overnight. This finding raises a question about the effects of animal handling involved in insects. The results also suggested that the recorded foragers’ behaviour can vary depending on the recording methods used. GPS and path-video recordings yielded the most consistent data, but the high level of labour-intensive work required makes data repeatability quite time-consuming. On the other hand, while the goniometer recoding technique provides easier data repeatability, it produces lower levels of data consistency. Lastly, the trackball technique emerged as the best cost-effective solution, with good data consistency and repeatability.  

Chapter 3 examines re-learning responses induced by physical aversion. Chapter 2 raised a question about the effect of animal handling on foragers’ behaviour. This study investigated ants’ ability to store and recall views associated with positive and aversive events, and how these aversive events affect navigation in Australian bull ants. I used a capture-and-release procedure as an aversive event and its absence as a positive event for the foragers. The results showed that after aversive encounters, foragers learn to circumvent the location with the views associated with an aversive event and increase meandering and scanning along their route.  

Chapter 4 examines the foragers’ responses to several capture-and-release procedures, coupled with spatial displacement back to the route or without any displacement. To accomplish a spatial displacement, I collected homing foragers in the proximity of the nest and then displaced them back on the route, a procedure known as rewinding; in capture-and-release without displacement, I released the captured foragers at the same spot. Through four repeated rewinding procedures, homing foragers accumulated an opposite nest-to-feeder vector, which allows the foragers to estimate the current position relative to the starting point; all repeated capture-and-release procedures lead the foragers to form aversive memories. Additionally, to test their vertical navigation behaviour, foragers were made to re-experience the foraging tree, a vertical surface, or the foraging corridor, a horizontal surface, or both. As a result of all conditions of capture, regardless of whether spatial displacement was involved or not, the foragers exhibited a similar increase in meandering and U-turns. Furthermore, in all conditions, foragers exhibited higher meandering and scanning behaviour in their vertical navigation, suggesting a bigger challenge for them on the tree than on the ground. The results of this study suggest that in this species, all the effects of rewinding the ants, up to 4 times, stem from the aversiveness of capture rather than the increasing vector in the nest-to-feeder direction.  

This research advances our understanding of ants’ sensitivity to physical manipulations. It establishes that physical handling induces an aversive response in Australian bull ants. In contrast, similar physical handling does not induce an aversive response in desert ants. The different sensitivity to physical manipulation in Australian desert ants raises questions about the origin and mechanism behind these behavioural differences. Since Australian bull ants live over 3 years against a few weeks in desert ants, I speculate that M. midas has evolved a higher sensitivity to physical manipulation as a means of safeguarding the colony’s investment in the foraging force, which is more valuable owing to their longer average lifespan. This thesis lays the groundwork for future research on aversive learning in solitarily foraging navigators.  

This work contributed to expand our understanding of two understudied areas: insect’s navigation at night and their capacity to learn and respond to negative experiences. Myrmecia foragers demonstrated remarkable ability to navigate home even in morning twilight. In dim light conditions, foragers exhibited the ability to generalise learned views enabling them to return to their nest from novel locations, even when having a conflicting path integration vector. This work also revealed that nocturnal Myrmecia bull ants possess the capability to learn and recall views associated with negative experiences, such as being captured, leading them to avoid these locations. In addition, I found that different recording techniques might provide varied data consistency and accuracy. This highlights the significance of considering the ants' experience and sensitivity when designing experiments, particularly in terms of recording techniques and animal handling.