Search

In large animal groups, collective decision-making is often self-organised, i.e. decisions emerge from local interactions between individuals following simple behavioural rules. It is often assumed that no great cognitive complexity or diversity is required in individuals to produce complex, finely tuned collective choices. In this thesis, I investigate whether collective decisions can benefit from individual memories and previous experience in house-hunting by the rock ant Temnothorax albipennis. Rock ants emigrate readily into a new home if their current nest is damaged. During emigrations, colonies can select the best available site using distributed decision-making mechanisms. Here, I show that rock ants continually gather information about available nest sites, even when their home nest is still intact. This leads to a generally improved collective performance in later emigrations. Prior familiarisation with high-quality nest sites indeed allows colonies to emigrate faster, be more cohesive and/or choose more accurately than naïve colonies, thus leading to a better compromise between speed and accuracy. Additionally, rock ants appear to adjust their preference criteria according to the respective qualities of their home nest and of available nest sites. This confers colonies with high flexibility in their choices and allows them to tune collective decisions according to external conditions. I also provide evidence that workers memorise the position and suitability of high-quality, available nest sites. They then retrieve that memory, to play a key role in later emigrations. Additionally, nest marking chemicals and social interactions within previously visited, candidate nest sites ensure effective transfer of information to naïve workers during emigrations. These results indicate that ants have high individual cognitive abilities, as they can memorise information and retrieve it later, and that certain individuals are disproportionately influential in subsequent collective choices. This suggests that distributed decision-making may greatly benefit from both individual cognitive complexity and inter-individual variability.