Etude expérimentale et modélisation des déplacements collectifs chez le mouton Mérinos, Ovis aries

This thesis presents a comprehensive set of results, obtained through an innovative experimental methodology, that have important and extensive implications for the fields of integrative biology and complex systems. The main objective of the thesis is to study the inter-individual interactions involved during the initiation and coordination of movement in gregarious vertebrates, and in particular in the sheep Merinos d’Arles (Ovis aries). Key questions are, when an individual initiates a movement, what information is taken into account by conspecifics, how is this information spread across the group, and what mechanisms underlie the collective decision processes? To answer these questions, we created an experimental paradigm to trigger, in a standardized way, the movement of trained individuals that were then placed in a group of naïve conspecifics. Using two types of stimuli, a sound (public) and a vibration (private), we could evaluate the individual response of followers, and the effect of the behavioural state on this response. An additional set of experiments also provided recordings of spontaneous initiations of movement.

Our results suggest that every individual in a group can initiate a collective movement. Our quantitative analysis then showed that, in Mérinos sheep, the individual decision to follow depends on a double mimetic effect; individuals take into account both the number of already departed individuals and the number of individuals which have not yet departed. A comparison between three experimental situations reveals that the decision rule is unique and that the behavioural state of potential followers only slightly affects the collective dynamics.

Our approach, a combination of experimentation and modelling, provides original results that contribute to the understanding of individual and collective decision-making processes, and of the mechanisms involved during collective movement. The experimental paradigm that was proposed here, and the mathematical tools that were used, open interesting perspectives for new experimental studies and for the generalization of the behavioural rules exposed in this thesis.