The hallmark of social insect colonies is reproductive division of labour which is often associated with dramatic morphological and behavioural differences between queens, workers and males. The aim of this proposal is three-fold. First, we will use our recently developed fiducial identification system to investigate the general principles of social organisation and division of labour. The video tracking of workers labelled with markers derived from the augmented reality library ARTag allows us for the first time to distinguish up to 2000 individuals and precisely locate them every 500ms, hence allowing large-scale experiments addressing the question of how the behaviour of individual workers is influenced by the joint effects of environmental factors and social interactions. The second related aim is to investigate how the level of altruism within colonies and the reliability of communication systems are shaped by colony kin structure. Because it is not possible to conduct artificial evolution with social insects we will use a new experimental system consisting of colonies of small mobile robots with simple vision and communication abilities. This system permits to conduct hundreds of generations of experimental evolution in colonies with variable group composition to identify the factors affecting the evolution of altruism and communication. Finally, we will complement these studies with a genetic perspective using a remarkable genetic social polymorphism that we recently discovered in the fire ant Solenopsis invicta. The advent of new ultra high-throughput sequencing techniques will allow us to document the steps involved in the evolution of this genetic social polymorphism and test the suggestion that the chromosome involved in the social polymorphism has the properties of a sex chromosome. This project will be highly interdisciplinary, involving skills in evolutionary biology, the study of animal behaviour, bioinformatics, engineering and molecular biology