Two populations of rabbits (Oryctolaqus cuniculus) were investigated to see whether polygynous, multi-male groups formed in the absence of large multi-entranced warrens. They did not. Rabbits neither gathered in space nor time. The small warrens were spread out evenly across homogeneous patches and the females were well spaced out. Monogamy, distinguished by a battery of tests, was prevalent, with the more dominant males as 'mate' rabbits. That the polygyny frequently mentioned in the literature was a result of male dominance and female defense was considered. The genetic structure of each population was investigated by taking blood from rabbits and having it analysed electrophoretically and for immunoglobulins. A method for assessing relatedness between groups of pairs of animals was implemented, then validated and developed with Monte Carlo simulations. With the seven polymorphic allele obtained, no non-zero relatedness was found but it was sometimes possible to exclude high relatedness. The bearing of sociality on vigilance during feeding was investigated. Although a rabbit's vigilance decreased as its 'mate' approached, the presence of other rabbits was correlated with increased vigilance. It was concluded that the need for social vigilance outweighed the benefit of 'many eyes' watching for predators. This conclusion was tested by experiment, using stuffed animals as stimuli. Rabbits increased their vigilance during grazing bouts both by increasing the length and frequency of scans. Scans could be short or long: the probability of ending a scan decreased sharply at a certain point; a form of positive feedback. The durations of short 'maintenance' scans were dependent on chewlength (the amount of food in the mouth). This fitted a timesharing definition as supported by experiment. Long scans in response to a visible threat did not involve chewing.