Group size (GS), density and distribution of resources in space and time are manipulated over a wide range in populations of domestic animals kept for agricultural production. Different numbers of animals may be kept in enclosures of the same or different sizes, thereby affecting density. Food resources may be limited and clumped in time or space. Alternatively, food resources may by unlimited and evenly distributed in space to facilitate feed intake and growth. It is important from economic and animal welfare perspectives to understand the potential effects of manipulating GS, density and distribution of resources on aggression in captive populations of domestic animals. Three experiments were conducted to investigate the effect of GS on the frequencies of different types of aggressive acts performed by 11-21 wk old White Leghorn pullets. In these experiments, aggression was measured in the context of competition for a limited supply of highly attractive food provided in two patches (central and peripheral) in addition to the regular (ad libitum) food supply. The birds were observed in their original groups of 15, 30, 60, and 120 birds (3-4 replicate groups per GS), reared together since hatch, or in temporary sub-groups of the original groups. In the first experiment (space per bird constant, pen size and GS history carrying), the number of aggressive acts delivered per food patch declined as GS increased from 15 to 120 birds. In the second experiment, utilizing birds from the groups of 120 only (space per bird varying, pen size and GS history constant), the number of aggressive pecks delivered per patch was increased as temporary GS increased from 15 to 60 birds. In the third experiment (space per bird and pen size constant, GS history varying), the number of aggressive pecks delivered per patch was constant among groups of 15 birds obtained form the groups of 15, 30, 60, and 120 birds. In all experiments, the central food patch was exploited sooner than the peripheral one. The actual number of birds at a food patch at a given time was more important in determining the frequency of aggression at the food patch than the number of birds present in an enclosure or previous GS experience. When a high number of individuals was present at a food patch, aggression at that patch was relatively was relatively infrequent. The results support the hypothesis that aggression in the domestic fowl is a dynamic process, with decisions about aggressive behaviour being made facultatively according to the relative costs and benefits of different behavioural strategies at a given time and place rather than being fixed obligatorily at a constant level according to the number of birds present in an enclosure and their ability of form dominance relationships.