How organisms allocate limited resources to reproduction is critical to their fitness. The size and number of offspring produced have been the focus of many studies. Offspring size affects survival and growth and determines offspring number in the many species where there is a trade-off between size and number. Many social insects reproduce by colony fission, whereby young queens and accompanying workers split off from a colony to form new colonies. The size of a new colony (number of workers) is set at the time of the split, and this may allow fine tuning size to local conditions. Despite the prevalence of colony fission and the ecological importance of social insects, little is known of colony fission except in honey bees. We studied colony fission in the ant Cataglyphis cursor. For clarity, "colony" and "nest" refer to colonies before and after colony fission, respectively (i.e., each colony fissions into several nests). The reproductive effort of colonies was highly variable: Colonies that fissioned varied markedly in size, and many colonies that did not fission were as large as some of the fissioning colonies. The mother queen was replaced in half of the fissioning colonies, which produced 4.0 +/- 1.3 (mean +/- SD) nests of markedly varied size. Larger fissioning colonies produced larger nests but did not produce more nests, and resource allocation among nests was highly biased. When a colony produced several nests and the mother queen was not replaced, the nest containing the mother queen was larger than nests with a young queen. These results show that the pattern of resource allocation differs between C. cursor and honey bees. They also suggest that C. cursor may follow a bet-hedging strategy with regard to both the colony size at which fission occurs and the partitioning of resources among nests. In addition, colony fission may be influenced by the age and/or condition of the mother queen, and the fact that workers allocating resources among nests have incomplete knowledge of the size and number of nests produced. These results show that the process of colony fission is more diverse than currently acknowledged and that studies of additional species are needed.