The production of queen cups, queen cells, worker brood and drone brood in a total of 81 colonies was recorded regularly throughout four summers. All these processes are shown to occur in annual cycles whose peaks were approximately synchronized. All colonies produced queen cups at some time during the summer, and the great majority of colonies possessed them by the end of June. The number of queen cups varied considerably in different colonies but tended to fluctuate with the amount of brood during the season. About half the total number of colonies produced occupied queen cells, but queen rearing was abortive in more than half of these colonies. Many queen cells and some virgin queens were destroyed by the bees. Of 25 colonies with sealed queen cells, 19 replaced their queens, 16 superseding and 3 swarming, and there was some evidence that older queens were more likely to be replaced than younger ones. No queen rearing took place in colonies with queens of the current year. No significant relationship was found between queen rearing and amounts of brood, although colonies with larger amounts of brood in May showed indications of an increased tendency to produce queen cells subsequently. Most supersedure was found to occur during the swarming season.