Serum samples were obtained from 602 calves (from 19 groups in four feedlots: three in Ontario, and one in Alberta) upon arrival at the feedlot and 28 d later. Of these calves, 202 developed bovine respiratory disease (BRD) and 400 did not develop BRD. Based on high antibody titers noted upon arrival, we infer that most calves were exposed to Haemophilus somnus prior to arrival at the feedlot. Within a group, calves with high titers on arrival had a reduced risk of developing BRD later. Most calves did not experience titer increases after arrival; however, calves that had stable or increasing titers had a relatively low risk of contracting BRD. The calves at greatest risk of BRD were those with titers on arrival of less than 6.8 units and subsequent titer decreases of more than 1 unit. The effects of both the titer on arrival and the titer change after arrival were stable when the serologic effects of a number of viruses and Mycoplasma agents were considered. Neither antibody titer on arrival nor titer change was related to weight gain differences among calves. Calves with BRD or calves with lower weight on arrival had decreased weight gains in the first 28-day feeding period. The high titers on arrival may have protected most calves against further infection with H. somnus. However, since the calves that developed BRD had large titer increases to a number of viruses and to Pasteurella haemolytica, while having decreased antibody titers to H. somnus, we infer that the existing antibodies were "used up" in combatting the agents, including H. somnus, which may have "caused" the BRD. Calves which were able to increase their antibody levels to H. somnus tended to have a reduced risk of BRD.