Three field trials were conducted in a large commercial feedlot in Saskatchewan to determine the prevalence of Haemophilus somnus in calves and to evaluate the efficacy of prophylactic mass medication with long-acting oxytetracycline on day 17 (1990, n = 1336), day 11 (1991, n = 4372), or day 8 (1992, n = 5632) postarrival. Hemophilosis accounted for > 40% of the mortality in feedlot calves each year. Haemophilus somnus was cultured from the blood of one febrile calf on day 1 (0.1%, n = 895), but it was not cultured from nasal swabs on day 1 or day 11 (n = 881) or from blood samples on day 11 (n = 883). Similarly, it was not cultured from nasal swabs or blood samples from sick calves first treated for bovine respiratory disease (BRD) (n = 219). Serological titers to H. somnus increased (p < 0.05) in unvaccinated calves from day 1 (Geometric mean titer = 11,846) to day 96 (Geometric mean titer = 63,712), indicating natural infection following feedlot entry. Calves that relapsed twice with BRD or died from BRD +/- hemophilosis had significantly (p < 0.06) lower titers to H. somnus on days 1 and 96 than those that did not relapse twice or die. Postarrival mass medication with long-acting oxytetracycline did not reduce (p > 0.05) the risk of hemophilosis mortality. However, it reduced (p < 0.05) the risk of BRD treatment by 14% and the risk of BRD mortality by 71%. Additional epidemiological studies of H. somnus are needed so that we can develop strategic medication and vaccination programs to reduce losses from hemophilosis.