This project investigated the use of ultrasonography at first diagnosis of presumptive early bovine respiratory disease (BRD) in feedlot cattle from western Canada. One hundred seventy-four cattle (116 cases and 58 controls) at high risk of developing BRD were enrolled in a prospective longitudinal study over 2 y (2006-2007). Cattle with clinical signs relating to the respiratory system and assessed as sick at the time of feedlot arrival (arrival fever cases) or assessed as sick in the pen 3 to 30 d post-arrival (post-arrival fever cases, post-arrival no fevers cases) were eligible for enrollment. Control animals were identified at the time of case enrollments. Ultrasonography was done using a 3.5 sector transducer at enrollment and at 2, 4, and 6 wk post-enrollment. Lung lesions were identified at least 1 time over the course of the trial in 32/116 (28%) cases and 9/58 (16%) controls. At enrollment, lung lesions were identified in 20/115 (17%) cases and 2/55 (4%) controls (data unreadable n = 4). Post-arrival fever cases (14/48) were the most likely to have a lesion identified using ultrasound. In arrival fever cases, average daily gain (enrollment to last ultrasound, average 34 d) was improved (P = 0.007) in cattle identified with a lesion at enrollment using ultrasound compared with those not identified with a lesion at that time, potentially demonstrating the effects of gut fill at arrival weighing, as these sicker animals may have eaten less prior to arrival and, therefore, had more room for improvement in weight over time due to restoration of normal gut fill. None of the ultrasound time points explored (enrollment, 2, 4, or 6 wk post-enrollment) were associated with the animal health outcomes of interest (subsequent treatment, chronicity, wastage, or mortality) for cattle enrolled at arrival or post-arrival.