In the summer of 1996, we screened 18,931 calves in 128 beef herds located in five US states for persistent bovine viral diarrhea virus (BVDV) infection. Of these, 76 herds were randomly selected from the client database of collaborating veterinary practices, and 52 herds were suspected by the collaborating veterinarians to have BVDV infection based on history or clinical signs. Serum was obtained from each calf in the cooperating herds prior to 4 months of age and tested for the presence of BVDV by microtiter virus isolation. Information about each of the herds (including management practices, vaccination history, and breeding- and calving-season production measures) were collected by the collaborating veterinarians using standardized questionnaires. A total of 56 BVDV-positive calves in 13 herds were identified on initial screening. Ten (19%) of the BVDV-suspect herds and three (4%) of the randomly selected herds had > or = 1 BVDV-positive calf at initial screening. Multiple BVDV-positive calves were identified in 10 of those 13 herds. Follow-up information was obtained for 54 of the 56 positive calves. Ten out of 54 (18%) died prior to weaning, and 1 (2%) was sold because of unusually poor growth. Thirty-three out of 54 (61%) of the initially positive calves remained BVDV positive at 6 months of age - confirming persistent-infection (PI) status. Dams of 45 of the 56 positive calves were tested, with 3 (7%) identified as positive - indicating most PI calves were products of acute dam infection during gestation. The proportion of cows that were pregnant at the fall 1995 pregnancy examination was 5% lower in herds with PI calves born during the 1996 calving season than in randomly selected herds without PI calves. Most of the calves we identified with persistent BVDV infections survived to weaning, and could provide a constant source of virus to the herd throughout the breeding season and early gestation.