Rotaviruses are accepted as enteric pathogens of calves but many natural infections are subclinical. In the present paper, the outcome of inoculation of gnotobiotic calves of three ages (the second day of life, the second week of life and calves aged 6 weeks and over) with doses of 10(5.0) to 10(6.5) TCID50 was compared for three bovine rotavirus isolates (C3-160, 17/4 and 39/58). The clinical outcome of infection was dependent on both calf age and rotavirus isolate. Age-dependent resistance to infection was not found. By contrast, age-dependent resistance to disease was found with rotavirus isolates C3-160 and 17/4 but not with 39/58. All three isolates caused disease in calves inoculated on the second day of life but only one, 39/58, caused disease in the two older groups. Peak levels and duration of virus excretion were similar in clinically normal (10(6.7 +/- 1.08) TCID50 per g of faeces for 4.6 +/- 1.2 days) and diseased (10(7.45 +/- 0.94) TCID50 per g of faeces for 5.3 +/- 0.98 days) calves of all ages, but the onset of virus excretion occurred sooner in clinically affected calves (1.6 +/- 0.63 days in clinically affected compared with 3.7 +/- 1.5 days in clinically normal calves, P < 0.01). The present study confirmed the findings of an earlier study (Bridger & Pocock, 1986) which showed that bovine rotaviruses differ in virulence for calves in the second week of life and that older calves are susceptible to rotavirus infection and disease. In addition the present study demonstrated for the first time, that differences in rotavirus virulence are not apparent with calves inoculated on the second day of life, an age which has been used commonly to assess rotavirus virulence. It is suggested that rotaviruses that cause disease in calves only on the second day of life should be described as of low virulence whereas those that cause disease in all ages should be described as virulent.