Vectorial capacity of Culicoides sonorensis Wirth & Jones for the transmission of bluetongue (BLU) virus was examined at a southern California dairy from January 1995 to December 1997. Insects were collected one to two times per week in five CDC-type suction traps (without light) baited with CO2 at a constant release rate of 1,000 ml/min. BLU virus was detected in midges collected from May through December with an estimated overall infection rate of 0.08%. The BLU virus infection rate of field-captured midges was not correlated with sentinel calf seroconversions to BLU virus. Sentinel calf seroconversions were highly seasonal, occurring from August through November with most calves seroconverting during September and October. Vector competence of field-collected nulliparous flies fed a locally acquired serotype of BLU virus in the laboratory was stable among years (17-23%). Vectorial capacity was strongly correlated with BLU virus transmission (measured by sentinel calf seroconversions) during 1995 and 1996, but not during 1997. Host biting rate estimated for traps nearest to the sentinel calves was the index best correlated with BLU virus transmission for all study years and was most highly correlated with sentinel seroconversions 4 wk later. The utility of vectorial capacity and its component variables is discussed for this system.