We studied the risk of herpes simplex virus (HSV) infections in neonates exposed to HSV at the time of vaginal delivery to mothers with a history of recurrent genital HSV infections. None of 34 infants exposed to HSV type 2 acquired an HSV infection. On the basis of this sample, the 95 percent confidence limit for the theoretical maximum infection rate is 8 percent. Cord blood or blood obtained during the first two weeks of life was available from 33 of the 34 exposed, uninfected neonates. All 33 of the samples possessed demonstrable neutralizing antibody to HSV type 2, and 79 percent had titers above 1:20. These results were compared with those in a previously studied group of neonates with HSV infections; the latter infants were significantly less likely at the onset of symptoms to have demonstrable neutralizing antibody to HSV type 2 (P = 0.000148) or to have titers above 1:20 (P less than 0.00001). We conclude that given the low attack rate, empirical antiviral therapy is not warranted in all infants of mothers with recurrent genital HSV infection who are exposed to the virus in the birth canal. Our findings suggest that the presence and titer of neutralizing antibody to HSV contribute to the low attack rate.