To answer questions related to the usefulness of premarital testing for human immunodeficiency virus type 1 (HIV-1), two “blinded” or “nonlinked” HIV-1 serosurveys were done in New Jersey, a state with a high incidence of AIDS, on blood specimens submitted for a premarital serologic test for syphilis. The first survey involved premarital blood specimens submitted to the New Jersey Department of Health laboratory for the year starting September 1987. The second survey involved premarital specimens submitted to five private or hospital clinical laboratories in the spring of 1989, of which approximately 1,000 consecutive premarital specimens from each laboratory were sent to the Department of Health laboratory for HIV-1 testing. Of 4,247 specimens tested in the 1987–1988 survey, 21 (0.49%) were positive for antibodies to HIV-1, while among 4,696 specimens in the 1989 survey, 29 (0.62%) were positive. When the survey results were weighted by the number of marriages by geographic regions of the state, the weighted premarital HIV-1 seroprevalence was 0.55% for the 1987–1988 survey and 0.62% for the 1989 survey. The male/female ratio of positive tests was 2.7:1 in 1987–1988 and 1.6:1 in 1989. Of the 8,943 specimens in both surveys, 5 (0.06%) gave an indeterminate immunoblot result, compared with 50 positive results. These percentages of premarital HIV-1 infections are much higher than earlier estimates and reports and are of the same magnitude as recently reported blinded premarital HIV-1 testing elsewhere. Results of this magnitude support a recommendation in New Jersey of voluntary HIV-1 counseling and testing for marriage applicants.