From 1988 to 1991, annual questionnaires and serosurveys were performed in a cohort of outdoor workers in New Jersey at high risk for Lyme disease to 1) evaluate temporal trends in seroprevalence and seroconversion of antibody to Borrelia burgdorferi; 2) identify risk factors for B. burgdorferi seroconversion during these years; and 3) examine associations between such seroconversion in 1989-1990 and anti-tick saliva antibody (ATSA, a biologic marker of tick exposure) seropositivity in 1990. A total of 1,519 workers participated in at least 1 year of the study. Lyme disease seroprevalence and seroconversion increased from 1988 to 1990 and then decreased in 1991. Years at residence, rural residence, and a history of medical problems were observed to be risk factors for seroconversion from 1988 to 1991. An interaction between pet ownership and rural residence was observed in that rural residents were only at an elevated risk if they owned pets. B. burgdorferi seroconversion from 1989 to 1990 was associated with ATSA seropositivity in 1990; in subjects reporting low tick exposure, the odds ratio was 8.2 (95% confidence interval 1.5-44.7). Associations between ATSA and B. burgdorferi serologic status suggested that educational programs may have contributed to the decline in Lyme seroprevalence and seroconversion in 1991.