This report describes the methods used in conducting a long-term follow-up study and the factors associated with the successful location of subjects. In 1987-91, we selected, from among the women who were born in 1959-66 as members of the Philadelphia and Providence cohorts of the Collaborative Perinatal Project (CPP), 391 who were preterm at birth, 413 who were small-for-gestational age (SGA) and 978 who were term and appropriately grown; 1314 were from Philadelphia and 468 from Providence. Tracing was accomplished using original CPP records, telephone directories, voter registration and driver's licence records, birth certificate files and credit bureaux. A total of 61% of the women were located, 55% at Philadelphia and 77% at Providence. After adjustment, the odds ratio for ease of location, compared with women who were term and appropriately grown, was 0.85 [95% CI 0.66-1.11] for women who were preterm and 1.02 [0.78-1.34] for women who were SGA. In Philadelphia, African-American women were more likely than white women to be located, but in Providence, African-American women were less likely than white women to be located. Women whose own mothers were older at the time of their birth were more likely to be located, as were those who had more siblings followed in the CPP. Compared with women who did not complete their original CPP examinations at ages 4 and 7 years, those whose original CPP follow-up was more complete were more likely to be located. These findings demonstrate that long-term follow up in a US urban setting can be accomplished, and that successful location was not associated with birth status.