Background: Silverthorn and Frick argued that there is no pathway to criminal activity among girls that is analogous to the early-onset/life-course- persistent-type one among boys. Instead, they argued for a female-specific theory in which even girls with the same high-risk backgrounds and criminal outcomes as early-onsetllife-course-persistent boys show delayed-onset offending. Aims: To test the Silverthorn-Frick model of gender-specific pathways to offending in an independent cohort. Methods: A cohort of 987 urban, African-American males and females participated in the Philadelphia portion of the National Collaborative Perinatal Project which collected biological, psychological and sociological variables. By age 18, 220 of these individuals had incurred a least one police contact, of which 64 did so under the age of 13. A criminal history follow-up was conducted in 1998 when cohort members had all reached their mid-thirties. Results: Females and males were equally likely to experience early-onset offending (police contact by age 13), but female late-onset offenders resembled male early-onset offenders on nine of 10 risk factors. Male early-onset offenders had worse criminal outcomes compared with female and male late-onset offenders, but did not differ from female early-onset offenders. Female late-onset offenders showed many similarities on key risk factors to male late-onset offenders. Conclusions: Silverthorn and Frick may have over-estimated the outcome similarity of late-onset offending females to early-onset males and under-estimated occurrence of early-onset-life-course persistent offending in females. In so far as cohort studies such as these inform social policy, it is likely to be important that they are interpreted with sensitivity to minority needs.