Purpose: To describe official adult-onset offenders, investigate their antisocial histories and test hypotheses about their origins. Methods: We defined adult-onset offenders among 931 Dunedin Study members followed to age 38, using criminal-court conviction records. Results: Official adult-onset offenders were 14% of men, and 32% of convicted men, but accounted for only 15% of convictions. As anticipated by developmental theories emphasizing early-life influences on crime, adult-onset offenders' histories of antisocial behavior spanned back to childhood. Relative to juvenile-offenders, during adolescence they had fewer delinquent peers and were more socially inhibited, which may have protected them from conviction. As anticipated by theories emphasizing the importance of situational influences on offending, adult-onset offenders, relative to non-offenders, during adulthood more often had schizophrenia, bipolar disorder, and alcohol-dependence, had weaker social bonds, anticipated fewer informal sanctions, and self-reported more offenses. Contrary to some expectations, adult-onset offenders did not have high IQ or high socioeconomic-status families protecting them from juvenile conviction. Conclusions: A tailored theory for adult-onset offenders is unwarranted because few people begin crime de novo as adults. Official adult-onset offenders fall on a continuum of crime and its correlates, between official non-offenders and official juvenile-onset offenders. Existing theories can accommodate adult-onset offenders.