The impacts of quarterly adult arrest rates on later male serious delinquency prevalence rates were investigated in Philadelphia police districts (N = 23) over several years using all male delinquents aged 10–15 years who were mandated to more than “straight” probation. An ecological deterrence model expects more arrests to lead to less delinquency later. A community justice or mass incarceration model, the ecological version of general strain theory, and an ecologized version of the procedural justice model, each anticipates more arrests lead to more delinquency later. Investigating quarterly lags from 3 to 24 months between adult arrests and later delinquency, the results showed a time-dependent relationship. Models with short lags showed the negative relationship expected by ecological deterrence theory. Models with lags of about a year and a half showed the positive relationship expected by the other three theories. Indicators needed so future works can gauge the relative merits of each theoretical perspective more accurately are described. The spatial distributions of current and 1920s delinquency rates were compared.