Civil service systems are a key analytic unit in modern public administration, but theory about them lags behind their practical importance and the attention they have received from scholars. This article argues that theory building has focused on bureaucracy rather than civil service systems per se and that empirical research has been dominated by functionalist logic. Future research needs to define more rigorously and measure the properties of civil service systems and study them more holistically, particularly in relation to their environments. Three ideal types are discussed to illustrate some of the implications of these recommendations.