Pointing out that the central subject of Th. Mommsen's research was Roman public law, presented by him as a system, the author puts forward a question: what makes Mommsen's «system» still important for the study of ancient history? To answer this question she considers the judgements of his contemporaries and scholars of later times up to our days, especially those who compare Mommsen's conception with other comprehensive studies of the Roman state system. As the author shows, detractors of Mommsen's «system» were nearly as numerous as its supporters. But what proved to be lasting, she maintains, was the very method of theoretical modelling of historical reality by means of legal categories. This method could not be rejected, consciously or not, by the generations of scholars that followed, for it proved to be fruitful: it provided a key for summarising various facts and gave grounds for a conceptual understanding of sources. Th. Mommsen constructed a logically consistent (free of inner contradictions) model of the Roman state system. A century after Mommsen one can see that his «system» turned out to be too global, too general to «work» equally well for any comparatively short period of Roman history, because it does not take into consideration the specificity of the Roman state system of every particular period. Special analysis shows, e.g., that many legal mechanisms and state institutions of the Early Republic cannot be fitted into Mommsen's «universal scheme». Most probably, the scheme is adequate for the Classical and Late Republic, for it was drawn on the basis of sources referring mainly to that time. But using the method put forward by Mommsen one can construct a theory of public law valid for the Early Republic, for the Principate and for any other period. This is why the importance of Mommsen's «system» consists not so much in the system itself (which is still far from being completely obsolete), as in the method of its construction. The author admits that in the years to come, when specific models of particular state mechanisms together with more general models of Roman state system have been built, one will be able to undertake an effort of constructing a new theory of Roman public law as comprehensive as Mommsen's and based upon a new corpus of special studies. But this will only be possible if one still resorts to the method which Theodor Mommsen was the first to apply, i.e. legal modelling of the Roman state system.