Summary — In his last book. Comparative Methods in the Social Sciences, Professor Smelser develops a well-grounded analysis of these procedures, usually referred to as «comparative method». The basic premise of Smelser's work is that the methodological problems of comparative analysis are not different in nature from those involved in scientific enterprise as such, provided that the latter be conceived as the effort to explain regularities and variations between phenomena. In this connection, Smelser traces the procedures of empirical hypothesis-validation followed by Tocqueville, Durkheim and Weber, who— albeit of different epistemological persuasions — all attributed a prominent place to comparative method. From this analysis, and from the analysis of more recent contributions as well, emerges that comparative method constitutes a logical equivalent of the canons of experimental inquiry formulated by J.S. Mill. The difference between the former and the latter is only a pragmatical one, lying in the impossibility of carrying on artificial manipulation in the case of comparative method. The a. agrees with Smelser's argument, though rejecting some of its more debatable implications, particularly assimilating comparative and statistical methods.