The author analyses the concept and the aim of models and drags the attention upon those which have no correlative in the reality, as they are only approximate and are representative of a scheme only in a normative and not in a descriptive sense. He then makes a distinction between causal and functional models and makes clear how some of them can mutually exclude one another, whilst others can integrate themselves either alternatively or reciprocally. He shows how those that are mutually exclusive, can be used in the problems of "hypothesis election" and deals with those which, though apparently exclusive, can not only be conciliate but even applied simultaneously or sometimes alternatively. After mentioning the inconvenience of using static models with a view to explain dynamic phenomena, the author recalls how the organistic conception could arrive to an ample vision of this problem. At the end he gives some examples of integrative models pointing out their applicability to the economic phenomena.El artículo reproduce el texto de la conferencia brindada por el autor el 15 de julio de 1955 en la Facultad de Ciencias Económicas (UNLP).Instituto de Investigaciones Económicas