Since the great atmospheric pollution accidents from sulfur oxides and suspended particulate matters, the concentrations of these pollutants have notably reduced. The experimental concentrations which are necessary to obtain results are even more important than these "historical" concentrations. Among the aggregate of epidemiologic surveys regarding these problems, some find a relationship between strong concentrations of pollution and morbidity and mortality. Such results are explained by references to experimental and metrologic knowledge. As far as epidemiological problems are concerned infectious and meteorological factors have to be eliminated to determine which is the place of atmospheric pollution. But such a place in all respects being so small, has to be defined in a more precise way. One cannot be certain whether the improvement of epidemiological and metrological methods may give more conclusive results, taking into account the foreseeable reduction of the pollution in the years to come.