Résumé
Molluscs remains are generally numerous in loess sequences and a rich literature deals with them. But since the proposal of Lozek's method which is widely used in Quaternary molluscan studies, no new methods have been proposed apart from some adjustments. It is a fundamental fact that to propose a new approach implies consideration of the studied material with other concepts. Nevertheless to renew existing methods and concepts needs the opportunities of new techniques, new physical supports but also depends on the scientific hard core of scientists. The generalization and the rising power of computers and micro-computers was one opportunity for the geosciences. More precisely, the renewal of concepts in paleontology leads to new hard cores based on multidisciplinary approaches implying a quantification of data. That is the reason why morphometry appears in evolution studies, quantitative paleoclimatology firstly developed on marine micro-organisms, quantitative paleoecology to characterize the different type of associations, quantitative biostratigraphy — all these approaches widely using multivariate methods. Concerning Quaternary molluscs, I have attempted to develop new concepts because these organisms allow precise studies dealing among others with evolution, biogeography, climatology, and biostratigraphy. They are generally present in the loess sequences which record several climatic cycles, that is to say variations in geosphere-biosphere relations. Consequently analyses of loess molluscs are able, using statistical analyses, to contribute to the understanding of the geosphere-biosphere system.