Limestone-marl alternations (LMA) are rhythmical successions of carbonate-rich sedimentary rocks. They are often assumed to record cyclic sedimentation linked to Milankovitch cycles. In spite of the importance of LMA for a range of questions in geosciences, it is not unequivocally understood how they originate. The two models explaining their origin both assume carbonate redistribution, either by late diagenetic pressure dissolution amplifying primary depositional differences, or through early diagenetic aragonite dissolution and reprecipitation as calcite, creating LMA even in the absence of primary differences. The latter model is known as differential diagenesis. As both models can imply different interpretations of paleoenvironmental conditions, the identification of the generating process is essential. This study addresses the question how to distinguish the generating process through statistical comparison of taphonomic characteristics of marls and limestones in thin sections by: (1) measuring the relative abundance of originally aragonitic and calcitic components in the fossil assemblages, and (2) by analysis of their orientations. Based on four sets of thin sections from different paleoenvironments from the Upper Ordovician to the Permian, the model of late diagenetic pressure-induced carbonate redistribution is ruled out. The results point towards early diagenetic aragonite dissolution and reprecipitation as calcite as the main diagenetic process generating LMA. Furthermore, the influence of primary sedimentary differences is demonstrated. This approach offers a tool to gauge conditions during sedimentation and a way to assess the systematically poorer preservation of aragonitic components in marine deposystems (aragonite bias) quantitatively.