Toward the end of the 1930s, Bernhard Rensch (1900-1990) turned from Lamarckism and orthogenesis to selectionism and became one of the key figures in the making of the Synthetic Theory of Evolution (STE). He contributed to the Darwinization of biological systematics, the criticism of various anti-Darwinian movements in the German lands, but more importantly founded a macroevolutionary theory based on Darwinian gradualism. In the course of time, Rensch's version of the STE developed into an all-embracing metaphysical conception based on a kind of Spinozism. Here we approach Rensch's "selectionist turn" by outlining its context, and by analyzing his theoretical transformation. We try to reconstruct the immanent logic of Rensch's evolution from a "Lamarckian Synthesis" to a "Darwinian Synthesis". We will pay close attention to his pre-Darwinian works, because this period has not been treated in detail in English before. We demonstrate an astonishing continuity in topics, methodology, and empirical generalizations despite the shift in Rensch's views on evolutionary mechanisms. We argue that the continuity in Rensch's theoretical system can be explained, at last in part, by the guiding role of general methodological principles which underlie the entire system, explicitly or implicitly. Specifically, we argue that Rensch's philosophy became an asylum for the concept of orthogenesis which Rensch banned from evolutionary theory. Unable to explain the directionality of evolution in terms of empirically based science, he "pre-programmed" the occurrence of human-level intelligence by a sophisticated philosophy combined with a supposedly naturalistic evolutionary biology.