An evolutionary progression leading toward replication is resolved into several phases; (a) the replication of RNA segments by self-priming and -templating, (b) the replication of single stranded molecules by elongation and controlled scission, (c) replication of complementary duplexes and (d) replication of DNA. The initial phase is suggested by evidence for the existence of tandem repeats in an early population of molecules presumed to be ancestral to today's structural RNAs. Relics of these repeats are seen in the positioning of sequence matches between transfer and ribosomal RNAs. Conservation of the positions of the matches is indicated by persistence of a periodicity in their spacings along the molecules. Selection is viewed as a vector, with a source and a focus. The evolutionary progression entails shifts in the source of selection, from external catalysts to the replicating molecule itself, and in its focus, from substrate to replicator, to the products of the replicator's activity. When the source and focus of selection are the same selection becomes internalized, and replication and Darwinian evolution follow. Catalytic specificity is regarded as an antecedent to natural selection. Shifting of the source and focus of selection and switches in evolution's 'vehicle', the most fundamental thing that evolves, result in profound changes in the modes of evolution. Control provides a conceptual framework within which entry into a Darwinian mode of evolution, and ultimately liberation from Darwinian evolution might be explained.