The explanatory system developed by Skinner culminated in the formulation of an explanatory mechanism that should address the behavior as an analogy between natural selection of Darwin and operant conditioning. Skinner believed that this analogy would be enough to sort the behavioral disciplines into three explanatory levels: phylogenetic, ontogenetic and cultural. However, before Darwin’s theory became a universal paradigm for biology, it was necessary to find a substrate on which selection could act in order to test the limits and scope given by Darwinian formulation, i.e., Mendel’s discoveries about genetic transmission of hereditary characters. Beyond neodarwinist synthesis, the Experimental Analysis of Behavior still does not have a biological basis for the test of the Skinner hypothesis about the selection of operant behavior as an analogue of natural selection. In addition, there is not a mathematical model to predict the distribution of variability of individual repertoire in an analogue of Hardy-Weinberg Law. What is the impact of these inconsistencies on the theory of selection by consequences? Before accepting the analogy between operant conditioning and natural selection, it is necessary to understand the laws of variation and retention of behavior and to explain how this sensitivity occurs and how it affects behavior.