The Cretaceous–Cenozoic history of angiosperms resulted in a certain character of the distribution of the number of taxa belonging to different ranks (number of species and genera within a family, S/G ratio in families, and the number of species within a genus). In most cases, such distributions are satisfactorily described by a power law (a Pareto distribution). In logarithmic coordinates, the power function represents a straight line. Empirical curves coincide with this line fairly well, except for the right part of the graphs (the area of small-sized taxa), where empirical curves noticeably deviate from theoretical ones. This fact means that the volumes of small-sized taxa should be significantly larger to fit the theoretical curves properly. The performed modeling of the S/G ratio in families has shown a satisfactory correspondence between the observed and calculated number of species within a wide iteration range only when the dynamic factor of extinction was applied. An assumption has been made that a differentiated extinction of species took place in the course angiosperm evolution. In this case, the extinction rate should be minimal for genera with a large number of species. On the contrary, a decrease in the number of species per a genus may drastically (by orders) increase the extinction rate. As a result, large-sized genera become larger in size, while small-sized genera become smaller; the frequency distribution of species within genera changes according to the power law. The initial divergence in the number of taxa, which determines their further division into larger and smaller sizes, could be caused by the emergence and expansion of herbs characterized by better functional and adaptive potential.