On careful examination of existing kinetic data for correct and incorrect dNTP incorporations by a variety of DNA polymerases, it is apparent that these enzymes resist a unified description. Instead, the picture that emerges is a rather complex one: for most polymerases, there is evidence for a noncovalent step preceding phosphoryl transfer, but there are less reliable data for determining whether the noncovalent step or phosphoryl transfer is rate-limiting during misincorporation. Although the structural conservation in the polymerase superfamily is probably reflected in a common set of intermediates along the reaction pathway, the energetics of these species vary even when closely related polymerases are compared. Consequently, some polymerases apparently show more discrimination between correctly paired and mispaired dNTPs in the binding step, and polymerases may differ in terms of which step of the reaction is rate-limiting in correct and incorrect insertion reactions. Because of the higher energy barrier in the misincorporation reaction, at least some of the intermediates both before and after the rate-limiting step in the misincorporation pathway will have higher energies than the corresponding intermediates in correct incorporation; consequently, these steps can serve as kinetic checkpoints.