It is shown that stoichiometric rather than thermodynamic dissociation constants can and should be used in the buffer capacity theory. The mean valence of bivalent weak acids and bases is deduced in a form with dimensionless parameters. The expressions for the molar buffer capacities of bivalent weak acids and bases are identical with that for a bivalent ampholyte, and the theory for the latter given earlier is thus valid in all details for bivalent weak acids and bases.