Given that DNA replication is achieved by bumbling Brownian motion and ordinary thermodynamics in a biochemical porridge at a temperature of 35 C, it's astonishing that the error-rate of DNA replication is about 10-9 per replicated nucleotide. How can this reliability be achieved, given that the energetic difference between a correct base-pairing and an incorrect one is only one or two hydrogen bonds and the thermal energy kT is only about a factor of four smaller than the free energy associated with a hydrogen bond? If ordinary thermodynamics is what favours correct base-pairing, surely the frequency of incorrect base-pairing should be about