The evolutionary biologist "studies the steps by which the miraculous adaptations so characteristic of every aspect of the organic world have evolved". But the general nature of such adaptive steps is still unclear. Evolution is often thought to be random and dependent on unpredictable events. In this light, one might expect the steps taken by adaptation to be completely random, both biologically and temporally. Here I present a mathematical derivation to show that, on the contrary, adaptive steps can have fairly strong rhythm. I find that the strength of the adaptive rhythm, that is its relative temporal regularity, is equal to a constant that is the same for all microbial populations. As a consequence, numbers of accumulated adaptations are predicted to have a universal variance/mean ratio. The theory derived here is potentially applicable to the study of molecular evolution.