Anthropologists think the change in human stature we see before us, the ”˜human secular trend,’ is too fast to be evolution. Geneticists think rapid change in laboratory selection experiments is artificial and couldn’t be sustained in nature. Paleontologists think evolution is slow because rates calculated for change in the fossil record are infinitesimal. Who is right and who is wrong? How fast is evolution?

Many phenotypic traits of interest in evolution are normally (log-normally) distributed because they are polygenic: differences between populations are naturally measured in standard deviation units. Evolution is a process that takes place through time. Natural selection acts generation by generation on phenotypic variance: evolutionary time is appropriately measured in generations. More time usually means more difference. Hence rates are calculated for comparison, in units called haldanes (not darwins): standard deviations per generation, not factors of e per million years.

Here things get interesting because a rate per generation on a time scale of one generation is rarely the same as a rate per generation on a time scale of ten, or a hundred, or a thousand generations. Scale is important. Further, there are different kinds of selection. Stabilizing selection and directional selection have cumulative fractal scaling effects, meaning they leave fingerprints and can be recognized by characteristic rate distributions.

How fast is evolution? Ask the question carefully because the answer depends on the question. Are you interested in the evolutionary process or the history of life?