JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
Back to the library
Biology 1968

Evolutionary Rate at the Molecular Level

Motoo Kimura

Most of evolution's molecular fine print is written not by selection but by sheer chance.

Choose your version
In depth · the introduction

Most of the tiny changes that pile up in our genes over millions of years may be neither good nor bad — just lucky accidents that happened to stick.

The big idea

At the level of molecules, Darwin's survival-of-the-fittest is not the whole story. Most changes to DNA and to proteins are “neutral”: they neither help nor harm the organism, so natural selection simply doesn't notice them. Whether such a change spreads to every individual or vanishes is then settled by sheer chance, generation after generation — a process called genetic drift.

Kimura argued that the great majority of the molecular differences separating species accumulated exactly this way: not because they were useful, but because they were invisible to selection and lucky in the lottery of inheritance.

How it came about

By the 1960s biologists could at last read the sequences of proteins like haemoglobin, letter by letter. Comparing species, they were startled by two things: how fast the sequences had changed, and how steadily. Motoo Kimura — working at Japan's National Institute of Genetics and one of the great mathematical biologists of the century — saw that the pace was simply too fast for natural selection to be paying for every change, an argument he borrowed from J. B. S. Haldane.

His answer, in a note of barely a page and a half in Nature in 1968, was radical: most of it is random. That same year, an ocean away and entirely independently, Jack King and Thomas Jukes reached the same conclusion, publishing it under the pointed title “Non-Darwinian Evolution.” The claim set off one of biology's longest and fiercest arguments.

Why it mattered

It handed biology a clock. If neutral changes accumulate at a steady rate, then the number of molecular differences between two species measures how long ago they shared an ancestor — a “molecular clock” that now dates the entire tree of life.

And it supplied the essential null hypothesis. To claim that natural selection has shaped a particular gene, a biologist must first show its pattern of change is more than chance alone would produce. Kimura's chance became the yardstick against which all selection is measured.

A way to picture it

Imagine copying a long book by hand, again and again, across centuries. Most slips of the pen don't change the meaning — “colour” becomes “color” — so no one bothers to correct them or to keep them on purpose; they drift in and out of the copies at random. Now compare two libraries whose editions descend from the same lost original, and count the harmless differences. That count tells you how long ago the two copying lines split. It ticks at the rate slips are made — not at how carefully anyone reads.

Interactive molecular clock: neutral changes appear as ticks along a lineage from a common ancestor to today. Slide the mutation rate to add or remove ticks; slide the population size across a huge range and the ticks — the clock's beat — stay put.

Where it sits

A century after Darwin placed natural selection at the centre of life, and decades after Fisher, Haldane and Wright rebuilt evolution as mathematics, Kimura insisted that — at the molecular level — selection must share the stage with pure chance. He never denied that selection sculpts eyes and wings; he argued that the silent churn deep in the DNA is mostly drift. The dispute reshaped how biologists read every genome, and the molecular clock it gave us underlies work from human origins to tracking a virus as it spreads.

The original document
Original source text
M. Kimura · Nature 217 (1968): 624–626
The puzzle: molecules change fast — and steadily
Drawing on the first protein sequences, Kimura compares the haemoglobin α and β chains and cytochrome c across mammals. The amino-acid sequences have diverged at a striking, near-constant pace — on the order of one substitution per 100-residue chain every ten million years. Extrapolated across a mammalian genome of some four billion base pairs, that implies, he argues, roughly one new nucleotide substitution becoming fixed every two years.
Calculating the rate of evolution in terms of nucleotide substitutions seems to give a value so high that many of the mutations involved must be neutral ones.
Why selection cannot pay for it all
Kimura invokes Haldane's “cost of natural selection.” Every gene substitution driven by positive selection demands a quota of selective deaths — individuals lacking the favoured type that must be removed. At the molecular rate just estimated, the cumulative cost of driving every substitution by selection would be unbearably large. He concludes that the great majority of these substitutions must instead be selectively neutral, fixed not by selection but by random genetic drift in finite populations.
[ … ]
The consequence: a clock
A neutral mutant's chance of eventual fixation equals its starting frequency, and new mutants arise in exact proportion to population size — so population size cancels, and the long-run substitution rate equals the neutral mutation rate. Molecular evolution should therefore tick at a roughly constant rate per year: a molecular evolutionary clock. The full communication, with its substitution-rate figures and its load calculation, runs to under three pages and is available in full at the source below.
National Institute of Genetics, Mishima · 1968