If differences in a gene build up at a roughly steady rate, the number of differences estimates how long ago two lineages split, once the rate is calibrated with fossils.
In Evidence: DNA and Proteins you counted differences between species: more differences, more distant relatives. If differences in a gene build up at a roughly steady rate, they can do more than rank relatives. They can estimate when two lineages split. Emile Zuckerkandl and Linus Pauling proposed this "molecular clock" in the early 1960s.
🔑 Key idea
If a gene changes at a roughly steady rate, then $$\text{time since the split} = \frac{\text{number of differences}}{\text{rate of differences}}$$ The rate is calibrated with a pair of species whose split is dated by fossils. Estimates carry wide ranges, because rates vary between genes and between lineages.
Differences between pairs of species plotted against fossil dates: a rough straight line is what makes a clock (illustrative data).
💡 Why it works
Why would changes be steady? Many mutations are neutral, changing nothing that matters to survival, and neutral mutations become fixed in a lineage at a rate set by the mutation rate. Since mutations occur at roughly steady rates, neutral differences pile up like ticks of a clock. The clock is rough because some genes are under strong selection (and change slowly), generation times differ, and mutation rates differ between groups.
Worked examples
✏️ Worked example: Calibrate, then estimate
Fossils show that species A and B split 20 million years ago; a gene differs at 40 sites between them. The same gene differs at 30 sites between A and C. When did A and C split?
Rate: 40 differences / 20 million years = 2 differences per million years.
Time: 30 / 2 = 15 million years.
Answer: About 15 million years ago (with a wide range of uncertainty).
✏️ Worked example: Two genes, two answers
Using one gene, two lineages split about 12 million years ago; using another, about 18. Is one estimate wrong?
Each gene has its own rate, and each rate was calibrated with uncertain fossil dates.
Selection, generation time and chance all vary.
Scientists combine many genes and several fossil calibrations to narrow the range.
Answer: Not necessarily. Both are estimates; the true date is likely in the range they span, and more data would narrow it.
Different clocks suit different questions. A gene that changes slowly, such as one for ribosomal RNA, can still be compared across bacteria, plants and animals that split billions of years ago. A fast-changing stretch of DNA, such as many parts of mitochondrial DNA, is better for splits within the last few million years, because a slow gene would show almost no differences over so short a time.
⚠️ Common mistake
Treating the clock as exact for every gene. Rates differ between genes (a gene for a vital enzyme changes slowly) and between lineages, so a clock must be calibrated and its estimates carry ranges. When a problem gives you a rate, use it; when you report, say "about".
Practice
Pass with 3 right answers, the last 2 in a row. Each right answer earns 3 XP.