Each branch of the tree accumulates substitutions as an independent Poisson process: expected count per site λ = r·L·m, where r is the substitution rate, L the branch length in time, and m a per-branch rate multiplier (1 for a strict clock, random for a relaxed clock).
Jukes-Cantor observed difference: p = (3/4)(1 - e^(-4λ/3))
JC corrected distance: d = -(3/4) ln(1 - 4p/3)
Molecular-clock age estimate: t_est = d / (2r)
- Species sampled — number of present-day tips (lineages) descending from one common ancestor.
- Root age — true elapsed time since the common ancestor; the tree's radius is normalised so the shape stays framed at any age.
- Substitution rate — expected substitutions per site per million years along a strict-clock branch.
- Clock model — strict gives every branch the same rate; relaxed multiplies each branch by an independent random factor, the real-world violation that molecular dating has to correct for.
- Small orange dots along each branch are individual substitutions; branch colour (blue→red) encodes how much divergence that branch accumulated.
Real-world relevance: this is the logic behind dating the human–chimpanzee split (~6–7 Mya) or building the COI "DNA barcode" tree of life — genetic distance between living species, corrected for hidden multiple substitutions, is converted into a divergence time via an assumed substitution rate.