This last result has an interesting-- but extremely controversial--
consequence. If the underlying neutral mutation rate is constant, then,
by definition, so is the rate of substitution. Then we have what is known
as a ``molecular clock", steadily clicking off substitutions along
lineages, a clock which may be calibrated to real time if we can
accurately date divergence times (usually, from paleontological data) for
at least one pair of lineages. Zuckerkandl and Pauling[16] first
proposed this idea in 1965 after observing what appeared to be remarkably
steady rates of protein evolution in cytochrome c and
hemoglobin
but since then, it
has become clear that, although portions of a phylogeny may exhibit
clock-like behavior, it is risky to generalize this behavior too far. Rate
variation among lineages is widespread, and for many phylogenetic
reconstruction methods, failure to account for this variation can lead to
tremendous difficulties and inaccurate conclusions.