1965). Natural selection is rather fickle, subject to the vagaries of environmental change
and organism interactions, whereas mutation rate for a given gene is likely to remain
constant through time and across lineages. The number of amino acid replacements (or
nucleotide substitutions) between species would, then, reflect the time elapsed since their
last common ancestor. The time of remote events, as well as the degree of relationship
among contemporary lineages could, thus, be determined on the basis of amino acid (or
nucleotide) differences. A notable feature of the hypothesis of the molecular evolutionary
clock is multiplicity: every one of the thousands of proteins or genes of an organism is an
independent clock, each ticking at a different rate but all measuring the same events
(Ayala 1986; Gillespie 1991; Li 1997).
Early investigations showed that the evolution of the globins in vertebrates conformed
fairly well to the clock hypothesis, which allowed reconstruction of the history of globin
gene duplications (Zuckerkandl and Pauling 1965). Fitch and Margoliash (1967) would
soon provide a ‘genetic distance’ method that was effectively used for reconstruction of
the history of 20 organisms, from yeast to moth to human, based on the amino acid
sequence of a small protein, cytochrome c. A theoretical foundation for the clock was
provided by Kimura (1968), who developed a ‘neutral theory of molecular evolution’,
with great mathematical simplicity; notably, the theory states that the rate of substitution
of adaptively equivalent (‘neutral’) alleles, k, is precisely the rate of mutation, u, of neutral
alleles, k=u. The neutrality theory predicts that molecular evolution behaves like a
stochastic clock, such as radioactive decay, with the properties of a Poisson distribution, in
which the mean, M, and variance, V, are expected to be identical, so that V/M=1. The
‘index of dispersion’, measuring the deviation of this ratio from the expected value of 1,
is a way to test whether observations fit the theory.
Experimental data have shown that often the rate of molecular evolution is
‘overdispersed’, that is, that the index of dispersion is often significantly greater than 1
(Gillespie 1991; Li 1997). Deviations from rate constancy occur between lineages, for
example between rodents and mammals, as well as at different times along a given lineage,
both factors having significant effects (Langley and Fitch 1974). Consequently, several
modifications of the neutral theory have been proposed, seeking to account for the excess
variance of the molecular clock. It has been proposed, for example, that most protein
evolution involves ‘slightly deleterious’ replacements rather than strictly neutral ones
(Ohta 1973); or that certain ‘biological properties’, such as the effectiveness of the error-
correcting polymerases, vary among organisms, so that mutation rates for a given gene
vary from one organism to another (Kimura and Ohta 1972; Kimura 1980, 1983;
Gillespie 1991; Li and Graur 1997). A ‘population size’ hypothesis proposes that
organisms with larger effective population size have a slower rate of evolution than
organisms with smaller population size, because the time required to fix new mutations
increases with population size (Ohta 1972; Kimura 1983). Another supplementary
hypothesis invokes a generation-time effect. Protein evolution has been extensively
investigated in primates and rodents with the common observation that the number of
replacements is greater in the rodents (Kohne 1970; Li et al. 1996). In plants, the overall
rate at the rbcL locus is more than five times greater in annual grasses than in palms, which
have much longer generations (Gaut et al. 1992). These rate differences could be
accounted for, according to the generation-effect hypothesis, by assuming that the time-rate
6 FRANCISCO RODRÍGUEZ-TRELLES ET AL.