gENETIC dRIfT: EvolUTIoN AT RANdoM 177
size of about 7 million individuals in Australia. Analysis of their genetic variation
suggests that they are descendants of a much smaller population of roughly 19,000
birds that lived some 2 Mya (see Figure 7.6) [2].
There are large differences in the average values of π among species of animals.
Ants and vertebrates tend to have relatively little genetic variation, for example,
while butterflies and bivalves (clams and their relatives) are highly heterozygous.
Differences in effective population size account for some of this variation, but
variation in mutation rates, selective sweeps, background selection, and other fac-
tors must contribute as well. Intriguingly, species with high fecundity and small
propagules tend to have high heterozygosity (FIGURE 7.12). Exactly how those life
history factors affect heterozygosity is not yet clear [9].genetic drift and Natural Selection
As an advantageous mutation sweeps through a population, natural selection is
not the only process that determines its fate. Genetic drift adds a random com-
ponent to its trajectory. As a result, the mutation’s frequency can increase more
rapidly than is expected from selection alone. But sometimes drift causes its fre-
quency to increase less rapidly, or even to decrease.
We can see drift interacting with selection in FIGURE 7.13. In these simulations,
the favored allele has a selective advantage of s = 0.01. In a population of size of Ne
= 500,000, the allele frequency increases along a smooth trajectory much like those
back in Chapter 5 (see Figure 5.7), but with small random fluctuations caused by
genetic drift. When the population size is Ne = 5000, the fluctuations become more
obvious. With a population size of Ne = 50, the effects of drift are so strong that
the tendency of the advantageous allele to spread is no longer obvious. In fact, that
allele is completely lost in two of the five simulations.
These examples illustrate a general point about how genes evolve when both
selection and drift are at work. In some cases, the selection is so much stronger
than drift that the effects of drift can be largely ignored. In other cases, drift over-
whelms selection. A simple rule of thumb tells us if an allele is in one situation or
the other. A natural measure for the strength of selection is the selection coefficient
s (see Chapter 5). Since large populations have weak drift, a natural measure for the
strength of genetic drift is 1/Futuyma Kirkpatrick Evolution, 4eNe. We then simply compare these two numbers. The
Sinauer Associates
Troutt Visual Services
Evolution4e_07.12.ai Date 01-12-20171 4
Heterozygosity (π) (%)0.2100Propagule size (cm)10.010.1 100
Fecundity (offspring per day)100,000FIGURE 7.12 The heterozygosity averaged across the
genome varies among groups of animals. Dots are
colored according to their heterozygosity (π) at syn-
onymous sites in coding regions. Ants and vertebrates
have low heterozygosity, typically between 0.2 per-
cent and 1 percent. Other groups of animals, includ-
ing butterflies and bivalves, have higher heterozygos-
ity, typically between 1 percent and 10 percent. These
differences result from variation in population sizes,
as predicted by Equation 7.1, and from other factors.
Fecundity and propagule size (shown on the x- and
y-axes) are also strongly correlated with heterozygos-
ity, perhaps because they influence population size,
mutation rates, the frequency of selective sweeps,
and the strength of background selection. (After [9].)07_EVOL4E_CH07.indd 177 3/23/17 9:09 AM