36 CHAPTER 2In some cases, fossils provide very important information about the evolution-
ary history of a group, including relationships among its members (see Chapter
16). But many groups of organisms have left no fossil record at all, and even in the
best cases, the fossil record is incomplete. We will concern ourselves mostly with
how to infer phylogenies from data on living organisms. Each trait of an organ-
ism (e.g., the number of toes on a hindlimb) is called a character, which may have
various character states (e.g., five toes in humans, three in rhinoceroses, one in
horses). All kinds of phenotypic characters have been used, especially morpho-
logical features (which are usually the only features we can use for fossilized
extinct taxa). Phylogenetic study has been revolutionized by DNA sequencing,
which reveals variation at thousands or even millions of base pair positions in
homologous DNA sequences. Each position (“site”) on one strand of the double
helix represents a character, and the identity of its nucleotide base (A, T, C, or G)
represents a character state.
Homologous character states that are shared among species provide evidence of
common ancestry if they evolved only once. Using DNA sequence data, we begin
our discussion of how phylogenies are estimated with an example. Imagine that we
want to find the phylogeny showing the evolutionary relations among three species
of squirrels in the genus Sciurus. We have sequenced part of the hemoglobin gene
from an individual of each of four species: the eastern gray squirrel (Sciurus caroli-
nensis), western gray squirrel (S. griseus), fox squirrel (S. niger), and ground squirrel
(Spermophilus citellus) (FIGURE 2.10A). The homologous sequence fragments from
the hemoglobin gene are shown in FIGURE 2.10B. The ground squirrel serves as
an outgroup. This is a taxon that we are quite sure (based on prior evidence) is
more distantly related to the three species of interest. Those three species are the
ingroup. The outgroup/ingroup distinction immediately gives us a basic framework
for the phylogenetic tree: the outgroup and the ingroup form two branches from the
common ancestor of all the species. Given this framework, there are three possible
evolutionary trees for these species (FIGURE 2.10C).
We know from many studies of the hemoglobin gene that changes to the
sequence are rare over short evolutionary time spans. This means that if we
compare possible phylogenies, those that require fewer evolutionary changes are
more likely to reflect actual relationships than are those with more changes. That
logic makes it simple to find the evolutionary tree that most likely represents the
history of the four DNA sequences and hence the four species of squirrels.
Look at tree 1 in Figure 2.10C. At site 3, the eastern and western gray squirrels
share an A, and they differ from the other two species, which share a T. Starting
with the DNA sequence at the root of the tree, the evolution from T to A (shown
by the red bar on the tree) happened in the common ancestor of these two spe-
cies. At site 9, the evolution from A to T occurred in the ancestor of the fox squir-
rel (shown by the blue bar). Tree 2 therefore involves two evolutionary changes.
Now consider tree 2. At site 3, there were two changes from T to A (shown by
the two red bars) and again one change at site 9, for a total of three changes. The
same conclusion applies to tree 3: at least three changes must have occurred to
produce the data at the tips of the tree, that is, the DNA sequences from the four
species. (We can imagine other scenarios for where and when changes occurred
on the tree, but they all require at least three changes.)
To sum up, the phylogeny that requires the fewest evolutionary changes is tree 1.
Given our assumption that evolutionary changes to the hemoglobin sequence are
rare, this is the most likely phylogeny. This logic for estimating phylogenies is called
parsimony. A final question you may have is how we could possibly have known the
DNA sequence of the ancestor at the root of the tree. For example, that species could
have had an A rather than a T at site 1. But if it did, all three phylogenies require at02_EVOL4E_CH02.indd 36 3/23/17 8:59 AM