20

CHAPTER

Genes Within Populations

Part IV Evolution

Chapter Contents
20.1 Genetic Variation and Evolution
20.2 Changes in Allele Frequency
20.3 Five Agents of Evolutionary Change
20.4 Quantifying Natural Selection
20.5 Reproductive Strategies
20.6 Natural Selection’s Role in Maintaining
Variation
20.7 Selection Acting on Traits Affected by
Multiple Genes
20.8 Experimental Studies of Natural Selection
20.9 Interactions Among Evolutionary Forces
20.10 The Limits of Selection

tamoncity/Shutterstock

Visual Outline Introduction

Individuals in most populations
Genetic variation in
vary greatly in phenotype, as the
populations
pigeons shown above illustrate;
acted these differences are often the re-
on by sult of genetic differences among
Rosita So Image/Getty Images
individuals. Often the particular
Processes that cause characteristics of an individual
evolutionary change have an important bearing on its
survival, on its chances to repro-
include
duce, and on the success of its
offspring. Evolution is driven by
such factors, as the frequency of
Mutation Gene flow Nonrandom mating Genetic drift Natural selection
different alleles rises and falls in
occurs when populations. These deceptively
phenotypes differ in simple matters lie at the core of
Average Tendency to Fly Toward Light

Light 11
10
9 evolutionary biology, which is the
8
7
Fitness topic of this chapter and of chap-
6
5
ters 21–24.
4
differences result from
3
2
Dark 1
0 2 4 6 8 10 12 14 16 18 20 Number of offspring
Survival Mating success
Number of Generations per mating event
 20.1 Genetic Variation and
Evolution

Learning Outcomes
1. Define evolution and population genetics.
2. Explain the importance and extent of variation within
populations.

The term genetic variation refers to the different alleles of genes

found within individuals of a population. Natural populations
contain a wealth of such variation. In this chapter, we explore ge-
netic variation in natural populations and consider the evolutionary
forces that cause allele frequencies in natural populations to change.
The word evolution is widely used in the natural and social
sciences. It refers to how an entity—be it a social system, a gas, or
a planet—changes through time. Although development of the
modern concept of evolution in biology can be traced to Darwin’s
Figure 20.1 Polymorphic variation. This natural population
of lupines, Lupinus, exhibits considerable variation in flower color.
landmark work, On the Origin of Species, the first five editions of
Individual differences are inherited and passed on to offspring.
his book never actually used the term. Rather, Darwin used the Rosita So Image/Getty Images
phrase “descent with modification.”
Although many more complicated definitions have been evolution, therefore, without also considering population genet-
proposed, Darwin’s words probably best capture the essence of ics, the study of genetic variation within populations.
biological evolution. Through time, species accumulate differ-
ences; as a result, descendants differ from their ancestors. In this Populations contain ample genetic variation
way, new species arise from existing ones. Genetic variation is required for evolution to occur. Consequently,
biologists have always wanted to know how much genetic variation
Many processes can lead exists in natural populations. The ability to ask this question has
to evolutionary change been limited by the techniques available to analyze variation at dif-
ferent levels: proteins, genes, and now genomes. The story of ge-
You have already learned about the development of Darwin’s ideas netic variation in natural populations is told using increasingly
in chapter 1. Darwin was not the first to propose a theory of sophisticated tools for detecting differences.
evolution. Rather, he followed a long line of earlier philosophers Initial approaches examined the most obvious differences,
and naturalists who deduced that the many kinds of organisms morphological variation. In many cases, such phenotypic variation
around us were produced by a process of evolution. is the result of underlying genetic differences. In such cases, natu-
Unlike his predecessors, however, Darwin proposed natural ral populations usually show substantial genetic variation, as
selection as the mechanism of evolution. Natural selection produces figure 20.1 illustrates.
evolutionary change when some individuals in a population possess
certain inherited characteristics and then produce more surviving
offspring than individuals lacking these characteristics. As a result,
the population gradually comes to include more and more individu-
als with the advantageous characteristics. In this way, the popula-
tion evolves and becomes better adapted to its local circumstances.
One way to monitor how populations change through time is
to look at changes in the frequencies of alleles of a gene from one
generation to the next. Natural selection, by favoring individuals
with certain alleles, can lead to change in such allele frequencies,
but it is not the only process that can do so. Allele frequencies can
also change when mutations occur repeatedly, changing one allele
to another, and when migrants bring alleles into a population. In
addition, the frequencies of alleles can change randomly as the
result of chance events, particularly when populations are small.
Often, natural selection overwhelms the effects of these other pro-
cesses, but as you will see, this is not always the case.
Evolution can result from any process that causes a change Figure 20.2 Variation within a population of
in the genetic composition of a population. We cannot talk about humans. Diverse Images/Universal Images Group/Getty Images

chapter 20 Genes Within Populations 417

 Other examples are closer to home. Consider all of the varia- some regions of the genome were more variable than others—just as
tion in a human population: skin, eye, and hair color; hair type; with humans—and patterns of variation differed between the popula-
height; eye and nose shape; blood type; navel innie or outie; and tions. In another study, 133 rhesus macaque (Macaca mulatta) ge-
many more. Most human variation is the result of underlying ge- nomes were sequenced, revealing substantial variation among these
netic differences (figure 20.2). monkeys. Undoubtedly, the number of species for which many indi-
The first approach to directly assay genetic variation within viduals are sequenced will skyrocket in the next few years, substan-
populations was the use of electrophoresis to separate proteins tially enhancing our understanding of population variation.
produced by alternative alleles of enzyme-encoding genes (see
­chapter 17). As each new DNA analysis tool was developed in the Learning Outcomes Review 20.1
laboratory, it was quickly adapted for use with samples collected from
Evolution can be described as descent with modification. Variation
natural populations. This led to, in order of increasing detail, examin-
is required for evolution to occur, and most populations contain
ing restriction fragment length polymorphisms (RFLPs), sequencing extensive phenotypic and genotypic variation. Population genetics
specific genes, and, most recently, sequencing entire genomes. studies this variability.
One of the most useful tools, both for genetic mapping and
for the analysis of population-level variation, has been single- ■■ Why is genetic variation in a population necessary for
nucleotide polymorphisms (SNPs). These are defined as single-base evolution to occur?
differences between individuals that exist in the population at more
than 1% (see chapter 24). An initial large-scale international effort
identified several million SNPs in the human genome, a number
we now know to be a vast underestimate. 20.2 Changes in Allele Frequency
Such variation is also being assayed in many other species of
interest. The results for most species are similar to those in humans:
the closer you look, the more variation you see. One reason for the
appeal of SNPs is that we have been able to automate the analysis of Learning Outcomes
multiple samples. SNP data are now available for thousands of spe- 1. Explain the Hardy–Weinberg principle.
cies, revealing the near ubiquity of genetic variation and providing 2. Describe the characteristics of a population that is in
tools for assessing patterns in human and natural populations. Hardy–Weinberg equilibrium.
As genome sequencing moves into a phase where multiple ge- 3. Demonstrate how the operation of evolutionary processes can
nomes are being sequenced for many different species, a landscape of be detected.
massive genetic variation is being revealed. Not surprisingly, the spe-
cies with the most genomes sequenced is our own: by the time you Genetic variation within natural populations was a puzzle to
read this, more than 100,000 human genomes will have been partially Darwin and his contemporaries in the mid-1800s. Although
or entirely sequenced. As we saw in chapter 18, this revealed an enor- Mendel performed his experiments during this same time period,
mous amount of genetic variation individuals will have around 3% of his work was largely unknown. Selection, scientists then thought,
their genome varying from the reference genome. should always favor an optimal form, and so tend to eliminate
As well as extensive variation within human populations, these variation. Moreover, the theory of blending inheritance—in
studies are also documenting substantial variation between popula- which offspring were expected to be phenotypically intermediate
tions. This kind of work is in progress looking at genomes from spe- relative to their parents—was widely accepted. If blending in-
cifically Asian and African populations, and in one case a Dutch heritance were correct, then the effect of any new genetic variant
population. The early results from these forays into human genetic would quickly be diluted to the point of disappearance in subse-
variation are significant and surprising. Not only do these populations quent generations. For example, if a mutation made an individual
have additional genetic variation not previously seen, but they also 4 inches taller, its offspring, the result of a mating with an indi-
have entire genomic regions that are not represented in the reference vidual without that mutation, would be only 2 inches taller, and
genome which comes from individuals in North America. We have in the next generation’s offspring would be only 1 inch taller, and
long known that African populations have greater genetic diversity so on, until the effect completely disappeared.
than European and Asian populations. This is a result of human his-
tory: our species originated in Africa, and the populations there repre- The Hardy–Weinberg principle allows
sent the “oldest” human genomes. A recent study that involved
sequencing over 900 African genomes, from different regions of Af-
prediction of genotype frequencies
rica, showed that taking all of these genomes into account, there was Following the rediscovery of Mendel’s research, two people in 1908
more than 300 Mb of DNA not found in the reference genome. It is solved the puzzle of why genetic variation persists—Godfrey
becoming very clear that within a species, different populations have H. Hardy, an English mathematician, and Wilhelm Weinberg, a
unique evolutionary trajectories and that substantial variation within a German physician. These workers were initially confused about why,
species is distributed across the species’ geographic range. after many generations, a population didn’t come to be composed
Studies of genomic variation in natural populations are under solely of individuals with the dominant phenotype. The conclusion
way as well. More than 1000 genomes of the laboratory fruit fly, they independently came to was that the original proportions of the
Drosophila melanogaster, have been completed. Comparisons of genotypes in a population will remain constant from generation to
these genomes revealed extensive variation, although interestingly, generation, as long as the following assumptions are met:

418 part IV Evolution

 1. No mutation takes place. pened is p * p = p2 (figure 20.3). By the same reasoning, the prob-
2. No genes are transferred to or from other sources (no ability that an individual will have two b alleles is q2.
immigration or emigration takes place). What about the probability that an individual will be a hetero-
3. Mating is random (individuals do not choose mates based zygote? There are two ways this could happen: the individual could
on their phenotype or genotype). receive a B from its father and a b from its mother, or vice versa.
4. The population size is very large. The probability of the first case is p * q and the probability of the
5. No selection favoring one genotype over another occurs. second case is q * p. Because the result in either case is that the in-
If these assumptions are met, the genotypes’ proportions will dividual is a heterozygote, the probability of that outcome is the sum
not change, and the population is said to be in Hardy–Weinberg of the two probabilities, or 2pq.
equilibrium. So, to summarize, if a population is in Hardy–Weinberg
equilibrium with allele frequencies of p and q, then the probabil-
ity that an individual will have one of the three possible geno-
The Hardy–Weinberg equation with two alleles: types is p2 + 2pq + q2. You may recognize this as the binomial
abinomial expansion expansion:
In algebraic terms, the Hardy–Weinberg principle is written as an
(p + q)2 = p2 + 2pq + q2
equation. Consider a population of 100 cats in which 84 are black and
16 are white. The frequencies of the two phenotypes would be 0.84 Finally, we may use these probabilities to predict the distri-
(or 84%) black and 0.16 (or 16%) white. Based on these phenotypic bution of genotypes in the population, again assuming that the
frequencies, can we deduce the underlying frequency of genotypes? population is in Hardy–Weinberg equilibrium. If the probability
If we assume that the white cats are homozygous recessive that any individual is a heterozygote is 2pq, then we would expect
for an allele we designate as b, and the black cats are either the proportion of heterozygous individuals in the population to be
homozygous dominant BB or heterozygous Bb, we can calculate 2pq; similarly, the frequency of BB and bb homozygotes would
the frequencies of the two alleles in the population from the pro- be expected to be p2 and q2.
portion of black and white individuals, assuming that the population Let us return to our example. Remember that 16% of the
is in Hardy–Weinberg equilibrium. cats are white. If white is a recessive trait, then this means that
Let the letter p designate the frequency of the B allele and the such individuals must have the genotype bb. If the frequency of
letter q the frequency of the alternative allele. We call these allele this genotype is q2 = 0.16 (the frequency of white cats), then q
frequencies. Because there are only two alleles, p plus q must (the frequency of the b allele) = 0.4 (0.4 is the square root of
always equal 1 (that is, the total population). In addition, we know 0.16). Because p + q = 1, therefore, p, the frequency of allele B,
that the sum of the three genotype frequencies must also equal 1. would be 1.0 – 0.4 = 0.6 (remember, the frequencies must add up
The probability that an individual will have two B alleles is simply to 1). We can now easily calculate the expected genotype fre-
the probability that each of its alleles is a B. The probability of two quencies: homozygous dominant BB cats would make up the p2
events happening independently is the product of the probability of group, and the value of p2 = (0.6)2 = 0.36, or 36 homozygous
each event; in this case, the probability that the individual received dominant BB individuals in a population of 100 cats. The hetero-
a B allele from its father is equal to the frequency of the allele in zygous cats have the Bb genotype and would have the frequency
the population, p, and the probability the individual received a B corresponding to 2pq, or (2 * 0.6 * 0.4) = 0.48, or 48 heterozy-
allele from its mother is also p, so the probability that both hap- gous Bb individuals.

Generation One Generation Two

B b
Phenotypes 84% 16% Eggs
p = 0.60 q = 0.40

B BB Bb
p = 0.60 p2 = 0.36 pq = 0.24

Sperm
Genotypes BB Bb bb b Bb bb
Frequency of q = 0.40 pq = 0.24 q2 = 0.16
0.36 0.48 0.16
genotype in population

Frequency of gametes 0.36 + 0.24 = 0.60B 0.24 + 0.16 = 0.40b p2 + 2 pq + q2 = 1

Figure 20.3 The Hardy–Weinberg equilibrium. In the absence of factors that alter them, the frequencies of gametes, genotypes, and
phenotypes remain constant generation after generation.

Data analysis If all white cats died, what proportion of the kittens in the next generation would be white?

chapter 20 Genes Within Populations 419

Using the Hardy–Weinberg equation to predict more bb homozygotes than expected; then clearly, the population
frequencies in subsequent generations is not in Hardy–­Weinberg equilibrium.
The Hardy–Weinberg equation is another way of expressing the What could cause such an excess of homozygotes and deficit
Punnett square described in chapter 12, with two alleles assigned of heterozygotes? A number of possibilities exist, including
frequencies, p and q. Figure 20.3 allows you to trace genetic reas- (1) natural selection favoring homozygotes over heterozygotes,
sortment during sexual reproduction and see how it affects the (2) individuals choosing to mate with genetically similar individu-
frequencies of the B and b alleles during the next generation. als (because BB * BB and bb * bb matings always produce homo-
In constructing this diagram, we have assumed that the union zygous offspring, but only half of Bb * Bb produce heterozygous
of sperm and egg in these cats is random. The alleles are therefore offspring, such mating patterns would lead to an excess of
represented in the next generation in proportion to their original homozygotes), or (3) an influx of homozygous individuals from
occurrence. Each individual egg or sperm in each generation has a outside populations (or conversely, emigration of heterozygotes to
0.6 chance of receiving a B allele (  p = 0.6) and a 0.4 chance of other populations). By detecting a lack of Hardy–Weinberg equi-
receiving a b allele (q = 0.4). librium, we can generate potential hypotheses that we can then
In the next generation, therefore, the chance of combining investigate directly.
two B alleles is p2, or 0.36 (that is, 0.6 * 0.6), and approximately The operation of evolutionary processes can be detected in a
36% of the individuals in the population will continue to have the second way. If all of the Hardy–Weinberg assumptions are met,
BB genotype. The frequency of bb individuals is q2 (0.4 * 0.4) and then allele frequencies will stay the same from one generation to
so will continue to be about 16%, and the frequency of Bb indi- the next. Changes in allele frequencies between generations would
viduals will be 2pq (2 * 0.6 * 0.4), or on average, 48%. indicate that one of the assumptions is not met.
Phenotypically, if the population size remained at 100 cats, Suppose, for example, that the frequency of b was 0.53 in
we would still see approximately 84 black individuals (with either one generation and 0.61 in the next. Again, there are a number of
BB or Bb genotypes) and 16 white individuals (with the bb geno- possible explanations: for example, (1) selection favoring individ-
type). Allele, genotype, and phenotype frequencies have remained uals with b over B, (2) immigration of b into the population or emi-
unchanged from one generation to the next, despite the reshuffling gration of B out of the population, or (3) high rates of mutation that
of genes that occurs during meiosis and sexual reproduction. more commonly occur from B to b than vice versa. Another pos-
One important point to remember is that the terms “domi- sibility is that the population is a small one, and that the change
nant” and “recessive” refer only to the phenotype of the heterozy- represents the random fluctuations that result because, simply by
gote and in no way imply that one allele is selectively superior chance, some individuals pass on more of their genes than others.
to another. For this reason, in Hardy-Weinberg equilibrium, when se- We will discuss how each of these processes is studied in the rest
lection is not operating, allele frequencies are not expected to change of the chapter.
regardless of whether one trait is dominant over another.
Learning Outcomes Review 20.2
Hardy–Weinberg predictions can be The Hardy–Weinberg principle states that in a large population
with no selection and random mating, the proportion of alleles
applied to data to find evidence does not change through the generations. Finding that a
of evolutionary processes population is not in Hardy–Weinberg equilibrium indicates that
one or more evolutionary agents are operating.
The lesson from the example of black and white cats is that if all
five of the assumptions listed earlier in this section hold true, the ■■ If you know the genotype frequencies in a population, how
allele and genotype frequencies will not change from one genera- can you determine whether the population is in Hardy–
tion to the next. But in reality, most populations in nature will not Weinberg equilibrium?
fit all five assumptions. The primary utility of this method is to ■■ What would you conclude if you found a population not
determine whether some evolutionary process or processes are in Hardy–Weinberg equilibrium? What would be your
operating in a population and, if so, to suggest hypotheses about next step?
what they may be.
Suppose, for example, that the observed frequencies of the
BB, bb, and Bb genotypes in a different population of cats were
0.6, 0.2, and 0.2, respectively. We can calculate the allele frequen-
cies for B as follows: 60% (0.6) of the cats have two B alleles, 20%
have one, and 20% have none. This means that the average number 20.3 Five Agents of Evolutionary
of B alleles per cat is 1.4 [(0.6 × 2) + (0.2 × 1) + (0.2 × 0) = 1.4].
Because each cat has two alleles for this gene, the frequency is
Change
1.4/2.0 = 0.7. Similarly, you should be able to calculate that the
frequency of the b allele = 0.3. Learning Outcomes
If the population were in Hardy–Weinberg equilibrium, 1. Define the five processes that can cause evolutionary change.
then, according to the equation earlier in this section, the frequen- 2. Explain how these processes can cause populations to
cy of the BB genotype would be 0.72 = 0.49, lower than it really is. deviate from Hardy–Weinberg equilibrium.
Similarly, you can calculate that there are fewer heterozygotes and

420 part IV Evolution

The five assumptions of the Hardy–Weinberg principle also stages of plants or marine animals from one place to another.
indicate the five agents that can lead to evolutionary change in ­Pollen, the male gamete of flowering plants, is often carried great
populations. They are mutation, gene flow, nonrandom mating, distances by insects and other animals that visit flowers (figure
genetic drift in small populations, and the pressures of natural se- 20.4b). Seeds may also blow in the wind or be carried by animals
lection. Any one of these may bring about changes in allele or to new populations far from their place of origin.
genotype proportions, causing the population to not be in Hardy- Consider two populations initially different in allele frequen-
Weinberg equilibrium. cies: in population 1, p = 0.2 and q = 0.8; in population 2, p = 0.8
and q = 0.2. Gene flow brings alleles into a population in propor-
Mutation changes alleles tion to their frequency in the source population. In this case, it
would disproportionately bring in alleles that are rare in the recepi-
Mutation from one allele to another can obviously change the pro- ent population because those alleles are common in the source.
portions of particular alleles in a population. Mutation rates are Thus, allele frequencies would change from generation to genera-
generally so low that they have little effect on the Hardy–Weinberg tion, and the populations would not be in Hardy–Weinberg equilib-
proportions of common alleles. A typical gene mutates about once rium. Only when allele frequencies reach 0.5 for both alleles in
per 100,000 cell divisions. Because this rate is so low, other evolu- both populations will equilibrium be attained (assuming that rates
tionary processes are usually more important in determining how of gene flow are the same in both directions). This example also
allele frequencies change. indicates that gene flow tends to homogenize allele frequencies
Nonetheless, mutation is the ultimate source of genetic among populations.
variation and thus makes evolution possible (figure 20.4a). It is
important to remember, however, that the likelihood that a particu-
lar mutation will occur is not affected by natural selection; that is,
Nonrandom mating shifts genotype
particular mutations do not occur more frequently in situations in frequencies
which they would be favored by natural selection. Individuals with certain genotypes sometimes mate with one an-
other more commonly than would be expected on a random basis,
Gene flow occurs when alleles move a phenomenon known as nonrandom mating (­figure 20.4c).
Assortative mating, in which phenotypically similar individuals
between populations mate, is a type of nonrandom mating that causes the frequencies of
Gene flow is the movement of alleles from one population to another. particular genotypes to differ greatly from those predicted by the
It can be a powerful agent of change. Sometimes gene flow is obvi- Hardy–Weinberg principle.
ous, as when an animal physically moves from one place to another. Assortative mating does not change the frequency of the indi-
If the characteristics of the newly arrived individual differ from those vidual alleles because it does not change the reproductive success of
of the animals already there, and if the newcomer is adapted well individuals, but rather changes with whom they mate. Because phe-
enough to the new area to survive and mate successfully, the genetic notypically similar individuals are likely to be genetically similar
composition of the receiving population may be altered. and thus are also more likely to produce offspring with two copies
Other important kinds of gene flow are not as obvious. These of the same allele, assortative mating will increase the proportion of
subtler movements include the drifting of gametes or the immature homozygotes in the next generation.

Mutation Gene Flow Nonrandom Mating Genetic Drift Selection

Mutagen DNA Self-fertilization

C
G
T

C
G
A

a. The ultimate source of b. A very potent agent of c. Inbreeding, of which mating d. Statistical accidents. The e. The only agent that
variation. Individual change. Individuals or with one's self is the most random fluctuation in allele produces adaptive
mutations occur so rarely gametes move from one extreme form, is a common frequencies increases as evolutionary changes.
that mutation alone usually population to another. type of nonrandom mating. It population size decreases.
does not change allele does not alter allele frequency,
frequency much. but reduces the proportion of
heterozygotes.

Figure 20.4 Five agents of evolutionary change. a. Mutation. b. Gene flow. c. Nonrandom mating. d. Genetic drift. e. Selection.

chapter 20 Genes Within Populations 421

 Inbreeding results when closely related individuals mate
with each other. Because relatives are genetically similar, in-
breeding will have the same result as assortative mating and in-
crease the proportion of homozygotes. In fact, because relatives
tend to be phenotypically similar, assortative mating can lead to
inbreeding.
By contrast, disassortative mating, in which pheno­typically
different individuals mate, produces an excess of heterozygotes.

Genetic drift may alter allele

frequencies in small populations Parent Bottleneck Surviving Next
population (drastic reduction individuals generation
In small populations, frequencies of particular alleles may change in population)
drastically by chance alone. Such changes in allele frequencies oc-
cur randomly, as if the frequencies were drifting from their values. Figure 20.5 Genetic drift: a bottleneck effect. The
These changes are thus known as genetic drift (figure 20.4d). For parent population contains roughly equal numbers of green and yellow
this reason, a population must be large to be in Hardy–Weinberg individuals and a small number of red individuals. By chance, the few
equilibrium: although drift occurs in all populations, the smaller remaining individuals that contribute to the next generation are mostly
the population, the greater the change is in allele frequencies from green and the red allele has disappeared entirely. The bottleneck occurs
generation to generation as a result of genetic drift. because so few individuals form the next generation, as might happen
If the gametes of only a few individuals form the next gen- after an epidemic or a catastrophic storm.
eration, the alleles they carry may by chance not be representative
of the parent population from which they were drawn, as illustrated
in figure 20.5. In this example, a small number of individuals are
? Inquiry question Why are rare alleles particularly likely
to be lost in a population bottleneck?
removed from a bottle. By chance, most of the individuals removed
are green, so the new population has a much higher population of present in the source population. Thus, some alleles may be lost
green individuals than the parent generation had. from the new population, and others may change drastically in
Suppose, for example, that from a population in which B and frequency. In some cases, previously rare alleles in the source pop-
b are equally represented (that is, p = q = 0.5), four individuals ulation may be a significant fraction of the new population’s
formed the next generation, and that by chance they were two Bb genetic endowment. This phenomenon, in which the allele fre-
heterozygotes and two BB homozygotes—that is, the allele fre- quencies of the founding individuals are different from the allele
quencies in the next generation would be p = 0.75 and q = 0.25. In frequencies in the population from which they came, is called the
fact, if you were to replicate this experiment 1000 times, each time founder effect.
randomly­drawing four individuals from the parental population, Founder effects are not rare in nature. Many ­self-­pollinating
then in about 8 of the 1000 experiments, one of the two alleles plants start new populations from a single seed. Founder effects
would be missing entirely. have been particularly important in the evolution of organisms on
This result leads to an important conclusion: genetic drift distant oceanic islands, such as the Hawaiian and Galaˊpagos
can lead to the loss of alleles in isolated populations. Alleles that Islands. Most of the organisms in such areas probably derive from
initially are uncommon are particularly vulnerable (figure 20.5). one or a few initial founders. Although rare, such events are
A set of small populations that are isolated from one an- occasionally observed, such as when a mass of vegetation carrying
other may come to differ strongly as a result of genetic drift, even several iguanas washed up on the shore of the Caribbean island of
if the forces of natural selection are the same for all. Because of Anguilla in 1996, leading to the establishment of a population that
genetic drift, sometimes harmful alleles may increase in frequen- still occurs there to this day.
cy in small populations, despite selective disadvantage, and In a similar way, isolated human populations begun by
favorable alleles may be lost even though they are selectively relatively few individuals are often dominated by genetic features
advantageous. characteristic of their founders. Amish populations in the United
Although genetic drift occurs in any population, its effect is States, for example, were established from relatively few individu-
inversely related to population size. Thus, it is particularly likely in als and have unusually high frequencies of a number of conditions,
populations that were founded by a few individuals or in which the such as polydactylism (the presence of a sixth finger).
population was reduced to a very small number at some time in the
past. Such founder effects or bottlenecks are particularly clear The bottleneck effect
examples of genetic drift. Even if organisms do not move from place to place, occasionally
their populations may be drastically reduced in size. This may re-
The founder effect sult from flooding, drought, epidemic disease, and other natural
Sometimes one or a few individuals disperse and become the forces, or from changes in the environment. Just as we have seen
founders of a new, isolated population at some distance from their with founder effects, the few surviving individuals may by chance
place of origin. These pioneers are not likely to carry all the alleles have allele frequencies different from those in the entire

422 part IV Evolution

 population, and some alleles may be lost entirely (natural selection Evolution by natural selection occurs when the following
may also be responsible for change in allele frequencies, but here conditions are met:
we are referring to random changes resulting from small popula-
1. Phenotypic variation must exist among individuals
tion size). The resulting alterations and loss of genetic variability
in a population. Natural selection works by favoring
have been termed the bottleneck effect.
individuals with some traits over individuals with
The genetic variation of some living species appears to be
alternative traits. If no variation exists, natural selection
severely depleted, probably as the result of a bottleneck effect in
cannot operate.
the past. For example, the northern elephant seal, which breeds on
2. Variation among individuals must result in
the western coast of North America and nearby islands, was nearly
differences in the number of offspring surviving
hunted to extinction in the 19th century and was reduced to a sin-
in the next generation. This is natural selection.
gle population containing perhaps no more than 20 individuals on
Because of their phenotype or behavior, some
the island of Guadalupe off the coast of Baja California
individuals are more successful than others in
(figure 20.6). As a result of this bottleneck, the species has lost
producing offspring. Although many traits are
almost all of its genetic variation, even though the seal populations
phenotypically variable, individuals exhibiting
have rebounded and now number in the tens of thousands and
variation do not always differ in survival and
breed in locations as far north as near San Francisco.
reproductive success.
Any time a population becomes drastically reduced in num-
3. Phenotypic variation must have a genetic basis. For
bers, such as in endangered species, the bottleneck effect is a po-
natural selection (see item 2) to result in evolutionary
tential problem. Even if population size rebounds, the lack of
change, the selected differences must have a genetic
variability may mean that the species remains vulnerable to extinc-
basis. Not all variation has a genetic basis—even
tion—a topic to which we will return in chapter 58.
genetically identical individuals may be phenotypically
quite distinctive if they grow up in different
Selection favors some environments. Such environmental effects are common
phenotypes over others in nature. In many turtles, for example, individuals
that hatch from eggs laid in moist soil are heavier, with
As Darwin pointed out, some individuals leave behind more prog-
longer and wider shells, than individuals from nests in
eny than others, and the rate at which they do so is affected by their
drier areas.
phenotype. We describe the results of this process as selection,
when individuals with one phenotype leave, on average, more sur- When phenotypically different individuals do not differ
viving offspring in the next generation than individuals with an genetically, differences in the number of their offspring will not
alternative phenotype (see figure 20.4e). In artificial selection, a alter the genetic composition of the population in the next gen-
breeder selects for the desired characteristics. In natural selection, eration, and thus, no evolutionary change will have occurred.
environmental conditions determine which individuals in a popu- It is important to remember that natural selection and evolu-
lation produce the most offspring. tion are not the same—the two concepts often are incorrectly

population in 1890,
reduced to inhabiting
Guadalupe only
UNITED current population
S TAT E S

Figure 20.6 Bottleneck effect: case study. Because northern elephant seals,
Mirounga angustirostris, live in very cold waters, they have thick layers of fat, for which
they were hunted nearly to extinction late in the 19th century. At the low point, only one
population remained, on Guadalupe Island, with perhaps as few as 20 individuals; during
this time, genetic variation was lost. Since being protected, the species has reclaimed
most of its original range and now numbers in the tens of thousands, but genetic
variation will recover only slowly over time as mutations accumulate.

Guadalupe

MEXICO

chapter 20 Genes Within Populations 423

equated. Natural selection is a process, whereas evolution is the Selection to match climatic conditions
historical record, or outcome, of change through time. Natural Many studies of selection have focused on genes encoding en-
selection (the process) can lead to evolution (the outcome), but zymes, because in such cases the investigator can directly assess
natural selection is only one of several processes that can result in the consequences to the organism of changes in the frequency of
evolutionary change. Moreover, natural selection can occur alternative enzyme alleles.
without producing evolutionary change; only if variation is geneti- Often investigators find that enzyme allele frequencies vary
cally based will natural selection lead to evolution. with latitude, so that one allele is more common in northern
populations, but is progressively less common at more southern
Selection to avoid predators locations. A superb example is seen in studies of a fish, the mum-
A common result of evolution driven by natural selection is that michog (Fundulus heteroclitus), which ranges along the eastern
populations become better adapted to their environment. Many of coast of North America. In this fish, geographic variation occurs in
the most dramatic documented instances of adaptation involve allele frequencies for the gene that produces the enzyme lactate
genetic changes that decrease the probability of capture by a preda- dehydrogenase, which catalyzes the conversion of pyruvate to lac-
tor. The caterpillar larvae of the common sulphur butterfly Colias tate (see section 7.8).
eurytheme usually exhibit a pale green color, providing excellent Biochemical studies show that the enzymes formed by these
camouflage against the alfalfa plants on which they feed. An alleles function differently at different temperatures, thus explain-
alternative bright yellow color morph is reduced to very low fre- ing their geographic distributions. The form of the enzyme more
quency because this color renders the larvae highly visible on the frequent in the north is a better catalyst at low temperatures than is
food plant, making it easier for bird predators to see them (see the enzyme from the south. Moreover, studies indicate that at low
figure 20.4e). temperatures, individuals with the northern allele swim faster, and
One of the most dramatic examples of background match- presumably survive better, than individuals with the alternative
ing involves ancient lava flows in the deserts of the American allele.
Southwest. In these areas, the black rock formations produced
when the lava cooled contrast starkly with the surrounding bright Selection for pesticide and microbial resistance
glare of the desert sand. Populations of many species of animals A particularly clear example of selection in natural populations
occurring on these rocks—including lizards, rodents, and a vari- is provided by studies of pesticide resistance in insects. The wide-
ety of insects—are dark in color, whereas sand-dwelling popula- spread use of insecticides has led to the rapid evolution of resistance
tions in surrounding areas are much lighter (figure 20.7). in more than 500 pest species. The cost of this evolution, in terms
Predation is the likely cause for these differences­in color. of crop losses and increased pesticide use, has been estimated at
Laboratory studies have confirmed that predatory birds such as $3–8 billion per year.
owls are adept at picking out individuals occurring on backgrounds In the housefly, the resistance allele of the pen gene decreas-
to which they are not adapted. es the uptake of insecticide, whereas alleles of the kdr and dld-r
genes decrease the number of target sites, thus decreasing the bind-
ing ability of the insecticide (figure 20.8). Other alleles enhance
the ability of the insects’ enzymes to identify and detoxify insec-
ticide molecules.
Single genes are also responsible for resistance in other or-
ganisms. For example, Norway rats are normally susceptible to the
pesticide warfarin, which diminishes the clotting ability of the rat’s
Light coat color pocket mouse blood and leads to fatal hemorrhaging. However, a resistance allele
is vulnerable on lava rock
of the VKORC1 gene reduces the ability of warfarin to bind to its
Light coat color favored by
natural selection because
target enzyme and thus renders it ineffective.
it matches sand color Selection imposed by humans has also led to the evolution of
resistance to antibiotics in many disease-causing pathogens. For
example, Staphylococcus aureus, which causes staph infections,
was initially treated by penicillin, which latched onto S. aureus,
degrading cell walls and causing the death of the bacteria. However,
Dark coat color favored by within four years of mass-production of the drug, evolutionary
natural selection because change in S. aureus modified an enzyme, penicillinase, so that it
it matches black lava rock would attack penicillin, making the drug unable to attach to S. aureus
and thus rendering it ineffective. Since that time, several other drugs
Figure 20.7 Pocket mice from the Tularosa Basin of have been developed to attack the microbe, and each time resistance
New Mexico whose color matches their background. Black has evolved. As a result, staph infections have re-emerged as a major
lava formations are surrounded by desert, and selection favors coat health threat. The speed by which adaptation occurs may be surpris-
color in pocket mice that matches their surroundings. Genetic studies ing, but remember that microbes have enormous population sizes.
indicate that the differences in coat color are the result of small Even though mutation rates are low, the vast size of these popula-
differences in the DNA of alleles of a single gene. tions guarantees a steady supply of new mutations for natural

424 part IV Evolution

 Selection occurs when individuals with one phenotype leave more
Pesticide molecule Target site surviving offspring in the next generation than individuals with an
alternative phenotype. Evolutionary biologists quantify reproduc-
tive success as fitness, the number of surviving offspring left in
the next generation.
Fitness is a relative concept; the most fit phenotype is simply
Resistant target site Insect cell membrane the one that produces, on average, the greatest number of
offspring.
a. Insect cells with resistance allele at pen gene:
decreased uptake of the pesticide.
A phenotype with greater fitness
usually increases in frequency
Target site
Suppose, for example, that in a population of toads, two pheno-
types exist: green and brown. Suppose, further, that green toads
leave, on average, 4.0 offspring in the next generation, but brown
toads leave only 2.5. By custom, the most fit phenotype is assigned
a fitness value of 1.0, and other phenotypes are expressed as rela-
tive proportions. In this case, the fitness of the green phenotype
would be 4.0/4.0 = 1.000, and the fitness of the brown phenotype
b. Insect cells with resistance allele at kdr gene: would be 2.5/4.0 = 0.625. The difference in fitness would therefore
decreased number of target sites for the pesticide.
be 1.000 – 0.625 = 0.375. A dif­ference in fitness of 0.375 is quite
Figure 20.8 Selection for pesticide resistance. large; natural selection in this case strongly favors the green
Resistance alleles of genes such as pen and kdr allow insects to be phenotype.
more resistant to pesticides. Insects that possess these resistance If differences in color have a genetic basis, then we would
alleles have become more common through selection. expect evolutionary change to occur; the frequency of green toads
should be substantially greater in the next generation. Further, if
the fitness of two phenotypes remained unchanged, we would
selection to utilize. The effect of such evolution on human health is expect alleles for the brown phenotype eventually to disappear
enormous. In the United States alone, 2 million people each year from the population.
become ill due to antibiotic-resistant bacteria, and 23,000 die from
such infections.
? Inquiry question Why might the frequency of green
toads not increase in the next generation, even if color
differences have a genetic basis?
Learning Outcomes Review 20.3
Five factors can bring about deviation from the predicted
Hardy–Weinberg genotype frequencies. Of these, only Fitness may consist of many components
selection regularly produces adaptive evolutionary change, but Although selection is often characterized as “survival of the
the genetic constitution of populations, and thus the course ­fittest,” differences in survival are only one component of fitness.
of evolution, can also be affected by mutation, gene flow,
Even if no differences in survival occur, selection may
nonrandom mating, and genetic drift.
operate if some individuals are more successful than others in
■■ How does each of these processes cause populations to attracting mates. In many territorial animal species, for example,
vary from Hardy–Weinberg equilibrium? large males mate with many females, and small males rarely get
to mate. Selection with respect to mating success is termed
sexual selection; we describe this topic more fully in the next
section.
In addition, the number of offspring produced per reproduc-
tive event is also important. Large female frogs and fish lay more
20.4 Quantifying Natural Selection eggs than do smaller females, and thus they may leave more off-
spring in the next generation.
Fitness is therefore a combination of survival, mating success,
and number of offspring per mating. Selection favors phenotypes
Learning Outcomes with the greatest fitness, but predicting fitness from a single compo-
1. Define evolutionary fitness. nent can be tricky because traits favored for one component of fit-
2. Explain the different components of fitness. ness may be at a disadvantage for others. As an example, in water
3. Demonstrate how the success of different phenotypes can be striders, larger females lay more eggs per day (figure 20.9). Thus,
compared by calculating their relative fitness. natural selection at this stage favors large size. However, larger
females also die at a younger age and thus have fewer opportunities

chapter 20 Genes Within Populations 425

 8 50 200

Number of Eggs Laid

40

Life Span of Adult

6 150
Number of Eggs

During Lifetime
Female (days)
Laid per Day

30
4 100
20
2 50
10

0 0 0
12 13 14 15 16 12 13 14 15 16 12 13 14 15 16
Length of Adult Female Water Strider (mm) Length of Adult Female Water Strider (mm) Length of Adult Female Water Strider (mm)

Figure 20.9 Body size and egg-laying in water striders. Larger female water striders lay more eggs per day (left panel), but also
survive for a shorter period of time (center panel). As a result, intermediate-sized females produce the most offspring over the course of their
entire lives and thus have the highest fitness (right panel).

Inquiry question What evolutionary change in body size might you expect? If the number of eggs laid per day was not affected by
? body size, would your prediction change?

Data analysis Assuming that the values on the x-axis represent the final body size of different animals and that the y-axis
represents egg-laying rate and survival once they reach that size, how many eggs would you expect a 12-mm water strider to lay?
And a 15-mm strider?

to reproduce than smaller females. Overall, the two opposing direc- they attempt to maximize fitness. Such a difference in reproductive
tions of selection cancel each other out, and the intermediate-sized behavior is clearly seen in mate choice. Darwin was the first to ob-
females leave the most offspring in the next generation. serve that females often do not mate with the first male they encoun-
ter, but instead seem to evaluate a male’s quality and then decide
whether to mate. Peahens prefer to mate with peacocks that have
Learning Outcomes Review 20.4 more eye­spots on their elaborate tail feathers (figure 20.10b, c). Simi-
Fitness is defined by an organism’s reproductive success relative to larly, female frogs prefer to mate with males having more acousti-
other members of its population. This success is determined cally complex, and thus attractive, calls. This behavior, called mate
by how long it survives, how often it mates, and how many offspring choice, is well known in many invertebrate and vertebrate species.
it produces per mating. Relative fitness assigns numerical values to
different phenotypes relative to the most fit phenotype. The sexes often have different
■■ Is one of these factors always the most important in reproductive strategies
determining reproductive success? Explain.
Males are selective in choosing a mate much less frequently than
females. Why should this be? Many of the differences in reproduc-
tive strategies between the sexes can be understood by comparing
the parental investment made by males and females. Parental
investment refers to the energy and time each sex “invests” in pro-
20.5 Reproductive Strategies ducing and rearing offspring; it is, in effect, an estimate of the en-
ergy expended by males and females in each reproductive event.
Numerous studies have shown that females generally have a
higher parental investment. One reason is that eggs are much larger
Learning Outcomes than sperm—195,000 times larger in humans! Eggs contain
1. Explain parental investment and the prediction it makes about proteins and lipids in the yolk and other nutrients for the develop-
mate choice. ing embryo, but sperm are little more than mobile DNA packages.
2. Understand the difference between intra- and intersexual In some groups of animals (mammals, for example), females are
selection. responsible for gestation and lactation, costly reproductive
3. Describe how sexual selection leads to the evolution of functions only they can carry out.
secondary sexual characteristics. The consequence of such inequalities in reproductive invest-
ment is that the sexes face very different selective pressures. Be-
cause any single reproductive event is relatively inexpensive for
Males and females have the common goal of improving the quantity males, they can best increase their fitness by mating with as many
and quality of offspring they produce, but usually differ in the way females as possible. This is because male fitness is likely not

426 part IV Evolution

 Number of Mates
5

0
140 150 160
Number of Eyespots in Male Tail Feathers

c.

Figure 20.10 Products of sexual selection. In bird species

such as a. the peacock, Pavo cristatus, and b. the raggiana bird of
paradise, Paradisaea raggiana, males use their much longer tail
feathers in courtship displays. c. Female peahens prefer to mate with
males having greater numbers of eyespots in their tail feathers. (a) Carole_
R/Flickr Flash/Getty Images; (b) Bruce Beehler/Science Source

Inquiry question Why do females prefer males with

? more spots?
a. b.
Data analysis Suppose a male had 155 eyespots. How
many females would you expect him to mate with? If you had
multiple males with 155 eyespots, do you think they all would
limited by the amount of sperm they can produce. By contrast, each mate with the same number of females?
reproductive event for females is much more costly, and the number
of eggs that can be produced often limits reproductive success. For
this reason, a female should be choosy, trying to pick the male that
can provide the greatest benefit to her offspring and thus improve
her fitness. consider sexual selection to be distinctive from natural selection,
These conclusions hold only when female reproductive but others see it as a subset of natural selection, just one of the
investment is much greater than that of males. In species with many factors affecting an organism’s fitness.
biparental care, males may contribute equally to the cost of raising Sexual selection involves both intrasexual selection, or
young; in this case, the degree of mate choice should be more competitive interactions between members of one sex (“the
equal between the sexes. power to conquer other males in battle,” as Darwin put it), and
In some cases, male investment exceeds that of females. intersexual selection, which is another name for mate choice
For example, male Mormon crickets transfer a protein-contain- (“the power to charm”). Sexual selection leads to the evolution
ing packet (a spermatophore) to females during mating. Almost of structures used in combat with other males, such as a deer’s
30% of a male’s body weight is made up by the spermatophore, antlers and a ram’s horns, as well as ornaments used to
which provides nutrition for the female and helps her develop “persuade” members of the opposite sex to mate, such as long
her eggs. As we might expect from our model of mate choice, in tail feathers and bright plumage (see figure 20.10a, b). These
this case it is the females that compete with one another for ac- traits are called secondary sexual characteristics.
cess to males, which are the choosy sex. I­ ndeed, males are quite Selection strongly favors any trait that confers greater ability
selective, favoring heavier females. Heavier females have more in mate competition. Larger body size is a great advantage if
eggs; thus, males that choose larger females leave more offspring dominance is important, as it is in territorial species. Males may thus
(figure 20.11). be considerably larger than females. Such differences between the
Males care for eggs and developing young in many species, sexes are referred to as sexual dimorphism. In other species,
including seahorses and a number of birds and insects. As with structures used for fighting, such as horns, antlers, and large canine
Mormon crickets, these males are often choosy, and females com- teeth, have evolved to be larger in males because of the advantage
pete for mates. they give in intrasexual competition.
Sometimes sperm competition occurs between the sperm of
Sexual selection occurs through mate different males if females mate with multiple males. This type of
competition, which occurs after mating, has selected for features that
competition and mate choice increase the probability that a male’s sperm will fertilize the eggs.
The reproductive success of an individual is determined by how Such features include sperm-transfer organs designed to remove the
long the individual lives, how frequently it mates, and how many sperm of a prior mating, large testes to produce more sperm per
offspring it produces per mating. The second of these factors, com- mating, and sperm that hook themselves together to swim more
petition for mates, is termed sexual selection. Some people rapidly.

chapter 20 Genes Within Populations 427

 Indirect benefits of mate choice. In many species, however,
males provide no direct benefits of any kind to females. In such
120 cases, it is not intuitively obvious what females have to gain by be-
ing “choosy.” Moreover, what could be the possible benefit of
choosing a male with an extremely long tail or a complex song?
100
One possible indirect benefit of mate choice is that it could lead to
Number of Mature Eggs

higher-quality offspring.
80 A number of theories have been proposed as to how
this might work. One idea is that females choose the male
60 that is the healthiest or oldest. Large males, for example, have
probably been successful at living long, acquiring a lot of food,
40 and resisting parasites and disease. In other species, features
other than size may indicate a male’s condition. In guppies and
20
some birds, the brightness of a male’s color reflects the quality of
his diet and overall health. Females may gain two benefits from
mating with the healthiest males. First, healthy males are less
0 likely to be carrying diseases, which might be transmitted to the
Female Body Weight
female during mating; this would be a direct benefit. Second, to
the extent that the males’ success in living long and prospering is
Figure 20.11 The advantage of male mate the result of his genetic makeup, the female will be ensuring that
choice. Male Mormon crickets, Anabrus simplex, her offspring receive good genes from their father, an indirect
choose heavier females as mates, and larger females benefit.
have more eggs. Thus, male mate selection Several experimental studies in fish and moths have examined
increases fitness. whether female mate choice leads to greater reproductive success. In
these experiments, females in one group were allowed to choose the
Inquiry question Is there a benefit to
? females for mating with large males?
males with which they would mate, whereas males were randomly
mated to a different group of females. Offspring of females that
chose their mates were more vigorous and survived better than
offspring from females given no choice, which suggests that females
preferred males with a better genetic makeup.
A variant of this theory goes one step further. In some
Intrasexual selection cases, females prefer mates with traits that appear to be detri-
In many species, individuals of one sex—usually mental to survival (see figure 20.10c). The long tail of the
males—compete with one another for the opportunity
to mate. Competition among males can occur for a ter-
ritory in which females feed or bear young. Males may
also directly compete for the females themselves. A few
successful males may engage in an inordinate number of
matings, whereas most males do not mate at all. For example,
elephant seal males control territories on breeding beaches and a
few dominant males do most of the breeding (figure 20.12). On one
beach, for example, eight males impregnated 348 females, whereas
the remaining males mated rarely, if at all.

Intersexual selection
Intersexual selection concerns the active choice of a mate. Mate
choice has both direct and indirect benefits.

Direct benefits of mate choice. In some cases, the female

benefits directly from her choice of mates. If males help raise off-
spring, females benefit by choosing the male that can provide the
best care—the better the parent, the more offspring she is likely to
rear. In other species, males provide no care, but maintain territo-
ries that provide food, nesting sites, and predator refuges. In red
deer, males that hold territories with the highest-quality grasses Figure 20.12 Male northern elephant seals (Mirounga
mate with the most females. In this case, there is a direct benefit of angustirostris) compete for mates. Male northern elephant seals
a female mating with such a territory owner: she feeds with little fight with one another for possession of territories. Only the largest males
disturbance on quality food. can hold territories, which contain many females. ©Cathy & Gordon ILLG

428 part IV Evolution

 Frequency (kHz)

Frequency (kHz)
3 3
2 2
1 1
0 0
0.2 0.4 0.2 0.4
Time (seconds) Time (seconds)

a. b. c.
Figure 20.13 Male Túngara frog, Physalaemus pustulosus, calling. Female frogs of several species in the genus Physalaemus
prefer males that include a “chuck” in their call. However, only males of the Túngara frog a. produce such calls b. males of other species
do not c. W. Perry Conway/Getty Images

peacock is a hindrance in flying and makes males more vulner- with white crests, but not green or red ones. Why this might be is
able to predators. Why should females prefer males with such not at all clear; it suggests there is something in the zebra finch’s
traits? The handicap hypothesis states that only genetically neural hardwiring that for some reason finds white much more
superior mates can survive with such a handicap. By choosing attractive than green or red. Regardless of the cause of the prefer-
a male with the largest handicap, the female is ensuring that her ence, this finding suggests that if a mutation arose that caused a
offspring will receive these quality genes. Of course, the male male to have a white crest on his head, the mutation would be-
offspring will also inherit the genes for the handicap. For this stow a great advantage and would quickly spread through the
reason, evolutionary biologists are still debating the merit of population. Perhaps there are many such latent preferences in
this hypothesis. species—preferences we might never be able to predict and that
might be hard to explain. But if appropriate variation arises in a
Alternative theories about the evolution of mate choice. population, females may prefer it. This is a fascinating area of
Some courtship displays appear to have evolved from a research that is just in its infancy, and will require coupling stud-
predisposition in the female’s sensory system to respond to ies of neurophysiology and behavior to understand how and why
certain stimuli. For example, females may be better able to detect such mate choice preferences evolve.
particular colors or sounds at a certain frequency, and thus be
attracted to such signals. Sensory exploitation involves the evo-
lution in males of a signal that “exploits” these preexisting bias-
es. For example, if females are particularly adept at detecting red
objects, then red coloration may evolve in males as part of a
courtship display.
To understand the evolution of courtship calls, consider
the vocalizations of the Túngara frog (figure 20.13). Unlike re-
lated species, males include a short burst of sound, termed a
“chuck,” at the end of their calls. Recent research suggests that
not only are females of this species particularly attracted to
calls of this sort, but so are females of related species, even
though males of these species do not produce “chucks.” This
pattern suggests that the preference evolved for some reason in
the ancestor of all of the species, and then subsequently led to
the evolution of the chuck only in males of the Túngara frog.
Why males of the other species haven’t evolved to produce
chucks is a good question.
The opportunity for sensory exploitation may be wide-
spread. For example, researchers conducted an experiment to
see if birds had latent preferences for particular stimuli using a
bird common in the pet trade, the zebra finch. To conduct the
experiment, the researchers glued long feathers vertically to the
tops of the heads of male birds (figure 20.14). The result was that
the males had very tall crests, completely unlike anything seen in Figure 20.14 Male zebra finches with artificial crests.
nature. They then presented males with different colored crests to Researchers glued different colored feathers on the heads of male zebra
females to see if the females had a preference for one color over finches to see if females exhibited a preference for one color over
another. Surprisingly, females were strongly attracted to males another.

chapter 20 Genes Within Populations 429

 A great variety of other hypotheses have been proposed to SCIENTIFIC THINKING
explain the evolution of mating preferences. Many of these
hypotheses may be correct in some circumstances, but none seems Question: Does negative frequency-dependent selection maintain
capable of explaining all of the variation in mating behavior in the variation in a population?
animal world. This is an area of vibrant research, with new Hypothesis: Fish may disproportionately capture water boatmen (a type
discoveries appearing regularly. of aquatic insect) with the most common color.
Experiment: Place predatory fish in different aquaria with the different
frequencies of the color types in each aquarium.
Learning Outcomes Review 20.5 Result: Fish prey disproportionately on the common color in each
The sex that invests more in reproduction (parental investment) aquarium. The rare color in each aquarium generally survives best.
tends to exhibit mate choice. Females or males can be
selective, depending on the energy and time they devote to 100
parental care. Sexual selection refers to differences among
individuals of the same sex in mating success. Intrasexual
selection results when individuals of the same sex compete 80
for matings, whereas intersexual selection occurs when

Taken by Fish Predators

Percent of Color Type
members of one sex choose which individuals of the other
sex with whom to mate. Sexual selection governs evolution of
60
secondary sex characteristics in that mates are chosen on the
basis of phenotype and competitive success.
■■ Pipefish males incubate young in a brood pouch, a process 40
that takes two weeks, during which time males cannot Color type of
accept further eggs. Which sex would you expect to show water boatman
mate choice? Why? 20 dark brown
medium brown
light brown

20 40 60 80 100
Color Type Frequency in Population

20.6 Natural Selection’s Role in

Maintaining Variation Figure 20.15 Frequency-dependent selection.

Data analysis How can this experiment distinguish

between frequency-dependent and directional selection?
Learning Outcomes
1. Define frequency-dependent selection, oscillating selection,
and heterozygote advantage. phenotype frequency, hence the term negative frequency-dependent.
2. Explain how these processes affect the amount of genetic Assuming a genetic basis for phenotypic variation, such selection
variation in a population. will have the effect of making rare alleles more common, thus main-
taining variation.
Negative frequency-dependent selection can occur for
Natural selection has been discussed as a process that removes varia- many reasons. For example, it is well known that animals or
tion from a population by favoring one allele over others at a gene lo- people searching for something form a “search image.” That is,
cus. However, in some circumstances, selection can do exactly the they become particularly adept at picking out certain objects.
opposite and actually maintain population variation. Consequently, predators may form a search image for common
prey phenotypes. Rare forms may thus be preyed upon less
Frequency-dependent selection may favor frequently.
either rare or common phenotypes An example is fish predation on an insect, the water boat-
man, which occurs in three different colors. Experiments indicate
In some circumstances, the fitness of a phenotype depends on that each of the color types is preyed upon disproportionately when
its frequency within the population, a phenomenon termed it is the most common one; fish eat more of the common-colored
frequency-dependent selection. This type of selection favors insects than would occur by chance alone (figure 20.15).
certain phenotypes depending on how commonly or uncommon- Another cause of negative frequency dependence is resource
ly they occur. competition. If genotypes differ in their resource requirements, as
occurs in many plants, then the rarer geno­type will have fewer com-
Negative frequency-dependent selection petitors. When the different resource types are equally abundant,
In negative frequency-dependent selection, rare phenotypes are the rarer genotype will be at an advantage relative to the more
favored by selection—thus, fitness has a negative relationship with common genotype.

430 part IV Evolution

 change in frequencies themselves that leads to the changes in fit-
ness of the different phenotypes.

In some cases, heterozygotes may exhibit

greater fitness than homozygotes
If heterozygotes are favored over homozygotes, then natural selection
actually tends to maintain variation in the population. This
heterozygote advantage favors individuals with copies of both al-
leles, and thus works to maintain both alleles in the population. Some
evolutionary biologists think that heterozygote advantage is pervasive
and can explain the high levels of genetic variation observed in natural
populations. Others, however, consider it to be relatively rare.
The best-documented example of heterozygote advantage is
sickle cell anemia, a hereditary disease affecting hemoglobin in
humans. Individuals with sickle cell anemia exhibit symptoms
of severe anemia and abnormal red blood cells that are irregular
Figure 20.16 Positive frequency-dependent selection. in shape, with a great number of long, sickle-shaped cells
In some cases, rare individuals stand out from the rest and draw the (­figure 20.17). Chapter 13 discusses why the sickle cell mutation
attention of predators; thus, in these cases, common phenotypes have (S) causes red blood cells to sickle.
the advantage (positive frequency-dependent selection). The average incidence of the S allele in central African popu-
lations is about 0.12, far higher than that found among Americans
of African descent. From the Hardy–Weinberg principle, you can
Positive frequency-dependent selection
Positive frequency-dependent selection has the opposite effect; by
favoring common forms, it tends to eliminate variation from a pop-
ulation. For example, predators don’t always select common indi-
viduals. In some cases, “oddballs” stand out from the rest and
attract attention (figure 20.16).
The strength of selection should change through time as a
result of frequency-dependent selection. In negative frequency-­
dependent selection, rare genotypes should become increasingly
common, and their selective advantage will decrease correspond-
Normal red
ingly. Conversely, in positive frequency dependence, the rarer a blood cells
geno­type becomes, the greater the chance it will be selected
against.
Sickled red
blood cells
In oscillating selection, the favored phenotype
changes as the environment changes
In some cases, selection favors one phenotype at one time and
another phenotype at another time, a phenomenon called
Sickle cell
oscillating selection. If selection repeatedly oscillates in this allele in Africa
fashion, the effect will be to maintain genetic variation in the 1–5%
5–10%
population. 10–20%
One example, discussed in chapter 21, concerns the medium Geographic
distribution of
ground finch of the Galaˊpagos Islands. In times of drought, the sup- P. falciparum
ply of small, soft seeds is depleted, but there are still enough large malaria
seeds around. Consequently, birds with big bills are favored. How-
ever, when wet conditions return, the ensuing abundance of small
seeds favors birds with smaller bills.
Oscillating selection and frequency-dependent selection are Figure 20.17 Frequency of sickle cell allele and
similar because in both cases the form of selection changes through distribution of Plasmodium falciparum malaria. The red
time. But it is important to recognize that they are not the same: in blood cells of people homozygous for the sickle cell allele collapse
oscillating selection, the fitness of a phenotype does not depend on into sickled shapes when the oxygen level in the blood is low. The
its frequency; rather, environmental changes lead to the oscillation distribution of the sickle cell allele in Africa coincides closely with
in selection. In contrast, in ­frequency-dependent selection, it is the that of P. falciparum malaria.

chapter 20 Genes Within Populations 431

calculate that 1 in 5 central African individuals is heterozygous at
the S allele, and 1 in 100 is homozygous and develops the fatal form 20.7 Selection Acting on Traits
of the disorder. People who are homozygous for the sickle cell al-
lele almost never reproduce because they usually die before they Affected by Multiple Genes
reach reproductive age.
Why, then, is the S allele not eliminated from the central
African population by selection rather than being maintained at such
Learning Outcomes
high levels? As it turns out, one of the leading causes of illness and
death in central Africa, especially among young children, is malaria. 1. Define and contrast disruptive, directional, and stabilizing
selection.
People who are heterozygous for the sickle cell allele (and thus do not
2. Explain the evolutionary outcome of each of these types of
suffer from sickle cell anemia) are much less susceptible to malaria.
selection.
The reason is that when the parasite that causes malaria, Plasmodium
falciparum, enters a red blood cell, it causes extremely low oxygen
tension in the cell, which leads to sickling in cells of individuals either
homozygous or heterozygous for the sickle cell allele (but not in In nature, many traits—perhaps most—are affected by more than
individuals that do not have the sickle cell allele). Such cells are one gene. The interactions between genes are typically complex, as
quickly filtered out of the bloodstream by the spleen, thus eliminating you saw in chapter 12. For example, alleles of hundreds of genes
the parasite (the spleen’s filtering effect is what leads to anemia in play a role in determining human height (see figure 12.10). In such
persons homozygous for the sickle cell allele because large numbers cases, selection operates on all the genes, influencing most strong-
of red blood cells become sickle-shaped and are removed; in the ly those that make the greatest contribution to the phenotype. Such
case of heterozygotes, only those cells containing the Plasmodium traits exhibit a continuous distribution of phenotypes, often con-
parasite sickle, whereas the remaining cells are not affected, and thus forming to a normal (“bell”) curve, rather than to several discretely
anemia does not occur). Thus, homozygotes with the sickle cell allele different phenotypes (such as “red” and “black”).
all die of anemia, homozygotes for the non-sickle cell allele are
vulnerable to malaria, and heterozygotes have the highest fitness. Disruptive selection removes intermediates
Consequently, even though most homozygous recessive
individuals die at a young age, the sickle cell allele is maintained In some situations, selection acts to eliminate intermediate types, a
at high levels in these populations because it is associated with phenomenon called disruptive selection (figure 20.18a). A clear
resistance to malaria in heterozygotes and also, for reasons not yet example is the different beak sizes of the ­African black-­bellied
fully understood, with increased fertility in female heterozygotes. seedcracker finch (figure 20.19). Populations of these birds con-
Figure 20.17 shows the overlap between regions where sickle cell tain individuals with large and small beaks, but very few individu-
anemia is found and those where malaria is prevalent. als with intermediate-sized beaks.
For people living in areas where malaria is common, having As their name implies, these birds feed on seeds, and the
the sickle cell allele in the heterozygous condition has adaptive available seeds fall into two size categories: large and small. Only
value. However, in the United States, the environment does not large-beaked birds can open the tough shells of large seeds, where-
place a premium on resistance to malaria because the disease is as birds with the smaller beaks are more adept at handling small
now essentially absent from North America. Consequently, no seeds. Birds with intermediate-sized beaks are at a disadvantage
adaptive value counterbalances the ill effects of the disease; in this with both seed types—they are unable to open large seeds and too
nonmalarial environment, selection is acting to eliminate the S al- clumsy to efficiently process small seeds. Consequently, selection
lele. Only 1 in 375 Americans of African descent, many of whose acts to eliminate the intermediate phenotypes, in effect partition-
ancestors lived in central Africa, develops sickle cell anemia, far ing (or “disrupting”) the population into two phenotypically dis-
fewer than in central Africa. tinct groups.

Directional selection eliminates phenotypes

Learning Outcomes Review 20.6 at one end of a range
Selection can maintain variation within populations in a number
of ways. Negative frequency-dependent selection tends to When selection acts to eliminate one extreme from an array of
favor rare phenotypes. Oscillating selection favors different phenotypes, the genes promoting this extreme become less fre-
phenotypes at different times. In some cases, heterozygotes quent in the population and may eventually disappear. This
have a selective advantage that may act to retain alleles that are form of selection is called directional selection (see
deleterious in the homozygous state. ­f igure 20.18b). Thus, in the Drosophila population illustrated
■■ How would genetic variation in a population change if in ­f igure 20.20, eliminating flies that move toward light causes
heterozygotes had the lowest fitness? the population over time to contain fewer individuals with al-
■■ Explain the difference between negative frequency- leles promoting such behavior. If you were to pick an individual
dependent and oscillating selection. Over long periods at random from a later generation of flies, there is a smaller
of time, which is more likely to maintain variation in a chance that the fly would spontaneously move toward light than
population? if you had selected a fly from the original population. Artificial
selection has changed the population in the direction of being

432 part IV Evolution

Number of Individuals

0 25 50 75 100 125 0 25 50 75 100 125 0 25 50 75 100 125

Body Size (g) Body Size (g) Body Size (g)

Selection for small and large individuals Selection for larger individuals Selection for mid-size individuals

Two peaks Peak shifts Distribution

Number of Individuals

form gets narrower

0 25 50 75 100 125 0 25 50 75 100 125 0 25 50 75 100 125

Body Size (g) Body Size (g) Body Size (g)

a. Disruptive selection b. Directional selection c. Stabilizing selection

Figure 20.18 Three kinds of selection. The top panels show the populations before selection has occurred (under the solid red line).
Within the population, those favored by selection are shown in light brown. The bottom panels indicate what the populations would look like in the
next generation. The dashed red lines are the distribution of the original population and the solid, dark brown lines are the true distribution of the
population in the next generation. a. In disruptive selection, individuals in the middle of the range of phenotypes of a certain trait are selected
against, and the extreme forms of the trait are favored. b. In directional selection, individuals concentrated toward one extreme of the array of
phenotypes are favored. c. In stabilizing selection, individuals with midrange phenotypes are favored, with selection acting against both ends of the
range of phenotypes.

SCIENTIFIC THINKING

Question: Does disruptive selection promote differences in beak size Light 11

Average Tendency to Fly Toward Light

in the African black-bellied seedcracker finches (Pyrenestes ostrinus)? 10

9
8
7
6
5
4
3
2
Dark 1
0 2 4 6 8 10 12 14 16 18 20
Field Study: Capture, measure, and release birds in a population. Number of Generations
Follow the birds through time to determine how long each lives.
Result: Large- and small-beaked birds have higher survival rates Figure 20.20 Directional selection for negative
than birds with intermediate-sized beaks. phototropism in Drosophila. Flies that moved toward light were
Interpretation: What would happen if the distribution of seed size discarded, and only flies that moved away from light were used as
and hardness in the environment changed? parents for the next generation. This procedure was repeated for
20 generations, producing substantial evolutionary change.

Figure 20.19 Disruptive selection for large and small

beaks. Differences in beak size in the black-bellied seedcracker ? Inquiry question What would happen if after 20
generations, experimenters started keeping flies that moved
finch of west Africa are the result of disruptive selection. toward the light and discarded the others?

chapter 20 Genes Within Populations 433

less attracted to light. Directional selection often occurs in na-
ture when the environment changes, favoring traits at one end of Learning Outcomes Review 20.7
the phenotypic distribution. In disruptive selection, intermediate forms of a trait diminish;
in stabilizing selection, intermediates increase, whereas in
Stabilizing selection favors individuals disruptive selection they decrease. Directional selection shifts
frequencies toward one end or the other and may eventually
with intermediate phenotypes eliminate alleles entirely.
When selection acts to eliminate both extremes from an array of ■■ How does directional selection differ from frequency-

phenotypes, the result is to increase the frequency of the already dependent selection?
common intermediate type. This form of selection is called
stabilizing selection (see figure 20.18c). In effect, selection is
operating to prevent change away from this middle range of values.
Selection does not change the most common phenotype of the
population, but rather makes it even more common by eliminating
extremes. Many examples are known. In humans, infants with 20.8 Experimental Studies
intermediate weight at birth have the highest survival rate
(figure 20.21). In ducks and chickens, eggs of intermediate weight
of Natural Selection
have the highest hatching success.

Learning Outcome
1. Explain how experiments can be used to test evolutionary
hypotheses.
births in population
infant mortality

20 100
To study evolution, biologists have traditionally investigated what
has happened in the past, sometimes many millions of years ago.
70
To learn about dinosaurs, a paleontologist looks at dinosaur fos-
sils. To study human evolution, an anthropologist looks at human
Percent of Births in Population

15 50
Percent Infant Mortality

fossils and, increasingly, examines the “family tree” of mutations

30
that have accumulated in human DNA over millions of years. In
20 this traditional approach, evolutionary biology is similar to astron-
omy and history, relying on observation rather than experimenta-
10 10
tion to examine ideas about past events.
Nonetheless, evolutionary biology is not entirely an observa-
7 tional science. Darwin was right about many things, but one area
5 5 in which he was mistaken concerns the pace at which evolution
occurs. Darwin thought that evolution occurred at a very slow, al-
3 most imperceptible pace. But in recent years many case studies
2
have demonstrated that in some circumstances, evolutionary
change can occur rapidly. Consequently, experimental studies can
2 3 4 5 6 7 8 9 10 be devised to test evolutionary hypotheses.
Birth Weight in Pounds Although laboratory studies on fruit flies and other organ-
isms have been common for nearly a century, scientists have only
Figure 20.21 Stabilizing selection for birth weight in recently started conducting experimental studies of evolution in
humans. The death rate among babies (red curve; right y-axis) is nature. One excellent example of how observations of the natural
lowest at an intermediate birth weight; both smaller and larger babies
world can be combined with rigorous experiments in the lab and in
have a greater tendency to die than those around the most frequent
the field concerns research on guppies.
weight (tan area; left y-axis) of between 7 and 8 pounds. Recent medical
advances have reduced mortality rates for small and large babies. Guppy color variation in different environments
suggests natural selection at work
Inquiry question As improved medical technology leads
? to decreased infant mortality rates, how would you expect the
The guppy is a popular aquarium fish because of its bright color-
ation and prolific reproduction. In nature, guppies are found in
distribution of birth weights in the population to change?
small streams in northeastern South America and in many moun-
Data analysis Sketch what a graph would look like that tain streams on the nearby island of Trinidad. One interesting
showed the absolute number of infant mortality deaths versus feature of several of the streams is that they have waterfalls.
birth weight. Amazingly, guppies and some other fish are capable of coloniz-
ing portions of the stream above the waterfall.

434 part IV Evolution

 The killifish is a particularly­good colonizer; apparently on
rainy nights, it will wriggle out of the stream and move through the
damp leaf litter. Guppies are not so proficient, but they are good at
Guppy
swimming upstream. During flood seasons, rivers sometimes over- Killifish
(Poecilia reticulata)
flow their banks, creating secondary channels that move through the (Rivulus hartii )
forest. On these occasions, guppies may be able to swim upstream in
the secondary channels and invade the pools above waterfalls.
By contrast, some species are not capable of such dispersal
and thus are found only in streams below the first waterfall. One spe-
cies whose distribution is restricted by waterfalls is the pike cichlid,
a voracious predator that feeds on other fish, including guppies.
Because of these barriers to dispersal, guppies can be found in
two very different environments. In pools just below the waterfalls,
predation by the pike cichlid is a substantial risk, and rates of survival
are relatively low. But in similar pools just above the waterfall, the
only predator present is the killifish, which rarely preys on guppies.
Guppy populations above and below waterfalls exhibit
many differences. In the ­high-predation pools, guppies exhibit
drab coloration. Moreover, they tend to reproduce at a younger age
and attain relatively smaller adult sizes. Male fish above the
waterfall, in contrast, are colorful (figure 20.22), ma-
ture later, and grow to larger sizes.
These differences suggest the operation of
natural selection. In the low-predation environ-
ment, males display gaudy colors and spots that
they use to court females. Moreover, larger males Pike cichlid Guppy
(Crenicichla alta) (Poecilia reticulata)
are most successful at holding territories and mating
with females, and larger females lay more eggs. Thus, in the ab- Figure 20.22 The evolution of protective coloration in
sence of predators, larger and more colorful fish may have pro- guppies. In pools below waterfalls where predation is high, male
duced more offspring, leading to the evolution of those traits. guppies are small and drab in color. In the absence of the highly
In pools below the waterfall, natural selection would favor predatory pike cichlid in pools above waterfalls, male guppies are
different traits. Colorful males are likely to attract the attention of larger and much more colorful and attractive to females. The killifish
the pike cichlid, and high predation rates mean that most fish live is also a predator, but it only rarely eats guppies. The evolution of
short lives. Individuals that are more drab and shunt energy into these differences in guppies can be experimentally tested.
early reproduction, rather than growth to a larger size, are therefore
likely to be favored by natural selection. These results established that predation can lead to rapid­
evolutionary change, but do these laboratory experiments reflect
Experimentation reveals the agent of selection what occurs in nature?
Although the differences between guppies living above and below
The field experiment
the waterfalls suggest evolutionary responses to differences in the
strength of predation, alternative explanations are possible. Perhaps, To find out whether the laboratory results were an accurate reflec-
for example, only very large fish are capable of crawling past the tion of natural processes, the scientists located two streams that
waterfall to colonize pools. If this were the case, then a founder ef- had guppies in pools below a waterfa ll, but not above it. As in
fect would occur in which the new population was established solely other Trinidadian streams, the pike cichlid was present in the lower
by individuals with genes for large size. The only way to rule out pools, but only the killifish was found above the waterfalls.
such alternative possibilities is to conduct a controlled experiment. The scientists then transplanted guppies to the upper pools and
returned at several-year intervals to monitor the populations. Despite
The laboratory experiment originating from populations in which predation levels were high, the
The first experiments were conducted in large pools in laboratory transplanted populations rapidly evolved the traits characteristic of
greenhouses. At the start of the experiment, a group of 2000 guppies low-predation guppies: they matured late, attained greater size, and
was divided equally among 10 large pools. Four weeks later, pike had brighter colors. The control populations in the lower pools, by
cichlids were added to four of the pools and killifish to another four, contrast, continued to be drab and to mature early and at a smaller
with the remaining two pools left to serve as “no-predation” controls. size. Laboratory analysis confirmed that the variations between the
Fourteen months later (which corresponds to 10 guppy gen- populations were the result of genetic differences.
erations), the scientists compared the populations. The guppies in These results demonstrate that substantial evolutionary
the killifish and control pools were indistinguishable—brightly change can occur in just a few years. More generally, these studies
colored and large. In contrast, the guppies in the pike cichlid pools indicate how scientists can formulate hypotheses about how evolu-
were smaller and drab in coloration (figure 20.23). tion occurs and then test these hypotheses in natural conditions.

chapter 20 Genes Within Populations 435

 SCIENTIFIC THINKING

Question: Does the presence of predators affect the evolution of 20.9 Interactions Among
guppy color?
Hypothesis: Predation on the most colorful individuals will cause a
Evolutionary Forces
population to become increasingly dull through time. Conversely, in
populations with few or no predators, increased color will evolve.
Experiment: Establish laboratory populations of guppies in large Learning Outcomes
pools with or without predators. 1. Discuss how evolutionary processes can work simultaneously,
Result: The populations with predators evolved to have fewer spots, but in opposing ways.
while the populations in pools without predators evolved more spots. 2. Evaluate what determines the evolutionary outcome when
multiple processes are operating simultaneously.

no predation
low predation
14 high predation The amount of genetic variation in a population may be deter-
13 mined by the relative strength of different evolutionary processes.
Sometimes these processes act together, and in other cases they
12 work in opposition.
Spots per Fish

11

10
Mutation and genetic drift may counter
selection
9
In theory, if allele B mutated to allele b at a high enough rate,
8 allele b could be maintained in the population, even if natural
selection strongly favored allele B. In nature, however, mutation
0 4 8 12
rates are rarely high enough to counter the effects of natural
Duration of Experiment (months) selection.
The effect of natural selection also may be countered by
genetic drift. Both of these processes may act to remove variation
Interpretation: Why does color increase in the absence of predators?
from a population. But selection is a nonrandom process that op-
How would you test your hypothesis?
erates to increase the representation of alleles that enhance sur-
vival and reproductive success, whereas genetic drift is a random
Figure 20.23 Evolutionary change in spot number. process in which any allele may increase. Thus, in some cases,
Guppy populations raised for 10 generations in low-predation or drift may lead to a decrease in the frequency of an allele that is
no-predation environments in laboratory greenhouses evolved a favored by selection. In some extreme cases, drift may even lead
greater number of spots, whereas selection in more dangerous to the loss of a favored allele from a population.
environments, such as the pools with the highly predatory pike Remember, however, that the magnitude of drift is inversely
cichlid, led to less conspicuous fish. The same results are seen in field related to population size; consequently, natural selection is
experiments conducted in pools above and below waterfalls. expected­to overwhelm drift, except when populations are very
small.
Inquiry question How do these results depend
? on the manner by which the guppy predators locate their prey?
Gene flow may promote or constrain
evolutionary change
The results give strong support to the theory of evolution by natural
selection. Gene flow can be either a constructive or a constraining force. On
one hand, gene flow can spread a beneficial mutation that arises in
one population to other populations. On the other hand, gene flow
Learning Outcomes Review 20.8 can impede adaptation within a population by the continual flow
Although much of evolutionary theory is derived from of inferior alleles from other populations.
observation, experiments are sometimes possible in natural Consider two populations of a species that live in different
settings. Studies have revealed that traits can shift in populations environments. In this situation, natural selection might favor dif-
in a relatively short time. The data obtained from evolutionary ferent alleles—B and b—in the two populations. In the absence
experiments can be used to refine theoretical assumptions. of other evolutionary processes such as gene flow, the frequency
■■ What experiments could you design to test other examples of B would be expected to reach 100% in one population and 0%
of natural selection, such as the evolution of pesticide in the other. However, if gene flow occurred between the two
resistance or background color matching? populations, then the less favored allele would continually be re-
introduced into each population. As a result, the frequency of the

436 part IV Evolution

alleles in the populations would reflect a balance between the
rate at which gene flow brings the inferior allele into a popula- Pollen

Index of Copper Tolerance

tion, and the rate at which natural selection removes it. Prevailing wind
A classic example of gene flow opposing natural selection Non- Mine site Nonmine
60 mine
occurs on abandoned mine sites in Great Britain. Although mining
activities ceased hundreds of years ago, the concentration of metal 40 Bent grass
ions in the soil is still much greater than in surrounding areas. Large (Agrostis tenuis)
concentrations of heavy metals are generally toxic to plants, but al- 20
leles of certain genes confer the ability to grow on soils high in heavy 0
metals. The ability to tolerate heavy metals comes at a price, how- 20 0 0 20 40 60 80 100 120 140 160
ever; individuals with the resistance allele exhibit lower growth rates Distance Upwind Distance Away
on nonpolluted soil. Consequently, we would expect the resistance from Mine (m) from Mine (m)
allele to occur with a frequency of 100% on former mine sites and
0% elsewhere. Figure 20.24 Degree of copper tolerance in grass plants
Heavy-metal tolerance has been studied intensively in the on and near ancient mine sites. Individuals with tolerant alleles
slender bent grass, in which the resistance allele occurs at interme- have decreased growth rates on unpolluted soil. Thus, we would expect
diate levels in many areas (figure 20.24). The explanation relates copper tolerance to be 100% on mine sites and 0% on nonmine sites.
to the reproductive system of this grass, in which pollen, the floral However, prevailing winds blow pollen-containing nontolerant alleles
equivalent of sperm, is dispersed by the wind. As a result, pollen onto the mine site and tolerant alleles beyond the site’s borders. The
grains—and the alleles they c­arry—can move great distances, amount of pollen received decreases with distance, which explains the
leading to levels of gene flow between mine sites and unpolluted changes in levels of tolerance. The index of copper tolerance is calculated
areas high enough to counteract the effects of natural selection. as the growth rate of a plant on soil with high concentrations of copper
In general, the extent to which gene flow can hinder the relative to growth rate on soils with low levels of copper; the higher the
effects of natural selection should depend on the relative strengths index, the more tolerant the plant is of heavy metal pollution.
of the two processes. In species in which gene flow is generally
strong, such as in birds and wind-pollinated plants, the frequency Data analysis Examine the index of copper tolerance
of the allele less favored by natural selection may be relatively on nonmine areas. What does it suggest about the factor
responsible? In particular, how does the index change as a
high. In more sedentary species that exhibit low levels of gene
function of distance from the mine on the right-hand side? And
flow, such as snails, the favored allele should occur at a frequency how does the relationship between distance and index value
near 100%. differ on the two sides of the mine? What process might be
responsible for such patterns?

Learning Outcomes Review 20.9

Allele frequencies sometimes reflect a balance between
opposing processes. Gene flow, for example, may increase some
alleles while natural selection decreases them. Where several limits result from multiple phenotypic effects of alleles, lack of ge-
processes are involved, observed frequencies depend on the netic variation upon which selection can act, and interactions be-
relative strength of the processes. tween genes.
■■ Under what circumstances might evolutionary processes
operate in the same direction, and what would be the
outcome? Genes have multiple effects
Alleles often affect multiple aspects of a phenotype (the phenom-
enon of pleiotropy; see chapter 12). These multiple effects tend
to set limits on how much a phenotype can be altered.
For example, in chickens, the same gene affects the size of a

The Limits of Selection

hen’s comb and the rate at which she lays—alleles that produce
20.10 large combs also lead to faster egg production. As a result of this
linkage, selection for hens that lay many eggs, but have a small
comb, would be very difficult.
Learning Outcomes
1. Define pleiotropy and epistasis. Evolution requires genetic variation
2. Explain how these phenomena may affect the evolutionary
response to selective pressure. Genetic variation is a prerequisite for evolutionary change; without
it, natural selection cannot produce evolution.
Over 80% of the gene pool of the thoroughbred horses racing
Although selection is the most powerful of the principal agents of today goes back to 31 ancestors from the late 18th century. Despite
genetic change, there are limits to what it can accomplish. These intense directional selection on thoroughbreds, their performance

chapter 20 Genes Within Populations 437

 170

160

150
Finish time (seconds)

140

130

120

110
1880 1900 1920 1940 1960 1980 2000 2020

Figure 20.25 Selection for increased speed in racehorses is no longer effective. Kentucky Derby winning speeds have not
improved significantly since 1950.

? Inquiry question What might explain the lack of change in winning speeds?

times have not improved for more than 60 years (figure 20.25). De-
cades of intense selection presumably have removed variation from
the population at a rate greater than mutation can replenish it, such that
little genetic variation now remains, and evolutionary change is not
possible.
In some cases, phenotypic variation for a trait may never Right eye Left eye
of insect of insect
have had a genetic basis. The compound eyes of insects are made
up of hundreds of visual units, termed ommatidia (described in
Ommatidia
chapter 33). In some individuals, the left eye contains more om-
matidia than the right. In other individuals, the right eye contains
more than the left (figure 20.26). However, despite intense se-
lection experiments in the laboratory, scientists have never been
able to produce a line of fruit flies that consistently has more
Figure 20.26 Phenotypic variation in insect ommatidia.
ommatidia in the left eye than in the right.
In some individuals, the number of ommatidia in the left eye is
The reason is that separate genes do not exist for the left
greater than the number in the right.
and right eyes. Rather, the same genes affect both eyes, and dif-
ferences in the number of ommatidia result from differences
that occur as the eyes are formed in the development process. advantage of an allele at one gene may vary from one genotype to
Thus, despite the existence of phenotypic variation, no underly- another. If a population is polymorphic for a second gene, then
ing genetic variation is available for selection to favor. selection on the first gene may be constrained because different
alleles are favored in different individuals of the same population.
Studies on bacteria illustrate how selection on alleles for
Gene interactions affect fitness of alleles one gene can depend on which alleles are present at other genes.
As discussed in chapter 12, epistasis is the phenomenon in which In E. coli, two biochemical pathways exist to break down gluco-
an allele for one gene may have different effects, depending on al- nate, each using enzymes produced by different genes. One gene
leles present at other genes. Because of epistasis, the selective produces the enzyme 6-PGD, for which there are several alleles.

438 part IV Evolution

When the common allele for the second gene, which codes for
the other biochemical pathway, is present, selection does not fa- Learning Outcomes Review 20.10
vor one allele over another at the 6-PGD gene. In some E. coli, In pleiotropy, a single gene affects multiple traits; in epistasis,
however, an alternative allele at the second gene occurs that is interaction between alleles of different genes affects a single trait.
not functional. The bacteria with this alternative allele are forced Both these conditions can constrain the effects of natural selection.
to rely only on the 6-PGD pathway, and in this case, selection ■■ How can epistasis and pleiotropy constrain the evolutionary
favors one 6-PGD allele over another. Thus, epistatic interactions response to natural selection?
exist between the two genes, and the outcome of natural selection
on the 6-PGD gene depends on which alleles are present at the
second gene.

Chapter Review

tamoncity/Shutterstock

20.1 Genetic Variation and Evolution Genetic drift may alter allele frequencies in small populations.
Genetic drift refers to random shifts in allele frequency. Its effects may
Many processes can lead to evolutionary change. be severe in small populations.
Darwin proposed that evolution of species occurs by the process
of natural selection. Other processes can also lead to evolutionary Selection favors some phenotypes over others.
change. For evolution by natural selection to occur, genetic variation must
exist, it must result in differential reproductive success, and it must be
Populations contain ample genetic variation. inheritable.
For a population to be able to evolve, it must contain genetic variation.
DNA testing shows that natural populations generally have substantial 20.4 Quantifying Natural Selection
variation.
A phenotype with greater fitness usually increases in frequency.
20.2 Changes in Allele Frequency (figure 20.3) Fitness is defined as the reproductive success of an individual. Relative
fitness refers to the success of one phenotype relative to others in a
The Hardy–Weinberg principle allows prediction of genotype population. Usually, the phenotype with highest relative fitness increases
frequencies. in frequency in the next generation, assuming that phenotypic differences
Hardy–Weinberg equilibrium exists when observed genotype frequencies are the result of genetic differences.
match the prediction from calculated frequencies. It occurs only when Fitness may consist of many components.
evolutionary processes are not acting to shift the distribution of alleles or
genotypes in the population. Reproductive success is determined by how long an individual survives,
how often it mates, and how many offspring it has per reproductive event.
Hardy–Weinberg predictions can be applied to data to find
evidence of evolutionary processes. 20.5 Reproductive Strategies
If genotype frequencies are not in Hardy–Weinberg equilibrium, then
evolutionary processes must be at work. The sexes often have different reproductive strategies.
One sex may be choosier than the other, and which one often depends on
the degree of parental investment.
20.3 Five Agents of Evolutionary Change (figure 20.4)
Sexual selection occurs through mate competition and mate
Mutation changes alleles. choice.
Mutations are the ultimate source of genetic variation. Because mutation Intrasexual selection involves competition among members of the same sex
rates are low, mutation usually is not responsible for deviations from for the chance to mate. Intersexual selection is one sex choosing a mate.
Hardy–Weinberg equilibrium.
Mate choice may provide direct benefits (increased resource availability
Gene flow occurs when alleles move between populations. or parental care) or indirect benefits (genetic quality of the mate).
Gene flow is the migration of new alleles into a population. It can
introduce genetic variation and can homogenize allele frequencies 20.6 Natural Selection’s Role in Maintaining Variation
between populations.
Frequency-dependent selection may favor either rare or
Nonrandom mating shifts genotype frequencies. common phenotypes.
Assortative mating, in which similar individuals tend to mate, increases Negative frequency-dependent selection favors rare phenotypes and
homozygosity; disassortative mating increases the frequency of maintains variation within a population. Positive frequency-dependent
heterozygotes. selection favors the common phenotype and leads to decreased variation.

chapter 20 Genes Within Populations 439

In oscillating selection, the favored phenotype changes as the Experimentation reveals the agent of selection.
environment changes. Guppies in natural populations subject to different predators were shown
If environmental change is cyclical, selection would favor first one to undergo color change over generations.
phenotype, then another, maintaining variation.
In some cases, heterozygotes may exhibit greater fitness
20.9 Interactions Among Evolutionary Forces
than homozygotes. Mutation and genetic drift may counter selection.
Heterozygote advantage favors individuals with both alleles. In theory, a high rate of mutation could oppose natural selection, but this
rarely happens. Genetic drift also can work counter to natural selection.
20.7 Selection Acting on Traits Affected by Multiple Genes
Gene flow may promote or constrain evolutionary change.
Disruptive selection removes intermediates.
Gene flow can spread a beneficial mutation to other populations, but it
When intermediate phenotypes are at a disadvantage, a population may
can also impede adaptation due to influx of alleles with low fitness in a
exhibit a bimodal trait distribution.
population’s environment.
Directional selection eliminates phenotypes at one end of a range.
Directional selection tends to shift the mean value of the population 20.10 The Limits of Selection
toward the favored end of the distribution. Genes have multiple effects.
Stabilizing selection favors individuals with intermediate phenotypes. Pleiotropic genes, which have multiple effects, set limits on how much
Stabilizing selection eliminates both extremes and increases the a phenotype can be altered. Even if one affected trait is favored, other
frequency of an intermediate type. The population may have the same affected traits may not be.
mean value, but with decreased variation. Evolution requires genetic variation.
Intense selection pressure may remove genetic variation.
20.8 Experimental Studies of Natural Selection
Gene interactions affect fitness of alleles.
The hypothesis that natural selection leads to evolutionary change can be
tested experimentally. In epistasis, fitness of one allele may vary depending on the genotype of
a second gene.
Guppy color variation in different environments suggests natural
selection at work.
In the presence of predators, males are much less colorful.

Visual Summary
Single genes
Genetic variation in Phenotypic
produces affected by
populations variation
Multiple genes
changes leading to

Pleiotropy Deviations from Hardy-

Weinberg equilibrium
Factors that
Epistasis indicating action of
limit evolution

Lack of Processes that cause

constrained by
genetic evolutionary change
variation
include

Mutation Gene flow Nonrandom mating Genetic drift Natural selection

occurs when phenotypes differ in

Fitness

differences result from

Number of offspring per

Survival Mating success
mating event

determined by

Reproductive strategies

440 part IV Evolution

 Review Questions

tamoncity/Shutterstock
U N D E R S TA N D A P P LY
1. Assortative mating 1. In a population of red (dominant allele) or white flowers
a. affects genotype frequencies expected under Hardy– in Hardy–Weinberg equilibrium, the frequency of red flowers
Weinberg equilibrium. is 91%. What is the frequency of the red allele?
b. affects allele frequencies expected under Hardy–Weinberg a. 9% c. 91%
equilibrium. b. 30% d. 70%
c. has no effect on the genotypic frequencies expected under 2. Genetic drift and natural selection can both lead to rapid rates of
Hardy–Weinberg equilibrium because it does not affect the evolution. However,
relative proportion of alleles in a population.
d. increases the frequency of heterozygous individuals above a. genetic drift works fastest in large populations.
Hardy–Weinberg expectations. b. only drift leads to adaptation.
c. natural selection requires genetic drift to produce new
2. When the environment changes from year to year and different variation in populations.
phenotypes have different fitness in different environments, d. both processes of evolution can be slowed by gene flow.
a. natural selection will operate in a frequency-dependent manner. 3. Suppose that the relationship between birth weight and infant
b. the effect of natural selection may oscillate from year to year, mortality, instead of being at a minimum at intermediate sizes,
favoring alternative phenotypes in different years. changed such that babies born at 5 or 10 pounds had the lowest
c. genetic variation is not required to get evolutionary change by mortality, with an increase in between such that 7.5-pound babies
natural selection. had higher mortality. How would you expect the distribution of
d. None of the choices is correct. birth weights to change over time?
3. Many factors can limit the ability of natural selection to cause a. It would not change.
evolutionary change, including b. The distribution would shift to the right.
a. a conflict between traits favored by reproduction and survival. c. The distribution would become bimodal, with two peaks and
b. lack of genetic variation. the mean value unchanged.
c. pleiotropy. d. The distribution would become bimodal, with two peaks and
d. All of the choices are correct. the mean value shifted to the right.
4. Stabilizing selection differs from directional selection because
a. in the former, phenotypic variation is reduced but the average
phenotype stays the same, whereas in the latter both the births in population
infant mortality
variation and the mean phenotype change.
20 100
b. the former requires genetic variation, but the latter does not.
c. intermediate phenotypes are favored in directional selection. 70

d. None of the choices is correct.

Percent of Births in Population

15 50

Percent Infant Mortality

5. Founder effects and bottlenecks are 30
20
a. expected only in large populations.
b. mechanisms that increase genetic variation in a population. 10 10
c. two different modes of natural selection.
7
d. forms of genetic drift.
5 5
6. Relative fitness
3
a. refers to the survival rate of one phenotype compared to that 2
of another.
b. is the physical condition of an individual’s siblings and cousins. 2 3 4 5 6 7 8 9 10
Birth Weight in Pounds
c. refers to the reproductive success of a phenotype.
d. None of the choices is correct.
7. For natural selection to result in evolutionary change,
a. variation must exist in a population. 4. Peahens prefer to mate with peacocks that have more eyespots in
b. reproductive success of different phenotypes must differ. their tail feathers (that is, longer tail feathers). It has also been
c. variation must be inherited from one generation to suggested that the longer the tail feathers, the more impaired the
the next. flight of the males. One possible hypothesis to explain such a
d. All of the choices are correct. preference by females is that the males with the longest tail
8. The elaborate tail feathers of a male peacock evolved because they feathers experience the most severe handicap, and if they can
a. improve reproductive success of males and females. nevertheless survive, it reflects their “vigor.” Suggest some
b. improve male survival. studies that would allow you to test this idea. Your description
c. reduce survival. should include the kinds of traits that you would measure and
d. None of the choices is correct. why. (Refer to figure 20.10.)

chapter 20 Genes Within Populations 441

SYNTHESIZE 3. Based on a consideration of how strong artificial selection has
1. In Trinidadian guppies a combination of elegant laboratory and helped eliminate genetic variation for speed in thoroughbred
field experiments builds a very compelling case for predator- horses, we are left with the question of why, for many traits
induced evolutionary changes in color and life history traits. It is like speed (continuous traits), there is usually abundant genetic
still possible, although not likely, that there are other differences variation. This is true even for traits we know are under strong
between the sites above and below the falls aside from whether selection. Where does genetic variation ultimately come from,
predators are present. What additional studies could strengthen the and how does the rate of production compare with the strength
interpretation of the results? of natural selection? What other mechanisms can maintain and
increase genetic variation in natural populations?
2. On large black lava flows in the deserts of the southwestern
United States, populations of many types of animals are
composed primarily of black individuals. By contrast, on small
lava flows, populations often have a relatively high proportion
of light-colored individuals. How can you explain this
difference?

442 part IV Evolution

 21
CHAPTER

The Evidence for Evolution

Chapter Contents
21.1 The Beaks of Darwin’s Finches: Evidence
of Natural Selection
21.2 Peppered Moths and Industrial Melanism:
More Evidence of Selection
21.3 Artificial Selection: Human-Initiated Change
21.4 Fossil Evidence of Evolution
21.5 Anatomical Evidence for Evolution
21.6 Convergent Evolution and the
Biogeographical Record
21.7 Darwin’s Critics

Georgia Southern University Museum

Visual Outline Introduction

As we discussed in chapter 1, when Darwin
Evidence for evolution proposed his revolutionary theory of evolu-
tion by natural selection, little actual evi-
includes dence existed to bolster his case. Instead,
evidence from Darwin relied on observations of the natural
world, logic, and results obtained by breed-
ers working with domestic animals. Since
Convergent Biogeographical his day, however, the evidence for Darwin’s
Selection Fossils Anatomy
evolution record theory has become overwhelming.
categorized The case is built upon two pillars: first,
as
evidence that natural selection can produce
Artificial Natural
selection selection
evolutionary change, and second, evidence
from the fossil record that evolution has
occurred. The whale skeleton pictured here
is the Vogtle whale (Georgiacetus vogtlen-
Human Cat Bat
sis), which is the oldest whale fossil found
in North America (40 million years old). The
pelvic bones and hindlimbs reveal a link
between land mammals and whales. In
addition to evidence from studies of natural
Porpoise Horse selection and fossils, information from many
Jason Edwards/National Geographic/Getty Images different areas of biology—fields as different
as anatomy, molecular biology, and bioge-
ography—is only interpretable scientifically
as being the outcome of evolution.
 that his collection contained wrens, “gross-beaks,” and blackbirds,
21.1 The Beaks of Darwin’s all birds with which he was familiar from his native England.
Upon Darwin’s return home, ornithologist John Gould
Finches: Evidence of informed Darwin that the species he had collected were not closely
Natural Selection related to different English birds, but rather were a closely related
group of distinct species, all similar to one another except for their
beaks. In all, 14 species are now recognized.

Learning Outcomes
Galápagos finches exhibit variation
1. Describe how the species of Darwin’s finches have adapted to
feed in different ways. related to food gathering
2. Explain how climatic variation drives evolutionary change in The diversity of Darwin’s finches is illustrated in figure 21.1. The
the medium ground finch. ground finches feed on seeds that they crush in their powerful
beaks; species with smaller and narrower beaks, such as the
warbler finch, eat insects. Other species include fruit and bud
As you learned in chapter 20, a variety of processes can produce evo- eaters, and species that feed on cactus fruits and the insects they
lutionary change. Most evolutionary biologists, however, agree with attract; some populations of the sharp-beaked ground finch even
Darwin’s thinking that natural selection is the primary process respon- include “vampires” that sometimes creep up on seabirds and use
sible for evolution. Although we cannot travel back through time, their sharp beaks to pierce the seabirds’ skin and drink their blood.
modern-day evidence allows us to test hypotheses about how evolution Perhaps most remarkable are the tool users, woodpecker finches
proceeds and confirms the power of natural selection as an agent of that pick up a twig, cactus spine, or leaf stalk, trim it into shape with
evolutionary change. This evidence comes from both the field and the their beaks, and then poke it into dead branches to pry out grubs.
laboratory and from both natural and human-altered situations. The correspondence between the beaks of the finch species
Darwin’s finches are a classic example of evolution by natu- and their food sources suggested to Darwin that natural selection
ral selection. When he visited the Galápagos Islands off the coast had shaped them. In The Voyage of the Beagle, written several
of Ecuador in 1835, Darwin collected 31 specimens of finches years after his return to England, Darwin wrote, “Seeing this gra-
from three islands. Not being an expert on birds, Darwin had trou- dation and diversity of structure in one small, intimately related
ble identifying the specimens, believing by examining their beaks group of birds, one might really fancy that from an original paucity

Woodpecker finch (Cactospiza pallida) Large ground finch (Geospiza magnirostris) Cactus finch (Geospiza scandens)

Figure 21.1 Darwin’s finches. These species show

differences in beaks and feeding habits among Darwin’s
finches. This diversity arose when an ancestral finch
colonized the islands and diversified into habitats
lacking other types of small birds. The beaks of
several species resemble those of different families of
birds on the mainland. For example, the warbler finch has a
beak very similar to that of warblers, to which it is not closely
related.

Warbler finch (Certhidea olivacea) Vegetarian tree finch (Platyspiza crassirostris)

444 part IV Evolution

of birds in this archipelago, one species has been taken and modi- Measuring many birds every year, they were able to assemble for
fied for different ends.” the first time a detailed portrait of evolution in action. The Grants
found that not only did a great deal of variation in beak depth exist
Modern research has verified among members of the population, but the average beak depth
changed from one year to the next in a predictable fashion.
Darwin’s selection hypothesis During droughts, plants produced few seeds, and all avail-
Darwin’s observations suggest that differences among species in able small seeds were quickly eaten, leaving large seeds as the
beak size and shape have evolved as the species adapted to use major remaining source of food. As a result, birds with shorter and
different food resources, but can this hypothesis be tested? In deeper, more powerful beaks survived better, because they were
chapter 20, you read that the theory of evolution by natural selection better able to break open these large seeds (figure 21.2a). Con-
requires that three conditions be met: versely, in particularly wet years, plants flourished, producing an
abundance of small seeds; as a result, birds with long and shallow
1. Phenotypic variation must exist in the population.
beaks were favored, and beaks became pointier.
2. This variation must lead to differences among individuals
Could these changes in beak dimension be evidence of evolu-
in lifetime reproductive success.
tion by natural selection? If so, the variation among individuals in beak
3. Phenotypic variation among individuals must be
size must be genetically based. An alternative possibility might be that
genetically transmissible to the next generation.
the changes in beak depth do not reflect changes in gene frequencies,
The key to successfully testing Darwin’s proposal proved to but rather are simply a response to diet—for example, perhaps crush-
be patience. For more than 40 years, starting in 1973, Peter and ing large seeds causes a growing bird to develop a larger beak.
Rosemary Grant of Princeton University and their students studied To rule out this possibility, the Grants tested for a genetic basis
the medium ground finch on a tiny island in the center of the to beak size variation by measuring the relationship of parent beak
Galápagos called Daphne Major. These finches feed preferentially size to offspring beak size, examining many broods over several years.
on small, tender seeds, produced in abundance by plants in wet The depth of the beak was very similar between parents and offspring
years. During long periods of dry weather, the birds resort to larg- regardless of environmental conditions (figure 21.2b), suggesting that
er, drier seeds, which are harder to crush. the differences among individuals in beak size reflect genetic differ-
The Grants quantified beak shape among the medium ground ences. This result satisfies condition number three mentioned above
finches of Daphne Major by carefully measuring beak depth and therefore indicates that the year-to-year changes in average beak
(height of beak, from top to bottom, at its base) on individual birds. depth represent evolutionary change resulting from natural selection.

10.5
G. fortis
11
•
10.0
•• • ••
Beak depth of offspring (mm)

10 • • ••
Beak depth (mm)

9.5
•
• •
••
•
9.0 9 •
Drought, large seeds
become abundant
8.5
Wet years, small seeds 8
become abundant

8.0 8 9 10 11
1970 1980 1990 Mean beak depth of parents (mm)

a. b.
Figure 21.2 Evidence that natural selection alters beak shape in the medium ground finch, Geospiza fortis. a. In dry
years, when only large, tough seeds are available, the mean beak depth increases. In wet years, when many small seeds are available, mean beak
depth decreases. b. Beak depth is inherited from parents to offspring. Like many quantitative traits (see section 12.6), beak depth is probably
determined by many genes and, on average, offspring tend to have a beak depth equal to the mean of their parents’ beak depth.

Inquiry question What would the relationship in figure 21.2b look like if differences in beak shape were affected not by genetic
? differences, but rather by an environmental factor, such as what a nestling bird ate during its growth period?

Data analysis Suppose that a male with a beak depth of 10 mm mated with a female with a beak depth of 8 mm. What would the
expected beak depth of the offspring be? Would it matter if the female’s beak was 10 mm and the male’s 6 mm?

chapter 21 The Evidence for Evolution 445

 all-black individuals are the descendants of a single mutation.
Learning Outcomes Review 21.1 This dominant allele was present but very rare in populations be-
Among Darwin’s finches, natural selection has been responsible fore 1850. From that time on, dark individuals increased in fre-
for changes in the shape of the beak corresponding to quency in moth populations near industrialized centers until they
characteristics of the available food supply. Because beak made up almost 100% of these populations. Biologists also no-
morphology is a genetically transmissible trait, a beak better ticed that in industrialized regions where the dark moths were
suited to the distribution of available seed types would become common, the tree trunks were darkened almost black by the soot
more common in subsequent generations.
of pollution, which also killed many of the light-colored lichens
■■ Suppose that the act of eating hard seeds caused birds to on tree trunks.
develop bigger beaks. Would this lead to an evolutionary
increase in beak size after a drought? Light-colored moths decreased in polluted areas
Why did dark moths gain a survival advantage around 1850? In
1896, a British amateur moth collector named J. W. Tutt proposed
what became the most commonly accepted hypothesis explaining

21.2 Peppered Moths and the decline of the light-colored moths. He suggested that pep-
pered forms were more visible to predators on sooty trees that
Industrial Melanism: More have lost their lichens. Consequently, birds ate the peppered
moths resting on the trunks of trees during the day. The black
Evidence of Selection forms, in contrast, had an advantage because they were camou-
flaged (figure 21.3).
Although Tutt initially had no evidence, British ecologist Ber-
nard Kettlewell tested the hypothesis in the 1950s by releasing equal
Learning Outcomes numbers of dark and light individuals into two sets of woods: one
1. Explain the relationship between pollution and color evolution near heavily polluted Birmingham, and the other in unpolluted Dor-
in peppered moths. set. If Tutt was correct, then dark moths should have survived better
2. Distinguish between demonstrating that evolution has in the Birmingham woods and light moths in the Dorset woods. Ket-
occurred and understanding the mechanism that caused it. tlewell then set up lights in the woods to attract moths to traps to see
how many of both kinds of moths survived. To identify the moths he
had released, he had marked the released moths with a dot of paint on
When the environment changes, natural selection often may favor the underside of their wings, where birds could not see it.
a trait that previously wasn’t favored. One classic example con- In the polluted area near Birmingham, Kettlewell recaptured
cerns the peppered moth, Biston betularia. Adults come in a range only 19% of the light moths, but 40% of the dark ones. This indicated
of shades, from light gray with black speckling (hence the name that dark moths survived better in these polluted woods, where tree
“peppered” moth) to jet black (melanic). trunks were dark. In the relatively unpolluted Dorset woods, Ket-
Extensive genetic analysis has shown that the moth’s body tlewell recovered 12.5% of the light moths but only 6% of the dark
color is a genetic trait that reflects different alleles of a single ones. This result indicated that where the tree trunks were still light-­
gene. Recent molecular genetic studies have demonstrated that colored, light moths’ survival was twice as great as that of dark moths.

Figure 21.3 Tutt’s hypothesis

explaining industrial melanism.
These photographs show preserved
specimens of the peppered moth,
Biston betularia, placed on trees. Tutt
proposed that the dark melanic
variant of the moth is more visible to
predators on unpolluted trees (left),
whereas the light “peppered” moth is
more visible to predators on bark
blackened by industrial pollution
(right). (Left) IanRedding/Shutterstock;
(right) The Natural History Museum/Alamy
Stock Photo

446 part IV Evolution

 Kettlewell later solidified his argument by placing moths on
100
trees and filming birds looking for food. Sometimes the birds actually
90
passed right over a moth that was the same color as its background.

Percentage of melanic moths

80
Recently, an enormous six-year study involving the release of
nearly 5000 moths confirmed Kettlewell’s findings. Conducted in an 70
unpolluted forest, the study found that dark-colored moths disappeared 60
at a rate 10% higher than that of light-­colored moths. In addition, di- 50
rect observations of 250 feeding events revealed that dark moths were 40
captured by birds substantially more often than light-colored moths. 30
20

When environmental conditions reverse, 10

0
so does selection pressure ‘59 ‘63 ‘67 ‘71 ‘75 ‘79 ‘83 ‘87 ‘91 ‘95 ‘99 ‘03
In industrialized areas throughout Eurasia and North America, dozens Year
of other species of moths have evolved in the same way as the pep-
pered moth. The term industrial melanism refers to the phenomenon Figure 21.4 Selection against melanism. The red circles
in which darker individuals come to predominate over lighter ones. indicate the frequency of melanic Biston betularia moths at Caldy
In Great Britain, the air pollution that promoted industrial Common in Great Britain. Yellow diamonds indicate frequencies of
melanism began to reverse following enactment of the Clean Air melanic B. betularia in Michigan, and the blue squares indicate
Act in 1956. Beginning in 1959, the Biston population at Caldy corresponding frequencies in Pennsylvania.
Common outside Liverpool has been sampled each year. The fre-
quency of the melanic (dark) form has dropped from a high of 93%
Inquiry question What can you conclude from the fact
in 1959 to less than 5% in 2002 (figure 21.4).
The drop is consistent with a 15% selective disadvantage act- ? that the frequency of melanic moths decreased to the same
degree in the three locations?
ing against moths with the dominant melanic allele; it correlates
well with a significant drop in air pollution, particularly with a
Researchers supporting the bird predation hypothesis point
lowering of the levels of sulfur dioxide and suspended particulates,
out that a bird’s ability to detect moths may depend less on the
both of which act to darken trees.
presence or absence of lichens, and more on other ways in which
Interestingly, the same reversal of melanism occurred in the
the environment is darkened by industrial pollution. Pollution
United States. Of 576 peppered moths collected at a field station near
tends to cover all objects in the environment with a fine layer of
Detroit from 1959 to 1961, 515 were melanic, a frequency of 89%. The
particulate dust, which tends to decrease how much light is re-
American Clean Air Act, passed in 1963, led to significant reductions
flected by the surface. In addition, pollution has a particularly se-
in air pollution. Resampled in 2001, the frequency of melanics at the
vere effect on birch trees, which are light in color. Both effects
Detroit field station also had dropped to less than 5%; a similar decreas-
would tend to make the environment appear darker, and thus would
ing trend was also observed in Pennsylvania (figure 21.4). The moth
favor darker moths by protecting them from predation by birds.
populations in Liverpool, Detroit, and Pensylvania, all part of the same
Despite this uncertainty over the agent of selection, the
natural experiment, exhibit strong evidence for natural selection.
overall pattern is clear. Kettlewell’s experiments established indis-
putably that selection favors dark moths in polluted habitats and
The agent of selection may be light moths in pristine areas. The increase and subsequent decrease
difficult to pin down in the frequency of melanic moths, correlated with levels of pol-
lution independently on two continents, demonstrates clearly that
Although the evidence for natural selection in the case of the
this selection drives evolutionary change.
peppered moth is strong, Tutt’s hypothesis that predation on
The current reconsideration of the agent of natural selection
color-mismatched moths is the underlying mechanism of selection
illustrates well the way in which scientific progress is achieved: hy-
is currently being reevaluated. Researchers have noted­that the
potheses, such as Tutt’s, are put forth and then tested. If they are re-
recent selection against melanism does not appear to correlate with
jected, new hypotheses are formulated, and the process begins anew.
changes in the abundance of light-colored tree lichens.
At Caldy Common, the light form of the peppered moth
began to increase in frequency long before lichens began to Learning Outcomes Review 21.2
reappear on the trees. At the Detroit field station, the lichens never Natural selection has favored the dark form of the peppered moth in
changed significantly as the dark moths first became dominant and areas subject to severe air pollution, perhaps because on darkened
then declined over a 30-year period. In fact, investigators have not trees they are less easily seen by moth-eating birds. As pollution
been able to find peppered moths on Detroit trees at all, whether has abated, selection has in turn shifted to favor the light form.
covered with lichens or not. Some evidence suggests the moths rest Although selection is clearly occurring, further research is required to
understand whether predation by birds is the agent of selection.
on leaves in the treetops during the day, but no one is sure. Could
poisoning by pollution rather than predation by birds be the agent ■■ How would you test the idea that predation by birds is the
of natural selection on the moths? Perhaps—but to date, only agent of selection on moth coloration?
selection resulting from bird predation has been demonstrated.

chapter 21 The Evidence for Evolution 447

 SCIENTIFIC THINKING
21.3 Artificial Selection: Question: Can artificial selection lead to substantial evolutionary change?
Human-Initiated Hypothesis: Strong directional selection will quickly lead to a large

Change
shift in the mean value of the population.
Experiment: In one population, every generation pick out the 20% of
the population with the most bristles and allow them to reproduce to
form the next generation. In the other population, do the same with the
Learning Outcomes 20% with the smallest number of bristles.

1. Contrast the processes of artificial and natural

selection.
2. Explain what artificial selection demonstrates about the Initial
population
power of natural selection.
Low
population

Number of Individuals
Humans have imposed selection upon plants and animals since the High
dawn of civilization. Just as in natural selection, such a­ rtificial population
selection operates by favoring individuals with certain phenotypic
traits, allowing them to reproduce and pass their genes on to the
next generation. Assuming that phenotypic differences are geneti-

Mean

Mean

Mean
cally determined, this directional selection should lead to evolu-
tionary change, and indeed it has. 0 10 20 30 40 50 60 70 80 90 100 110
Artificial selection, imposed in laboratory experiments, agri- Bristle number in Drosophila
culture, and the domestication process, has produced substantial
change in almost every case in which it has been applied. This suc- Result: After 35 generations, mean number of bristles has changed
cess is strong proof that selection is an effective evolutionary substantially in both populations.
process. Interpretation: Note that at the end of the experiment, the range of
variation lies outside the range seen in the initial population. Selection
Experimental selection produces can move a population beyond its original range because mutation and
recombination continuously introduce new variation into populations.
changes in populations
With the rise of genetics as a field of science in the 1920s and
Figure 21.5 Artificial selection can lead to rapid and
1930s, researchers began conducting experiments to test the hypoth-
substantial evolutionary change.
esis that selection can produce evolutionary change. A favorite sub-
ject was the laboratory fruit fly, Drosophila melanogaster.
Inquiry question What would happen if, within a
Geneticists have imposed selection on just about every conceivable
aspect of the fruit fly—including body size, eye color, growth rate, ? population, both small and large individuals were allowed to
life span, and exploratory behavior—with a consistent result: breed, but middle-sized ones were not?
selection for a trait leads to a strong and predictable evolutionary
response.
In one classic experiment, scientists selected for fruit flies
with many bristles (stiff, hairlike structures) on their abdomens. At
the start of the experiment, the average number of bristles was 9.5. Agricultural selection has led to extensive
Each generation, scientists picked out the 20% of the population
with the greatest number of bristles and allowed them to repro-
modification of crops and livestock
duce, thus establishing the next generation. After 86 generations of Familiar livestock, such as cattle and pigs, and crops, such as corn
this directional selection, the average number of bristles had qua- and strawberries, are greatly different from their wild ancestors
drupled, to nearly 40! In another experiment, fruit flies in one (figure 21.6). These differences have resulted from generations of
population were selected for high numbers of bristles, while fruit human selection for desirable traits, such as greater milk produc-
flies in the other cage were selected for low numbers of bristles. tion and larger corn ear size.
Within 35 generations, the populations did not overlap at all in An experiment with corn demonstrates the ability of artifi-
range of variation (figure 21.5). cial selection to rapidly produce major change in crop plants. In
Similar experiments have been conducted on a wide variety 1896, agricultural scientists began selecting for the oil content of
of other laboratory organisms. For example, by selecting for rats corn kernels, which initially was 4.5%. Just as in the fruit fly
that were resistant to tooth decay, in less than 20 generations scien- experiments, the top 20% of all individuals were allowed to repro-
tists were able to increase the average time for onset of decay from duce. By 1986, at which time 90 generations had passed, average
barely over 100 days to greater than 500 days. oil content of the corn kernels had increased approximately 450%.

448 part IV Evolution

 Figure 21.6 Corn
looks very different
from its ancestor.
Teosinte, which can be
found today in a remote
part of Mexico, is very
similar to the ancestor
of modern corn.
Artificial selection has
transformed it into the
Figure 21.8 Domesticated foxes. After 40 years of
Modern form we know today.
Teosinte Intermediates corn selectively breeding the tamest individuals, artificial selection has
produced silver foxes that not only are as friendly as domestic dogs,
but also exhibit many physical traits seen in dog breeds. Artyom
Domesticated breeds have arisen Geodakyan/ITAR-TASS News Agency/Alamy Stock Photo

from artificial selection

Human-imposed selection has produced a great variety of breeds of
cats, dogs (figure 21.7), pigeons, and other domestic animals. In
some cases, breeds have been developed for particular purposes. licking their caretakers (­figure 21.8). In many respects, they had
Greyhound dogs, for example, resulted from s­ election for maximal become no different from domestic dogs.
running ability, resulting in an animal with long legs, a long tail for It was not only their behavior that changed, however. These
balance, an arched back to increase stride length, and great muscle foxes also began to exhibit other traits seen in some dog breeds,
mass. By contrast, the odd proportions of the ungainly dachshund such as different color patterns, floppy ears, curled tails, and short-
resulted from selection for dogs that could enter narrow holes in er legs and tails. Presumably, the genes responsible for docile
pursuit of badgers. In other cases, varieties have been selected pri- behavior either affect these traits as well or are closely­linked to the
marily for their appearance, such as the many colorful breeds of pi- genes for these other traits (the phenomena of pleiotropy and link-
geons or cats. age, which are discussed in chapters 12 and 13).
Domestication also has led to unintentional selection for
some traits. In recent years, as part of an attempt to domesticate the
Can selection produce major evolutionary changes?
silver fox, Russian scientists chose the most docile animals in each
generation and allowed them to reproduce. Within 40 years, most Given that we can observe the results of selection operating over
foxes were exceptionally tame, not only allowing themselves to a relatively short time, most scientists think that natural selec-
be petted, but also whimpering to get attention and sniffing and tion is the process responsible for the evolutionary changes doc-
umented in the fossil record. Some critics of evolution accept
that selection can lead to changes within a species, but contend
that such changes are relatively minor in scope and not equiva-
lent to the substantial changes documented in the fossil record.
In other words, it is one thing to change the number of bristles
Fox on a fruit fly or the size of an ear of corn, and quite another to
produce an entirely new species.
Chihuahua This argument does not fully appreciate the extent of
change produced by artificial selection. Consider, for example,
the existing breeds of dogs, all of which have been produced in
the last few thousand years after wolves were domesticated,
Coyote
Dachshund perhaps 15,000 years ago. If the various dog breeds did not
exist and a paleontologist found fossils of animals similar to
dachshunds, greyhounds, mastiffs, and chihuahuas, there is
no question that they would be considered different species.
Indeed, the differences in size and shape exhibited by these
breeds are greater than those between members of different gen-
Wolf Greyhound era in the family Canidae—such as coyotes, jackals, foxes, and
wolves—which have been evolving separately for 5 to 10 mil-
lion years. Consequently,­the claim that artificial selection pro-
Figure 21.7 Breeds of duces only minor changes is clearly incorrect. If selection
dogs. The differences among operating over a ­ period of only a few thousand years can
dog breeds are greater­than the produce such substantial d­ifferences, it should be powerful
differences displayed among enough, over the course of many millions of years, to produce
wild species of canids. Mastiff the diversity of life we see around us today.

chapter 21 The Evidence for Evolution 449

 radioactive decay
Learning Outcomes Review 21.3
1 parent daughter
In artificial selection, humans choose which plants or animals isotope isotope
to mate in an attempt to conserve desirable traits. Rapid and
substantial results can be obtained over a very short time, often 0.75 Amount of daughter isotope

Proportion of parent
isotope remaining
in a few generations. From this we can see that natural selection
1
is capable of producing major evolutionary change. 0.50 2
■■ In what circumstances might artificial selection fail to 1
produce a desired change? 0.25 4
1
Amount of 8 1
parent isotope 16
0
0 1 2 3 4 5
Time in half-lives
21.4 Fossil Evidence of Evolution
Figure 21.9 Isotopic decay. Isotopes decay at a known rate,
called their half-life. After one half-life, one-half of the original
amount of parent isotope has transformed into a daughter isotope.
Learning Outcomes After each successive half-life, one-half of the remaining amount of
1. Describe how fossils are formed. the parent isotope is transformed.
2. Explain the importance of the discovery of transitional fossils.
3. Name the evolutionary trends revealed by the study of horse
evolution. For example, potassium is one of the most common atoms in
organisms. All potassium (K) atoms have the same number of
protons, but the isotopes of K vary in the number of neutrons they
The most direct evidence that evolution has occurred is found in have. 40K has 19 protons and 21 neutrons and is less stable than 39K
the fossil record. Today we have a far more complete understand- or 41K. 40K is converted (decays) over time and forms 40Ar (argon).
ing of this record than was available in Darwin’s time. The 40K half-life is 1.25 billion years. That is, it takes 1.25 billion
Fossils are the preserved remains of once-living organisms. years for the amount of 40K to decrease by 50%. The long half-life
They include specimens preserved in amber, Siberian permafrost, and makes it useful for dating ancient fossils by determining the ratio
dry caves, as well as the more common fossils preserved as rocks. of 40K to 40Ar (figure 21.9).
Rock fossils are created when three events occur. First, the For events that occurred more recently, radiocarbon dating
organism must become buried in sediment; then, the calcium in can be used. Carbon in the form of atmospheric CO2, with a mix of
bone or other hard tissue must mineralize; and finally, the sur- isotopes 14C and 12C, is incorporated into plants via photosynthesis.
rounding sediment must eventually harden to form rock. The relative amount of 14C to 12C decreases with a half-life of
The process of fossilization occurs only rarely. Usually, about 5700 years. Other isotopes can be used for more interme-
animal or plant remains decay or are scavenged before the process diate dates.
can begin. In addition, many fossils occur in rocks that are inacces-
sible to scientists. When they do become available, they are often Fossils present a history of evolutionary change
destroyed by erosion and other natural processes before they can
be collected. As a result, only a very small fraction of the species When fossils are arrayed according to their age, from oldest to
that have ever existed (estimated by some to be as many as youngest, they often provide evidence of successive evolutionary
500 million) are known from fossils. Nonetheless, the fossils that change. At the largest scale, the fossil record documents the course
have been discovered are sufficient to provide detailed information of life through time, from the origin of prokaryotic and then eu-
on the course of evolution through time. karyotic organisms, through the evolution of fishes, the rise of
land-dwelling organisms, the reign of the dinosaurs, and on to the
The age of fossils can be estimated origin of humans. In addition, the fossil record shows the waxing
and waning of biological diversity through time, such as the
By dating the rocks in which fossils occur, we can get an accurate periodic mass extinctions that have reduced the number of living
idea of how old the fossils are. In Darwin’s day, rocks were dated species. These topics are discussed at greater length in chapter 25.
by their position with respect to one another (relative dating);
rocks in lower strata are generally older because young rocks form
on top of older ones. Knowing the relative positions of sedimen-
Fossils document evolutionary transitions
tary rocks and the rates of erosion of different kinds of sedimentary Given the low likelihood of fossil preservation and recovery, it is
rocks in different environments, geologists of the 19th century not surprising that there are gaps in the fossil record. Nonetheless,
derived a fairly accurate idea of the relative ages of rocks. intermediate forms are often available to illustrate how the major
Today, geologists can determine the absolute age of rocks transitions in life occurred.
using isotopic dating. At the time a rock forms, some elements exist Undoubtedly the most famous of these is the oldest known
as different isotopes. Over time the less stable isotope is converted bird, Archaeopteryx (meaning “ancient feather”), which lived
into the other isotope and the ratio of the two forms changes. around 165 million years ago (mya) (figure 21.10). This species is

450 part IV Evolution

 200 mya, oysters underwent a change from having small, curved
shells to having larger, flatter ones, with progressively flatter fos-
sils seen in the fossil record over a period of 12 million years. A
host of other examples illustrate similar records of successive
change. The demonstration of this successive change is one of the
strongest lines of evidence that evolution has occurred.

The evolution of horses is a prime

example of evidence from fossils
One of the most studied cases in the fossil record concerns the
evolution of horses. Modern-day members of the family Equidae
include horses, zebras, donkeys, and asses, all of which are large,
long-legged, fast-running animals adapted to living on open
Figure 21.10 Fossil of Archaeopteryx, the first bird. grasslands. These species, all classified in the genus Equus, are
The remarkable preservation of this specimen reveals soft parts usually the last living descendants of a long lineage that has produced
not preserved in fossils; the presence of feathers like those of modern- 34 genera since its origin in the Eocene period, approximately
day birds, as well as other features, make clear that Archaeopteryx was 55 mya. Examination of these fossils has provided a particularly
a bird, despite the presence of many dinosaurian traits. To see a picture
of what this animal may have looked like in life, see figure 34.29.
Fossil evidence documenting the evolutionary descent of birds from
dinosaurs is considered in greater detail in chapter 23 (see figure 23.12).
Jason Edwards/National Geographic/Getty Images

clearly intermediate between birds and dinosaurs. Its feathers, simi-

lar in many respects to those of birds today, and other features
­clearly reveal that it is a bird. Nonetheless, in many other respects— Modern toothed whales
for ­example, possession of teeth, a bony tail, and other anatomical
characteristics—it is indistinguishable from some carnivorous
­
dinosaurs. Indeed, it is so similar to these dinosaurs that several Rodhocetus kasrani's
specimens lacking preserved feathers were misidentified as ­dinosaurs reduced hindlimbs
and lay in the wrong natural history museum cabinet for ­several could not have aided it
in walking or swimming.
decades before the mistake was discovered! Rodhocetus swam with
Archaeopteryx reveals a pattern commonly seen in interme- an up-and-down motion,
diate fossils—rather than being intermediate in every trait, such as do modern whales.
fossils usually exhibit some traits like their ancestors and others
like their descendants. In other words, traits evolve at different
Ambulocetus natans
rates and different times; expecting an intermediate form to be in- probably walked on land
termediate in every trait would not be correct. (as do modern sea lions)
The first Archaeopteryx fossil was discovered in 1859, the and swam by flexing
its backbone and paddling
year Darwin published On the Origin of Species. Since then, pale- with its hindlimbs
ontologists have continued to fill in the gaps in the fossil record. (as do modern otters).
Today, the fossil record is far more complete, particularly among
the vertebrates; fossils have been found linking all the major groups.
Recent years have seen spectacular discoveries, closing
Pakicetus attocki lived on land,
some of the major remaining gaps in our understanding of verte- but its skull differed from that of
brate evolution. For example, four-legged aquatic mammals have its ancestors and exhibited
been discovered that provide important insights concerning the many characteristics seen in
the skulls of whales today.
evolution of whales and dolphins from land-dwelling, hoofed an-
cestors (figure 21.11). Similarly, fossil snakes with legs have shed
light on the evolution of snakes, which are descended from lizards Figure 21.11 Whale “missing links.” The discoveries of
that gradually became more and more elongated with the simulta- Ambulocetus, Rodhocetus, and Pakicetus have filled in the gaps
neous ­reduction and eventual disappearance of the limbs. In chap- between whales and their hoofed mammal ancestors. The features of
ter 34, we discuss another recent discovery, Tiktaalik, a species Pakicetus illustrate that intermediate forms are not intermediate in all
that bridged the gap between fish and the first land-living verte- characteristics; rather, some traits evolve before others. In the case of
brates (see figure 34.17). the evolution of whales, changes occurred in the skull prior to
On a finer scale, evolutionary change within some types of evolutionary modification of the limbs. All three fossil forms
animals is known in exceptional detail. For example, about occurred in the Eocene period, 45–55 mya.

chapter 21 The Evidence for Evolution 451

well-documented case of how evolution has proceeded through Pleistocene browsers
adaptation to changing environments. Pliocene
grazers
mixed feeders
5 MYA
The first horse
The earliest known members of the horse family, species in the ge-

Megahippus
10 MYA

Hypohippus
nus Hyracotherium, didn’t look much like modern-day horses at all.

Anchitherium
Miocene
Small, with short legs and broad feet, these species occurred in 15 MYA
wooded habitats, where they probably browsed on leaves and herbs
and escaped predators by dodging through openings in the forest 20 MYA
vegetation. The evolutionary path from these diminutive creatures to
the workhorses of today has involved changes in a variety of traits, 25 MYA

Miohippus
including size, toe reduction, and tooth size and shape (figure 21.12). Oligocene
30 MYA
Changes in size

Mesohippus
35 MYA
The first species of horses were as big as a large house cat or a me-
dium-sized dog. By contrast, modern equids can weigh more than
40 MYA
500 kg. Examination of the fossil record reveals that horses changed

Epihippus
little in size for their first 30 million years, but since then, a number Eocene 45 MYA

Orohippus
of different lineages have exhibited rapid and substantial increases.

Hyracotherium
However, evolution has not been unidirectional and trends toward 50 MYA
decreased size were also exhibited in some branches of the equid
evolutionary tree, as revealed, for example, by Nannippus. 55 MYA

Toe reduction 60 MYA

The feet of modern horses have a single toe enclosed in a tough,
bony hoof. By contrast, Hyracotherium had four toes on its front
feet and three on its hind feet. Rather than hooves, these toes were
encased in fleshy pads like those of dogs and cats.
Examination of fossils clearly shows the transition through
time: a general increase in length of the central toe, development of
the bony hoof, and reduction and loss of the other toes (figure 21.12).
As with body size, these trends occurred concurrently on several
different branches of the horse evolutionary tree and were not ex-
Hyracotherium Mesohippus Anchitherium
hibited by all lineages. (browsers) (browsers) (browsers)
At the same time that toe reduction was occurring, these
horse lineages were evolving changes in the length and skeletal
Figure 21.12 Evolutionary change in body size of
structure of their limbs, leading to animals capable of running long
horses. Lines indicate evolutionary relationships of the horse family.
dis­tances at high speeds.
Horse evolution is more like a bush than like a single-trunk tree;
Tooth size and shape diversity was much greater in the past than it is today. In general,
there has been a trend toward larger size, more complex molar teeth,
The teeth of Hyracotherium were small and relatively simple in
and fewer toes, but this trend has exceptions. For example, a relatively
shape. Through time, horse teeth have increased greatly in length
recent form, Nannippus, evolved in the opposite direction, toward
and have developed a complex pattern of ridges on their molars and
decreased size.
premolars. The effect of these changes is to produce teeth better
capable of chewing tough and gritty vegetation, such as grass, which
Inquiry question Why might the evolutionary line
?
tends to wear teeth down. As with body size, evolutionary change
has not been constant through time. Rather, much of the change in leading to Nannippus have experienced an evolutionary
tooth shape has occurred within the past 20 million years, and decrease in body size?
changes have not been constant across all horse lineages.
All of these changes may be understood as adaptations to
changing global climates. In particular, during the late Miocene the same time, horses were eating grasses and other vegetation that
and early Oligocene epochs (approximately 20 to 25 mya), grass- contained more grit and other hard substances, thus favoring teeth
lands became widespread in North America, where much of horse better suited for withstanding such materials.
evolution occurred. As horses shifted from forests to grasslands,
high-speed locomotion probably became more important to escape
Evolutionary trends
predators. By contrast, the greater flexibility provided by multiple For many years, horse evolution was held up as an example of
toes and shorter limbs, which was advantageous for ducking constant evolutionary change through time. Some even saw in the
through complex forest vegetation, was no longer beneficial. At record of horse evolution evidence for a progressive, guiding force,

452 part IV Evolution

 Equus
Onohippidion
Astrohippus
Neohipparion

Cormohipparion
Nannippus
Pseudhipparion

Dinohippus
Pliohippus
Protohippus
Hipparion

Calippus
Desmatippus

Merychippus
Archaeohippus

Parahippus
Kalobatippus

Merychippus Neohipparion Nannippus Equus

(mixed feeders) (grazers) (grazers) (grazers)

consistently pushing evolution in a single direction, toward longer Hyracotherium to modern-day Equus. But today’s limited horse
limbs, fewer toes, and larger and more complex teeth. We now diversity—only one surviving g­ enus—is unusual. In fact, at the peak
know that such views are misguided, and that the course of evolu- of horse diversity in the Miocene epoch, 13 genera of horses could be
tionary change over millions of years is rarely so simple. found in North America alone. These species differed in body size
Rather, the fossils demonstrate that even though overall and in a wide variety of other characteristics. Presumably, they lived
trends have been evident in a variety of characteristics, evolution- in different habitats and exhibited different dietary preferences. Had
ary change has been far from constant and uniform through time. this diversity existed to modern times, early evolutionary biologists
Instead, rates of evolution have varied widely, with long periods of would likely have had a different outlook on horse evolution.
little observable change and some periods of great change. More-
over, when changes happen, they often occur simultaneously in
multiple lineages of the horse evolutionary tree. Learning Outcomes Review 21.4
Finally, even when a trend exists, exceptions, such as the evo-
Fossils form when an organism is preserved in a matrix such
lutionary decrease in body size exhibited by some lineages, are not as amber, permafrost, or rock. They can be used to construct
uncommon. These patterns are usually discovered for any group of a record of evolutionary transitions over long periods of time,
plants and animals for which we have an extensive fossil record, as which allows us to understand how major changes in evolution
you will see when we discuss human evolution in chapter 34. occur. The extensive fossil record for horses provides a detailed
view of evolutionary diversification of this group, although trends
Horse diversity are not constant and uniform and may include exceptions.
One reason that horse evolution was originally conceived of as linear ■■ Why might rates and direction of evolutionary change vary
through time may be that modern horse diversity is relatively limited. through time?
For this reason it is easy to mentally picture a straight line from

chapter 21 The Evidence for Evolution 453

 should these very different structures be composed of the same
21.5 Anatomical Evidence bones—a single upper forearm bone, a pair of lower forearm
bones, several small carpals, and one or more digits? If evolution
for Evolution had not occurred, this would indeed be a riddle. But when we con-
sider that all of these animals are descended from a common an-
cestor, it is easy to understand that natural selection has modified
the same initial starting blocks to serve very different purposes.
Learning Outcomes
1. Explain the evolutionary significance of homologous and
vestigial structures Early embryonic development shows
2. Describe how patterns of early development provide evidence similarities in some groups
for evolution.
Some of the strongest anatomical evidence supporting evolution
comes from comparisons of how organisms develop. Embryos of
different types of vertebrates, for example, often are similar early on,
Much of the power of the theory of evolution is its ability to pro- but become more different as they develop. Early in their develop-
vide a sensible framework for understanding the diversity of life. ment, vertebrate embryos possess pharyngeal pouches, which de-
Many observations from throughout biology simply cannot be velop into different structures. In humans, for example, they become
understood in any meaningful way except as a result of evolution. various glands and ducts; in fish, they turn into gill slits. At a later
stage, all primate embryos have a long tail, but whereas monkeys
Homologous structures suggest and most primates keep the tail, all we and our ape relatives retain is
common derivation the coccyx at the end of our spine. Human fetuses even possess a
fine fur (called lanugo) during the fifth month of development.
As vertebrates have evolved, the same bones have sometimes been
Similarly, although most frogs go through a tadpole stage,
put to different uses. Yet the bones are still recognizable, their
some species develop directly and hatch out as little, fully formed
presence betraying their evolutionary past. For example, the fore-
frogs. However, the embryos of these species still exhibit tadpole
limbs of vertebrates are all homologous structures—structures
features, such as the presence of a tail, which disappear before the
with different appearances and functions that all derived from the
froglet hatches (figure 21.14).
same body part in a common ancestor.
These relict developmental forms suggest strongly that our
You can see in figure 21.13 how the bones of the forelimb
development has evolved, with new instructions modifying ances-
have been modified in different ways for different mammals. Why
tral developmental patterns. We have discussed the topic of embry-
onic development and evolution in chapter 19.

Humerus Some structures are imperfectly

suited to their use
Because natural selection can work only on the variation present in
Radius
a population, it should not be surprising that some organisms do
Ulna
not appear perfectly adapted to their environments. For example,
most animals with long necks have many neck vertebrae for
Carpals enhanced flexibility: geese have up to 25, and plesiosaurs, the
Metacarpals long-necked reptiles that patrolled the seas during the age of dino-
Phalanges saurs, had as many as 76. By contrast, giraffes have only 7 very
long neck vertebrae. Why haven’t they evolved more, like other
Human Cat Bat
long-necked animals? It turns out that almost all mammals have
only 7 neck vertebrae. Because mammal species have no variation
in vertebra number among individuals in a population, natural
selection has nothing to work with, and thus there is no way for
selection to lead to increased numbers of vertebrae (why other
types of animals are more variable is a question for which we cur-
Figure 21.13 Homology rently don’t have an answer). In the absence of variation in vertebra
of the bones of the forelimb number, selection led to an evolutionary increase in vertebra size
of mammals. Although these to produce the long neck of the giraffe.
structures show considerable An excellent example of an imperfect design is the eye of
differences in form and function, vertebrate animals, in which the photoreceptors face backward,
the same basic bones are present toward the wall of the eye (figure 21.15a). As a result, the nerve
in the f­ orelimbs of humans, cats, fibers extend not backward, toward the brain, but forward into
bats, porpoises, and horses. Porpoise Horse the eye chamber, where they slightly obstruct light. Moreover,

454 part IV Evolution

 Such examples illustrate that natural selection is like a tin-
kerer, working with whatever material is available to craft a work-
able solution, rather than like an engineer, who can design and
build the best possible structure for a given task. Workable, but
imperfect, structures such as the vertebrate eye are an expected
outcome of evolution by natural selection.

Vestigial structures can be explained

as holdovers from the past
Many organisms possess vestigial structures that have no appar-
ent function, but resemble structures their ancestors possessed.
Humans, for example, possess a complete set of muscles for wig-
gling their ears, just as many other mammals do. Although these
muscles allow other mammals to move their ears to pinpoint
sounds such as the movements or growl of a predator, they have
little purpose in humans other than amusement.
As other examples, boa constrictors have hip bones and rudi-
mentary hind legs. Manatees (a type of aquatic mammal often
referred to as “sea cows”) have fingernails on their flippers, which
evolved from legs (figure 21.16). Some blind cave fish, which
never see the light of day, have small, nonfunctional eyes (other
blind cave fish have lost their eyes entirely).
The human vermiform appendix has long been thought to be
vestigial; it represents the degenerate terminal part of the cecum, the
blind pouch or sac in which the large intestine begins. In other
­mammals, such as mice, the cecum is the largest part of the large
Figure 21.14 Developmental features reflect intestine and functions in storage—usually of bulk cellulose in
evolutionary ancestry. Some species of frogs have lost the herbivores. Although greatly diminished in size, some recent
­
tadpole stage. Nonetheless, tadpole features first appear and then ­evidence suggests that the human appendix may not technically be
disappear during development in the egg. ©James Hanken, Museum of vestigial because it may harbor beneficial intestinal bacteria. ­Vestigial
Comparative Zoology, Harvard University, Cambridge or not, the human appendix can be a dangerous organ: appendicitis,
which results from infection of the appendix, can be fatal.
these fibers bundle together to form the optic nerve, which exits Vestigial traits can also be seen in the genomes of many
through a hole at the back of the eye, creating a blind spot. organisms. For example, the icefish (figure 21.17) is a bizarre-
By contrast, the eyes of mollusks—such as squid and looking, nearly see-through fish that lives in the frigid waters of
­octopuses—are better arranged: the photoreceptors face forward, the Antarctic. The icefish’s transparency results not only from a
and the nerve fibers exit at the back, neither obstructing light nor lack of pigment in its body structures, but also from the near invis-
creating a blind spot (figure 21.15b). ibility of its blood. Our blood is red due to the presence of red

Blind
spot

Photopigment Photoreceptor cells

Photoreceptor cells Interneuron Nerve fibers

Light Light
Photopigment Nerve fibers
To brain via to brain
Nerve impulse
optic nerve
a. b.
Figure 21.15 The eyes of vertebrates and mollusks. a. Photoreceptors of vertebrates point backward, whereas (b) those of mollusks
face forward. As a result, vertebrate nerve fibers pass in front of the photoreceptor—and where they bundle together and exit the eye, a blind spot
is created. Mollusks’ eyes have neither of these problems.

chapter 21 The Evidence for Evolution 455

 the cold waters of the Antarctic and lost the need for hemoglobin,
mutations that would prevent the production of hemoglobin and
thus would be filtered out by natural selection in other species
were able to persist in the population. Just by chance, some of
these mutations increased in frequency in the population through
time (the process of genetic drift, discussed in chapter 20), eventu-
ally becoming established in all individuals and knocking out the
fish’s ability to produce hemoglobin.
Pseudogenes, sometimes called fossil genes because they
are traces of previously functioning genes, such as the hemoglobin
gene in the icefish, are actually quite common in the genomes of
most organisms and are discussed in chapter 24: when a trait disap-
pears, the gene does not just vanish from the genome; rather, some
mutation renders it inactive, and once that occurs, other mutations
can accumulate.
It is difficult to understand vestigial structures such as these
as anything other than evolutionary relicts, holdovers from the past.
However, the existence of vestigial structures argues strongly for
the common ancestry of the members of the groups that share
Figure 21.16 Vestigial structures. The flippers of the West
them, regardless of how different those groups have subsequently
Indian manatee (Trichechus manatus), a relative of the elephant and
become.
descended from a terrestrial mammal, have retained nails, even
All of these anatomical lines of evidence—homology,
though the manatee never leaves the water.
development, and imperfect and vestigial structures—are readily
understandable as a result of descent with modification, that is,
blood cells, which contain hemoglobin, the molecule that trans- evolution.
ports oxygen from the lungs to the tissues (see chapter 47).
­However, oxygen concentration in water increases as temperature
decreases. The waters of the Antarctic, which are about 0°C, con-
tain so much oxygen that the fish do not need special molecules to Learning Outcomes Review 21.5
carry oxygen. The result is that these fish do not have hemoglobin, Comparisons of the anatomy of different living animals often reveal
and consequently their blood is colorless. Nonetheless, when the evidence of shared ancestry. In cases of homology, the same organ
DNA of icefish was examined, scientists discovered that they have has evolved to carry out different functions. In other cases, an
the same gene that produces hemoglobin as that found in other organ is still present, usually in diminished form, even though it has
vertebrates. However, the icefish hemoglobin gene has a variety of lost its function altogether; such an organ or structure is termed
mutations that render it nonfunctional, and thus the icefish does vestigial.
not produce hemoglobin. The presence of this inoperative version ■■ What are alternative explanations for homologous and
of the hemoglobin gene in icefish means that its ancestors had vestigial structures?
hemoglobin; however, once the icefish’s progenitors occupied

21.6 Convergent Evolution

and the Biogeographical
Record

Learning Outcomes
1. Explain the principle of convergent evolution.
2. Demonstrate how the biogeographical distribution of plant
and animal species on islands provides evidence of
evolutionary diversification.
Figure 21.17 Photo of an icefish. This nearly transparent
fish, photographed from above, is found in the frigid waters of the Biogeography, the study of the geographic distribution of spe-
Antarctic. British Antarctic Survey/Science Source cies, reveals that different geographical areas sometimes exhibit

456 part IV Evolution

groups of plants and animals of strikingly similar appearance, evolved in different, isolated areas because of similar selective
even though the organisms may be only distantly related. pressures in similar environments.
It is difficult to explain so many similarities as the result of
coincidence. Instead, natural selection appears to have favored paral- Convergent evolution
lel evolutionary adaptations in similar environments. Because selec- is a widespread phenomenon
tion in these instances has tended to favor changes that made the two
groups more alike, their phenotypes have converged. This form of When species interact with the environment in similar ways, they
evolutionary change is referred to as convergent evolution. often are exposed to similar selective pressures, and they there-
fore frequently develop the same evolutionary adaptations. Con-
Marsupials and placentals sider, for example, fast-moving marine predators (­figure 21.19).
The hydrodynamics of moving through water require a stream-
demonstrate convergence lined body shape to minimize friction. It is no coincidence that
In the best-known case of convergent evolution, two major groups dolphins, sharks, and tuna—among the fastest of marine
of mammals—marsupials and placentals—have evolved in very ­species—have all evolved to have the same basic shape. We can
similar ways in different parts of the world. Marsupials are a group infer as well that ichthyosaurs—marine reptiles that lived during
in which the young are born in a very immature condition and held the Age of the Dinosaurs—exhibited a similar lifestyle.
in a pouch until they are ready to emerge into the outside world. In Island trees exhibit a similar phenomenon. Most islands are
placentals, by contrast, offspring are not born until they can safely covered by trees (or were until the arrival of humans). Careful
survive in the external environment (with varying degrees inspection of these trees, however, reveals that they are not closely
of parental care). related to the trees with which we are familiar. Although they have
Australia separated from the other continents more than all the characteristics of trees, such as being tall and having a tough
70 mya; at that time, both marsupials and placental mammals had outer covering, in many cases island trees are members of plant
evolved, but in different places. In particular, only marsupials families that elsewhere exist only as flowers, shrubs, or other small
occurred in Australia. As a result of this separation, the only pla- bushes. For example, on many islands, the native trees are mem-
cental mammals in Australia today are bats and a few colonizing bers of the sunflower family.
rodents (which arrived relatively recently), and the continent is Why do these plants evolve into trees on islands? Probably
dominated by marsupials. because seeds from trees rarely make it to isolated islands. As a re-
What are the Australian marsupials like? To an astonishing sult, those species that do manage to colonize distant islands face an
degree, they resemble the placental mammals living today on the empty ecological landscape upon arrival. In the absence of other
other continents (figure 21.18). The similarity between some indi- treelike plants, natural selection often favors individual­plants that
vidual members of these two sets of mammals argues strongly that can capture the most sunlight for photosynthesis, and the result is the
they are the result of convergent evolution, similar forms having evolution of similar treelike forms on islands throughout the world.

Nocturnal Stalking Chasing

Niche Burrower Anteater Climber Glider
Insectivore Predator Predator

Placental
Mole Grasshopper
Mammals
mouse Flying squirrel Wolf

Lesser anteater Ring-tailed lemur Ocelot

Australian
Marsupials Numbat
Thylacine
Tree
kangaroo
Marsupial mole Marsupial Flying phalanger Tasmanian
mouse quoll

Figure 21.18 Convergent evolution. Many marsupial species in Australia resemble placental mammals occupying similar ecological
niches elsewhere in the rest of the world. Marsupials evolved in isolation after Australia separated from other continents.

chapter 21 The Evidence for Evolution 457

Figure 21.19 Convergence among fast-swimming predators. Fast movement through water requires a streamlined body form,
which has evolved numerous times.

Convergent evolution is even seen in humans. People in most

populations stop producing lactase, the enzyme that digests milk,
sometime in childhood. However, individuals in African and 21.7 Darwin’s Critics
European populations that raise cattle produce lactase throughout
their lives. DNA analysis indicates that the retention of lactase pro-
duction into adulthood is the result of different mutations in Africa Learning Outcomes
and Europe, which indicates that the populations have indepen-
1. Characterize the criticisms of evolutionary theory and list
dently (convergently) acquired this adaptation.
counterarguments that can be made.
2. Distinguish between hypothesis and theory in scientific usage.
Biogeographical studies provide
further evidence of evolution
In the century and a half since he proposed it, Darwin’s theory of
Darwin made several important observations during his voyage evolution by natural selection has become nearly universally accept-
around the world. He noted that islands often are missing plants ed by biologists, but has been a source of controversy among some
and animals common on continents, such as frogs and land mam- members of the general public. Here we discuss seven principal ob-
mals. Accidental human introductions have proved that these spe- jections that critics raise to the teaching of evolution as biological
cies can survive if they are released on islands, so lack of suitable fact, along with some answers that scientists present in response:
habitat is not the cause. In addition, those species that are present
on islands often have diverged from their continental relatives and 1. Evolution is not solidly demonstrated. “Evolution is just
sometimes—as with Darwin’s finches and the island trees just a theory,” Darwin’s critics point out, as though theory
discussed—occupy ecological niches used by other species on meant a lack of knowledge, or some kind of guess.
continents. Lastly, island species usually are more closely related Scientists, however, use the word theory in a very
to species on nearby continents, even though the environment on different sense than the general public does. They use
continents and nearby islands often is not very similar. theory only for those ideas that are strongly supported by
Darwin deduced the explanation for these phenomena. Many many lines of evidence, like the theories of gravity and
islands have never been connected to continental areas. The species evolution. It is important to recognize that both evolution
that occur there arrived by dispersing across the water. Some species— and gravitation are “just theories.” The term theory is
those that can fly, float, or swim—are more likely to get to the island not the equivalent of a “hunch” or a “notion,” or even an
than others. Some, like frogs, are particularly vulnerable to dehydra- initial hypothesis. Rather, it is a well-tested phenomenon
tion in salt­water and have almost no chance of island colonization. that rationalizes the data available.
The absence of some types of plants and animals pro- 2. There are no fossil intermediates. “No one ever saw a fin
vides opportunity to those that do arrive; as a result, colonizers, on the way to becoming a leg,” critics claim, pointing to
which usually come from nearby areas, often evolve into many spe- the many gaps in the fossil record in Darwin’s day.
cies exhibiting great ecological and morph­ological diversity. This Since that time, however, many fossil intermediates
phenomenon, termed adaptive radiation, is discussed in chapter 22. in vertebrate evolution have indeed been found. A clear
line of fossils now traces the transition between hoofed
mammals and whales, between reptiles and mammals,
Learning Outcomes Review 21.6 between dinosaurs and birds, and between apes and
Convergence is the evolution of similar forms in different lineages humans. The fossil evidence of evolution from one major
when exposed to similar selective pressures. The biogeographical type to another is compelling.
distribution of species often reflects the outcome of evolutionary 3. The intelligent design argument. “The organs of living
diversification with closely related species in nearby areas. creatures are too complex for a random process to have
■■ Why does convergent evolution occur and why might produced them—the existence of a clock is evidence of
species occupying similar environments in different the existence of a clockmaker.”
localities sometimes not exhibit it? Evolution by natural selection is not a random
process. Quite the contrary, by favoring those variations
458 part IV Evolution
 that lead to the highest reproductive fitness, natural built the structure because intermediate stages, with not
selection is a nonrandom process that can construct all parts in place, would not function and thus would not
highly complex organs by incrementally improving them be favored by natural selection.
from one generation to the next. What’s wrong with this argument is that each part
For example, the intermediates in the evolution of a complex molecular machine evolves as part of the
of the mammalian ear can be seen in fossils, and many whole system. Natural selection can act on a complex
intermediate “eyes” are known in various invertebrates. system because at every stage of its evolution, the
These intermediate forms arose because they have system functions. Parts that improve function are added.
value—being able to detect light slightly is better than Subsequently, other parts may be modified or even
not being able to detect it at all. Complex structures such lost, so that parts that were not essential when they first
as eyes evolved as a progression of slight improvements. evolved become essential. In this way, an “irreducibly
Moreover, inefficiencies of certain designs, such as the complex” structure can evolve by natural selection. The
vertebrate eye and the existence of vestigial structures, same process works at the molecular level.
do not support the idea of an intelligent designer. For example, snake venom initially evolved as
4. Evolution violates the Second Law of Thermodynamics. enzymes to increase the ability of snakes to digest large
“A jumble of soda cans doesn’t by itself jump neatly into a prey items, which were captured by biting the prey
stack—things become more disorganized due to random and then constricting them with coils. Subsequently,
events, not more organized.” the digestive enzymes evolved to become increasingly
Biologists point out that this argument ignores what lethal. Rattlesnakes kill large prey by injecting them
the second law really says: disorder increases in a closed with venom, letting them go, and then tracking them
system, which the Earth most certainly is not. Energy down and eating them after they die. To do so, they
continually enters the biosphere from the Sun, fueling life have evolved extremely toxic venom, highly modified
and all the processes that organize it. syringe-like front teeth, and many other characteristics.
5. Proteins are too improbable. “Hemoglobin has 141 Take away the fangs or the venom and the rattlesnakes
amino acids. The probability that the first one would be can’t feed—what initially evolved as nonessential parts
leucine is 1/20, and that all 141 would be the ones they are now indispensable; irreducible complexity has
are by chance is (1/20)141, an impossibly rare event.” evolved by natural selection.
This argument illustrates a lack of understanding The mammalian blood clotting system similarly has
of probability and statistics—probability cannot be evolved from much simpler systems. The core clotting
used to argue backward. The probability that a student system evolved at the dawn of the vertebrates more than
in a classroom has a particular birthdate is 1/365; 500 mya, and it is found today in primitive fishes such as
arguing this way, the probability that everyone in a lampreys. One hundred million years later, as vertebrates
class of 50 would have the birthdates that they do is continued to evolve, proteins were added to the clotting
(1/365)50, and yet there the class sits, all with their system, making it sensitive to substances released
actual birthdates. from damaged tissues. Fifty million years later, a third
6. Natural selection cannot account for major ­changes in component was added, triggering clotting by contact with
evolution. “No scientist has come up with an experiment in the jagged surfaces produced by injury. At each stage,
which fish evolve into frogs and leap away from predators.” as the clotting system evolved to become more complex,
Can we extrapolate from our understanding that its overall performance came to depend on the added
natural selection produces relatively small changes elements. Thus, blood clotting has become “irreducibly
that are observable in populations within species complex” as the result of Darwinian evolution.
to explain the major differences observed between Statements that various structures could not have
species? Most biologists who have studied the problem been built by natural selection have repeatedly been
think so. The differences between breeds produced by made over the past 150 years. In many cases, after
artificial selection—such as chihuahuas, mastiffs, and detailed scientific study, the likely path by which such
greyhounds—are more distinctive than the differences structures have evolved has been discovered.
between some wild species, and laboratory selection
experiments sometimes create forms that cannot
interbreed and thus would in nature be considered
different species. Thus, production of radically different Learning Outcomes Review 21.7
forms has indeed been observed, repeatedly. These Darwin’s theory of evolution is controversial to some in the general
changes usually take millions of years, and they are seen public. Objections are often based on a misunderstanding of the
clearly in the fossil record. theory. In scientific usage, a hypothesis is an educated guess,
whereas a theory is an explanation that fits available evidence and
7. The irreducible complexity argument. Because each part
has withstood rigorous testing.
of a complex cellular mechanism such as blood clotting
is essential to the overall process, the intricate machinery ■■ Suppose someone suggested that humans originally came
of the cell cannot be explained by evolution from simpler from Mars. Would this be a hypothesis or a theory, and how
stages. Without all of the parts functioning, the system could it be tested?
would not work. Consequently, evolution couldn’t have
chapter 21 The Evidence for Evolution 459
 Chapter Review

Georgia Southern University Museum

21.1 The Beaks of Darwin’s Finches: Evidence Fossils present a history of evolutionary change.
of Natural Selection Fossils document evolutionary transitions.
The history of life on Earth can be traced through the fossil record.
Galápagos finches exhibit variation related to food gathering. In recent years, new fossil discoveries have provided more detailed
The correspondence between beak shape and its use in obtaining food understanding of major evolutionary transitions.
suggested to Darwin that finch species had diversified and adapted to eat
different foods. The evolution of horses is a prime example of evidence from
fossils.
Modern research has verified Darwin’s selection hypothesis. The fossil record indicates that horses have evolved from small, forest-
Natural selection acts on variation in beak morphology, favoring larger- dwelling animals to the large and fast plains-dwelling species alive today.
beaked birds during extended droughts and smaller-beaked birds during
Over the course of 50 million years, evolution has not been constant and
long periods of heavy rains.
uniform. Rather, change has been rapid at some times, slow at others.
Because this variation is heritable, evolutionary change occurs in the Although a general trend toward increase in size is evident, some species
frequencies of beak sizes in subsequent generations. evolved to smaller sizes.

21.2 Peppered Moths and Industrial Melanism: More 21.5 Anatomical Evidence for Evolution
Evidence of Selection
Homologous structures suggest common derivation.
Light-colored moths decreased in polluted areas. Homologous structures may have different appearances and functions
In polluted areas where soot built up on tree trunks, the dark-colored even though derived from the same common ancestral body part.
form of the peppered moth became more common. In unpolluted areas,
Early embryonic development shows similarities in some groups.
light-colored forms remained predominant.
Embryonic development shows similarity in developmental patterns
Experiments suggested that predation by birds was the cause; light- among species whose adult phenotypes are very different.
colored moths stand out on dark trunks, and vice versa.
Species that have lost a feature that was present in an ancestral
When environmental conditions reverse, so does selection form often develop and then lose that feature during embryological
pressure. development.
In the last 50 years, pollution has decreased in many areas and the
Some structures are imperfectly suited to their use.
frequency of light-colored moths has rebounded.
Natural selection can influence only the variation present in a
The agent of selection may be difficult to pin down. population; because of this, evolution often results in workable,
Some have questioned whether bird predation is the agent of selection, but imperfect, structures, such as the vertebrate eye.
but recent research supports this hypothesis. Regardless, the observation
Vestigial structures can be explained as holdovers from the past.
that the dark-colored form has increased during times of pollution and
then declined as pollution abates indicates that natural selection has The existence of vestigial structures supports the concept of common
acted on moth coloration. ancestry among organisms that share them.

21.3 Artificial Selection: Human-Initiated Change 21.6 Convergent Evolution and the
(figure 21.5) Biogeographical Record
Experimental selection produces changes in populations. Marsupials and placentals demonstrate convergence.
Laboratory experiments in directional selection have shown that Convergent evolution may occur in species or populations exposed
substantial evolutionary change can occur in these controlled to similar selective pressures. Marsupial mammals in Australia have
populations. converged upon features of their placental counterparts elsewhere.

Agricultural selection has led to extensive modification of crops Convergent evolution is a widespread phenomenon.
and livestock. Examples include hydrodynamic streamlining in marine species and the
evolution of tree species on islands from ancestral forms that were not
Domesticated breeds have arisen from artificial selection. treelike.
Crop plants and domesticated animal breeds are often substantially
different from their wild ancestors. Biogeographical studies provide further evidence of evolution.
If artificial selection can rapidly create substantial change over short Island species usually are closely related to species on nearby continents
periods of time, then it is reasonable to assume that natural selection even if the environments are different. Early island colonizers often evolve
could have created the Earth’s diversity of life over millions of years. into diverse species because other, competing species are scarce.

21.4 Fossil Evidence of Evolution 21.7 Darwin’s Critics

Darwin’s theory of evolution by natural selection is almost universally
The age of fossils can be estimated. accepted by biologists. Many criticisms have been made both historically and
Specimens become fossilized in different ways. Fossils in rock can be recently, but most stem from a lack of understanding of scientific principles,
dated by calculating the extent of radioactive decay based on half-lives the theory’s actual content, or the time spans involved in evolution.
of known isotopes.

460 part IV Evolution

Visual Summary

Evidence for evolution

includes
evidence from

Convergent Biogeographical
Natural selection Artificial selection Fossils Anatomy
evolution record

exemplified by include seen in

used to
modify

Darwin’s Peppered Transitional exemplified by Marsupials revealed

finches moths species and through
placental
mammals
exemplified by

Crops Livestock Horses

differ from Island vs.

Wild Homologous Embryological Imperfect Vestigial mainland
ancestors structures development structures structures species

Review Questions

Georgia Southern University Museum

U N D E R S TA N D 4. Homologous structures
1. Artificial selection is different from natural selection because a. are structures in two or more species that originate as the
a. artificial selection is not capable of producing large changes. same structure in a common ancestor.
b. artificial selection does not require genetic variation. b. are structures that look the same in different species.
c. natural selection cannot produce new species. c. cannot serve different functions in different species.
d. breeders (people) choose which individuals reproduce based d. must serve different functions in different species.
on desirability of traits. 5. Convergent evolution
2. Gaps in the fossil record a. is an example of stabilizing selection.
a. demonstrate our inability to date geological sediments. b. depends on natural selection to independently produce similar
b. are expected since the probability that any organism will phenotypic responses in different species or populations.
fossilize is extremely low. c. occurs only on islands.
c. have not been filled in as new fossils have been discovered. d. is expected when different lineages are exposed to vastly
d. weaken the theory of evolution. different selective environments.
3. The evolution of modern horses (Equus) is best described as 6. Darwin’s finches are a noteworthy case study of evolution
by natural selection because evidence suggests
a. the constant change and replacement of one species
by another over time. a. they are descendants of many different species that colonized
b. a complex history of lineages that changed over time, the Galápagos.
with many going extinct. b. they radiated from a single species that colonized the Galápagos.
c. a simple history of lineages that have always resembled extant c. they are more closely related to mainland species than
horses. to one another.
d. None of the choices is correct. d. None of the choices is correct.

chapter 21 The Evidence for Evolution 461

 7. The possession of fine fur in 5-month human embryos indicates SYNTHESIZE
a. that the womb is cold at that point in pregnancy. 1. What conditions are necessary for evolution by natural selection?
b. that humans evolved from a hairy ancestor.
c. that hair is a defining feature of mammals. 2. Explain how data shown in figure 21.2a and b relate to the
d. that some parts of the embryo grow faster than others. conditions identified by you in question 1.
3. On figure 21.2b, draw the relationship between offspring beak
A P P LY depth and parent beak depth, assuming that there is no genetic
basis to beak depth in the medium ground finch.
1. In Darwin’s finches,
a. occurrence of wet and dry years preserves genetic variation for 4. Refer to figure 21.5, artificial selection in the laboratory. In this
beak size. experiment, one population of Drosophila was selected for low
b. increasing beak size over time proves that beak size numbers of bristles and the other for high numbers. Note that not
is inherited. only did the means of the populations change greatly in 35
c. large beak size is always favored. generations, but also all individuals in both experimental
d. All of the choices are correct. populations lie outside the range of the initial population. What
would happen if the direction of selection were reversed, such that
2. Artificial selection experiments in the laboratory such as in
a greater number of bristles was selected for in the low-bristle
figure 21.5 are an example of
population, and vice versa? How would the rate of evolutionary
a. stabilizing selection. change compare with that in the initial part of the experiment
b. negative frequency-dependent selection. before selection was reversed?
c. directional selection.
d. disruptive selection. 5. The ancestor of horses was a small, many-toed animal that lived in
forests, whereas today’s horses are large animals with a single hoof
3. Convergent evolution is often seen among species on different
that live on open plains. A series of intermediate fossils illustrate
islands because
how this transition has occurred, and for this reason, many old
a. island populations are usually smaller and more affected by treatments of horse evolution portrayed it as a steady increase
genetic drift. through time in body size accompanied by a steady decrease in toe
b. disruptive selection occurs commonly on islands. number. Why is this interpretation incorrect?
c. island species are usually most closely related to species in
similar habitats elsewhere.
d. when islands are first colonized, many ecological resources
are unused, allowing descendants of a colonizing species to
diversify and adapt to many different parts of the environment.

462 part IV Evolution

 22
CHAPTER

The Origin of Species

Chapter Contents
22.1 The Nature of Species and the Biological
Species Concept
22.2 Natural Selection and Reproductive Isolation
22.3 The Role of Genetic Drift and Natural
Selection in Speciation
22.4 The Geography of Speciation
22.5 Adaptive Radiation and Biological Diversity
22.6 The Pace of Evolution
22.7 Speciation and Extinction Through Time

JohnMernick/iStock/Getty Images

Visual Outline Introduction

Although Darwin titled his
Biological species Reproductive can be Natural
Species defined by
concept
requires
isolation shaped by selection
book On the Origin of Spe-
cies, he never actually dis-
leads to affects
cussed what he referred to as
that “mystery of mysteries”—
Geography affects Speciation may affect Genetic drift
how one species gives rise to
another. Rather, his argu-
can lead to
ment concerned evolution by
natural selection—that is, how
Evolutionary Fluctuations in one species evolves through
resulting in
PACIFIC diversification species numbers time to adapt to its changing
OCEAN
environment. Although an im-
1000

producing
portant mechanism of evolu-
800
NEW
tionary change, the process of
Number of families

GUINEA
600
Cretaceous adaptation does not explain
Adaptive 400
how one species becomes
radiation Devonian

another, a process we call

200 Ordovician
Permian
Triassic

speciation. As we shall see,

0
600 500 400 300 200 100 0
Millions of years ago
adaptation may be involved in
the speciation process, but it
does not have to be.
Before we can discuss
how one species gives rise
 to another, we need to understand exactly what a species is.
Even though the definition of a species is of fundamental
importance to evolutionary biology, this issue has still not
been completely settled and is currently the subject of consid-
erable research and debate.

22.1 The Nature of Species Black

and the Biological Intergrade

Species Concept
Yellow

Learning Outcomes Gray

1. Understand the biological species concept and why it does

not explain all observations. Figure 22.1 Geographic variation in the eastern rat
2. Define the two kinds of reproductive isolating mechanisms. snake, Pantherophis alleghaniensis. Although populations at
3. Describe the relationship of reproductive isolating the eastern, western, and northern ends of the species’ range are
mechanisms to the biological species concept. phenotypically quite distinctive from one another, they are connected by
populations, labelled “intergrade,” that are phenotypically intermediate.
Any concept of a species must account for two phenomena: the
distinctiveness of species that occur together at a single locality, distant populations may appear distinct, they are usually con-
and the connection that exists among different populations belong- nected by intervening populations that are intermediate in their
ing to the same species. characteristics.

Sympatric species inhabit the same The biological species concept focuses
locale but remain distinct on the ability to exchange genes
Put out a birdfeeder on your balcony or in your back yard, and What can account for both the distinctiveness of sympatric species
you will attract a wide variety of birds (especially if you include and the connectedness of geographically separate populations of
different kinds of foods). In the midwestern United States, for the same species? One obvious possibility is that each species ex-
example, you might routinely see cardinals, blue jays, downy changes genetic material only with other members of its species. If
woodpeckers, house finches—even hummingbirds in the summer. sympatric species commonly exchanged genes, which they gener-
Although it might take a few days of careful observation, ally do not, we might expect such species to rapidly lose their dis-
you would soon be able to readily distinguish the many different tinctions, as the gene pools (that is, all of the alleles present in a
species. The reason is that species that occur together (termed species) of the different species became homogenized. Conversely,
sympatric) are distinctive entities that are phenotypically differ- the ability of geographically distant populations of a single species
ent, utilize different parts of the habitat, and behave differently. to share genes through the process of gene flow may keep these
This observation is generally true not only for birds, but also for populations integrated as members of the same species.
most other types of organisms. Based on these ideas, in 1942 the evolutionary biologist
Occasionally, two species occur together that appear to be Ernst Mayr set forth the biological species concept, which defines
nearly identical. In such cases, we need to go beyond visual simi- species as “. . . groups of actually or potentially interbreeding
larities. When other aspects of the phenotype are examined, such natural populations which are reproductively isolated from other
as the mating calls or the chemicals exuded by each species, they such groups.”
usually reveal great differences. In other words, even though we In other words, the biological species concept says that a
might have trouble distinguishing them, the organisms themselves species is composed of populations whose members mate with
have no such difficulties. each other and produce fertile offspring—or would do so if they
came into contact. Conversely, populations whose members do not
Populations of a species exhibit mate with each other or who cannot produce fertile offspring are
said to be reproductively isolated and, therefore, are members of
geographic variation different species.
Within a single species, individuals in populations that occur in What causes reproductive isolation? If organisms cannot in-
different areas may be distinct from one another. In areas where terbreed or cannot produce fertile offspring, they clearly ­belong to
these populations occur close to each other, individuals often different species. However, some populations that are considered
exhibit combinations of features characteristic of both popula- separate species can interbreed and produce fertile offspring, but
tions (figure 22.1). In other words, even though geographically they ordinarily do not do so under natural conditions. They are still

464 part IV Evolution

considered reproductively isolated because genes from one species Prezygotic isolating mechanisms prevent
generally will not enter the gene pool of the other.
Table 22.1 summarizes the steps at which barriers to suc-
the formation of a zygote
cessful reproduction may occur. Such barriers are termed Mechanisms that prevent formation of a zygote include ecological
reproductive isolating mechanisms because they prevent gen­etic or environmental isolation, behavioral isolation, temporal isola-
exchange between species. We will discuss examples of these next, tion, mechanical isolation, and prevention of gamete fusion.
beginning with those that prevent the formation of zygotes, which
are called prezygotic isolating mechanisms. Mechanisms that Ecological isolation
prevent the proper functioning of zygotes after they form are called Even if two species occur in the same area, they may utilize differ-
postzygotic isolating mechanisms. ent portions of the environment and thus not hybridize because
they do not encounter each other. For example, lions and tigers can
produce hybrid offspring in zoos (figure 22.2), but even though the
Reproductive Isolating ranges of the two species overlapped in India until about 150 years
TA B L E 2 2 .1 ago, no natural hybrids were ever discovered. Lions stayed mainly
Mechanisms
in the open grassland and hunted in groups called prides; tigers
Mechanism Description tended to be solitary creatures of the forest. Because of their eco-
PRE Z YG OTI C I SO L ATI N G M E CH A N I S M S logical and behavioral differences, lions and tigers rarely came into
direct contact with each other, even though their ranges overlapped
Ecological isolation Species occur in the same over thousands of square kilometers.
area, but they occupy In another example, the ranges of two toads, Bufo wood­
different habitats and housei and B. americanus, overlap in some areas. Although these
rarely encounter
two species can produce viable hybrids, they usually do not inter-
each other.
breed because they utilize different portions of the habitat for
breeding. B. woodhousei prefers to breed in streams, and B. ameri­
Behavioral isolation Species differ in their canus breeds in rainwater puddles.
mating rituals. Similar situations occur among plants. Two species of oaks
occur widely in California: the valley oak, Quercus lobata, and the
scrub oak, Q. dumosa. The valley oak, a graceful deciduous tree
that can be as tall as 35 m, occurs in the fertile soils of open
grassland on gentle slopes and valley floors. In contrast, the scrub
oak is an evergreen shrub, usually only 1 to 3 m tall, which often
Temporal isolation Species reproduce in
different seasons or at
forms the kind of dense scrub known as chaparral. The scrub oak
different times of the day. is found on steep slopes in less fertile soils. Hybrids between
these different oaks do occur and are fully fertile, but they are rare.

Mechanical isolation Structural differences

between species
prevent mating.

Prevention of Gametes of one species

gamete fusion function poorly with the
gametes of another species
or within the reproductive
tract of another species.

POSTZYGOTIC
I SO L ATI N G M E CH A N I S M S

Hybrid inviability Hybrid embryos do not Figure 22.2 Lions and tigers are ecologically isolated.
or infertility develop properly, hybrid
The ranges of lions and tigers overlap in India. However, lions and
adults do not survive in
tigers do not hybridize in the wild because they utilize different
nature, or hybrid adults
are sterile or have portions of the habitat. Hybrids, such as this tiglon, have been
reduced fertility. successfully produced in captivity, but hybridization does not occur in
the wild. Alexander Bayburov/Shutterstock

chapter 22 The Origin of Species 465

 Chrysoperla
plorabunda

Amplitude (dB)
Chrysoperla
adamsi

Chrysoperla
johnsoni

0 1 2 3 4 5 6 7 8 9 10 11 12
Time (seconds)

Figure 22.3 Differences in courtship rituals can isolate

related bird species. These Gala ́pagos blue-footed boobies select Figure 22.4 Differences in courtship song of sympatric
their mates only after an elaborate courtship display. This male is species of lacewings. Lacewings are small insects that rely on
lifting his feet in a ritualized high-step that shows off his bright blue signals produced by moving their abdomens to vibrate the surface
feet. The display behavior of the two other species of boobies that on which they are sitting to attract mates. As these recordings
occur in the Gala ́pagos is very different, as is the color of their feet. indicate, the vibration patterns produced by sympatric species differ
Rene Baars/Shutterstock greatly. Females, which detect the calls as they are transmitted
through solid surfaces such as branches, are able to distinguish calls
The sharply distinct habitats of their parents limit their occurrence of different species and respond only to individuals producing their
together, and there is little intermediate habitat where the hybrids own species’ call.
might flourish.

Behavioral isolation Some species even use electroreception. African and South
Chapter 53 describes the often elaborate courtship and mating ritu- Asian electric fish independently have evolved specialized organs
als of some groups of animals. Related species of organisms such in their tails that produce electrical discharges and electroreceptors
as birds often differ in their courtship rituals, which tends to keep on their skins to detect them. These discharges are used to com-
these species distinct in nature even if they inhabit the same places municate in social interactions; field experiments indicate that
(figure 22.3). Sympatric species avoid mating with members of the males can distinguish between signals produced by their own and
wrong species in a variety of ways; every mode of communication other, co-occurring species, probably on the basis of differences in
imaginable appears to be used by some species. Differences in vi- the timing of the electrical pulses.
sual signals, as just discussed, are common; however, other types
of animals rely more on other sensory modes for communication. Temporal isolation
Many species, such as frogs, birds, and a variety of insects, use Two species of wild lettuce, Lactuca graminifolia and L. cana­
sound to attract mates. Predictably, sympatric species of these ani- densis, grow together along roadsides throughout the southeastern
mals produce different calls. Similarly, the “songs” of lacewings United States. Hybrids between these two species are easily made
are produced when they vibrate their abdomens against the surface experimentally and are completely fertile. But these hybrids are
on which they are sitting, and sympatric species produce different rare in nature because L. graminifolia flowers in early spring and
vibration patterns (figure 22.4). L. canadensis flowers in summer. When their blooming periods
Other species rely on the detection of chemical signals, overlap, as happens occasionally, the two species do form hybrids,
called pheromones. The use of pheromones in moths has been which may become locally abundant.
particularly well studied. When female moths are ready to mate, Many species of closely related amphibians have different
they emit a pheromone that males can detect at great distances. breeding seasons that prevent hybridization. For example, five spe-
Sympatric species differ in the pheromone they produce: either cies of frogs of the genus Rana occur together in most of the east-
they use different chemical compounds, or, if they are using the ern United States, but hybrids are rare because the peak breeding
same compounds, the proportions used are different. Laboratory time is different for each of them.
studies indicate that males are remarkably adept at distinguishing
the pheromones of their own species from those of other species or Mechanical isolation
even from synthetic compounds similar, but not identical, to that of Structural differences prevent mating between some related spe-
their own species. cies of animals. Aside from such obvious features as size, the

466 part IV Evolution

structures of the male and female copulatory organs may be in-
compatible. In many insect and other arthropod groups, the sexual
organs, particularly those of the male, are so diverse that they are 1
used as a primary basis for distinguishing species.
Similarly, flowers of related species of plants often differ
significantly in their proportions and structures. Some of these dif-
ferences limit the transfer of pollen from one plant species to an-
other. For example, bees may carry the pollen of one species on a
certain place on their bodies; if this area does not come into ­contact 2 3
with the receptive structures of the flowers of another plant spe-
cies, the pollen is not transferred.

Prevention of gamete fusion

In animals that shed gametes directly into water, the eggs and
sperm derived from different species may not attract or fuse with
4
one another. Many animals that have internal fertilization may not
hybridize successfully because the sperm of one species functions
so poorly within the reproductive tract of another that fertilization
never takes place. In plants, the growth of pollen tubes may be
impeded in hybrids between different species. In both plants and 1. Rana pipiens
2. Rana blairi
animals, isolating mechanisms such as these prevent the union of 3. Rana sphenocephala
gametes, even following successful mating. Figure 22.5 Postzygotic 4. Rana berlandieri

isolation in leopard frogs.

Postzygotic isolating mechanisms prevent These four species resemble one another
normal development into reproducing adults closely in their external features. Their status as separate species first
was suspected when hybrids between some pairs of these species were
All of the factors we have discussed so far tend to prevent hybrid- found to produce defective embryos in the laboratory. Subsequent
ization. If hybrid matings do occur and zygotes are produced, research revealed that the mating calls of the four species differ
many factors may still prevent those zygotes from developing into substantially, indicating that the species have both pre- and
normally functioning, fertile individuals. postzygotic isolating mechanisms.
As you saw in chapter 19, development is a complex pro-
cess. In hybrids, the genetic complements of two species may be
so different that they cannot function together normally in em-
Inquiry question Can more than one type of reproductive
bryonic development. For example, hybridization between
sheep and goats usually produces embryos that die in the earliest
? isolating mechanism be operating between two species?
developmental stages.
The leopard frogs (Rana pipiens complex) of the eastern The biological species concept
United States are a group of similar species, assumed for a long
time to constitute a single species (figure 22.5). Careful examina-
does not explain all observations
tion, however, revealed that although the frogs appear similar, suc- The biological species concept has proved to be an effective way of
cessful mating between them is rare because of problems that occur understanding the existence of species in nature. Nonetheless, it
as the fertilized eggs develop. Many of the hybrid combinations fails to take into account all observations, leading some biologists
cannot be produced even in the laboratory. Examples of this kind, to propose alternative species concepts.
in which the existence of multiple species has been recognized only One criticism of the biological species concept concerns the
as a result of hybridization experiments, are common in plants. extent to which all species truly are reproductively isolated. By
Even when hybrids survive the embryo stage, they may still definition, under the biological species concept, species should not
not develop normally. If the hybrids are less physically fit than their interbreed and produce fertile offspring. But in recent years, biolo-
parents, they will almost certainly be eliminated in nature. Even if a gists have detected much greater amounts of interspecific (i.e., be-
hybrid is vigorous and strong, as in the case of the mule, which is a tween different species) hybridization than was previously thought
hybrid between a female horse and a male donkey, it may still be to occur between populations that seem to coexist as distinct bio-
sterile and thus incapable of contributing to succeeding generations. logical entities.
Hybrids may be sterile because the development of sex or- Botanists have always been aware that plant species often un-
gans is abnormal, because the chromosomes derived from the re- dergo substantial amounts of hybridization. More than 50% of Cali-
spective parents cannot pair properly during meiosis, or due to a fornia plant species included in one study, for example, were not
variety of other causes. In the case of the mule, for example, the well defined by genetic isolation. This coexistence without genetic
parental species have different numbers of chromosomes, and as a isolation can be long-lasting: fossil data show that balsam poplars
result, the mule’s chromosomes cannot align properly during mei- and cottonwoods have been phenotypically distinct for 12 million
osis, thus rendering the animal infertile. years, but they also have routinely produced hybrids throughout this

chapter 22 The Origin of Species 467

time. Consequently, many botanists have long felt that the biologi- reproduce without mating. Reproductive isolation therefore has no
cal species concept is misnamed and applies only to animals. meaning for such organisms.
New evidence, however, increasingly indicates that hybrid- For these reasons, a variety of other ideas have been put for-
ization is not all that uncommon in animals, either. In recent years, ward to establish criteria for defining species. Many of these are
many cases of substantial hybridization between animal species specific to a particular type of organism, and none has universal
have been documented. One survey indicated that almost 10% of applicability. In reality, there may be no single explanation for
the world’s more than 10,000 bird species are known to have what maintains the identity of species. Given the incredible varia-
hybridized in nature. tion evident in plants, animals, and microorganisms in all aspects
The Galápagos finches provide a particularly well-­studied of their biology, it would not be surprising to find that different
example. Three species on the island of Daphne Major—the processes are operating in different organisms.
medium­ground finch, the cactus finch, and the small ground In addition, some scientists have turned from emphasizing
finch—are clearly distinct morphologically, and they occupy dif- the processes that maintain species distinctions to examining the
ferent ecological niches. Studies over the past 30 years by Peter and evolutionary history of populations. These genealogical species
Rosemary Grant found that, on average, 2% of the medium ground concepts are currently a topic of great debate and are discussed
finches and 1% of the cactus ground finches mated with other spe- further in chapter 23.
cies every year. Furthermore, hybrid offspring appeared to be at no
disadvantage in terms of survival or subsequent reproduction. This
is not a trivial amount of genetic exchange, and one might expect to Learning Outcomes Review 22.1
see the species coalesce into one genetically variable popula- Species are populations of organisms that are distinct from other,
tion—but the species are maintaining their distinctiveness. co-occurring species, and are interconnected geographically.
Hybridization is not rampant throughout the animal world, The biological species concept therefore defines species based
on their ability to interbreed. Reproductive isolating mechanisms
however. Most bird species do not hybridize, and probably even
prevent successful interbreeding between species. An alternative
fewer experience significant amounts of hybridization. Still, hybrid-
approach emphasizes the role of adaptation and natural selection
ization is common enough to cast doubt on whether reproductive as a force for maintaining separation of species.
isolation is the only force maintaining the integrity of species.
■■ How does the ability to exchange genes explain why
Natural selection and the maintenance of species’ sympatric species remain distinct and geographic
differences populations of one species remain connected?
An alternative species concept, the ecological species concept, ■■ How could natural selection explain this phenomenon?
proposes that the distinctions among species are maintained by
natural selection. The idea is that each species has adapted to its
own specific part of the environment. Stabilizing selection, de-
scribed in chapter 20, then maintains the species’ different adapta-
tions. Hybridization has little ­effect because alleles introduced into 22.2 Natural Selection and
one species’ gene pool from other species are quickly eliminated Reproductive Isolation
by natural selection.
You probably recall from chapter 20 that the interaction be-
tween gene flow and natural selection can have many outcomes. In
some cases, strong selection can overwhelm any effects of gene Learning Outcomes
flow—this is how natural selection can maintain species’ differ- 1. Define reinforcement in the context of reproductive isolation.
ences in the face of gene flow. But, in other situations, gene flow 2. Explain the possible outcomes when two populations that are
can prevent natural selection from eliminating less successful al- partially reproductively isolated become sympatric.
leles from a population; in this case, the biological species concept
would be applicable.
One of the oldest questions in the field of evolution is: How does
Other weaknesses of the biological species concept one ancestral species become divided into two descendant species?
If species are defined by the existence of reproductive isolation,
The biological species concept has been criticized for other reasons
then the process of speciation is identical to the evolution of repro-
as well. For example, it can be difficult to apply the concept to
ductive isolating mechanisms.
populations that are geographically separated in nature. Because
individuals of these populations do not encounter each other, it is
not possible to observe whether they would interbreed naturally.
Selection may reinforce isolating mechanisms
Although experiments can determine whether fertile hybrids The formation of species is a continuous process, and as a result,
can be produced, this information is not enough. Many species that two populations may have gone only part of the way toward be-
coexist without interbreeding in nature will readily hybridize in the coming different species and thus be only partially reproductively
artificial settings of the laboratory or zoo. Consequently, evaluat- isolated. For example, because of behavioral or ecological differ-
ing whether such populations constitute different species is ulti- ences, individuals of two populations may be more likely to mate
mately a judgment call. In addition, the concept is more limited with members of their own population, but ­between-population
than its name would imply. Many organisms are asexual and matings may still occur. If mating occurs and fertilization produces

468 part IV Evolution

a zygote, postzygotic barriers may also be incomplete: develop- Pointed phlox
mental problems may result in lower embryo survival or reduced Drummond’s phlox
in sympatry
fertility, but some individuals may survive and ­reproduce.
Drummond’s phlox
in allopatry
How reinforcement can complete the speciation process
What happens when two populations come into contact thus de-
pends on the extent to which isolating mechanisms have already
evolved. If isolating mechanisms have not evolved at all, then the
two populations will interbreed freely, and whatever differences
have evolved between them should disappear over the course of
time as genetic exchange homogenizes the populations. Converse-
ly, if the populations are completely reproductively isolated, then
no genetic exchange will occur and the two populations will re-
main different species.
More interesting is the situation in which hybrids are pro-
duced that are either only partly sterile or not as well adapted to the
existing habitats as their parents. Selection would favor any alleles
in the parental populations that prevented hybridization, because
individuals that did not engage in hybridization would produce
more successful offspring.
The result would be that selection would improve prezygotic
isolating mechanisms until the two populations were completely Figure 22.6 Reinforcement in phlox. Drummond’s phlox,
reproductively isolated. This process is termed r­ einforcement be- Phlox drummondii, is light blue with a pink tinge throughout most
cause initially incomplete isolating mechanisms are reinforced by of its range (bottom left). However, where it is sympatric with the
natural selection until they are completely effective. pointed phlox, D. cuspidata (upper right), it is deep red in color
An example of reinforcement is provided by wildflowers (lower right), making it easier for the butterflies that pollinate them to
called phlox (figure 22.6). Most phlox flowers are a pale blue col- distinguish between the two species. ©Robin Hopkins
or, sometimes with a pink tinge. Drummond’s phlox, a beautiful
species that grows profusely along highways, is no exception. The
one exception is in certain parts of Texas, where it is sympatric How gene flow may counter speciation
with another species, the pointed phlox. When the two species are Reinforcement is not inevitable, however. When incompletely iso-
found together, the pointed phlox maintains the pale blue color- lated populations come together, gene flow immediately begins to
ation, but Drummond’s phlox is a deep shade of red. occur between them. Although hybrids may be inferior, they are not
How has this color difference evolved, and why? The color of completely inviable or infertile—if they were, the species would
these flowers is a result of red and blue color pigments, produced already be completely reproductively isolated. When the surviving
by different genes and leading to the light blue, slightly pinkish hybrids reproduce with members of either population, they will
color, and another gene that affects the amount of all pigments pro- serve as a conduit of genetic exchange from one population to the
duced, affecting the intensity of the color. In sympatric populations other, and the two populations will tend to lose their genetic distinc-
of Drummond’s phlox, an allele that causes an increase in pigment tiveness. Thus, a race ensues: can complete reproductive isolation
production has replaced the normal allele. By itself, this change evolve before gene flow erases the differences between the popula-
would lead to increased production of both red and blue pigments, tions? The outcome depends on the initial conditions and the par-
leading to a much deeper shade of the same color. However, a new ticular natural history of the species involved, but many experts
allele at a second gene has knocked out production of the blue pig- consider reinforcement to be the less common outcome.
ments, leaving the red pigment to predominate. The result is the
deep red color of the sympatric Drummond’s phlox flowers.
But why has this change occurred? The answer lies in the Learning Outcomes Review 22.2
interaction between the two phlox species. The flowers are able Natural selection may favor the evolution of increased prezygotic
to interbreed, but their offspring have very low fertility. Conse- reproductive isolation between sympatric populations when
quently, any factor that would diminish interbreeding would be postzygotic isolation initially exists and prezygotic isolation is
favored by natural selection. And that’s just what color does. only partial. This phenomenon is termed reinforcement, and it
Both species are pollinated by butterflies that go from one flow- may lead to the complete reproductive isolation of populations.
er to another, and it turns out that individual butterflies have In contrast, however, genetic exchange between populations may
color preferences. As a result, some butterflies visit only the decrease genetic differences among populations, thus preventing
light blue flowers and others only the dark red ones, and thus speciation from occurring.
pollen from one species does not fertilize flowers of the other ■■ How might the initial degree of reproductive isolation affect
species. This example illustrates how reproductive isolation ini- the probability that reinforcement will occur when two
tially resulting from postzygotic infertility can be reinforced by populations come into sympatry?
the evolution of premating isolation.

chapter 22 The Origin of Species 469

 cases like this, contrary to what occurs during reinforcement,
22.3 The Role of Genetic Drift natural selection isn’t acting to favor those individuals that avoid
hybridizing with the other species. Rather, increased reproductive
and Natural Selection in isolation evolves as an incidental consequence of the evolution of
Speciation different adaptations in the different populations. For example, if
one population of flies adapts to wet habitats and another to dry
areas, then natural selection will favor a variety of corresponding
differences in physiological and sensory traits. These differences
Learning Outcomes may produce ecological and behavioral isolation and may cause
any hybrids the two populations produce to be poorly adapted to
1. Describe the effects of genetic drift on a population.
either habitat.
2. Explain how genetic drift and natural selection can lead to
Selection on mating behavior might also incidentally lead
speciation.
to the evolution of reproductive isolation. Male Anolis lizards,
for example, court females by extending a colorful flap of skin,
What role does natural selection play in the speciation process? called a dewlap, located under their throats (­figure 22.7). The
Certainly, the process of reinforcement is driven by natural selec- ability of one lizard to see the dewlap of another lizard depends
tion, favoring the evolution of complete reproductive isolation. But not only on the color of the dewlap, but on the environment in
reinforcement may not be common. In situations other than rein- which the lizards live. A light-colored dewlap is most effective
forcement, does natural selection play a role in the evolution of in reflecting light in a dim forest, whereas dark colors are more
reproductive isolating mechanisms? apparent in the bright glare of open habitats. As a result, when
these lizards occupy new habitats, natural selection favors evolu-
Random changes may cause tionary change in dewlap color because males whose dewlaps
cannot be seen will not attract many mates. But the lizards also
reproductive isolation distinguish members of their own species from other species by
As mentioned in chapter 20, populations may diverge as a result of the color of the dewlap. Adaptive change in mating signals in
genetic drift. Random change in small populations, founder ef- new environments to enhance visibility to conspecifics could
fects, and population bottlenecks all may lead to changes in traits therefore produce reproductive isolation from populations in the
that cause reproductive isolation. ancestral environment; if individuals from the two populations
For example, in the Hawaiian Islands, closely related species ever came into contact, because of differences in their dewlap
of Drosophila often differ greatly in their courtship behavior. Col- colors, they might not recognize each other as members of the
onization of new islands by these fruit flies probably involved a same species.
founder effect, in which one or a few flies—perhaps only a single Laboratory scientists have conducted experiments on fruit
pregnant female—were blown by strong winds to the new island. flies and other fast-reproducing organisms in which they isolate
Changes in courtship behavior between ancestor and descendant populations in different laboratory chambers and measure how
populations may be the result of such founder events. much reproductive isolation evolves. These experiments indi-
Given enough time, any two isolated populations will diverge cate that genetic drift by itself can lead to some degree of repro-
because of genetic drift (remember that even large populations expe- ductive isolation, but in general, reproductive isolation evolves
rience drift, but at a lower rate than in small populations). In some more rapidly when the populations are forced to adapt to differ-
cases, this random divergence may affect traits responsible for repro- ent laboratory environments (such as differences in temperature
ductive isolation, and speciation may occur. or food type). Although natural selection in the experiment
does not directly favor traits because they lead to reproductive
isolation, the incidental effect of adaptive divergence is that
Adaptation can lead to speciation populations in different environments become reproductively
Although random processes may sometimes be responsible, in isolated. For this reason, some biologists believe that the term
many cases natural selection probably plays a role in the specia- isolating mechanisms is misguided, because it implies that the
tion process. As populations of a species adapt to different cir- traits evolved specifically for the purpose of genetically isolat-
cumstances, they likely accumulate many differences that may ing a species, which in most cases—except reinforcement—is
lead to reproductive isolation. It is important to realize that in probably incorrect.

Figure 22.7 Dewlaps of different species of Caribbean Anolis lizards. Males use their dewlaps in both territorial and courtship
displays. Coexisting species almost always differ in their dewlaps, which are used in species recognition. Darker-colored dewlaps, such as those of
the two species on the left, are easier to see in open habitats, whereas lighter-colored dewlaps, like those of the two species on the right, are more
visible in shaded environments. Jonathan Losos
a. b. c.

Figure 22.8 Populations can become geographically isolated for a variety of reasons. a. Colonization of remote areas by one
or a few individuals can establish populations in a distant place. b. Barriers to movement can split an ancestral population into two isolated
populations. c. Extinction of intermediate populations can leave the remaining populations isolated from one another.

Allopatric speciation takes place when

Learning Outcomes Review 22.3 populations are geographically isolated
Genetic drift refers to randomly generated changes in a
population’s genetic makeup. Isolated populations will eventually Ernst Mayr was the first biologist to demonstrate that geo­­
diverge because of genetic drift. Adaptation to different graphically separated, or allopatric, populations appear much
environments may also lead to the reproductive isolation of more likely to have evolved substantial differences leading to
populations from each other and, in general, is a more potent speciation. Marshalling data from a wide variety of organisms
force producing reproduction isolation. and localities, Mayr made a strong case for allopatric speciation
■■ How is the evolution of reproductive isolation in populations as the primary means of speciation.
adapting to different environments different from the For example, the little paradise kingfisher varies little
process of reinforcement? throughout its wide range in New Guinea, despite the great
variation in the island’s topography and climate. By contrast,
isolated populations on nearby islands are strikingly different
from one another and from the mainland population (figure 22.9).
Thus, geographic isolation seems to have been an important pre-
22.4 The Geography requisite for the evolution of differences between populations.
Many other examples indicate that speciation can occur un-
of Speciation der allopatric conditions. Because we would expect isolated popu-
lations to diverge over time, by either drift or selection, this result
is not surprising. Rather, the more intriguing question becomes: Is
geographic isolation required for speciation to occur?
Learning Outcomes
1. Compare and contrast sympatric and allopatric speciation. Sympatric speciation occurs without
2. Explain the conditions required for sympatric speciation to geographic separation
occur.
For decades, biologists have debated whether one species can split
into two at a single locality, without the two new species ever
Speciation is a two-part process. First, initially identical popula- having been geographically separated. Investigators have suggested
tions must diverge, and second, reproductive isolation must evolve that this sympatric speciation could occur either instantaneously or
to maintain these differences. The difficulty with this process, as over the course of multiple generations. Although most of the
we have seen, is that the homogenizing effect of gene flow between hypotheses suggested so far are highly controversial, one type of
populations is constantly acting to erase any differences that may instantaneous sympatric speciation is known to occur commonly,
arise by genetic drift or natural selection. Gene flow occurs only as the result of polyploidy.
between populations that are in contact, however, and populations
can become geographically isolated for a variety of reasons Instantaneous speciation through polyploidy
(figure 22.8). Consequently, evolutionary biologists have long Instantaneous sympatric speciation occurs when an individual is
recognized that speciation is much more likely in geographically born that is reproductively isolated from all other members of its
isolated populations. species. In most cases, a mutation that would cause an individual

chapter 22 The Origin of Species 471

Isolated island Isolated island including many of great commercial importance, such as bread
population population wheat, cotton, tobacco, sugarcane, bananas, and potatoes. Specia-
of kingfishers of kingfishers
tion by polyploidy is also known to occur in a variety of animals,
including insects, fish, and salamanders, although much more
rarely than in plants.
Mainland
PACIFIC
population Sympatric speciation by disruptive selection
OCEAN
of kingfishers
Some investigators believe that sympatric speciation can occur over
the course of multiple generations through the process of disruptive
selection. As noted in chapter 20, disruptive selection can cause a pop-
ulation to contain individuals exhibiting two different phenotypes.
NEW One might think that if selection were strong enough, these
GUINEA
two phenotypes would evolve over a number of generations into

Species 2 Species 1

Generation
Isolated island Mainland Mainland

Parent
population population population
of kingfishers of kingfishers of kingfishers

2n=4 2n=6
Figure 22.9 Phenotypic differentiation in the little
paradise kingfisher, Tanysiptera hydrocharis, in
New Guinea. Isolated island populations (left) are quite distinctive,
showing variation in tail feather structure and length, plumage Gametes

coloration, and bill size, whereas kingfishers on the mainland

(right) show little variation. n=2 n=3

to be greatly different from others of its species would have many

adverse side effects due to the many pleiotropic effects of most F1 Generation: Hybrid Offspring
genes (see chapter 12), and the individual likely would not survive.
One exception often seen in plants, however, occurs through the
process of ­polyploidy, which ­produces individuals that have more
than two sets of chromosomes.
Polyploid individuals can arise in two ways. In autopoly-
ploidy, all of the chromosomes come from a single species. This Doubling of No doubling of
might happen, for example, due to an error in cell division that chromosome number chromosome number
causes a doubling of chromosomes. Such individuals, termed tetra­
ploids because they have four sets of chromosomes, can self-fertilize
or mate with other tetraploids, but cannot mate and produce fertile
offspring with normal diploids. The reason is that the tetraploid spe-
cies produce diploid gametes that produce triploid offspring (having
2n=10
three sets of chromosomes) when combined with haploid gametes
Pairing now possible during meiosis Chromosomes either
from normal diploids. Triploids are sterile because the odd number cannot pair or go
of chromosomes prevents proper pairing during meiosis. through erratic meiosis
A more common type of polyploid speciation is
allopolyploidy, which may happen when two species hybridize
(figure 22.10). The resulting offspring, having one copy of the
chromosomes of each species, is usually infertile because the chro-
n=5 No gametes, or sterile
mosomes do not pair correctly in meiosis. However, such individu-
Viable gametes – sexual reproduction gametes – no sexual
als are often otherwise healthy. Sometimes, the chromosomes of possible with other tetraploid reproduction possible
such an individual spontaneously double, as just described for
autopolyploidy. Consequently, the resulting tetraploid has two cop-
ies of each set of chromosomes and pairing during meiosis is no Figure 22.10 Allopolyploid speciation. Hybrid offspring
longer a problem. As a result, such tetraploids are able to interbreed from parents with different numbers of chromosomes often cannot
with other similar tetraploids, and a new species has been created. reproduce sexually. Sometimes, the number of chromosomes in such
It is estimated that about half of the approximately 260,000 hybrids doubles to produce a tetraploid individual that can undergo
species of plants have a polyploid episode in their history, meiosis and reproduce with similar tetraploid individuals.

472 part IV Evolution

different species. But before the two phenotypes could become dif- and insects, both of which allowed descendant species to diversify
ferent species, they would have to evolve reproductive isolating and adapt to many newly available parts of the environment.
mechanisms. Initially, the two phenotypes would not be reproduc- Adaptive radiation requires both speciation and adaptation
tively isolated at all, and genetic exchange between individuals of to different habitats. A classic model postulates that a species colo-
the two phenotypes would tend to prevent genetic divergence in nizes multiple islands in an archipelago. Speciation subsequently
mating preferences or other isolating mechanisms. As a result, the occurs allopatrically, and then the newly arisen species colonize
two phenotypes would be retained as polymorphisms within a sin- other islands, producing multiple species per island (figure 22.11).
gle population. For this reason, most biologists consider sympatric
speciation of this type to be a rare event.
In recent years, however, a number of cases have appeared that
are difficult to interpret in any way other than as sympatric speciation. 1. An ancestral
species flies
An example occurs on Lord Howe island, a small (16 km2) island, from mainland
remotely located in the Pacific Ocean 600 km from the nearest large to colonize
landmass, Australia. The palm genus Howea contains two species, one island.
both found on Lord Howe and adapted to living on different types of
soil. Given the small size of the island and that the pollen of these
trees is dispersed by the wind, little opportunity would have existed 2. The ancestral
for divergence of two populations in isolation from each other. The species spreads
to different
most reasonable explanation is that the ancestral Howea species colo-
islands.
nized the island and subsequently underwent sympatric speciation.

Learning Outcomes Review 22.4

3. Populations on
Sympatric speciation occurs without geographic separation, different islands
whereas allopatric speciation occurs in geographically isolated evolve to become
populations. Polyploidy and disruptive selection are two ways by different species.
which a species may undergo sympatric speciation.
■■ How do polyploidy and disruptive selection differ as ways in
which sympatric speciation can occur? or
a. b.

22.5 Adaptive Radiation and

Biological Diversity
4. Species evolve different 4. Colonization of islands.
adaptations in allopatry.

Learning Outcomes
1. Describe adaptive radiation.
2. List conditions that may lead to adaptive radiation.

One of the most visible manifestations of evolution is the existence

of groups of closely related species that have recently evolved from
5. Colonization of islands. 5. Species evolve different adap-
a common ancestor and have adapted to many different parts of the tations to minimize competition
environment. These adaptive radiations are particularly common with other species (character
in situations in which a species occurs in an environment with few displacement).
other species and many available resources. One example is the
creation of new islands through volcanic activity, such as the Ha- Figure 22.11 Classic model of adaptive radiation on
waiian and Galápagos Islands. Another example is a catastrophic island archipelagoes. (1) An ancestral species colonizes an island
event leading to the extinction of most other species, a phenome- in an archipelago. Subsequently, the population colonizes other islands
non termed mass extinction that we discuss in section 22.7. (2), after which the populations on the different islands speciate in allopatry
Adaptive radiation can also result when a new trait, called a (3). Then some of these new species colonize other islands, leading to
key innovation, evolves within a species, allowing it to use local communities of two or more species. Adaptive differences can
resources or other aspects of the environment that were previously evolve either when species are in allopatry in response to different
inaccessible. Classic examples of key innovation leading to adap- environmental conditions (a) or as the result of ecological interactions
tive radiation are the evolution of lungs in fish and of wings in birds between species (b) by the process of character displacement.

chapter 22 The Origin of Species 473

 Adaptation to new habitats can occur either during the habitats throughout the lake and evolved to be different from their
allopatric phase, as the species respond to different environments marine ancestors. So different, in fact, that in the few lakes in which
on the different islands, or after two species become sympatric. In a second colonization occurred, the two stickleback populations had
the latter case, this adaptation may be driven by selective pressures evolved a high degree of reproductive isolation. As a result, natural
to minimize competition for available resources with other species. selection led to the evolution of differences in the two populations to
In this process, termed character ­displacement, two reproduc- minimize competition for food. One stickleback adapted to using
tively isolated but ecologically similar species come into contact. open water by evolving a streamlined body form for greater speed
Because the two species use the same resources, natural selection and larger gill structures to more effectively strain out the zooplank-
in each species favors those individuals that use resources not used ton that occur in the open. At the same time, the other population
by the other species. Because those individuals will have greater adapted to foraging near the margins and bottom of the lake, evolving
fitness, whatever traits cause the differences in resource use will a larger and stouter body and altering the shape of the mouth to better
increase in frequency (assuming that a genetic basis exists for these feed on large invertebrates from the lake floor (figure 22.13). Similar
differences), and, over time, the species will diverge (figure 22.12). character displacement occurred in seven doubly colonized lakes.
One example of character displacement involves three-spined An alternative possibility is that adaptive radiation occurs
sticklebacks, small fish found in temperate lakes throughout the through repeated instances of sympatric speciation, producing a
northern hemisphere. In British Columbia, many lakes were created suite of species adapted to different habitats. As discussed in sec-
in the past 12,000 years when the glaciers melted at the end of the tion 22.4, such scenarios are hotly debated.
last Ice Age. Some of these lakes near the ocean have been invaded In this section, we discuss four exemplary cases of adaptive
by marine populations of the stickleback. These fish tend to occur in radiation.

Inquiry question How would the scenario for adaptive

SCIENTIFIC THINKING
? radiation differ depending on whether speciation is allopatric
Question: Does competition for resources cause character displacement? or sympatric? What is the relationship between character
Hypothesis: Competition with similar species will cause natural displacement and sympatric speciation?
selection to promote evolutionary divergence.
Experiment: Place a species of fish in a pond with another, similar
fish species and measure the form of selection. As a control, place a Hawaiian plants and animals exploited
population of the same species in a pond without the second species. Note a rich, diverse habitat
that the size of food that these fish eat is related to the size of the fish.
More than 1000 species in the fly genus Drosophila occur on the
Result: In the pond with two species, directional selection favors those
Hawaiian Islands. New species of Drosophila are still being dis-
individuals which have phenotypes most dissimilar from the other
covered in Hawaii, although the rapid destruction of the native
species, and thus are most different in resource use. Directional vegetation is making the search more difficult.
selection does not occur in the control population.

species 1 species 2
Frequency

Frequency

Displacement

Body size Body size

a. b.
Interpretation: Would you expect character displacement to occur if
resources were unlimited?

Figure 22.12 Character displacement. a. Two species are

initially similar and thus overlap greatly in resource use, as might
happen if the two species were similar in size (in many species, body
size and food size are closely related). Individuals in each species that
are most different from the other species (circled) will be favored by
natural selection, because they will not have to compete with the other
species. For example, the smallest individuals of one species and the Figure 22.13 Stickleback fish adapted to using different
largest of the other would not compete with the other species for food habitats. Fish living in open water are more streamlined in body
and thus would be favored. b. As a result, the species will diverge in form, whereas those that occur near the substrate are larger and bulkier,
resource use and minimize competition between the species. with a mouth better equipped to pick invertebrates off the lake bottom.

474 part IV Evolution

a. b.
Figure 22.14 Hawaiian Drosophila. The hundreds of
species that have evolved on the Hawaiian Islands are extremely
variable in appearance; note in particular the differences in head
shape and body color between these two species. a. Drosophila
heteroneura. b. Drosophila grimshawi. Kevin T. Kaneshiro
a. b.

Aside from their sheer number, Hawaiian Drosophila

s­pecies are unusual because of their incredible diversity of
­morphological and behavioral traits (figure 22.14). When their
ancestors first reached these islands, they encountered many
“empty” habitats that other kinds of insects and other animals oc-
cupied elsewhere. As a result, the species have adapted to all man-
ners of fruit fly life and include predators, parasites, and
herbivores, as well as species specialized for eating the detritus in
leaf litter and the nectar of flowers. The larvae of various species
live in rotting stems, fruits, bark, leaves, or roots, or feed on sap.
No comparable diversity of Drosophila species is found anywhere
else in the world.
The great diversity of Hawaiian species is a result of the geo-
logical history of these islands. New islands have continually c. d.
arisen from the sea in this region. As they have done so, they Figure 22.15 Hawaiian lobeliads. a. A coastal cliff-dwelling
appear to have been invaded successively by the various Drosophila species, Brighamia rockii, on the cliffs of Molokaì, Hawaiì, b. large
groups present on the older islands. New species thus have evolved tree form of Cyanea hamatiflora from Maui, c. small tree of Cyanea
as new islands have been colonized. koolauensis from Oahu, and d. bog habitat species Lobelia villosa
In addition, the Hawaiian Islands are among the most volca- from Kauai. ©K. R. Wood/NTBG
nically active islands in the world. Periodic lava flows often have
created patches of habitat within an island surrounded by a “sea”
of barren rock. These land islands are termed kipukas. Drosophila
populations isolated in these kipukas often undergo speciation. In
different sets of selective pressures. Under these circumstances,
these ways, rampant speciation combined with ecological opportu-
and aided by the geographic isolation afforded by the many islands
nity has led to an unparalleled diversity of insect life.
of the Galaˊpagos archipelago, the ancestral finches rapidly split
Similar patterns are shown by many other animal and plant
into a series of diverse populations, some of which evolved into
groups. For example, lobeliads are a group of 125 species that
separate species. These species now occupy many different habi-
have radiated to produce plants ranging in growth form from vines
tats on the Galápagos Islands, which are comparable to the habitats
and shrubs to tall trees and that live in habitats as diverse as cliff-
several distinct groups of birds occupy on the mainland. As illus-
sides, high-elevation bogs, rainforests, deserts, and coastlines
trated in figure 22.16, the 14 species fall into four groups:
(figure 22.15).
1. Ground and cactus finches. There are six species of
Darwin’s finch species adapted Geospiza ground finches. Most of the ground finches
feed on seeds. The size of their bills is related to the size
to use different food types of the seeds they eat. Some of the ground finches feed
The diversity of Darwin’s finches on the Galápagos Islands was primarily on cactus flowers and fruits, and they have a
first mentioned in chapter 21. Presumably, the ancestor of Darwin’s longer, larger, and more pointed bill than the others.
finches reached these islands before other land birds, and as a 2. Tree finches. There are five species of insect-eating tree
result many of the types of habitats that other types of birds use on finches. Four species have bills suitable for feeding on
the mainland were unoccupied. insects. The woodpecker finch has a chisel-like beak.
As the new arrivals moved into these vacant ecological nich- This unusual bird carries around a twig or a cactus spine,
es and adopted new lifestyles, they were subjected to many which it uses to probe for insects in deep crevices.

chapter 22 The Origin of Species 475

 Vegetarian Warbler
Ground and Cactus Finches Tree Finches
Tree Finch Finches

Geospiza Geospiza Geospiza Cactospiza Camarhynchus Camarhynchus Certhidea

magnirostris scandens fortis heliobates psittacula pauper fusca

Geospiza Geospiza Geospiza Cactospiza Camarhynchus Platyspiza Certhidea

conirostris fuliginosa difficilis pallida parvulus crassirostris olivacea

Figure 22.16 An evolutionary tree of Darwin’s finches. This evolutionary tree, derived from the complete genome sequences of
Darwin’s finches, indicates that warbler finches are an early offshoot. Ground and tree finches subsequently diverged, and then species within
each group specialized to use different resources. Analysis of these genomic data suggested that for several species, such as G. difficilis, what we
recognize as a single species that occurs on different islands may instead represent several different species, and these species may not be closely
related to each other. Further study is needed to examine this hypothesis.

3. Vegetarian tree finch. The very heavy bill of this species genes affecting these aspects of beak growth: Bmp4, CaM,
is used to wrench buds from branches. TGFβIIr, β-catenin, and Dkk3. In addition, two other genes, ALX1
4. Warbler finches. These unusual birds play the same and HMGA2, have been shown to regulate the activity of other
ecological role in the Gala ́pagos woods that warblers genes that affect beak shape and size. Genetic changes in different
play on the mainland, searching continuously over the combinations of these genes are responsible for the interspecific
leaves and branches for insects. They have slender, differences in adult beaks. In addition, one study documented that
warblerlike beaks. when character displacement in beak size occurred between two
ground finch species, selection acted on HMGA2 to produce the
Recently, scientists have examined the complete genome differences in beak size.
sequences of Darwin’s finches to study their evolutionary history.
These studies suggest that the deepest branches in the finch evolution- Lake Victoria cichlid fishes diversified
ary tree lead to warbler finches, which implies that warbler finches
were among the first types to evolve after colonization of the islands. very rapidly
The ground species are closely related to one another, and the same is Lake Victoria is an immense, shallow, freshwater sea about the
true for all of the tree finches. Nonetheless, within each group, spe- size of Switzerland in the heart of equatorial East Africa. Until
cies differ in beak size and other attributes, as well as in resource use. recently, the lake was home to an incredibly diverse collection of
Field studies, conducted in conjunction with those discussed over 450 species of cichlid fishes.
in chapter 21, demonstrate that ground species compete for re-
sources; the differences between species likely resulted from char- Geologically recent radiation
acter displacement as initially similar species diverged to minimize The cluster of cichlid species appears to have evolved recently and
competitive pressures. quite rapidly. By sequencing the cytochrome b gene in many of the
Recent studies are beginning to uncover the underlying ge- lake’s fish, scientists have been able to estimate that the first cich-
netic and developmental basis for Darwin’s finch diversification. lids entered Lake Victoria only 200,000 years ago.
Work to date has focused on the beaks of the ground finches. Beak Dramatic changes in water level encouraged species forma-
variation among species is a result of differential rates of growth in tion. As the lake rose, it flooded new areas and opened up new
beak height, width, and length. Researchers have identified five habitats. Many of the species may have originated after the lake

476 part IV Evolution

dried down 14,000 years ago, isolating local populations in small a second set of functioning jaws (figure 22.17a). This trait occurs
lakes until the water level rose again. in many other fish, but in cichlids it is greatly enlarged. The ability
of these second jaws to manipulate and process food has freed the
Cichlid diversity oral jaws to evolve for other purposes, and the result has been the
Cichlids are small, perchlike fishes ranging from 5 to 25 centime- incredible diversity of ecological roles filled by these fish.
ters (cm) in length, and the males come in endless varieties of col- In recent years, we have begun to discover the genetic differ-
ors. The ecological and morphological diversity of these fish is ences underlying the great morphological variety of these fish. For
remarkable, particularly given the short span of time over which example, extensive genetic analyses have implicated several genes
they have evolved. in determining the size and shape of the cichlid jaw (figure 22.17b),
We can gain some sense of the vast range of types by looking including β-catenin, Bmp4, Ptch1, and Lbh. Mutations in these
at how different species eat. There are mud biters, algae scrapers, genes in the ancestral cichlid likely allowed individuals to access
leaf chewers, snail crushers, zooplankton eaters, insect eaters, food in different parts of the habitat, leading to evolutionary diver-
prawn eaters, and fish eaters. Snail shellers pounce on slow- gence and adaptive radiation.
crawling snails and spear their soft parts with long, curved teeth
before the snails can retreat into their shells. Scale scrapers rasp Abrupt extinction in the last several decades
slices of scales off other fish. There are even cichlid species that Recently, much of the cichlid diversity has disappeared. In the
are “pedophages,” eating the young of other cichlids. 1950s, the Nile perch, a large commercial fish with a voracious
Cichlid fish have a remarkable key innovation that may appetite, was introduced to Lake Victoria. Since then, it has spread
have been instrumental in their evolutionary radiation: they carry through the lake, eating its way through the cichlids.

Scale scraper
Leaf eater

Second
set of jaws

Snail eater

Fish eater

Zooplankton eater

Algae scraper

Insect eater

a.

Labeotropheus fuelleborni Maylandia zebra

Short snout Long snout

b.

Figure 22.17 Cichlid fishes of Lake Victoria. a. These fishes have evolved adaptations to use a variety of different habitats. The
enlarged second set of jaws located in the throat of these fish has provided evolutionary flexibility, allowing oral jaws to be modified in many
ways. b. A difference in two genes is responsible for a short snout in Labeotropheus fuelleborni and a long snout in Maylandia zebra.
(b left) Roberto Nistri/Alamy Stock Photo; (b right) Frank Hecker/Alamy Stock Photo

chapter 22 The Origin of Species 477

 By 1990, many of the open-water cichlid species had be-
come extinct, as well as others living in rocky shallow regions.
Over 70% of all the named Lake Victoria cichlid species had disap-
peared, as well as untold numbers of species that had yet to be de-
scribed. We will revisit the story of Lake Victoria when we discuss
conservation biology in chapter 58.

Learning Outcomes Review 22.5

Adaptive radiation occurs when a species diversifies, producing
descendant species that are adapted to use many different
parts of the environment. Adaptive radiation may occur under

Time
conditions of recurrent isolation, which increases the rate at
which speciation occurs, and by occupation of areas with few
competitors and many types of available resources, such as on
volcanic islands. The evolution of a key innovation may also allow
adaptation to parts of the environment that previously couldn’t be
utilized.
■■ In contrast to the archipelago model, how might an adaptive
radiation proceed in a case of sympatric speciation by
disruptive selection?

22.6 The Pace of Evolution a. Gradualism b. Punctuated equilibrium

Figure 22.18 Two views of the pace of macroevolution.

a. Gradualism suggests that evolutionary change occurs slowly
through time and is not linked to speciation, whereas b. punctuated
Learning Outcomes equilibrium requires that phenotypic change occurs in bursts associated
1. Define stasis and compare it to gradual evolutionary change. with speciation, separated by long periods of little or no change.
2. Explain the components of the punctuated equilibrium
hypothesis.

We have discussed the manner in which speciation may occur, but time intervals. This phenomenon is termed punctuated ­equilibrium
we haven’t yet considered the relationship between speciation and (­figure 22.18b); some have argued that these periods of rapid
the evolutionary change that occurs within a species. Two hypoth- change occurred only during the speciation process.
eses, gradualism and punctuated equilibrium, have been advanced We have seen in chapters 20 and 21 that when natural selec-
to explain the relationship. tion is strong, evolutionary change can occur rapidly, so the “punc-
tuated” part of the theory is not controversial. A more difficult
question involves the long periods of stasis (the equilibrium): Why
Gradualism is the accumulation would species exist for thousands, or even millions, of years with-
of small changes out changing?
For more than a century after the publication of On the Origin of Although a number of possible reasons have been suggested,
Species, the standard view was that evolution occurred very slowly. most researchers now believe that a combination of stabilizing and
Such change would be nearly imperceptible from generation to oscillating selection is responsible for stasis. If the environment
generation, but would accumulate such that, over the course of does not change over long periods of time, or if environmental
thousands and millions of years, major changes­could occur. This changes oscillate back and forth, then stasis may occur for long
view is termed gradualism (figure 22.18a). periods. One factor that may enhance this stasis is the ability of
species to shift their ranges; for example, during the ice ages, when
the global climate cooled, the geographic ranges of many species
Punctuated equilibrium is long periods shifted southward, so that the species continued to experience sim-
of stasis followed by relatively rapid change ilar environmental conditions.
An alternative possibility is that species experience long periods of
Inquiry question Why would changes in geographic
?
little or no evolutionary change (termed stasis), punctuated by
bursts of evolutionary change occurring over geologically short ranges promote evolutionary stasis?

478 part IV Evolution

Evolution may include both types of change
The fossil record shows that some well-documented groups,
22.7 Speciation and Extinction
such as African mammals, clearly have evolved gradually, not Through Time
in spurts. Other groups, such as marine bryozoa, seem to show
the irregular pattern of evolutionary change predicted by the
punctuated equilibrium model. It appears, in fact, that gradual-
ism and punctuated equilibrium are two ends of a continuum. Learning Outcomes
Although some groups have evolved solely in a gradual manner 1. Describe the pattern of species diversity through time.
and others only in a punctuated mode, many other groups show 2. Define mass extinction and identify when major mass
evidence of both gradual and punctuated episodes at different extinctions have occurred.
times in their evolutionary history.
The idea that speciation is necessarily linked to phenotypic
change has not been supported, however. On one hand, it is now Biological diversity has increased vastly over the last 600 million
clear that speciation can occur without substantial phenotypic years, but the trend has been far from consistent. After a rapid rise,
change. For example, many closely related salamander species are diversity reached a plateau for about 200 million years, but since
nearly indistinguishable. On the other hand, it is also clear that then has risen steadily. Because changes in the number of species
phenotypic change can occur within species in the absence of reflect the rate of origin of new species relative to the rate at which
speciation. existing species disappear, this long-term trend reveals that specia-
tion has, in general, surpassed extinction.
Nonetheless, speciation has not always outpaced extinction.
In particular, interspersed in the long-term increase in species
Learning Outcomes Review 22.6
diversity have been a number of sharp declines, termed mass
Gradualism is the accumulation of almost imperceptible extinctions.
changes that eventually results in major differences. Punctuated
equilibrium proposes that long periods of stasis are interrupted
(punctuated) by periods of rapid change. Evidence for both Five mass extinctions have occurred
gradualism and punctuated equilibrium has been found in
different groups. Stasis refers to a period in which little or no in the distant past
evolutionary change occurs. Stasis may result from stabilizing or Five major mass extinctions have been identified, the most severe
oscillating selection.
one occurring at the end of the Permian period, approximately
■■ Could evolutionary change be punctuated in time (that is, 250 million years ago (mya) (figure 22.19). At that time, more than
rapid and episodic), but not linked to speciation? half of all plant and animal families and as many as 96% of all
species may have perished.

Figure 22.19
Biodiversity through
1000
time. The taxonomic diversity
of families of marine animals
has increased through time,
800 although occasional dips have
occurred. The fossil record is
Number of families

most complete for marine

600 organisms because they are
more readily fossilized than
Cretaceous
terrestrial species. Families are
400 shown, rather than species,
because many species are
Devonian known from only one specimen,
200 Ordovician thus introducing error into
Permian counts of the number of species
Triassic present at a particular point in
time. Arrows indicate the five
0
major mass extinction events.
600 500 400 300 200 100 0
Millions of years ago

chapter 22 The Origin of Species 479

 The most famous and well-studied extinction, although not do not immediately increase after an extinction pulse, but rather
as drastic, occurred at the end of the Cretaceous period (66 mya), take about 10 million years to reach their maximum. The cause of
at which time the dinosaurs (except for birds, which are one type of this delay is not clear, but it may result because it takes time for
dinosaur—see chapter 34) and a variety of other organisms went ecosystems to recover and for the processes of speciation and
extinct. Recent findings have supported the hypothesis that this adaptive diversification to begin. Consequently, species diversity
extinction event was triggered when a large asteroid slammed into may require 10 million years, or even much longer, to attain its
Earth, perhaps causing global forest fires and obscuring the Sun for previous level.
months by throwing particles into the air. The cause of other mass
extinction events is less certain. Some scientists suggest that aster- A sixth mass extinction is under way
oids may have played a role in at least some of the other mass
extinction events; other hypotheses implicate global climate change, The number of species in the world in recent times is greater than
massive volcanic eruptions, and other causes. it has ever been. Unfortunately, that number is decreasing at an
One important result of mass extinctions is that not all groups alarming rate due to human activities (see chapter 58).
of organisms are affected equally. For example, in the extinction at Some estimate that as much as one-fourth of all species will
the end of the Cretaceous, not only most dinosaurs, but also marine become extinct in the near future, a rate of extinction not seen on
and flying reptiles and ammonites (a type of mollusk) went extinct. Earth since the Cretaceous mass extinction. Moreover, the rebound
Marsupials, flowering plants, birds, and some forms of plankton in species diversity may be even slower than following previous
were greatly reduced in diversity. In contrast, turtles, crocodilians, mass extinction events because, instead of the ecologically impov-
and amphibians seemed to have been unscathed. Why some groups erished but energy-rich environment that existed after previous
were harder hit than others is not clear, but one hypothesis suggests mass extinction events, a large proportion of the world’s resources
that survivors were those animals that could shelter underground or will be taken up already by human activities, leaving few resources
in water, and that either could scavenge or required little food in the available for adaptive radiation. The current species extinction cri-
cool temperatures that resulted from the blockage of sunlight. sis is discussed in greater detail in chapter 58.
A consequence of mass extinctions is that previously domi-
nant groups may perish, thus changing the course of evolution.
This is certainly true of the Cretaceous extinction. During the Learning Outcomes Review 22.7
Cretaceous period, placental mammals were a minor group com- The number of species has increased through time, although
posed of species that were mostly no larger than a house cat. not at a constant rate. Five major extinction events have
When the dinosaurs, which had dominated the world for more substantially, although briefly, reduced the number of species.
than 100 million years, disappeared at the end of this period, the Diversity rebounds, but the recovery is not rapid, and the groups
placental mammals underwent a significant adaptive radiation. It making up that diversity are not the same as those that existed
is humbling to think that humans might never have arisen had that before the extinction event. Unfortunately, humans are currently
causing a sixth mass extinction event.
asteroid not struck Earth 65 mya.
As the world around us illustrates today, species diversity ■■ In what ways is the current mass extinction event different
does rebound after mass extinctions, but this recovery is not rapid. from those that have occurred in the past?
Examination of the fossil record indicates that rates of speciation

Chapter Review

JohnMernick/iStock/Getty Images

22.1 The Nature of Species and the Biological Populations of a species exhibit geographic variation.
Species Concept Populations that differ greatly in phenotype or ecologically are usually
connected by geographically intermediate populations.
Sympatric species inhabit the same locale but remain
distinct. The biological species concept focuses on the ability to
Most sympatric species are readily distinguishable phenotypically exchange genes.
and ecologically. Others usually can be identified by careful In the biological species concept, species are defined as those
study. populations that interbreed, or have the potential to do so, and produce

480 part IV Evolution

fertile offspring. Reproductive isolating mechanisms prevent genetic one species to divide into two, but scientists debate how commonly it
exchange between species. occurs.

Prezygotic isolating mechanisms prevent the formation 22.5 Adaptive Radiation and Biological Diversity
of a zygote.
Prezygotic isolating mechanisms prevent a viable zygote from being Adaptive radiation occurs when a species finds itself in a new or
created. These mechanisms include ecological, behavioral, temporal, suddenly changed environment with many resources and few competing
and mechanical isolation, as well as failure of gametes to unite to form a species (figure 22.11).
zygote. The evolution of a new trait that allows individuals to use previously
inaccessible parts of the environment, termed a “key innovation,” may
Postzygotic isolating mechanisms prevent normal development also trigger an adaptive radiation.
into reproducing adults.
Postzygotic isolating mechanisms prevent a zygote from developing into Hawaiian plants and animals exploited a rich, diverse
a viable and fertile individual. habitat.
At least 1000 species of Hawaiian Drosophila have been identified.
The biological species concept does not explain all Several groups of plants have also diversified to occupy many
observations. habitats.
An alternative explanation for the existence of distinct species focuses
on the role of natural selection and differences among species in their Darwin’s finch species adapted to use different
ecological requirements. food types.
In reality, no one species concept can explain all the diversity of life. Fourteen species in four genera have evolved to exploit four different
habitats based on type of food.

22.2 Natural Selection and Reproductive Isolation Lake Victoria cichlid fishes diversified very rapidly.
Cichlids underwent rapid radiation to form more than 450 species,
Selection may reinforce isolating mechanisms. although today more than 70% of these species are now
Populations may evolve complete reproductive isolation in allopatry. extinct.
If populations that have evolved only partial reproductive isolation
come into contact, natural selection can lead to increased reproductive
isolation, a process termed “reinforcement.”
22.6 The Pace of Evolution
Alternatively, genetic exchange between the populations can lead to Gradualism is the accumulation of small changes.
homogenization and thus prevent speciation from occurring. Historically, scientists took the view that speciation occurred gradually
through very small cumulative changes.
22.3 The Role of Genetic Drift and Natural Selection in
Punctuated equilibrium is long periods of stasis followed by
Speciation relatively rapid change.
Random changes may cause reproductive isolation. The punctuated equilibrium hypothesis contends that not only is
In small populations, genetic drift may cause populations to diverge; change rapid and episodic, but also it is only associated with the
the differences that evolve may cause the populations to become speciation process.
reproductively isolated.
Evolution may include both types of change.
Adaptation can lead to speciation. Scientists generally agree that evolutionary change occurs on a
Adaptation to different situations or environments may incidentally lead continuum, with gradualism and punctuated change being the
to reproductive isolation. In contrast to reinforcement, these differences extremes.
are not directly favored by natural selection because they prevent
hybridization. 22.7 Speciation and Extinction Through Time
In general, the number of species has increased through time
22.4 The Geography of Speciation (figure 22.9) (figure 22.19).

Allopatric speciation takes place when populations are

Five mass extinctions have occurred in the distant past.
geographically isolated.
Mass extinctions have led to dramatic decreases in species diversity
Allopatric, or geographically isolated, populations may evolve into
followed by recovery that takes millions of years. Five such mass
separate species because no gene flow occurs between them. Most
extinctions have occurred in the distant past due to asteroids hitting the
speciation probably occurs in allopatry.
Earth and global climate change, among other events.
Sympatric speciation occurs without geographic separation.
Sympatric speciation can occur in two ways. One is polyploidy, which A sixth mass extinction is under way.
instantly creates a new species. Disruptive selection also can cause Humans are currently causing a sixth mass extinction.

chapter 22 The Origin of Species 481

Visual Summary

Genetic drift

may evolve due to

Biological species Reproductive can be Natural

Species defined by requires
concept isolation shaped by selection

limitations lead to
can occur by affects
Altenative concepts
Prezygotic Postzygotic
isolation isolation
such as

Ecological species concept leads to

Allopatry

occurs in Geography affects Speciation may affect Genetic drift

Sympatry and extinction produce

Evolutionary
varies in decreases during Mass extinctions
diversification

can produce
Pace of evolution

Adaptive
occurs rapidly occurs slowly radiation

Punctuated
Gradualism
equilibrium

Review Questions

JohnMernick/iStock/Getty Images
U N D E R S TA N D c. prezygotic isolating mechanisms are extremely rare.
1. Prezygotic isolating mechanisms include all of the following except d. All of the choices are correct.
a. hybrid sterility. c. habitat separation. 4. Allopatric speciation
b. courtship rituals. d. seasonal reproduction. a. is less common than sympatric speciation.
2. Reproductive isolation is b. involves geographic isolation of some kind.
c. is the only kind of speciation that occurs in plants.
a. a result of individuals not mating with each other.
d. requires polyploidy.
b. a specific type of postzygotic isolating mechanism.
c. required by the biological species concept. 5. Gradualism and punctuated equilibrium are
d. None of the choices is correct. a. two ends of the continuum of the rate of evolutionary change
3. Problems with the biological species concept include the over time.
fact that b. mutually exclusive views about how all evolutionary change
takes place.
a. many species reproduce asexually.
c. mechanisms of reproductive isolation.
b. postzygotic isolating mechanisms decrease hybrid
d. None of the choices is correct.
viability.

482 part IV Evolution

 6. Prezygotic isolation c. gene flow between populations will be impossible.
a. always involves mechanisms that prevent interbreeding of d. the speciation will be more likely than if hybrids were
members of different species. completely infertile.
b. includes the death of a zygote shortly after fertilization. 2. Natural selection can
c. occurs only in plants. a. enhance the probability of speciation.
d. becomes stronger as the result of reinforcement. b. enhance reproductive isolation.
7. Speciation by allopolyploidy c. act against hybrid survival and reproduction.
a. takes a long time. d. All of the choices are correct.
b. is common in birds. 3. Hybridization between incompletely isolated populations
c. leads to reduced numbers of chromosomes. a. always leads to reinforcement due to the inferiority of
d. occurs after hybridization between two species. hybrids.
8. Adaptive radiation b. can serve as a mechanism for preserving gene flow
a. is the result of enriched uranium used in power plants. between populations, thus preventing speciation.
b. is the evolution of closely related species adapted to use c. occurs only in plants.
different parts of the environment. d. never affects rates of speciation.
c. results from genetic drift. 4. Natural selection can lead to speciation
d. is the outcome of stabilizing selection favoring the a. by causing small populations to diverge more than large
maintenance of adaptive traits. populations.
9. Leopard frogs from different geographic populations b. because the evolutionary changes that two populations acquire
of the Rana pipiens complex while adapting to different habitats may have the effect of
a. are members of a single species because they look very making them reproductively isolated.
similar to one another. c. by favoring the same evolutionary change in multiple
b. are different species shown to have pre- and postzygotic populations.
isolating mechanisms. d. by favoring intermediate phenotypes.
c. frequently interbreed to produce viable hybrids.
d. are genetically identical due to effective reproductive SYNTHESIZE
isolation. 1. Natural selection can lead to the evolution of prezygotic isolating
10. Character displacement mechanisms, but not that of postzygotic isolating mechanisms.
a. arises through competition and natural selection, favoring Explain.
divergence in resource use. 2. What role does natural selection play in the origin of species?
b. arises through competition and natural selection, favoring How might your answer depend on how you define what a
convergence in resource use. species is?
c. does not promote speciation.
d. reduced speciation rates in Galápagos finches. 3. Refer to figure 22.6. In Texas, Drummond’s phlox is a dark shade
of red where it occurs with the pale blue pointed phlox. Elsewhere,
11. During the history of life on Earth
where it occurs by itself, it is a pale pinkish-blue. In this case, there
a. there have been major extinction events. is no competition for ecological resources as in other cases of
b. species diversity has steadily increased. character divergence discussed. Why has the color of Drummond’s
c. species diversity has stayed relatively constant. phlox evolved when it is sympatric with pointed phlox?
d. extinction rates have been completely offset by speciation
rates. 4. Refer to figure 22.16. Geospiza fuliginosa and Geospiza fortis are
found in sympatry on at least one island in the Galápagos and in
allopatry on several islands in the same archipelago. Compare your
A P P LY
expectations about degree of morphological similarity of the two
1. If reinforcement is weak and hybrids are not completely infertile, species in these two contexts, given the hypothesis that competition
a. genetic divergence between populations may be overcome by for food played a large role in the adaptive radiation of this group.
gene flow. Would your expectations be the same for a pair of finch species
b. speciation will occur 100% of the time. that are not as closely related? Explain.

chapter 22 The Origin of Species 483

 23
CHAPTER

Systematics, Phylogenies, and

Comparative Biology
Chapter Contents
23.1 Systematics
23.2 Cladistics
23.3 Systematics and Classification
23.4 Phylogenetics and Comparative Biology
23.5 Phylogenetics and Disease Evolution

Imagemore Co, Ltd./Imagemore/Getty Images

Visual Outline Introduction

Evolutionary
All organisms share many biologi-
relationships cal characteristics. They are com-
posed of one or more cells, carry
out metabolism and transfer energy
depicted by with ATP, and encode hereditary in-
formation in DNA. Yet, there is also
a tremendous diversity of life, rang-
Cladistics inferred by Phylogenies ordered into Classifications ing from bacteria and amoebas to
blue whales and sequoia trees. The
photograph above shows variation
Lamprey Shark Salamander Lizard Tiger Gorilla Human
used to infer Domain
Eukarya in one particularly diverse group
Kingdom
Animalia of organisms, the beetles. For
Evolution
Phylum generations, biologists have tried to
Chordata

Subphylum
group organisms based on shared
Vertebrata
characteristics. The most mean-
ingful groupings are based on the
Class
Mammalia

Order
Rodentia
study of evolutionary relationships
Family among organisms. New methods
Sciuridae
for constructing evolutionary trees
Genus
Sciurus
and a sea of molecular sequence
Species
Sciurus
carolinensis
Sciurus
carolinensis data are leading to improved evolu-
tionary hypotheses to explain life’s
diversification.
 by new data, leading to the formation of better, more accurate sci-
23.1 Systematics entific ideas.
The study of evolutionary relationships is called systematics.
By looking at the similarities and differences between species, sys-
tematists can construct an evolutionary tree, or phylogeny, which rep-
Learning Outcomes resents a hypothesis about patterns of relationship among species.
1. Understand what a phylogeny represents.
2. Explain why phenotypic similarity does not necessarily Branching diagrams depict
indicate a close evolutionary relationship.
evolutionary relationships
Darwin envisioned that all species were descended from a single
One of the great challenges of modern science is to understand the common ancestor, and that the history of life could be depicted as a
history of ancestor–descendant relationships that unites all forms branching tree (figure 23.1). In Darwin’s view, the twigs of the tree
of life on Earth, from the earliest single-celled organisms to the represent existing species. As one works down the tree, the joining
complex organisms we see around us today. If the fossil record of twigs and branches reflects the pattern of common ancestry back
were perfect, we could trace the evolutionary history of species in time to the single common ancestor of all life. The process of de-
and examine how each arose and proliferated. However, as dis- scent with modification from common ancestry results in all species
cussed in chapter 21, the fossil record is far from complete, an- being related in this branching, hierarchical fashion, and their evolu-
swering many questions about life’s diversification, but leaving tionary history can be depicted using branching diagrams referred to
many others unsettled. as phylogenetic trees. Figure 23.1b shows how evolutionary rela-
Consequently, scientists must rely on other types of evidence tionships are depicted with a branching diagram. Humans and chim-
to establish the best hypothesis of evolutionary relationships. Bear panzees are descended from a common ancestor and are each other’s
in mind that the outcomes of such studies are hypotheses, and as closest living relative (the position of this common ancestor is indi-
such, they require further testing. All hypotheses may be disproved cated by the node labeled 1). Humans, chimps, and gorillas share an

Variations of a Cladogram

Gibbon Human Chimp Gorilla Orangutan Gibbon Orangutan Gorilla Human Chimp

1 1

2 2

Chimp
3 3
Version 1 1 Version 2
2 Human

3 Gorilla

Orangutan

Gibbon

Version 3
a. b.
Figure 23.1 Phylogenies depict evolutionary relationships. a. A drawing from one of Darwin’s notebooks, written in 1837 as he
developed his ideas that led to On the Origin of Species. Darwin viewed life as a branching process akin to a tree, with species on the twigs, and
evolutionary change represented by the branching pattern displayed by a tree as it grows. b. An example of a phylogeny. Note that these three
versions convey the same information despite the differences in arrangement of species and orientation. Humans and chimpanzees are more
closely related to each other than they are to any other living species. This is apparent because they share a common ancestor (the node labeled 1)
that was not an ancestor of other species. Similarly, humans, chimpanzees, and gorillas are more closely related to one another than any of them is
to orangutans because they share a common ancestor (node 2) that was not ancestral to orangutans. Node 3 represents the common ancestor of all
apes. At each node, the two descendants can be rotated without changing the meaning. For example, one difference between versions 1 and 2 is
that the descendants of node 2 have been rotated so that gorilla branches to the right in version 1 and to the left in version 2. However, this does
not affect the interpretation that humans and chimps are more closely related to each other than either species is to gorillas. (a) Letz/SIPA/Newscom

chapter 23 Systematics, Phylogenies, and Comparative Biology 485

older common ancestor (node 2), and all apes share a more distant
common ancestor (node 3).
One key to interpreting a phylogeny is to look at how recently
23.2 Cladistics
species share a common ancestor, rather than looking at the ar-
rangement of species across the top of the tree. If you compare the
three versions of the phylogeny of figure 23.1b, you can see that Learning Outcomes
the relationships are the same: regardless of where they are 1. Describe the difference between ancestral and derived
positioned, chimpanzees and humans are still more closely related similarities.
to each other than to any other species. 2. Explain why only shared derived characters indicate close
Moreover, even though humans are placed next to gibbons in evolutionary relationship.
version 1 of figure 23.1b, the pattern of relationships still indicates 3. Demonstrate how a cladogram is constructed.
that humans are more closely related (that is, share a more recent
common ancestor) with gorillas and orangutans than with gibbons.
Phylogenies are also sometimes displayed on their side, rather than Because phenotypic similarity may be misleading, most system-
upright (figure 23.1b, version 3), but this arrangement also does not atists no longer construct their phylogenetic hypotheses solely on
affect its interpretation. this basis. Rather, they distinguish similarity among species that is
inherited from the most recent common ancestor of an entire
Similarity may not accurately predict group, which is called ancestral, from similarity that arose more
recently—that is, is not inherited from the most recent common
evolutionary relationships ancestor of the group—and is shared only by a subset of the spe-
We might expect that the greater the time since two species di- cies; this similarity is called derived. In this approach, termed
verged from a common ancestor, the more different they would be. cladistics, only shared ­derived characters are considered infor-
Early systematists relied on this reasoning and constructed phylog- mative in determining evolutionary relationships.
enies based on overall similarity. If, in fact, s­ pecies evolved at a
constant rate, then the amount of divergence between two species The cladistic method requires that character
would be a function of how long they had been diverging, and thus
phylogenies based on degree of similarity would be accurate. As a
variation be identified as ancestral or derived
result, we might think that chimps and gorillas are more closely To employ the method of cladistics, systematists first gather data
related to each other than either is to humans. on a number of characters for all the species in the analysis. Char-
But as chapter 22 revealed, evolution can occur very rapidly acters can be any aspect of the phenotype, including morphology,
at some times and very slowly at others. Species invading new physiology, behavior, and DNA. As chapters 18 and 24 show, the
habitats are likely to experience new selective pressures and may revolution in genomics is now providing a vast body of data revo-
change greatly; those staying in the same habitats as their ancestors lutionizing our ability to identify and study character variation.
may change only a little. For this reason, rates of change may vary To be useful, the characters should exist in recognizable
markedly among species. Consequently, some will diverge more character states. For example, consider the character “tail” in ver-
than others, and as a result, similarity is not necessarily a good tebrates. This character has two states: presence in most verte-
predictor of how long it has been since two species shared a com- brates and absence in humans, apes, and some other groups, such
mon ancestor. as frogs.
A second fundamental problem exists as well: evolution is A cladistic analysis begins by determining the states for a
not always divergent. In chapter 21, we discussed convergent number of characters for each taxon (taxa are species or higher-
evolution, in which two species independently evolve the same level groups, such as genera or families) in the analysis. In the ex-
features. Often, species evolve convergently because they use simi- ample in figure 23.2, six characters are used for seven taxa. The
lar environments, in which similar adaptations are favored. As a first character, presence of jaws, is scored as absent in lamprey
result, two species that are not closely related may end up more (denoted by 0) and present in the other species (denoted with a 1).
similar to each other than they are to their close relatives. Evolu- Another character, hair, is absent in all of the taxa except the three
tionary reversal, the process in which a species re-evolves the char- mammal species.
acteristics of an ancestral species, also has this effect.
Examples of ancestral versus derived characters
The presence of hair is a shared derived feature of mammals
Learning Outcomes Review 23.1 (figure 23.2); in contrast, the presence of lungs in mammals is an an-
Systematics is the study of evolutionary relationships. cestral feature because it is also present in amphibians and reptiles
Phylogenies, or phylogenetic trees, are graphic representations (represented by a salamander and a lizard) and therefore presumably
of relationships among species. Similarity of organisms alone evolved prior to the common ancestor of mammals (figure 23.2). The
does not necessarily correlate with their relatedness because presence of lungs, therefore, does not tell us that mammal species are
evolutionary change is not constant in rate and direction.
all more closely related to one another than to reptiles or amphibians,
■■ Why might a species be most phenotypically similar to a but the shared, derived feature of hair suggests that all mammal spe-
species that is not its closest evolutionary relative? cies share a common ancestor that existed more recently than the com-
mon ancestor of mammals, amphibians, and reptiles.

486 part IV Evolution

 Traits: Jaws Lungs Amniotic Hair No Tail Bipedal Lamprey Shark Salamander Lizard Tiger Gorilla Human
Organism Membrane

Lamprey 0 0 0 0 0 0

Shark 1 0 0 0 0 0

Bipedal
Salamander 1 1 0 0 0 0
Tail loss
Lizard 1 1 1 0 0 0
Hair

Tiger 1 1 1 1 0 0 Amniotic
membrane

1 1 1 1 1 0 Lungs
Gorilla
Jaws
Human 1 1 1 1 1 1

a. b.
Figure 23.2 A cladogram. a. Morphological data for a group of seven vertebrates are tabulated. A “1” indicates possession of the derived
character state, and a “0” indicates possession of the ancestral character state (note that the derived state for character “no tail” is the absence of a tail;
for all other traits, absence of the trait is the ancestral character state). b. A tree, or cladogram, diagrams the relationships among the organisms based
on the presence of derived characters. The derived characters between the cladogram branch points are shared by all organisms above the branch
points and are not present in any below them. The outgroup (in this case, the lamprey) does not possess any of the derived characters.

To return to the question concerning the relationships of hu- Construction of a cladogram

mans, chimps, and gorillas, a number of morphological and DNA Once all characters have been polarized, systematists use this infor-
characters exist that are derived and shared by chimps and humans, mation to construct a cladogram, which depicts a hypothesis of
but not by gorillas or other great apes. These characters suggest evolutionary relationships (a cladogram is the type of phylogenetic
that chimps and humans diverged from a common ancestor tree produced by cladistic analysis). Species that share a common
(figure 23.1b, node 1) that existed more recently than the common ancestor, as indicated by the possession of shared derived charac-
ancestor of gorillas, chimps, and humans (node 2). ters, are said to belong to a clade. Clades are thus evolutionary
units and refer to a common ancestor and all of its descendants. A
Determination of ancestral versus derived derived character shared by clade members is called a synapomor-
Once the data are assembled, the first step in a cladistic analysis is phy of that clade. Figure 23.2b illustrates that a simple cladogram
to polarize the characters—that is, to determine whether particular is a nested set of clades, each characterized by its own synapomor-
character states are ancestral or derived. To polarize the character phies. For example, amniotes are a clade for which the evolution
“tail,” for example, systematists must determine which state— of an amniotic membrane is a synapomorphy. Within that clade,
presence or absence—was exhibited by the most recent common mammals are a clade, with hair as a synapomorphy, and so on.
ancestor of this group. Ancestral states are called plesiomorphies, and shared ances-
Usually, the fossils available do not include the most recent tral states are called symplesiomorphies. In contrast to synapomor-
common ancestor—or at least we cannot be confident that they do. phies, symplesiomorphies are not informative about phylogenetic
As a result, the method of outgroup comparison is used to assign relationships.
character polarity. To use this method, a species or group of spe- Let’s return to the character state “presence of a tail,” which
cies that is closely related to, but not a member of, the group under is exhibited by lampreys, sharks, salamanders, lizards, and tigers.
study is designated as the outgroup. Does this mean that tigers are more closely related to—and shared
When the group under study exhibits multiple character a more recent common ancestor with—lizards and sharks than to
states, and one of those states is exhibited by the outgroup, then apes and humans, their fellow mammals? The answer, of course, is
that state is considered to be ancestral and other states are consid- no: because symplesiomorphies reflect character states inherited
ered to be derived. However, outgroup species also evolve from from a distant ancestor, they do not imply that species exhibiting
their ­ancestors, so the outgroup species will not always exhibit the that state are closely related.
ancestral condition. For this reason, polarity assignments are most
reliable when the same character state is exhibited by several dif-
ferent outgroups. In the preceding example, a tail is generally pres-
Homoplasy complicates cladistic analysis
ent in the nearest outgroups of vertebrates. Consequently, the In real-world cases, phylogenetic studies are rarely as simple as the
presence of a tail in vertebrates is considered ancestral, and its ab- examples we have shown so far. The reason is that in some cases,
sence in some species is considered derived. the same character has evolved independently in several species.

chapter 23 Systematics, Phylogenies, and Comparative Biology 487

Salamander Frog Lizard Tiger Gorilla Human Salamander Lizard Tiger Frog Gorilla Human

Hair loss
Tail loss Amniotic
membrane
Tail loss loss
Tail loss
Hair
Hair
Amniotic Amniotic
membrane membrane

a. b.
Figure 23.3 Parsimony and homoplasy. a. The placement of frogs as closely related to salamanders requires that tail loss evolved twice,
an example of homoplasy. b. If frogs are closely related to gorillas and humans, then tail loss had to evolve only once. However, this arrangement
would require two additional evolutionary changes: frogs would have to have lost the amniotic membrane and hair (alternatively, hair could have
evolved independently in tigers and the clade of humans and gorillas; this interpretation would require two evolutionary changes, hair evolving twice,
just like the interpretation shown in the figure, in which hair evolved only once, but then was lost in frogs). Based on the principle of parsimony, the
cladogram that requires the fewest number of evolutionary changes is favored; in this case the cladogram in (a) requires four changes, whereas that in
(b) requires five, so (a) is considered the preferred hypothesis of evolutionary relationships.

Data analysis Construct a data matrix like the one in figure 23.2, showing the character states for all six species for the traits hair,
amniotic membrane, and tail.

These characters would be categorized as shared derived char- sequence data in the same manner as any other type of data: char-
acters, but they would be false signals of a close evolutionary rela- acter states are polarized by reference to the sequence of an out-
tionship. In addition, derived characters may sometimes be lost as group, and a cladogram is constructed that minimizes the amount
species within a clade re-evolve to the ancestral state. of character evolution required ­(figure 23.4).
Homoplasy refers to a shared derived character state that has
not been inherited from a common ancestor exhibiting that charac- Other phylogenetic methods work better
ter state. Homoplasy can result from convergent evolution or from
evolutionary reversal. For example, adult frogs do not have a tail.
than cladistics in some situations
Thus, absence of a tail is a synapomorphy that unites not only go- If characters evolve from one state to another at a slow rate com-
rillas and humans, but also frogs. However, frogs are not closely pared with the frequency of speciation events, then the principle of
related to gorillas and humans; they have neither an amniotic parsimony works well in reconstructing evolutionary relationships.
membrane nor hair, both of which are synapomorphies for clades In this situation, the principle’s underlying assumption—that shared
that contain gorillas and humans. derived similarity is indicative of recent common ancestry—is
In cases such as this, when there are conflicts among the usually correct. In recent years, however, systematists have real-
characters, systematists rely on the principle of parsimony, ized that some characters evolve so rapidly that the principle of
which favors the hypothesis that requires the fewest assump- parsimony may be misleading.
tions. As a result, the phylogeny that requires the fewest evolu-
tionary events is considered the best hypothesis of phylogenetic Rapid rates of evolutionary change and homoplasy
relationships (figure 23.3). In the example just stated, therefore, Of particular interest is the rate at which some parts of the genome
grouping frogs with salamanders is favored because it requires evolve. As discussed in chapter 18, some stretches of DNA do not
only one instance of homoplasy (the multiple origins of tailless- appear to have any function. As a result, mutations that occur in
ness), whereas a phylogeny in which frogs were most closely these parts of the DNA are not eliminated by natural selection, and
related to humans and gorillas would require two homoplastic thus the rate of evolution of new character states can be quite high
evolutionary events (the loss of both amniotic membranes and in these regions as a result of genetic drift.
hair in frogs). Moreover, because only four character states are possible for
The examples presented so far have all involved morphologi- any nucleotide base (A, C, G, or T), there is a high probability that
cal characters, but systematists increasingly use DNA sequence two species will independently evolve the same derived character
data to construct phylogenies because of the large number of char- state at any particular base position. If such homoplasy dominates
acters that can be obtained through sequencing. Cladistics analyzes the character data set, then the assumptions of the principle of

488 part IV Evolution

 DNA Sequence Outgroup Species B Species D Species A Species C

Site 1 2 3 4 5 6 7 8 9 10

Species A G C A T A G G C G T
8:T C

Species B A C A G C C G C A T 4:T G 10:T G 1:A G

8:T C 5:C A
Species C G C A T A G G T G T
6:C G
9:A G Homoplastic
evolutionary
Species D A C A T C G G T G G changes
2:T C Nonhomoplastic
evolutionary
Outgroup A T A T C C G T A T changes

Figure 23.4 Cladistic analysis of DNA sequence data. Sequence data are analyzed just like any other data. The most parsimonious
interpretation of the DNA sequence data requires nine evolutionary changes. Each of these changes is indicated on the phylogeny. Change in site 8
is homoplastic: species A and B independently evolved from thymine to cytosine at that site. Nonhomoplastic changes are those in which all
species that possess the trait derived it from the same common ancestor. Such traits are referred to as homologous.

parsimony are violated: similarity of species in this situation would make such estimates, the timing of one or more divergence events
more likely result from homoplasy than from inheritance from a must be confidently estimated. For example, the fossil record may
common ancestor. As a result, under these circumstances, indicate that two clades diverged from a common ancestor at a
phylogenies inferred using the cladistic method are likely to be particular time. Alternatively, the timing of separation of two
inaccurate. clades may be estimated from geological events that likely led to
their divergence, such as the rise of a mountain that now separates
the two clades. With this information, the amount of DNA diver-
? Inquiry question Why do high rates of evolutionary
change and a limited number of character states cause
gence separating two clades can be divided by the length of time
separating the two clades, which produces an estimate of the rate
problems for parsimony analyses?
of DNA divergence per unit of time (usually, per million years).
Assuming a molecular clock, this rate can then be used to date
Statistical approaches other divergence events in a cladogram.
Because evolution can sometimes proceed rapidly, systematists in Although the molecular clock appears to hold true in some
recent years have been exploring other methods based on statistical cases, in many others the data indicate that rates of evolution
approaches, such as maximum likelihood, to infer phylogenies. have not been constant through time across all branches in an
These methods start with an assumption about the rate at which evolutionary tree. For this reason, evolutionary dates derived
characters evolve and then fit the data to these models to derive the from molecular data must be treated cautiously. Recently, meth-
phylogeny that best accords (that is, “is maximally likely”) with ods have been developed to date evolutionary events without as-
these assumptions. suming that molecular evolution has been clocklike. These
One advantage of these methods is that different assump- methods hold great promise for providing more reliable estimates
tions of rate of evolution can be used for different characters. If of evolutionary timing.
some DNA characters evolve more slowly than others—for exam-
ple, because they are constrained by natural selection—then the
methods can employ different models of evolution for the different Learning Outcomes Review 23.2
characters. This approach is more effective than parsimony in deal- In cladistics, derived character states are distinguished from
ing with homoplasy when rates of evolutionary changes are high. ancestral character states, and species are grouped based
on shared derived character states. Derived characters are
The molecular clock determined from comparison to a group known to be closely
In general, cladograms such as the one in figure 23.2 indicate only related, termed an outgroup. A clade contains all descendants of
the order of evolutionary branching events; they do not contain a common ancestor. A cladogram is a hypothetical representation
information about the timing of these events. In some cases, how- of evolutionary relationships based on derived character states.
ever, branching events can be timed, either by reference to fossils, Homoplasies may give a false picture of relationships.
or by making assumptions about the rate at which characters ■■ Why is cladistics more successful at inferring phylogenetic
change. One widely used but controversial method is the molecular relationships in some cases than in others?
clock, which states that the rate of evolution of a molecule is con- ■■ Why are only shared derived, instead of all derived,
stant through time. In this model, divergence in DNA can be used characters useful in cladistics for reconstructing phylogenies?
to calculate the times at which branching events have occurred. To

chapter 23 Systematics, Phylogenies, and Comparative Biology 489

 grouped based on their phylogenetic relationships. A monophy-
23.3 Systematics and Classification letic group includes the most recent common ­ancestor of the group
and all of its descendants. By definition, a clade is a monophyletic
group. A paraphyletic group includes the most recent common
ancestor of the group, but not all its descendants, and a polyphy-
Learning Outcomes letic group does not include the most recent common ancestor of
1. Explain the taxonomic classification system. all members of the group (figure 23.6).
2. Differentiate among monophyletic, paraphyletic, and Taxonomic hierarchies are based on shared traits, and ideally
polyphyletic groups. they should reflect evolutionary relationships. Traditional taxonom-
3. Explain the meaning of the PSC and why it is controversial. ic groups, however, do not always fit well with the new understand-
ing of phylogenetic relationships. For example, birds have historically
been placed in the class Aves, and dinosaurs have been considered
Whereas systematics is the reconstruction and study of evolution- part of the class Reptilia. But recent phylogenetic advances make
ary relationships, classification refers to how we place species and clear that birds evolved from dinosaurs. The last common ancestor
higher groups—genus, family, class, and so forth—into the taxo- of all birds and a dinosaur was a meat-eating dinosaur (figure 23.6).
nomic hierarchy. Therefore, having two separate monophyletic groups, one
for birds and one for reptiles (including dinosaurs and crocodiles,
The taxonomic classification as well as lizards, snakes, and turtles), is not possible. And yet the
terms Aves and Reptilia are so familiar and well established that
system is hierarchical suddenly referring to birds as a type of dinosaur, and thus a type of
Taxonomy is the science of classifying living things. By agree- reptile, is difficult for some. Nonetheless, biologists increasingly
ment among taxonomists throughout the world, the first word of refer to birds as a type of dinosaur and hence a type of reptile.
the binomial name is the genus to which the organism belongs. Situations like this are not uncommon. Another example
This word is always capitalized. The second word refers to the par- concerns the classification of plants. Traditionally, three major
ticular species and is not capitalized. The two words together groups were recognized: green algae, bryophytes, and vascular
are called the species name (or scientific name) and are written in plants (figure 23.7). However, recent research reveals that neither
italics—for example, Homo sapiens. Once a genus has been used the green algae nor the bryophytes constitutes a mono­phy­let­ic
in the body of a text, it is often abbreviated in later uses. For group. Rather, some bryophyte groups are more closely related to
example, the dinosaur Tyrannosaurus rex becomes T. rex. vascular plants than they are to other bryophytes, and some green
Also by agreement among taxonomists, no two plant, animal, algae are more closely related to bryophytes and vascular plants
or microbial species can have the same scientific name. The scien- than they are to other green algae. As a result, systematists no lon-
tific name of an organism is the same anywhere in the world and ger recognize green algae or bryophytes as evolutionary groups,
avoids confusion caused by common names that may differ from and the classification system has been changed to reflect evolu-
one place to the next (oddly, the same scientific name can be used tionary relationships.
for species in different kingdoms, but such duplication is very rare).
The taxonomic system is hierarchical, with taxa at a lower level The phylogenetic species concept focuses
grouped into a smaller number of taxa in the next higher level. Just as
on shared derived characters
species are grouped into genera, genera with similar characteristics
are grouped into families, and similar families are placed into the In chapter 22, you read about a number of different ideas concern-
same order. Orders with common properties are placed into the same ing what determines whether two populations belong to the same
class, and classes with similar characteristics into the same phylum species. The biological species concept (BSC) defines species as
(plural, phyla). Finally, the phyla are assigned to one of several great groups of interbreeding populations that are reproductively isolat-
groups, the kingdoms. In addition, an eighth level of classification, ed from other groups. In recent years, a phylogenetic perspective
called a domain, is now frequently used. Figure 23.5 illustrates the has emerged and has been applied to the question of species con-
classification system, using the gray squirrel as an example. cepts. Advocates of the phylogenetic species concept (PSC) pro-
The categories at the different levels may include many, a few, or pose that the term species should be applied to groups of
only one taxon. For example, there is only one living genus of the fam- populations that have been evolving independently of other groups
ily Ornithorhynchidae (Ornithorhynchus, containing the species O. of populations. Moreover, they suggest that phylogenetic analysis
anatinus, the duck-billed platypus), but several living genera of Faga- is the way to identify such species. In this view, a species is a popu-
ceae (the birch family). To someone familiar with classification or hav- lation or set of populations characterized by one or more shared
ing access to the appropriate reference books, each taxon implies both derived character(s)—the rationale being that the possession of a
a set of characteristics and a group of organisms belonging to the taxon. derived character implies a period of separate evolution.
This approach solves two of the problems with the BSC that
Current classification sometimes does not were discussed in chapter 22. First, the BSC cannot be applied to
allopatric populations because scientists cannot determine whether
reflect evolutionary relationships individuals of the populations would interbreed and produce
Systematics and traditional classification are not always congru- fertile offspring if they ever came together. The PSC solves this
ent; to understand why, we need to consider how species may be problem: instead of trying to predict what would happen in the

490 part IV Evolution

 Domain
Eukarya

Kingdom
Animalia

Figure 23.5 The

Phylum
Chordata hierarchical system
used in classifying
an organism. The
organism, in this case
the eastern gray squirrel,
Subphylum is first recognized as a
Vertebrata eukaryote (domain Eukarya).
Within this domain, it is an
animal (kingdom Animalia).
Among the different phyla of
animals, it is a vertebrate (phylum
Class Chordata, subphylum Vertebrata).
Mammalia The organism’s fur characterizes it as a
mammal (class Mammalia). Within this
class, it is distinguished by its gnawing teeth
(order Rodentia). Next, because it has four front
toes and five back toes, it is a squirrel (family
Order Sciuridae). Within this family, it is a tree squirrel (genus
Rodentia Sciurus), with gray fur and white-tipped hairs on the tail
(species Sciurus carolinensis, the eastern gray squirrel).

future if allopatric populations ever came into contact (that is, would
Family
Sciuridae the populations be reproductively isolated?), the PSC looks to the
past to determine whether a population (or groups of populations)
has evolved independently for a long enough time to develop its own
derived characters.
Second, the PSC can be applied equally well to both sexual
and asexual species, in contrast to the BSC, which deals only with
Genus
Sciurus sexual forms.

The phylogenetic species concept

also has drawbacks
Species The PSC is controversial, however, for several reasons. First,
Sciurus some critics contend that it will lead to the recognition of every
Sciurus
carolinensis
carolinensis slightly differentiated population as a distinct species. In M
­ issouri,
for example, open, desertlike habitat patches called glades are
distributed throughout much of the state. These glades contain a

chapter 23 Systematics, Phylogenies, and Comparative Biology 491

 Archosaurs Figure
23.6 Monophyletic,
Giraffe Bat Turtle Crocodile Stegosaurus Tyrannosaurus Velociraptor Hawk
paraphyletic, and
polyphyletic groups.
a. A monophyletic group consists
of the most recent common
ancestor and all of its descendants.
For example, the name
“Archosaurs” is given to the
monophyletic group that includes
crocodiles, Stegosaurus,
Tyrannosaurus, Velociraptor, and
Monophyletic Group a hawk. b. A paraphyletic group
consists of the most recent
a. common ancestor and some of its
Dinosaurs descendants. For example, some,
but not all, taxonomists
Giraffe Bat Turtle Crocodile Stegosaurus Tyrannosaurus Velociraptor Hawk
traditionally give the name
“dinosaurs” to the paraphyletic
group that includes Stegosaurus,
Tyrannosaurus, and Velociraptor.
This group is paraphyletic because
one descendant of the most recent
ancestor of these species, birds, is
not included in the group. Other
taxonomists include birds within
the Dinosauria because
Paraphyletic Group Tyrannosaurus and Velociraptor
are more closely related to birds
b. than to other dinosaurs. c. A
Flying Vertebrates polyphyletic group does not
contain the most recent common
Giraffe Bat Turtle Crocodile Stegosaurus Tyrannosaurus Velociraptor Hawk
ancestor of the group. For
example, bats and birds could be
classified in the same group,
which we might call “flying
vertebrates,” because they have
similar shapes, anatomical
features, and habitats. However,
their similarities reflect convergent
evolution, not common ancestry.

Polyphyletic Group

c.

? Inquiry question Based on this phylogeny, are there any alternatives to convergence to explain
the presence of wings in birds and bats? What types of data might be used to test these hypotheses?

variety of warmth-loving species of plants and animals that do not A second problem is that species may not always be mono-
occur in the forests that separate the glades. Glades have been phyletic, contrary to the definition of some versions of the PSC.
isolated from one another for a few thousand years, allowing Consider, for example, a species composed of five populations,
enough time for populations on each glade to evolve differences in with evolutionary relationships like those indicated in figure 23.8.
some rapidly evolving parts of the genome. Does that mean that Suppose that population C became isolated and evolved differ-
each of the hundreds, if not thousands, of Missouri glades con- ences that made it qualify as a species by any concept (for example,
tains its own species of lizards, grasshoppers, and scorpions? reproductively isolated, ecologically differentiated). But this dis-
Some scientists argue that if one took the PSC to its logical tinction would mean that the remaining populations, which might still
extreme, that is exactly what would result. be perfectly capable of exchanging genes, would be paraphyletic,

492 part IV Evolution

 Red Algae Green Algae Bryophytes Vascular Plants
Chlorophytes Charophytes Liverworts Hornworts Mosses

Figure 23.7 Phylogenetic information transforms plant classification. The traditional classification included two groups that we
now realize are not monophyletic: the green algae and bryophytes. For this reason, plant systematists have developed a new classification of plants
that does not include these groups (discussed in chapter 29).

rather than monophyletic. Such situations probably occur often in

the natural world. Learning Outcomes Review 23.3
Phylogenetic species concepts, of which there are many The traditional classification is hierarchical, placing each species
different permutations, are increasingly used, but are also conten- into a nested set of taxonomic categories. This classification is not
tious for the reasons just discussed. Evolutionary biologists are always consistent with modern understanding of phylogenetic
trying to find ways to reconcile the historical perspective of relationships. By definition, a clade is monophyletic. A paraphyletic
the PSC with the process-oriented perspective of the BSC and group contains the most recent common ancestor, but not all its
descendants; a polyphyletic group does not contain the most
other species concepts.
recent common ancestor of all members. The PSC focuses on the
possession of shared derived characters, in contrast to the BSC,
which emphasizes reproductive isolation. The PSC solves some
A B C D E
problems of the BSC but has difficulties of its own.
■■ Under the BSC, is it possible for a species to be
polyphyletic?

23.4 Phylogenetics and

Comparative Biology

Learning Outcomes
1. Explain the importance of homoplasy for interpreting patterns
Figure 23.8 Paraphyly and PSC. The five populations of evolutionary change.
initially were all members of the same species, with their historical 2. Describe how phylogenetic trees can reveal the existence
relationships indicated by the cladogram. Then, population C evolved of homoplasy.
in some ways to become greatly differentiated ecologically and 3. Discuss how a phylogenetic tree can indicate the timing of
reproductively from the other populations. By all species concepts, species diversification.
this population would qualify as a different species. However, the
remaining four species do not form a clade; they are paraphyletic
because population C has been removed and placed in a different Phylogenies not only provide information about evolutionary rela-
species. This scenario may occur commonly in nature, but most tionships among species, but also are indispensable for understanding
versions of the PSC do not recognize paraphyletic species. how evolution has occurred. By examining the distribution of traits

chapter 23 Systematics, Phylogenies, and Comparative Biology 493

 Figure 23.9 Parental care in
dinosaurs and crocodiles a. Fossil
dinosaur incubating its eggs. This
re­markable fossil of Oviraptor shows
the dinosaur sitting on its nest of eggs
just as chickens do today. Not only is
the dinosaur squatting on the nest, but
its forelimbs are outstretched, perhaps
to shade the eggs. b. A mother exhibiting
parental care. Female crocodilians build
nests and then remain nearby, guarding
them, while the eggs incubate. When
they are ready to hatch, the baby
a. b. crocodiles vocalize; females respond by
digging up the eggs and carrying the
babies to the water. (a) ©David Clark/
Dinosaurs Alive/Giant Screen Films; (b) Shah,
Anup/Nature Picture Library/Alamy Stock Photo

among species in the context of their phylogenetic relationships, Homoplastic convergence: saber teeth
much can be learned about how and why evolution has proceeded. In and plant conducting tubes
this way, phylogenetics is the basis of all comparative biology. In other cases, by contrast, phylogenetic analysis can indicate that
Homologous features are derived from the similar traits have evolved independently in different clades. This
convergent evolution from different ancestral sources indicates that
same ancestral source; homoplastic features such traits represent homoplasies. As one example, the fossil record
are not reveals that extremely elongated canine teeth (saber teeth) occurred
In chapter 21, we pointed out that homologous structures are those in a number of different groups of extinct carnivorous mammals.
that are derived from the same body part in a common ancestor. Although how these teeth were actually used is still debated, all
Thus, the forelegs of a dolphin (flipper) and of a horse (leg) are saber-toothed carnivores had body proportions similar to those of
homologous because they are derived from the same bones in an an- cats, which suggests that these different types of carnivores all
cestral vertebrate. By contrast, the wings of birds and those of drag- evolved into a similar predatory lifestyle. Examination of the saber-
onflies are homoplastic structures because they are not derived from toothed character state in a phylogenetic context reveals that it most
the same structure in a common ancestor. Phylogenetic analysis can likely evolved independently at least four times (figure 23.10).
help determine whether structures are homologous or homoplastic. Conducting tubes in plants provide a similar example. The
tracheophytes, a large group of land plants discussed in chapter 29,
Homologous parental care in dinosaurs, transport photosynthetic products, hormones, and other molecules
crocodiles, and birds over long distances through elongated, tubular cells that have per-
Recent fossil discoveries have revealed that many species of dinosaurs forated walls at the end. These structures are stacked upon each
exhibited parental care. They incubated eggs laid in nests and took other to create a conduit called a sieve tube. Sieve tubes facilitate
care of the growing baby dinosaurs, many of which could not have long-distance transport that is essential for the survival of tall
fended for themselves. Some recent fossils show dinosaurs sitting on a plants on land.
nest in exactly the same posture used by birds today (figure 23.9a)! Most members of the brown algae, which includes kelp, also
Initially, these discoveries were treated as remarkable and have sieve elements (see figure 23.11 for a comparison of the sieve
unexpected—dinosaurs apparently had independently evolved behav- plates in brown algae and angiosperms) that aid in the rapid
iors similar to those of modern-day organisms. But examination of the transport of materials. The land plants and brown algae are
phylogenetic position of dinosaurs (see figure 23.6) indicates that they distantly related (figure 23.11), and their last common ancestor
are most closely related to two living groups of animals—crocodiles was a single-celled organism that could not have had a multicellu-
and birds—both of which exhibit parental care (figure 23.9b). lar transport system. This indicates that the strong structural and
It appears likely, therefore, that the parental care exhibited by functional similarity of sieve elements in these plant groups is an
crocodiles, dinosaurs, and birds did not evolve convergently from example of convergent evolution.
different ancestors that did not exhibit parental care; rather, the be-
haviors are homologous, inherited by each of these groups from Complex characters evolve through
their common ancestor that cared for its young. a sequence of evolutionary changes
Inquiry question What if only some types of dinosaurs Most complex characters do not evolve, fully formed, in one step.
? exhibited parental care? In light of the phylogeny in figure 23.6, Rather, they are often built up, step-by-step, in a series of evolu-
how would that change our inferences about the evolution of tionary transitions. Phylogenetic analysis can help discover these
parental care? evolutionary sequences.

494 part IV Evolution

 SCIENTIFIC THINKING

Question: How many times have saber teeth evolved in mammals?

Hypothesis: Saber teeth are homologous and have only evolved once in mammals (or, conversely, saber teeth are convergent and have evolved
multiple times in mammals).
Phylogenic Analysis: Examine the distribution of saber teeth on a phylogeny of mammals, and use parsimony to infer the history of saber tooth
evolution (note that not all branches within marsupials and placentals are shown on the phylogeny).

Phylogeny of Mammals Phylogeny of Carnivores and Creodonts

weasels, canids,

Saber-toothed

Saber-toothed

Saber-toothed
Saber-toothed

Saber-toothed

and raccoons
Bears, seals,

Mongooses
Carnivores

Creodonts
marsupial

nimravid

nimravid
creodont

Hyenas
Civets

cat
Felines
Marsupials
Placentals Nimravids

Monotremes

Creodonts Carnivores

Result: Saber teeth have evolved at least four times in mammals: once within marsupials, once in felines, once in an extinct group called creodonts, and
at least once in another group of now-extinct catlike carnivores called nimravids.
Interpretation: Note that it is possible that saber teeth evolved twice in nimravids, but another possibility that requires the same number of evolutionary
changes (and thus is equally parsimonious) is that saber teeth evolved only once in the ancestor of nimravids and then were subsequently lost in one
group of nimravids. (Note that for clarity, not all branches within marsupials and placentals are shown in this illustration.)

Figure 23.10 Distribution of saber-toothed mammals.

Stramenopiles Alveolates Red Algae Chlorophytes Chlorophytes Liverworts Hornworts Mosses Tracheophytes

60 μm 2 μm

Brown Alga Angiosperms

Figure 23.11 Convergent evolution of conducting tubes. Sieve tubes, which transport hormones and other substances throughout
the plant, have evolved in two distantly related plant groups (brown algae are stramenopiles and angiosperms are tracheophytes).
(Left) ©Lee W. Wilcox; (Right) Clouds Hill Imaging Ltd./Getty Images

chapter 23 Systematics, Phylogenies, and Comparative Biology 495

 Other dinosaurs Coelophysis Tyrannosaurus Sinosauropteryx Velociraptor Caudipteryx Archaeopteryx Modern Birds

Loss of teeth and

reduction of tail
Arms longer than legs
Long, aerodynamic
feathers

Long arms, highly mobile wrist, feathers

with vanes, shafts, and barbs

Downy feathers

Wishbone, breastbone,
loss of fingers 4 and 5

Light bones

Figure 23.12 The evolution of birds. The traits we think of as characteristic of modern birds have evolved in stages over many millions
of years.

Modern-day birds—with their wings, feathers, light bones, Larval dispersal in marine snails
and breastbone—are exquisitely adapted flying machines. Fossil An example of this use of phylogenetic analysis concerns the evo-
discoveries in recent years now allow us to reconstruct the evolu- lution of larval forms in marine snails. Most species of snails pro-
tion of these features. When the fossils are arranged phylogeneti- duce microscopic larvae that drift in the ocean currents, sometimes
cally, it becomes clear that the features characterizing living birds traveling hundreds or thousands of miles before becoming estab-
did not evolve simultaneously. ­­Figure 23.12 shows how the fea- lished and transforming into adults. Some species, however, have
tures important to flight evolved sequentially, probably over a long evolved larvae that settle to the ocean bottom very quickly and thus
period of time, in the ancestors of modern birds. don’t disperse far from their place of origin. Studies of fossil snails
One important finding often revealed by studies of the evo- indicate that the proportion of species that produce nondispersing
lution of complex characters is that the initial stages of a character larvae has increased through geological time (figure 23.13).
evolved as an adaptation to some environmental selective pressure
different from that for which the character is currently adapted.
Examination of figure 23.12 reveals that the first feathery struc- 40 MYA
tures evolved deep in the theropod phylogeny, in animals with fore-
arms clearly not modified for flight. Therefore, the initial Eocene
45 MYA
feather-like structures must have evolved for some other reason,
50 MYA
perhaps serving as insulation or decoration. Through time, these
structures have become modified to the extent that modern feath- 55 MYA
ers produce excellent aerodynamic performance.
Paleocene 60 MYA
Phylogenetic methods can be used to 65 MYA
distinguish between competing hypotheses 0 50 100
Understanding the causes of patterns of biological diversity ob- Percentage of species with
nondispersing larvae
served today can be difficult because a single pattern often could
have resulted from several different processes. In many cases, sci- Figure 23.13 Larval dispersal. Increase through time in the
entists can use phylogenies to distinguish between competing proportion of species whose larvae do not disperse far from their
hypotheses. place of birth.

496 part IV Evolution

 Two processes could produce an increase in nondispersing than changes in the opposite direction (indicated by minuses) in figure
larvae through time. First, if evolutionary change from dispersing 23.14a. In contrast, if nondispersing species underwent greater
to nondispersing occurred more often than change in the opposite speciation, then clades of nondispersing species would contain more
direction, then the proportion of species that are nondispersing species than clades of dispersing species, as shown in figure 23.14b.
would increase through time. Evidence for both processes was revealed in an examination of
Alternatively, if species that were nondispersing speciated the phylogeny of marine snails in the genus Conus, in which 30% of
more frequently, or became extinct less frequently, than dispersing species are nondispersing (figure 23.14c). The phylogeny indicates that
species, then through time the proportion of nondispersers would possession of dispersing larvae was the ancestral state; nondispersing
also increase (assuming that the descendants of nondispersing spe- larvae are inferred to have evolved eight times, with no evidence for
cies also were nondispersing). This latter case is a reasonable hy- evolutionary reversal from nondispersing to dispersing larvae.
pothesis because nondispersing species probably have lower At the same time, clades of nondispersing larvae tend to
amounts of gene flow than dispersing species, and thus might have on average 3.5 times as many species as dispersing larvae,
more easily become geographically isolated, increasing the likeli- which suggests that in nondispersing species, rates of speciation
hood of allopatric speciation (see chapter 22). are higher, rates of extinction are lower, or both.
These two processes would result in different phylogenetic pat- This analysis therefore indicates that the evolutionary in-
terns. If evolution from a dispersing ancestor to a nondispersing de- crease in nondispersing larvae through time may be a result both of
scendant occurred more often than the reverse, then an excess of such a bias in the direction in which evolution proceeds and of an in-
changes should be evident in the phylogeny, as shown by more transi- crease in rate of diversification (that is, speciation rate minus ex-
tions from dispersing to nondispersing larvae (indicated by pluses) tinction rate) in nondispersing clades.

+ –
–
+
–
+

+
+
+
Dispersing larvae
Nondispersing larvae

a. b.

+
+

+ +
+

+ +

Nondispersing larvae Dispersing larvae

c.
Figure 23.14 Phylogenetic investigation of the evolution of nondispersing larvae. a. In this hypothetical example, the
evolutionary transition from dispersing to nondispersing larvae occurs more frequently (four times, indicated by a plus sign) than the converse
(once, indicated by a minus sign). By contrast, in (b) the number of evolutionary changes in each direction is the same, but clades that have
nondispersing larvae diversify to a greater extent due to higher rates of speciation or lower rates of extinction (assuming that extinct forms are not
shown). c. Phylogeny for Conus, a genus of marine snails. Nondispersing larvae have evolved eight separate times from dispersing larvae, with no
instances of evolution in the reverse direction. This phylogeny does not show all species, however; nondispersing clades contain on average
3.5 times as many species as dispersing clades.

chapter 23 Systematics, Phylogenies, and Comparative Biology 497

 Figure 23.15 Evolution
of direct development in a
+ + family of limpets. a. Direct
– development evolved many
– – + times (beige lines stemming
+ + + from a red ancestor, indicated by a
+ plus sign), and three instances of
+ + reversed evolution from direct
+ + +
+ + +
development to larval development
+ are indicated (red lines from a
+
+ + beige ancestor, indicated by a
+ +
+ minus sign). b. A less
parsimonious interpretation of
evolution in the clade in the light
blue box is that, rather than two
Direct development + evolutionary reversals, six
Larval stage instances of the evolution of
development occurred without any
a. b. evolutionary reversal.

? Inquiry question How might you distinguish between the hypotheses in parts a and b
of this figure?

The lack of evolutionary reversal is not surprising because reversal—should be considered. For example, studies of the morphol-
when larvae evolve to become nondispersing, they often lose a variety ogy or embryology of direct-developing species might shed light on
of structures used for feeding while drifting in the ocean current. In whether such structures are homologous or convergent. In some cases,
most cases, once a structure is lost, it rarely re-evolves, and thus the artificial selection experiments in the laboratory or genetic manipula-
standard view is that the evolution of nondispersing larvae is a one- tions can test the hypothesis that it is difficult for lost structures to re-
way street, with few examples of re-evolution of dispersing larvae. evolve. Conclusions from phylogenetic analyses are always stronger
when supported by results of other types of studies.
Loss of the larval stage in marine invertebrates
A related phenomenon in many marine invertebrates is the loss of the Phylogenetics helps explain
larval stage entirely. Most marine invertebrates—in groups as diverse
as snails, sea stars, and anemones—pass through a larval stage as
species diversification
they develop. But in a number of different types of organisms, the One of the central goals of evolutionary biology is to explain pat-
larval stage is omitted, and the eggs develop directly into adults. terns of species diversity: Why do some types of plants and ani-
The evolutionary loss of the larval stage has been suggested mals exhibit more species richness—a greater number of species
as another example of a nonreversible evolutionary change because per clade—than others? Phylogenetic analysis can be used both to
once the larval stages are lost, it is difficult for them to re-evolve— suggest and to test hypotheses about such differences.
or so the argument goes. A recent study on one group of marine
limpets, shelled marine organisms related to snails, shows that this Species richness in beetles
is not necessarily the case. Among these limpets, direct develop- Beetles (order Coleoptera) are the most diverse group of animals.
ment has evolved many times; however, in three cases, the phylog- Approximately 60% of all animal species are insects, and there are
eny strongly suggests that evolution reversed and a larval stage far more species of beetles than of any other type of insect. Among
re-evolved (figure 23.15a). beetles, families that are herbivorous are particularly species-rich.
It is important to remember that patterns of evolution sug- Examination of the phylogeny provides insight into beetle evo-
gested by phylogenetic analysis are not always correct—evolution lutionary diversification (figure 23.16). Among the Phytophaga, the
does not necessarily occur parsimoniously. In the limpet study, for clade which contains most herbivorous beetle species, the deepest
example, it is possible that within the clade in the light blue box, branches belong to beetle families that specialize on conifers. This
presence of a larva was retained as the ancestral state, and direct finding agrees with the fossil record because conifers were among
development evolved independently six times (figure 23.15b). the earliest seed plant groups to evolve. By contrast, the flowering
Phylogenetic analysis cannot rule out this possibility, even if it is plants (angiosperms) evolved more recently, in the Cretaceous, and
less phylogenetically parsimonious. beetle families specializing on them have shorter evolutionary
If the re-evolution of lost traits seems unlikely, then the alterna- branches, indicating their more recent evolutionary appearance.
tive hypothesis that direct development evolved six times—rather than This correspondence between phylogenetic position and
only once at the base of the clade, with two instances of evolutionary timing of plant origins suggests that beetles have been remarkably

498 part IV Evolution

 25,000

33,400
41,602
Number
Phylogenetics and Disease

2000
23.5

1500

400
of Species

150
20
85

24

78
10

18
8
3
Tertiary Evolution

Cretaceous Learning Outcome

1. Discuss how phylogenetic analysis can help identify patterns
of disease transmission.

Jurassic The examples so far have illustrated the use of phylogenetic analy-
sis to examine relationships among species. Such analyses can also
be conducted on virtually any group of biological entities, as long
Conifer
Cycad as evolutionary divergence in these groups occurs by a branching
Triassic Angiosperm process, with little or no genetic exchange between different
groups. No example illustrates this better than recent attempts to
Figure 23.16 Evolutionary diversification of the understand the evolution of viruses. In this section, we will first
Phytophaga, the largest clade of herbivorous beetles. see how phylogenetic analysis has been important in discovering
Clades that originated deep in the phylogenetic tree feed on conifers; clades the origin and evolution of HIV, the virus that causes acquired
that feed on angiosperms, which evolved more recently, originated more immunodeficiency syndrome (AIDS). Then, we will see how the
recently. Age of clades is established by examination of fossil beetles. same approach has proven critical in understanding the recent
coronavirus pandemic.

conservative in their diet. The family Nemonychidae, for example,

appears to have remained specialized on conifers since the begin-
HIV has evolved from a viral counterpart
ning of the Jurassic, approximately 210 mya. present in another primate species
Phylogenetic explanations for beetle diversification AIDS was first recognized in the early 1980s, and it rapidly became
epidemic in the human population. Current estimates are that
The phylogenetic perspective suggests factors that may be respon-
more than 38 million people are infected with the human immuno-
sible for the incredible diversity of beetles. The phylogeny for the
deficiency virus (HIV), of whom nearly 700 thousand die each year.
Phytophaga indicates that it is not the evolution of herbivory itself
At first, scientists were perplexed about where HIV had
that is linked to great species richness. Rather, specialization on
originated and how it had infected humans. In the mid-1980s, how-
angiosperms seems to have been a prerequisite for great species
ever, scientists discovered a related virus in laboratory ­monkeys,
diversification. Specialization on angiosperms appears to have
termed simian immunodeficiency virus (SIV). In b­iochemical
arisen five times independently within herbivorous beetles; in each
terms, the viruses are very similar, although genetic differences
case, the angiosperm-specializing clade is substantially more species-
exist. At last count, SIV has been detected in more than 40 species
rich than the clade to which it is most closely related (termed a
of primates, but only in species found in sub-Saharan Africa. Inter-
sister clade) and which specializes on some other type of plant.
estingly, SIV—which appears to be transmitted sexually—does
Why specialization on angiosperms has led to great species
not appear to cause illness in some of these species.
diversity is not yet clear and is the focus of much current research.
Based on the degree of genetic differentiation among strains
One possibility is that this diversity is linked to the great species-
of SIV, scientists estimate that SIV may have been around for more
richness of angiosperms themselves. With more than 300,000
than a million years in these primates, perhaps providing enough
species of angiosperms, beetle clades specializing on them may
time for these species to adapt to the virus and thus prevent it from
have had a multitude of opportunities to adapt to feed on different
having adverse effects.
species, thus promoting divergence and speciation.
Phylogenetic analysis identifies
Learning Outcomes Review 23.4
the path of transmission
Homologous traits are derived from the same ancestral character
states, whereas homoplastic traits are not, even though they may Phylogenetic analysis of strains of HIV and SIV reveals three clear
have similar function. Phylogenetic analysis can help determine findings. First, HIV obviously descended from SIV. All strains of
whether homology or homoplasy has occurred. By correlating HIV are phylogenetically nested within clades of SIV strains, indi-
phylogenetic branching with known evolutionary events, the cating that HIV is derived from SIV (figure 23.17).
timing and cause of diversification can be inferred. Second, a number of different strains of HIV exist, and they
■■ Does the possession of the same character state by all appear to represent independent transfers from different primate
members of a clade mean that the ancestor of that clade species. Each of the human strains is more closely related to a
necessarily possessed that character state? strain of SIV than it is to other HIV strains, indicating separate ori-
gins of the HIV strains.

chapter 23 Systematics, Phylogenies, and Comparative Biology 499

 relatives, are plummeting toward extinction. A

SIV- greater spot-nasal monkey

second consequence of this hunting is that hu-
mans are increasingly brought into contact

SIV- red-capped mangabey

SIV- De Brazza’s monkey

with bodily fluids of these animals, and it is

SIV- sooty mangabey

SIV- sooty mangabey

easy to imagine how during the butchering

SIV- Syke’s monkey

SIV- vervet monkey
process, blood from a recently killed animal
SIV- chimpanzee

SIV- chimpanzee
SIV- chimpanzee
SIV- chimpanzee
SIV- chimpanzee

SIV- chimpanzee

HIV- 2 group A
HIV- 1 group M

HIV- 2 group B
HIV- 1 group O
HIV- 1 group N

HIV- 1 group P
might enter the human bloodstream through

SIV- tantalus
open sores or cuts in the skin.

SIV- gorilla

SIV- drill
Establishing the crossover time line
and location
Where and when did this cross-species trans-
mission occur? HIV strains are most diverse
in Africa, and the incidence of HIV is higher
there than elsewhere in the world. Combined
with the evidence that HIV is related to SIV
in African primates, it seems certain that
AIDS appeared first in Africa.
As for when the jump from other pri-
mates to humans occurred, the fact that AIDS
was not recognized until the 1980s suggests
that HIV probably arose recently. Descen-
dants of slaves brought to North America
from West Africa in the 19th century lacked
Indicates
transmission the disease, indicating that it probably did
to humans not occur at the time of the slave trade.
Once the disease was recognized in the
Figure 23.17 Evolution of HIV and SIV. HIV has evolved multiple times and from 1980s, scientists scoured repositories of old
strains of SIV in different primate species (each primate species indicated by a different blood samples to see whether HIV could be
color; drill and tantalus are species of monkey). Red rectangles indicate transmission to detected in blood samples from the past. The
humans from other primate species. earliest HIV-positive result was found in a
sample from 1959, pushing the date of origin
back at least two decades. Based on the
Finally, humans have acquired HIV from different host spe- amount of genetic difference between strains of HIV-1, including
cies. HIV-1, which is the virus responsible for the global epidemic, the 1959 sample, and assuming the operation of a molecular clock,
has four subtypes. Two of these subtypes are most closely related scientists estimate that the deadly strain of AIDS probably crossed
to strains in chimpanzees, whereas a third is most closely related into humans in the 1920s.
to a strain in gorillas, indicating transmission from both ape spe-
cies. The origin of the fourth subtype is not clear; further sam- Phylogenies can be used to track the evolution
pling will likely reveal it to be the sister taxon to a currently of AIDS among individuals
undiscovered chimp or gorilla strain.
By contrast, subtypes of HIV-2, which is much less wide- The AIDS virus evolves extremely rapidly, so much so that dif-
spread (in some cases known from only one individual), are related ferent strains can exist within a single individual in a single popu-
to SIV found in West African monkeys, primarily the sooty mang- lation. As a result, phylogenetic analysis can be applied to
abey (Cercocebus atys). Moreover, the subtypes of HIV-2 also ap- answer very specific questions; just as phylogeny proved useful
pear to represent multiple cross-species transmissions to humans. in determining the source of HIV, it can also pinpoint the source
of infection for particular individuals.
Transmission from other primates to humans This ability became apparent in a court case in Louisiana in
Several hypotheses have been proposed to explain how SIV jumped 1998, in which a dentist was accused of injecting his former girl-
from chimps and monkeys to humans. The most likely idea is that friend with blood drawn from an HIV-infected patient. The dentist’s
transmission occurred as the result of blood-to-blood contact that records revealed that he had drawn blood from the patient and had
may occur when humans kill and butcher monkeys and apes. Recent done so in a suspicious manner. Scientists sequenced the viral strains
years have seen a huge increase in the rate at which primates are from the victim, from the patient, and from a large number of HIV-
hunted for the “bushmeat” market, particularly­in central and west- infected people in the local community. The phylogenetic analysis
ern Africa. This increase has resulted from growing human popula- clearly demonstrated that the victim’s viral strain was most closely
tions desiring greater amounts of protein combined with increased related to the patient’s (figure 23.18). This analysis, which for the
access to the habitats in which these primates live as the result of first time established phylogenetics as a legally admissible form of
road building and economic development. The unfortunate result is evidence in courts in the United States, helped convict the dentist,
that population sizes of many primate species, including our closest who is now serving a 50-year sentence for attempted murder.
500 part IV Evolution
 Figure 23.18 Evolution of HIV strains reveals the source
of infection. HIV mutates so rapidly that multiple genotypes often

Community
circulate within the body of a single person infected with HIV. As a

Patient
Victim
result, it is possible to create a phylogeny of HIV strains and to identify
the source of infection of a particular individual. In this case, the HIV
strains of the victim (pink) clearly are derived from strains in the body
of another individual, the patient (blue). Other HIV strains are from
HIV-infected individuals in the local community.

? Inquiry question What would the phylogeny look like if

the victim had not contracted HIV from the patient?

California early in the pandemic indicated that already by March

2020, at least seven different strains had arrived in the state, indi-
cating introduction of the disease from multiple localities on sev-
eral continents. On the other hand, clusters of patients possessing
the same strain indicated the occurrence of community transmis-
sion from one individual to another. A similar phylogenetic analy-
sis of patients in New York City in March 2020 identified at least
nine introductions, most from Europe or elsewhere in North Amer-
ica, and multiple cases of community transmission.

Phylogenies also have been used to study the Learning Outcomes Review 23.5
origin and spread of COVID-19 Modern phylogenetic techniques and analysis can track the
evolution of disease strains, uncovering sources and progression.
When SARS-CoV-2 (COVID-19) emerged in early 2020, scien- The HIV virus provides a prime example: analysis of viral strains has
tists were well prepared to investigate its origin and spread follow- shown that the progression from simian immunodeficiency virus
ing the approaches established to study HIV and other emerging (SIV) into human hosts has occurred several times. Phylogenetic
diseases over the previous several decades. Moreover, the capacity analysis is also used to track the transmission of human disease.
to collect DNA sequence data rapidly and in large quantities had ■■ Could HIV have arisen in humans and then been transmitted
greatly expanded since the emergence of earlier diseases. to other primate species?
Very quickly, the virus responsible for the disease was iden-
tified as a coronavirus, similar to the one that had caused the SARS
outbreak of 2003 (see chapter 26). Phylogenetic analysis indicated
that the most closely related viral lineages all occurred in bat spe-
cies, suggesting that the disease was the result of a virus that had
jumped from bats to humans (figure 23.19).
One complication was that the disease apparently first emerged
in a market in Wuhan, China, where live animals of many species Human Bat Bat Pangolin Pangolin Bat Bat
were kept for sale in very crowded conditions. Presumably, the virus
had been passed to humans at that market. However, bats were not
on sale at the market, so it was suspected that an intermediate spe-
cies, which had gotten the virus from a bat and then passed it on to
humans, was involved. Exactly that scenario happened in 2003 with
SARS—genetic analysis identified the civet, a cat-sized Asian car-
nivorous mammal, as the intermediate species and horseshoe bats as
the natural reservoir from which the disease had emerged.
The unexpected discovery that closely related viruses occur in
the pangolin, an obscure ant-eating species endangered by illegal wild-
life trafficking, suggested that this species might be the conduit be-
tween bats and humans, but subsequent research has been equivocal.
At this point, the intermediate species still has not been conclusively
identified and researchers continue the search. In addition, the possi- Figure 23.19 Phylogenetic relationships of
bility that the virus escaped from a research laboratory is still under coronaviruses. The most closely related virus to the one that
investigation, though evidence supporting this hypothesis is scant. causes COVID-19 in humans occurs in bat. Because pangolins harbor
Phylogenetic analysis of genetically different strains of a similar, though phylogenetically more distant, virus, some have
COVID-19 in different patients has been used to understand how suggested that pangolins may be the intermediate host that
the disease spread, just as with HIV. One study among patients in transmitted COVID-19 from bats to humans. EcoPic/Getty Images
chapter 23 Systematics, Phylogenies, and Comparative Biology 501
 Chapter Review

Imagemore Co, Ltd./Imagemore/Getty Images

23.1 Systematics The PSC focuses on shared derived characters.

The PSC emphasizes the possession of shared derived characters, whereas
Branching diagrams depict evolutionary relationships. the BSC focuses on reproductive isolation. Many versions of the PSC
Systematics is the study of evolutionary relationships, which are depicted recognize species that are monophyletic.
on branching evolutionary trees, called phylogenies.
The PSC also has drawbacks.
Similarity may not accurately predict evolutionary relationships. Among criticisms of the PSC are that it recognizes as a distinct species
The rate of evolution can vary among species and can even reverse any population with any degree of genetic difference from other
direction. Closely related species can therefore be dissimilar in populations and that evolution routinely produces paraphyletic groups
phenotypic characteristics. that would be considered to be species by most other species concepts.
Conversely, convergent evolution results in the phenotypic similarity of
distantly related species. 23.4 Phylogenetics and Comparative Biology
23.2 Cladistics Homologous features are derived from the same ancestral
source; homoplastic features are not.
The cladistic method requires that character variation be Phylogenetic analysis can indicate whether different structures have
identified as ancestral or derived. been built from the same ancestral structure and thus may be considered
Derived character states are those that differ from the ancestral condition. homologous.
Only shared derived characters are useful for inferring phylogenies.
Character polarity is established using an outgroup comparison in which Complex characters evolve through a sequence of evolutionary
the outgroup consists of one or a group of species known to be outside of changes.
the group under study. Most complex features do not evolve in a single step but include stages of
transition. They may have begun as an adaptation to a selective pressure
Character states exhibited by the outgroup are assumed to be ancestral,
different from the one for which the feature is currently adapted.
and other character states are considered derived.
Phylogenetic methods can be used to distinguish between
Homoplasy complicates cladistic analysis.
competing hypotheses.
Homoplasy refers to a shared derived character state, such as wings of
Different evolutionary scenarios can be distinguished by phylogenetic
birds and wings of insects, that has not been inherited from a common
analysis. The minimum number of times a trait may have evolved can be
ancestor.
established, and the direction of trait evolution, the timing, and the cause
Cladograms are constructed based on the principle of parsimony, which of diversification can be inferred.
indicates that the phylogeny requiring the fewest evolutionary changes is
accepted as the best working hypothesis. Phylogenetics helps explain species diversification.
Questions regarding the causes of species richness may be addressed
Other phylogenetic methods work better than cladistics in with phylogenetic analysis.
some situations.
When evolutionary change is rapid, other methods, such as statistical
approaches and the use of the molecular clock, are sometimes more useful. 23.5 Phylogenetics and Disease Evolution
HIV has evolved from a viral counterpart present in another
23.3 Systematics and Classification primate species.
Phylogenetic methods have indicated that HIV is related to SIV.
The taxonomic classification system is hierarchical.
Species are grouped into genera, genera into families, and so on through Phylogenetic analysis identifies the path of transmission.
the ranks of order, class, phylum, kingdom, and domain. It is clear that HIV has descended from SIV, and that independent transfers
from other primate species to humans have occurred several times.
Current classification sometimes does not reflect evolutionary
relationships. Phylogenies can be used to track the evolution of AIDS
A monophyletic group consists of the most recent common ancestor and among individuals.
all of its descendants. Because HIV evolves rapidly, phylogenetic analysis can trace
A paraphyletic group consists of the most recent common ancestor and the origin of a current strain to a specific source of infection.
some of its descendants. Phylogenies also have been used to study the origin and spread
A polyphyletic group does not contain the most recent ancestor of the group. of COVID-19.
Some currently recognized taxa are not monophyletic, such as reptiles, COVID-19 is derived from viruses found in bats. Phylogenetic analysis
which are paraphyletic with respect to birds. indicates multiple introductions and community transmission.

502 part IV Evolution

Visual Summary
Evolutionary
may not represent Hierarchy
relationships

depicted arranged
Ancestral by as a
characters
inferred
identifies Cladistics by Phylogenies Classification
Derived
characters
complicated by used to infer

Homoplasy Evolution

understood is key to
by studying

Species with Comparative Studying disease

explains tracks
similar features biology dynamics

as a result of enlightens
patterns of identifies

Species Path of Evolution of

Homoplasy Homology diversification transmission HIV and COVID-19

Review Questions

Imagemore Co, Ltd./Imagemore/Getty Images

U N D E R S TA N D 4. Parsimony suggests that parental care in birds, crocodiles, and
1. Overall similarity of phenotypes may not always reflect some dinosaurs
evolutionary relationships a. evolved independently multiple times by convergent evolution.
a. due to convergent evolution. b. evolved once in an ancestor common to all three groups.
b. because of variation in rates of evolutionary change of c. is a homoplastic trait.
different kinds of characters. d. is not a homologous trait.
c. due to homoplasy. 5. The forelimb of a bird and the forelimb of a rhinoceros
d. All of the choices are correct. a. are homologous and symplesiomorphic.
2. Cladistics b. are not homologous but are symplesiomorphic.
a. is based on overall similarity of phenotypes. c. are homologous and synapomorphic.
b. requires distinguishing similarity due to inheritance d. are not homologous but are synapomorphic.
from a common ancestor from similarity for other reasons. 6. In order to determine polarity for different states of a character
c. is not affected by homoplasy. a. there must be a fossil record of the groups in question.
d. None of the choices is correct. b. genetic sequence data must be available.
3. The principle of parsimony c. an appropriate name for the taxonomic group must be
a. helps evolutionary biologists distinguish among competing selected.
phylogenetic hypotheses. d. an outgroup must be identified.
b. does not require that the polarity of traits be determined. 7. In a paraphyletic group
c. is a way to avoid having to use outgroups in a phylogenetic a. all species are more closely related to each other than they are
analysis. to a species outside the group.
d. cannot be applied to molecular traits. b. evolutionary reversal is common.

chapter 23 Systematics, Phylogenies, and Comparative Biology 503

 c. polyphyly also usually occurs. SYNTHESIZE
d. some species are more closely related to species 1. List the synapomorphy and the taxa defined by that synapomorphy
outside the group than to some species within the group. for the groups pictured in figure 23.2. Name each group defined
8. A paraphyletic group includes by a set of synapomorphies in a way that might be construed as
a. an ancestor and all of its descendants. informative about what kind of characters define the group.
b. an ancestor and some of its descendants. 2. Identifying “outgroups” is a central component of cladistic
c. descendants of more than one common ancestor. analysis. As described on page 487, a group is chosen that is
d. All of the choices are correct. closely related to, but not a part of, the group under study. If one
9. Sieve tubes and sieve elements are does not know the relationships of members of the group under
study, how can one be certain that an appropriate outgroup is
a. homoplastic because they have different function.
chosen? Can you think of any approaches that would minimize the
b. homologous because they have similar function.
effect of a poor choice of outgroup?
c. homoplastic because their common ancestor was
single-celled. 3. As noted in your reading, cladistics is a widely utilized method of
d. structures involved in transport within animals. systematics, and our classification system (taxonomy) is
increasingly becoming reflective of our knowledge of evolutionary
A P P LY relationships. Using birds as an example, discuss the advantages
and disadvantages of recognizing them as reptiles versus as a group
1. A taxonomic group that contains species that have similar
separate from and equal to reptiles.
phenotypes due to convergent evolution is
4. Across many species of limpets, loss of larval development and
a. paraphyletic. c. polyphyletic.
reversal from direct development appear to have occurred multiple
b. monophyletic. d. a good cladistic group.
times. Under the simple principle of parsimony, are changes in
2. Rapid rates of character change relative to the rate of speciation either direction merely counted equally in evaluating the most
pose a problem for cladistics because parsimonious hypothesis? If it is much more likely to lose a larval
a. the frequency with which distantly related species evolve the mode than to re-evolve it from direct development, should that be
same derived character state may be high. taken into account? If so, how?
b. evolutionary reversals may occur frequently. 5. Birds, pterosaurs (a type of flying reptile that lived in the
c. homoplasy will be common. Cretaceous period), and bats all have modified their forelimbs to
d. All of the choices are correct. serve as wings, but they have done so in different ways. Are these
3. Species recognized by the PSC structures homologous? If so, how can they be both homologous
a. sometimes also would be recognized as species by the BSC. and convergent? Do any organisms possess wings that are
b. are sometimes paraphyletic. convergent, but not homologous?
c. are characterized by symplesiomorphies. 6. In what sense does the BSC focus on evolutionary mechanisms and
d. are more frequent in plants than in animals. the PSC on evolutionary patterns? Which, if either, is correct?

504 part IV Evolution

 24
CHAPTER

Genome Evolution

Chapter Contents
24.1 Comparative Genomics
24.2 Genome Size
24.3 Evolution Within Genomes
24.4 Gene Function and Expression Patterns
24.5 Applying Comparative Genomics

Stephen Robinson/NHPA/Photoshot/Newscom

Visual Outline Introduction

Genomic analysis Genomes contain the raw material for
evolution, and many clues to evolution
includes are hidden in the ever-changing nature
of genomes. As more genomes have
been sequenced, the new and excit-
Comparative
Genome size
Genome Functional ing field of comparative genomics has
genomics evolution genomics
emerged and has yielded some surprising
results as well as many, many questions.
acquiring applications
informs increases examining in Comparing whole genomes, not just
new function
due to
reveals individual genes, enhances our ability to
Rearranged Species-specific understand the workings of evolution, to
Phylogeny Duplication
DNA genes improve crops, and to identify the genetic
and
basis of disease so that we might develop
whole affected and complicated
genome by by more effective treatments with minimal
Rate of Noncoding Species-specific
evolution expression
side effects. The focus of this chapter
DNA
is the role of comparative genomics in
Polyploidy enhancing our understanding of genome
Nicotiana sylvestris Nicotiana tomentosiformis Transposable Horizontal
Medicine
evolution and how this new knowledge
elements gene transfer
can be applied to improve our lives.
Diploid (SS) Diploid (TT)

and Bacteria Archaea Eukarya

ST hybrid Noncoding Conservation

DNA biology
Genome duplication
asts
opl
lor
Nicotiana tabacum Ch ndri
a
Allopolyploid (SSTT) cho
ito
M
 Comparison between human and mouse genomes
24.1 Comparative Genomics The vertebrate other than human with the most genomic data is the
mouse. The importance of the mouse as a model system has led to
equivalent projects for virtually every human genome–related project.
A comparison of the genomes reveals that both have about 20,000
Learning Outcomes
protein-coding genes, essentially all of which are shared by both
1. Describe the kinds of differences that can be found between species. There is more divergence in genes that encode long non-
genomes. coding RNAs (lncRNA), and microRNAs (miRNA), where only
2. Explain how we can find human-specific differences. 10–20% of these genes appear to be conserved between the two spe-
3. Explain why genomes may evolve at different rates. cies. As these are either known (miRNA) or thought to be (lncRNA)
involved in regulating gene expression, these may explain more of
A key challenge of modern evolutionary biology is finding a the differences between species than the protein-coding genes.
way to link changes in DNA sequences, which we are now able to
study in great detail, with the evolution of the complex morpho- The protovertebrate genome
logical characters used to construct a traditional phylogeny. Many
Studies like those detailed above, and other mapping studies, are be-
different genes contribute to complex characters. Making the
ing used to reassemble a karyotype for the original protovertebrate
connection between a specific change in a gene and a modification
genome. These studies take advantage of the fact that we now have
in a morphological character is particularly difficult.
genome sequences for the coelocanth, a representative of lobe-finned
Comparing genomes provides a powerful tool for exploring
fish, which gave rise to all vertebrate tetrapods (see chapter 34), and
evolutionary divergence among organisms to connect DNA-level
to nonvertebrate chordates, which branched off from the vertebrate
changes with morphological differences. Genomes are more than
lineage prior to any vertebrate-specific rearrangements. One model
instruction books for building and maintaining an organism; they
proposes an ancestral chordate genome that consisted of 17 chromo-
also record the history of life. The growing number of fully
somes. This ancient genome then underwent a whole-genome dupli-
sequenced genomes in all kingdoms is leading to a revolution in
cation event to produce 34 chromosomes. These 34 chromosome
comparative evolutionary biology (figure 24.1). Genetic differences
then underwent a series of fusion events, followed by a second round
between species can be explored in a very direct way, examining the
of whole-genome duplication, to produce 54 chromosomes in the
footprints on the evolutionary path between different species.
ancestor to vertebrates. In the lineage leading to amniotes (birds,
reptiles, and mammals; see chapter 34), more fusion events occurred
Evolution of the vertebrate genome leading to 49 chromosomes in the amniote genome. This protoamni-
ote genome has undergone further rearrangements in different
Since the first human genome was sequenced, literally thousands lineages to produce the karyotypes we see today.
of human genomes have been sequenced. But while this allows
inferences about variation within humans, it tells us nothing about
the evolution of the human genome. To address this, there has been The human genome shares similarities
a large-scale effort to sequence as many chordate genomes as with those of great apes
possible. As vertebrates make up the majority of chordates (see
chapter 34), this is a very broad data set for comparative genomics. Genome projects have been completed or are under way for the
There are literally sequences for hundreds of vertebrates, from majority of extant primates. Among these, the chimpanzee, Pan
aardvark to zebrafish, available. This has allowed the curation of troglodytes, and bonobo, Pan paniscus, share the most recent com-
almost 16,000 orthologous genes, that is, genes that are considered mon ancestor with humans. Comparisons among these genomes
to have been inherited from a common ancestor, in vertebrates. reveal 1.4% divergence in single-site substitutions between humans
This effort is ongoing, and these data are publicly available, which and chimps, and 1.5% between humans and bonobos. There is
will accelerate vertebrate genomics studies. another 1.5% of the genome that differs in small insertions and
deletions (indels) between humans and chimps, and a similar
percentage between humans and bonobos.
Comparison between human and pufferfish genomes Comparing all sequenced primates indicates that the preva-
It may seem strange that the first animal genome compared to human lence of transposable elements observed in humans (chapter 18) is
was the pufferfish (Fugu rubripes), but it was the second vertebrate also true for other primates. About 50% of all primate genomes con-
genome sequenced. These two animals last shared a common ances- sist of various transposons. Species-specific insertions of these
tor 450 mya. Some human and pufferfish genes have been conserved elements varies considerably in different lineages and different
during evolution, but others are unique to each species. About 25% elements have expanded in some lineages more than others. In hu-
of human genes have no counterparts in Fugu. Also, extensive mans, Alu insertions have expanded almost twice as fast as in the
genome rearrangements have occurred during the 450 million years chimp and bonobo genomes. As these elements can also facilitate
since the mammal lineage and the teleost fish diverged, indicating a deletions or duplications, they can have effects on genome evolu-
considerable scrambling of gene order. Finally, the human genome tion outside of insertion events.
contains about 50% repetitive DNA, but repetitive DNA accounts The majority of protein-coding genes found in humans
for less than one-sixth of the Fugu sequence. have homologues among both great apes and old-world apes.

506 part IV Evolution

 Pan troglodytes
Fugu rubripes (chimpanzee)
(pufferfish)

Anopheles gambiae
(mosquito) Ornithorynchus anatinus
(duck-billed platypus)

2,996 Mb Homo sapiens

393 Mb 18,759 genes
18,523 genes (human)

Drosophila melanogaster
(fruit fly) Bos taurus
265 Mb (domestic cow)
14,086 genes 1,918 Mb
21,698 genes

Mus musculus 3,384 Mb

Caenorhabditis elegans (mouse) 20,300 genes
(a translucent worm)

142 Mb
13,918 genes 2,650 Mb
19,994 genes

Schizosaccharomyces
3,482 Mb pombe (fission yeast)
103 Mb 22,606 genes
20,447 genes
Zea mays
(maize)

Oryza sativa
(rice)
Arabidopsis thaliana 12.6 Mb
(wall cress) 6,991 genes 2 μm

Saccharomyces cerevisiae
2,067 Mb (baker’s yeast)
39,475 genes

373 Mb
39,045 genes
120 Mb
35,378 genes

Thermoplasma volcanium 12.2 Mb

5,906 genes 25 μm
(archaea)

Plasmodium falciparum
(malaria parasite)

Escherichia coli
(bacteria)
1.6 Mb
1,550 genes 1 μm

23.3 Mb
5,500 genes 1 μm

4.6 Mb 8,800×
4,377 genes Figure 24.1 Milestones of comparative eukaryotic genomics. The sizes
of genomes listed are the latest found in a variety of genome databases. The number of
genes refers specifically to protein coding genes, or so-called coding sequences (CDS).
This does not include genes for microRNAs and other forms of RNA.

(mosquito) Source: James Gathany/CDC; (pufferfish) Stephen Frink/Digital Vision/Getty Images; (duck-billed platypus) Paulo Oliveira/Alamy Stock Photo; (chimpanzee) Eric
Gevaert/Shutterstock; (a translucent worm) Heiti Paves/Alamy stock photo; (fruit fly) Thomas Deerinck, NCMIR/Getty Images; (mouse) G.K. & Vikki Hart/Photodisc/Getty Images;
(domestic cow) Alan and Sandy Carey/Getty Images; (human) James Stevenson/Science Source; (wall cress) Nigel Cattlin/Alamy Stock Photo; (maize) stevanovicigor/iStock/
Getty Images; (rice) Photo by Gary Kramer, U.S. Department of Agriculture Natural Resources Conservation Service; (fission yeast) Steve Gschmeissner/Science Source;
(baker’s yeast) Dr. Jeremy Burgess/Science Source; (bacteria) BSIP/age fotostock; (archaea) Dr. Harald Huber and Prof. Dr. Reinhard Rachel, University of Regensburg,
Regensburg, Germany; (malaria parasite) T. Brain/Science Source
 Dates of occupation
Type of sites
30,000–45,000
Neanderthal sites 45,000–135,000
Modern human sites 135,000–250,000

Figure 24.2 Europeans and Asians, but not Africans, share between 1% and 4% of their genomes with Neanderthals.
From 30,000 to 45,000 years ago humans and Neanderthals coexisted in Europe, and possibly in the Middle East as early as 80,000 years ago.
Interbreeding likely occurred after European and Asian ancestors left Africa.

Despite this, some gene families have experienced lineage-specific (figure 24.2). As Neanderthals preceded humans out of Africa into
expansion and contraction in different lineages. Genes in the major Europe, this must have occurred when early humans and Neander-
histocompatibility complex (MHC) are expanded in macaques thals coexisted on the European continent prior to the extinction
relative to humans, perhaps related to differential exposure to of Neanderthals around 28,000 years ago.
pathogens. Humans and chimps show interesting lineage-specific As if this was not surprising enough, DNA isolated from a
gains and losses of the zinc finger transcription factor family of single finger bone from a cave in the Altai mountains in Siberia,
genes. The effect of these differences in transcription factors on proved to be from a previously unknown hominin we now call
patterns of gene expression in chimps and humans is not clear, but Denisovans, for the name of the cave. Populations from Oceania
might potentially be involved in species differences. (Australia, Papua New Guinea, and nearby islands) inherited up to
A recent analysis of lineage-specific differences in structural 5% of their genomes from Denisovans. Populations on mainland
variation among primates has shown that there are close to 18,000 Asia have a much lower percentage of Denisovan DNA.
human-specific structural variants. These include around 12,000 The final surprise was the recent sequencing of another fossil
different fixed human-specific insertions, and almost 6,000 fixed genome that appears to represent an individual with a Denisovan
human-specific deletions. There is some correlation between the father and a Neanderthal mother. This would be the first such hybrid
location of these variants and regions known to be involved in regu- identified, but it fits with the idea that there was much more inter-
lating gene expression. One tantalizing example is the deletion of a breeding among these extinct primates than previously thought.
region thought to regulate two cell-cycle genes active in cortical In searching for genetic features unique to humans, researchers
neurons. As one human-specific morphological difference is an en- have taken advantage of the thousands of human genomes sequenced,
larged brain, this kind of alteration could lead to more cell division and the improving quality of the Neanderthal reference genome, to
during brain development. All of these functional studies are in their search for human genetic variation missing in Neanderthals. Looking
infancy, but they indicate the direction this research is going. at variation shared by 90% of humans analyzed, 105,757 SNPs and
3,900 indels are unique to humans. For protein-coding genes, there
The human genome contains sequences are 96 fixed substitutions in a total of 87 proteins. There are also
about 3000 changes that potentially affect gene expression. These
from extinct primates results are just a beginning, but eventually this may lead to a genetic
It is now possible to sequence DNA isolated from fossil starting definition of what it is to be human.
material. This has led to what sounds like a science fiction movie: The amount of archaic DNA found in modern humans varies
the Neanderthal reference genome. Yes, we now have sequenced across the genome. This implies that negative selection may have
not one, but several genomes from Homo neanderthalensis. been acting to remove archaic DNA. Regions that are thought to be
The biggest surprise from this work is that our genomes contain functionally important appear to have less Neanderthal DNA. This
the genetic remains of ancient hybridization between humans and is illustrated by the X chromosome, which has about five-fold less
Neanderthals. Present-day humans from non-African populations Neanderthal DNA than is found in autosomes. Further, testis-
inherited from 1 to 3% of their DNA from Neanderthals specific genes are found in regions that lack Neanderthal DNA,

508 part IV Evolution

perhaps indicating some level of male sterility in hybrids. There is plant history was in the gene families required for colonizing land,
also evidence for positive selection for Neanderthal sequences a completely new environment.
containing genes involved in skin and hair pigmentation.
Comparison of plants with animals and fungi
Genomes evolve at different rates About one-third of the genes in Arabidopsis and rice appear to be
plant-specific, that is, genes not found in any animal or fungal ge-
As noted earlier in this section, the genomes of viruses and bacte- nome sequenced so far. These include the many thousands of genes
ria can evolve in a matter of days. Evolution of insect genomes involved in photosynthesis and photosynthetic anatomy.
occurs more rapidly than that of mammalian genomes, which Among the remaining genes found in plants, many are very
evolve over millions of years. At least a part of this is due to differ- similar to those found in animal and fungal genomes, particularly
ences in generation time. However, even in some complex eukary- the genes involved in basic intermediary metabolism, in genome
otic organisms, genome evolution can occur more rapidly. replication and repair, and in transcription and protein synthesis.
A comparison using a selection of plant and animal genomes Prior to the availability of whole-genome sequences, assessment
yielded a surprising finding. Plant genomes change much more rap- of the extent of genetic similarity and difference among diverse
idly than animal genomes. This is especially evident in noncoding organisms had been difficult at best.
DNA, which tends to change more rapidly than protein-coding DNA.
Many conserved, noncoding regions of DNA can be identified when
a fish and a primate genome are aligned, even though the two
Learning Outcomes Review 24.1
diverged 400 mya. But very few similarities are found between the
Genomes vary in number of similar genes, arrangement of genes,
noncoding DNA of Arabidopsis and rice, Oryza sativa, although the
total number of base-pairs, and base-pair differences. Closely
two diverged only 200 mya. The sequencing of the corn genome has
related animals such as humans and chimps exhibit highly similar
revealed that two varieties can differ by as much as 20%. Contrast genomes, but greater variation is found in related plant genomes.
this with the differences between human and chimp genomes.
What accounts for all the variation in plant genomes? ■■ Would you expect a high degree of similarity between
Transposable elements and other mobile elements frequently remodel genomes of a bony fish, such as a swordfish, and a
the genome. Different bits of different genes come together in new cartilaginous fish, such as a shark? What genes might
be different?
combinations, and as a result, new regulatory elements are created.

Plant, fungal, and animal genomes

have unique and shared genes
We now step back farther and consider genomic differences among 24.2 Genome Size
the eukaryotic kingdoms that diverged long before the examples just
discussed. You have already seen that many genes are highly con-
served in animals. Plant genes are also highly conserved, but new
Learning Outcomes
families of genes have emerged at each stage of plant evolution.
1. Differentiate between autopolyploidy and allopolyploidy.
Evolution of plant-specific genes 2. Explain why most crosses between two species do not result
in a new polyploid species.
Sequenced plant genomes offer a window into a billion years of
3. Explain why the genome of a polyploid is not identical to the
evolution. A total of 3814 gene families are shared by every plant,
sum of the two parental genomes.
from the algae to grapes and corn. All the essentials for photosyn-
4. Explain why genome size and gene number do not correlate.
thesis are found in the genome of the single-celled alga, Chlam-
ydomonas reinhardtii, along with a gene that is now used to make
flowers in the flowering plants. Although multicellularity was a Both genome size and gene number vary greatly among the eu-
huge evolutionary step, the multicellular alga Volvox has just about karyote species (see figure 24.1). One contributing factor is whole-
the same number of genes as Chlamydomonas. Lots of extra genes genome duplication, which results in polyploidy, the focus of this
were not required for this major innovation. section. Genome duplication alone, however, cannot explain why
The move onto land is represented by the genome of the rice has about 40,000 genes distributed among 370 Mb of DNA,
moss Physcomitrella patens that arose 450 mya. New genes that while within the about 3400 Mb of the human genome, only about
allowed the plant to survive in an arid environment arose, resulting 20,000 genes are found. To more fully understand the relationship
in 3006 new gene families. Although moss are very different from between gene number and genome size in eukaryotes, the noncod-
flowering plants, making spores instead of seed, for example, they ing DNA must also be considered.
share 80% of the toolkit of developmental genes found in the flow-
ering plant Arabidoposis. The shift from mosses to plants with
Data analysis Estimate and compare the number of
internal transport systems required another 516 gene families. The
genes per mega base-pair of genomic DNA in humans and rice.
transition to flowering plants, which now dominate land-based Which genomes in figure 24.1 are most like humans in terms of
flora, required another 1350 gene families. It is perhaps not the ratio of genes to base-pairs?
surprising that the biggest genomic change in the billion years of

chapter 24 Genome Evolution 509

 Nicotiana sylvestris Nicotiana tomentosiformis
Diploid (SS) Diploid (TT)

ST hybrid
Figure 24.3 Allopolyploidy occurred in
tobacco 5 mya, but can be approximated by crossing
Genome duplication the progenitor species and initiating a doubling of
chromosomes, often through tissue culture followed by
plant regeneration, which can lead to chromosome
Nicotiana tabacum
doubling. Tobacco species have many chromosomes,
Allopolyploid (SSTT)
not all of which are visible in a single plane of a cell.

Ancient and newly created polyploids

guide studies of genome evolution
Average number of gene pairs

2
Polyploidy (three or more chromosome sets) can give rise to new
species, as you learned in chapter 22. Autopolyploids arise by
genome duplication within a single lineage, and allopolyploids
arise by hybridization of two lineages followed by genome
duplication (figure 24.3). Genome duplication is more prevalent genome duplication
in plants than in animals, although animal polyploids are known, 1 loss of duplicate genes
especially in fish and amphibians. In fact, the vertebrate lineage 0
is thought to have undergone two ancient whole-genome dupli- 75 50 25 0
cation events. Time (MYA)
Two avenues of research have led to intriguing insights into
genome alterations following polyploidization. The first approach
studies ancient polyploids, called paleopolyploids. Sequence Figure 24.4 Sequence comparisons of numerous
comparisons and phylogenetic tools establish the time and patterns genes in a polyploid genome tell us how long ago
of polyploidy events (figure 24.4). Sequence divergence between allopolyploidy or autopolyploidy events occurred. Complex
homologues, as well as the presence or absence of duplicated gene analyses of sequence divergence among duplicate gene pairs and
pairs from the hybridization, can be used for historical reconstruc- presence or absence of duplicate gene pairs provide information about
tions of genome evolution. All copies of duplicate gene pairs arising when both genome duplication and gene loss occurred. The graph
through polyploidy are not necessarily found thousands or millions reveals multiple polyploidy events over evolutionary time.
of years after polyploidization. The loss of duplicate genes is con-
sidered later in this section.
The second approach is to create synthetic polyploids by such as asexually propagated commercial bananas, since three sets
crossing plants most closely related to the ancestral species and of chromosomes can’t be evenly divided between two cells. Triploid
then chemically inducing chromosome doubling. Unless the hy- bananas are seedless. The aborted ovules appear as the little brown
brid genome becomes doubled, the plant will be sterile because it dots in the center of any cross section of a banana.
will lack homologous chromosomes that need to pair during
metaphase I of meiosis.
Because meiosis requires an even number of chromosome ? Inquiry question Sketch out what would happen in
meiosis in a 3n banana cell, referring back to chapter 11 if
sets, species with ploidy levels that are multiples of two can repro- necessary.
duce sexually. However, meiosis is problematic in a 3n organism,

510 part IV Evolution

Figure 24.5
Polyploidy has
occurred

Lycopersicon

Arabidopsis
(sugarcane)

(sunflower)
Oryza (rice)
Zea (maize)
Saccharum

Gossypium
Helianthus
numerous times

truncatula
(soybean)

Medicago
Hordeum
Sorghum

(tomato)

Solanum

(lettuce)

Brassica
(potato)

thaliana
(cotton)
(barley)
Triticum

Lactuca
(wheat)

Glycine
in the evolution
of the flowering
plants (the 2.5– 1–2
3–5
numbers indicate 4.5 11
millions of years 11–14
13–15
ago—mya).
25–40

50–70

150–170
Inferred polyploidy event
(MYA) 225–300

While polyploidy is much more frequent in plants, it is not limited

to the plant kingdom. For example, triploid populations of whiptail M. truncatula Soybean
lizards (­Cnemidophorus tesselatus) are found in southeastern
Genome size: Genome size:
Colorado. These lizards are all female, and they reproduce parthe-
315 Mb 974 Mb
nogenetically (see chapter 51). That is, the triploid eggs are able to
produce identical triploid, female offspring through mitosis, with
no meiosis or fertilization needed for reproduction.

Evidence of ancient polyploidy is found

in plant genomes
Polyploidy has occurred numerous times in the evolution of the
flowering plants (figure 24.5). The legume clade that includes the
soybean (Glycine max), the plant Medicago truncatula (a forage
15
legume often used in research), and the garden pea (Pisum sativum)
underwent a major polyploidization event 44 to 58 mya and again
15 mya. 44–58
A quick comparison of the genomes of soybean and
M. truncatula reveal a huge difference in genome size (figure 24.6).
In addition to increasing genome size through polyploidization,
genomes like M. truncatula definitely downsized over evolutionary Polyploidy event (MYA)

time as well. The total size of a genome cannot be explained solely

on the basis of polyploidization. Figure 24.6 Genome downsizing. Genome downsizing must
A number of independent whole-genome duplications in have occurred in Medicago truncatula (the numbers indicate millions of
plants cluster around 65 mya, coinciding with a mass extinction years ago—mya).
event caused by a catastrophic change due to an asteroid impact or
increased volcanic activity. Polyploids may have had a better In some species a great deal of gene loss occurs in the first few
chance of surviving the extinction event with increased genes and generations after polyploidization.
alleles available for selection. Modern tobacco, Nicotiana tabacum, arose from the hybridiza-
tion and genome duplication of a cross between Nicotiana sylvestris
Polyploidy induces elimination (female parent) and N. tomentosiformis (male parent). Researchers
have reconstructed these events and followed chromosome loss in this
of duplicated genes synthetic N. tabacum. Curiously, the loss of chromosomes is not even.
Formation of an allopolyploid from two different species is often More N. tomentosiformis chromosomes are lost than those of N.
followed by a rapid loss of genes (figure 24.7) and rearrangement sylvestris. Similar unequal chromosome loss has been observed in
or loss of chromosomes. In some polyploids, however, loss of one synthetic wheat hybrids. This pattern has also been observed in the
copy of many duplicated genes arises over a much longer period. closely related Arabidopsis thaliana and Arabidopsis lyrata, where A.

chapter 24 Genome Evolution 511

 SCIENTIFIC THINKING Polyploidy can alter gene expression
Hypothesis: Some duplicated genes may be eliminated after polyploidy. A striking discovery is the change in gene expression that occurs in
Prediction: More duplicate genes will be present when an allopolyploid the early generations after polyploidization. Some of this may be
forms than a few generations later. connected to an increase in the methylation of cytosines in the
Test: Make a synthetic polyploid from two Nicotiana species and look DNA. Methylated genes are usually not transcribed, as described
at the chromosomes under a microscope then and after three in chapter 16. Simply put, polyploidization can lead to a short-term
generations of self-fertilizing offspring. silencing of some genes. In subsequent generations, methylation
decreases.
Nicotiana sylvestris Nicotiana tomentosiformis
Diploid Diploid
Transposons are mobilized by polyploidization
Barbara McClintock, who won the Nobel Prize for her discovery
and characterization of transposable elements, hypothesized that
transposons could be mobilized by what she called genome shock.
There are some data on transposon activity following hybridiza-
tion that support this hypothesis. During the early generations
following a polyploidization event, transposition events appear to
increase. Gene mutations arise if a new insertion is in the coding
Polyploidy
region of a gene. Insertions in promoter and other regulatory
regions can change gene expression. Transposition can also result
Nicotiana tabacum in chromosomal rearrangements when a portion of the genome is
Allopolyploid relocated. All of these changes increase genetic variation, poten-
tially leading to rapid evolution.
The relative number of transposable elements in a species
may influence the rate of genome evolution whether or not hy-
bridization occurs. The orangutan’s genome has evolved more
slowly than either of its close relatives—humans and chimps.
Comparisons of the recently sequenced orangutan genome with
chimp and humans revealed many fewer transposable elements in
Duplicate gene loss
the orangutan.

Nicotiana tabacum Polyploidy alone cannot account

for variation in genome size
It is difficult to account for all genome size variation with poly-
ploidy. Even with gene or chromosome elimination following
hybridization, polyploidy is unlikely to result in such a range of
ratios of gene number to genome size. Humans have nine times
the amount of DNA found in the 3.93 × 108-bp pufferfish
Result: Over time, N. tomentosiformis chromosomes are lost or shortened. genome, but about the same number of genes. Plants have an
Conclusion: Chromosomes and genes are preferentially eliminated even greater range of genome sizes. As much as a 200-fold differ-
following polyploidy. ence has been found, yet all these plants weigh in with about
Further Experiments: Why might the chromosomes and genes of one 24,000 to 60,000 genes. Tulips, for example, have 170 times
species be preferentially eliminated? How could you test your explanation? more DNA than Arabidopsis.

Noncoding DNA inflates genome size

Figure 24.7 Polyploidy may be followed by the unequal
loss of duplicate genes from the combined genomes. Why do humans have so much extra DNA compared with puffer-
fish? Much of it appears to be in the form of introns, noncoding
lyrata has 500 more genes than A. thaliana. In the 10,000 years since segments (ncDNA) within a gene’s sequence, that are substantially
the two species diverged, hundreds of thousands of small regions of bigger than those in pufferfish. The Fugu genome has only a hand-
the A. thaliana genome have been deleted. The differential loss may ful of “giant” genes containing long introns; studying them should
be due to different rates of genome replication, as is true for synthetic provide insight into the evolutionary forces that have driven the
human–mouse hybrid cultured cells. change in genome size during vertebrate evolution.
Large expanses of retrotransposon DNA contribute to the
differences in genome size from one species to another. Although
? Inquiry question Why does the number of duplicate
genes decrease after multiple rounds of polyploidization? part of the genome, ncDNA does not contain genes in the usual
sense. As another example, Drosophila exhibits less ncDNA than

512 part IV Evolution

Anopheles, although the evolutionary force driving this reduction better able to tolerate aneuploidy than animals, but the explanation
in noncoding regions is unclear. for this difference is elusive.
The 370-Mb rice genome and the 2-Gb maize genome are
now fully sequenced. Both have about 40,000 protein-coding DNA segments may be duplicated
genes, roughly twice the number of genes in the human genome.
Maize contains lots of repetitive DNA, which has increased its One of the greatest sources of novel traits in genomes is duplica-
DNA content, but not its gene content. Comparisons between the tion of segments of DNA. Two genes within an organism that have
rice, maize, and wheat genomes should provide clues about the arisen from the duplication of a single gene in an ancestor are
genome of their common ancestor and the dynamic evolutionary called paralogs. In contrast, orthologs reflect the inheritance of a
balance between opposing forces that increase genome size (poly- single gene from a common ancestor—a conserved gene.
ploidy, transposable element proliferation, and gene duplication) When a gene is duplicated, the most likely fates of the dupli-
and those that decrease genome size (mutational loss). cate gene are (1) losing function through subsequent mutation
(becomes a pseudogene), (2) gaining a novel function through sub-
sequent mutation (neofunctionalization), and (3) having the total
Learning Outcomes Review 24.2 function of the ancestral gene partitioned into the two duplicates
Autopolyploids arise from duplication of a species’ genome; (subfunctionalization). Gene families grow through gene duplica-
allopolyploids occur from hybridization between two species. tion. In reality, however, most duplicate genes lose function.
Unless the number of chromosomes doubles, an allopolyploid Why, then, do we think that gene duplication is a major
will likely be sterile because of lack of homologous pairs evolutionary force for genes to gain new function? Part of the
at meiosis. Polyploidization can lead to major changes in answer lies in where gene duplication is most likely to occur in
genome structure, including loss of genes, alteration of gene the genome. In humans, the highest rates of duplication lie in the
expression, increased transposon hopping, and chromosomal three most gene-rich chromosomes. Conversely, the seven
rearrangements. Polyploidy is considered important in the chromosomes with the fewest genes show the least amount of
generation of biodiversity and adaptation, especially in plants,
duplication. Note that gene-poor does not necessarily mean less
but it cannot explain why increases or decreases in genome size
do not correlate with the number of genes. Evidently DNA content total DNA.
is not the same as gene content. Polyploidy in plants does not by Even more compelling, certain types of genes appear more
itself explain differences in genome size. Often a greater amount likely to be duplicated: growth and development genes, immune
of DNA is explained by the presence of introns and non-protein- system genes, and cell-surface receptors. Finally, and importantly,
coding sequences than by gene duplicates. gene duplication is thought to be a major evolutionary force for
gene innovation because the duplicated genes are found to have
■■ How might a genome with a small number of genes and
different patterns of gene expression. For example, the two dupli-
a small number of total base-pairs evolve into a genome
with the same small number of genes and a thousand-fold cated copies may be expressed in different or overlapping sets of
larger genome? tissues or organs during development.
About 5% of the human genome consists of segmental dupli-
cations, which are duplications of blocks of genes (figure 24.8).
Segmental duplications may account for differences between
human and chimp. Species-specific segmental duplications tend to
24.3 Evolution Within Genomes contain genes that are differentially expressed between the species.
A simple explanation is that there are more copies of the gene be-
ing expressed, but experimental evidence shows that it is more
likely that the regulatory DNA near the duplicate is different.
Learning Outcomes Both rice and Arabidopsis have higher copy numbers for
1. Define the terms segmental duplication, genome gene families (multiple slightly divergent copies of a gene) than
rearrangement, and pseudogene. are seen in animals or fungi, suggesting that these plants may have
2. Explain why horizontal gene transfer can complicate undergone numerous episodes of segmental duplication during the
evolutionary hypotheses. 150 to 200 million years since rice and Arabidopsis diverged from
a common ancestor. As whole-genome duplications have also
occurred since they diverged, additional analysis will be needed to
Genome evolution can occur at the level of whole-genome dupli- determine the relative effects of polyploidy and segmental duplica-
cation, but also occurs within genomes. Alterations are seen at all tion in plant evolution.
levels from individual genes to chromosomal rearrangements.
Genomes may become rearranged
Individual chromosomes may be duplicated Humans have one fewer chromosome than chimpanzees, gorillas,
Aneuploidy refers to the gain or loss of an individual chromosome and orangutans (figure 24.9). It’s not that we have lost a chromo-
rather than of an entire genome. Failure of a pair of homologous some. Rather, at some point in time, two midsized ape chromosomes
chromosomes or sister chromatids to separate during meiosis is the fused to make what is now human chromosome 2, the second-
most common way that aneuploidy occurs. In general, plants are largest chromosome in our genome.

chapter 24 Genome Evolution 513

Figure 24.8 Segmental duplication on the human Y chromosome. Each orange region has 98% sequence similarity with a
sequence on a different human chromosome. Each dark blue region has 98% sequence similarity with a sequence elsewhere on the Y chromosome.

Figure 24.9 Living great

apes. All living great apes,
Orangutan Gorilla Chimpanzee Hominid
with the exception of humans,
have a haploid chromosome
number of 24. Humans have not
lost a chromosome; rather, two
smaller chromosomes fused to
make a single chromosome.

24 chromosomes 24 chromosomes 24 chromosomes 23 chromosomes

Apes

The fusion leading to human chromosome 2 is an example of 1500 OR genes—about 40% more OR genes than humans. Only 20%
the sort of genome reorganization that has happened in the verte- of the mouse OR genes are pseudogenes, compared to 63% in humans,
brate lineage since the 54-chromosome protovertebrate genome which are inactive pseudogenes (genes that do not produce a func-
described earlier. The remains of some of these rearrangements are tional product due to premature stop codons, missense mutations, or
visible in the structural variation observed between species. deletions). In contrast, half the chimpanzee and gorilla OR genes
Chromosome rearrangements are common, yet over long function effectively, and over 95% of New World monkey OR genes
segments of chromosomes, the linear order of mouse and human appear to be functional. The most likely explanation for these differ-
genes is the same—the common ancestral sequence has been ences is that humans came to rely on other senses, reducing the selec-
preserved in both species. This conservation of synteny (see tion pressure against loss of OR gene function by random mutation.
chapter 18) was anticipated from earlier gene mapping studies, and An older question about the possibility of positive selection
it provides strong evidence that evolution actively shapes the orga- for OR genes in chimps was resolved with the completion of the
nization of the eukaryotic genome. As seen in figure 24.10, the chimp genome. A careful analysis indicated that both humans and
conservation of synteny allows researchers to more readily locate a chimps are gradually losing OR genes to pseudogenes and that
gene in a different species using information about synteny, thus there is no evidence to support positive selection for any of the OR
underscoring the power of a comparative genomic approach. genes in the chimp.

Gene inactivation results in pseudogenes Rearranged DNA can acquire new functions
The loss of gene function is another important way genomes evolve. Errors in meiosis that rearrange parts of genes most often create
Consider the olfactory receptor (OR) genes that are responsible for pseudogenes, but occasionally a broken piece of a gene can end up
our sense of smell. These genes code for receptors that bind odor- in a new spot in the genome where it acquires a new function. One
ants, initiating a cascade of signaling events that eventually lead to of the most intriguing examples occurred within a family of fish in
our perception of scents. the suborder Notothenioidae found in the Antarctic Ocean. These
Gene inactivation seems to be the best explanation for our re- fish are called icefish because they survive the frigid temperatures
duced sense of smell relative to that of the great apes and other mam- in the Antarctic, in part because of a protein in their blood that
mals. Mice have the largest mammalian OR gene family, with about works like antifreeze. Reconstructing evolutionary history using

514 part IV Evolution

 Interchromosomal duplications
Intrachromosomal duplications
Not sequenced
Heterochromatin that is not expressed

Soybean

d2 K c2 b2 c1

2 3

M. truncatula

Figure 24.10 Synteny and gene identification. Genes sequenced in the model legume, Medicago truncatula, can be used to identify
homologous genes in soybean, Glycine max, because large regions of the genomes are syntenic, as illustrated for some of the linkage groups
(chromosomes) of the two species. Regions of the same color represent homologous genes.

comparative genomics reveals that a 9-bp fragment of a gene cod- species seem to have been less firm than they are now and DNA
ing for a digestive enzyme moved to a new location where it more readily moved among different organisms. Earlier in the his-
evolved to encode part of an antifreeze protein. The series of errors tory of life, gene swapping between species was rampant. Today
that gave rise to the new protein persisted only because the change HGT continues in prokaryotes and eukaryotes, but less frequently.
coincided with a massive cooling of the Antarctic waters. Natural An intriguing example of more recent HGT between moss and a
selection acted on this mutation over millions of years. flowering plant is described in chapter 30.

Gene swapping in early lineages

Noncoding DNA can acquire
The extensive gene swapping among early organisms has caused
regulatory functions many researchers to reexamine the base of the tree of life. Early
So far, we have primarily compared genes that code for proteins. phylogenies based on ribosomal RNA (rRNA) sequences indi-
As more genomes are sequenced, we learn that much of the genome cate that an early prokaryote gave rise to two major domains: the
is composed of ncDNA. The repetitive DNA is often retrotranspo- Bacteria and the Archaea. From one of these lineages, the do-
son DNA, contributing to as much as 30% of animal genomes and main Eukarya emerged; its organelles originated as unicellular
40 to 80% of plant genomes. (Refer to chapter 18 for more informa- organisms engulfed specialized prokaryotes.
tion on repetitive DNA in genomes.) Yet there are conserved non- This rRNA phylogeny is being revised as more microbial
coding regions (CNCs) that evolved much more slowly than genomes are sequenced. In 2015, the National Center for Bio-
expected, assuming there was no associated function. An analysis technology Information (NCBI) databases contained 33,543 pro-
of CNCs in 40 vertebrates led to the conclusion that as the rate of karyotic genomes. With new sequencing technology, microbial
base-pair substitution slowed down over evolutionary time, these genomes can be sequenced in less than a day. Phylogenies built
CNCs had begun to regulate the expression of transcription factors, with rRNA sequences suggest that the domain Archaea is more
genes that specifically influence development, and genes that closely related to the Eukarya than to the Bacteria. But as more
encode receptor-binding proteins that are important in signaling. microbial genomes are sequenced, investigators find bacterial and
archaeal genes showing up in the same organism! The most likely
conclusion is that organisms swapped genes, even absorbing DNA
Horizontal gene transfer complicates matters obtained from a food source. Perhaps the base of the tree of life is
Evolutionary biologists build phylogenies on the assumption that better viewed as a web than as a branch (figure 24.11).
genes are passed from generation to generation, a process called
vertical gene transfer (VGT). Hitchhiking genes from other Evidence for gene swapping in the human genome
species, a process referred to as horizontal gene t­ ransfer (HGT) Let’s move closer to home and look at the human genome, which
and sometimes called lateral gene transfer, can lead to phyloge- is riddled with foreign DNA, often in the form of transposons. The
netic complexity. HGT was likely most prevalent very early in the many transposons of the human genome provide a paleontological
history of life, when the boundaries between individual cells and record over several hundred million years.

chapter 24 Genome Evolution 515

 Bacteria Archaea Eukarya 24.4 Gene Function and
Expression Patterns

Learning Outcomes
1. Explain how species with nearly identical genes can look very
different.
2. Describe the action of the FOXP2 gene across species.

s
last
rop Gene function can be inferred by comparing genes in different
C hlo
ria species. When the human genome was first compared to another
o nd
to ch mammalian genome, the mouse, 1000 genes with unknown func-
Mi
tion had a function assigned. One of the major puzzles arising from
comparative genomics is that organisms with very different forms
can share so many conserved genes in their genomic toolkit.
The best explanation for why a mouse develops into a mouse
and not a human is that the same or similar genes are expressed at
different times, in different tissues, and in different amounts and
Figure 24.11 Horizontal gene transfer. Early in the history combinations. For example, the cystic fibrosis gene (cystic fibrosis
of life, organisms may have freely exchanged genes beyond multiple
transmembrane conductance regulator, CFTR), which has been
endosymbiotic events. To a lesser extent, this transfer continues today.
identified in both species and affects a chloride ion channel, illus-
The tree of life may be more like a web or a net.
trates this point. Defects in the human CFTR gene cause especially
devastating effects in the lungs, but mice with the mutant CFTR
Comparisons of versions of a transposon that has duplicated
gene do not have lung symptoms. Possibly variations in expression
many times allow researchers to construct a “family tree” to identify
of CFTR between mouse and human explain the difference in lung
the ancestral form of the transposon. The percentage of sequence
symptoms when CFTR is defective.
divergence found in duplicates allows an estimate of the time at
which that particular transposon originally invaded the human
genome. In humans, most of the DNA hitchhiking seems to have Chimp and human gene transcription
occurred millions of years ago in very distant ancestor genomes.
Our genome carries many more ancient transposons than do
patterns differ
the genomes of Drosophila, C. elegans, and Arabidopsis. One ex- One approach to analyzing transcription patterns in humans and
planation for the observed low level of transposons in Drosophila chimps has been to use RNA sequencing (RNA-seq; see chapter 18)
is that fruit flies somehow eliminate unnecessary DNA from their to measure mRNA from single-cells taken from cerebral organoids.
genome 75 times faster than humans do. Our genome has simply Cerebral organoids are cultured from pluripotent stem cells that are
hung on to hitchhiking DNA more often. induced to differentiate into organized neural tissue that include
The human genome has had minimal transposon activity radial glial cells and excitatory neurons.
in the past 50 million years; mice, by contrast, are continuing to Comparing chimp and human cerebral organoids, investi-
acquire new transposable elements. This difference may explain in gators found 383 genes up-regulated in radial glial cells and
part the more rapid change in chromosome organization in mice 220 genes up-regulated in excitatory neurons. Conversely, 285 genes
than in humans. in radial glial cells, and 165 genes in excitatory neurons, were
down-regulated. A subset of these genes proved to be associated
with human-specific structural variations. While this currently falls
Learning Outcomes Review 24.3 into the category of intriguing possibilities, it highlights one ap-
In segmental duplication, part of a chromosome and the genes it proach to finding functional significance in genomic differences.
contains are duplicated. In genome rearrangement, segments of
chromosomes may change places or chromosomes may fuse with
one another. Pseudogenes have become inactivated in the course
The FOXP2 gene is involved in speech
of evolution but still persist in the genome. All these changes Language, conveyed by complex vocalizations, is one of the defin-
have evolutionary consequences. Horizontal gene transfer has ing characteristics of humans. This has inspired generations of
led to an unexpected mixing of genes among organisms, creating geneticists to search for genes involved in all aspects of language
many phylogenetic questions.
skills. While the use and acquisition of language show high herita-
■■ How would you determine whether a gene was a bility, the search for the source of this variation has produced mixed
pseudogene or an example of horizontal gene transfer? results. One clear exception is the FOXP2 gene, which encodes a
transcription factor expressed in the brain. Individuals heterozygous

516 part IV Evolution

for a mutant allele of FOXP2 show difficulties in learning and se-
quencing the facial movements necessary to produce speech, but not Learning Outcomes Review 24.4
impaired language comprehension. The function of FOXP2 is nec- To understand functional differences between genes shared by
essary for the development of brain circuits in the neocortex, basal species, one must look beyond sequence similarity. Alterations
ganglia, and cerebellum, which are involved in language and fine in the time and place of gene expression can lead to marked
motor control. differences in phenotype. As an example, the FOXP2 gene
Analysis of the evolution of this gene has only added to appears to be involved in sound production in mice, chimps,
interest in its role. In comparisons between humans and mice, gorillas, macaques, humans, and songbirds, and only very small
FOXP2 is among the most conserved 5% of genes. Despite this differences may have led to human speech.
conservation, the rate of amino acid substitution in FOXP2 pro- ■■ How can a single-nucleotide difference in a gene lead to a
tein in humans is increased 50-fold over the average rate in the noticeably different phenotype? Give examples.
genome. It is also found in a region of the genome that has little
archaic DNA. There are three amino acid substitutions in the
human protein compared to the protein in mice. Of these, two
occurred after humans diverged from chimps and bonobos. The
FOXP2 protein is identical in all nonhuman primates examined,
with the exception of a single amino acid substitution in the
orangutan (figure 24.12). 24.5 Applying Comparative
The downstream targets of the FOXP2 transcription factor
have not been characterized, but there are interesting hints that its
Genomics
role in communication may not be limited to humans. Song birds
with the level of FOXP2 protein reduced by RNA interference
show disturbed development of song variability, especially learn- Learning Outcomes
ing by imitation. Mice communicate via squeaks, with lost young 1. Describe how comparative genomics can reveal the genetic
mice emitting high-pitched squeaks. FOXP2 mutations leave mice basis for disease.
squeakless and transgenic baby mice carrying the human FOXP2 2. Explain how genome comparisons between a pathogen and
squeak differently. Both downstream targets of FOXP2 protein and its host can aid drug development.
its role in other species are areas of active research that should 3. Describe how genome comparisons can be useful when
shed further light on the role of FOXP2 in human evolution. working with endangered species.

Comparisons among individual human genomes continue to pro-

Mouse Chimp
p vide information on genetic disease detection and the best course
H
Human of treatment. An even broader array of possibilities arises when
comparisons are made among species. There are advantages to
comparing both closely and distantly related pairs of species, as
Gorilla
well as comparing the genomes of a pathogen and its host. Genome-
level comparisons are also assisting conservation biologists in
developing breeding programs. Examples of the benefits of each
0/5
Rhesus type of genome comparison follow.
2/0
monkey
Orangutan 0/2 0/2 Distantly related genomes offer clues
for causes of disease
Sequences that are conserved between humans and pufferfish pro-
0/7
vide valuable clues for understanding the genetic basis of many
1/2 human diseases. Amino acids critical to protein function tend to be
0/5
preserved over the course of evolution, and changes at such sites
1/131 0/2 within genes are more likely to cause disease.
It is difficult to distinguish functionally conserved sites when
Synonymous changes comparing human proteins with those of other mammals because not
Nonsynonymous changes
enough time has elapsed for sufficient changes to accumulate at non-
conserved sites. A promising exception is the duck-billed platypus
Figure 24.12 Evolution of FOXP2. Beige bars represent (Ornithorhynchus anatinus), which diverged from other mammals
synonymous nucleotide substitutions and brown bars represent about 166 mya and whose genome provides clues to the evolution of
nonsynonymous changes. Numbers indicate nonsynonymous over the immune system. Because the pufferfish genome is only distantly
synonymous nucleotide substitutions unique to each lineage. Humans related to the human genome, conserved sequences are far more eas-
have two amino acid substitutions not found in other primates. ily distinguished than even in the platypus.

chapter 24 Genome Evolution 517

Closely related organisms enhance
medical research
Because studies in humans must be strictly regulated for ethical
reasons, it is much easier to design experiments to identify gene
function in an experimental system like that of the mouse. Com-
paring mouse and human genomes quickly revealed the function of
1000 previously unidentified human genes. The effects of these
genes can be studied in mice, and the results can be used in poten-
tial treatments for human diseases.
A draft of the rat genome has been completed, and even
more exciting news about the evolution of mammalian genomes
may emerge from comparisons of these species. One of the most
exciting aspects of comparing rat and mice genomes is the poten- 2 μm
tial to capitalize on the extensive research on rat physiology, espe-
cially heart disease, and the long history of genetics in mice. Figure 24.13 Plasmodium apicoplast. Drugs targeting
Linking genes to disease has become much easier. enzymes used for fatty acid biosynthesis within Plasmodium
As described in section 24.3, the human genome contains apicoplasts (colored dark green) offer hope for treating malaria.
segmental duplications that are absent in the chimp genome. Some Courtesy of Dr. Lewis G. Tilney and Dr. David S. Roos, University of Pennsylvania

of these duplications contain genes with alleles that produce human

disease. For example, some of the regions duplicated only in humans
and its relatives, appears to be derived from a chloroplast appropri-
correspond to regions implicated in Prader–Willi syndrome and
ated from algae engulfed by the parasite’s ancestor (figure 24.13).
spinal muscular atrophy. The significance of these observations is
Analysis of the Plasmodium genome reveals that about 12%
not clear, but there is great interest in analyzing regions that differ
of all the parasite’s proteins, encoded by the nuclear genome, head
between humans and closely related species.
for the apicoplast. These proteins act there to produce fatty acids.
The apicoplast is the only location in which the parasite makes the
Pathogen–host genome differences fatty acids, suggesting that drugs targeted at this biochemical path-
reveal drug targets way might be very effective against malaria.
With genome sequences in hand, pharmaceutical researchers are Another disease-prevention possibility is to look at chloroplast-
more likely to find suitable drug targets to eliminate pathogens specific herbicides, which might kill Plasmodium by targeting the
without harming the host. Diseases—including malaria and Cha- chloroplast-derived apicoplast. Coupled with a newer vaccine, this
gas disease—in many developing countries have both human and treatment could substantially reduce the incidence of malaria.
insect hosts. Both of these infections are caused by protists (see
chapter 28), and the value of comparative genomics in drug dis- Chagas disease
covery to treat them is illustrated later in this section. Trypanosoma cruzi, an insect-borne protozoan, kills about 21,000
people in Central and South America each year. As many as
Malaria 18 million suffer from this infection, called Chagas disease, the
Anopheles gambiae, the malaria-carrying mosquito, along with symptoms of which include damage to the heart and other internal
Plasmodium falciparum, the protistan parasite it transmits, togeth- organs. Genome sequencing of T. cruzi was completed in 2005.
er have an enormous effect on human health, resulting in 1.7 to A surprising and hopeful finding is that a common core of
2.5 million deaths each year from malaria. The genomes of both 6200 genes is shared among T. cruzi and two other insect-borne
Anopheles and Plasmodium were sequenced in 2002. pathogens: Trypanosoma brucei and Leishmania major. T. brucei
P. falciparum, which causes malaria, has a relatively small causes African sleeping sickness, and L. major infections result in
genome of 2.33 × 107 bp that proved very difficult to sequence. It lesions of the skin of the limbs and face. These core genes are be-
has an unusually high proportion of adenine and thymine, making ing considered as possible targets for drug treatments.
it hard to distinguish one portion of the genome from the next. The Currently, no effective vaccines and only a few drugs with
project took five years to complete. P. falciparum appears to have limited effectiveness are available to treat any of these diseases.
about 5500 genes, with those of related function clustered togeth- The genomic similarities may not only aid in targeting drug de-
er, suggesting that they might share the same regulatory DNA. velopment, but perhaps also result in a treatment or vaccine that
P. falciparum is a particularly crafty organism that hides is effective against all three devastating illnesses (figure 24.14).
from our immune system inside red blood cells, regularly changing
the proteins it presents on the surface of the red blood cell. This Genome comparisons inform
“cloaking” has made developing a vaccine or other treatment for
malaria particularly difficult.
conservation biology
Recently, a link to chloroplast-like structures in P. falci- Genomes of species on the brink of extinction are being mined for
parum has raised other possibilities for treatment. An odd subcel- information that could contribute to disease reduction and con­
lular component called the apicoplast, found only in Plasmodium servation efforts. Here we consider the application of genomics

518 part IV Evolution

Figure 24.14 Comparative Chagas disease African sleeping sickness Leishmania infection
genomics may aid in drug
development. The organisms that
cause Chagas disease, African
sleeping sickness, and leishmaniasis,
which claim millions of lives in
developing nations each year, share
6200 core genes. Drug development
targeted at proteins encoded by the
shared core genes could yield a 2 μm 3 μm 5 μm
single treatment for all three
diseases. (left) Eye of Science/Science
Source; (middle) David Spears FRPS FRMS/
Corbis Documentary/Getty Images; 6200 shared core genes
(right) KATERYNA KON/Science Photo
Library/Alamy Stock Photo

Target for drug development

to three endangered species, the Tasmanian devil (Sarcophilus panda diet, and habitat destruction is a factor in the decline of pan-
harrisii), the giant panda (Ailuropoda melanoleura), and the polar das. Curiously, the panda’s relatives are carnivores, and pandas
bear (Ursus maritimus). maintain the genes associated with carnivores. Furthermore, they
The Tasmanian devil, a marsupial on the Australian island of lack the genes coding for enzymes needed to fully digest bamboo.
Tasmania, faces decimation from devil facial tumor disease Attention is now focused on the microbes in the panda’s gut that
(DFTD, figure 24.15). Close to 90% of the population is affected, digest bamboo. Collectively, this information may assist conserva-
and since 1996, 60% of the devils have been wiped out due to tionists in keeping the population from dipping below its current
DFTD. A comparison of genomes from two devils, named Cedric level of 2500 to 3000 individuals.
and Spirit, from distant corners of Tasmania revealed extremely An unexpected finding arose when the mitochondrial ge-
low genetic diversity that can be traced back 100 years before the nomes of modern and fossil polar and brown bears (including
DFTD outbreak. Only the now extinct Tasmanian tiger had less DNA from the jaw of a 110,000- to 130,000-year-old polar bear)
genetic diversity (­figure 24.16). Breeding programs can use the were sequenced. Polar bears evolved about 150,000 years ago. Ge-
genomic information to preserve what diversity there is in the neticists were shocked to find that the entire maternal line of polar
population. bears can be traced back to a brown bear living in Ireland between
Sequencing of the genome of the giant panda offered more 20,000 and 50,000 years ago. Despite the close relation to brown
promising news about population diversity (figure 24.16). The bears and the ability to interbreed, this finding does not offer much
panda has 2.7 million SNPs, which is three times as many as were hope for the polar bears facing extinction as warmer temperatures
identified in the Tasmanian devil. Bamboo is the mainstay of a melt their sea ice homes. Average temperatures for the next
50 years are predicted to be much warmer than anything polar
bears have experienced in their 150,000 years on Earth or the
20,000 or so years since they mated with brown bears.

Learning Outcomes Review 24.5

DNA sequences conserved over evolutionary time tend to be
those critical for protein function and survival. Variations in
these conserved sequences may provide clues to diseases with
a hereditary component. Knowledge of a pathogenic organism’s
genome and its differences from a host’s genome may allow
targeting of drugs and vaccines that affect the invader but leave
the host unharmed. Comparative genomics within species can
provide information about population diversity of endangered
species. Conservation biologists can use this information to
develop breeding strategies.
■■ A pathogen makes a critical protein that differs from the
human version by only seven amino acids. What approaches
Figure 24.15 DFTD is a fatal condition for Tasmanian might lead to an effective drug against the pathogen? What
devils. Tumors interfere with eating, and devils die within months of drawbacks might be encountered?
the first appearance of lesions. Dave Watts/Alamy Stock Photo

chapter 24 Genome Evolution 519

 Figure 24.16 Genetic
100 diversity of mitochondrial
Endangered
90
genomes based on average
92 Extinct
Not endangered number of differences
80 85 between pairs of individuals
77 in the populations indicated.
Genetic diversity within species

70
67
60

50

40 44 43

30
8
28
25 25
20
20 19
15
10
10
5
0
southern Africa

Wolf

Panda

Polar bear

Mammoth
clade II

Whale

Mammoth
clade I

Brown bear

European
human

Stellar sea lion

Bison

Tasmanian
devil
Tasmanian
tiger
Gorilla

Bushman of

Chapter Review

Stephen Robinson/NHPA/Photoshot/Newscom

24.1 Comparative Genomics Plant, fungal, and animal genomes have unique and shared
genes.
Evolution of the vertebrate genome. Plant genomes have changed much more rapidly than animal genomes.
The human genome has been compared to many other vertebrates. It appears that about one-third of plant genes are unique to plants.
This has led to identification of vertebrate-specific genes. This Of the remaining plant genes, many are also found in animal and
can be taken farther to hypothesize a protovertebrate karyotype of fungal genomes and are required for metabolism and gene
54 chromosomes. expression.

The human genome shares similarities with those of 24.2 Genome Size
great apes.
Human genomes are very similar to the genomes of great apes. Primate Ancient and newly created polyploids guide studies of genome
genomes share the feature of a large amount of repetitive DNA, evolution.
including transposable elements. These transposons show species- Autopolyploidy results from an error in meiosis that leads to a duplicated
specific expansion and deletion. These comparisons are beginning to genome; allopolyploidy is the result of hybridization between species
find possible functional differences. (figure 24.3).
The human genome contains sequences from extinct Evidence of ancient polyploidy is found in plant genomes.
primates.
Polyploidy has occurred numerous times in the evolution of flowering
The human genome contains sequences that are the result of ancient plants, and downsizing of genomes is common.
hybridization events with Neanderthal and Denisovans, two extinct
primates. Analyzing these differences shed light on human Polyploidy induces elimination of duplicated genes.
evolution. Downsizing of a polyploid genome can be caused by unequal loss of
duplicated genes (figure 24.7).
Genomes evolve at different rates.
Viral, bacterial, and even insect genomes evolve more rapidly than Polyploidy can alter gene expression.
mammalian genomes. Plant genomes change more quickly than animal Polyploidization can lead to short-term silencing of genes via
genomes, possibly as the result of massive genome remodeling from methylation of cytosines in the DNA.
extensive transposition of mobile elements.

520 part IV Evolution

Transposons are mobilized by polyploidization. Noncoding DNA can acquire regulatory functions.
Transposons become highly active after polyploidization; their insertions DNA that has no function in an organism can acquire mutations without
into new positions may lead to new phenotypes. any detrimental effect. However, conserved noncoding regions (CNCs)
evolved more slowly than expected. These CNC regions appear to have
Polyploidy alone cannot account for variation in genome size. gained functions that regulate the expression of neighboring protein-
The vast range of ratios of gene number to genome size, even in closely coding genes.
related species, cannot be explained by polyploidy alone.
Horizontal gene transfer complicates matters.
Noncoding DNA inflates genome size. Horizontal gene transfer creates many phylogenetic questions, such as
Genome size is most often inflated due to the presence of introns and the origins of the three major domains (figure 24.11).
non-protein-coding sequences. Genome size does not correlate with the
number of genes. Some species have very little noncoding DNA, and 24.4 Gene Function and Expression Patterns
others have extensive amounts of ncDNA, often retrotransposon DNA.
Two species, such as rice and maize, can vary substantially in the Chimp and human gene transcription patterns differ.
amount of ncDNA and yet have similar numbers of genes. Even when species have highly similar genes, expression of these genes
may vary greatly. Posttranscriptional differences may also contribute to
24.3 Evolution Within Genomes species differences.

Individual chromosomes may be duplicated. The FOXP2 gene is involved in speech.

Aneuploidy, the duplication or loss of individual chromosomes, Small evolutionary changes in the FOXP2 protein and its expression may
results from errors in meiosis. It is tolerated better in plants than in have influenced the evolution of human speech (figure 24.12).
animals.
24.5 Applying Comparative Genomics
DNA segments may be duplicated (figure 24.8).
Paralogs are duplicated ancestral genes; orthologs are conserved Distantly related genomes offer clues for causes of disease.
ancestral genes. Duplicated DNA is common for genes associated with Changes in amino acid sequences of critical proteins are a likely
growth and development, immunity, and cell-surface receptors. cause of diseases, and these differences can be identified by genome
comparison.
Genomes may become rearranged.
Genomes may be rearranged by moving gene locations within a Closely related organisms enhance medical research.
chromosome or by the fusion of two chromosomes. By comparing related organisms, researchers can focus on genes that
Conservation of synteny refers to preservation of long segments cause diseases and devise possible treatments.
of ancestral chromosome sequences identifiable in related species Pathogen–host genome differences reveal drug targets.
(figure 24.10).
Analysis of the genomes of pathogenic organisms may provide new
Gene inactivation results in pseudogenes. avenues of treatment and prevention (figure 24.14).
Some ancestral genes become inactivated as they acquire mutations and Genome comparisons inform conservation biology.
are termed pseudogenes.
Comparative genomics within species can provide information about
Rearranged DNA can acquire new functions. population diversity of endangered species. Conservation biologists can
use this information to develop breeding strategies.
Occasionally part of a gene can end up in a new spot in the genome
where its function changes.

chapter 24 Genome Evolution 521

Visual Summary

Genomic analysis

includes

Comparative Genome
reveals Genome size Functional genomics
genomics evolution

informs Rate of acquiring

complicated by
evolution increases affected new function examining
due to by
more slower
Phylogeny
variation Horizontal gene Species-specific
Transposable Rearranged
Animal Duplication transfer genes
less elements DNA
Plants variation faster whole and applications
and and can in
genome
Plants occur by
Animal Noncoding Species-specific
Polyploidy Noncoding DNA expression
identified DNA
orthologs often changes in
Polyploidy
followed by Transcription
Vertebrates Loss of genes activity
Aneuploidy Medicine
duplications, fusions, Gene
Rearrangements expression
rearrangements Chromosome
fusion Timing Conservation
Loss of differences biology
Amniotes
chromosomes Segmental in
transposable duplication Tissues
elements
loss of Amounts
Pseudogene
Primates function

522 part IV Evolution

 Review Questions

Stephen Robinson/NHPA/Photoshot/Newscom
U N D E R S TA N D c. It cannot be explained with current genetic theory.
1. Humans and pufferfish diverged from a common ancestor about d. The differences are caused by random effects during
450 mya, and these two genomes have development.
a. very few of the same genes in common. 2. You are offered a summer research opportunity to investigate a
b. all the same genes. region of ncDNA in maize. A friend politely smiles and says that
c. a large proportion of the genes in common. only graduate students get to work on the coding regions of DNA.
d. no nucleotide divergence. How would you critique your friend’s statement?
2. Genome comparisons have suggested that mouse DNA has mutated a. The friend has a point; ncDNA is “junk” DNA and therefore
about twice as fast as human DNA. What is a possible explanation not very important.
for this discrepancy? b. The ncDNA produces protein through mechanisms other than
transcription.
a. Mice are much smaller than humans.
c. Most ncDNA is usually translated.
b. Mice live in much less sanitary conditions than humans and
d. Often ncDNA produces RNA transcripts that themselves have
are therefore exposed to a wider range of mutation-causing
regulatory function.
substances.
c. Mice have a smaller genome size. 3. Analyze the conclusion that the Medicago truncatula genome has
d. Mice have a much shorter generation time. been downsized relative to its ancestral legume, and circle the
evidence that is consistent with this conclusion.
3. Polyploidy in plants
a. Medicago has a proportional decrease in the number
a. has only arisen once and therefore is very rare.
of genes.
b. only occurs naturally when there is a hybridization event
b. Medicago has a proportional increase in the number
between two species.
of genes.
c. is common, but never occurs in animals.
c. Medicago has an increase in the amount of DNA.
d. is common, and does occur in some animals.
d. Medicago has a decrease in the amount of DNA.
4. Homologous genes in distantly related organisms can often be
4. Analyze why an herbicide that targets the chloroplast is effective
easily located on chromosomes due to
against malaria.
a. horizontal gene transfer.
a. Because Plasmodium needs a functional apicoplast
b. conservation of synteny.
b. Because the main vector for malaria is a plant
c. gene inactivation.
c. Because mosquitoes require plant leaves for food
d. pseudogenes.
d. Because Plasmodium mitochondria are very similar to
5. All of the following are believed to contribute to genomic diversity chloroplasts
among various species, except
a. gene duplication.
b. gene transcription. SYNTHESIZE
c. lateral gene transfer. 1. The FOXP2 gene is associated with speech in humans. It is also
d. chromosomal rearrangements. found in chimpanzees, gorillas, orangutans, rhesus macaques, and
6. What is the fate of most duplicated genes? even the mouse, yet none of these mammals speak. Develop a
hypothesis that explains why FOXP2 supports speech in humans
a. Gene inactivation but not other mammals.
b. Gain of a novel function through subsequent mutation
c. They are transferred to a new organism using lateral gene transfer. 2. One of the common misconceptions about sequencing projects
d. They become orthologues. (especially the high-profile Human Genome Project) is that
creating a complete road map of the DNA will lead directly to
A P P LY cures for genetically based diseases. Given the percentage
similarity in DNA between humans and chimps, is this simplistic
1. Chimp and human DNA whole-genome sequences differ by about
view justified? Explain.
1.23%. Determine which of the following explanations is most
consistent with the substantial differences in morphology and 3. How does horizontal gene transfer (HGT) complicate phylogenetic
behavior between the two species. analysis?
a. It must be due largely to gene expression.
b. It must be due exclusively to environmental differences.

chapter 24 Genome Evolution 523