class="bi x0 y0 w0 h1"
Evolutionary Ecology
This page intentionally left blank
Evolutionary Ecology
Concepts
and
Case
Studies
Edited
by
CHARLES
W.
FOX,
DEREK
A.
ROFF,
AND
DAPHNE
J.
FAIRBAIRN
OXPORD
UNIVERSITY
PRESS
2001
OXFORD
UNIVERSITY
PRESS
Oxford
New
York
Athens Auckland Bangkok Bogota Buenos Aires Cape Town

York,
New
York 10016
Oxford
is a
registered trademark
of
Oxford University Press.
All
rights reserved.
No
part
of
this publication
may be
reproduced,
stored
in a
retrieval system,
or
transmitted,
in any
form
or by any
means,
electronic, mechanical, photocopying, recording,
or
otherwise,
without
the

0-19-513154-1;
0-19-513155-X
(pbk.)
1.
Ecology.
2.
Evolution
(Biology)
I.
Fox, Charles
W. II.
Roff,
Derek
A.,
1949-
.
III.
Fairbairn, Daphne
J.
QH541
.E86
2001
577—dc21
00-053758
Cover art:
A
female broad-tailed hummingbird
(Selasphoms
platycercus)
pollinating Delphinium

its
functional basis. Within this com-
mon
framework, evolutionary biologists empha-
size
historical
and
lineage-dependent processes
and
hence
often
incorporate phylogenetic reconstruc-
tions
and
genetic models
in
their analyses.
Ecolo-
gists, while cognizant
of
historical processes, tend
to
explain variation
in
terms
of the
contemporary
effects
of
biotic

to
variation among com-
munities
or
major
taxonomic groups.
The
overlap
between evolutionary ecology
and
ecology
is so
broad
that
some
previous
treatments
of the
field
(e.g.,
Pianka 1994) have little
to
distinguish them
from
standard ecology textbooks. However, recent
advances
in
molecular genetics, quantitative genet-
ics
(e.g., multivariate models, analyses

tested, rather than
a
priori
as-
sumptions.
As the
chapters
in
this volume
attest,
contemporary evolutionary ecologists have assem-
bled
a
very diverse
and
effective
array
of
tech-
niques
and
approaches
to
test these hypotheses.
Our
primary
objective
in
organizing this book
was to

we
recognized
the
need
for
such
a
volume
and
discov-
ered that many
of our
colleagues, including some
of
the
contributors
to
this volume,
felt
the
same
way.
We
hope
that
this
book
will
fill
this

this audi-
ence. When writing
the
first
part
of the
book, enti-
tled "Recurring Themes," authors have assumed
that
students
have
the
equivalent
of at
least
one
undergraduate course
in
ecology
and one
course
in
genetics,
but
they have
not
assumed
any
back-
ground

II-V),
we
assume
that
stu-
dents have
a
basic understanding
of the
evolution-
ary
processes
and
concepts discussed
in
part
I.
Authors
in all
sections have also assumed that stu-
dents have read
the
preceding chapters
in the
vol-
ume, allowing chapters
to
build
on
each other

ensure
that
a
reference
to the
appropriate preceding chapter
is
included.
The
chapters
in
this volume have each been
written
by a
different
author;
all
authors
are
lead-
ing
researchers
in
their
field.
Chapters thus repre-
sent
the
current stage
of

team-taught lecture course presents
challenges.
Authors vary
in the
level
at
which they
present their material
and in the
amount
of
back-
ground
that
they expect students
to
have when
reading their chapter. Authors also vary
in
their
writing styles
and
vary somewhat
in the way
that
they
organize their chapters.
We
have attempted
to

another. Perhaps
our
major
chal-
lenge
as
editors
was to
keep
the
volume
to a
rea-
sonable length, given
28
independently written
chapters. Each author
was
asked
to
contribute
no
more than 8000 words
of
text,
using
no
more than
six
figures,

full
responsi-
bility
for the
resulting (and necessary) omission
of
many
additional references, perhaps equally appro-
priate
as
examples
or
case studies.
We
hope
that
readers will
be
inspired
to
delve more
fully
into
at
least
some
of the
research areas
and
will thus have

second
objective—to
pro-
duce
a
volume that
is
valuable
to all
researchers
in
ecology, evolution,
and
genetics.
It is
largely
for
this reason
that
we
opted
for a
multiauthored vol-
ume
rather than
a
traditional textbook style. This
volume
is a
collection

effective
updates
and
insights that will,
we
hope, encourage cross-fertilization.
Evolutionary
ecology
is a
very broad
and di-
verse
field
that
includes much
of
modern ecology
and
evolutionary biology. Unfortunately,
we
have
only
one
volume within which
to
cover
the
field.
We
have tried

most
substantial decisions involved what topics should
be
left
out of the
book. Undoubtedly,
our
personal
interests
and
biases have influenced some
of
these
decisions,
but
most omissions
are for
practical rea-
sons.
For
example,
we
have opted
not to
include
chapters
on
speciation because numerous edited
volumes
have been dedicated

as
chapters
on
molecular meth-
ods, methods
of
phylogenetic analysis,
or
methods
for
measuring genetic variance components. These
are
important techniques
for us all to
understand
but are
best acquired
from
specialized volumes
(e.g.,
Brooks
and
McLennan 1991; Harvey
and Pa-
gel
1991; Avise 1994;
Roff
1997).
Instead,
we fo-

gratitude
to all of the
authors contributing
to
this volume. Writing
a
book chapter
is
often
a
thankless task,
and our
stringent requirements have made
the
task espe-
cially
difficult.
We
have been
uniformly
impressed
not
only
by the
very
high quality
of the
contribu-
tions,
but

as a
whole.
To the
extent
that
we
have succeeded
in
this,
and in our
overall goal
of
providing
a
state-of-the-art introduction
to
evo-
lutionary ecology,
we
must thank
the
individual
chapter authors.
Charles
W.
Fox
Derek
A.
Roff
Daphne

Damuth
3.
Natural Selection
29
Daphne
J.
Fairbairn
Jeff
P.
Reeve
4.
Adaptation
44
David Reznick
Joseph
Travis
5.
Phenotypic
Plasticity
58
Massimo Pigliucci
6.
Population Structure
70
Leonard Nunney
7.
Inbreeding
and
Outbreeding
Nickolas

128
Marc Tatar
11.
Life
Cycles
142
Jan A.
Pechenik
99
113
12. Sex and
Gender
Turk
Rhen
David
Crews
154
13. Sex
Ratios
and Sex
Allocation
Steven
Hecht
Orzack
14.
Ecological Specialization
and
Generalization
177
Douglas

232
Donald
L.
Kramer
19.
The
Evolutionary Ecology
of
Movement
247
Hugh
Dingle
Marcel
Holyoak
Part
IV.
Interspecific Interactions
20.
Ecological
Character
Displacement
265
Dolph
Schluter
21.
Predator-Prey Interactions
277
Peter
A.
Abrams

Anthropogenic
Change
26.
Pesticide Resistance
347
John
A.
McKenzie
27.
Predicting
the
Outcome
of
Biological
Control
361
Judith
H.
Myers
28.
Evolutionary Conservation Biology
371
Philip
W.
Hedrick
References
385
Index
415
Contributors

of
Ecology
and
Evolutionary
Biology, University
of
Arizona,
Tucson,
Arizona
85721
USA.
Crews,
David.
Section
of
Integrative Biology,
University
of
Texas,
Austin, Texas
78712
USA.
Damuth,
John.
Department
of
Ecology, Evolution
and
Marine Biology, University
of

Entomology,
S-225 Agricultural Science Center
North,
University
of
Kentucky, Lexington, Kentucky
40546-0091
USA.
Futuyma, Douglas
J.
Department
of
Ecology
and
Evolutionary Biology, State University
of New
York, Stony Brook,
New
York
11794-5245
USA.
Hedrick,
Philip
W.
Department
of
Biology,
Arizona State University,
Tempe,
Arizona

Biology,
McGill University, 1205 Avenue Docteur
Penfield,
Montreal, Quebec
H3A 1B1
Canada.
Mazer,
Susan
J.
Department
of
Ecology,
Evolution
and
Marine Biology, University
of
California,
Santa Barbara, California
93106
USA.
McKenzie,
John
A.
Center
for
Environmental
Stress
and
Adaptation
Research,

V6T 1Z4
Canada.
Nunney, Leonard. Department
of
Biology,
University
of
California, Riverside, California
92521
USA.
Orzack,
Steven
Hecht.
Fresh Pond Research
Institute,
64
Fairfield
St., Cambridge,
Massachusetts
02140
USA.
Pechenik,
Jan A.
Department
of
Biology, Tufts
University, Medford, Massachusetts
02155
USA.
Pigliucci,

California, Riverside, California
92521
USA.
Rhen,
Turk.
Laboratory
of
Signal Transduction,
National Institute
of
Environmental Health
Sciences,
National
Institute
of
Health,
Research
Triangle Park,
North
Carolina
27709
USA.
Roff,
Derek
A.
Department
of
Biology, University
of
California,

Zoology,
University
of
British Columbia,
6270
University
Blvd.,
Vancouver,
British Columbia,
V6T 1Z4
Canada.
Tatar,
Marc.
Department
of
Ecology
and
Evolutionary Biology, Brown University,
Box
G-W,
Providence, Rhode Island
02912
USA.
Thompson,
John
N.
Department
of
Ecology
and

USA.
Westneat,
David
F.
Center
for
Ecology, Evolution
and
Behavior, School
of
Biological Sciences,
101
Morgan Building, University
of
Kentucky,
Lexington, Kentucky
40506-0225
USA.
Williams, Charles
F.
Department
of
Biological
Sciences, Idaho State University, Pocatello, Idaho
83209
USA.
Wilson,
David Sloan. Department
of
Biological

core
the
study
of
variation within individuals, among
in-
dividuals,
among populations,
and
among species.
For
several reasons, evolutionary ecologists need
to
know
the
causes
and the
effects
of
variation
in
traits
that
influence
the
performance, behavior,
longevity,
and
fertility
of

of a
trait
is
determined
by
the
genetic constitution
(or
genotype)
of an
indi-
vidual
and by the
environment
in
which
an
individ-
ual
is
raised. Second,
to
predict whether
and how
natural selection will cause
the
mean phenotype
of
a
trait

to
understand
why
the
phenotype
of a
given trait
influences
an
indi-
vidual's
fitness,
we
need
to
know
how the
trait
af-
fects
an
individual's ability
to
garner resources
for
growth
or
reproduction,
to
avoid predation,

must understand
how
different
phenotypes perform under
different
environmental
conditions.
In
sum, with
an
understanding
of the
causes
and
consequences
of
phenotypic variation
within
and
among populations,
we can
detect evo-
lutionary processes operating
at a
variety
of
eco-
logical levels: within random-mating populations;
within
and

evolutionary ecology
has
increased
in
direct
proportion
to our
understand-
ing of the
multiple causes
of
intraspecific
pheno-
typic
variation.
Before
reviewing these sources
of
variation,
it is
worth
considering
briefly
a
funda-
mental question:
What
kind
of
variation

determined
by
nuclear genes
operating
in an
additive
manner
(i.e.,
alleles
whose
effects
are
inde-
pendent
of the
genetic background
in
which they
are
expressed)
are
most
likely
to
fulfill
this crite-
rion. Indeed,
the
importance
of

what about traits
that
are
partly
or
largely
3
4
Recurring Themes
influenced
by
nonnuclear genes, nonadditive inter-
actions among
alleles
or
loci,
the
maternal environ-
ment,
an
individual's current environment,
the in-
teraction between
an
individual's genotype
and its
environment,
or the age or
developmental stage
of

to
these
effects
can be
difficult
to
predict.
The
rate
or
direction
of
their evolution
can
depend
on
the
degree
and
nature
of
population
structure,
in-
teractions among individuals,
and
nonrandom
mating.
In
addition, genetic

important
not
only
to
determine whether traits
are
transmitted
from
parents
to
offspring,
but
whether they
are
transmitted
in a
predictable fashion.
An
under-
standing
of all
potential sources
of
phenotypic
variation
in a
trait
helps
to
achieve

1 and
2
consider
the
causes
and
evolutionary conse-
quences
of
variation
in
both unstructured
and
structured populations,
and we
highlight
our
view
that
new
insights into
the
potential
for
natural
se-
lection
to
cause phenotypic change
in

being able
to
determine
the
quantitative relationships between
phenotype, genotype,
and
fitness.
If the
phenotypic
variation
in a
trait
is
genetically based
and
corre-
lated with
the
fitness
of
individuals
in a way
that
can be
expressed mathematically, then
it is
possible
to
predict

measure
this relationship.
Discrete
Traits
Traits
whose phenotypes
can be
assorted into dis-
tinct, nonoverlapping classes exhibit discrete varia-
tion.
The
phenotypic
frequency
distributions
of
such
categorical
or
qualitative traits
are
usually
de-
picted
as
histograms,
where
the
phenotypic catego-
ries
are

Mendel's peas
and the
wing
color
of the
peppered moth
(Biston
betularia)
pro-
vide excellent
if
time-worn examples
for
introduc-
tory biology students, many other discretely inher-
ited
traits provide evidence
for the
potential
for (or
the
limitations
of)
natural selection
to
mold genetic
variation.
When
the
frequencies

morphs
may
be
identical with respect
to
both survivorship
and
reproduction,
or the
morphs
may
enjoy
equal
fitnesses
because where one, say,
has an
advantage
in
fertility,
another
has an
advantage
in
survivor-
ship. Other possibilities
are
that each
morph
enjoys
a

polymorphisms
in
natural
populations
is a
chal-
lenge
for
evolutionary ecologists,
and
numerous
studies have aimed
to do so.
In
many animals,
for
example, body color
is a
discrete trait that varies among individuals
and af-
fects
their vulnerability
to
their predators.
For ex-
ample,
the
adder,
Viperus
berus, exhibits

eastern
Ontario
population
of the
tristylous perennial plant,
Decodon
vertidllatus
(Lythraceae). This population exhibits
a
marked deficiency
of
mid-style morphs
that
is
persistent over time (Eckert
and
Barrett 1995).
(B)
Mean frequencies
of
style morphs
in the
tristylous
perennial
herb
Lythrum
salicaria
(purple loosestrife; Lythraceae) sampled from
populations
in

and
Ericson 1996).
N
represents
the
number
of
populations
sampled from each region.
(C)
Diploid genotype frequencies
for the
malate dehydrogenase locus sampled
in
three Eastern Ontario populations
of
Decodon vertidllatus (Eckert
and
Barrett
1993).
The
genotypic
frequencies
do not
differ
significantly from
Hardy-Weinberg
expectations.
(D)
Frequencies

on
body color patterns
in
this
sample
(King
1987).
6
Recurring Themes
ulations
in and
near Sweden, Forsman (1995) found
that
the
relative performance
of
these
two
morphs
differed
between males
and
females.
In
male snakes,
the
black form
suffered
lower survival (apparently
due

gesting that
the
fitness
of a
given morph depends
on
the
behavior
of the
individual exhibiting
it.
Individ-
uals
of the
Lake Erie water snake,
Nerodia
sipedon
insularum,
also exhibit qualitative variation
in
body
color pattern
(King
1987).
Differences
in the
relative
abundances
of
banded

assess
than visible polymorphisms,
but
several
studies
suggest
that
selection acts
on
allozymes
1
(or
on
closely linked genes)
that
appear
to
influence
fitness
through
their
physiological
effects.
Carter
(1997)
found
that
among populations
of
gilled

ponds. This geographic pattern
is
consistent
with environment-specific selection under controlled
conditions
and
with
temporal patterns
of
selection
within ponds; relative
to the
Adh-FF
(Fast/Fast
ho-
mozygotes)
and
Adh-FS
(Fast/Slow
heterozygotes)
genotypes, Adh-SS genotypes
are
selected against
under low-oxygen
conditions.
Discrete traits controlled
by one or a few
loci
are of
special interest

testing evo-
lutionary
hypotheses.
Quantitative
Traits
In
contrast
to
discrete traits
are
those
for
which
the
phenotype
varies along
a
continuum
and is
deter-
mined
by
alleles
at
multiple loci.
The
frequency
distribution
of
such quantitative

1.2).
(Indeed,
the
width
of the
categories
can
have
a
very
strong
effect
on the
shape
of the
resulting distribu-
tion.)
The
y-axis
shows
the
proportion
or
number
of
all
sampled individuals whose "phenotypic
value"
for the
trait

curve.
Where
the
value
of a
trait must
be
expressed
as an
integer (e.g.,
the
number
of
anthers
per
flower),
the
boundaries between
the
categories
are
less
arbitrary,
but the
trait
may
nevertheless behave
as
a
quantitative trait

locus
is
independent
of
both
the
other allele
at
that locus
and the
genotypes expressed
at
other loci. Alterna-
tively, alleles
and
loci
may
interact
so
that
the ef-
fect
on
phenotype
of an
allelic change
at a
locus
depends
on the

the
similarity between parents
and
offspring
or
the
phenotypic response
to
selection.
Many evolutionary ecologists focus exclusively
on the
evolution
of
quantitative traits simply because
so
many traits
of
known ecological importance
and
with strong
effects
on
fitness
are
continuously
distributed
or
known
to
behave

photosynthetic rate
or
meta-
bolic
efficiency;
and
fitness-related morphological
traits, such
as
size
at
birth
and
adulthood, bill size
Nature
and
Causes
of
Variation
Figure
1.2
Frequency distributions
of
quantitative
traits found
in a
greenhouse-raised experimental
population
of the
annual salt marsh plant

mean area
of all
petals produced
by
a
flower.
N=
1179
individuals
for all
traits.
The
normal distribution corresponding
to the
mean
and
variance
of
each
trait
is
superimposed
on the
histogram
of the
actual
frequency
distribution
of
phenotypic values.

known
(figure
1.3; Falconer
and
MacKay
1996).
This means that
it is a
simple matter
to ask
wheth-
er
population
or
species means
differ
significantly,
potentially
reflecting
the
direct outcome
of
evolu-
tion
by
natural selection
(figure
1.4; Mazer
and
Lebuhn 1999; Reznick

the re-
sponse
to
artificial
and
natural selection
on
nor-
mally
distributed traits (Fairbairn
and
Reeve, this
volume).
Currently, agriculturalists
and
evolution-
ary
ecologists alike
use
these methods both
to
esti-
mate genetic
and
environmental causes
of
phe-
notypic variance
and to
predict

eters
that
can be
measured given
a
sample
of
data
representing
a
continuous variable.
Any set of ob-
servations
of a
quantitative trait
can be
summa-
rized
by its
mean
(or
average), variance, standard
deviation, standard error
of the
mean,
and
coeffi-
cient
of
variation (among others)

to
char-
acterize
a
trait emerges when
one
aims
to
compare
the
variability
of two or
more
traits.
This
is
often
a
first
step
when
attempting
to
predict
which
traits
may
most
easily respond
to

grams, length
in
linear units, vol-
ume in
cubed units, color
in
wavelengths, etc.),
it
is
often
meaningless
to
use, say,
the
standard devi-
7
Figure
1.3
Statistical properties
of
quantitative traits. Top:
The
shape
of a
normal distribution,
for a
hypothetical trait whose
values
range between
40 and 180

field
populations
in the
Bishop Creek
Drainage
(Inyo
County, California). Flowers
of
Aquilegia for-
mosa
(N=
129) were sampled between 1950
and
2780
m
in
ele-
vation; flowers
of A.
pubescence
(N
-
236) were collected
at
elevations
of
between
3400
and
3950

and
developmentally
normal anthers
per
flower
are
shown
for
each
of
four
greenhouse-raised
populations
of
Spergularia
marina. Each population
was
derived
from
seeds collected
from
a
distinct wild population. There
is
signif-
icant variation among populations with respect
to the
mean number
of
ovules

Various solutions have been
proposed
to
solve
this problem, including
the use of
dimension-
less
parameters such
as the
coefficient
of
variation.
Because
the
statistical
and
mathematical prop-
erties
of
normal distributions
are
well known
and
tractable, theoretical models
of the
evolution
of
quantitative traits (for which
a

many
continuously distributed traits (Falconer
and
MacKay
1996).
A
summary
of the
statistical meth-
ods
used
to
estimate
the
heritability
of
quantitative
traits
and to
predict their evolutionary trajectories
is
beyond
the
scope
of
this chapter
but is
available
in
several recent volumes (Falconer

conducted routinely
to
detect
the
causes
of
variation
in
quantitative traits
and
their
effects
on
fitness
and on
each
other,
and to
identify
pat-
terns
of
temporal
and
geographic variation would
not be
possible without
Sir
Ronald Fisher's inven-
tion

that
are
determined
by
alleles
at
multiple loci.
The
loci
affecting
a
threshold trait
each
have
a
relatively small
effect
on
some under-
lying
trait
that varies continuously, such
as the
concentration
of a
chemical product,
the
rate
of
development,

continuous
underlying
variable.
Two-class (dimorphic) threshold traits
may ex-
hibit
the
expected
(3:1)
Mendelian ratios
in the F2
generation produced
by
crosses among
the
Fl
progeny
of
parents representing
the two
pheno-
typic
classes,
but the
expected ratios
do not
appear
when conducting backcrosses.
The
underlying trait

threshold trait
and
quantitative
traits
(such
as
size, fecundity,
and
fit-
ness)
are
often
nonlinear,
and the use of
highly con-
trolled breeding designs
and
selection experiments
to
evaluate these relationships
can
strengthen con-
clusions
concerning their inheritance
and
covaria-
tion
(Roff
et
al.

winged
flightless
females
have smaller
flight
mus-
cles
but
higher
fecundity
than
the
long-winged
morph,
and
that
both wing morph
and
fecundity
have
a
quantitative genetic basis. Moreover,
artifi-
cial
selection experiments confirmed
the
interpreta-
tion
that
there

natural populations because spatial
and
temporal heterogeneity
in
habitat persistence con-
tinually
shifts
the
balance between selection
for
movement among patches
(flight)
and
rapid popu-
lation growth within patches
(flightlessness).
As
in the
case
of
quantitative traits, dimor-
phisms
or
polyphenisms (where multiple pheno-
typic states exist)
that
behave
as
threshold traits
can be

presence
or
absence
of
horns
is
determined
by the
quality
and
quantity
of the
food they receive
from
their parents.
In
other species, dimorphisms
are
strongly
associated with
a
highly heritable
trait.
Quantitative genetic analyses
of
juvenile
hormone
esterase
in the
crickets Gryllus firmus

Sexually
Dimorphic
Traits
In
organisms with separate sexes,
it is
common
to
observe
that
many traits
differ
between males
and
females.
Sexually dimorphic traits often play
a
role
in
attracting
or
competing
for
mates
or in
raising
offspring.
Where behavior
and its
concordant risks

nuptial
gifts;
and
parental care (Andersson
1994; Fairbairn
and
Reeve, this volume; Savalli,
this volume). Dimorphisms
may
also evolve where
the
sexes
differ
in
other social behaviors
or in
habi-
tat
preferences.
In
either case, gender-specific traits
are
usually interpreted
as
being
the
direct
or
indi-
rect result

ship
behaviors
or
visually attractive traits will
be
restricted
to
males. Where
the
outcome
of
direct
competition among males determines their repro-
ductive
success,
sexual
selection
favoring large size
or
aggressive behavior
may be
stronger
in
males
than
in
females.
Similarly,
where there
are

phenotype favored
in one
sex is
actually selected against
in the
other
sex are
termed sexually antagonistic characters
(Rice
1984).
When
the
direction
of
selection
is
gender-specific,
if
there
is a
genetic mechanism (such
as
X-linkage)
that
permits
the
expression
of a
trait
to be re-

that
the
dimorphism
is the
result
of
gender-specific
patterns
of
sexual
or
natu-
ral
selection (e.g., Fairbairn
and
Reeve, this vol-
ume; Savalli, this volume).
For
example, Grether
(1996)
reports evidence
that
the red
wing spots
re-
stricted
to
male rubyspot
damselflies
are the

of the
mating advantage
enjoyed
by
rela-
tively
small males during most
of the
reproductive
season. Gwynne
and
Jamieson
(1998)
suggest
that
the
evolution
of the
huge mandibles
in
male alpine
wetas
(Hemideina
maori,
Orthoptera)
of New
Zealand represent "cephalic weaponry"
that
have
evolved

1997).
Above
a
given body size,
female
marine iguanas allocate resources
to ad-
ditional
egg
production rather than
to
increased
growth, although both sexes grow
to be
larger than
the
apparent naturally selected optimum. Balmford
et
al.
(1994) provide comparative data suggesting
that sexual dimorphism
in
wing length among
57
species
of
sexually dimorphic long-tailed birds
is
the
result

sexual selec-
tion
is not
restricted
to
animals with complex
so-
cial interactions. Dioecious plant species also
ex-
hibit
marked sexual dimorphism
in
traits related
to
mating success
(Delph
et al.
1996).
Gender-specific
sexual
selection
may be
expected
in
species
in
which male success
in
delivering pollen
to

visit.
By
contrast,
reproductive success
by fe-
males
is
often
limited
not by
pollen
but by
other
12
Recurring Themes
resources;
the
result
is
that
relatively
few
pollinator
visits
are
required
to
achieve maximum seed set,
and
females

differ
from
those favored
in
females
(smal-
ler-
or
fewer-flowered inflorescences). Accordingly,
the
smaller
but
more numerous flowers
in the in-
florescences
of
male relative
to
female
Silene
lati-
folia
(Meagher 1992)
may be the
result
of
competi-
tion among males
to
attract

evolution
of
sexually dimorphic wing
size
in
birds,
the
dimorphism
in
sexually selected
traits
in
plants seems
to
result
in the
evolution
of
gender-specific
life-history traits. Male
and
female
plants have also been found
to
differ
in
growth
rates, phenology, frequency
of
reproduction,

Evolutionary
Consequences
Regardless
of the
kind
of
variation exhibited
by a
trait,
predicting
its
evolutionary
trajectory
requires
knowledge
of its
environmental
and
genetic basis.
In
the
remainder
of
this chapter,
we
consider
the
causes
and
evolutionary consequences

at
random,
and the re-
lationship between genotype
and
fitness
can be
highly sensitive
to the
identity
of the
genotypes
with which
an
individual interacts (Nunney, this
volume; Wilson, this volume).
Variation
within Individuals
Ontogenetic
Variation
Ontogenetic variation
is the
component
of
phenotypic
variation
in a
trait
ex-
pressed

changes
in
the
phenotype exhibited
by
sequentially produced
or
"modular"organs.
For
example,
the
size
of se-
quentially
produced leaves, flowers,
or
fruits
may
change over time. Similarly,
the
size
or
number
of
seeds
or
eggs produced
in
successive
fruits

the
proportion
of
phenotypic variance that
is
genetically based
may
be
obscured
to the
point
of
being undetectable
un-
less
ontogenetic sources
of
variation
are
taken into
account.
Consider
a
population
of
individuals
for
which
a
trait's

rate,
then
the
phenotypic
variance
will
not
change over time
(figure
1.5A).
On the
other hand,
if
individuals
differ
in
their ontoge-
netic
trajectories,
the
phenotypic variance exhib-
ited
by the
cohort
may
either increase
or
decrease
(figures
1.5B-D:

genotypes exhibit
different
pat-
terns
of
ontogenetic variation,
the
amount
of
inter-
genotypic variation
may
also vary over time (fig-
ures
1.5B-D).
An
example
of
this phenomenon
is
observed
in
floral
traits among successively produced flowers
of
a
short-lived,
self-fertilizing,
annual species
in