Difference in defense strategy in flower heads and leaves of Asteraceae: multiple-species approach


Oguro, M.; Sakai, S.

Oecologia 174(1): 227-239

2014


Although a vast number of studies have investigated defenses against herbivores in leaves, relatively little is known about defenses in flowers. Using wild individuals of 34 species of Asteraceae, we investigated differences in five traits that are thought to affect the intensity of herbivory (C, N, P, water, and total phenolic contents). Combinations of these traits between flower heads and leaves were studied as well. We also evaluated phylogenetic patterns of flower head and leaf traits. Flower heads had higher P and lower total phenolics than leaves. Water and C contents were negatively correlated both in the flower heads and leaves. N, P, and water contents were positively correlated in the flower heads, whereas this pattern was not found in the leaves. Thus, the traits we measured were more tightly inter-correlated in flower heads than in leaves. Because the flower heads had a lower total phenolic content, the relative allocation of defensive compounds could not be explained solely by fitness values of the organs. Perhaps plants employ an escape strategy rather than a defense strategy to cope with floral herbivores and higher allocation in P may enhance their escape from herbivores by improving the growth rate of flower heads, though our result might be affected in part by the plasticity of plants growing at different sites. Moreover, we found weak phylogenetic signals in the defensive traits. Because we found significant differences in the flower head traits, these weak signals may imply that the traits we measured evolved frequently.

Oecologia
DOI
10.1007/s00442-013-2765-x
PLANT-MICROBE-ANIMAL
INTERACTIONS
-
ORIGINAL
RESEARCH
Difference
in
defense
strategy
in
flower
heads
and
leaves
of
Asteraceae:
multiple-species
approach
Michio
Oguro
Satoki
Sakai
Received:
13
July
2012
/Accepted:
28
August
2013
©
Springer-Verlag
Berlin
Heidelberg
2013
Abstract
Although
a
vast
number
of
studies
have
investi-
gated
defenses
against
herbivores
in
leaves,
relatively
little
is
known
about
defenses
in
flowers.
Using
wild
individu-
als
of
34
species
of
Asteraceae,
we
investigated
differences
in
five
traits
that
are
thought
to
affect
the
intensity
of
her-
bivory
(C,
N,
P,
water,
and
total
phenolic
contents).
Com-
binations
of
these
traits
between
flower
heads
and
leaves
were
studied
as
well.
We
also
evaluated
phylogenetic
patterns
of
flower
head
and
leaf
traits.
Flower
heads
had
higher
P
and
lower
total
phenolics
than
leaves.
Water
and
C
contents
were
negatively
correlated
both
in
the
flower
heads
and
leaves.
N,
P,
and
water
contents
were
positively
correlated
in
the
flower
heads,
whereas
this
pattern
was
not
found
in
the
leaves.
Thus,
the
traits
we
measured
were
more
tightly
inter-correlated
in
flower
heads
than
in
leaves.
Because
the
flower
heads
had
a
lower
total
phenolic
con-
tent,
the
relative
allocation
of
defensive
compounds
could
not
be
explained
solely
by
fitness
values
of
the
organs.
Per-
haps
plants
employ
an
escape
strategy
rather
than
a
defense
Communicated
by
Evan
DeLucia.
The
information
of
the
tree
is
uploaded
on
the
TreeBASE
database:
http://purLorg/phylo/treebase/phylows/study/
TB2:512703?x-access-code=cc289f93145a0fc596d04767d8368
1f5&format=html.
Electronic
supplementary
material
The
online
version
of
this
article
(doi:10.1007/s00442-013-2765-x)
contains
supplementary
material,
which
is
available
to
authorized
users.
M.
Oguro
(RI)
S.
Sakai
Graduate
School
of
Life
Sciences,
Tohoku
University,
Aoba
6-3,
Sendai
980-8578,
Japan
e-mail:
mogu@biology.tohoku.ac.jp
S.
Sakai
e-mail:
satoki@m.tohoku.ac.jp
strategy
to
cope
with
floral
herbivores
and
higher
allocation
in
P
may
enhance
their
escape
from
herbivores
by
improv-
ing
the
growth
rate
of
flower
heads,
though
our
result
might
be
affected
in
part
by
the
plasticity
of
plants
growing
at
dif-
ferent
sites.
Moreover,
we
found
weak
phylogenetic
signals
in
the
defensive
traits.
Because
we
found
significant
dif-
ferences
in
the
flower
head
traits,
these
weak
signals
may
imply
that
the
traits
we
measured
evolved
frequently.
Keywords
Florivory
Defense
syndrome
Optimal
defense
Phylogenetic
analysis
Plant
apparency
Introduction
A
growing
body
of
studies
has
shown
that
floral
herbivory
(or
florivory)
is
common
in
the
wild
and
is
one
of
the
important
determinants
of
plant
fitness
(McCall
and
Irwin
2006).
Floral
herbivory
directly
reduces
components
of
plant
fitness
such
as
fruit
production
(Dugal
Wallace
and
O'Dowd
1989;
Schemske
and
Horvitz
1988).
Moreover,
floral
herbivory
indirectly
reduces
plant
fitness
by
reduc-
ing
the
attractiveness
of
flowers
to
pollinators
(Karban
and
Strauss
1993;
Krupnick
and
Weis
1999;
Krupnick
et
al.
1999;
but
see
McCall
2008).
In
some
cases,
these
reduc-
tions
in
plant
fitness
cause
detrimental
effects
on
popula-
tion
dynamics
(Louda
and
Potvin
1995;
Maron
et
al.
2002).
Plants
have
evolved
a
wide
variety
of
defense
traits
against
herbivores.
Since
Dethier
(1954)
and
Fraenkel
(1959)
noticed
their
importance,
plant
secondary
metabo-
lites
have
been
a
focus
of
attraction
in
the
study
of
defense
against
herbivores
(Stamp
2003).
Also,
a
higher
alloca-
tion
in
mechanical
traits
such
as
leaf
dry
matter
content
is
considered
to
act
as
defense
against
herbivores
(Cornelis-
sen
et
al.
2003).
Likewise,
low
nutritional
quality
has
been
Published
online:
14
September
2013
4tZ
Springer
Oecologia
considered
to
act
as
a
defense
(Feeny
1976;
Moran
and
Hamilton
1980;
Williams
1999).
These
defense
components
are
not
evenly
distributed
across
tissues
or
organs
(McCall
and
Fordyce
2010).
In
particular,
the
differences
in
the
allocations
between
repro-
ductive
and
vegetative
organs
are
one
of
the
central
themes
of
intra-plant
allocation
of
defenses.
Most
previous
studies
have
discussed
the
difference
in
allocation
of
secondary
metabolites
between
flowers
and
leaves
in
the
framework
of
optimal
defense
theory.
The
optimal
defense
theory
[sensu
optimal
defense
within
plants
(Stamp
2003)]
states
that
the
intra-plant
allocation
of
defenses
is
determined
by
the
value
of
the
tissue
in
terms
of
loss
of
fitness,
probabil-
ity
of
attack
and
cost
of
defense
(McKey
1974).
Because
floral
herbivory
has
greater
consequences
on
plant
fitness
than
foliar
herbivory
(Garcia
and
Ehrlen
2002;
Rose
et
al.
2005;
Strauss
et
al.
2004),
various
authors
have
predicted
that
plants
allocate
a
higher
level
of
constitutive
defense
to
flowers
(Kessler
and
Halitschke
2009;
McCall
and
Irwin
2006;
Strauss
et
al.
2004).
However,
recent
meta-analysis
(McCall
and
Fordyce
2010)
has
shown
no
significant
dif-
ference
in
the
chemical
defense
allocation
between
flowers
and
leaves
(but
see
also
Kessler
and
Halitschke
2009).
This
finding
might
have
resulted
because
their
study
(McCall
and
Fordyce
2010)
did
not
use
the
same
kind
of
defensive
chemicals
for
all
species
and
because
of
its
small
sample
size
(n
=
12).
Since
different
chemical
compounds
might
evolved
under
different
ecological/evolutional
conditions,
using
small
samples
of
different
kinds
of
defensive
chemi-
cals
may
prevent
us
from
detecting
significant
differences
between
flowers
and
leaves.
Therefore,
studies
using
a
suf-
ficient
number
of
species
to
examine
the
differences
in
the
same
defense
traits
between
flowers
and
leaves
are
needed.
Also,
there
is
increasing
awareness
that
plant
defense
is
not
a
single
trait,
but
a
combination
of
several
traits
(Agrawal
2007;
Agrawal
and
Fishbein
2006).
Plants
are
likely
to
invest
in
multiple
types
of
defense
due
to
the
potential
advantage
of
a
combination
of
defenses
(Broad-
way
and
Duffey
1988;
Rasmann
and
Agrawal
2009)
and
to
cope
with
a
wide
array
of
herbivores
attacking
the
same
host
species
(Agrawal
2007,
2011;
Agrawal
and
Fishbein
2006).
However,
there
has
been
little
study
of
the
pattern
of
the
combination
of
defenses
in
flowers
(but
see,
Armbruster
1997;
Armbruster
et
al.
2009).
Although
phylogenetic
patterns
provide
valuable
infor-
mation
about
the
evolution
of
plant
traits
(Losos
2011),
previous
reviews
and
meta-analysis
of
floral
defense
did
not
incorporate
phylogenetic
information
into
their
analysis
(Kessler
and
Halitschke
2009;
McCall
and
Fordyce
2010).
Also,
studies
that
evaluate
both
floral
and
foliar
chemi-
cal
defenses
using
a
phylogenetic
comparative
framework
are
rare
(but
see
Adler
et
al.
2012;
Armbruster
et
al.
2009;
Manson
et
al.
2012).
Therefore,
using
multiple
species
and
their
phylogenetic
relationship
may
promote
our
under-
standing
of
evolution
of
floral
defenses.
In
this
study,
we
focused
on
the
chemical
defense
(total
phenolics)
and
the
nutritional
traits
(water,
C,
N
and
P
content)
of
flower
heads
because
chemical
defense
and
nutritional
quality
can
have
synergistic
effects
on
the
per-
formance
of
herbivores
(Broadway
and
Duffey
1988)
and
have
possibly
evolved
together.
Plant
phenolic
compounds
are
known
to
reduce
the
performance
of
herbivores
(Reese
et
al.
1982),
and
also
play
several
roles
other
than
defense
against
herbivores
such
as
ultra-violet
protective
agents
and
contributors
to
plant
colors
(Hattenschwiler
and
Vitousek
2000).
Among
major
biomolecules,
it
is
known
that
pro-
teins
have
high
N
content,
nucleic
acids
and
adenosine-
5
1
-triphosphate
(ATP)
have
high
P
content,
and
lignin
and
lipids
have
high
C
content
(Sterner
and
Elser
2002a).
These
biomolecules
have
essential
functions
in
cells
such
as
metabolism,
structure,
energy
transportation
and
storage
of
genetic
information
(Sterner
and
Elser
2002a).
Because
biomass
C:N
and
C:P
ratios
of
animals
are
generally
lower
than
those
of
plants
(Elser
et
al.
2000),
lower
N
and
P
and
higher
C
contents
limit
the
performance
of
herbivores
(Sterner
and
Elser
2002b).
Water
mediates
several
meta-
bolic
reactions
in
plants
and
contributes
to
their
essential
functions
such
as
growth,
synthesis
of
protein,
photosyn-
thesis
and
respiration
(Larcher
1999).
Tissue
water
content
is
negatively
correlated
with
tissue
density
(Cornelissen
et
al.
2003)
and
low
dietary
water
content
reduces
perfor-
mance
of
herbivores
(Scriber
1979;
Scriber
and
Slansky
1981).
Because
these
chemical
and
nutritional
traits
are
known
to
be
related
to
foliar
herbivory
(Kurokawa
et
al.
2010;
Schadler
et
al.
2003),
it
may
be
possible
that
these
traits
play
significant
roles
in
plant-floral
herbivore
inter-
action.
However,
although
it
is
known
that
pollen
and
the
gynoecium
have
higher
N
and
P
contents
than
other
floral
parts
(Ashman
and
Baker
1992)
and
that
some
compounds
in
floral
fragrance
contain
N
(Levin
et
al.
2001),
selec-
tive
agents,
including
herbivores
and
pollinators,
that
have
acted
on
the
traits
of
flowers
we
measured
are
unknown.
The
objective
of
this
study
was
to
investigate
the
differ-
ences
in
defensive
traits
and
combinations
of
these
traits
between
flowers
and
leaves.
We
examined
34
species
of
Asteraceae
because
several
studies
have
reported
that
spe-
cies
in
this
family
are
likely
to
be
attacked
by
floral
her-
bivores
(e.g.,
Fenner
et
al.
2002;
Louda
and
Potvin
1995;
Maron
et
al.
2002;
Ohashi
and
Yahara
2000).
Because
actual
trait-mediated
plant-animal
interactions
occur
in
wild
populations,
we
measured
plant
traits
using
wild-
growing
individuals
rather
than
plants
growing
in
a
com-
mon
garden.
However,
to
maximize
the
number
of
spe-
cies,
plant
samples
for
each
species
were
collected
from
a
single
population
at
different
times.
Therefore,
our
results
might
be
affected
in
part
by
plasticity
of
plants
growing
4
Springer
Oecologia
at
different
sites.
The
measurements
of
flower
head
traits
were
conducted
in
three
developmental
stages,
namely,
bud,
flower
and
fruit
stages,
because
our
previous
study
(Oguro
and
Sakai
2009)
showed
that
effects
of
floral
her-
bivory
on
plant
reproduction
differ
among
the
developmen-
tal
stages
of
floral
organs.
First,
we
compared
each
trait
values
between
flowers
and
leaves.
Next,
we
evaluated
the
association
of
flower
and
leaf
traits.
Finally,
we
evaluated
phylogenetic
patterns
of
defensive
traits
to
obtain
insights
into
the
evolution
of
floral
defenses.
Materials
and
methods
Study
sites
and
study
species
This
study
was
conducted
in
Sendai,
Miyagi
Prefecture,
located
in
the
northern
part
of
Honshu
Island,
Japan.
The
study
sites
are
in
the
temperate
zone
with
a
mean
annual
temperature
of
12.1
°C
and
a
mean
annual
rainfall
of
1,242
mm.
We
sampled
18
native
and
16
exotic
herbaceous
species,
including
both
monocarpic
and
polycarpic
species,
from
23
genera
of
Asteraceae
(for
information
about
each
species,
see
Table
S1
in
Online
Resource
1).
These
species
were
selected
based
on
their
abundance
and
accessibility
in
the
fields.
In
the
species
of
Asteraceae,
the
functional
unit
that
serves
as
a
flower
is
a
flower
head,
a
group
of
florets
borne
on
a
common
receptacle
surrounded
by
sepal-like
involu-
cral
bracts
(Fenner
et
al.
2002).
Because
not
only
the
florets
but
also
the
receptacles
of
some
species
are
attacked
(M.
Oguro,
personal
observation),
the
floral
trait
values
using
whole
flower
heads
including
receptacles
were
measured.
The
flower
heads
of
our
study
species
are
attacked
by
sev-
eral
species.
The
main
floral
herbivores
are
lepidopteran
and
dipteran
larvae
and
land
snails
(M.
Oguro,
personal
observation).
Plant
sample
collection
From
September
2008
to
October
2010,
plant
samples
were
collected
from
a
single
wild
population
of
each
spe-
cies.
Because
we
could
not
obtain
enough
samples
for
trait
measurements,
some
species
were
sampled
in
more
than
1
year.
Plant
samples
were
collected
from
several
types
of
habitats,
including
forest
floors,
abandoned
golf
courses,
cemeteries,
riversides,
and
waysides
(Online
Resource
1).
The
maximum
distance
between
the
sampling
sites
was
ca.
7
km.
For
each
species,
11-108
plant
individuals
were
randomly
sampled
according
to
the
number
of
flower
heads
of
an
individual
in
that
species:
for
species
that
have
a
small
number
of
flower
heads,
a
large
number
of
individu-
als
was
sampled,
and
vice
versa
(Online
Resource
1).
The
aboveground
parts
of
individuals
were
cut
and
immediately
taken
to
the
laboratory.
In
the
laboratory,
the
plant
samples
were
placed
in
buckets
of
water
to
avoid
wilting
until
meas-
urement
of
fresh
weight
(see
below).
Because
some
spe-
cies
(e.g.,
Solidago
gigantea)
create
rhizome
and
reproduce
clonally,
individuals
growing
sufficiently
apart
from
each
other
were
sampled
for
such
species.
Flower
heads
were
sampled
in
three
reproductive
stages:
the
bud,
flower
and
fruit
stages.
Each
stage
was
defined
as
follows:
the
bud
stage
as
the
period
from
bud
formation
to
flowering
of
the
earliest
floret
in
the
flower
head;
the
flower
stage
as
the
period
from
anthesis
of
the
earliest
floret
to
the
time
when
all
florets
terminate
their
flowering;
and
the
fruit
stage
as
the
time
after
the
termination
of
flowering
of
all
florets
in
the
flower
head.
Hereafter,
we
refer
to
the
flower
head
in
the
bud
stage
as
"bud,"
the
flower
head
in
the
flower
stage
as
"flower"
and
the
flower
head
in
the
fruit
stage
as
"fruit."
The
basal
points
of
receptacles
were
cut
by
a
scissor
and
the
flower
heads
were
used
for
each
measurement.
Since
not
all
individuals
of
each
species
have
buds,
flowers,
fruits
and
leaves
at
the
same
time,
only
a
subset
of
tissues
was
sampled.
Each
trait
was
measured
using
flower
heads
and
leaves
that
were
not
apparently
damaged
by
herbivores
and
pathogens.
The
numbers
of
individuals
used
for
each
meas-
urement
in
each
stage
are
shown
in
Online
Resource
1.
Flower
head
and
leaf
traits
measurement
Water
content
In
the
laboratory,
the
fresh
weight
of
the
flower
heads
was
measured.
We
also
collected
several
leaves
from
an
indi-
vidual
of
each
species
and
measured
their
fresh
weight.
The
samples
were
dried
in
an
oven
(60
°C)
for
at
least
1
week
and
then their
dry
weight
was
measured.
Using
fresh
and
dry
weight,
the
water
content
was
calculated
[(1
—dry
weight/fresh
weight)
x
100].
This
measurement
was
con-
ducted
for
each
individual
separately.
After
weighing,
the
samples
were
stored
at
room
temperature
until
C,
N
and
P
measurement.
C,
N
and
P
contents
The
dried
samples
used
for
measurement
of
water
content
were
ground
and
their
C,
N
and
P
contents
(dry
weight
basis)
were
analyzed
for
each
individual.
C
and
N
contents
were
analyzed
using
an
NC
analyzer
(vario
EL
III;
Elemen-
tar
Analysensysteme,
Hanau,
Germany).
Standard
curves
were
prepared
using
acetanilide.
P
content
was
determined
by
spectrophotometric
methods
after
oxidation
with
persul-
fate
(Wetzel
and
Likens
2000).
Standard
curves
were
pre-
pared
using
KH
2
PO
4
.
4
Springer
Oecologia
Total
phenolic
content
In
2008
and
2009,
the
samples
of
flower
heads
not
used
for
the
measurements
of
water,
C,
N
and
P
contents
were
stored
in
freezers
until
the
spring
of
2010.
In
2010,
these
samples
were
thawed
and
dried
using
silica
gel.
The
leaf
samples
and
the
flower
head
samples
taken
in
2010,
which
were
not
used
for
the
measurements
of
water,
C,
N
and
P
contents,
were
dried
using
silica
gel.
These
dried
samples
(those
collected
in
2008-2010)
were
stored
at
room
tem-
perature
until
the
measurement
of
total
phenolics.
Samples
were
ground
before
the
measurements.
Samples
of
15
mg
of
the
flower
heads
or
leaves
were
extracted
by
50
%
aqueous
methanol.
The
Folin-Ciocalteu
method
(Julkunen-Tiitto
1985)
was
used
to
determine
the
total
phenolic
content
(dry
weight
basis)
using
tannic
acid
as
a
standard
curve.
In
this
measurement,
leaf
samples
of
Bischofia
javanica,
which
is
known
to
have
high
total
phe-
nolic
content
in
its
leaves,
were
used
as
positive
controls.
Condensed
tannins
were
also
analyzed,
but
were
not
found
in
all
species
except
in
the
positive
controls
(for
values
of
the
traits,
see
Table
S2
in
Online
Resource
2).
In
most
species,
samples
were
collected
from
single
individuals.
However,
for
several
species
(Ainsliaea
api-
culata,
Aster
ageratoides,
Bidens
bipinnata,
Carpesium
glossophyllum,
Conyza
canadensis,
Youngia
denticulata,
Erigeron
annuus,
Gnaphalium
affine,
Ixeridium
dentatum,
Leucanthemum
vulgare,
Senecio
vulgaris
and
Solidago
virgaurea),
not
enough
samples
could
be
prepared
for
the
measurement
of
phenolics
from
single
individuals.
Hence,
for
these
species,
the
samples
from
several
individuals
were
pooled
for
the
measurement
of
phenolics.
Because
A.
api-
culata
produces
cleistogamous
flowers
and
it
rarely
opens
its
flowers,
it
was
not
possible
to
obtain
enough
samples
for
the
measurement
of
phenolics
in
the
flowers
of
this
species.
Thus,
this
species
was
removed
from
analyses
using
the
data
of
phenolics
in
flower
stage. Measurement
was
con-
ducted
2-4
times
for
each
species.
Statistical
analyses
All
statistical
analyses
were
conducted
using
R
2.13.1
(R
Development
Core
Team
2011)
unless
otherwise
noted.
All
data
were
averaged
for
each
organ
(bud,
flower,
fruit,
and
leaf)
of
each
species
and
used
for
the
analyses.
The
data
of
the
flower
heads
and
leaves
used
for
the
analyses
were
the
C,
N,
P,
water
and
total
phenolic
contents.
To
detect
the
differences
in
each
trait
between
flower
heads
in
each
stage
and
leaves,
two
approaches
were
employed.
First,
before
the
univariate
analysis
of
the
indi-
vidual
traits,
two-way
multivariate
ANOVA
(MANOVA)
was
conducted
using
the
manova
function
to
investigate
overall
difference
between
plant
traits
among
species
and
plant
organs.
Next,
data
were
analyzed
using
two-
way
ANOVA
with
the
main
factors
being
species
and
plant
organ.
When
a
significant
effect
of
plant organ
was
detected,
multiple
comparisons
were
conducted
using
Tukey's
test.
Associations
among
the
traits
for
the
flower
heads
and
leaves
were
evaluated
by
calculating
Pearson's
product
moment
correlation
coefficient
(r).
To
evaluate
phylogenetic
patterns
of
floral
defense,
three
types
of
analyses
were
conducted.
The
phylogenetic
rela-
tionship
of
the
species
was
estimated
using
DNA
sequences
published
in
GenBank.
For
the
details
of
the
estimation,
see
Online
Resource
3.
First,
to
evaluate
phylogenetic
signals
of
each
trait,
Blomberg's
K-statistics
(Blomberg
et
al.
2003),
which
measure
phylogenetic
signal
(Blomberg
and
Garland
2002;
Blomberg
et
al.
2003),
were
calculated
for
each
trait
of
each
organ
using
picante
package
version
1.3.0
(Kembel
et
al.
2010)
in
R.
If
K
is
<1,
it
implies
that
relatives
resem-
ble
each
other
less
than
expected
under
Brownian
motion
evolution.
If
K
is
>1,
it
implies
that
relatives
resemble
each
other
more
than
expected
under
Brownian
motion
evolution.
Next,
to
determine
whether
closely
related
species
share
similar
combinations
of
traits
or
whether
combina-
tions
of
traits
have
independently
evolved,
we
compared
the
phylogenetic
relationship
and
the
similarity
of
the
traits
among
the
species.
Similarities
of
all
measurements
for
each
organ
were
detected
using
cluster
analysis.
All
means
of
each
measured
datum
in
each
stage
of
flower
heads
and
leaves
were
transformed
to
Z-scores
(mean
=
0,
SD
=
1)
and
used
for
the
analysis.
Hierarchical
cluster
analyses
were
conducted
using
squared
Euclidian
distance
and
Ward's
method
for
linkage.
To
evaluate
the
associa-
tion
between
phylogeny
and
similarity
of
the
traits,
the
distance
matrix
between
each
tip
of
the
phylogenetic
tree
and
results
of
the
cluster
analysis
was
calculated.
Then,
Pearson's
correlation
coefficients
(r)
between
the
distance
matrix
of
the
phylogenetic
tree
and
results
of
the
cluster
analyses
of
each
stage
of
flower
heads
and
leaves
were
calculated.
Finally,
to
evaluate
correlated
evolution
of
flower
head
and
leaf
traits,
the
correlations
between
phylogenetic
independent
contrasts
(PICs)
(Felsenstein
1985)
of
the
traits
were
calculated.
Because
estimated
phylogeny
(Fig.
51
in
Online
Resource
3)
have
polytomies,
we
randomly
assigned
zero
length
branches
and
resolved
them
using
the
multi2bi
function
in
ape
package
(Paradis
et
al.
2004).
Then
Pearson's
product
moment
correlation
coefficients
(r)
between
PICs
and
P-values
were
calculated
using
the
cortable
function
in
picante
package.
This
procedure
was
repeated
10,000
times
and
the
r-
and
P-values
were
averaged.
4)
Springer
Oecologia
Table
1
Results
from
two-
way
ANOVA
testing
for
the
difference
in
plant
traits
among
species and
plant
organs
for
34
herbaceous
species
of
df
SS
MS
F
P
N
Species
33
34.89
1.06
10.92
<0.001
***
Asteraceae
Organ
3
8.52
2.84
29.34
<0.001
***
Residuals
99
9.58
0.10
C
Species
33
246.43
7.47
4.50
<0
.
00
1
***
Organ
3
260.74
86.91
52.43
<0
.
00
1
***
Residuals
99
164.11
1.66
P
Species
33
3.09
0.09
18.84
<0.001
***
Organ
3
0.83
0.28
55.42
<0.001
***
Residuals
99
0.49
0.00
Water
Species
33
2,106.03
63.82
5.12
<0.001
***
Organ
3
1,205.99
402.00
32.22
<0.001
***
Residuals
99
1,235.06
12.48
Phenol
Species
32
334.15
10.44
3.25
<0.001
***
Organ
3
416.90
138.97
43.28
<0.001
***
*p
<
0.05;
**
p
<
0.01;
***p
<
0.001
Residuals
96
308.26
3.21
Results
Differences
in
defensive
traits
among
plant
organs
and
stages
A
significant
difference
in
plant
traits
among
species
and
plant
organs
was
found
(MANOVA:
species,
Pillai-
Bartlett
statistic
=
30
3
-
-
-,
-
F
160,
480
=
5.74,
P
<
0.0001;
organ,
Pillai-Bartlett
statistic
=
1.63,
F
15,
282
=
22.40,
P
<
0.0001).
In
agreement
with
this
result,
all
ANOVAs
showed
a
significant
difference
among
the
species
and
the
plant
organs
(buds,
flowers,
fruits
and
leaves)
for
each
trait
(Table
1).
The
C
content
was
higher
in
the
flower
heads
than
in
the
leaves
in
all
the
stages
(Fig.
la).
The
buds,
flowers
and
fruits
had
6-8
%
more
C,
on
average,
than
leaves.
The
N
content
was
higher
in
the
buds
and
leaves
than
in
the
flowers
and
fruits
(Fig.
lb).
The
flowers
and
fruits
had
15-20
%
less
N,
on
average,
than
leaves.
The
P
content
was
highest
in
the
buds,
and
flowers
and
fruits
had
higher
P
content
than
the
leaves
(Fig.
lc).
The
buds,
flowers,
and
fruits
had
50-75
%
more
P,
on
average,
than
leaves.
The
water
content
was
lower
in
the
fruits
than
in
the
buds,
flowers
and
leaves
(Fig.
1d).
Also,
the
water
content
of
the
leaves
was
higher
than
that
of
the
buds.
The
buds
and
fruits
had
2-10
%
less
water,
on
average,
then
the
leaves.
The
total
phenolic
content
was
lower
in
the
flower
heads
than
in
the
leaves
in
all
the
stages
(Fig.
le).
The
buds,
flowers
and
fruits
had
41-47
%
less
total
phenolics,
on
average,
than
the
leaves.
Association
of
the
traits
among
the
flower
heads
and
the
leaves
Significant
correlations
were
found
between
different
traits
(Table
2).
Significant
positive
correlations
were
found
between
the
N
and
P
contents,
between
the
P
and
water
contents
and
between
the
water
and
N
contents
in
the
buds
and
flowers.
Significant
negative
correlations
were
also
found
between
the
water
and
C
contents
in
the
flowers,
fruits,
and
leaves.
However,
the
total
phenolic
content
of
the
flower
heads
in
all
the
stages
and
in
the
leaves
did
not
show
a
significant
correlation
with
any
other
traits.
Relationship
between
plant
traits
and
phylogeny
Related
species
did
not
have
similar
traits
of
leaves
nor
of
flower
heads
in
any
of
the
stages:
all
the
K-statistics
calcu-
lated
for
each
trait
of
each
organ
were
<1
(Table
3).
Also,
results
from
hierarchical
cluster
analysis
that
represent
similarities
of
the
five
traits
did
not
match
the
phylogenetic
relationship
of
the
species
(Fig.
2;
for
the
bud
and
fruit
stages,
see
Fig.
S2
in
Online
Resource
4).
Accordingly,
cor-
relations
between
the
distance
matrixes
of
the
phylogenetic
tree
and
the
results
of
cluster
analysis
were
generally
low
(r
=
0.13, 0.31, 0.19,
and
0.05
for
buds,
flowers,
fruits,
and
leaves,
respectively).
4)
Springer
Oecologia
(a)
(b)
(c)
1.0
50
48
a
a
a
0
4.5-
4.0-
3.5
-
0
O
0
0
a
0.8
e
46
44
b
3.0-
2.5-
b
13_
0.6
C
0.4
42
2.0-
O
1.5-
0.2
40
0
0
B
FL
FR
L
FL
FR
L
FL
FR
L
Organ
Organ
Organ
(d)
90
b
(e)
0
0
b
85
a
ab
9
0
e
80
C
0
rs
75
,e
.c
0
6
a
a
a
70
3
0
65
0
B
FL
FR
L
B
FL
FR
L
Organ
Fig.
1
Box-and-whisker
plot
of
a
C,
b
N,
c
P,
d
water
and
e
total
phenolic
contents
(%)
of
each
developmental
stage
of
flower
heads
and
leaves
of
34
species
of
Asteraceae.
The
horizontal
thick
lines
rep-
resent
the
median
values,
the
boxes
represent
the
first
and
third
quar-
The
correlations
between
PICs
are
somewhat
differ-
ent
from
those
of
standard
correlation
analysis
(Table
2).
According
to
standard
correlation
analysis,
there
were
significant
positive
correlations
between
N
and
water
and
between
P
and
water
in
buds
and
flowers,
and
significant
negative
correlations
between
C
and
water
in
flowers,
fruits
and
leaves.
On
the
other
hand,
significant
correla-
tions
were
found
that
were
not
revealed
by
standard
corre-
lation
analysis
in
leaves.
Interestingly,
although
a
correla-
tion
was
shown
between
the
same
trait
in
flower
heads
and
leaves,
this
pattern
was
not
found
in
an
analysis
of
total
phenolics.
Discussion
Differences
in
defensive
traits
between
flower
heads
and
leaves
Our
results
showed
significant
differences
in
most
traits
between
flower
heads
and
leaves
(Fig.
1;
Online
Resource
2).
In
particular,
flower
heads
had
a
remarkably
higher
P
content
(52-75
%)
and
a
lower
total
phenolic
content
Organ
tiles,
and
the
whiskers
represent
the
range
of
the
data.
The
outliers
are
plotted
as
open
circles.
Boxes
with
different
letters
are
significantly
different
(Tukey's
test
P
<
0.05).
B
Flower
heads
in
the
bud
stage,
FL
flower
stage,
FR
fruit
stage,
L
leaves
(41-47
%)
than
leaves.
Also,
flower
heads
had
a
lower
N
content
(15-20
%)
in
the
flowers
and
fruits.
The
differences
in
C
(6-8
%)
and
water
contents
(9.8
%)
were
lower
than
those
of
N,
P
and
total
phenolics.
Our
results
did
not
match
the
prediction
regarding
the
relative
allocation
of
chemical
defenses
between
flow-
ers
and
leaves
(Kessler
and
Halitschke
2009;
McCall
and
Fordyce
2010;
McCall
and
Irwin
2006;
Strauss
et
al.
2004).
Several
authors
have
attempted
to
explain
the
difference
in
defense
allocation,
in
particular
secondary
metabo-
lites,
between
flowers
and
leaves.
They
have
focused
on
the
relative
value
of
flowers
compared
with
that
of
leaves,
and
hence
have
predicted
flowers
to
have
a
higher
defense
allocation
than
leaves.
In
agreement
with
this
prediction,
several
authors
have
shown
a
higher
allocation
of
defen-
sive
chemicals
(e
g
,
alkaloids,
tannins
and
glucosinolates)
in
flowers
(see
Kessler
and
Halitschke
2009;
McCall
and
Fordyce
2010
and
references
therein).
However,
a
lower
allocation
of
total
phenolics
was
found
in
flower
heads
than
in
leaves.
Why
were
our
results
not
consistent
with
the
pre-
diction?
We
think
that
there
is
a
possibility
that
the
rela-
tive
allocation
of
defensive
chemicals
between
flowers
and
leaves
depends
on
the
class
of
chemicals.
4)
Springer
Table
2
Pearson's
product
moment
correlation
coefficient
between
each
flower
head
and
leaf traits
of
34
herbaceous
species
of
Asteraceae
C
N
P
Water
Phenol
Bud
Flower
Fruit
Leaf
Bud
Flower
Fruit
Leaf
Bud
Flower
Fruit
Leaf
Bud
Flower
Fruit
Leaf
Bud
Flower
Fruit
Leaf
C
Bud
0.75* 0.40*
0.27
-0.03
-0.06
-0.15
-0.04
-0.07
0.03
0.16
-0.01
-0.27
-0.35*
0.01
-0.16
-0.01
0.01
0.16
-0.21
Flower
0.56*
0.66* 0.45*
-0.10
-0.04
-0.13
-0.12
-0.09
0.07
0.14
0.18
-0.31
-0.45*
-0.23
-0.20
-0.03
0.03
0.09
0.12
Fruit
0.35*
0.62*
0.38*
-0.31
-0.23
-0.10
-0.23
-0.06
0.09
0.09
0.14
-0.35*
-0.50*
-0.68*
-0.27
0.06
0.11
0.09
0.04
Leaf
0.15
0.59* 0.46*
-0.16
-0.12
-0.10
-0.03
-0.15
-0.06
0.01
-0.01
-0.41*
-0.39*
-0.23
-0.61*
0.04 0.04
0.03
0.18
N
Bud
-0.18
-0.14
-0.23
0.00
0.94* 0.80* 0.63*
0.60*
0.45*
0.33
0.20
0.64*
0.61* 0.39*
0.26
-0.14
-0.13
-0.09
-0.30
Flower
0.02
0.14
-0.08
-0.13
0.94*
0.84* 0.67*
0.56*
0.45*
0.33
0.27
0.57* 0.48*
0.26
0.25
-0.06 -0.06
0.00
-0.28
Fruit
-0.32
-0.17
-0.03 -0.33
0.74* 0.84*
0.52* 0.41*
0.27
0.20 0.10
0.48* 0.34*
0.13
0.25
0.00
-0.04
-0.06
-0.34*
Leaf
0.02
0.08
-0.01
-0.01
0.48* 0.48* 0.48*
0.37*
0.24
0.07
0.10
0.54* 0.47*
0.32
0.29
-0.10
-0.15
-0.01
-0.30
P
Bud
-0.12
0.18
0.13
0.15
0.34
0.32
0.17
-0.25
0.90* 0.84* 0.78* 0.60* 0.49*
0.24
0.19
-0.27
-0.24
-0.13
-0.09
Flower
0.03
0.34
0.34
0.31
0.17
0.24
0.08
-0.28
0.87*
0.91* 0.80* 0.50* 0.36*
0.24
0.07
-0.19
-0.17
-0.04
0.12
Fruit
0.12
0.44* 0.40*
0.34 0.14
0.20
0.04
-0.41*
0.86* 0.93*
0.82* 0.39*
0.24
0.25
0.08
-0.16
-0.12
0.02 0.02
Leaf
-0.03
0.46*
0.33
0.30
0.12
0.21
0.04
-0.39*
0.86* 0.85* 0.86*
0.28
0.14
0.00
0.12
-0.13
-0.11
0.05
0.09
Water
Bud
-0.15
-0.16
-0.30
-0.26
0.68* 0.59* 0.43* 0.37* 0.42*
0.27
0.19
0.15
0.84* 0.55* 0.65*
-0.23
-0.22
-0.15
-0.06
Flower
-0.39*
-0.37*
-0.48*
-0.29
0.57* 0.43*
0.25
0.22
0.28
0.13
0.03
0.01
0.72*
0.53* 0.46*
-0.31
-0.32
-0.27
-0.08
Fruit
0.09
-0.26
-0.64*
-0.26
0.23
0.08
-0.08
0.40*
-0.09
-0.09
-0.18
-0.27
0.43*
0.30
0.32
-0.16
-0.17
-0.07
0.14
Leaf
-0.04
-0.35*
-0.49*
-0.61*
0.39*
0.34
0.31
0.51*
-0.22
-0.35*
-0.40*
-0.36*
0.56* 0.37* 0.43*
0.02
0.03
0.08
-0.24
Phenol
Bud
0.00
-0.07
-0.19
-0.06
0.21
0.24
0.29
0.31
0.04
0.09
-0.02
0.05
0.15
0.00
0.23
0.20
0.96* 0.89*
-0.23
Flower
0.04
-0.02
-0.15
-0.01
0.21
0.18
0.16
0.21
0.10
0.12
0.06
0.05
0.17
0.00
0.28
0.16
0.94*
0.90*
-0.18
Fruit
0.16
-0.02 -0.22 -0.02
0.25
0.21
0.13
0.35*
0.13
0.15
0.09
0.05
0.27
0.04
0.44*
0.23
0.87*
0.92*
-0.26
Leaf
0.06
0.47* 0.37* 0.42*
-0.38* -0.39*
-0.54*
-0.48*
0.02
0.22
0.23
0.23
-0.32
-0.23
-0.14
-0.55*
-0.28
-0.14
-0.20
Upper
diagonal
represents
standard
correlation
coefficients
and
lower
diagonal
represents
correlations
between
phylogenetic
independent
contrasts
*
P
<
0.05
Oecologia
Table
3
K-statistics
of
each
traits
of
each
organ
calculated
from
34
herbaceous
species
of
Asteraceae
C
N
P
Water
Phenol
Bud
0.89
0.53
0.51
0.77
0.38
Flower
0.33
0.48
0.51
0.74
0.46
Fruit
0.33 0.53
0.29
0.38
0.38
Leaf
0.36
0.27
0.24
0.37
0.25
Traditionally,
plant
secondary
metabolites
have
been
largely
categorized
into
two
classes:
"quantitative
defense"
and
"qualitative
defense"
(Feeny
1975,
1976;
Stamp
2003).
Quantitative
defenses
such
as
tannins
and
polyphenols
have
a
relatively
large
molecular
weight
and
may
require
a
rela-
tively
high
concentration
to
have
a
negative
impact
on
her-
bivores
(Feeny
1975,
1976;
Stamp
2003).
Thus,
quantita-
tive
defense
imposes
a
large
initial
allocation
cost
(Coley
et
al.
1985).
However,
because
of
their
low
metabolic
activ-
ity,
their
maintenance
cost
is
relatively
low
although
it
is
difficult
for
plants
to
recycle
the
resources
used
for
quan-
titative
defense
(Coley
et
al.
1985).
The
larger
initial
and
lower
maintenance
costs
of
quantitative
defense
are
suit-
able
for
the
protection
of
leaves.
This
is
because,
in
most
plants,
leaves
exist
for
most
of
the
growing
season,
whereas
reproductive
organs
exist
for
a
limited
duration.
Hence,
allocating
quantitative
defenses
to
leaves
is
more
effective
in
terms
of
resource
use.
Qualitative
defenses
such
as
alkaloids
and
glucosinolates
have
a
relatively
low
molecular
weight
and
are
toxic
at
low
dosage
(Feeny
1975,
1976;
Stamp
2003).
Thus,
qualitative
defense
imposes
a
lower
initial
allocation
cost
(Coley
et
al.
1985).
However,
because
of
their
high
metabolic
activity,
the
maintenance
cost
of
qualitative
defense
is
relatively
high,
although
plants
can
easily
recover
resources
allocated
to
them
from
senescing
tissues
(Coley
et
al.
1985).
Because
flowers
have
a
shorter
life
span,
lower
initial
and
higher
maintenance
costs
of
qualitative
defense
are
suitable
for
their
protection.
Moreover,
higher
recyclability
of
qualita-
tive
defense
is
appropriate
for
flowers
because
flowers
are
aborted
before
becoming
fruits
when
sufficient
pollination
does
not
occur.
A
recent
review
(Endara
and
Coley
2011)
noted
a
pat-
tern
that
unapparent
plants
invested
in
qualitative
defenses
and
apparent
plants
in
quantitative
defenses,
i.e.,
the
rela-
tionship
between
plant
lifetime
and
the
class
of
chemicals
is
fairly
well
supported.
Therefore,
although
the
idea
of
categorizing
defenses
as
"qualitative"
versus
"quantita-
tive"
is
a
classic
one,
there
may
be
differences
in
some
properties
between
the
two
classes
of
defensive
chemi-
cals.
However,
Endara
and
Coley
(2011)
questioned
the
initial
explanation
of
the
mechanism
involved
in
the
rela-
tionship
between
plant
lifetime
and
the
class
of
chemi-
cal
defenses
in
terms
of
the
different
effects
of
the
two
classes
of
defenses
on
specialist
and
generalist
herbivores
(Feeny
1975,
1976;
Rhoades
and
Cates
1976).
Also,
con-
trary
to
the
initial
idea
of
the
difference
in
metabolic
activ-
ity
(Coley
et
al.
1985),
both
qualitative
and
quantitative
defenses
are
known
to
turn
over
(e.g.,
Kleiner
et
al.
1999;
Petersen
et
al.
2002).
However,
though
it
may
be
true
that
molecular
weights
of
quantitative
defenses
are
larger
than
those
of
qualitative
defenses,
general
patterns
of
differ-
ences
between
them
in
allocation
cost
and
recyclability
of
the
chemicals
are
still
unknown.
Studies
investigating
allo-
cation
cost
and
recyclability
of
the
two
classes
of
defenses
may
promote
an
understanding
of
the
mechanisms
under-
lying
the
relationship
between
the
lifetime
of
tissues
and
the
class
of
chemicals.
Fig.
2
Phylogenetic
relation-
ships
(left
side)
and
their
asso-
ciation
with
the
trait
similarity
(right
side)
calculated
by
the
hierarchical
cluster
analysis
of
the
five
traits
of
a
flower
heads
in
flower
stage
and
b
leaves
of
34
Asteraceae
species.
The
traits
used
for
the
cluster
analysis
were
C,
N,
P,
water
and
total
phenolic
contents.
For
abbrevia-
tions,
see
Online
Resource
1.
For
the
bud
and
fruit
stages,
see
Fig.
S2
in
Online
Resource
4
(a)
Flower
head
in
the
flower
stage
A.acr
A.glh
A.hml
E.phl
P.hrc
Lind
t
.
ar
C.dvr
Ydnt
C.rs1
Tplt
C.gls
Toff
A.mcr
Talb
S.vrg
.
0_
Ldnt
Y.dnt
_C
H.rdc
A.acr
co
S.vIg
S.cnd
C
O
C.crp
E.ann
'
ITS
E.hrc
C.cnd
C.rs1
Gaff
:
1
2
C.gls
C.dvr
S.p
c
g;
o
G.qdr
A.hml
t
R.Icn
C.crp
H.tbr
E.mkn
CD
E.mk
E.hrc
CD
B.fm
G.qdr
o
B.bpn
B.bpn
>,
Gaff
n
_C
L.vIg
S.vIg
0-
S.vrg
Lind
S.ggn
B.fm
S.cnd
H.rdc
E.phl
H.tbr
E.ann
Tplt
C.cnd
Toff
A.mcr
Talb
A.lh
P.hrc
A.a
g
gr
L.vIg
sr
)
CD
CD
O
_C
0
(b)
Leaf
A.apc
S.cnd
A.acr
S.ggn
A.hml
Lind
P.hrc
B.bpn
Lind
R.Icn
Y.dnt
E.mkn
T.
It
B.fm
Toff
S.vrg
T.alb
Y.dnt
Ldnt
A.apc
ilgr
H.rdc
S.vIg
C.cnd
C.crp
C.rs!
E.hrc
Ldnt
C.rs1
A.mcr
C.gls
C.dvr
C.dvr
A.acr
G.qdr
Gaff
R.Icn
E.ann
H.tbr
C.crp
E.m
P.hrc
B.fm
E.hrc
B.bpn
T.plt
G.aff
Talb
Lvlg
C.gls
S.vrg
Toff
S.ggn
Lvlg
S.cnd
H.r
dc
E.phl
A.glh
E.ann
E.phl
C.cnd
A.agr
A.mcr
G.qdr
A.glh
A.hml
A.agr
S.vIg
AIP
ED
WI
S
P
il
4
Springer
Oecologia
Consistent
with
the
above
prediction,
a
lower
allocation
of
quantitative
defense
(total
phenolics)
in
flower
heads
than
in
leaves
was
found.
Furthermore,
several
studies
have
shown
a
higher
concentration
of
qualitative
defense
in
flowers
than
in
leaves
(Fordyce
2000;
Gleadow
and
Wood-
row
2000;
but
see
Irwin
and
Adler
2006;
see
the
review
and
Table
Si
of
Kessler
and
Halitschke
2009;
Strauss
et
al.
2004).
These
results
are
also
consistent
with
our
view.
Thus,
the
difference
in
the
allocation
of
defensive
chemi-
cals
between
flowers
and
leaves
cannot
be
explained
solely
by
the
difference
in
the
value
of
the
organs,
and
taking
the
cost
of
defense
and
the
lifetime
of
organs
into
account
is
required.
However,
there
is
a
possibility
that
the
observed
differ-
ence
in
the
total
phenolic
content
was
caused
by
the
various
roles
of
phenolics
in
plants.
Plant
phenolics
are
known
to
have
several
roles
such
as
UV
protective
agents,
defensive
compounds
against
herbivores
and
pathogens,
and
contribu-
tors
to
plant
colors
(Hattenschwiler
and
Vitousek
2000).
Thus,
if
plant
phenolics
have
different
roles
in
flowers
and
leaves,
this
difference
in
roles
could
cause
the
observed
difference
in
the
amount
of
total
phenolics
between
flower
heads
and
leaves.
However,
because
several
phenolic
com-
pounds,
including
floral
pigment
(Johnson
et
al.
2008),
are
known
to
have
a
defensive
ability
(e.g.,
Chen
2008;
Reese
et
al.
1982),
we
believe
that
the
difference
in
the
total
phe-
nolic
content
represents
a
difference
in
defensive
allocation.
Also,
the
observed
difference
in
the
total
phenolic
con-
tent
between
flower
heads
and
leaves
can
be
explained
by
the
difference
in
inducibility
of
defenses
between
them.
Several
authors
predicted
that
flowers
might
have
lower
levels
of
induced
defense
than
leaves
based
on
their
rela-
tive
value
(McCall
and
Irwin
2006;
Strauss
et
al.
2004).
However,
the
optimal
defense
theory
also
predicts
that
plants
with
a
low
probability
of
attack
may
have
higher
inducibility
than
those
with
a
high
probability
of
attack
(Stamp
2003).
If
flowers
in
fact
adopt
an
escape
strat-
egy
and
have
a
low
probability
of
attacks,
they may
be
defended
by
induced
defenses
rather
than
constitutive
defenses.
Although
induced
defenses
of
flowers
are
not
well
known
(McCall
and
Karban
2006),
some
studies
have
supported
this
hypothesis
(Adler
et
al.
2006)
while
others
have
not
(Strauss
et
al.
2004).
Further
studies
investigating
differences
in
the
probability
of
attack
and
inducibility
of
defenses
both
in
flowers
and
leaves
may
help
our
under-
standing
of
plant
defense
strategies.
However,
other
selection
agents
might
have
acted
on
the
difference
in
the
traits
between
flowers
and
leaves.
For
example,
defensive
chemicals
in
the
flowers
are
known
to
reduce
pollinator
visits
(Euler
and
Baldwin
1996;
Strauss
1997;
Strauss
et
al.
2002).
Therefore,
the
allocation
of
defensive
chemicals
in
flowers
may
be
determined
by
a
bal-
ance
between
advantages
of
deterring
floral
herbivores
and
disadvantages
of
deterring
pollinators.
Then,
if
the
disad-
vantage
of
the
allocation
of
defensive
chemicals
in
flowers
is
quite
large,
plants
may
not
allocate
a
high
level
of
defen-
sive
chemicals
in
them.
For
this
reason,
flowers
might
have
a
lower
content
of
total
phenolics
than
leaves.
In
addition
to
the
lower
allocation
of
phenolics
in
the
flower
heads,
several
differences
in
the
nutritional
traits
between
the
flower
heads
and
leaves
were
found.
The
dif-
ferences
in
N
(15-20
%)
and
C
(6-8
%)
contents
between
flower
heads
and
leaves
are
much
lower
than
the
differ-
ence
in
P
content
(52-75
%).
Therefore,
the
difference
in
P
content
may
have
a
greater
impact
on
interactions
between
plants
and
herbivores
than
N
and
C
contents.
Although
the
leaves
showed
a
slightly
higher
water
content
than
buds,
mean
water
contents
of
the
buds,
flowers
and
leaves
(ca.
80
%)
are
suitable
for
the
growth
of
immature
insects
(Scriber
and
Slansky
1981),
which
are
the
main
floral
her-
bivores
in
our
system.
Thus,
considering
that
a
higher
P
content
improves
the
performance
of
herbivores
(Chen
et
al.
2007;
Huberty
and
Denno
2006;
Schade
et
al.
2003),
the
above
difference
in
nutritional
quality
and
phenolics
may
result
in
the
buds
and
flowers
being
more
acceptable
as
food
for
herbivores
than
the
leaves.
On
the
other
hand,
because
of
the
lower
water
content,
the
fruits
may
have
lower
food
quality
than
the
buds
and
flowers,
and
the
rela-
tive
food
value
of
fruits
compared
to
leaves
may
depend
on
the
relative
importance
of
the
lower
water
content
and
the
higher
P
content.
On
the
other
hand,
if
a
higher
allocation
of
P
in
flower
heads
represents
a
higher
allocation
of
ribosomes
(Elser
et
al.
1996),
it
may
enable
flower
heads
to
grow
rapidly.
If
this
is
the
case,
although
higher
allocation
of
P
may
increase
the
acceptability
of
flower
heads
for
herbivores,
their
contribution
to
the
escape
from
herbivores
may
in
part
counter
the
negative
effects
of
the
allocation.
Plants
may
employ
an
escape
strategy
to
defend
their
flowers
and
a
higher
allocation
in
P
may
help
them
escape
from
her-
bivores.
However,
although
the
roles
of
P
in
flowers
are
unknown,
it
is
known
that
pollen
and
the
gynoecium
have
higher
P
content
than
other
floral
parts
(Ashman
and
Baker
1992).
Therefore,
it
is
also
possible
that
the
reproductive
function
per
se
requires
P,
resulting
in
a
higher
allocation
of
P
in
flower
heads.
Future
study
of
the
relationship
between
P
content
of
flowers
and
floral
herbivory
may
help
clarify
the
role
of
P
in
flowers.
Combinations
of
defensive
traits
Several
combinations
of
defensive
traits
in
flower
heads
and
leaves
were
found
in
this
study.
In
general,
several
flower
head
traits
were
found
to
correlate
with
each
other,
while
correlations
among
traits
in
leaves
were
weaker
than
those
in
flower
heads
(Table
2):
significant
correlations
4
Springer
Oecologia
between
N,
P
and
water
contents
and
between
C
and
water
contents
were
observed
in
flower
heads,
whereas
only
a
significant
correlation
between
C
and
water
contents
was
observed
in
leaves.
In
general,
N,
P
and
water
contents
correlated
positively
in
the
buds
and
flowers
although
these
relationships
were
not
found
in
fruits
and
leaves
(Table
2).
This
positive
cor-
relation
is
congruent
with
the
positive
correlation
previ-
ously
shown
in
leaves
(e.g.,
Bakker
et
al.
2011;
Kerkhoff
et
al.
2006;
Scriber
and
Feeny
1979;
Wright
et
al.
2004).
When
plants
grow
their
organs,
they
must
synthesize
pro-
teins,
which
requires
allocation
of
ribosomes
(Elser
et
al.
1996).
Proteins
contain
a
high
concentration
of
N
and
ribo-
somes
contain
a
high
concentration
of
P
(Elser
et
al.
1996).
Likewise,
because
turgor
pressure
is
regarded
as
the
driv-
ing
force
for
cell
extension
(Green
and
Cummins
1974),
sufficient
water
is
required
for
plant
growth.
Therefore,
simultaneous
allocation
of
N,
P
and
water
to
flower
heads
enhances
their
growth
rate
and
may
help
them
escape
from
floral
herbivores.
However,
another
explanation
is
also
possible.
It
is
known
that
several
leaf
traits
including
N
and
P
con-
verged
into
a
single
economic
spectrum,
possibly
due
to
a
trade-off
between
leaf
photosynthetic
rates,
construc-
tion
cost,
and
leaf
longevity
(Shipley
et
al.
2006;
Wright
et
al.
2004),
and
in
some
cases,
water
content
also
corre-
lates
with
these
traits
(Pringle
et
al.
2011;
Ricklefs
2008).
These
correlations
were
also
found
in
our
study
system
(Table
2).
Therefore,
if
plants
cannot
control
the
alloca-
tion
of
these
nutrients
between
flower
heads
and
leaves
as
large
as
their
inter-specific
difference,
the
correlations
between
these
nutrients
might
automatically
be
observed
despite
the
adaptive
significances
of
the
combination.
In
fact,
these
nutrients
of
flower
heads
and
leaves
are
corre-
lated
with
each
other
(Table
2).
To
see
if
this
combination
in
fact
enhances
the
growth
of
flowers
and
helps
the
plant
to
escape
from
herbivores,
additional
tests
to
investigate
the
relationships
between
this
combination
and
the
growth
rate
of
flower
heads,
and
this
combination
and
floral
herbivory,
are
required.
Significant
negative
correlations
were
also
shown
between
water
and
C
contents
both
in
leaves
and
flower
heads
except
for
buds
(Table
2).
One
possible
explana-
tion
of
this
relationship
is
that
increasing
tissue
density
(i.e.,
decreasing
water
content)
requires
an
allocation
of
C-rich
materials.
For
example,
if
an
increase
of
tissue
den-
sity
requires
a
high
allocation
of
lignins,
it
automatically
increases
their
C
content
because
of
the
high
C
content
of
lignins
(Sterner
and
Elser
2002a).
Another
explanation
is
the
combination
of
synergistic
effects
against
herbivores.
If
the
increase
of
C
content
is
in
fact
brought
about
by
a
higher
allocation
of
lignins,
it
may
increase
the
structural
toughness
of
tissues
(Sterner
and
Elser
2002a).
If
this
is
the
case,
our
result
may
imply
that
species
having
tougher
flower
heads
tend
to
have
flower
heads
with
a
lower
water
content.
Because
a
lower
content
of
water
contributes
to
tissue
toughness, having
both
of
these
traits
may
synergis-
tically
increase
defensive
ability.
However,
other
biomol-
ecules
such
as
lipids
and
proteins
also
have
high
C
content
(Sterner
and
Elser
2002a).
Future
studies
investigating
what
biomolecules
explain
the
observed
difference
in
C
contents
would
help
clarify
the
mechanisms
underlying
the
observed
patterns
and
their
effects
on
plant-floral
herbivore
interactions.
Phylogenetic
patterns
of
plant
traits
In
each
of
the
traits,
K-statistics
showed
a
value
lower
than
one
(Table
3).
This
result
implies
that
relatives
resem-
ble
each
other
less
than
expected
under
Brownian
motion
evolution
along
the
tree
(Blomberg
et
al.
2003).
Also,
the
hierarchical
cluster
analyses
that
represent
the
similarities
of
the
five
traits
did
not
match
the
phylogenetic
relation-
ship
of
the
species
(Fig.
2;
Online
Resource
4).
This
result
may
imply
that
combinations
of
plant
traits
independently
evolved
several
times
within
the
Glade.
However,
inferring
an
evolutionary
process
from
a
phy-
logenetic
signal
is
difficult
because
different
evolutional
processes
result
in
the
same
phylogenetic
patterns
(Rev-
ell
et
al.
2008).
For
example,
low
phylogenetic
signals
are
observed
when
traits
are
under
strong
constant
stabilizing
selection
among
taxa
(Revell
et
al.
2008)
although
this
sce-
nario
is
unlikely
because
significant
differences
in
the
traits
were
found
among
species
(Fig.
1;
Table
1).
On
the
other
hand,
low
phylogenetic
signals
are
also
expected
for
the
traits
that
have
evolved
frequently
and
substantially
within
species
(Losos 2011).
It
is
known
that
herbivores
can
make
a
rapid
evolutionary
change
in
response
to
defensive
traits
of
plants
even
in
a
single
species
(Agrawal
et
al.
2013;
Zilst
et
al.
2012).
Also,
as
noted
in
the
"Introduction,"
the
meas-
ured
traits
have
potential
roles
in
the
plant
floral
herbivore
interaction.
Therefore,
it
may
be
possible
that
observed
phy-
logenetic
patterns
represent
the
situation
in
which
measured
traits
evolved
rapidly
due
to
the
selection
of
herbivores.
Interestingly,
correlated
evolution
of
the
same
trait
in
flower
heads
and
leaves
was
found:
PICs
of
the
same
traits
of
flower
heads
in
different
stages
and
leaves
were
corre-
lated
with
each
other
in
most
traits
(Table
2).
One
possible
explanation
for
the
observed
correlation
is
that
plant
spe-
cies
adopt
a
similar
defense
strategy
in
flower
heads
and
leaves
to
cope
with
herbivores.
If
flower
heads
in
three
stages
and
leaves
have
tended
to
be
attacked
by
the
same
community
of
herbivores,
the
same
defense
strategies
would
have
evolved.
However,
in
Solanum
carolinense,
it
is
known
that
it
is
rare
for
herbivores
to
attack
both
flow-
ers
and
leaves,
and
flowers
in
different
stages
(Imura
2003;
4)
Springer
Oecologia
Wise
2007).
Therefore,
although
general
patterns
of
differ-
ences
in
species
composition
of
herbivores
between
flow-
ers
and
leaves,
and
between
flowers
in
different
stages,
are
unknown,
this
scenario
seems
to
be
unlikely.
Another
possible
explanation
of
the
correlations
is
that
plants
cannot
control
allocations
of
these
compounds
between
flower
heads
and
leaves
and
between
flowers
in
different
stages.
In
some
species,
it
is
known
that
the
pro-
duction
of
secondary
metabolites
is
correlated
across
tissues
(Adler
et
al.
2006;
Kessler
and
Halitschke
2009).
If
this
is
also
the
case
for
the
traits
we
measured,
the
pattern
noted
above
might
be
observed.
However,
in
contrast
to
other
traits,
the
PICs
of
total
phenolics
in
flower
heads
did
not
correlate
with
those
of
leaves
(Table
2).
This
may
mean
that
the
plants
might
be
able
to
control
the
relative
allocation
of
phenolic
compounds
between
flowers
and
leaves,
or
at
least
their
total
amount.
Although
theories
of
floral
defense
rely
on
the
notion
that
plants
can
independently
control
the
expression
of
defenses
in
different
tissue
types
(McCall
and
Irwin
2006),
the
correlated
evolution
of
defense
expression
across
tissue
types
has
rarely
been
studied
(Adler
et
al.
2012;
Kessler
and
Halitschke
2009).
Therefore,
our
results
add
to
the
limited
knowledge
of
evolution
of
defense
expression
in
different
tissues.
Our
results
may
imply
that
plants
can
con-
trol
the
allocation
of
chemical
compounds
such
as
phenolics
between
flowers
and
leaves,
but
cannot
control
the
alloca-
tion
of
macronutrients
such
as
N
and
P
and
cannot
control
allocation
between
flowers
in
different
stages.
Therefore,
evolution
of
floral
defenses
could
not
be
understood
solely
by
their
relationship
with
floral
herbivores,
but
considering
selection
agents
acting
on
leaf
traits
may
be
required.
Caveats
There
are
some
caveats
of
this
study.
First,
to
maximize
the
number
of
species,
we
measured
plant
traits
using
wild-
growing
individuals
rather
than
plants
growing
in
a
com-
mon
garden.
Because
environmental
factors
often
affect
plant
traits
(Jansson
and
Ekbom
2002;
Nordin
et
al.
1998),
measured
trait
values
contain
not
only
genetic
differences
of
the
traits
but
also
environmental
effects
on
the
traits.
Therefore,
if
there
are
correlations
among
environmental
factors,
e.g.,
fertile
environments
have
both
high
N
and
P,
it
is
possible
that
this
environmental
correlation
brought
about
the
correlations
between
the
traits
that
we
observed.
On
the
other
hand,
different
species
appear
in
different
environmental
conditions
(e.g.,
Tilman
1984).
Therefore,
if
plants
grow
in
the
same
conditions,
traits
of
some
species
may
be
different
from
those
of
the
same
species
growing
in
wild
populations.
Because
plant-herbivore
interactions
occur
in
wild
populations,
measuring
plant
traits
in
wild
populations
may
also
be
important
for
understanding
plant
defense
strategies.
Second,
because
we
collected
each
species
from
a
sin-
gle
population,
the
measured
trait
values
may
possibly
have
deviated
from
the
"true"
trait
values
of
the
species.
This
might
have
increased
the
deviation
from
Brownian
motion
evolution
and
might
have
brought
about
observed
small
values
of
K.
Further
studies
elucidating
the
phylogenetic
signals
in
flower
head
traits
in
a
common
garden
using
multiple
seed
sources
may
provide
a
valuable
insight
into
the
evolution
of
floral
defenses.
Conclusion
We
showed
that
the
allocation
pattern
of
defense,
espe-
cially
quantitative
defense
in
flower
heads
and
leaves,
is
not
solely
explained
by
the
value
of
organs
but
requires
consideration
of
the
cost
of
defense
and
the
lifetime
of
organs.
However,
we
did
not
measure
qualitative
defenses,
and
neither
there
have
been
any
studies
that
simultaneously
measured
qualitative
and
quantitative
defenses
as
far
as
we
know.
Future
studies
that
measure
both
qualitative
and
quantitative
defenses
may
help
clarify
patterns
of
intra-
plant
allocation
of
defenses.
The
convergences
among
traits
were
stronger
in
flower
heads
than
in
leaves.
Moreover,
these
patterns
were
not
explained
by
the
phylogenetic
relationship
among
the
spe-
cies
(Fig.
2).
However,
which
selection
agents
have
acted
to
form
this
pattern
is
unknown.
While
the
traits
we
meas-
ured
are
known
to
have
a
significant
relationship
with
the
intensity
of
foliar
herbivory
(Bosu
and
Wagner
2008;
Coley
1987;
Guerra
et
al.
2010;
Kurokawa
et
al.
2010;
Perez-
Harguindeguy
et
al.
2003;
Poorter
et
al.
2004),
the
rela-
tionship
between
these
traits
and
floral
herbivory
is
largely
unknown.
Future
studies
that
investigate
the
relationship
between
these
traits
of
flowers
and
intensity
of
floral
her-
bivory
are
needed.
Acknowledgments
We
wish
to
thank
Haruka
Imai,
Hiroko
Kurok-
awa
and
Hirofumi
Onodera
for
their
help
in
the
fieldwork
and
chemi-
cal
analyses
and
Tomonori
Yamamoto
for
his
advice
on
phylogenetic
analyses.
We
thank
all
colleagues
in
our
laboratory,
especially
Tohru
Nakashizuka,
Kouki
Hikosaka,
Hiroko
Kurokawa
and
Tomoyuki
Itagaki
for
their
valuable
suggestions
throughout
the
study.
We
also
thank
anonymous
reviewers
for
their
valuable
comments
and
sugges-
tions
on
earlier
versions
of
this
manuscript.
References
Adler
LS,
Wink
M,
Distl
M,
Lentz
AJ
(2006)
Leaf
herbivory
and
nutrients
increase
nectar
alkaloids.
Ecol
Lett
9:960-967
Adler
LS,
Seifert
MG,
Wink
M,
Morse
GE
(2012)
Reliance
on
pol-
linators
predicts
defensive
chemistry
across
tobacco
species.
Ecol
Lett
15:1140-1148
Agrawal
AA
(2007)
Macroevolution
of
plant
defense
strategies.
Trends
Ecol
Evol
22:103-109
4)
Springer
Oecologia
Agrawal
AA
(2011)
Current
trends
in
the
evolutionary
ecology
of
plant
defence.
Funct
Ecol
25:420-432
Agrawal
AA,
Fishbein
M
(2006)
Plant
defense
syndromes.
Ecology
87:S132—S149
Agrawal
AA,
Johnson
MTJ,
Hastings
AP,
Maron
JL
(2013)
A
field
experiment
demonstrating
plant
life-history
evolution
and
its
eco-evolutionary
feedback
to
seed
predator
populations.
Am
Nat
0:S1—S11
Armbruster
WS
(1997)
Exaptations
link
evolution
of
plant-herbivore
and
plant-pollinator
interactions:
a
phylogenetic
inquiry.
Ecology
78:1661-1672
Armbruster
WS,
Lee
J,
Baldwin
BG
(2009)
Macroevolutionary
pat-
terns
of
defense
and
pollination
in
Dalechampia
vines:
adapta-
tion,
exaptation,
and
evolutionary
novelty.
Proc
Nail
Acad
Sci
USA
106:18085-18090
Ashman
TL,
Baker
I
(1992)
Variation
in
floral
sex
allocation
with
time
of
season
and
currency.
Ecology
73:1237-1243
Bakker
MA,
Carrefio-Rocabado
G,
Poorter
L
(2011)
Leaf
econom-
ics
traits
predict
litter
decomposition
of
tropical
plants
and
differ
among
land
use
types.
Funct
Ecol
25:473-483
Blomberg
SP,
Garland
T
Jr
(2002)
Tempo
and
mode
in
evolution:
phylogenetic
inertia,
adaptation
and
comparative
methods.
J
Evol
Biol
15:899-910
Blomberg
SP,
Garland
T
Jr,
Ives
AR
(2003)
Testing
for
phylogenetic
signal
in
comparative
data:
behavioral
traits
are
more
labile.
Evo-
lution
57:717-745
Bosu
PP,
Wagner
MR
(2008)
Anatomical
and
nutritional
fac-
tors
associated
with
susceptibility
of
elms
(Ulmus
spp.)
to
the
elm
leaf
beetle
(Coleoptera:
Chrysomelidae).
J
Econ
Entomol
101:944-954
Broadway
RM,
Duffey
SS
(1988)
The
effect
of
plant
protein-quality
on
insect
digestive
physiology
and
the
toxicity
of
plant
protein-
ase-inhibitors.
J
Insect
Physiol
34:1111-1117
Chen
M-S
(2008)
Inducible
direct
plant
defense
against
insect
herbi-
vores:
a
review.
Insect
Sci
15:101-114
Chen
Y,
Opit
GP,
Jonas
VM,
Williams
KA,
Nechols
JR,
Margolies
DC
(2007)
Twospotted
spider
mite
population
level,
distribution,
and
damage
on
ivy
geranium
in
response
to
different
nitrogen
and
phosphorus
fertilization
regimes.
J
Econ
Entomol
100:1821-1830
Coley
PD
(1987)
Interspecific
variation
in
plant
anti-herbivore
prop-
erties:
the
role
of
habitat
quality
and
rate
of
disturbance.
New
Phytol
106:251-263
Coley
PD,
Bryant
JP,
Chapin
FS
(1985)
Resource
availability
and
plant
antiherbivore
defense.
Science
230:895-899
Cornelissen
JHC
et
al
(2003)
A
handbook
of
protocols
for
standard-
ised
and
easy
measurement
of
plant
functional
traits
worldwide.
Aust
J
Bot
51:335-380
Dethier
V
(1954)
Evolution
of
feeding
preferences
in
phytophagous
insects.
Evolution
8:33-54
Dugal
Wallace
D,
O'Dowd
DJ
(1989)
The
effect
of
nutrients
and
inflorescence
damage
by
insects
on
fruit-set
by
Banksia
spinu-
losa.
Oecologia
79:482-488
Elser
JJ,
Dobberfuhl
DR,
MacKay
NA,
Schampel
JH
(1996)
Organism
size,
life
history,
and
N:P
stoichiometry.
Bioscience
46:674-684
Elser
JJ
et
al
(2000)
Nutritional
constraints
in
terrestrial
and
freshwa-
ter
food
webs.
Nature
408:578-580
Endara
M-J,
Coley
PD
(2011)
The
resource
availability
hypothesis
revisited:
a
meta-analysis.
Funct
Ecol
25:389-398
Euler
M,
Baldwin
TT
(1996)
The
chemistry
of
defense
and
apparency
in
the
corollas
of
Nicotiana
attenuata.
Oecologia
107:102-112
Feeny
P
(1975)
Biochemical
coevolution
between
plants
and
their
insect
herbivores.
In:
Gilbert
LE,
Raven
PH
(eds)
Coevolution
of
animals
and
plants.
University
of
Texas
Press,
Austin,
pp
3-19
Feeny
P
(1976)
Plant
apparency
and
chemical
defense.
In:
Wallace
JW,
Mansell
RL
(eds)
Recent
advances
in
phytochemistry,
vol
10.
Plenum
Press,
New
York,
pp
1-40
Felsenstein
J
(1985)
Phylogenies
and
the
comparative
method.
Am
Nat
125:1-15
Fenner
M,
Cresswell
JE,
Hurley
RA,
Baldwin
T
(2002)
Relation-
ship
between
capitulum
size
and
pre-dispersal
seed
predation
by
insect
larvae
in
common
Asteraceae.
Oecologia
130:72-77
Fordyce
JA
(2000)
A
model
without
a
mimic:
aristolochic
acids
from
the
California
pipevine
swallowtail,
Battus
philenor
hir-
suta,
and
its
host
plant,
Aristolochia
califomica.
J
Chem
Ecol
26:2567-2578
Fraenkel
GS
(1959)
The
raison
d'être
of
secondary
plant
substances.
Science
129:1466-1470
Garcia
MB,
Orden
J
(2002)
Reproductive
effort
and
herbivory
tim-
ing
in
a
perennial
herb:
fitness
components
at
the
individual
and
population
levels.
Am
J
Bot
89:1295-1302
Gleadow
RM,
Woodrow
IE
(2000)
Temporal
and
spatial
variation
in
cyanogenic
glycosides
in
Eucalyptus
cladocalyx.
Tree
Physiol
20:591-598
Green
PB,
Cummins
WR
(1974)
Growth-rate
and
turgor
pressure—
auxin
effect
studied
with
an
automated
apparatus
for
single
coleoptiles.
Plant
Physiol
54:863-869
Guerra
PC,
Becerra
J,
Gianoli
E
(2010)
Explaining
differential
her-
bivory
in
sun
and
shade:
the
case
of
Aristotelia
chilensis
saplings.
Arthropod-Plant
Interact
4:229-235
Hfittenschwiler
S,
Vitousek
PM
(2000)
The
role
of
polyphenols
in
ter-
restrial
ecosystem
nutrient
cycling.
Trends
Ecol
Evol
15:238-243
Huberty
A,
Denno
R
(2006)
Consequences
of
nitrogen
and
phos-
phorus
limitation
for
the
performance
of
two
planthoppers
with
divergent
life-history
strategies.
Oecologia
149:444-455
Imura
0
(2003)
Herbivorous
arthropod
community
of
an
alien
weed
Solanum
carolinense
L.
Appl
Entomol
Zool
38:293-300
Irwin
RE,
Adler
LS
(2006)
Correlations
among
traits
associated
with
herbivore
resistance
and
pollination:
implications
for
pollination
and
nectar robbing
in
a
distylous
plant.
Am
J
Bot
93:64-72
Jansson
J,
Ekbom
B
(2002)
The
effect
of
different
plant
nutrient
regimes
on
the
aphid
Macrosiphum
euphorbiae
growing
on
petu-
nia.
Entomol
Exp
Appl
104:109-116
Johnson
E,
Berhow
M,
Dowd
P
(2008)
Colored
and
white
sectors
from
star-patterned
petunia
flowers
display
differential
resist-
ance
to
corn
earworm
and
cabbage
looper
larvae.
J
Chem
Ecol
34:757-765
Julkunen-Tiitto
R
(1985)
Phenolic constituents
in
the
leaves
of
north-
ern
willows:
methods
for
the
analysis
of
certain
phenolics.
J
Agric
Food
Chem
33:213-217
Karban
R,
Strauss
SY
(1993)
Effects
of
herbivores
on
growth
and
reproduction
of
their
perennial
host,
Erigeron
glaucus.
Ecology
74:39-46
Kembel
SW
et
al
(2010)
Picante:
R
tools
for
integrating
phylogenies
and
ecology.
Bioinformatics
26:1463-1464
Kerkhoff
AJ,
Fagan
WF,
Elser
JJ,
Enquist
BJ
(2006)
Phylogenetic
and
growth
form
variation
in
the
scaling
of
nitrogen
and
phosphorus
in
the
seed
plants.
Am
Nat
168:E103—E122
Kessler
A,
Halitschke
R
(2009)
Testing
the
potential
for
conflicting
selection
on
floral
chemical
traits
by
pollinators
and
herbivores:
predictions
and
case
study.
Funct
Ecol
23:901-912
Kleiner
KW,
Raffa
KF,
Dickson
RE
(1999)
Partitioning
of
C-14-1a-
beled
photosynthate
to
allelochemicals
and
primary
metabo-
lites
in
source
and
sink
leaves
of
aspen:
evidence
for
secondary
metabolite
turnover.
Oecologia
119:408-418
Krupnick
GA,
Weis
AE
(1999)
The
effect
of
floral
herbivory
on
male
and
female
reproductive
success
in
Isomeris
arborea.
Ecology
80:135-149
Krupnick
GA,
Weis
AE,
Campbell
DR
(1999)
The
consequences
of
floral
herbivory
for
pollinator
service
to
Isomeris
arborea.
Ecol-
ogy
80:125-134
Kurokawa
H,
Peltzer
DA,
Wardle
DA
(2010)
Plant
traits,
leaf
palat-
ability
and
litter
decomposability
for
co-occurring
woody
species
4)
Springer
Oecologia
differing
in
invasion
status
and
nitrogen
fixation
ability.
Funct
Ecol
24:513-523
Larcher
W
(1999)
Okophysiologie
der
Pflanzen
(in
Japanese),
5th
edn.
Springer,
Tokyo
Levin
RA,
Raguso
RA,
McDade
LA
(2001)
Fragrance
chemistry
and
pollinator
affinities
in
Nyctaginaceae.
Phytochemistry
58:429-440
Losos
JB
(2011)
Seeing
the
forest
for
the
trees:
the
limitations
of
phy-
logenies
in
comparative
biology
(American
Society
of
Naturalists
Address).
Am
Nat
177:709-727
Louda
SM,
Potvin
MA
(1995)
Effect
of
inflorescence-feeding
insects
on
the
demography
and
lifetime
fitness
of
a
native
plant.
Ecology
76:229-245
Manson
JS,
Rasmann
S,
Halitschke
R,
Thomson
JD,
Agrawal
AA
(2012)
Cardenolides
in
nectar
may
be
more
than
a
consequence
of
allocation
to
other
plant
parts:
a
phylogenetic
study
of
Ascle-
pias.
Funct
Ecol
26:1100-1110
Maron
JL,
Combs
JK,
Louda
SM
(2002)
Convergent
demographic
effects
of
insect
attack
on
related
thistles
in
coastal
vs.
continen-
tal
dunes.
Ecology
83:3382-3392
McCall
AC
(2008)
Florivory
affects
pollinator
visitation
and
female
fitness
in
Nemophila
menziesii.
Oecologia
155:729-737
McCall
AC,
Fordyce
JA
(2010)
Can
optimal
defence
theory
be
used
to
predict
the
distribution
of
plant
chemical
defences?
J
Ecol
98:985-992
McCall
AC,
Irwin
RE
(2006)
Florivory:
the
intersection
of
pollination
and
herbivory.
Ecol
Lett
9:1351-1365
McCall
AC,
Karban
R
(2006)
Induced
defense
in
Nicotiana
attenuata
(Solanaceae)
fruit
and
flowers.
Oecologia
146:566-571
McKey
D
(1974)
Adaptive
patterns
in
alkaloid
physiology.
Am
Nat
108:305
Moran
N,
Hamilton
WD
(1980)
Low
nutritive
quality
as
defense
against
herbivores.
J
Theor
Biol
86:247-254
Nordin
A,
Nasholm
T,
Ericson
L
(1998)
Effects
of
simulated
N
dep-
osition
on
understorey
vegetation
of
a
boreal
coniferous
forest.
Funct
Ecol
12:691-699
Oguro
M,
Sakai
S
(2009)
Floral
herbivory
at
different
stages
of
flower
development
changes
reproduction
in
Iris
gracilipes
(Iridaceae).
Plant
Ecol
202:221-234
Ohashi
K,
Yahara
T
(2000)
Effects
of
flower
production
and
predis-
persal
seed
predation
on
reproduction
in
Cirsium
purpuratum.
Can
J
Bot
78:230-236
Paradis
E,
Claude
J,
Strimmer
K
(2004)
APE:
analyses
of
phylogenet-
ics
and
evolution
in
R
language.
Bioinformatics
20:289-290
Perez-Harguindeguy
N,
Diaz
S,
Vendramini
F,
Comelissen
JHC,
Gur-
vich
DE,
Cabido
M
(2003)
Leaf
traits
and
herbivore
selection
in
the
field
and
in
cafeteria
experiments.
Aust
Ecol
28:642-650
Petersen
BL,
Chen
SX,
Hansen
CH,
Olsen
CE,
Halkier
BA
(2002)
Composition
and
content
of
glucosinolates
in
developing
Arabi-
dopsis
thaliana.
Planta
214:562-571
Poorter
L,
de
Plassche
MV,
Willems
S,
Boot
RGA
(2004)
Leaf
traits
and
herbivory
rates
of
tropical
tree
species
differing
in
succes-
sional
status.
Plant
Biol
6:746-754
Pringle
EG
et
al
(2011)
Distinct
leaf-trait
syndromes
of
evergreen
and
deciduous
trees
in
a
seasonally
dry
tropical
forest.
Biotropica
43:299-308
R
Development
Core
Team
(2011)
R:
A
language
and
environment
for
statistical
computing.
In:
2.13.2
edn.
R
Foundation
for
Statis-
tical
Computing,
Vienna
Rasmann
S,
Agrawal
AA
(2009)
Plant
defense
against
herbivory:
progress
in
identifying
synergism,
redundancy,
and
antagonism
between
resistance
traits.
Curr
Opin
Plant
Biol
12:473-478
Reese
JC,
Chan
BG,
Waiss
AC
(1982)
Effects
of
cotton
condensed
tannin,
maysin
(corn)
and
pinitol
(soybeans)
on
Heliothis
zea
growth
and
development.
J
Chem
Ecol
8:1429-1436
Revell
U
.
,
Harmon
U
.
,
Collar
DC
(2008)
Phylogenetic
signal,
evolu-
tionary
process,
and
rate.
Syst
Biol
57:591-601
Rhoades
DF,
Cates
RG
(1976)
Toward
a
general
theory
of
plant
anti-
herbivore
chemistry.
In:
Wallace
JW,
Mansell
RL
(eds)
Recent
advances
in
phytochemistry,
vol
10.
Plenum
Press,
New
York,
pp
168-213
Ricklefs
RE
(2008)
Foliage
chemistry
and
the
distribution
of
Lepi-
doptera
larvae
on
broad-leaved
trees
in
southern
Ontario.
Oeco-
logia
157:53-67
Rose
KE,
Louda
SM,
Rees
M
(2005)
Demographic
and
evolutionary
impacts
of
native
and
invasive
insect
herbivores
on
Cirsium
cane-
scens.
Ecology
86:453-465
Schade
JD,
Kyle
M,
Hobbie
SE,
Fagan
WF,
Elser
JJ
(2003)
Stoichio-
metric
tracking
of
soil
nutrients
by
a
desert
insect
herbivore.
Ecol
Lett
6:96-101
Schadler
M,
Jung
G,
Auge
H,
Brandl
R
(2003)
Palatability,
decom-
position
and
insect
herbivory:
patterns
in
a
successional
old-field
plant
community.
Oikos
103:121-132
Schemske
DW,
Horvitz
CC
(1988)
Plant-animal
interactions
and
fruit
production
in
a
neotropical
herb:
a
path
analysis.
Ecology
69:1128-1137
Scriber
JM
(1979)
Effects
of
leaf-water
supplementation
upon
post-
ingestive
nutritional
indices
of
forb-,
shrub-,
vine-,
and
tree-feed-
ing
Lepidoptera.
Entomol
Exp
Appl
25:240-252
Scriber
JM,
Feeny
P
(1979)
Growth
of
herbivorous
caterpillars
in
relation
to
feeding
specialization
and
to
the
growth
form
of
their
food
plants.
Ecology
60:829-850
Scriber
JM,
Slansky
F
(1981)
The
nutritional
ecology
of
immature
insects.
Annu
Rev
Entomol
26:183-211
Shipley
B,
Lechowicz
MJ,
Wright
I,
Reich
PB
(2006)
Fundamental
trade-offs
generating
the
worldwide
leaf
economics
spectrum.
Ecology
87:535-541
Stamp
N
(2003)
Out
of
the
quagmire
of
plant
defense
hypotheses.
Quart
Rev
Biol
78:23-55
Sterner
RW,
Elser
JJ
(2002a)
Biological
chemistry:
building
cells
from
elements.
In:
Ecological
stoichiometry:
the
biology
of
ele-
ments
from
molecules
to
the
biosphere.
Princeton
University
Press,
pp
44-79
Sterner
RW,
Elser
JJ
(2002b)
Imbalanced
resources
and
animal
growth.
In:
Ecological
stoichiometry:
the
biology
of
elements
from
molecules
to
the
biosphere.
Princeton
University
Press,
pp
179-230
Strauss
SY
(1997)
Floral
characters
link
herbivores,
pollinators,
and
plant
fitness.
Ecology
78:1640-1645
Strauss
SY,
Rudgers
JA,
Lau
JA,
Irwin
RE
(2002)
Direct
and
ecologi-
cal
costs
of
resistance
to
herbivory.
Trends
Ecol
Evol
17:278-285
Strauss
SY,
Irwin
RE,
Lambrix
VM
(2004)
Optimal
defence
theory
and
flower
petal
colour
predict
variation
in
the
secondary
chemis-
try
of
wild
radish.
J
Ecol
92:132-141
Tilman
GD
(1984)
Plant
dominance
along
an
experimental
nutrient
gradient.
Ecology
65:1445-1453
Wetzel
RG,
Likens
GE
(2000)
Limnological
analyses,
3rd
edn.
Springer,
New
York
Williams
IS
(1999)
Slow-growth,
high-mortality—a
general
hypoth-
esis,
or
is
it?
Ecol
Entomol
24:490-495
Wise
MJ
(2007)
Evolutionary
ecology
of
resistance
to
herbivory:
an
investigation
of
potential
genetic
constraints
in
the
multiple-
herbivore
community
of
Solanum
carolinense.
New
Phytol
175:773-784
Wright
IJ
et
al
(2004)
The
worldwide
leaf
economics
spectrum.
Nature
428:821-827
Zilst
T,
Heichinger
C,
Grossniklaus
U,
Harrington
R,
Kliebenstein
DJ,
Turnbull
LA
(2012)
Natural
enemies
drive
geographic
variation
in
plant
defenses.
Science
338:116-119
4)
Springer