Effect of hybridization of the Quercus crassifolia×Quercus crassipes complex on the community structure of endophagous insects

Abstract

In a previous study, we showed that the geographic proximity of hybrid plants to the allopatric areas of parental species increases their morphological and genetic similarity with them. In the present work, we explored whether the endophagous fauna of hybrid plants show the same pattern. We studied the canopy species richness, diversity and composition of leaf-mining moths (Lepidoptera: Tischeridae, Citheraniidae) and gall-forming wasps (Hymenoptera: Cynipidae) associated with two species of red oaks (Quercus crassifolia and Quercus crassipes) and their interspecific hybrid (Quercusxdysophylla Benth pro sp.) in seven hybrid zones in central Mexico, during four seasons in 2 years. The study was conducted on 194 oak trees with known genetic status [identified by leaf morphology and molecular markers (random amplified polymorphic DNAs)], and the results indicate a bidirectional pattern of gene flow. Hybrid plants supported intermediate levels of infestation of gall-forming and leaf-mining insects compared to their putative parental species. The infestation level of leaf-mining insects varied significantly following the pattern: Q. crassifolia>hybrids>Q. crassipes, whereas the gall-forming insects showed an inverse pattern. A negative and significant relationship was found between these two types of insect guilds in each host taxa, when the infestation percentage was evaluated. It was found that 31.5% (n=11) of the endophagous insects were specific to Q. crassipes, 22.9% (n=8) to Q. crassifolia, and 8.6% (n=3) to hybrid individuals. The hybrid bridge hypothesis was supported in the case of 25.7% (n=9) of insects, which suggests that the presence of a hybrid intermediary plant may favor a host herbivore shift from one plant species to another. Greater genetic diversity in a hybrid zone is associated with greater diversity in the endophagous community. The geographic proximity of hybrid plants to the allopatric site of a parental species increases their similarity in terms of endophagous insects and the Eje Neovolcánico acts as a corridor favoring this pattern.

Full text

COMMUNITY ECOLOGY
Efraı
´n Tovar-Sa
´nchez ÆKen Oyama
Effect of hybridization of the
Quercus crassifolia
·
Quercus crassipes
complex on the community structure of endophagous insects
Received: 9 October 2004 / Accepted: 2 December 2005 / Published online: 4 February 2006
Springer-Verlag 2006
Abstract In a previous study, we showed that the geo-
graphic proximity of hybrid plants to the allopatric
areas of parental species increases their morphological
and genetic similarity with them. In the present work, we
explored whether the endophagous fauna of hybrid
plants show the same pattern. We studied the canopy
species richness, diversity and composition of leaf-min-
ing moths (Lepidoptera: Tischeridae, Citheraniidae) and
gall-forming wasps (Hymenoptera: Cynipidae) associ-
ated with two species of red oaks (Quercus crassifolia
and Quercus crassipes) and their interspecific hybrid
(Quercus·dysophylla Benth pro sp.) in seven hybrid
zones in central Mexico, during four seasons in 2 years.
The study was conducted on 194 oak trees with known
genetic status [identified by leaf morphology and
molecular markers (random amplified polymorphic
DNAs)], and the results indicate a bidirectional pattern
of gene flow. Hybrid plants supported intermediate
levels of infestation of gall-forming and leaf-mining in-
sects compared to their putative parental species. The
infestation level of leaf-mining insects varied signifi-
cantly following the pattern: Q. crassifolia>hybrids>Q.
crassipes, whereas the gall-forming insects showed an
inverse pattern. A negative and significant relationship
was found between these two types of insect guilds in
each host taxa, when the infestation percentage was
evaluated. It was found that 31.5% (n=11) of the en-
dophagous insects were specific to Q. crassipes, 22.9%
(n=8) to Q. crassifolia, and 8.6% (n=3) to hybrid
individuals. The hybrid bridge hypothesis was supported
in the case of 25.7% (n=9) of insects, which suggests
that the presence of a hybrid intermediary plant may
favor a host herbivore shift from one plant species to
another. Greater genetic diversity in a hybrid zone is
associated with greater diversity in the endophagous
community. The geographic proximity of hybrid plants
to the allopatric site of a parental species increases their
similarity in terms of endophagous insects and the Eje
Neovolca
´nico acts as a corridor favoring this pattern.
Keywords Hybridization ÆGall forming Æ
Leaf mining ÆQuercus ÆCommunity structure
Introduction
Genetically controlled plant traits can affect the distri-
bution and abundance of associated insect species (Fritz
and Simms 1992), and the community-level conse-
quences of these same traits (Dungey et al. 2000). Ge-
netic differences among plant species (Knops et al.
1999), within plant species (Simms and Rausher 1993),
or within a hybridizing complex (Dungey et al. 2000;
Wimp et al. 2004,2005) affect the dependent herbivore
community. Also, population genetic diversity in plants
can affect dependent insect species, especially when these
plant species are so common as to characterize a habitat
type (Wimp et al. 2004). Plant hybrid zones represent
some of the most genetically diverse interbreeding sys-
tems, and thus represent ideal systems for an examina-
tion of the effects of host plant genetic variation on the
endophagous community. Hybridization can result in
novel host plant traits (Rieseberg and Ellstrand 1993)
and high levels of genetic variation within the hybrid
zone (Whitham et al. 1999).
Community level studies of different organisms be-
tween parental and hybrid plant taxa gave four results.
First, hybrid plants supported more insect species than
either parental species (Fritz et al. 1994; Whitham et al.
1994). The authors suggest an increased genetic sus-
Electronic Supplementary Material Supplementary material is
available for this article at http://dx.doi.org/10.1007/s00442-005-
0328-5
Communicated by Jim Cronin
E. Tovar-Sa
´nchez (&)ÆK. Oyama
Centro de Investigaciones en Ecosistemas, UNAM,
Campus Morelia, Antigua Carretera a Pa
´tzcuaro no. 8701,
Col. San Jose
´de la Huerta, Morelia, Michoaca
´n, Me
´xico
E-mail: efratosa@yahoo.com
Fax: +52-55-56232719
Oecologia (2006) 147: 702–713
DOI 10.1007/s00442-005-0328-5

ceptibility of hybrids due to hybrid breakdown; in-
creased stress in the hybrid zone resulting in greater
plant susceptibility; and a greater diversity of resources
in the hybrid zone, which could support more species.
Second, equal numbers of insect species were found on
hybrids compared to either parental species (Fritz et al.
1994,1998; Messina et al. 1996). Generally, the dif-
ference in species number occurs when comparing dif-
ferent habitats over an elevation gradient, and common
garden studies at different elevations revealed no dif-
ferences in hybrid susceptibility. Third, fewer insects
occurred on hybrids than on either parental plant
species (Boecklen and Spellenberg 1990; Aguilar and
Boecklen 1992; Fritz et al. 1999). Host architecture,
leaf morphology, and geographic range of the plant
were proposed as possible causes for differences in
community structure. Fourth, intermediate numbers of
insects occurred on hybrids in comparison to their
parental species. The authors suggest that hybrid
resistances are due to the additive inheritance of resis-
tance traits from both parents. The lack of a clear
community-level response to hybrid zones is due to the
complex interactions between several factors such as
geographical range of hybrid zones, environmental
gradients, hybrid zone age, and genetic status of hy-
brids (Boecklen and Spellenberg 1990; Hanhima
¨ki et al.
1994; Strauss 1994).
Floate and Whitham (1993) proposed the hybrid
bridge hypothesis, which suggests that an intermediate
hybrid plant facilitates the switch by endophagous in-
sects from one host species to another. Host shift may be
affected by plant hybridization patterns, therefore, four
hypothetical systems are proposed:
1. There is no hybridization. In the absence of hybrid-
ization the gap between species is greater, and models
of host shifting assume this is the general pattern.
2. There is hybridization, producing infertile F
1
off-
spring, where hybrid individuals will have 50% of
each parental genome. The production of sterile F
1
hybrids should facilitate a host shift, but two large
genetic gaps remain.
3. There is hybridization, producing fertile F
1
off-
spring where hybrids will have unidirectional
introgression. Unidirectional hybridization to one
parental species produces a continuum of hybrid
genotypes between one parental species and F
1
hy-
brids, but a large gap remains between F
1
and the
other parental species. This pattern should facilitate
host shifting halfway between the two species, but
the remaining gap still represents a barrier to host
shifting in gradual steps.
4. There is hybridization, producing fertile F
1
offspring
where hybrids will have bidirectional introgression.
F
1
hybrids may backcross to both of the parental
species in a pattern of bidirectional introgression
resulting in a complete continuum of hybrid geno-
types. A genetic continuum with no gaps should be
most likely to facilitate host shifting.
Hybridization between plants is considered a
prevalent phenomenon in nature (Anderson 1949), due
to which between 50 and 70% of the total number of
species may have originated (Stace 1987; Wendel et al.
1991). This phenomenon contributes to genetic diversity
and plant speciation (Grant 1981; Rieseberg and
Brunsfeld 1992; Rieseberg and Wendel 1993). Hybrid-
ization between species is found in localized areas
(Hardin 1975), and these hybrid zones are places of high
ecological and evolutionary activity, providing spots of
biodiversity forming new habitats for associate fauna.
Quercus species (Fagaceae) have scarce sterility barriers,
promoting a high level of hybridization among species
that are morphologically and physiologically very dif-
ferent from each other (Jensen and Eshbaugh 1976;
Guttman and Weight 1989; Whittemore and Schaal
1991; Bacilieri et al. 1995; Dumolin et al. 1997; Bruschi
et al. 2000).
The greatest richness of Quercus species is found in
Mexico, where 90% of the approximately 150 oak spe-
cies of North America, including 86 endemics, are dis-
tributed (Nixon 1993). Oak distribution and patterns of
species richness have played a major role in the distri-
bution and species richness of oak gall wasps (Hyme-
noptera: Cynipidae) (Stone et al. 2002), Mexico has the
greatest richness of oak gall wasps, with approximately
700 species included in 29 genera (Wendel 1960). Studies
of cynipid fauna in oak hybrid zones have shown that
gall wasps are highly sensitive to levels of introgression
between host species (Boecklen and Spellenberg 1990;
Moorehead et al. 1993; Dungey et al. 2000; Wimp et al.
2004,2005). Valuable insights into these processes can
be gained through the investigation of distribution,
abundance and diversity of canopy insects in response to
host hybrid plants in relation to their parental species
(e.g., Boecklen and Larson 1994; Fritz et al. 1994,1996;
Whitham et al. 1994; Morrow et al. 1994). It has been
found that different herbivore species respond in a
variety of ways to hybrids vs. their putative parentals
(Boecklen and Larson 1994; Fritz et al. 1994; Morrow
et al. 1994; Whitham et al. 1994).
Quercus crassifolia H. and B. and Quercus crassipes
H. and B. are two red oak species (subg. Erythrobalanus)
that overlap at the Eje Neovolca
´nico, and in the
southeastern part of the Sierra Madre Oriental in central
Mexico, where they meet to form hybrids (Quer-
cus·dysophylla Benth pro sp.). In a previous study, each
of these species was well characterized by leaf mor-
phology and genetic markers (random amplified poly-
morphic DNAs; RAPDs), and it was reported that the
geographic proximity of hybrid plants to the allopatric
site of a parental species increases their morphological
and genetic similarity with the latter (Tovar-Sa
´nchez and
Oyama 2004). If the parental species have species-spe-
cific parasitic insects (gall-forming and leaf-mining) and
if attractant and (or) survival factors with regard to
these parasites present co-dominant or intermediate
inheritance, hybrid host trees may harbor parasites of
703

both parental species. In such case, the endophagous
insect species associated with the hybrid plant will
increase its similarity with the closest parental species.
Questions that have not been addressed about a
hybrid system are:
1. Does geographic proximity of hybrid plants to the
allopatric area of a parental species increase the
similarity of their gall-forming insects with those of
the parental species?
2. Does a bidirectional introgression support the hybrid
bridge hypothesis?
3. Does population genetic diversity (generated by
interspecific hybridization) in a hybrid zone affect the
dependent insect community?
We compared endophagous insects among trees of
Q. crassipes,Q. crassifolia and derived hybrids in seven
hybrid zones to determine how hybrid trees modify the
patterns of species richness, diversity, and composition
of insects in time and space. Also, we determined the
effect of hybridization on patterns of infestation of gall-
forming and leaf-mining guilds, whether each host taxa
(Q. crassifolia,Q. crassipes and hybrids) has specific
endophagous insects, and if these patterns vary among
localities.
Materials and methods
Species
Quercus crassifolia includes large trees up to 23 m in
height with a trunk diameter of 1 m. Leaves are
deciduous, aristate (bearing an awn or bristle at the
tip), ovate (egg-shaped in outline and attached at the
broad end), obovate (inversely ovate, with the attach-
ment at the narrower end) or elliptic (in the shape of an
ellipse, or a narrow oval; broadest at the middle and
narrower at the two equal ends), with a coriaceous
(leathery texture) upper surface; the lower surface is
yellow tomentose (with a covering of short, matted or
tangled, soft, wooly hairs; with tomentum), orange or
brown. The flowering period is in April (Romero 1993).
This species presents a broad range of geographical
distribution in Mexico, occupying the major mountain
ranges (Sierra Madre Oriental, Sierra Madre Occi-
dental, Sierra Madre del Sur and Eje Neovolca
´nico). Q.
crassipes include trees up to 17 m tall and 0.40–1 m in
trunk diameter. Leaves are deciduous, coriaceous,
narrowly elliptic and lanceolate, their surface is barely
lustrous and the lower surface is tomentose, white-
grayish. The flowering period is in May. It is distributed
within the southeast part of the Sierra Madre Oriental
and at the Eje Neovolca
´nico. Where the two oak spe-
cies meet and overlap they form hybrid individuals with
intermediate characteristics. The flowering phenology
may explain the occurrence of hybridization between
these two species. We observed flowering overlap in the
last days of April.
The parental and hybrid plants used for this study
had been morphologically (leaf characters) and geneti-
cally (RAPDs) examined (Tovar-Sa
´nchez and Oyama
2004). Also, all the individuals were in mature stages and
without any apparent damage. Plants that were found to
be backcrosses were excluded from this study. Six
primers were used to estimate the genetic status of Q.
crassifolia and Q. crassipes plants. These primers yielded
49 distinct markers (bands). The maximum likelihood
(ML) hybrid index score from RAPD analysis using six
RAPD markers supported the field identification of 250
plants. Few plants had perfect marker additivity as may
be expected in F
1
(12 plants), but 54 individuals were
interpreted as F
1
, 26 deviated by only one character
(0.437–0.562), and 16 plants deviated by only two
characters (0.375–0.625). Twelve plants were interpreted
as backcrosses toward Q. crassipes (0.250–0.312) and 11
as backcrosses toward Q. crassifolia (0.687–0.750) (see
Fig. 5 in Tovar-Sa
´nchez and Oyama 2004).
Study sites
We found seven zones in which hybridization occurs
naturally between these two species, through the Eje
Neovolca
´nico and the southeastern part of the Sierra
Madre Oriental in central Mexico (Cantera, Canalejas,
Tlaxco, Acajete, La Esperanza, Agua Blanca, and Palo
Bendito). In each locality the parental species were
dominant while the hybrid individuals were infrequent
(between ten and 17 trees), requiring an extensive field
search. The altitude of hybrid zones varied between
1,790 (Agua blanca) and 2,772 m (Tlaxco). Hybrid
zones with high levels of disturbance (i.e., Canalejas,
Acajete, Esperanza; and Agua blanca) had the highest
number of hybrid individuals.
A total of 194 trees were sampled in the seven hybrid
zones, ten trees for each parental species, and a variable
number of trees of the hybrid (Cantera, n=9; Canalejas,
n=10; Tlaxco, n=6; Acajete, n=5; Esperanza, n=9;
Agua blanca, n=8; Palo bendito, n=7) per hybrid zone.
Leaf-mining moths (Lepidoptera: Tischeridae, Cithera-
niidae) and gall-forming wasps (Hymenoptera: Cynipi-
dae) sampled in each host tree were collected and
separated to morphospecies level and placed in plastic
containers previously vouchered (i.e., locality, host cat-
egory, season, etc.) and transported to the laboratory,
where the insects emerged. Insects were identified to the
most approximate taxonomical level. The total number
of individuals that emerged varied depending on the
species and gall and mine density (oscillating between
five and 834 individuals).
Infestation by leaf-mining and gall-forming insects
associated with each host tree was estimated using four
randomly selected branches and 200 leaves (50 leaves per
branch). For each insect species average infestation va-
lue was estimated (number of galls or miners)/200
leaves·100) over the four branches. A total of 155–200
leaves in seven hybrid zones and for four seasons were
704

sampled. The number and type of galls and mines were
recorded for each host tree in the seven hybrid zones.
Analysis
Nested ANOVAs were conducted (model I fixed effects;
Zar 1999) to determine the effects of oak host taxa (Q.
crassifolia,Q. crassipes and hybrid), hybrid zone
(locality) and tree (individuals) on the gall-forming and
leaf-mining infestation percentage. Trees were consid-
ered as a random factor nested within species, because
they were representative of each population. One-way
ANOVA was performed separately for each hybrid zone
to determine the effect of season on gall-forming and
leaf-mining infestation percentage. Percentage data were
corrected as: X=arcsin (%)½ (Zar 1999).
A Tukey analysis was conducted to determine dif-
ferences in mean infestation (%) between hybrid plants
and parental species for all hybrid zones and within each
hybrid zone. The correlation analysis was used to test
the relationship between infestation of leaf-mining and
gall-forming insects associated with each oak host taxa.
A cluster analysis was done to group individual oak
host taxa in each hybrid zone according to taxonomic
similarity of endophagous insects. Discontinuous data
were transformed as: X=(x)½+0.5 (Zar 1999).
Endophagous insect diversity (Shannon-Wiener in-
dex; H¢) and richness was estimated as follows: for each
hybrid zone we used five hybrid individuals sampled
corresponding to the same number of each parental
species (individuals for each parental species were ran-
domly sampled). Differences in endophagous species
composition among the different host categories was
tested using non-metric multidimensional scaling
(NMDS) based on the presence or absence of 35 endo-
phagous insects. NMDS is a robust ordination technique
for community analysis (Clarke 1993), which has been
used to analyze differences in arthropod community
composition (Dungey et al. 2000; Wimp et al. 2004,2005).
NMDS was used to create a dissimilarity matrix among
the host categories (Q. crassifolia,Q. crassipes and
hybrids) using the Bray–Curtis dissimilarity coefficient
(Faith et al. 1987). Analysis of similarity (ANOSIM) was
used to evaluate endophagous community composition
differences among host categories. Bootstrap analysis
and ANOSIM were employed to test for differences
among groups using 1,000 random reassignments and
determine whether the generated dissimilarity matrix is
significantly different than chance (Warwick et al. 1990).
Multiple comparisons in ANOSIM were made using a
sequential Bonferroni correction (Zar 1999).
Correlation analysis was used to test the relationship
between endophagous diversity (H¢) and genetic diver-
sity of the hybrid zone (H). Genetic diversity of each
hybrid zone was evaluated using three (Ccmp3, Ccmp4,
and Ccmp41) cpSSR loci, which were polymorphic in
210 individuals of Q. crassifolia,Q. crassipes and a
derived hybrid; a total of 16 alleles (having either five or
six alleles per locus) and 26 haplotypes were found
(E. Tovar-Sa
´nchez et al., in progress). We registered the
genetic variation of each hybrid zone using the allele
frequencies per locus in each population. Also, we reg-
istered for each taxon the haplotypic variation, which
was estimated using the unbiased genetic diversity
(H; Nei 1987).
Results
Infestation of gall-forming and leaf-mining insects
associated with the Q. crassifolia·Q. crassipes complex
The mean percentage±SE of infestation of leaf-mining
insects was 3.03±0.09 and for gall-forming insects was
2.34±0.07. The percentage of infestation for leaf-mining
and gall-forming insects varied significantly among oak
host taxa, hybrid zones, trees and between hybrid zo-
nes·tree (Table 1). Also, significant differences among
seasons for all seven hybrid zones were obtained for leaf
mining (Table 2) but not for gall-forming insects.
Infestation levels of endophagous insects varied
significantly among host taxa; in general, the highest
percentages of infestation by leaf-mining insects were
Table 1 Nested ANOVA results to determine the effects of oak host taxa (Quercus crassifolia,Q. crassipes and hybrid), hybrid zone (H),
tree (T), and interaction (H·T) on the gall-forming and leaf-mining infestation percentage
Source SS df MS F
Gall-forming insect
Hybrid zone (H) 135.66 6 22.61 9.54***
Host taxa 654.21 2 327.10 46.02***
Tree (T) 36.93 189 0.20 2.60***
H·T 31.36 12 2.61 2.21***
Error 14.22 840 0.02
Leaf-mining insect
H 67.89 6 11.32 3.17***
Host taxa 996.52 2 498.26 43.52***
T 59.05 189 0.31 2.76***
H·T 38.89 12 3.24 1.82*
Error 21.42 840 0.03
*P<0.05, ***P<0.001
705

recorded for Q. crassifolia (4.16±0.19), followed by the
hybrids (2.95±0.15), and Q. crassipes (2.10±0.12). In
contrast, infestation levels for gall-forming insects
showed an inverse pattern: the highest infestation was
recorded for Q. crassipes (2.96±0.14), followed by the
hybrid hosts (2.53±0.12), and Q. crassifolia (1.53±0.07)
(Fig. 1).
Leaf-mining infestation percentages ranged between
1.57±0.19% (mean±SE) (Q. crassipes at Canalejas)
and 5.45±0.58% (Q. crassifolia at Tlaxco). Gall-form-
ing infestation values ranged between 1.08±0.16% (Q.
crassifolia at Tlaxco) and 4.76±0.66% infestation (Q.
crassipes at Acajete). In general, patterns of infestation
varied from site to site (Fig. 2).
Negative and significant relationships were found
between the log gall infestation versus log mine infes-
tation for Q. crassifolia (r=0.49, F=107.90,
P<0.001), hybrids (r=0.55, F=152.83, P<0.001),
and for Q. crassipes (r=0.51, F=122.03, P<0.001)
(Fig. 3).
Species composition of endophagous insects associated
with the Q. crassifolia·Q. crassipes complex in seven
hybrid zones
Canopy endophagous communities in the Q. crassifoli-
a·Q. crassipes complex comprised 35 species belonging
to two orders [Hymenoptera (n=32) and Lepidoptera
(n=3)], three families (Cynipidae, Tischeridae and Ci-
theraniidae) and ten genera: Adleria,Amphibolips,An-
dricus,Atrusca,Biorhiza,Callirhytis,Conobius,
Disholcaspis,Thischeria,Anisota (Appendix 1). The
most important genera for relative species richness were:
Andricus (34%), Disholcaspis (20%), Conobius (14%),
Amphibolips (11%), and Thischeria (6%).
In general, we found significant differences in endo-
phagous insect species composition among host cate-
gories (ANOSIM r=0.4135, n=194, P<0.001; Fig. 4),
showing that these communities on each host category
are significantly different from one another. Also, the
differences among categories were significant after cor-
recting the critical value of alpha for inflated type-II
errors (P<0.02 for all comparisons): Q. crassifolia ver-
sus F
1
hybrid, r=0.4705, P<0.001; Q. crassifolia versus
Q. crassipes,r=0.5392, P<0.0001; Q. crassipes versus
F
1
hybrid, r=0.2748, P<0.001.
Table 2 ANOVA results to determine the effect of season on percentage of leaf-mining infestation associated with the Q. crassifolia·Q.
crassipes complex in seven hybrid zones in Mexico
Hybrid zone Source SS df SM F
Santa Clara Season 164.30 3 54.77 9.97***
Error 16.45 145
Canalejas Season 86.04 3 28.68 3.71**
Error 23.21 145
Tlaxco Season 98.09 3 24.52 2.95*
Error 33.25 145
Acajete Season 134.09 3 33.52 5.05***
Error 26.53 145
Esperanza Season 171.58 3 42.89 5.62***
Error 30.51 145
Agua Blanca Season 115.28 3 28.82 3.63**
Error 31.73 145
Palo Bendito Season 129.96 3 32.49 4.12**
Error 31.51 145
*P<0.05,** P<0.01, ***P<0.001
Fig. 1 Percent infestation (mean±SE) of aleaf-mining insects and
bgall-forming wasps associated with Quercus crassifolia,Q.
crassifolia·Q. crassipes hybrids and Q. crassipes in central Mexico.
Different letters show significant differences at P<0.05 (Tukey’s
honestly significant difference test)
706

Diversity
In four hybrid zones, Q. crassipes had the highest species
diversity followed by hybrids and Q. crassifolia. In three
hybrid zones, hybrids had higher species diversity than
Q. crassipes and Q. crassifolia. In all cases, Q. crassifolia
had the lowest diversity values (Table 3).
In general, species richness showed an unidirectional
gradient in all hybrid zones except in Esperanza where
Q. crassipes had the highest values, followed by hybrids
and then Q. crassifolia. In Esperanza, the hybrid indi-
viduals showed the highest value, followed by Q. crass-
ipes and Q. crassifolia.
We found a significant and positive relationship be-
tween endophagous diversity (H¢) and oak hybrid zone
genetic diversity (H)(r
2
=0.779, n=7 hybrid zones,
P=0.003, Fig. 5).
Specificity of the host plant
We found 31.46% of endophagous insects present on Q.
crassipes (Andricus linaria, five species of Andricus,
Amphibolips sp., Atrusca pumilio,Callirhytis sp., and two
species of Discholcaspis) of which 45.5% (two species of
Andricus,A. pumilio, and two species of Discholcaspis)
had colonized hybrid individuals. From the 22.88%
of insects that infested Q. crassifolia (Anisota sp.,
Amphibolips sp., Disholcaspis pallens, two species of
Disholcaspis,Callirhytis sp.), 66.6% (D. pallens, two
Fig. 2 Percent infestation (mean±SE) of leaf-mining insects (open
bars) and gall-forming wasps (shaded bars) associated with Q.
crassifolia,Q. crassifolia·Q. crassipes hybrids and Q. crassipes in
seven hybrid zones in central Mexico: aCantera, bCanalejas,
cTlaxco, dAcajete, eEsperanza, fAgua Blanca, and gPalo
Bendito. Different letters show significant differences at P<0.05
(Tukey’s honestly significant difference test)
707

species of Disholcaspis,Callirhytis sp.) had colonized
hybrid trees. Only 8.58% of the endophagous insects
were hybrid specific (Andricus sp., Conobius sp., and
Callirhytis sp.), 11.44% (Amphibolips sp., two species of
Thischeria, and Andricus peredurus) are generalist insects
that infest all host taxa in all localities, 2.86% (Dish-
olcaspis cinerosa) are associated with all the complexes in
the Eje Neovolca
´nico, 2.86% (Callirhytis sp.) with the
complex of Esperanza and Palo Bendito, and another
2.86% (Andricus sp.) with the complex of Palo Bendito;
the remaining 22.8% (Andricus tibialis, two species of
Andricus, Adleria sp., Amphibolips sp. Biorhiza sp.,
Callirhytis sp., Disholcaspis sp.) were not specific to a
host plant taxa.
Similarity between host taxa and among hybrid zones
Endophagous insect compositions could be used to
group the Q. crassifolia and Q. crassipes populations by
their taxonomic affinity and not by geographical posi-
tion of hybrid zones in central Mexico, except for Tlaxco
that is localized in the central region of the Eje Neo-
volca
´nico. The similarity of insects associated with hy-
brid hosts in comparison to their parental species,
changes between hybrid zones (Fig. 6). Endophagous
insects associated with hybrids in Cantera and Canalejas
showed more similarity to Q. crassifolia from Cantera,
Canalejas and Acajete. The host hybrids from Agua
Blanca and Palo Bendito were more similar to Q.
crassifolia from Esperanza, Palo Bendito and Agua
Blanca. On the other hand, the dependent insects of host
hybrids from Esperanza and Acajete were more similar
to those of Q. crassipes from Acajete, Canalejas, Espe-
ranza, Tlaxco, Palo Bendito and Agua Blanca. Finally,
in Tlaxco, insects associated with the host hybrids were
more similar to those of their putative parentals. We
performed a separate cluster analysis on the more
common species (Andricus pereduris,D. cinerosa,Calli-
rhytis sp., Amphibolips sp., Andricus sp., and two species
of Thischeria). These results are in agreement with
Fig. 6, which shows that the observed pattern is deter-
mined by only a few species.
Discussion
Our study reveals that genetic host taxa and locality
variation affect significantly the community structure of
canopy endophagous insects. Five gall-forming wasp
species that were specific to Q. crassipes (belonging to
three genera: Amphibolips,Andricus and Callirhytis) and
four wasp species that were specific to Q. crassifolia
(belonging to two genera: Callirhytis and Discholcaspis),
have already colonized hybrid individuals, favoring a
future host-species shift. Our findings support the hybrid
Table 3 Shannon diversity index (H¢) and species richness (S) of gall-forming insect for each taxa, seven hybrid zones, and four seasons in
central Mexico
Taxa March June September December Total
H¢SH¢SH¢SH¢SH¢S
Cantera
Q. crassifolio 0.84 9 0.79 10 0.71 9 0.71 8 0.74 10
Hybrids 1.01 7 0.86 8 0.91 7 0.84 8 0.90 8
Q. crassipes 0.95 11 0.81 12 0.77 13 0.79 13 0.87 14
Canalcjas
Q. crassifolia 0.39 3 0.47 4 0.50 2 0.32 3 0.48 4
Hybrids 0.56 5 0.51 6 0.59 4 0.50 5 0.65 6
Q. crassipes 0.88 10 0.90 12 0.89 11 0.88 9 0.89 12
Tlaxco
Q. crassifolia 0.58 5 0.65 6 0.70 6 0.66 5 0.69 6
Hybrids 0.80 8 0.77 8 0.75 8 0.80 8 0.77 8
Q. crassipes 0.68 8 0.76 9 0.70 9 0.72 8 0.74 9
Acajete
Q. crassifolia 0.69 6 0.65 7 0.63 8 0.62 7 0.67 8
Hybrids 0.82 11 0.67 12 0.66 12 0.67 12 0.70 12
Q. crassipes 0.82 12 0.82 15 0.84 14 0.83 13 0.87 15
Esperanza
Q. crassifolia 0.45 5 0.56 4 0.56 6 0.55 5 0.54 6
Hybrids 0.82 6 0.74 8 0.77 7 0.75 9 0.78 9
Q. crassipes 0.67 6 0.71 7 0.74 8 0.74 5 0.72 8
Agus blanca
Q. crassifolia 0.47 4 0.35 6 0.38 5 0.29 6 0.36 6
Hybrids 0.58 8 0.50 7 0.53 6 0.51 7 0.55 8
Q. crassipes 0.78 6 0.80 6 0.82 8 0.78 7 0.78 9
Palo bendito
Q. crassifolia 0.16 3 0.21 2 0.21 3 0.20 3 0.20 3
Hybrids 0.59 5 0.56 5 0.49 6 0.41 4 0.51 6
Q. crassipes 0.72 6 0.69 7 0.70 7 0.73 7 0.72 7
708

bridge hypothesis which suggests that the presence of a
hybrid intermediary plant may favor host–herbivore
shifts from one plant species to another (Floate and
Whitham 1993). Fifty-four percent of gall-forming in-
sects were associated with a particular host parental
species (Q. crassifolia or Q. crassipes), and then shifted
to the hybrid plant. This complex presents a bidirec-
tional introgression (Tovar-Sa
´nchez and Oyama 2004),
that results in a complete morphological and genetic
continuum of hybrids with no gaps that may inhibit a
host shift (Floate and Whitham 1993). Also, intermedi-
ate inheritance of leaf morphology and genetic traits has
been shown in hybrid plants of Q. crassifolia·Q. crass-
ipes (Tovar-Sa
´nchez and Oyama 2004). Therefore, the
new hybrid host should have very similar characteristics
(e.g., leaf morphology, defense, phenology and second-
ary compounds) to both parental species, which facili-
tates the shift of the endophagous insects from the
parental species to the hybrid individual. Currently,
mitochondrial DNA analyses of gall forming wasps are
being conducted in order to determine in which direction
the host shift took place, and the chemical analysis of
the leaf complex is being conducted to determine the
variation of secondary metabolites.
Endophagous community composition was signifi-
cantly different on F
1
hybrids relative to their Q.
crassifolia and Q. crassipes parents. These findings sug-
gest that genetic differences among oak host categories
exert a strong organizing influence on the endophagous
community. These findings agree with those of Fritz
et al. (1994), Dungey et al. (2000), Wimp et al. (2004,
2005), who found that the arthropod community com-
position on hybrid and parental species was significantly
different.
In general, the hybrid hosts support intermediate
percentages of infestation in relation to their parentals
(when analyzing all hybrid zones together), showing an
increasing and unidirectional gradient towards Q.
crassifolia for leaf-mining and towards Q. crassipes for
gall-forming insects. Our study supports the additive
Fig. 3 Relationships between gall and mine infestations in aQ.
crassifolia,bhybrids, and cQ. crassipes in Mexico. aLog
10
(gall
infestation)=0.54–0.33 log
10
(mine infestation), r=0.49,
r
2
=0.24; blog
10
(gall infestation)=0.68–0.48 log
10
(mine infesta-
tion), r=0.55, r
2
=0.31; clog
10
(gall infestation)=0.71–0.50 log
10
(mine infestation), r=0.51, r
2
=0.26
Fig. 4 Differences in endophagous (gall-forming and leaf-mining
insects) community composition among oak host Q. crassifolia,
hybrids, and Q. crassipes. Each point is a two-dimensional (axis 1
and axis 2) representation of endophagous species composition on
an individual tree based on global, non-metric multidimensional
scaling (NMDS). Distances between points reflect a dissimilarity
matrix created using the Bray–Curtis dissimilarity coefficient (Faith
et al. 1987). Points that are close together indicate endophagous
communities that are more similar in composition
709

hypothesis (Fritz et al. 1996), which predicts that hybrid
plants (F
1
) are intermediate between the herbivore
resistances of the parental species. Examples from other
studies showing the same patterns are: beetles on elms
(Hall and Towsend 1987), Pontonia on European wil-
lows (Soetens et al. 1991), beetles on willows (Soetens
et al. 1991), sawflies on birches (Hanhima
¨ki et al. 1994),
several insect species on oaks (Aguilar and Boecklen
1992; Boecklen and Larson 1994), and several species of
insects on American willows (Fritz et al. 1994,1996).
It has been well documented that the increase in the
size of oak hosts and their geographical range (e.g.,
Cornell and Washburn 1979; Cornell 1986), are corre-
lated with cynipid species richness. Still, none of these
factors appear to provide a significant explanation for
the observations made on Q. crassifolia, for which the
largest geographical range and tree size as well as the
lowest gall-forming insect richness were found in com-
parison to Q. crassipes. However, our data are sup-
ported by Price et al. (2004) who found a negative
species–area relationship for cynipids and oaks, and
also for host-plant height and cynipid richness. We
suggest that the greatest gall-forming insect richness on
Q. crassipes can be influenced by different crown
structures that may provide diverse resources and a
great array of niches; and also, the different chemical
substances among taxa can strongly affect the resistance
of plants to herbivores (Rowell-Rahier 1984; Orians
and Fritz 1995; Orians 2000; Osier and Lindroth 2001).
In general, one would expect host-plant species with
higher nitrogen levels and lower levels of compounds
which reduce digestibility such as tannins to be most
favorable for insects (e.g., Slansky and Scriber 1985).
Ultimately, it will be important to determine the vari-
ability of chemical substances among taxa and between
seasons.
In contrast, the highest number of leaf-mining insects
was recorded for Q. crassifolia in comparison to Q.
crassipes, probably because the former trees are larger
and their canopy structure is characterized by a larger
leaf size, which represents more varied habitats and re-
sources than do smaller tree hosts (Lawton 1978; Strong
et al. 1984). For example, Simberloff and Stiling (1987)
reported a positive relationship between leaf size and
oviposition rate in a leaf-mining moth (Stilbosis quad-
ricustatella)onQuercus geminate and Q. gambelii com-
pared to Q. grisea, and host plant height may increase
leaf mining oviposition rate of some species (Connor
et al. 1983).
In general, leaf-mining and gall-forming insects are
known to have narrow host preferences (Shoonhoven
et al. 1998; Stone et al. 2002). Oak gall wasps (Cynipi-
dae) are obligate parasites and are considered to be or-
gan species specific (Stone et al. 2002). The precise
causes of such differences in richness and abundance
between gall-forming and leaf-mining insects on Q.
crassipes and Q. crassifolia are not understood. How-
ever, the limited dispersal of cynipids and their complex
life cycle, requiring two different resources for the two
generations, may be contributing factors to their more
limited distribution.
Variation in infestation levels and species diversity of
endophagous insects on hybrid plants in relation to their
parental species among hybrid zones suggested that ge-
netic host taxa and locality variation factors play a role
in the infestation patterns of this system (Strauss 1994).
The same resistance traits can be expressed differently
among environments (hybrid zones). Therefore, envi-
ronmental changes (e.g., temperature, precipitation, soil,
humidity, solar radiation, etc.) could affect the expres-
sion of resistance (to parasites) among genotypes of
many plant species (Fritz et al. 1996).
Our data suggest that the conservation of genetic
diversity is much more than a species issue; it is also an
important community issue. This was demonstrated by
the fact that 78% of endophagous diversity could be
accounted by genetic diversity in oak hybrid zones.
Preserving genetic variation in plant populations may
preserve greater species richness in the dependent com-
munity (Whitham et al. 1999). Furthermore, application
of genetics to the preservation of biodiversity has the
potential to be broadly applicable because genes are the
products of long-term interactions, whereas ecological
patterns are more likely to vary from one year to the
next (Wimp et al. 2004).
The distribution of Q. crassifolia ranges from the
Sierra Madre Occidental to the center of Mexico along
the Eje Neovolca
´nico, while that of Q. crassipes ranges
from the Sierra Madre Oriental to the Eje Neovolca
´nico,
where both species overlap producing hybrid zones. The
Eje Neovolca
´nico is located in the central part of the
country, lying in an east–west direction, and is consid-
ered geologically the youngest mountain range in Mex-
ico. It is hypothesized that its evolutionary process
began during the Mid-Tertiary with the formation of the
Fig. 5 Oak genetic diversity complex predicted endophagous insect
diversity at a stand level. Each point in the scatterplot represents the
Shannon–Wiener index of endophagous diversity as a function of
the oak genetic diversity complex in the same hybrid zone; a total
of seven hybrid zones were analyzed in two regions in central
Mexico: Cantera, Canalejas, Tlaxco, Acajete, Esperanza in the Eje
Neovolca
´nico region and Agua Blanca and Palo Bendito in the
Sierra Madre oriental region
710

occidental portion followed by the development of the
central and oriental portion during the Quaternary–
Pliocene (Ferrusquı
´a-Villafranca 1993). Tovar-Sa
´nchez
and Oyama (2004) showed that in the Q. crassifolia·Q.
crassipes complex, the geographic proximity of hybrid
plants to the allopatric site of a parental species increases
their morphological and genetic similarity with their
parental species. In this study, the similarity of endo-
phagous insects of hybrid plants increases with the
proximity of parental species of an allopatric site. Thus,
the Eje Neovolca
´nico acts as a corridor in which its
proximity to an allopatric site favors the shift of parasite
species (gall forming and leaf mining) associated with
the parental species to closely related hybrid hosts. Also,
the genetic diversity of the hybrid zone decreases in an
east–west direction on the Eje Neovolca
´nico, supporting
the idea that the Eje Neovolca
´nico acts as a corridor.
Seasonal variation during a 1-year period did not
affect the infestation rate and species diversity of gall-
forming insects. We suggest that these results are due to
the fact that wasps (Cynipidae) associated with the
complex that develop galls on ephemeral structures (e.g.,
leaves, acorns, stems etc.) of the plant may prolong the
life cycle of these structures in the host, extending the
available period for insect development (Ananthakri-
sham 1984; Stone et al. 2002).
In this study, we have shown that genetic differences
among oak host taxa result in extended phenotypes that
structure the endophagous insect community. Also, the
pure species and F
1
hybrids supported different endo-
phagous communities. We found a significant and
positive relationship between endophagous diversity
(H¢) and oak hybrid zone genetic diversity. Therefore,
we conclude that in order to maintain insect diversity,
host plant genetic diversity must be conserved in the
hybrid zones. Furthermore, we suggest that in order to
gain a better understanding of speciation of endopha-
gous insects on oaks, it is important to conduct phylo-
genetic analyses of the insects on each host species.
Acknowledgements We thank Jim Cronin and two anonymous
reviewers for their comments that improved the original manuscript
considerably. The authors thank Patricia Mussali, Susana Valencia,
Rocio Esteban, Marco Romero, Mauricio Mora, and Maribel
Paniagua for technical assistance. This research was supported by a
PAEP–UNAM grant and a CONACYT scholarship to E. T. S.
References
Aguilar JM, Boecklen WJ (1992) Patterns of herbivory in the
Quercus grisea·Quercus gambelii species complex. Oikos
64:498–504
Anderson E (1949) Introgressive hybridization. Wiley, New York
Ananthakrisham NT (1984) The biology of gall insects. Arnold,
London
Bacilieri R, Ducousso A, Kremer A (1995) Genetic, morphological,
ecological and phenological differentiation between Quercus
petrea (Matt.) Liebl. and Quercus robur L. in a hybrid zone of
the northwest. Silvae Genet 44:1–1
Boecklen WJ, Spellenberg R (1990) Structure of herbivore com-
munities in two oak (Quercus spp.) hybrid zones. Oecologia
85:92–100
Boecklen WJ, Larson KC (1994) Gall-forming wasp (Hymenop-
tera: Cynipidae) in an oak hybrid zone: testing hypotheses
about hybrid susceptibility to herbivores. In: Price PW, Matt-
son WJ, Baranchikov YN (eds) The ecology and evolution of
gall-forming insects. North Central Forest Experimental Sta-
tion, Forest Service, USDA, St. Paul, Min., pp 110–120
Bruschi P, Vendramin GG, Bussotti F, Grossoni P (2000) Mor-
phological and molecular differentiation between Quercus pe-
trea (Matt.) Liebl. and Quercus pubescens Willd. (Fagaceae) in
northern and central Italy. Ann Bot 85:325–333
Clarke KR (1993) Non-parametric multivariate analyses of chan-
ges in community structure. Aust J Ecol 18:117–143
Connor EF, Faeth SH, Simberloff D (1983) Leaf-miners on oaks:
the role of immigration in situ reproductive recruitment. Ecol-
ogy 64:191–204
Fig. 6 Cluster analysis of
endophagous insects associated
with Q. crassifolia,Q.
crassifolia·Q. crassipes hybrids,
and Q. crassipes in seven hybrid
zones (listed in Fig. 5)in
Mexico
711

Cornell HV (1986) Oak species attributes and host size influence
cynipine wasp species richness. Ecology 67:1582–1592
Cornell HV, Washburn JO (1979) Evolution of the richness-area
correlation for cynipid gall wasp on oak tree: a comparison of
two geographic areas. Evolution 33:257–274
Dumolin-Lapegue
`S, Demesure B, Fineschi S, LeCorre V, Petit RJ
(1997) Phylogeographic structure of white oaks throughout the
European Continent. Genetics 146:1475–1487
Dungey HS, Potts BM, Whitham TG, Li HF (2000) Plant genetics
effects arthropod community richness and composition: evi-
dence from a synthetic eucalypt hybrid population. Evolution
54:1938–1946
Faith DP, Minchin PR, Belbin L (1987) Compositional dissimi-
larity as a robust measure of ecological distance. Vegetation
69:57–68
Ferrusquı
´a-Villafranca I (1993) Geology of Mexico: a synopsis. In:
Ramamoorthy TP, Bye R, Lot A, Fa J (eds) Biological diversity
of Mexico: origins and distribution. Oxford University Press,
New York, pp 3–107
Floate KD, Whitham TG (1993) The ‘‘hybrid bridge’’ hypothesis:
host shifting via plant hybrid swarms. Am Nat 141:651–662
Fritz RS, Simms EL (1992) Plant resistance to herbivores and
phatogens. University of Chicago Press, Chicago, Ill.
Fritz RS, Nichols-Orians CM, Brunsfeld SJ (1994) Interspecific
hybridization of plants and resistance to herbivores: hypothesis,
genetics, and variable responses in a diverse herbivore com-
munity. Oecologia 97:106–117
Fritz RS, Roche BM, Brunsfeld SJ, Orians CM (1996) Interspecific
and temporal variation in herbivores responses to hybrid wil-
lows. Oecologia 108:121–129
Fritz RS, Roche BM, Brunsfeld SJ (1998) Genetic variation in
resistance of hybrid willows to herbivores. Oikos 83:117–128
Fritz RS, Moulia C, Newcombe G (1999) Resistance of hybrid
plants and animals to herbivores, pathogens, and parasites.
Annu Rev Ecol Syst 30:565–591
Grant V (1981) Plant speciation. 2nd edn. Columbia University
Press, New York
Guttman SI, Weight LA (1989) Electrophoretic evidence of rela-
tionships among Quercus (oaks) of eastern North America. Can
J Bot 67:339–351
Hall RW, Towsend AM (1987) Suitability of Ulmus wilsoniana, the
‘‘urban’’ elm, and their hybrids for the elm leaf beetle, Xan-
thogaleruca luteola (Mu
¨ller) (Coleoptera: Chrysomelidae).
Environ Entomol 16:1042–1044
Hanhima
¨ki S, Senn J, Haukioja E (1994) Permanence of insect
herbivores on hybridizing trees: the case of the subarctic bir-
ches. J Anim Ecol 63:163–175
Hardin JW (1975) Hybridization and introgression in Quercus alba.
J Arnold Arbor 56:336–363
Jensen RJ, Eshbaugh WH (1976) Numerical taxonomic studies of
hybridization in Quercus. Populations of restricted aerial dis-
tribution and low taxonomic diversity. Syst Bot 1:1–10
Knops JMH, Tilman D, Haddad NM, Naeem S, Mitchell CE,
Haarstad J (1999) Effects of plant species richness on invasion
dynamics, disease outbreaks, insect abundance and diversity.
Ecol Lett 2:286–293
Lawton JH (1978) Host-plant influences on insect diversity: the
effects of space and time. Symp R Entomol Soc Lond 9:105–125
Messina FJ, Richards JH, McArthur ED (1996) Variable responses
of insects to hybrid versus parental sagebrush in common gar-
dens. Oecologia 107:513–521
Moorehead JR, Taper ML, Case TJ (1993) Utilization of hybrid
oak hosts by a monophaghous gall wasp: how little host char-
acter is sufficient? Oecologia 95:385–392
Morrow PA, Whitham TG, Potts BM, Ladiges P, Ashton DH,
Williams J (1994) Gall-forming insects concentrate on hybrid
phenotypes of Eucalyptus host. In: Price PW, Mattson WJ Jr,
Baranchikov YN (eds) The ecology and evolution of gall-
forming insects. North Central Forest Experimental Station,
Forest Service, USDA, St. Paul, Min.
Nei M (1987) Molecular evolutionary genetics. Columbia Univer-
sity Press, New York
Nixon KC (1993) The genus Quercus in Mexico. In: Nixon KC (ed)
Biological diversity of Mexico: origins and distributions,
Oxford University Press, New York, pp 447–458
Orians CM, Fritz RS (1995) Secundary chemistry of hybrid and
parental willows: phenolic glycosides and condensed tannins in
Salix sericea,S. eriocephala, and their hybrids. J Chem Ecol
21:1245–1253
Orians CM (2000) The effects of hybridization in plants on sec-
ondary chemistry: implications for the ecology and evolution of
plant-herbivore interactions. Am J Bot 87:1749–1756
Osier TL, Lindroth RL (2001) Effects of genotype, nutrient avail-
ability, and defoliation on aspen phytochemistry and insect
performance. J Chem Ecol 27:1289–1313
Price PW, Abrahamson WG, Hunter MD, Melika G (2004) Using
gall wasp on oaks to test broad ecological concepts. Conserv
Biol 18:1405–1416
Rieseberg LH, Brunsfeld SJ (1992) Molecular evidence and plant
introgression. In: Soltis PS, Soltis DE, Doyle JD (eds) Molec-
ular systematic of plants. Chapman and Hall, New York,
pp 151–176
Rieseberg LH, Elltrand NC (1993) What can molecular and mor-
phological markers tell us about plant hybridization? Crit Rev
Plant Sci 12:213–241
Rieseberg LH, Wendel JF (1993) Introgression and its conse-
quences in plants. In: Harrison RG (ed) hybrid zones and the
evolutionary process, Oxford University Press, Oxford, pp 70–
109
Romero RS (1993) El ge
´nero Quercus (Fagaceae) en el Estado de
Me
´xico. M.Sc. thesis, Universidad Nacional Auto
´noma de
Me
´xico
Rowell-Rahier M (1984) The food plant preferentes of Phratora
vitellinae (Coleoptera: Chrysomelinae). Oecologia 64:375–380
Shoonhoven LM, Jermy T, Van Loon JJA (1998) Insect–plant
biology: from physiology to evolution. Chapman and Hall,
London
Simberloff D, Stiling P (1987) Larval dispersion and survivorship in
a leaf-mining moth. Ecology 68:1647–1657
Simms EL, Rausher MD (1993) Patterns of selection on phyto-
phage resistance in Ipomoea purpurea. Evolution 47:970–976
Slansky F, Scriber JM (1985) Food composition and utilization. In:
Kerkurt G, Gilbert LI (eds) Comprehensive insect physiology,
biochemistry and pharmacology. Pergamon, Oxford, pp 87–163
Soetens P, Rowell-Rahier M, Pasteels JM (1991) Influence of
phenolglucosides and trichome density on the distribution
of insects herbivores on willows. Entomol Exp Appl 59:175–
187
Stace CA (1987) Hybridization and the plant species. In: Urbanska
KM (ed) Differentiation patterns in higher plants. Academic
Press, New York, pp 115–127
Stone GN, Scho
¨nrogge K, Atkinson RJ, Pujade-Villar J (2002) The
population biology of gall wasp (Hymenoptera: Cynipidae).
Annu Rev Entomol 47:633–668
Strauss SY (1994) Levels of herbivory and parasitism in host hy-
brid zones. Trends Ecol Evol 9:209–214
Strong DR, Lawton JH, Southwood TRE (1984) Insects on plants:
community patterns and mechanisms. Harvard University
Press, Cambridge, Mass.
Tovar-Sa
´nchez E, Oyama K (2004) Natural hybridization and
hybrid zones between Quercus crassifolia and Q. crassipes in
Mexico. Morphological and molecular evidence. Am J Bot
91:1352–1363
Warwick RM, Clarke KR, Suharsono A (1990) A statistical anal-
ysis of coral community responses to the 1982–1983 El Nin
˜oin
the Thousand Island, Indonesia. Coral Reefs 8:171–179
Wendel LH (1960) Cynipid Galls of the Southwest. Privately
printed, Ann Arbor, Mich.
Wendel JF, Steward JMcD, Rettig JH (1991) Molecular evidence
for homoploid reticulate evolution among Australian species of
Gossypium. Evolution 45:694–711
Whitham TG, Morrow PA, Potts BM (1994) Plant hybrid zones as
center of biodiversity: the herbivore community of two endemic
Tasmanian eucalypts. Oecologia 97:481–490
712

Whitham TG, Martinsen GD, Floate KD, Dungey H, Potts BM,
Keim P (1999) Plant hybrid zones affect biodiversity: tools for a
genetic-based understanding of community structure. Ecology
80:416–428
Whittemore AT, Schaal BA (1991) Interspecific gene flow in sym-
patric oaks. Proc Natl Acad Sci USA 88:2540–2544
Wimp GM, Young WP, Woolbright SA, Martinsen GD, Keim P,
Whitham TG (2004) Conserving plant genetic diversity for
dependent animal communities. Ecol Lett 7:776–780
Wimp GM, Gregory D, Martinsen GD, Floate KD, Bangert RK,
Whitham TG (2005) Plant genetic determinants of arthropod
community structure and diversity. Evolution 59:61–69
Zar JH (1999) Biostatistical analysis, 4th edn. Prentice Hall,
Englewood Cliffs, N.J.
713