Geographic divergence in a species-rich symbiosis: Interactions between Monterey pines and ectomycorrhizal fungi

Abstract

A key problem in evolutionary biology is to understand how multispecific networks are reshaped by evolutionary and coevolutionary processes as they spread across contrasting environments. To address this problem, we need studies that explicitly evaluate the multispecific guild structure of coevolutionary processes and some of their key outcomes such as local adaptation. We evaluated geographic variation in interactions between most extant native populations of Monterey pine (Pinus radiata) and the associated resistant-propagule community (RPC) of ectomycorrhizal (EM) fungi, using a reciprocal cross-inoculation experiment with all factorial combinations of plant genotypes and soils with fungal guilds from each population. Our results suggest that the pine populations have diverged in community composition of their RPC fungi, and have also diverged genetically in several traits related to interactions of seedlings with particular EM fungi, growth, and biomass allocation. Patterns of genetic variation among pine populations for compatibility with EM fungi differed for the three dominant species of EM fungi, suggesting that Monterey pines can evolve differently in their compatibility with different symbiont species.

Full text

Ecology, 93(10), 2012, pp. 2274–2285
Ó2012 by the Ecological Society of America
Geographic divergence in a species-rich symbiosis: interactions
between Monterey pines and ectomycorrhizal fungi
JASON D. HOEKSEMA,
1,7
JESUS VARGAS HERNANDEZ,
2
DEBORAH L. ROGERS,
3
LUCIANA LUNA MENDOZA,
4,5
AND JOHN N. THOMPSON
6
1
Department of Biology, 214 Shoemaker Hall, University of Mississippi, University, Mississippi 38677 USA
2
Colegio de Postgraduados, Km. 36.5 Carretera Mexico-Texcoco, Montecillo, Mexico 56230
3
Center for Natural Lands Management, 27258 Via Industria, Suite B, Temecula, California 92590 USA
4
Grupo de Ecologı´a y Conservacio
´n de Islas, A.C., Avenida Moctezuma 836, Zona Centro, Ensenada, Baja California, Mexico 22800
5
Centre for Biodiversity and Biosecurity, University of Auckland, Private Bag 92019, Auckland, New Zealand 1142
6
Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, California 95064 USA
Abstract. A key problem in evolutionary biology is to understand how multispecific
networks are reshaped by evolutionary and coevolutionary processes as they spread across
contrasting environments. To address this problem, we need studies that explicitly evaluate the
multispecific guild structure of coevolutionary processes and some of their key outcomes such
as local adaptation. We evaluated geographic variation in interactions between most extant
native populations of Monterey pine (Pinus radiata) and the associated resistant-propagule
community (RPC) of ectomycorrhizal (EM) fungi, using a reciprocal cross-inoculation
experiment with all factorial combinations of plant genotypes and soils with fungal guilds
from each population. Our results suggest that the pine populations have diverged in
community composition of their RPC fungi, and have also diverged genetically in several traits
related to interactions of seedlings with particular EM fungi, growth, and biomass allocation.
Patterns of genetic variation among pine populations for compatibility with EM fungi differed
for the three dominant species of EM fungi, suggesting that Monterey pines can evolve
differently in their compatibility with different symbiont species.
Key words: coevolution; ectomycorrhizal fungi; geographic divergence; Monterey pine; multispecific
mutualism; Pinus radiata.
INTRODUCTION
Natural selection on species varies among environ-
ments because gene expression and the fitness of
genotypes varies among environments, creating a
genotype by environment interaction for fitness (Wil-
liams 1966, Endler 1986). The result, if selection is
strong enough to counter gene flow, is the local
adaptation of populations (Hereford 2009) and, some-
times, ecological speciation (Schluter 2009). This process
can occur even more rapidly when the main agent of
selection is another species, because the species impose
reciprocal selection on each other, which often varies
among environments (genotype by genotype by envi-
ronment, or G 3G3E, interactions; Thompson 2005).
The result is a geographic mosaic of coevolution due to
three factors: (1) geographic variation in the structure of
coevolutionary selection, e.g., ranging from mutualism
to parasitism (selection mosaics, or G 3G3E
interactions); (2) geographic variation in the intensity
of coevolutionary selection, resulting in coevolutionary
hotspots and coldspots; and (3) trait remixing resulting
from gene flow, random genetic drift, and metapopula-
tion dynamics (Thompson 1994, 2005). The mosaics are
likely to increase in complexity as pairwise interactions
diversify into multispecific interactions. A key problem
in evolutionary biology is therefore to understand how
multispecific networks are reshaped by evolutionary and
coevolutionary processes as they spread across contrast-
ing environments.
Although multiple studies have now shown how
coevolution between two focal species may vary with
the presence of a third species (G 3G3G interactions;
e.g., Thompson and Cunningham 2002, Siepielski and
Benkman 2007, Thompson et al. 2010), a further
challenge is to understand how the outcomes of
coevolution vary among different members of diverse
interacting guilds (i.e., groups of species with similar
roles in communities). If we are to understand how
coevolution proceeds within multispecific assemblages,
we need to assess the extent to which all species in a
guild evolve in the same manner in response to evolution
in another guild, and which guilds are most likely to
drive the structure of selection within complex assem-
blages. Recent theoretical studies have suggested that
coevolution within mutualistic assemblages is driven
especially by supergeneralists interacting with species
across guilds (Guimaraes et al. 2011). We therefore need
Manuscript received 20 September 2011; revised 13 April
2012; accepted 16 April 2012. Corresponding Editor: C. V.
Hawkes.
7
E-mail: hoeksema@olemiss.edu
2274

empirical studies that assess not only geographic
differences in coevolution driven by other species, but
also the multispecific guild structure of coevolution and
its key outcomes such as local adaptation and varying
degrees of specialization.
We addressed this problem by experimentally manip-
ulating the ectomycorrhizal (EM) interaction between
most native populations of Monterey pine (Pinus radiata
D. Don) and members of the diverse guild of fungal
species that colonize the roots. Post-Pleistocene native
populations of Monterey pine are restricted to a small
set of geographically separated sites along the west coast
of California (USA) and two islands off Baja California
(Mexico) (Axelrod 1980), and therefore provide an
opportunity to study how geographically isolated
diverse interactions evolve amid limited gene flow
among populations. We focused on the guild of
ectomycorrhizal fungal species that typically colonize
pine roots via resistant propagules in the soil such as
spores and sclerotia, rather than via mycelial growth
from other colonized roots. In pine-dominated ecosys-
tems on the West Coast of North America, members of
the resistant propagule community (RPC) of ectomy-
corrhizal fungi are particularly important colonists of
pine seedlings and saplings in early secondary succes-
sion, after stand-replacing disturbances or during stand
expansion into adjacent habitat (Taylor and Bruns 1999,
Ashkannejhad and Horton 2006). Bishop pine (Pinus
muricata D. Don), a sister species to Monterey pine in
northern and central California, has been shown to have
a relatively low-diversity EM fungal community overall
(Horton and Bruns 2001), with an RPC fungal guild
dominated by members of the genera Rhizopogon,
Tomentella,Tuber, and Cenococcum (Gardes and Bruns
1996, Horton and Bruns 1998, Baar et al. 1999, Taylor
and Bruns 1999, Grogan et al. 2000).
In previous studies, we found that one of the RPC
fungi, Rhizopogon occidentalis, varies considerably in its
ecological interactions with Monterey pine, bishop pine,
and shore pine (Pinus contorta var. contorta Dougl. ex
Loud.), a more northerly coastal species. Using cross-
inoculation experiments with multiple genotypes of
fungi and pines, we found that within populations,
fungal and pine seedling performance frequently varied
with fungal genotype, pine genotype, and abiotic
factors, suggesting the potential for ongoing coevolution
(G 3G interactions) mediated by selection mosaics (G 3
G3E interactions) among populations (Piculell et al.
2008, Hoeksema et al. 2009). We also found evidence for
local adaptation of fungal populations to nearby pine
populations at a large geographic scale, suggesting pines
are exerting geographically variable selection on fungal
populations (Hoeksema and Thompson 2007). Never-
theless, much of the variation among populations in
pine seedling traits was independent of R. occidentalis
genotype and vice versa and was correlated with
latitude, pointing to the importance of abiotic selection
on plant and fungal traits (Hoeksema and Thompson
2007).
Rhizopogon occidentalis, however, is only one species
in this complex mycorrhizal assemblage. Here, we
expand this analysis by evaluating geographic variation
in the interaction between disjunct native Monterey pine
populations in coastal California and insular Mexico
and the associated RPC fungi. Using a reciprocal cross-
inoculation experiment with all factorial combinations
of plant genotypes and fungal guilds from each
population, we aimed to answer two questions:
1) Has the species composition of the RPC fungi
diverged among populations of Monterey pine?
2) Have populations of pines and fungi diverged
genetically in their compatibility or traits, and are
patterns of divergence consistent across different
members of the fungal guild?
METHODS
We used a factorial experiment in the laboratory to
assess the main and interactive effects of soils (contain-
ing resistant propagules of mycorrhizal fungi ) and
Monterey pine genotypes on plant performance traits
and the composition of ectomycorrhizal fungi colonizing
pine roots. We sampled soils and open-pollinated
families from five native stands of Monterey pine, and
combined them in all 25 possible combinations in a
growth chamber pot experiment. We measured plant
performance traits for 28 weeks and used molecular
methods to identify the mycorrhizal fungi that colonized
the roots of the experimental pine seedlings. We used
root colonization on these experimental seedlings
planted into field soils as a ‘‘bioassay’’ of the resistant
propagule community (RPC) of ectomycorrhizal fungi.
This procedure has been shown to provide a relatively
accurate assessment of the RPC of ectomycorrhizal
fungi that are ecologically important during early
succession in coastal pine forests (e.g., Baar et al.
1999, Taylor and Bruns 1999). Moreover, we posited
that variation among seedling genotypes in observed
relative abundances of fungal taxa in such a bioassay
would provide an estimate of plant genetic variation for
compatibility with those fungal taxa.
Collection and preparation of soil and seeds
Soil and seeds were collected from stands in two
native Monterey pine populations in mainland Califor-
nia (A˜
no Nuevo, 3783046.800 N, 122814016.800 W; Cam-
bria, 3583206.000 N, 12184048.000 W) and three stands in
insular Mexico (Guadalupe Island, 2989036.000 N,
118817060.000 W; northern Cedros Island, 28820 048.000
N, 115813024.000 W; southern Cedros Island, 28810045.900
N, 115812039.000 W). Although the latter two stands are
usually considered to be part of the same population
(Cedros Island), they are separated by ;20 km, and one
of our goals was to explore variation between these
isolated stands within the Cedros Island. Hereafter, we
October 2012 2275GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE

refer to the five sampled stands, including the two on
Cedros Island, as five ‘‘populations.’’ No sampling was
conducted in the pine population at Monterey, Cal-
ifornia, in an effort to reduce the complexity of the
experiment. Monterey pine was the dominant ectomy-
corrhizal tree species at all five sampling sites, although
occasional Douglas-firs (Pseudotsuga menziesii ) were
observed at A˜
no Nuevo, and occasional coast live oaks
(Quercus agrifolia) were observed at both Cambria and
A˜
no Nuevo. In addition, knobcone pine (P. attenuata)is
known to occur in the A˜
no Nuevo population and to
hybridize with Monterey pine, but no knobcone pines
were observed in the vicinity of our specific sampling
site. On average from 1950 to 2000, the Cedros Island
sites were the driest and warmest compared to the other
sites (;105 mm annual precipitation, ;188C annual
mean temperature), A˜
no Nuevo was the wettest and
coolest (841 mm, ;138C), with Cambria and Guadalupe
Island intermediate but closer to A˜
noNuevoin
temperature and closer to Cedros Island in precipitation
(Supplement). Additional characteristics of each sam-
pling site, including soil characteristics and bioclimatic
variables often used in ecological niche modeling (Hij-
mans et al. 2005), can be found in the Supplement. As
variation in environmental characteristics occurs within
each population, it should be noted that these charac-
teristics are representative of our particular sampling
sites. Collections were made from the Mexican popula-
tions (under permit from the Mexican government)
during an expedition on 18–26 May 2006, and from the
California sites in August, 2006 (see Appendix B for
photos of the Mexican sites).
At each of the two California sites, soil was obtained
by collecting 20 cores, each 10 cm in diameter and 20 cm
deep (excluding litter), spaced 30 m apart along two
randomly placed 300-m transects through the closed-
canopy pine forest. On the Mexican islands, the soils
were thin and rocky, and were not conducive to removal
of cylindrical soil cores. Consequently, equivalent
volumes of soil were removed using a hand trowel. In
addition, because tree density at the Mexican sites was
either low (Guadalupe Island) or irregular and clumped
(Cedros Island), at those sites we collected soil samples
within 10 m of specific individual trees. At all five sites,
we collected ;20 pine cones from each of 10 different
trees. At the California sites, those trees were chosen as
the nearest cone-bearing trees adjacent to every second
soil sampling point. In the Mexican populations, the 10
trees were chosen haphazardly at equally spaced
intervals across the populations. Soils and seeds were
imported to the United States under USDA APHIS
permits P526P-06-02719 and 37-86851, respectively.
Cones were immersed in hot water and dried to
extract seeds, and samples of 100 seeds from each cone
were weighed, so that seed mass could be used as a
covariate in analyses to control for potential maternal
effects on seedling growth traits. Soils were dry at the
time of collection, and were stored dry until use in
experiments, at which time the multiple soil samples
from each site were bulked (separately for each site),
homogenized, and sifted through a 5-mm sieve. A subset
of the soil was subjected to gamma irradiation
(Sterigenics, Hayward, California, USA) to reduce
densities of mycorrhizal fungal propagules and other
microbes for use in sterile controls. Subsamples of the
soil from each of the five populations were passed
through a 2-mm sieve and submitted to the University of
California, Davis Analytical Laboratory for determina-
tion of chemical and physical characteristics including
extractable nutrients, cation exchange capacity, percent-
age organic matter, percentage organic carbon, and
texture. Seeds were surface sterilized with 10%sodium
hypochlorite and cold stratified at 48C for four weeks.
Growth chamber experiment
Experimental units were established by filling cylin-
drical pots (6 cm diameter by 15 cm deep) with a 3:1
mixture of field-collected soil and sterile sand (to
improve drainage), and planting three seeds from the
appropriate tree (i.e., open-pollinated, or OP, seed
family) into each pot. After seed germination, each pot
was thinned to contain only one seedling. We created
different treatments by pairing soil from each of the five
populations with seedlings grown from seeds collected
from three different OP familes at each of the five
populations, resulting in 75 different (soil 3population
3OP family) combinations. Each combination was
replicated five times, for a total of 375 pots (hereafter
referred to as the ‘‘main experiment’’). In addition, three
gamma-irradiated replicates of each OP family with each
soil were established for an additional 225 pots, to
determine plant growth in soils with reduced microbial
densities. Beginning in February 2007, seedlings were
grown in a controlled-environment room at the Univer-
sity of California, Santa Cruz, on a 14:10 h day : night
cycle (20:108C day : night temperature, 225 lmolm
2
s
1
of light at plant height) and watered to capacity twice
weekly with deionized water. Pots were set in racks with
6.5 cm between pots, with pot locations assigned
randomly, and locations were re-randomized once every
4 weeks during the experiment. A 5-mm layer of sterile
sand was added to the surface of the soil in each pot to
minimize splashing of soil and microbes among pots.
Seedling height was measured after 12 weeks of
growth and at the end of the experiment by measuring
needle-bearing stem length, which we have found in
previous work to be tightly correlated with total biomass
for Monterey pine seedlings grown in the laboratory.
After 28 weeks, the experiment was harvested. Fourteen
plants in the main experiment did not survive to harvest
and were not included in any analyses. Soil was washed
from the roots of surviving seedlings on a 2-mm sieve,
and 5–10 root tips colonized by ectomycorrhizal fungi
(as indicated by a lack of root hairs) were randomly
selected from each root system under a dissecting
microscope by choosing the colonized root tips nearest
JASON D. HOEKSEMA ET AL.2276 Ecology, Vol. 93, No. 10

to sampling points on a regular grid. Uncolonized root
tips were rarely observed. Collected colonized root tips
were lyophilized to preserve DNA for molecular
identification of fungi. Ninety percent of seedlings
growing in the gamma-irradiated soils were observed
to be colonized by ectomycorrhizal fungi, and this
contamination did not differ among treatments; but
contamination only occurred on feeder roots in the
bottom 1 cm of each pot, suggesting that it took place
via spores carried in water that splashed into the
drainage holes on the bottom of these pots. Therefore,
when sampling ectomycorrhizal root tips on the main
experiment, we did not sample from any roots growing
in the bottom 5 cm of the pots. Representative root tips
from contaminant ectomycorrhizal root tips were also
saved for molecular identification. Root length of each
plant was estimated using the gridline intercept method
(Newman 1966). Aboveground (shoot) and below-
ground (root) biomass of seedlings were separated,
oven-dried for 48 h at 608C, and weighed. Four plant
performance variables were calculated and used for
analysis: relative growth rate of height over the final 16
weeks (RGR, cmcm
1
d
1
), final total biomass (MASS,
g), final ratio of root to shoot biomass (RSR), and final
specific root length (SRL, m/g of root).
Molecular identification of ectomycorrhizal fungi
We extracted DNA from each ectomycorrhizal root
tip, used PCR with fungal-specific primers to amplify
DNA from the nuclear ITS genomic region of the
ectomycorrhizal fungal symbionts, and used Sanger
sequencing and matching with public databases to
classify similar fungal sequences into operational taxo-
nomic units (OTUs, hereafter ‘‘species’’). Methods were
similar to recent published community studies of
ectomycorrhizal fungi (e.g., Izzo et al. 2006), and details
can be found in Appendix A.
Data analysis
Species richness of the RPC ectomycorrhizal fungi.—
All observed species (including ‘‘singlets,’’ i.e., species
observed only once) were included in analyses to
estimate species richness and diversity in each of the
five soils studied. We generated estimates of fungal
species richness in each soil (across all tree genotypes) by
conducting separate analyses (using EstimateS 8.2
software; Colwell 2009) for each of the five soils across
all of the replicate seedlings grown in each of those soils.
In each of these five separate analyses, the Mao Tau
rarefaction function was used to create sample-based
species accumulation curves, rescaled as a function of
the number of ectomycorrhizal root tips identified (as
recommended by Gotelli and Colwell 2001). Because
none of the species accumulation curves reached an
asymptote, we estimated asymptotic or total species
richness in each analysis by calculating a nonparametric
total richness estimator, ICE (the incidence-based
coverage estimator), and by functional extrapolation
of the Mao Tau rarefaction curve using the Michaelis-
Menten function (Colwell 2009). For all analyses,
parameters and measures of their uncertainty were
estimated from 50 randomizations of sample order.
Variation in composition of the dominant RPC
ectomycorrhizal fungi.—We tested for variation in
composition of the RPC ectomycorrhizal fungal guild
due to the main and interactive effects of soils and pine
genotypes by conducting a permutation-based multivar-
iate analysis of variance (PERMANOVA) based on the
Bray-Curtis dissimilarity of species relative abundances
among samples. These analyses excluded species that
occurred on fewer than 5%of seedlings, as these rare
taxa are expected to contribute little to multivariate
patterns of community composition (McCune and
Grace 2002). An initial full model for the PERMANO-
VA included soil, pine population, and their interactions
as fixed effects, and included plant family (nested within
plant population) and the interaction of plant family
with soil as random factors. In initial analyses, the
factors other than soil and plant population had
negative estimated variance components and were
eliminated from the final analysis. Pvalues were
calculated from 10 000 permutations, and significance
of all statistical tests was assessed using a¼0.05.
Multivariate analyses were conducted using PRIMER
version 6 (PRIMER-E, Plymouth, UK).
We explored the importance of each of the three
dominant species for the results of the PERMANOVA
analyses by testing for potential local adaptation of
fungal species to particular pine populations using
separate univariate mixed ANOVAs (in SAS PROC
MIXED version 9.2; SAS Institute, Cary, North
Carolina, USA) for each of the three dominant species.
Each univariate analysis on a particular species used only
data from soils in which that species occurred in the data
set. Hence, if a fungal species was never detected in a
particular soil, then all replicates for that soil were
excluded from univariate analyses of that species. In
these analyses, the response variables analyzed were the
relative abundances of the species being analyzed, and
the statistical models had the same fixed effects as the full
model for the PERMANOVA analyses (soil, pine
population, and their interactions). The corresponding
random effects were included only when their variance
components were estimated to be greater than zero. We
tested for patterns of local adaptation of fungal species to
pine populations by following any significant soil by pine
population interactions with planned contrasts compar-
ing the relative abundance of a particular species in
sympatric combinations (i.e., from the same population)
of soils plus tree genotypes vs. allopatric combinations
(i.e., from different populations). Because residuals in
these analyses were all highly non-normal, Pvalues were
obtained with randomization tests (10 000 iterations)
using a macro wrapper in SAS PROC MIXED.
Variation in plant growth traits.—Univariate mixed
ANOVA models identical to those used for the
October 2012 2277GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE

univariate analyses of dominant fungal species were
used to test for variation in plant performance in the
main experiment due to the main and interactive effects
of soils and pine genotypes on each plant performance
variable. Average seed mass for each open-pollinated
family was explored as a covariate to account for
potential effects of maternal environment on seedling
traits, but results of analyses were not qualitatively
different, so seed mass was not used as a factor in the
final analyses presented. When soil by pine population
interactions were significant, we also tested the overall
contrast between sympatric and allopatric combinations
of soils and tree populations. However, for some
measures of plant performance, their qualitative rela-
tionship with plant relative fitness may be context
dependent; thus, overall differences in such variables
between sympatric and allopatric combinations of soils
and plant populations may not be indicative of local
adaptation. For example, higher RGR may be favored
in some environments and lower RGR may be favored
in others; thus, finding consistently higher RGR in
sympatric combinations of soils and plant populations
would not necessarily indicate that plants are locally
adapted to soils. Due to the observation of some
contaminant ectomycorrhizal fungi on the lower 1 cm
of roots of seedlings in the gamma-irradiated soils,
comparisons of seedling growth between the main
experiment and the gamma-irradiated soils would be
difficult to interpret, i.e., they would not represent clear
comparisons of plant growth in sterilized and non-
sterilized soils. Thus, we do not present analyses of plant
performance in the gamma-irradiated soils.
RESULTS
Question 1: Has the species composition of the RPC fungi
diverged among populations of Monterey pine?
Richness and diversity of the RPC ectomycorrhizal
fungi.—Sequencing of the ectomycorrhizal fungi identi-
fied a rich mycorrhizal assemblage of 35 operational
taxonomic units (species) among the 789 samples,
including 2%(17 sequences) that occurred only once in
the experiment (Appendix C: Table C1). The high fungal
diversity found in these samples did not exhaust the full
diversity of fungal species associated with the pines at
any of these sites, after analyzing more than 40 seedlings
and 100 ectomycorrhizal root tips from each soil.
Asymptotic species richness of the RPC ectomycorrhizal
fungi was estimated to be more than twofold higher in
Guadalupe Island soil (ICE ¼26.8 63.9 taxa, mean 6
SD) than in Cambria soil (ICE ¼11.8 66.5), with
intermediate levels of richness in the other three soils
(Cedros Island South, 12.3 62.6; A˜
no Nuevo, 14.3 6
3.1; Cedros Island North, 16.1 65.7), although
uncertainty around these richness estimates was sub-
stantial, with Cambria and Guadalupe Island exhibiting
overlapping 95%confidence intervals on Mao Tau
rarefaction curves (Appendix C: Fig. C1).
Variation in composition of the dominant RPC
ectomycorrhizal fungi.—Fungal diversity showed a
strongly exponential distribution of abundance among
the samples, suggesting that the assemblage was
dominated by a few fungal species (Appendix C: Fig.
C2). Three dominant species (Wilcoxina1, Pyronemata-
ceae1, and Rhizopogon roseolus) occurred each on at
least 5%of experimental seedlings, and thus were used in
multivariate analyses. Among these species, Pyronema-
taceae1 and Rhizopogon roseolus occurred in all five soils
(Wilcoxina1 was never found in Cambria soil) and all
three dominants occurred on all five plant populations
of genotypes (Appendix C: Table C1), but their relative
abundances differed significantly among soils (pseudo-
F
4, 251
¼85.2, P¼0.0001; PERMANOVA) and among
populations of tree genotypes (pseudo-F
4, 251
¼2.3, P¼
0.0238; PERMANOVA). The sharpest break in compo-
sition among soils was between A˜
no Nuevo and the
other four sites (Fig. 1a). Wilcoxina1 dominated the
roots of plants growing in A˜
no Nuevo soil, but was a
minor component or did not appear in the others
(ANOVA, F
3, 185
¼139.8, P,0.0001). In contrast,
Pyronemataceae1 dominated the other four soils
(ANOVA, F
4, 225
¼94.3, P,0.0001), and Rhizopogon
roseolus was the second most abundant taxon in each
soil except Cedros Island South (ANOVA, F
4,49
¼2.02,
P¼0.078).
The only contaminants found in the gamma-irradiat-
ed soils were Rhizopogon occidentalis gr and Pyronema-
taceae1, both of which were also found in the main
experiment. This fact, combined with the observation
that contaminants only colonized the bottom 1 cm of
feeder roots in pots with gamma-irradiated soils,
suggests that contaminants were not from airborne
propagules and did not influence the RPC fungal
community data in the main experiment (collected only
from the upper 10 cm of those pots).
Question 2: Have populations of pines and fungi
diverged genetically in their compatibility or traits,
and are patterns of divergence consistent across different
members of the fungal guild?
Mycorrhizal compatibility traits of pine seedlings.—
Comparisons of relative abundances of the dominant
fungal species on seedlings suggest that the five plant
populations have diverged genetically for their relative
compatibility with two of the dominant species, Wilcox-
ina1 (univariate ANOVA, F
4, 185
¼2.8, P¼0.025) and
Rhizopogon roseolus (univariate ANOVA, F
4,49
¼2.6, P
¼0.0315). Specifically, Wilcoxina1 and Rhizopogon
roseolus were both moderately abundant on the Cal-
ifornia pine genotypes, but on the Mexican pine
genotypes Wilcoxina1 was more common and Rhizopo-
gon roseolus was only rarely encountered (Fig. 1b).
Pyronemataceae1 was similarly abundant across all five
populations of tree genotypes (univariate ANOVA, F
4,10
¼0.68, P¼0.56, average relative abundance ¼0.68 6
0.04, mean 6SE). We found no evidence that any of the
JASON D. HOEKSEMA ET AL.2278 Ecology, Vol. 93, No. 10

three dominant species was locally adapted (or mal-
adapted) to particular plant populations.
Variation in plant growth traits.—Several seedling
growth traits have also diverged among the five native
populations of Pinus radiata. Soils and plant genotypes
interacted to control three plant performance traits:
RGR, MASS, and SRL (Fig. 2). Genetic differences
among plant populations in RGR were strongest in
Guadalupe Island soils in which A˜
no Nuevo and Cedros
Island North genotypes grew the fastest, and were
weakest in Cedros Island South soils in which RGR was
equally low among all plant genotypes (Fig. 2a). RGR
was significantly lower in sympatric combinations of
soils and plant genotypes, mainly driven by the
Cambria, Cedros Island North, and Guadalupe soils in
which sympatric genotypes exhibited low RGRs relative
to other genotypes in those soils.
Patterns of genetic variation among plant populations
in seedling MASS differed among soils (Fig. 2b). This
pattern was largely driven by A˜
no Nuevo genotypes,
which had relatively high mass in California soils and
relatively low mass in Mexican island soils. Overall, the
other four populations of genotypes exhibited relatively
consistent high (Cambria and Guadalupe Islands) or
low (both Cedros Island populations) MASS, and
MASS was typically lower in the Mexican soils than
the California soils. Patterns of genetic variation in SRL
also varied among soils, and SRL was lower (i.e., roots
FIG. 1. Variation in relative abundance (proportion of root tips occupied per seedling, mean 6SE) of three dominant
ectomycorrhizal fungal species (a) among the five soils studied (each sampled across all five genetic populations of plants), and (b)
among the five Monterey pine (Pinus radiata) populations studied (each sampled across all five soils) in mainland California, USA,
and Guadalupe and Cedros Islands, Mexico. OTUs are operational taxonomic units. Pyronemataceae1 [not shown in panel (b)]
was equally abundant across all five pine populations of genotypes (mean relative abundance ¼0.68 60.04, mean 6SE ).
Wilcoxina1: soil, F
3, 185
¼139.8, P,0.0001; plant, F
4, 185
¼2.8, P¼0.025; soil 3plant, F
12, 185
¼0.82, P¼0.64. Rhizopogon roseolus
gr: soil, F
4,49
¼2.02, P¼0.078; plant, F
4,49
¼2.6, P¼0.032; soil 3plant, F
16,49
¼0.85, P¼0.55. Pyronemataceae1: soil, F
4, 225
¼94.3,
P,0.0001; plant, F
4,10
¼0.68, P¼0.56; soil 3plant, F
16, 225
¼0.73, P¼0.77.
October 2012 2279GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE

were coarser) in sympatric combinations of soils and
plant genotypes (Fig. 2c). Overall, SRL was lowest in
A˜
no Nuevo soils and highest in Cedros Island North
soils. Final root : shoot ratio of biomass varied signif-
icantly among plant genotype populations and also
among soils (Fig. 2d). Variation among genotypes was
greater than variation among soils, with Cedros Island
genotypes exhibiting the lowest root : shoot ratios.
DISCUSSION
The results presented here suggest that the five studied
populations of Monterey pine have diverged in the
community composition of their associated resistant-
propagule community (RPC) of ectomycorrhizal (EM)
fungi, and have also diverged genetically in several traits
related to interactions of seedlings with EM fungi,
growth, and biomass allocation. Patterns of genetic
variation among pine populations differed among three
dominant species of EM fungi (Rhizopogon roseolus,
Wilcoxina1, and Pyronemataceae1), suggesting that
Monterey pines can independently evolve in their
compatibility with different ectomycorrhizal symbiont
species.
Geographic variation among native Monterey pine
populations in guild composition of RPC
ectomycorrhizal fungi
Climatic variation and biogeographic processes such
as dispersal and extinction frequently cause geographic
variation in community composition and environmental
variables (Lomolino et al. 2010). Such geographic
divergence in biotic and abiotic context can drive
divergent direct selection on species traits (G 3E
interactions), resulting in local adaptation of a single
species (Williams 1966, Endler 1986), or can modify
interactions between species, resulting in a geographic
mosaic of coevolutionary selection (G 3G3E
interactions) that drives divergence in traits governing
those interactions (Thompson 1994, 2005). We sought
to characterize geographic variation among native
FIG. 2. Variation in Pinus radiata seedling growth traits among soils and plant genotypes from each of five populations: (a)
relative growth rate (RGR) of height during the final 16 weeks of the experiment, (b) final total biomass, (c) specific root length,
and (d) final root : shoot biomass ratio. All data are presented as mean 6SE. In panels (a)–(c), sympatric combinations of plant
genotypes and soils are marked with an uppercase ‘‘S.’’ Note that panels (c) and (d) are on the following page.
JASON D. HOEKSEMA ET AL.2280 Ecology, Vol. 93, No. 10

Monterey pine populations in one component of
community composition—the RPC ectomycorrhizal
fungi, in order to understand the potential for such
biotic variation to drive trait divergence among popu-
lations of pines and their ectomycorrhizal fungi. Our
results suggest that both diversity and composition of
the RPC ectomycorrhizal fungi vary among the five
studied soils, with the A ˜
no Nuevo soil exhibiting
contrasting relative abundances of the dominant fungal
taxa (Fig. 1a). Given that different EM fungal taxa can
have dramatically different ecological effects on their
host plants (Jones and Smith 2004), and that plant
growth and productivity often varies with the diversity
of mycorrhizal fungal taxa (Hoeksema et al. 2010), these
results suggest the potential for divergent selection from
variable RPC ectomycorrhizal fungi on Monterey pine
traits among populations, and/or geographic variation
in coevolutionary dynamics of particular ectomycor-
rhizal interactions. Any divergent selection likely de-
pends on the degree to which the dominant RPC fungal
taxa (Rhizopogon roseolus, Wilcoxina1, and Pyronema-
taceae1) differ significantly in key traits affecting
Monterey pine fitness or affecting each other’s interac-
tions with Monterey pine, and whether the low-diversity
community of RPC fungi found in Cambria soils has a
different suite of average traits that affect pines,
compared to the other communities.
Studies in other ectomycorrhizal systems have also
typically found significant variation in EM fungal
community composition among geographically isolated
populations of the same host plant species (e.g., Kjoller
and Bruns 2003, Walker et al. 2005, Rusca et al. 2006,
Moser et al. 2009), perhaps in part due to limited
dispersal among populations. Rusca et al. (2006) and
Kjoller and Bruns (2003) used bioassay methods similar
to those used in this study, but focused on variation in
relative abundances of Rhizopogon species in the RPC
ectomycorrhizal fungal guilds associated with pines in
California. They studied a series of geographically
isolated populations of bishop pine (Pinus muricata),
FIG. 2. Continued.
October 2012 2281GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE

which is a sister species of Monterey pine and
ecologically similar to it (Axelrod 1980), along a
latitudinal gradient overlapping in the north with the
one examined in this study. From their site in Monterey,
California, which is located geographically between our
A˜
no Nuevo and Cambria sites, they retrieved several
Rhizopogon species, the most abundant of which were R.
occidentalis and a member of the Amylopogon subgenus
(clade I); the soils from their Santa Barbara site (;120
km southeast of our Cambria site) were dominated by R.
brunsii (called Amylopogon clade II at the time;
Grubisha et al. 2005). In contrast, R. roseolus, which
was detected by Kjoller and Bruns on the coast only at
Point Reyes (;125 km north-northwest of our A˜
no
Nuevo site) but was frequently detected in pine-
dominated sites in the Sierra Nevada mountains (Izzo
et al. 2006, Rusca et al. 2006), was the most frequent
Rhizopogon species in all of our soils. R. brunsii and R.
occidentalis were detected only in small numbers in our
study (Appendix C: Fig. C2, Table C1), although our
observation of their occurrence in soils from Cedros
Island significantly extends the southern limit of the
known range of these widespread taxa. Our root tip
sampling method (collecting tips nearest to points on a
fixed grid) may have underestimated relative abundances
of taxa such as Rhizopogon with clumped mycorrhizal
morphology, and a better understanding of the bioge-
ography of Rhizopogon species among our study sites
would be facilitated by further studies designed to
enhance detection of Rhizopogon, including significant
dilution of soils for bioassays in a neutral soil medium
(as in Kjoller and Bruns 2003), which could also help
remove potentially confounding effects of differences
among soils in physicochemical factors and other
microbes. In general, deeper sampling within the soils
from each of our sites would likely reveal additional rare
species at each site, and molecular studies of available
propagules in the soil would help paint a more complete
picture of these communities.
In comparison to other laboratory bioassay studies of
RPC ectomycorrhizal fungi, taxonomic composition
and diversity in our five Monterey pine soils were most
similar to those associated with bristlecone pine (Pinus
longaeva) in the White Mountains of California (Bidar-
tondo et al. 2001). In both the bristlecone and the
Monterey pine study systems, which occur in relatively
arid climates, the guilds of RPC fungi were dominated
by three taxa: an unidentified species in the Pyronema-
taceae (different species in the two studies), a Wilcoxina
species (both very similar to W. rehmii ), and a
Rhizopogon species in the subgenus Roseoli (Grubisha
et al. 2002). The unidentified Pyronemataceous species
was the most abundant in the bristlecone pine soils and
in soils from all of our sites except the least arid one,
A˜
no Nuevo (Fig. 1a). Moreover, the observed species
richness in each of our five soils was relatively low
(Appendix C: Fig. C1), comparable to that found by
Bidartondo et al. (2001), and lower than has been
typically found in studies of bishop pine soils farther
north along the California coast (e.g., Taylor and Bruns
1999). Perhaps this guild structure of the RPC ectomy-
corrhizal fungi will prove to be typical of pine-
dominated habitats in arid climates. Future analyses of
the RPC fungal guild structure in this system will be
enhanced by consideration of phylogeographic structure
of the fungal taxa, as the fungal species identified here
likely contain multiple phylogenetic lineages.
Geographic divergence in growth and mycorrhizal traits
of Monterey pine
The five studied populations of Monterey pine have
diverged genetically in most of the traits we examined,
including growth and biomass allocation traits (Fig. 2)
and mycorrhizal traits, i.e., compatibility with particular
taxa of ectomycorrhizal fungi (Fig. 1b). Moreover, the
results for mycorrhizal traits demonstrate that patterns
of variation in Monterey pine compatibility with its
ectomycorrhizal symbionts differ among fungal taxa,
i.e., mycorrhizal colonization of Monterey pine seedlings
depends on a G
P
3G
F
interaction, where G
F
represents
different fungal taxa. Specifically, the Mexican pine
populations have evolved higher compatibility with
Wilcoxina1 and lower compatibility with Rhizopogon
roseolus, compared to California populations, and the
populations have not diverged with respect to compat-
ibility with Pyronemataceae1 (Fig. 1b). In a prior study,
Hoeksema and Thompson (2007) found yet a different
pattern of divergence among the A˜
no Nuevo, Guada-
lupe Island, and Cedros Island South populations for
their compatibility with another RPC ectomycorrhizal
fungus, Rhizopogon occidentalis. Specifically, the Cedros
Island South population of pines (see Plate 1) seems to
have evolved significantly lower compatibility with R.
occidentalis, compared to the A˜
no Nuevo and Guada-
lupe Island populations (as well as the Monterey,
California population, which was not included in the
current study).
The conclusion that Monterey pines can evolve
unique patterns of compatibility with different ectomy-
corrhizal symbiont species has significant implications
for how diverse mutualisms may coevolve, or for how
diverse guilds of species may exert selection on a focal
species (if the interaction is not coevolving). Different
fungal taxa, rather than exerting conflicting selection on
the same plant trait (e.g., overall plant compatibility
with mycorrhizal fungi), could have selective effects on
different plant traits, such as specific signaling and
recognition loci (Hoeksema 2010). A high degree of
specificity in such traits could allow coevolutionary
cycling to drive fluctuating polymorphisms at corre-
sponding recognition loci in plants and fungi, as has
been found for the major histocompatibility complex
(MHC) loci in vertebrates (Hughes and Nei 1988) and
self-incompatibility loci in fungi (May and Matzke
1995). Selection should strongly favor recognition by
plants of more beneficial vs. less beneficial taxa of
JASON D. HOEKSEMA ET AL.2282 Ecology, Vol. 93, No. 10

ectomycorrhizal fungi, and subsequent exclusion or
rejection of the less beneficial taxa (Kiers and van der
Heijden 2006, Bever et al. 2009). In turn, less beneficial
fungal taxa would be under selection to avoid recogni-
tion, which would drive further reciprocal evolution in
the recognition genes of the plant.
Divergence among Monterey pine populations in
plant growth and allocation traits was also substantial,
but differences among genotypes also interacted with
soils for all traits but root : shoot (Fig. 2), and
geographic patterns of seedling growth traits were more
correlated than were the mycorrhizal traits. Overall,
Cedros Island plants allocated relatively less biomass to
roots and made finer roots (i.e., higher SRL) compared
to the other three populations, although in some soils
the A˜
no Nuevo genotypes also exhibited relatively fine
roots. Similarly, mass was typically lowest for the
Cedros Island genotypes, but genetic variation in RGR
during the final 16 weeks depended heavily on soils.
These overall patterns are qualitatively similar to those
found in a previous study of variation in seedling growth
traits among native Monterey pine populations (Hoek-
sema and Thompson 2007). Although greater relative
allocation to aboveground biomass is often interpreted
as a response to light limitation in plants (Tilman 1988),
in native Monterey pine populations fog moisture is also
(in addition to light) an important ‘‘aboveground’’
resource (Rogers et al. 2005 ), and lower root : shoot
ratios of Cedros Island seedlings may reflect adaptation
for capturing fog moisture on needles in that extremely
arid climate. The observations that RGR and SRL were
both lower in sympatric combinations of pine genotypes
and soils compared to allopatric combinations suggests
the possibility of overall local adaptation or maladap-
tation of those traits to one or more soil characteristics,
e.g., chemical properties or composition of pathogenic
or ectomycorrhizal fungi. However, this idea depends on
the assumption that plant fitness has the same relation-
ship with those two plant traits in all five populations,
which may not be the case. For example, higher seedling
RGR may be less advantageous for survival and
reproduction in a drier climate than in a wetter climate.
It should be kept in mind that our estimates of plant
traits and fungal abundances would likely vary across
alternative light, moisture, and temperature regimes
compared to the conditions used in our experiment. In
addition, the presence of contaminant ectomycorrhizal
fungi on the lower 1 cm of feeder roots of some plants in
our experiment may have reduced the influence of
different soils on plant trait estimates.
CONCLUSIONS
The emerging story on evolution of interactions
between pines and their associated resistant-propagule
community of ectomycorrhizal fungi is that they are
highly variable geographically; community composi-
tion of the RPC fungi, soil and climatic factors, and
compatibility of pine seedlings with particular RPC
taxa have all diverged among communities. Moreover,
pine compatibility with particular RPC fungi appears
to evolve as a set of relatively independent traits. The
challenge now is to develop clear tests of hypotheses
PLATE 1. Groves of Pinus radiata in the Cedros Island South population (Mexico). The Cedros Island Pinus radiata habitats are
significantly drier than Guadalupe Island and sites in California, receiving just over 100 mm of precipitation each year on average.
Photo credit: J. D. Hoeksema.
October 2012 2283GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE

on the specific selective sources and other evolutionary
forces that may be driving these divergences. In
particular, if pine populations are diverging indepen-
dently in their compatibility with different ectomycor-
rhizal symbiont taxa, are fungal traits acting as
sources of selection driving this divergence? If so,
what are the fungal traits in question and is their
selection on plant traits modulated by environmental
factors? Answering these questions will lend insight
into how coevolution proceeds in widespread species-
rich interactions.
ACKNOWLEDGMENTS
This work was supported by a Collaborative Grant (Award
20050865) from the University of California Institute for
Mexico and the United States (UC MEXUS) and the Consejo
Nacional de Ciencia y Tecnologı
´a (CONACYT). J. D. Hoek-
sema was supported by a grant from the National Science
Foundation (DEB 0625120), the Department of Biology at the
University of Mississippi, and the National Evolutionary
Synthesis Center, which receives support from a National
Science Foundation grant (EF-0423641), Duke University, the
University of North Carolina–Chapel Hill, and North Carolina
State University. During the expedition to Mexico, a great deal
of logistical support and field assistance was provided by
Alfonso Aguirre and Grupo de Ecologı
´a y Conservacio
´nde
Islas, A.C. In addition, substantial logistical support and
lodging on Cedros Island was provided by the abalone and
lobster fishermen cooperative of Cedros Island ‘‘Pescadores
Nacionales de Abulo
´n.’’ We are grateful to Christopher
Schwind for assistance in all phases of the experimental work
at UCSC, and to the members of the Thompson laboratory at
UCSC, especially Catherine Fernandez, for logistical support
and conceptual input on the project.
LITERATURE CITED
Ashkannejhad, S., and T. R. Horton. 2006. Ectomycorrhizal
ecology under primary succession on coastal sand dunes:
interactions involving Pinus contorta, suilloid fungi and deer.
New Phytologist 169:345–354.
Axelrod, D. I. 1980. History of the maritime closed-cone pines,
Alta and Baja California. University of California Press,
Berkeley, California, USA.
Baar, J., T. R. Horton, A. M. Kretzer, and T. D. Bruns. 1999.
Mycorrhizal colonization of Pinus muricata from resistant
propagules after a stand-replacing wildfire. New Phytologist
143:409–418.
Bever, J. D., S. C. Richardson, B. M. Lawrence, J. Holmes, and
M. Watson. 2009. Preferential allocation to beneficial
symbiont with spatial structure maintains mycorrhizal
mutualism. Ecology Letters 12:13–21.
Bidartondo, M. I., J. Baar, and T. D. Bruns. 2001. Low
ectomycorrhizal inoculum potential and diversity from soils
in and near ancient forests of bristlecone pine (Pinus
longaeva). Canadian Journal of Botany 79:293–299.
Colwell, R. K. 2009. EstimateS: statistical estimation of species
richness and shared species from samples. Version 8.2. User’s
guide and application. http://purl.oclc.org/estimates
Endler, J. A. 1986. Natural selection in the wild. Princeton
University Press, Princeton, New Jersey, USA.
Gardes, M., and T. D. Bruns. 1996. Community structure of
ectomycorrhizal fungi in a Pinus muricata forest: above- and
below-ground views. Canadian Journal of Botany 74:1572–
1583.
Gotelli, N. J., and R. K. Colwell. 2001. Quantifying biodiver-
sity: procedures and pitfalls in the measurement and
comparison of species richness. Ecology Letters 4:379–391.
Grogan, P., J. Baar, and T. D. Bruns. 2000. Below-ground
ectomycorrhizal community structure in a recently burned
bishop pine forest. Journal of Ecology 88:1051–1062.
Grubisha, L. C., J. M. Trappe, A. R. Beyerle, and D. Wheeler.
2005. NATS truffle and truffle-like fungi 12: Rhizopogon ater
sp nov and R. brunsii sp nov (Rhizopogonaceae, Basidiomy-
cota). Mycotaxon 93:345–353.
Grubisha, L. C., J. M. Trappe, R. Molina, and J. W. Spatafora.
2002. Biology of the ectomycorrhizal genus Rhizopogon. VI.
Re-examination of infrageneric relationships inferred from
phylogenetic analyses of ITS sequences. Mycologia 94:607–
619.
Guimaraes, P. R., P. Jordano, and J. N. Thompson. 2011.
Evolution and coevolution within mutualistic networks.
Ecology Letters 14:877–885.
Hereford, J. 2009. A quantitative survey of local adaptation
and fitness trade-offs. American Naturalist 173:579–588.
Hijmans, R. J., S. E. Cameron, J. L. Parra, P. G. Jones, and A.
Jarvis. 2005. Very high resolution interpolated climate
surfaces for global land areas. International Journal of
Climatology 25:1965–1978.
Hoeksema, J. D. 2010. Tansley review. Ongoing coevolution in
mycorrhizal interactions. New Phytologist 187:286–300.
Hoeksema, J. D., et al. 2010. A meta-analysis of context-
dependency in plant response to inoculation with mycorrhi-
zal fungi. Ecology Letters 13:394–407.
Hoeksema, J. D., B. Piculell, and J. N. Thompson. 2009.
Within-population genetic variability in mycorrhizal interac-
tions. Communicative and Integrative Biology 2:110–112.
Hoeksema, J. D., and J. N. Thompson. 2007. Geographic
structure in a widespread plant–mycorrhizal interaction:
pines and false truffles. Journal of Evolutionary Biology
20:1148–1163.
Horton, T. R., and T. D. Bruns. 1998. Multiple-host fungi are
the most frequent and abundant ectomycorrhizal types in a
mixed stand of Douglas fir (Pseudotsuga menziesii ) and
bishop pine (Pinus muricata). New Phytologist 139:331–339.
Horton, T. R., and T. D. Bruns. 2001. The molecular revolution
in ectomycorrhizal ecology: peeking into the black box.
Molecular Ecology 10:1855–1871.
Hughes, A. L., and M. Nei. 1988. Pattern of nucleotide
substitution at major histocompatibility complex class-I loci
reveals overdominant selection. Nature 335:167–170.
Izzo, A., D. T. Nguyen, and T. D. Bruns. 2006. Spatial
structure and richness of ectomycorrhizal fungi colonizing
bioassay seedlings from resistant propagules in a Sierra
Nevada forest: comparisons using two hosts that exhibit
different seedling establishment patterns. Mycologia 98:374–
383.
Jones, M. D., and S. E. Smith. 2004. Exploring functional
definitions of mycorrhizas: Are mycorrhizas always mutual-
isms? Canadian Journal of Botany 82:1089–1109.
Kiers, E. T., and M. G. A. van der Heijden. 2006. Mutualistic
stability in the arbuscular mycorrhizal symbiosis: exploring
hypotheses of evolutionary cooperation. Ecology 87:1627–
1636.
Kjoller, R., and T. D. Bruns. 2003. Rhizopogon spore bank
communities within and among California pine forests.
Mycologia 95:603–613.
Lomolino, M. V., B. R. Riddle, R. J. Whittaker, and J. H.
Brown. 2010. Biogeography. Fourth edition. Sinauer Asso-
ciates, Sunderland, Massachusetts, USA.
May, G., and E. Matzke. 1995. Recombination and variation at
the A mating-type of Coprinus cinereus. Molecular Biology
and Evolution 12:794–802.
McCune, B., and J. B. Grace. 2002. Analysis of ecological
communities. MjM Software Design, Gleneden Beach,
Oregon, USA.
Moser, A. M., J. L. Frank, J. A. D’Allura, and D. Southworth.
2009. Ectomycorrhizal communities of Quercus garryana are
JASON D. HOEKSEMA ET AL.2284 Ecology, Vol. 93, No. 10

similar on serpentine and nonserpentine soils. Plant and Soil
315:185–194.
Newman, E. J. 1966. A method of estimating the total root
length of a root in a sample. Journal of Applied Ecology
3:139–145.
Piculell, B. J., J. D. Hoeksema, and J. N. Thompson. 2008.
Interactions of biotic and abiotic environmental factors in an
ectomycorrhizal symbiosis, and the potential for selection
mosaics. BMC Biology 6:23.
Rogers, D. L., J. J. V. Hernandez, A. C. Matheson, and J. J. G.
Santos. 2005. Conserving the pines of Guadalupe and Cedros
Islands, Mexico: an international collaboration. Pages 31–53
in A. Romero and S. West, editors. Environmental issues in
Latin America and the Caribbean. Springer-Verlag, Berlin,
Germany.
Rusca, T. A., P. G. Kennedy, and T. D. Bruns. 2006. The effect
of different pine hosts on the sampling of Rhizopogon spore
banks in five eastern Sierra Nevada forests. New Phytologist
170:551–560.
Schluter, D. 2009. Evidence for ecological speciation and its
alternative. Science 323:737–741.
Siepielski, A. M., and C. W. Benkman. 2007. Convergent
patterns in the selection mosaic for two North American
bird-dispersed pines. Ecological Monographs 77:203–220.
Taylor, D. L., and T. D. Bruns. 1999. Community structure of
ectomycorrhizal fungi in a Pinus muricata forest: minimal
overlap between the mature forest and resistant propagule
communities. Molecular Ecology 8:1837–1850.
Thompson, J. N. 1994. The coevolutionary process. University
of Chicago Press, Chicago, Illinois, USA.
Thompson, J. N. 2005. The geographic mosaic of coevolution.
University of Chicago Press, Chicago, Illinois, USA.
Thompson, J. N., and B. M. Cunningham. 2002. Geographic
structure and dynamics of coevolutionary selection. Nature
417:735–738.
Thompson, J. N., A. L. Laine, and J. F. Thompson. 2010.
Retention of mutualism in a geographically diverging
interaction. Ecology Letters 13:1368–1377.
Tilman, D. 1988. Plant strategies and the dynamics and
structure of plant communities. Princeton University Press,
Princeton, New Jersey, USA.
Walker, J. F., O. K. Miller, and J. L. Horton. 2005. Hyper-
diversity of ectomycorrhizal fungus assemblages on oak
seedlings in mixed forests in the southern Appalachian
Mountains. Molecular Ecology 14:829–838.
Williams, G. C. 1966. Adaptation and natural selection.
Princeton University Press, Princeton, New Jersey, USA.
SUPPLEMENTAL MATERIAL
Appendix A
Details of methods used for molecular identification of ectomycorrhizal fungi (Ecological Archives E093-212-A1).
Appendix B
Photographs of native Pinus radiata populations on Guadalupe and Cedros Islands, which are located in the Pacific Ocean off
the west coast of Baja California, Mexico (Ecological Archives E093-212-A2).
Appendix C
Figures showing a sample-based species accumulation curve for ectomycorrhizal fungal OTUs in each of the five soils and the
frequency of ectomycorrhizal fungal OTUs on seedlings across the entire study, and a table showing the frequencies of each fungal
OTU in each of the five soils along with Genbank best matches and accession numbers for each fungal OTU (Ecological Archives
E093-212-A3).
Supplement
Soil and climate data for each of the five sampled native populations of Pinus radiata (Ecological Archives E093-212-S1).
October 2012 2285GEOGRAPHIC DIVERGENCE IN ECTOMYCORRHIZAE