Niche Evolution and Adaptive Radiation: Testing the Order of Trait Divergence

Author(s): D. D. Ackerly, D. W. Schwilk and C. O. Webb

Reviewed work(s):
Source: Ecology, Vol. 87, No. 7, Supplement: Phylogenetic Approaches to Community Ecology
(Jul., 2006), pp. S50-S61
Published by: Ecological Society of America
Stable URL: http://www.jstor.org/stable/20069156 .
Accessed: 23/10/2012 18:12

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.

http://www.jstor.org
Ecology, 87(7) Supplement, 2006, pp. S50-S61
? 2006 by the Ecological Society of America

NICHE EVOLUTION AND ADAPTIVE RADIATION:

TESTING THE ORDER OF TRAIT DIVERGENCE
D. D. Ackerly,1'4 D. W. Schwilk,2 and C. O. Webb3'5

1
Department of Integrative Biology, University of California, Berkeley, California 94720 USA
2U.S. Geological Survey/WERC, Sequoia and Kings Canyon Field Station, Three Rivers, California 93271 USA
3Department of Ecology and Evolution, Yale University, New Haven, Connecticut 06511 USA

Abstract. In the course of an adaptive radiation, the evolution of niche parameters is of

particular interest for understanding modes of speciation and the consequences for coexistence
of related species within communities. We pose a general question: In the course of an

evolutionary radiation, do traits related to within-community niche differences (ot niche)

evolve before or after differentiation affinity or climatic
of macrohabitat tolerances (? niche)?
Here we introduce a new test to address
this question, based on a modification of the method
of independent contrasts. The divergence order test (DOT) is based on the average age of the
nodes on a tree, weighted by the absolute magnitude of the contrast at each node for a

particular trait. The comparison of these weighted averages reveals whether large divergences
for one trait have occurred earlier or later in the course of diversification, relative to a second

trait; significance is determined by bootstrapping from maximum-likelihood ancestral state

reconstructions. The method is applied to the evolution of Ceanothus, a woody plant group in
California, in which co-occurring species exhibit significant differences in a key leaf trait

(specific leaf area) associated with contrasting physiological and life history strategies. Co
occurring differ more for this trait than expected under a null model of community
species
assembly. This oc niche difference evolved early in the divergence of two major subclades
within Ceanothus, whereas climatic distributions (? niche traits) diversified later within each of
the subclades. However, rapid evolution of climate
parameters makes inferences of early

divergence events highly uncertain, and differentiation of the ? niche might have taken place

throughout the evolution of the group, without leaving a clear phylogenetic signal. Similar

patterns observed in several plant and animal groups suggest that early divergence of ot niche
traits might be a common feature of niche evolution in many adaptive radiations.

Key words: adaptive radiation; Ceanothus; Cerastes; Coast Range; community assembly; Euceanothus;
habitat, niche conservatism; phylogenetic comparative methods; specific leaf area; Sierra Nevada; trait
divergence; Transverse Range.

Introduction of species distributions. At large spatial scales, species

can occupy different macrohabitats or climatic enve
Ecologists have long considered niche differences
lopes; the resulting distributions will be largely allo
among species to be essential for species coexistence
patric, or, if they do overlap geographically, individuals
(Chesson 2000, Chase and Leibold 2003; but see of the two species would encounter one another
rarely
Hubbell [2001]). The evolution of niche differences due to habitat differentiation. At smaller related
scales,
among closely related species has received particular
species that co-occur in local communities usually
attention. Because close relatives tend to be ecologically exhibit spatial or temporal differentiation in micro
similar in many respects (Darwin 1859, Felsenstein resource or other factors. It is at this
habitat, use, diet,
1985, Harvey and Pagel 1991,Webb et al. 2002), those local scale, where the balance of intra- and interspecific
features that do diverge during speciation will provide interactions influences coexistence and community
important insights into ecological differentiation and structure, that the niche concept has played the most
consequences for coexistence of closely related species. role. Following Pickett and Bazzaz
important (1978)
It is useful in this context to distinguish two scales of and Silvertown et al. (2006), we employ the term ot niche
niche differentiation, corresponding to different scales to describe these small-scale of the niche
components
that differ among co-occurring species, corresponding

Manuscript received 21 January 2005; revised 9 August 2005;

toWhittaker's (1975) use of otdiversity for diversity of
11 August 2005. Editor: A. A. local communities. In contrast, the ? niche is defined as
accepted Corresponding
Agrawal. For reprints of this Special Issue, see footnote 1, p. macrohabitat and climate factors related to larger scale
SI.
4 distributions, corresponding to the ? component of
E-mail: [email protected]
5 habitats in a In this
Present affiliation: Arnold Arboretum of Harvard Uni diversity among landscape. paper,

versity. we do not distinguish the proposed ? and y niche, which

S50
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S51

refers to distributions at the scale of habitat vs. Recent developments of phylogenetic methodology
range et al. 2006), as these have offer outstanding opportunities to reevaluate these
geographic (Silvertown
in terms of interactions classic questions. Here, we present a comparative
equivalent implications species
at the community level. approach to the problem of niche evolution by
It has been that species within local commun introducing a new comparative method designed to test
argued
ities tend to be phylogenetically overdispersed; i.e., the relative timing of divergence in two ecological traits

closely related species co-occur less than expected, (e.g., ot vs. ? niche axes); we then use the test to evaluate

relative to an appropriate null model (Elton 1946, the sequence of trait divergence in the radiation of the

Williams 1964, Gotelli and Graves 1996, Cavender woody plant group Ceanothus in California. Our results

Bares et al. This would that ot indicate that ot niche traits related to local scale
2004). pattern suggest
niche are conserved coexistence diverge first in the radiation of this group,
parameters evolutionarily and/or ?
niche are such that close a pattern that is shared with several other plant and
parameters highly divergent,
relatives tend to occupy different macrohabitats and animal radiations.

hence different communities (Elton 1946,Williams 1964,

Divergence Order Test
Gotelli and Graves 1996, Cavender-Bares et al. 2004).
The alternative, if ot niche traits were more labile, would The divergence order test (DOT) was designed to
facilitate coexistence and resource use in address these questions of niche evolution, as well as
divergent
within local communities. of other questions regarding the relative sequence of
sibling species Divergence
macrohabitat taxa is compat diversification events in a clade. The test examines the
parameters among sibling
ible with the allopatric model of speciation, as disjunct relative timing of evolutionary divergence for two

in a heterogeneous are to continuous characters, and it is based on a modification

populations landscape likely
encounter distinct habitats (Graham et al. 2004; but see of the method of phylogenetic independent contrasts
Wiens [2004]). Differentiation of the a niche would then (Felsenstein 1985). Independent contrasts transform the

a later of evolutionary either trait data for N species into a set of N- 1 contrasts, each
represent stage divergence,
from or directly coexistence based on the difference between trait values across a
resulting promoting species
as the ranges of the now distinct into phylogenetic divergence. Under the assumption that the
species expand
each other's trait evolved independently at each divergence, the
territory.
In an important paper, Diamond contrasts a robust basis on which to test
(1986) argued for provide
this "habitat-first" model of based on his hypotheses of correlated evolution, addressing the
speciation,
observations that closely related bird species in New underlying historical pattern of trait evolution as well
Guinea tend to be allopatric and occupy distinct as better meeting the of standard para
assumptions
macrohabitats elevational or climatic metric statistics (Garland et al. 1992). For the DOT, we
along gradients
(see review by Schl?ter [2000]). In addition, Diamond modify this method and use the absolute differences
and Schl?ter both argued that the habitat-first speci between related nodes derived from maximum-like

ation model can be extended to the analysis of adaptive lihood estimates of ancestral trait reconstructions,

radiations, based on a parsimonious assumption that obtained using ANCML (Schl?ter et al. 1997). This
rates of evolution do not In other approach allows us to incorporate the uncertainty of
dramatically change.
if habitat the first of reconstructions in deeper nodes.
words, divergence represents stage
close then it would also be The divergence order test is based on two sets of
speciation among relatives,
characteristic of early speciation events at the base of an numbers: (1) the absolute value of the unstandardized

radiation. As a corollary, if differentiation of ot contrasts for trait / across the nodes (k= 1, 2,..., N) of a
adaptive
niche occurs late in speciation, or is observed among phylogeny (Cik), which measures the magnitude of the
distant then it would be characteristic of the divergence that occurred at each node regardless of the
relatives,
later of adaptive radiation. Streelman and Danley direction of change; and (2) the age of each node (Ak).
stage
a related model, that vertebrate We then calculate a weighted mean age of divergence for
(2003) presented arguing
radiations follow a trajectory each trait as follows:
adaptive of divergence

along three axes: habitat, trophic morphology, and n

communication, usually in that order. However, their
use of "habitat" refers more to microhabitats within a
=
Wi ?N- (i)
community (e.g., benthic vs. limnetic sticklebacks),
rather than Diamond's larger scale differentiation
k=\
among elevational bands and different forest types
occupied by New Guinea birds. This ambiguity over The result is an average age, in units of time, that
the use of the word habitat is unfortunate, as there is a indicates whether the large divergences in a trait tended
substantive difference between these models in their to occur early or late in the diversification of a group

emphasis on large-scale habitat differences, implying 1). Note that this age will not
(Fig. generally correspond
allopatric populations^ vs. microhabitat differentiation to the age of any single divergence; it is simply a
within local communities. statistical measure of the tendency toward or late
early
 D.S52
D. ACKERLY ET AL. Issue
Ecology Special

Trait A O the absolute value of standardized independent contrasts

+ 1 and the standard deviation of those contrasts (the square

root of subtending branch lengths) should be non
1
significant (Garland et al. 1992). A negative correlation
would indicate larger contrasts than
expected on rapid

O bifurcations, and vice versa a positive

for correlation.
The absolute values of standardized contrasts should
41
also fit a half-normal distribution, and this can be
1
checked visually using truncated normal probability
plots. In addition, if several distinct clades are present
(as in the Ceanothus case), homogeneity of evolutionary
TraitB rates can be tested a nonparametric of
using comparison
standardized contrasts between groups (Garland 1992),
t> or a recently introduced maximum-likelihood
approach
(O'Meara et al. 2005). In general, methods based on

independent contrasts are fairly robust to violations of

Brownian motion (Diaz-Uriarte and Garland 1996,
?7 Ackerly 2000), but this has not been evaluated for the
calculation of standard errors by ANCML or the DOT
analysis.
It can be useful to examine correlations between the
two traits under consideration, though DOT does not

require that the traits exhibit any particular pattern of

correlated or independent If
evolutionary change.
changes in the two traits are tightly linked, then DOT
will certainly not be significant, as the contrasts will be
similar in magnitude at each node. However, differences
in the magnitude of a few basal or distal nodes could
Node age (Ma) result in a significant DOT outcome, and trait evolution
could still be correlated overall on the tree.
Fig. 1.
Example of the divergence order test (DOT). Two
We have explored several approaches to significance
patterns of trait divergence are illustrated on a simple
phylogeny. Numbers at the tips of the phylogeny indicate two testing of the D statistic (see Appendix A). We present
trait states, 0 or 1.Numbers at interior nodes (in italics) here our preferred method, based on a
possible bootstrapping
show the contrast for that node. Trait A (open circles) exhibits a to obtain confidence intervals for the two
approach,
pattern of late divergence, whereas Trait B (shaded circles) estimates of W and their difference, D. The rationale for
exhibits a pattern of early divergence. The lower panel plots
vs. age and shows the calculated this approach is that comparative methods, particularly
contrast magnitude weighted
= 1Ma tend to underestimate the mag
divergence age for Trait A (WA [i.e., 1million years independent contrasts,
=
ago]) and Trait B (WB 2 Ma). nitude of older divergences for rapidly evolving traits.
As a simple
example, consider a bifurcating tree with

divergence for the trait in question. The DOT is then four at the tips.
species If each pair of sister taxa has
based on the comparison of the weighted divergence age divergent trait values, trait evolution,
reflecting rapid
for two traits (D = Wt ? WJ) to determine whether one then the averages for their respective common ancestors
trait diverged significantly earlier than the other, on could be virtually identical and the basal contrast will be
average. nearly or exactly zero (e.g., 1, Trait A). However,
Fig.
The DOT statistic is derived from contrasts between given the rapid rate of evolution for this trait, it is also
ancestral states and does depend on the accuracy of the possible that a large divergence occurred at the first
ancestral estimates themselves (Oakley and Cunningham node, followed by reversals at the subsequent nodes

2000). We do not consider the traits of outgroups in resulting in convergence among extant taxa. Maximum

calculating ancestral states, as these are only necessary likelihood estimates of ancestral states allow for this
to identify trends in trait evolution within a group, and possibility by placing confidence limits on the ancestors
not the magnitude of divergences. The maximum (Schl?ter et al. 1997). If a trait evolves rapidly, then the
likelihood algorithm of ANCML assumes that the confidence limits at deeper nodes will be large (see
pattern of trait evolution fits a model of Brownian Appendix A, Fig. Al).
motion. Global squared-change parsimony, which is We use ANCML (Schl?ter et al. 1997) to generate
also based on Brownian motion, provides the same maximum-likelihood estimates and confidence limits of
ancestral estimates, but no confidence limits. The fit to ancestral states at each node; we then create bootstrap
Brownian motion can be tested in several ways. First, distributions of the potential magnitude of each
using Felsenstein's (1985) algorithm, the correlation of divergence event (see Appendix A, Fig. Al). Hypo
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S53

thetical ancestral values are sampled from the distribu Despite these consistent differences in drought toler

tion for each trait at each node, and from each ance and fire response, both clades are widespread in the
sample
we calculate the corresponding values of Cik,
Cjk, Wh
California Floristic Province, inhabiting chaparral,
and D. We then examine the distribution of D values semiarid forests, and oak woodland. Often, species pairs
Wj,
to determine whether D = 0 falls outside the 95% representing one species from each of the subgroups co

confidence limits on the mean, significance of occur, and it has been suggested that differences in fire
indicating
the observed values at P < 0.05. The calculation and response and/or tolerance of water stress facilitate this

significance of the DOT method were implemented on coexistence (Keeley 1977,Keeley and Zedler 1978,Davis
Mac OS X, using awk scripts, R (R-project 2004), and et al.
1999). These patterns suggest that the basal

ANCML (see Supplement). divergence between.the two clades involved ot niche traits
related to drought tolerance or postfire regeneration
Branch lengths and node ages are represented
strategies. In contrast, both clades by
the objective of this test is to calculate relative
As species throughout California, more recent
suggesting

timing of divergence events, the analysis should ideally differentiation of the climatic niche envelope, represent
be conducted with ultrametric, calibrated phylogenies ing the ? niche (Knight and Ackerly 2001). Here we
based on a molecular or rate-smoothed branches undertake four new analyses to quantify and test these
clock,
if the branch violate a molecular clock observations. (1) We have assembled a co-occurrence
lengths (Sander
son Methods for relative are data set from the literature, and we use null models to
2002). obtaining ages
there can still be considerable test for nonrandom patterns of co-occurrence between
improving, although
due to rates of molecular species from the two subgroups. (2) We combine the co
uncertainty heterogeneous
evolution and difficulty in establishing calibration points occurrence data with a trait data set to test whether co

across different clades (Sanderson 2002, 2003). The occurring species differ significantly for traits related to

DOT method will be robust tomuch of this uncertainty, plant growth strategies (ot niche) and are more similar

because the same ages are used in the calculation of than expected for traits related to climatic envelopes (?
weighted divergence times for both traits. For the nodes niche). (3)We reanalyze the Ceanothus phylogeny, based
from the root to the tips, ages on internal transcribed spacer (ITS) sequence data, to
along any contiguous path
will be correctly even if the actual values obtain an ultrametric tree with branch lengths fit to a
always ordered,
are uncertain. Problems would most likely arise if molecular clock. (4)We use this phylogeny and the trait
incorrect calibrations were to two or more data set to apply the DOT analysis, testing the
applied
independent segments of the tree (i.e., along different prediction that ot niche traits diverged earlier than ?
in which different traits exhibited niche traits in the evolution of Ceanothus.
root-to-tip paths)
large evolutionary divergences. Attention to this prob Methods
lem is warranted in applications of the test.
Occurrence data
Case Study: Community Assembly and Niche
A matrix of co-occurrence data for Ceanothus in
Evolution In Ceanothus
California was obtained from a search of the literature
The woody plant group Ceanothus comprises ?55 and consultation with colleagues. A total of 51 sites were
minimum-rank taxa (species or subspecies) that have obtained that had two or more co-occurring Ceanothus,
primarily diversified within the California Floristic with a total of 16 different taxa (plots in the same
Province (McMinn 1942, Hardig et al. 2000). The group location with the same species composition were
is divided into two well-supported clades (Jeong et al. recorded as one site; Nicholson 1993). Of these sites,
1997, Hardig et al. 2000) that differ consistently in 35, with 13 taxa, were located in chaparral of the Coast
several morphological and physiological traits related to or Transverse ranges, while 16 sites with 7 taxa were

drought tolerance. The two groups are considered from the Sierra Nevada and
region (CT SN, respec
subgenera, and are currently designated Cerastes and
tively). Of the 51 sites, 48 had just two Ceanothus
Ceanothus; for greater clarity (and consistency with while the remainder had three. B
species, (See Appendix
phylogenetic naming conventions), we prefer the older for details of the occurrence data matrix.)
name Euceanothus for the latter group. Species in
Trait selection
Cerastes have thick leaves with stomatal crypts and have
shallower roots and more embolism-resistant xylem than coexistence of Ceanothus in a matrix
Although species,
do members of Euceanothus (McMinn 1942, Hellmers et of other taxa in the community, can involve contrasting
al. 1955, Davis et al. 1999). Additionally, species of physiological or regeneration strategies, no direct studies
Cerastes establish only from seed following fire, while of coexistence mechanisms in chaparral have been
Euceanothus species generally resprout as well (Wells conducted. To reflect differences in drought tolerance,
1969, Schwilk and Ackerly 2005). In a Mediterranean the ideal traits would be either a direct measure of xylem
type climate, seedlings of nonsprouting species must tolerance to embolism under water deficit or wood
survive an intense summer drought after winter or which is a close correlate of tolerance
period density, xylem
spring germination. (Hacke et al. 2001). These traits are known to vary
 D.S54D. ACKERLY ET AL. Ecology Special Issue

between the two (Davis et al. 1999), but they niche parameters show greater similarity among co
subgroups
are not available for large numbers of species. As a proxy occurring species than expected. The measure of trait
for specific we used dissimilarity within communities was simply the differ
contrasting physiological strategies,
leaf area (SLA), the ratio of fresh leaf area to dry mass. ence between species values for two-species samples, and
This well-studied leaf functional trait plays an important the mean of the successive differences among ranked
role in the "leaf economic spectrum" of variation in plant values in three-species samples. We also calculated the
metabolic rates et al. 2004). In general, mean trait value for each community, which allowed us
(Wright species
with higher SLA have shorter leaf life span and higher to test our additional null models. (4) The standard
photosynthetic rates. In chaparral shrubs such as deviation of site means should be higher for climate niche

Ceanothus, high SLA is associated with less drought parameters than expected by chance, due to turnover of

tolerant leaves (Ackerly 20046), and we have a large data species along climatic gradients. Finally, (5) among-site
set available for roughly two-thirds of the Ceanothus taxa standard deviation inmean SLA should be lower than
(Ackerly 2004?z).We do not claim that differences in SLA expected under the null, as the combinations of species
per se promote coexistence, but rather that they could be from the two clades result in high trait disparity within
associated with suites of traits that are related to niche sites and low disparity across sites. One-tailed tests were

partitioning and differentiation in ecological strategies of conducted for all hypotheses, based on these
predictions,
Data for SLA are species means comparing the observed data to 999 randomizations.
co-occurring species.
collected from field, herbarium, and botanic As a null model, we use the "independent-swap"
previously
garden specimens (Ackerly 2004a); three new taxa that algorithm (Gotelli and Entsminger 2001), preserving
in the occurrence data set were added to this both site and frequency of occurrence
appeared diversity species
trait data set (C. greggii, and C. sanguineus) while randomizing assignments of species to sites. This is
C.parvifolius,
based on measurements of specimens in the University critical to ensure that patterns of trait assembly do not

and Jepson Herbaria (University of California-Berkeley, simply reflect differential abundance of particular
California, Values were for all All calculations were carried out in R (R-project
USA). log-transformed species.
analyses, as relative differences are a better measure of 2004); the swap algorithm was implemented by S.
physiological differentiation (Reich et al. 1997). (See Kembel in C as part of the PHYLOCOM package
Appendix C for details of the trait data set.) (Webb et al. 2004).
At scales niche), Ceanothus are
larger spatial (?
Phylogeny
differentially distributed with respect to edaphic con
ditions (e.g., serpentine specialists), habitat (chaparral The Ceanothus phylogeny of Hardig et al. (2000) was
vs. conifer elevational and macroclimate to obtain an ultrametric tree fit to a
forests), range, reanalyzed
(precipitation and temperature) (Hickman 1993, Nich molecular clock for the taxon sample in our trait data

olson 1993,Davis et al. 1999).We quantified the realized set. Conflicting phylogenies for Ceanothus based on ITS
climatic niche, to characterize these distribu vs. matK sequence data could reflect lineage sorting
large-scale
distributions on climate radiations or hybridization (Hardig et al.
tions, by overlaying species during rapid
of California and mean climatic and ITS is a nuclear marker) was selected
maps calculating 2000), (which
parameters for each species (Knight and Ackerly as a more reliable estimator of the "true" species tree.

2001). We have selected the mean precipitation and the Limited sampling of ndhF (10 taxa; Jeong et al. 1997) is
mean within the geographic range insufficient to incorporate in the broader analysis
January temperatures
of each distributions along geographic considered here. ITS sequence data for 76 accessions
species, reflecting
and elevational in California. The climate (73 Ceanothus and 3 outgroups [Adolphia californica,
gradients
niche analysis serves as an indirect surrogate for Zizyphus obtusifolia, and Spyridium parvifolium]) were
unmeasured physiological traits related to species obtained from GenBank (accessions GBAN-AF048901
tolerances and distributions along climate gradients. through GBAN-AF048975; Hardig et al. 2000). Sequen
This assumes that distributions do not ces were aligned with ClustalX using default parameters,
interpretation
simply reflect historical factors and limited dispersal and alignments were checked by eye (no manual

potential. The contraction and expansion of chaparral adjustments were made); total aligned sequence length
during the interglacial periods (Graham 1999) argues for ITS1, ITS2, and the intervening 15S region was 627
a strong role for limitation as a nucleotides. Taxa with identical sequences were kept in
against dispersal long
term constraint on distributions. the analysis, for use later in comparative analyses.
species
sequences for individual taxa were pruned to
Multiple
Community assembly one sequence, based on
representative preliminary
We used five null models for assembly to analyses (Hardig et al. 2000). The resulting analysis
community
test several relative to patterns that would be included 56 Ceanothus sequences and 96 informative
predictions,
expected by chance: (1) Species from the two clades characters. Phylogenetic analysis was conducted with

and co-occur more often than PAUP*, criteria and a heuristic search
(Cerastes Euceanothus) using parsimony
of show variation addition with 10
expected. (2) Values SLA greater (random sequence replicates, TBR,

among co-occurring species than expected. (3) Climate MULTREES in effect, collapse zero-length branches in
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S55

effect, and steepest descent not in effect) (Swofford polytomy). Zero-length branches create problems for

2002). This analysis resulted in 174 equally parsimo both ANCML and independent contrasts, as they imply
nious trees of length 322 on one island, which was hit in instantaneous evolutionarydivergence (a hard polyto
all 10 replicate searches. The strict consensus of the my), whereas they
might reflect uncertainty of the

equally trees was very similar to the results sequence of speciation events (soft polytomy) as much
parsimonious
reported by Hardig et al. (2000). The most significant as rapid radiation. Zero-length branches represent
difference was that we obtained more resolution within truncated estimates of elapsed time since speciation,
Euceanothus, with the Western group (Hardig et al. since each branch will remain at zero length until the first
2000: Fig. 1)monophyletic and nested within a para fixed base change occurs. We addressed this problem by
Eastern group. adjusting zero-length branches to a small nonzero value
phyletic
For comparative analysis, ANCML (Schl?ter et al. (1.0 X 10~4) slightly lower than the shortest branches
1997) requires fully bifurcating phylogenies. It is resulting from the maximum-likelihood fit to the
possible to generate these by randomly resolving the molecular clock, which ranged from 1.03 X 10-4 to 4.94
trees obtained in the analysis just described (in which X 10~4 across the 100 trees. Based on our calibration, this

branches were but we are not adjustment represents an absolute time of ~5 X 104 yr.
zero-length collapsed),
aware of software that will provide alternative reso Note that these small divergences (especially if they occur
lutions while maintaining branch lengths. As an on both sides of a bifurcation) will lead to an inflated
alternative, we conducted a second search estimate of the Brownian motion rate parameter and,
parsimony
in PAUP* with the same but with zero hence, broader confidence intervals on ancestral states.
parameters,
not and MAXTREES = This will errors
length branches collapsed increase the standard of weighted
10000. Ceanothus which was divergence age (W) from our bootstrap procedure,
oliganthus ssp. oliganthus,
ourin data but missing from the leading to a more conservative test of significance for
present ecological
molecular data, was added to the matrix with the same the DOT statistic. We recognize that this is an imperfect
as C. oliganthus sorediatus. Given taxa solution and hope that the problem of zero-length
sequence ssp.
with identical this search reached branches will be addressed in future research on branch
sequences, quickly
= calibration for
the maximum number of trees, with length 322, as in length comparative analysis.
the prior case. We then pruned the trees to include only
Divergence order test analysis
the 39 taxa for which we had phenotypic data, resulting
We on
(outgroups were not
in 3254 unique conducted the DOT analysis based the average
topologies
considered in the of ancestral For age of internal nodes, weighted by the absolute value of
analysis states).
further 100 trees were selected from this set by the unstandardized independent contrasts for each trait.
analysis,
of the difference in ages was
sampling every 20th tree (because immediately adjacent Significance assessed by
trees tend to be similar to each other). This set of 100 bootstrapping trait histories from the mean and stan

alternative, trees was used for all dard deviations obtained from ANCML. Given the
fully bifurcating
strong preliminary data that we have introduced, we
subsequent analyses.
conducted one-tailed tests of the hypothesis that SLA
Branch lengths occurred earlier than in climate
divergence divergences
Maximum-likelihood methods were used to fit branch niche parameters.

lengths to the 100 topologies, based on the HKY85 Results

model with transition/transversion rates fit empirically
from the data (Swofford 2002). A molecular clock was Trait variation

not rejected using this model (P > 0.05 for all trees), so Across the entire trait data set leaf
(39 taxa), specific
the branch lengths fit with a molecular clock were used area (SLA) was significantly higher inEuceanothus than
for comparative analyses. Based on
an independent Cerastes. and SLA were correlated
Precipitation weakly
of rates of rbcL evolution in Ceanothus =
analysis (Jeong overall (R 0.25), and they were essentially independent
et al. 1997), the split between Cerastes and Euceanothus based on independent contrasts =
(R 0.04) (Fig. 2A).
was calibrated at 18-39 Ma. Fossil evidence and SLA were corre
provides January temperature negatively
=
independent confirmation of the minimum age, as taxa lated across species (R -0.35) and based on

assignable to both clades appear in the fossil record by independent contrasts (R

=
-0.31). This negative
18 Ma (Chaney 1927, Axelrod 1956). Using this relationship was also apparent within Euceanothus, but
calibration for the basal node, the clock-calibrated tree not in Cerastes (Fig. 2B). The decline in Euceanothus
suggests that radiation of each of the two (Cerastes and reflects the transition from deciduous taxa the
occupying
no more than 4-5 Ma, at coldest to evergreens
Euceanothus) began recently ranges (e.g., C. parvifolius) at lower
about the same time as the onset of the Mediterranean latitudes or elevations. The most of both
striking aspect
type climate in California. relationships is the marked difference in SLA between
The inclusion of taxa with identical sequences (a the clades that is maintained across the climatic
common occurrence for consistent with the of local co
rapidly speciating groups) gradients, prediction
results in zero-length branches in the phylogeny (i.e., a occurrence between that differ in SLA.
species
 S56 D. D. ACKERLY ET AL. Ecology Special Issue

250 500 750 1000 1250 1500-6-3 0 69 3

Mean annual precipitation (mm) Mean January temperature (?C)
Fig. 2. Specific leaf area (SLA; originally measured in mm2/mg) vs. (A) mean precipitation, and (B) January temperatures
within species ranges for species of Cerastes (solid circles) and Euceanothus (open circles).

Community assembly perhaps due to distribution of evergreen vs. deciduous

(low- vs. high-SLA) species elevational

Taxonomic co-occurrence.?Of the 48 sites with two along gradients.
one two major As expected, climate niche parameters are always
species, 38 had from each of the clades
and Euceanothus); this pattern was particularly similar among co-occurring species and significantly
(Cerastes
different sites.
striking in the Coast/Transverse (CT) chaparral sites (31 among
of 35 sites), but was not significant in the Sierra Nevada
Divergence order test analysis
(SN) (Table 1). For both the full data set and the coastal
partition, the elevated co-occurrence of species from These results support the selection of SLA and
different clades was highly significant, relative to the null realized climate niche parameters as traits reflecting
model that maintained both site diversity and species local (ot)vs. regional (?) differentiation, respectively. The
occurrence < The co-occurrence of are in the Coast and Transverse
(P 0.001). frequent patterns stronger
taxa from the two clades has long been noted in the ranges, and if there were a monophyletic group in
literature, but not previously quantified and tested Ceanothus restricted to these communities we would
relative to a null model. limit our to this the evolu
analyses group. However,
Trait disparity.?Across all sites, the mean difference radiation into the two and
tionary major clades,
in logio(SLA) between co-occurring species was 0.32
subsequently within clades, encompasses both geo
units, and as this value was
predicted significantly
graphic areas (and beyond). We feel it is important to
greater than the null expectation (P < 0.04; Table 2). In
contrast, differences between climate niche parameters
of co-occurring species were much smaller than expected Table 1. Relative frequency of local co-occurrence patterns
for Ceanothus, in terms of the number of species from each
(P < 0.001 in both cases; Table 2). When the data are = Coast and Transverse =
major subgroup (CT ranges; SN
partitioned geographically, the within-site disparity in Sierra Nevada).
SLA was still significantly greater in CT, but disparity in
SN was less than expected. Disparity in climate niche No. communities
??
???
:
parameters is significantly lower than the null, as Community -pattern Total CT SN
predicted, in both areas (Table 2). The among-site Cerastes Euceanothus (P < 0.001) (P < 0.001) (ns)
standard deviation in trait means was significantly 0 2 7 2 5
greater than expected for climate niche traits, as 11 38 31 7
predicted, in the entire data set and both geographic 2 0 3 2 1
For SLA, it was significantly lower than 0 3 2 0 2
partitions. 1 2 1 0 1
in CT, but the was reversed in SN
expected pattern 2 1 0 0 0
(Table 2). 3 0 0 0 0
Collectively, these analyses indicate that in Coast and Notes: of how to read Table 1: The first row
Example
Transverse range chaparral, co-occurring Ceanothus indicates that a total of seven communities were recorded (two
species exhibit greater disparity in SLA than expected in CT and five in SN) with two co-occurring Euceanothus
species and no Cerastes. P values indicate the probability of
under a null model of community assembly. In
obtaining the observed number of sites occupied by species
contrast, in the Sierra Nevada it appears that SLA from the two subgroups, relative to a null model of community
varies among sites, while within-site disparity is low, assembly (see text).
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S57

Table 2. Mean trait disparity among co-occurring species, and standard deviation of trait means among sites.

Mean disparity SD of site means

Trait Prediction All (N= 51) CT (N= 35) SN (N= 16) Prediction All (N= 51) CT (N= 35) SN (N= 16)

Specific leaf area

(mm2/mg, log)
Observed mean 0.32 0.35 0.25 0.158 0.093 0.170
Expected mean (sd) 0.29 (0.018) 0.28 (0.021) 0.28 (0.016) 0.17(0.012) 0.15(0.013) 0.14(0.014)
P <0.04 <0.001 ns ns <0.001 NS

Annual precipitation
(mm)
Observed mean < 117 123 103 > 217 159 83
Expected mean (sd) 280 (25.2) 187(17.3) 121 (8.7) 155(12.1) 124(10.5) 67 (8.22)
P <0.001 <0.002 <0.03 <0.001 <0.002 <0.02

January minimum
temperature (?C)
Observed mean 1.25 0.92 1.95 2.001 0.985 1.571
Expected mean (sd) 2.53(0.19) 1.24(0.093) 2.49(0.21) 1.50(0.11) 0.81(0.065) 1.23(0.14)
P <0.001 <0.002 <0.02 <0.001 < 0.01 <0.02

Notes: Observed values are compared to expectations based on a null model of community assembly. Analyses were conducted
for all sites, and for Coast and Transverse ranges (CT) and Sierra Nevada (SN) sites separately. Prior predictions regarding the
direction of the difference between observed and expected values are listed, and all significance tests are one-tailed.

conduct evolutionary analyses on all available data for For all 100 alternative phylogenies, weighted divergence
the entire group. ages were older for SLA than for both climate
The three traits considered here generally fit the parameters, and the DOT was significant in 94 of 100
Brownian motion model underlying the use of ANCML trees for the SLA vs. January temperature difference,
for ancestral states. For all three traits, inspection of and for 92 out of 100 trees for SLA vs. precipitation.
normal probability plots for the absolute value of
Discussion
standardized contrasts indicated a close fit to a

truncated normal distribution. For SLA and January This analysis of the radiation of Ceanothus demon

temperatures, correlations of standardized contrasts and strates an initial shift in otniche traits that subsequently
their standard deviation (i.e., the square root of the promote (or at least facilitate) coexistence among related
These traits were as the radiation
lengths) were not
then conserved
subtending branch significant. For species.
the correlation was progressed, and later speciation events were character
precipitation, significantly negative
across all trees, reflecting ized primarily by divergence in climate envelopes,
larger than expected divergen
ces across to geographic differentiation along lat
rapid speciation events, and a small diver corresponding
across itudinal and elevational gradients. The result is that co
gence the long branches between Cerastes and
occurring Ceanothus species within local communities
Euceanothus. This pattern will lead to a relatively high
are more distantly related than expected by chance,
Brownian motion rate parameter, to accommodate these
relative to the group as a whole.
rapid divergences, and will thus inflate the confidence
Similar patterns are evident in other clades that have
intervals around the ancestral estimates for precipita
been analyzed in a phylogenetic context. In the oaks
tion, making the DOT more conservative. Across most
(Quercus) of Florida, local communities tend to be
of the 100 trees, rates of trait evolution were homoge
phylogenetically overdispersed, often with one or two
neous between Cerastes and Euceanothus; in 3 and 24
members each drawn from three distinct clades (red,
trees, rates of SLA and January temperature evolution,
white, and live oaks). Habitat preferences diverge
respectively, were significantly higher in Euceanothus
repeatedly within each of these clades; deeper divergen
(0.04 < P < 0.05). ces between the clades involve traits, such as seed
Fig. 3 illustrates trait divergences on a randomly maturation time and wood density, which may promote
selected tree from the analysis. The largest contrast for
coexistence through differential regeneration or patho
SLA was located at the basal node between Cerastes and
gen tolerance, respectively (Cavender-Bares et al. 2004).
Euceanothus (Fig. 3). The weighted divergence age for In studies of Phylloscopus warblers in Kashmir, Rich
SLA, averaged across the 100 alternative phylogenies, man and Price (1992, Richman 1996) argued that deep
was 6.5 X 10~3 branch length units (Table 3). For the
divergences in the group involved differentiation in
climate parameters, the basal contrasts were much while more recent events
feeding strategies, speciation
smaller, and larger were noted within each were related to macrohabitat distributions
divergences (coniferous
of the two major clades (Fig. 3). The weighted vs. deciduous forests) (but see Forstmeier et al. [2001]).
divergence ages were 4.30 X 10~3 and 4.24 X 10~3 for In the radiation of Anolis distinctive
lizards, ecomorphs,
January temperatures and precipitation, respectively. which coexist by feeding in different parts of the canopy,
 D. S58
D. ACKERLY ET AL. Ecology Special Issue

groups (Glor et al. 2003). In our terminology, the earlier

divergence events in each of these cases involve shifts in

the ot niche, whereas ? niche traits continue to diverge
later in the radiation. Streelman and Danley (2003)
include the anoles as one of the case studies in their
review, considering the diversification of ecomorphs as
an example of the first stage of radiation, involving
microhabitat divergence. This is consistent with our

interpretation of ecomorphs as diversification of the ot

niche, although our terminology is different.

co 0.6n These case studies provide a contrast to

striking
B
T
Diamond's (1986) habitat-first model, which proposed
?? 0.5H O
^1
that the first stage of speciation and adaptive radiation
o$ "53 0.4
0 CO i involved allopatric divergence along habitat gradients.
0.3
? c These contrasting interpretations reflect differences in
JE O 0.2 the of fast- vs. slow-evolving
? o interpretation traits, and
CD ?
Q. 0.1 they highlight underlying philosophical differences in
CO the of data.
0.0' ?r-'v-r? & interpretation comparative
0.04 0.03 0.01 It is well known that phylogenetic inference works

0s_ best for slowly evolving traits (Felsenstein 1988). Rapid

8.0-1
C evolution erodes the
signal of early events on a
-i??
CO due toreversals on the same or adjacent
6.0 phylogeny,
branches and convergence among the terminal states
CLto
E 2-? 4.0
cd
(strictly speaking, this is only true for traits with a finite
number or range of states; Donoghue and Ree 2000,
b8 2.0 Ackerly and Nyffeler 2004). For this reason, indepen
dent contrasts will generally provide an extremely poor
estimate of the timing of divergence for rapidly evolving
0.04 0.03
traits, as itwill appear that there was little divergence in
800 deeper nodes (e.g., Fig. 1). The high level of uncertainty

H#H in ML ancestral state reconstructions reflects this

c 600 problem for rapidly evolving traits (Schl?ter et al.
g ^
'm W 1997, Cunningham et al. 1998) and led us to adopt the
.-S 2 400 O <3 method here. In this situation, if the
bootstrap proposed
-9-c
o o most recent events in an adaptive radiation
? speciation
2 200
involve divergence in macrohabitat or habitat-related

traits, it is parsimonious to assume that deeper events

~r \ T
0.04 0.03 0.01 also involved such divergences (T. Price, personal
communication), though evidence of this will not be
Node age available from phylogenetic analysis.
Fig. 3.
Divergence order test (DOT) for Ceanothus, In contrast, for slowly evolving traits phylogenetic
illustrated for one randomly chosen tree out of 100 equally methods are to a different
quite powerful, leading
parsimonious trees used for analysis. For all panels, ages along
interpretation of the comparative data. For this case,
the x-axis are clock-calibrated molecular branch length units; as in the contrasting SLA values in Ceanothus, or
the breaks in the axis indicate the long branches connecting the
two recently clades. in Anolis, the greatest trait differ
diversifying (A) Ceanothus phylogeny, divergent ecomorphs
based on reanalysis of internal transcribed spacer (ITS) ences will generally be observed among distantly related
sequence data from Hardig et al. (2000). Species names are The conservatism of these traits in diverse clades
species.
omitted for clarity. (B-D) Open circles indicate the magnitude within each that the underlying
group strongly argues
of unstandardized contrasts vs. node age for specific leaf area
divergences occurred during the initial speciation events
(SLA), January temperatures, and mean precipitation, respec
mean errors of in the overall radiation. The view that ot niche
tively (see Table 3 for weighted ages). Standard early
the contrasts (derived from the bootstrap procedure) are divergence occurs late in the course of species differ
illustrated for the basal contrast only. The solid diamonds because these traits tend to differ between
entiation,
indicate weighted mean divergence age for each trait (?se)
distantly related species, is not compatible with com
based on 200 bootstrap randomizations of the contrasts
of this point is arbitrary). parative phylogenetic analysis.
(vertical position
Taken together, these views lead to a possible
have evolved independently on multiple islands (Losos synthesis of contrasting interpretations. In the cases
et al. 1998, Knouft et al. 2006). On the larger islands, discussed here, adaptive radiation
have may
proceeded
however, there has been continued speciation involving by "ot niche early" and "? niche throughout." In other
diversification of macrohabitats within ecomorph words, habitat divergence can occur frequently at all
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S59

Table 3. Results of the divergence order test (DOT) for relative divergence times of specific leaf area (SLA) and climate niche
parameters in Ceanothus.

Mean
Trait age (se) D nonzero
No.
(se)

SLA 0.00654(0.00083)
January temperature 0.00430 (0.00071) 0.00225 (0.00108)94
Precipitation 0.00424(0.00083) 0.00230(0.00117) 92

Notes: Mean age is the average, over 100 alternative trees, of the weighted divergence times (W) for each trait. For each tree, Wis
derived by bootstrapping the ancestral states from maximum likelihood distributions for ancestral states, with 200 replicates; D is the'
mean difference in age of each climate parameter vs. SLA, over the 100 alternative trees; no. nonzero is the number of trees (out of
100) in which D was significantly different from zero, based on a one-tailed test of the hypothesis that mean age for SLA was older.

depths in the tree, although it can only be reconstructed generate stabilizing selection effects that lead to evolu
with confidence in recent speciation events. Habitat tionary stasis in niche parameters (Holt 1987, 2003,
differences clearly can play an important role in Kirkpatrick and Barton 1997, Case and Taper 2000).
allopatric speciation, though it is important to note that studies of these are needed. The
Empirical predictions
allopatric populations might also occupy similar envi roles of niche conservatism in speciation, the evolution of
ronments in different geographic areas (Peterson et al. biota and the of communities has
regional assembly
1999, Wiens 2004). In some cases, speciation also received increased attention et al. 2002,
recently (Webb
involves a shift in ot niche traits related to resource
Ackerly 2003, Wiens 2004), and each of these offers a
partitioning in local communities, but these events are to the emphasis on ecological as
counterpoint divergence
apparently less frequent. Divergence in ot niche traits a key component of evolutionary radiations.
might be due to incidental divergence under conditions The conclusion that ot niche traits evolve relatively
that favor differenttraits (e.g., island vs. mainland an radiation contrasts with the
slowly during adaptive
or divergence by character in
populations), displacement views of Silvertown et al. (2006?z, b) on niche evolution.
sympatry due to direct competitive interactions between based on several lines of evidence, that the ot
They argue,
the incipient species (Schl?ter 2000, Levin 2004). niche evolves and as a corollary local commun
rapidly,
If correct, "ot early,
the ? throughout" model presents ities usually show little phylogenetic in their
signal
two unresolved questions about the evolution of ot niche This. conclusion is
species composition. supported by
traits. First, why should these traits exhibit evolutionary their analysis of the structure of English
phylogenetic
transitions early on in adaptive radiations? (At the time meadow communities (Silvertown et al. 2001, 2006a, b)
of this initial divergence, the adaptive radiation is still an and of niche between
by comparison overlap congeneric
unrealized future of the clade.) And second, why do ot and within local
noncongeneric species communities.
niche traits often exhibit phylogenetic conservatism The conflict between their and our
apparent analyses
during subsequent diversification. With regard to the conclusions ismost likely due to the different scales of
first question, it is tempting to identify divergence in ot and of taxa. The of the English
analysis sampling species
niche traits as key innovations, or invasions of a novel meadow communities are across the
widely dispersed
adaptive zone (sensu Simpson 1953), contributing to the angiosperm phylogeny. When the phylogeny is pruned
subsequent radiation. For example, in the case of for of niche the closest
relatives
analysis distributions,
Ceanothus we have strong evidence that California on the tree are if ever immediate
remaining rarely
chaparral communities are capable of supporting at As a result, even the most recent "events"
sibling species.
least one sprouting and one nonsprouting Ceanothus on such a are old
represented phylogeny relatively
species. Thus, when this trait diverged early in the to our of adaptive radiations.
compared analysis Thus,
evolution of Ceanothus (the direction of evolution Silvertown et al.'s conclusions and our
(2006?z) analyses
between ancestral and derived states is unknown), the be but focused on different
may entirely compatible,
two descendent subclades were both able to diversify in scales of analysis.

parallel across a broad gradient of climatic conditions.

While such scenarios are plausible, and may Conclusions
be correct,
it is difficult to conduct rigorous analyses of diversifi The divergence order test (DOT) introduced here
cation hypotheses in terms of the timing of phenotypic provides a quantitative approach to test hypotheses
innovation and hypothesized shifts in speciation rates about the relative sequence of divergence for continuous
(Sanderson and Donoghue 1994, Hodges 1995). traits. Considering the potential pitfalls in comparative
The second
question addresses the important topic of analysis of rapidly evolving traits, the bootstrap method
niche conservatism: What are the mechanisms promoting the uncertainty of ancestral state recon
incorporating
evolutionary stasis in ecological traits through speciation structions provides a conservative for signifi
approach
and diversification of a clade? Recent theoretical analyses cance testing. Our application of the DOT to Ceanothus,

support the view that contrasting selection pressures in and interpretation of other cases in the literature, leads

heterogeneous environments, combined with gene flow, to the conclusion that ot niche traits often diverge early
habitat can in the course of adaptive radiations. The
interspecific competition, and/or selection, ? niche traits,
 D. S60
D. ACKERLY ET AL. Ecology Special Issue

which are related to macrohabitat distributions, might Elton, C. 1946. Competition and the structure of ecological
communities. Journal of Animal Ecology 15:54-68.
evolve rapidly throughout the radiation, although the
Felsenstein, J. 1985. Phylogenies and the comparative method.
signal of early divergences is erased by the high rates of American Naturalist 125:1-15.
evolution. Further application of the DOT, or improved J. 1988. Phylogenies and quantitative characters.
Felsenstein,
tests these questions, will an impor Annual Review of Ecology and Systematics. 19:445^171.
addressing provide
Forstmeier, W., O. Bourski, and B. Leisler. 2001. Habitat
tant step towards synthesis of niche evolution and
choice in Phylloscopus warblers: the role of morphology,
adaptive radiation. and competition. 128:566-576.
phylogeny Oecologia
Garland, T., Jr 1992. Rate tests for phenotypic evolution using
Acknowledgments
phylogenetically independent contrasts. American Naturalist
D. D. Ackerly thanks C. O. Webb and J. B. Losos for the 140:509-519.
invitation to contribute this paper to the ESA symposium and Garland, T., Jr., P. H. Harvey, and A. R. Ives. 1992. Procedures
this special issue on Phylogenetics and Community for the analysis of comparative data using phylogenetically
Ecology.
The authors thank D. Schl?ter and T. Price for valuable independent contrasts. Systematic Biology 41:18-32.
discussions and comments that improved the manuscript. This Glor, R. E., J. J. Kolbe, R. Powell, A. Larson, and J. B. Losos.
research was supported 2003. Phylogenetic analysis of ecological and morphological
by National Science Foundation grants
0212873 to C. O. Webb, M. J. Donoghue, and D. D. Ackerly diversification in Hispaniolan trunk-ground anoles (Anolis
and 0078301 to D. D. Ackerly. cybotes group). Evolution 57:2383-2397.
Gotelli, N. J., and G. Entsminger. 2001. Swap and fill
Literature Cited algorithms in null model analysis: rethinking the knight's
tour. Oecologia 129:281-291.
Ackerly, D. D. 2000. Taxon sampling, correlated evolution and and G. Graves.
Gotelli, N., 1996. Null models in ecology.
independent contrasts. Evolution 54:1480-1492. Smithsonian Institution Press, Washington D.C., USA.
Ackerly, D. D. 2003. Community assembly, niche conservatism
Graham, A. 1999. Late Cretaceous and Cenozoic history of
and adaptive evolution in changing environments. Interna North American vegetation. Oxford University Press, New
tional Journal of Plant Sciences 164:S165-S184.
York, New York, USA.
Ackerly, D. D. 2004a. Adaptation, niche conservatism and
Graham, C. H., S. R. Ron, J. C. Santos, C. J. Schneider, and C.
convergence: comparative studies of leaf evolution in the
Moritz. 2004. Integrating phylogenetics and environmental
California chaparral. American Naturalist 163:654-671.
niche models to explore speciation mechanisms in dendro
Ackerly, D. D. 2004&. Functional strategies of chaparral shrubs batid frogs. Evolution 58:1781-1793.
in relation to seasonal water deficit and disturbance.
Hacke, U., J. Sperry, W. Pockman, S. Davis, and K.
Ecological Monographs 75:25^4.
McCulloch. 2001. Trends in wood density and structure are
Ackerly, D. D., and R. Nyffeler. 2004. Evolutionary diversifi linked to prevention of xylem implosion by negative pressure.
cation of continuous traits: phylogenetic tests and application
Oecologia 126:457-461.
to seed size in the California flora. Evolutionary Ecology 18:
Hardig, T. M? P. S. Soltis, and D. E. Soltis. 2000.
249-272.
Diversification of the North American shrub genus Ceano
Axelrod, D. I. 1956. Mio-Pliocene floras from west-central
thus (Rhamnaceae): conflicting phylogenies from nuclear
Nevada. University of California Publications in Geological
ribosomal DNA and chloroplast DNA. American Journal of
Sciences 33:1?316.
Botany 87:108-123.
Case, T. J., and M. L. Taper. 2000. Interspecific competition, P. H., and M. Pagel. 1991. The comparative method in
environmental Harvey,
gradients, gene flow, and the coevolution of
evolutionary biology. Oxford University Press, Oxford, UK.
species' borders. American Naturalist 155:583-605.
Hellmers, H., J. S. Horton, G. Juhren, and J. O'Keefe. 1955.
Cavender-Bares, J., D. D. Ackerly, D. Baum, and F. A. Bazzaz.
some
Root systems of chaparral shrubs in Southern
2004. Phylogenetic repulsion in the assembly of Floridean
California. Ecology 36:667-678.
oak communities. American Naturalist 163:823-843.
Hickman, J. C. 1993. The Jepson manual: higher plants of
Chaney, R. W. 1927. Geology and palaeontology of the
California. University of California Press, Berkeley, Califor
Crooked River Basin with special reference to the Bridge
Creek Flora. Carnegie Institution of Washington Publication nia, USA.
346:45-138. Hodges, S.
1995. Spurring plant diversification: Are floral

Chase, J., and M. Leibold. 2003. Ecological niches:

nectar spurs a key innovation? Proceedings of the Royal
linking
Society (London) B 262:343-348.
classical and contemporary approaches. University of Chi
Holt, R. D. 1987. Population and evolutionary
cago Press, Chicago, Illinois, USA. dynamics
P. L. 2000. Mechanisms of maintenance of species processes: the manifold roles of habitat selection. Evolu
Chesson,
Annual Review of Ecology and Systematics 31: tionary Ecology 1:331-347.
diversity.
343-367. Holt, R. D. 2003. On the evolutionary ecology of species'
C. W., K. Omland, and T. H. Oakley. 1998. ranges. Evolutionary Ecology Research 5:159-178.
Cunningham,
ancestral character states: a critical Hubbell, S. P. 2001. The unified neutral theory of biodiversity
Reconstructing reap
and biogeography. Princeton
praisal. Trends in Ecology and Evolution 13:361-366. University Press, Princeton,
C. 1859. On the origin of species. Murray, New Jersey, USA.
Darwin, London,
UK. Jeong, S.-C, A. Liston, and D. D. Myrold. 1997. Molecular
S. D., F. W. Ewers, J.Wood, J. J. Reeves, and K. J. Kolb. phylogeny of the genus Ceanothus (Rhamnaceae) using rbcL
Davis,
1999. Differential to xylem cavitation and ndhF sequences. Theoretical and Applied Genetics 94:
susceptibility among
852-857.
three pairs of Ceanothus species in the Transverse mountain
ranges of Southern California. Ecoscience 6:180-186. Keeley, J. E. 1977. Fire-dependent reproductive strategies in
Diamond, J. M. 1986. Evolution of ecological in the Arctostaphylos and Ceanothus. Pages 391-396 in H. A.
segregation
New Guinea montane avifauna. 98-125 in J. M. Mooney and C. E. Conrad, editors. Proceedings of the
Pages
Diamond and T. J. Case, editors. Symposium on Environmental Consequences of Fire and
Community ecology.
and Row, Cambridge, Massachusetts, USA. Fuel Management in Mediterranean Ecosystems. USDA
Harper
Forest Service General Technical
Diaz-Uriarte, R., and T. Garland, Jr. 1996. Testing hypotheses Report WO-3, Washington,
of correlated evolution using phylogenetically independent D.C., USA.
contrasts: sensitivity to deviations from Brownian motion. Keeley, J. E., and P. H. Zedler. 1978. Reproduction of
Systematic Biology 45:27-47. chaparral shrubs after fire: a comparison of sprouting and
Donoghue, M. J., and R. Ree. 2000. Homoplasy and seeding strategies. American Midland Naturalist 99:142-161.
developmental constraint: a model and example from plants. Kirkpatrick, M., and N. H. Barton. 1997. Evolution of a
American Zoologist 40:759-769. species' range. American Naturalist 150:1-23.
 July 2006 NICHE EVOLUTION AND ADAPTIVE RADIATION S61

Knight, C, and D. D. Ackerly. 2001. Correlated evolution of Schl?ter, D. The ecology

2000. of adaptive radiations. Oxford
chloroplast heat shock protein expression in closely related Press, Oxford, UK.
University
plant species. American Journal of Botany 88:411-418. and D. Ludwig. 1997.
Schl?ter, D., T. Price, A. Mooers,
Knouft, J. H., J. B. Losos, R. E. Glor, and J. J. Kolbe. 2006. Likelihood of ancestor states in adaptive radiation. Evolution
Phylogenetic analysis of the evolution of the niche in lizards
51:1699-1711.
of the Anolis sagrei group. Ecology 87:S29-S38.
Schwilk, D. W., and D. D. Ackerly. 2005. Is there a cost to
Levin, D. A. 2004. Ecological speciation: crossing the divide.
resprouting? Seedling growth rate and drought tolerance in
Systematic Botany 29:807-816.
A. Larson, K. de Queiroz, and L. sprouting and nonsprouting Ceanothus (Rhamnaceae).
Losos, J., T. Jackman,
1998. Contingency and determinism in American Journal of Botany 92:404-410.
Rodriguez-Schettino.
replicated adaptive radiations of island lizards. Science 279: Silvertown, J., M. Dodd, and D. Gowing. 2001. Phylogeny and
2115-2118. the niche structure of meadow plant communities. Journal of
McMinn, H. 1942. Ceanothus. Volume II. A systematic study of 89:428-435.
Ecology
the genus Ceanothus. Santa Barbara Botanical Gardens, D. Gowing,
Silvertown, J., M. Dodd, C. Lawson, and K.
Santa Barbara, California, USA.
McConway. 2006a. Phylogeny and the hierarchical organ
Nicholson, P. 1993. Ecological and historical biogeography of
ization of plant diversity. Ecology 87:S39-S49.
Ceanothus (Rhamnaceae) in the Transverse Ranges of
Silvertown, J., K. McConway, D. Gowing, M. Dodd, M. F.
Southern California. Dissertation. University of California,
Los Angeles, USA. Fay, J. A. Joseph, and K. Dolphin. 20066. Absence of
California,
T. H., and C. W. Cunningham. 2000. phylogenetic signal in the niche structure of meadow plant
Oakley, Independent
contrasts succeed where ancestral reconstruction fails in a communities. Proceedings of the Royal Society (London) B
known bacteriophage phylogeny. Evolution 54:397-405. 273:39^14.
O'Meara, B. C, C. An?, M. J. Sanderson, and P. C. G. G. 1953. The major features of evolution.
Simpson,
Wainwright. In press. Testing for different rates of evolution Columbia University Press, New York, New York, USA.
using likelihood. Evolution.
Streelman, J. T., and P. D. Danley. 2003. The stages of
Peterson, A. T., J. Soberon, and V. S?nchez-Cordero. 1999.
vertebrate evolutionary radiation. Trends in Ecology and
Conservatism of ecological niches in evolutionary time.
Evolution 18:126-131.
Science 285:1265-1267.
of an assemblage Swofford, D. L. 2002. PAUP*: phylogenetic analysis using
Pickett, S., and F. Bazzaz. 1978. Organization
Version 4.10b. Massachu
of early successional species on a soil moisture gradient. parsimony. Sinauer, Sunderland,
Ecology 59:1248-1255. setts, USA.
R Development Core Team. 2004. R: a language and environ Webb, C. O., D. D. Ackerly, and S. Kembel. 2004. Phylocom:
ment for statistical computing. R Foundation for Statistical software for the analysis of community structure
phylogenetic
Computing, Vienna, Austria. and character evolution, (http://www.phylodiversity.net/
Reich, P. B., M. B. Walters, and D. S. Ellsworth. 1997. From
phylocom)
tropics to tundra: global convergence in plant functioning.
C. O., D. D. Ackerly, M. McPeek, and M. J. Donoghue.
Webb,
Proceedings of the National Academy of Sciences (USA) 94:
2002. Phylogenies and community ecology. Annual Review
13730-13734.
A. D. 1996. Ecological diversification and community of Ecology and Systematics 33:475-505.
Richman,
structure in the Old World leaf warblers Wells, P. V. 1969. The relation between mode of reproduction
(genus Phylloscopus):
a. phylogenetic perspective. Evolution 50:2461-2470. and extent of speciation in woody genera of the California
Richman, A., and T. Price. 1992. Evolution of ecological chaparral. Evolution 23:264-267.
differences in Old World warblers. Nature 355:817-821. Whittaker, R. 1975. Communities and ecosystems. Macmillan,
Sanderson, M. J. 2002. Estimating absolute rates of molecular New York, New York, USA.
evolution and divergence times: a penalized likelihood
Wiens, J. J. 2004. Speciation and ecology revisited: phylogenetic
approach. Molecular Biology and Evolution. 19:101-109.
niche conservatism and the origin of species. Evolution 58:
Sanderson, M. J. 2003. r8s: Inferring absolute rates of
193-197.
molecular evolution and divergence times in the absence of
a molecular clock. Bioinformatics 19:301-302. Williams, C. B. 1964. Patterns in the balance of nature.
Academic Press, New York, New York, USA.
Sanderson, M., and M. Donoghue. 1994. Shifts in diversifica
tion rate with the origin of angiosperms. Science 264:1590 Wright, I. J., et al. 2004. The leaf economic spectrum world
1593. wide. Nature 428:821-827.

APPENDIX A
Discussion of alternative null models (Ecological Archives E087-110-A1).

APPENDIX B
Ceanothus occurrence matrix (Ecological Archives E087-110-A2).

APPENDIX C
Ceanothus trait matrix (Ecological Archives E087-110-A3).

SUPPLEMENT
Source code for DOT (Ecological Archives E087-110-S1).