THE ROLE OF EVOLUTIONARY HISTORY AND NICHE DIFFERENTIATION
IN STRUCTURING SPECIES CO-OCCURRENCE IN
NEW ZEALAND PITTOSPORUM (PITTOSPORACEAE)
By
Kristen Marie Nolting

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Plant Biology – Master of Science
2014

	
  

	
  
ABSTRACT
THE ROLE OF EVOLUTIONARY HISTORY AND NICHE DIFFERENTIATION
IN STRUCTURING SPECIES CO-OCCURRENCE IN
NEW ZEALAND PITTOSPORUM (PITTOSPORACEAE)
By
Kristen Marie Nolting
Understanding the processes that shape the spatial distribution of species, and their co-existence
in communities, remains one of the biggest challenges that ecologists face today. Investigations
of co-existence typically treat species as independent entities, thereby ignoring their shared
evolutionary history, niche preferences and functional similarity. This limits the ability of
ecologists to make strong inferences regarding co-existence mechanisms. It is more useful to
employ a pluralistic approach that integrates phylogenetic information and species-specific
environmental and trait associations. In this study I evaluated the role of evolutionary history
and environmental and functional trait differentiation in predicting species co-occurrence in the
New Zealand plant genus Pittosporum. I hypothesized that co-occurring species would be more
distantly related than non co-occurring species, given that closely related species are likely to be
more ecologically similar as a result of their shared ancestry, and thus competitive interactions
would preclude them from co-occurring. Similarly, I predicted that co-occurring species would
be more divergent in their functional traits to enable co-existence. I found that co-occurring
species were no different than non co-occurring species with respect to their phylogenetic
dissimilarity, that co-occurring species had higher environmental niche overlap than non cooccurring species, and that for most traits measured there was no difference in trait dissimilarity
among co-occurring and non co-occurring species. Approximate maximum vessel length and
leaf nitrogen content, however, showed convergence among co-occurring species.

	
  
ACKOWLEDGEMENTS

I specifically would like to acknowledge my collaborators and colleagues at the University of
Waikato in Hamilton, New Zealand. Dr. Chrissen Gemmill guided me through the development
of this study and implementation of my data collection while she hosted me in New Zealand in
the summer of 2013. Additionally, I am grateful for the invaluable advice and instruction I
received from Dr. Michael Clearwater and Dr. Daniel Laughlin while collecting functional trait
data in the laboratory.

	
  

iii	
  

	
  
TABLE OF CONTENTS

LIST OF TABLES ...............................................................................................................v
LIST OF FIGURES ........................................................................................................... vi
INTRODUCTION ...............................................................................................................1
METHODS ..........................................................................................................................5
Study System ...........................................................................................................5
Phylogenetic Inference.............................................................................................6
Species Functional Traits .........................................................................................7
Functional Trait Measurements .............................................................................11
Species-Specific Climatic Niche Envelopes ..........................................................14
Quantifying Predicted Niche Occupancies ............................................................15
Determining Pairwise Species Differentiation.......................................................16
Determining Co-occurrence vs. Non Co-occurrence .............................................17
Statistical Analyses ................................................................................................18
RESULTS ..........................................................................................................................19
Phylogenetic Relatedness and Co-occurrence .......................................................19
Niche Overlap and Co-occurrence .........................................................................21
Trait Dissimilarity and Co-occurrence...................................................................22
DISCUSSION ....................................................................................................................26
Phylogenetic Relatedness and Species Associations .............................................27
Environmental Niche Similarity and Co-occurring Species ..................................27
Trait Convergence in Co-occurring Species ..........................................................28
Trait Dissimilarity and Species Associations ........................................................31
Community Assembly in the New Zealand Pittosporum ......................................32
APPENDIX ........................................................................................................................34
LITERATURE CITED ......................................................................................................37

	
  

iv	
  

	
  
LIST OF TABLES

Table 1: Results from the LMM analysis of trait dissimilarity differences among
co-occurring, non co-occurring and randomly distributed species pairs ...........................18
Table 2: Results from the LMM analysis of trait dissimilarity of the first four
principal components among co-occurring, non co-occurring and randomly
distributed species pairs .....................................................................................................19
Table A1: The percent of variation explained by each of the four first components
from the principal components analysis conducted on the 12 functional traits
investigated in this study....................................................................................................35
Table A2: The loadings from the principal components analysis conducted on the
12 functional traits investigated in this study ....................................................................35
Table A3: Results from the LMM analysis of environmental niche overlap
differences among co-occurring, non co-occurring and randomly distributed
species pairs .......................................................................................................................36

	
  

v	
  

	
  
LIST OF FIGURES

Figure 1: Phylogenetic relationships of the 24 New Zealand species of Pittosporum ........7
Figure 2: The mean phylogenetic distance, environmental niche overlap and
trait dissimilarity among co-occurring (C), non co-occurring (NC) and randomly
distributed (R) species pairs ...............................................................................................20
Figure 3: Mean environmental niche overlap among co-occurring (C),
non co-occurring (NC) and randomly distributed (R) species pairs for four
independent environmental variables ................................................................................22
Figure 4: Mean trait dissimilarity among co-occurring (C), non co-occurring (NC)
and randomly distributed (R) species pairs for the first four principal components .........23
Figure 5: Mean trait dissimilarity among co-occurring (C), non co-occurring (NC)
and randomly distributed (R) species pairs for four independent functional traits............24

	
  

vi	
  

	
  
INTRODUCTION

Describing patterns of biodiversity and identifying the processes that shape the geographic
distribution of species has been a major source of scientific intrigue for centuries (eg. Darwin
1859, Wallace 1869). Despite this interest, understanding the mechanisms underlying these
patterns remains one of the biggest challenges that biologists face today (Ricklefs 1987, Emerson
and Gillespie 2008, Sutherland et al. 2013). Addressing this problem requires us to understand
both the origins of biodiversity, as well as the factors that act to govern the distributions of
species across spatial and temporal scales. The spatial distribution of species itself is determined
by species’ interactions with the abiotic and biotic environment. In other words, ecologists are
challenged to explain why species occur where they do, as well as how they manage to co-exist
given that they may compete for the same limited resources.
A central goal in community ecology is to identify the processes that act to structure
community assemblages (Weiher and Keddy 1995), and similarly, those that are responsible for
maintaining species co-existence (reviewed in Mittelbach 2012). Traditionally, there have been
two main paradigms that have been suggested as processes acting to structure the distribution of
species into communities. First, competitive interactions preclude ecologically similar species
from occupying the same space, such that species must differ substantially along one or more
important niche axes in order to maintain stable co-existence, sensu limiting similarity
(MacArthur and Levins 1967). However, recent work suggests that competitive interactions can
lead to ecologically similar species co-existing when performance differences are small (Chesson
2000, Mayfield and Levine 2010), but this is thought to only occur at very local scales and when
the environment is highly homogenous (HilleRisLambers et al. 2012). Alternatively, species

	
  

1	
  

	
  
occupying the same habitat may also be ecologically similar as a result of an abiotic filter acting
to limit the types of phenotypes, and thus species, that are able to survive in the focal
environment (Keddy 1992). In any scenario, species assemblages at the local scale should be a
non-random subset of the species pool, or a larger set of species that can potentially colonize this
study site (Diamond 1975).
Early studies aimed at understanding community assembly used species lists, distribution
records, and a null model approach to identify instances of non-random species associations (eg.
Diamond 1975, Gilpin and Diamond 1982, Connor and Simberloff 1983, Gotelli 2000). More
recently, researchers have incorporated additional information related to species’ phylogenetic
relationships (eg. Webb 2000), environmental associations (John et al. 2007), and speciesspecific traits (eg. Weiher et al. 1998) to further elucidate the mechanisms acting to structure
species distributions. Evaluation of the phylogenetic structure of communities provides a way to
potentially test whether abiotic filters or biotic interactions are more important in the assembly
process. Under the assumption that closely related species are more ecologically similar than
more distantly related species given their shared ancestry (Darwin 1859), it is inferred that
communities that are more phylogenetically overdispersed than a null model are indicative of
competitive interactions. Alternatively, those communities that are more phylogenetically
clustered are inferred to be indicative of abiotic filtering (Webb 2000). In a similar way, we can
evaluate the phylogenetic dissimilarity of individual species pairs that are either co-occurring
(showing positive associations) or non co-occurring (showing negative associations). If cooccurring species pairs are less phylogenetically similar than non co-occurring pairs, competitive
interactions are inferred to be acting, whereas abiotic filtering is attributed if co-occurring
species are more phylogenetically related than non co-occurring pairs (eg. Winston 1995, Peres-

	
  

2	
  

	
  
Neto 2004, Tobias et al. 2013). However, it is important to also evaluate species’ traits
themselves as they may not be conserved, and thus patterns of trait distributions may indicate
different mechanisms of assembly and coexistence (Cavender-Bares et al. 2004).
Although a phylogenetic approach provides a quick and potentially informative way to
evaluate community assembly mechanisms, it is also valuable to assess the distributions of
species-specific traits themselves in assemblages (Swenson 2013). In particular, community
phylogenetics investigations assume that the traits associated with abiotic filtering and biotic
interactions are phylogenetically conserved (Webb 2000). While there has been evidence
presented for broad patterns of niche conservatism across taxa and spatial scales (Peterson et al.
1999, Wiens et al. 2010), there is also a great deal of evidence for trait divergence among closely
related species (Losos 2008). Thus it is preferable to test for phylogenetic signal in traits as
opposed to assuming trait conservatism is the norm (Losos 2008). Additionally, by evaluating
the distribution of traits themselves, we can identify which ecological aspect of a species is
acting to structure its distribution (Swenson 2013). It is even possible for some traits to be
structured by abiotic filtering and others to be structured by biotic interactions, as demonstrated
in multiple studies of assembly in plant (Weiher et al. 1998, Cavender-Bares et al. 2004,
Swenson and Enquist 2009) and animal (Tobias et al. 2013) communities. Lastly, it is important
to consider the phylogenetic and trait distributions in the context of species’ environmental
associations, as this provides information about the abiotically-defined aspects of a species niche
(Grinnell 1917). This allows one to determine whether or not traits are associated with
environmental variables, and thus strengthens our understanding of how abiotic filters are
operating to limit species distributions.

	
  

3	
  

	
  
In this study I integrate knowledge regarding the phylogenetic relatedness among species,
species-specific functional traits, and environmental niche occupancy to explore the mechanisms
of co-existence in the Pacific Basin plant genus, Pittosporum (Pittosporaceae), in New Zealand.
Instead of looking at whole plant communities, I evaluate the mechanisms contributing to cooccurrence and non co-occurrence of species pairs in a single group of closely related species.
This is advantageous as it is likely that closely related species are more strongly interacting given
their shared ancestry and thus putative ecological similarity. Further, most community
phylogenetics studies to date have evaluated distributions of species that did not evolve together
in the study region, that are highly divergent and thus have been unable to robustly link the
evolution of traits to co-existence due to poor taxonomic sampling. By investigating
mechanisms of co-existence in a single, densely sampled, and endemic group of species, we can
make stronger inferences about how evolutionary processes have contributed to current patterns
of species distributions and co-existence (Swenson 2013).
Specifically, I aim to evaluate (1) whether or not phylogenetic similarity predicts species
co-occurrence, and (2) how niche dissimilarity, as measured through trait differences and
environmental niche overlap, contributes to species co-occurrence. I predict that closely related
species will not co-occur, and that species that do co-occur will differ substantially along one or
more important niche axes.

	
  

4	
  

	
  
METHODS
Study System

This study examined the role that phylogenetic relatedness and niche differentiation – as
quantified through functional trait dissimilarity and environmental niche overlap – play in
structuring species distributions and co-occurrence in the genus Pittosporum (Pittosporaceae)
throughout New Zealand. The genus Pittosporum consists of approximately 200 species that are
widely distributed throughout the Pacific Basin, in tropical to temperate climates. Species are
found in disparate geographic regions such as Africa, Asia and Malesia, and Hawai’i, but the
largest centers of diversity are located in New Caledonia (50 species), and New Zealand (26
species) (Gemmill et al. 2002). The genus shows a tremendous amount of variation in
morphology and habitat preference within New Caledonia and New Zealand (Cooper 1956;
Gemmill et al. 2002), ranging from prostrate shrubs to subcanopy trees, and occupying coastal to
subalpine regions (Cooper 1956; Gemmill et al. 2002; Chandler et al. 2007).
The New Zealand Pittosporum represent an ideal group for investigating how
phylogenetic dissimilarity and niche differentiation contribute to species co-existence. The New
Zealand species are all endemic, and recent work has shown that the diversity of the genus
throughout the country is the result of only two colonization events (Gemmill et al. in prep).
This high level of endemism and presence of just two monophyletic groups mean that the
observed species diversity, and diversity in traits and habitat associations, is the result of
divergence within New Zealand in which all species arose in the study region and have
interacted with a similar suite of species and environments. This allows for the investigation not
only of present-day co-existence among species pairs, but also the role that evolution has played

	
  

5	
  

	
  
in generating the diversity in this genus. Lastly, the morphological diversity and the wide range
of habitats and climates occupied by species suggest that niche differentiation could be acting to
maintain co-existence.

Phylogenetic Inference

A recent phylogenetic treatment of the group indicates that the Pittosporum as a whole are
monophyletic (Gemmill et al. in prep). The New Zealand Pittosporum have arisen as a result of
two colonization events and subsequent radiations. The main radiation, consisting of 20 species,
arose approximately 20 million years (my) ago, with the putative ancestor colonizing from
Australia (Gemmill et al. in prep). The second radiation, consisting of just two species and one
subspecies, is approximately 10 my old and is hypothesized to have arisen as a result of a back
colonization event from New Caledonia (Gemmill et al. in prep). The present work will utilize
the chronogram produced by Gemmill et al. (in prep). The phylogeny was generated using the
ITS gene region and a Bayesian reconstruction. The phylogeny contains 22 of the 24 known
species of New Zealand Pittosporum (Figure 1).
For this study I evaluated the degree of differentiation between all pairwise species
combinations in the New Zealand Pittosporum. As the New Zealand Pittosporum are not
monophyletic, I conducted the analyses with and without the clade representing the second
colonization event. The results showed indistinguishable patterns so I only present data from the
analysis combining both clades.

	
  

6	
  

	
  
Species Functional Traits

A total of 12 functional traits known to be associated with various aspects of plant ecological
strategies were selected to evaluate the functional differentiation among species. Previous work

Figure 1. Phylogenetic relationships of the 24 New Zealand species of Pittosporum. The tree is
based off of ITS sequence data and was constructed using a Bayesian inference. The scale bar at
the bottom represents age since the origination of the genus.

	
  

7	
  

	
  
has demonstrated that many of these traits are often correlated with climatic gradients and are
therefore expected to relate to the climatic niches of species (eg. Reich et al. 1997, Wright et al.
2004, Moles and Westoby 2006, Swenson and Enquist 2007, Moles et al. 2009). In this way,
functional traits provide a link from the environment to the organism itself. Plant ecologists use
functional traits to evaluate how species are partitioning resources differently, through the
comparison of these ecologically important traits. In this study, four traits associated with wood
characteristics and hydraulics were measured (wood density, stem hydraulic conductance,
approximate maximum vessel length, and bark thickness), in addition to eight leaf traits
(chlorophyll content, leaf thickness, leaf density, leaf dry matter content, leaf mass per area
(LMA), leaf area, leaf nitrogen content and leaf carbon content).
The LMA and leaf nitrogen content are commonly referred to in the context of the ‘leaf
economics spectrum’ (Wright et al. 2004), a suite of traits that represent a continuum along
which all plant species lie (Reich et al. 1997; Wright et al. 2004). Leaves with high LMA tend to
have a longer lifespan than those with low LMA (eg. Reich et al. 1997, Wright et al. 2004). Leaf
nitrogen content relates to photosynthetic capacity, as nitrogen is the major component of many
compounds, such as RuBisCO, that are essential for, and constrain the rate of, photosynthesis
(Lambers et al. 1998). Leaf thickness and leaf density together can be used to calculate LMA,
and in turn are related to photosynthetic capacity (Niinemets 2001). Both traits, however,
represent independent axes of function along which species can differentiate. Increases in leaf
thickness are primarily driven by increases in numbers and lengths of photosynthetically active
mesophyll cells, thus increasing photosynthetic capacity (Pieters 1974, De Lillis and Valletta
1985, Korner et al. 1989, Niinemets 1999, Niienemets 2001). In contrast, increases in leaf
density are often correlated with decreases in photosynthetic capacity. Higher leaf density

	
  

8	
  

	
  
relates to mesophyll cells with thicker cell walls, leading to fewer photosynthetically active cells
that may also decrease gas exchange into the cells (Syvertsen et al. 1995). Higher density leaves
are thought to represent those species that are shade tolerant (Alvarez-Clare and Kitajima 2007)
and/or more resistant to herbivores (Kitajima and Poorter 2010). Leaf dry matter content
(LDMC) has also been shown to be an important trait representing variation along a resource-use
gradient (Wilson et al. 1999), where species with high LDMC are fast-growing in less resource
limited environments (Grime et al. 1997, Wilson et al. 1999, Garnier et al. 2001). Although
chlorophyll content of leaves is not included in the ‘leaf economics spectrum,’ it also positively
relates the photosynthetic capacity of leaves. Leaf carbon content represents the ability of leaves
to assimilate and store carbon (Elser et al. 2010), and is negatively correlated with vegetative
growth rate (Lambers et al. 1998). Lastly, leaf area relates to resource capture, with larger area
leaves being able to capture more sunlight. Larger leaves are also associated with lower
boundary layer conductance, meaning that there is less transfer of heat from the leaf to the
surrounding environment, resulting in an increase in leaf temperature (Lambers et al. 1998).
Increases in sunlight capture and leaf temperature lead to higher photosynthetic rates, but high
irradiance and high leaf temperatures may slow photosynthetic processes due to photoinhibition
and decreased stomatal conductance, respectively.
In response to the development of the ‘leaf economics spectrum,’ there has been a push to
evaluate wood traits in a similar trade-off framework. In this ‘wood economics spectrum,’
Chave et al. (2009) suggest that woody plants have developed and optimized anatomical traits
and physiological functions that lead to increased performance and survival in different habitats.
As with the ‘leaf economics spectrum’ it is suggested that there is a trade-off in wood traits that
either contribute to fast growth or to slower growth but increased survival. Species that have the

	
  

9	
  

	
  
ability to maximize resource transport and grow quickly typically have lower density wood and
high hydraulic conductance (Baralato et al. 2010), whereas species that are resistant to
disturbance and have greater mechanical support have higher density wood, but lower hydraulic
conductance and thus grow more slowly (Hacke et al. 2001). It is valuable to evaluate wood
traits in addition to leaf traits, as they have been shown to directly relate to plant growth and
survival (Poorter et al. 2009).
In this study I quantified wood density and hydraulic conductance, in addition to
approximate maximum vessel length (AMVL) and bark thickness of stems. As described above,
wood density and hydraulic conductance relate to competitive ability and resistance to
mechanical failure. In a study of 42 rainforest trees Poorter et al. (2009) found that growth rate
was positively correlated with hydraulic conductance and negatively related to wood density, but
wood density was positively related to survival. Although less emphasized in studies of
functional ecology, AVML and bark thickness are also indicators of different plant strategies.
Indeed, Comstock and Sperry (2000) argue that xylem conduit length, here referred to as AMVL,
represents a ‘neglected dimension’ for understanding the hydraulic architecture of plants and
thus a major performance axis. Their review of the literature suggests that variation in vessel
length is indicative of a trade-off between hydraulic conductance and safety. While long vessels
minimize resistance and thus maximize flow, trees with long vessels are more susceptible to
mechanical failure that results from cavitation or damage due to herbivores and pathogens
(Comstock and Sperry 2000). Bark thickness may also be important for plant performance in
different environments, and thus be an indicator of species niche differentiation. There is
evidence to suggest that thicker bark may be beneficial for species that are in habitats frequented
by fires (Hare 1965, Harmon 1984), but it may also be related to structural support and herbivore

	
  

10	
  

	
  
and pathogen resistance. In their study of wood and leaf functional traits of tree species in
French Guiana, Baralato et al. (2010) found a negative relationship between bark thickness and
wood density. They suggest that species with less dense wood may benefit from the structural
support of thicker bark, but more likely fast growing trees with less dense wood may be more
susceptible to herbivore and pathogen attacks, and thus have thicker bark to mitigate the effects
of this damage (Baralato et al. 2010, Paine et al. 2010).

Functional Trait Measurements

Leaf functional trait data were collected from 72 individuals and wood trait data were collected
from 67 individuals representing 22 of the 24 species of Pittosporum in New Zealand. Samples
were taken from saplings grown in the shadehouse at the University of Waikato in Hamilton,
New Zealand, from the Oratia Native Plant Nursery, or from native plant gardens in the
surrounding region. When possible, I collected wood and leaf samples from three or more
individuals per species, but for several uncommon species I was only able to collect from one to
two individuals.
Leaf traits were measured for three leaves per individual and these values were averaged
to obtain an individual mean trait value. All measurements were conducted on fresh, fully
mature sun leaves as suggested in Cornelissen et al. (2003). I measured traits on leaves both
with and without petioles, but as there was no significant difference in trait means between the
samples, I present data only from leaves with petioles. Leaf chlorophyll content was measured
using a SPAD-502 chlorophyll meter. Three measurements were taken on each leaf, avoiding
the mid-vein when possible. These values were averaged to obtain a single mean leaf value.

	
  

11	
  

	
  
Leaves from P. divaricatum and P. turneri were too small to obtain accurate measurements.
Leaf thickness was measured using a digital micrometer, and as with chlorophyll content, three
measurements were taken to obtain an average value for each leaf. Leaf area, defined as the
surface area of the upper lamina of the leaf, was measured using LI-3100 leaf area meter
(Lincoln, Nebraska USA). In order to calculate LMA, I weighed leaves that had been dried in a
70 degree Celsius oven for three or more days. For very small leaves, I determined the total area
and dry mass of 20 – 30 leaves and took an average to determine LMA. Leaf density was
calculated as LMA divided by leaf thickness (Kitajima and Poorter 2010), and leaf dry matter
content was calculated as the oven dry mass of a leaf divided by its fresh mass (Cornelissen et al.
2003, Perez-Harguindeguy et al. 2013). Mass-based leaf percent nitrogen and carbon were
analyzed at the Cornell Isotope Laboratory using a Thermo Delta V isotope ratio mass
spectrometer (IRMS) interfaced to a NC2500 elemental analyzer.
Wood traits were measured on a single stem per individual, and all traits were measured
on the same sample. When possible, measurements were taken on samples from lateral
branching stems, but for a few individuals with small diameter stems I sampled from the main
stem. Wood density was measured as dry mass per unit fresh volume of stems with the outer
cortex removed (Cornelissen et al. 2003, Perez-Harguindeguy et al. 2013). I quantified stem
volume using the dimensional method (Perez-Harguindeguy et al. 2013), with volume calculated
as π x L x (0.5D)2, where L is the length of the stem and D is its diameter. Stem sections
sampled were of regular shape making this an appropriate way to measure stem wood density.
To further ensure the accuracy of our values, diameter was measured using vernier calipers at
five points along the stem and averaged. Stems were dried at 100 degrees Celsius for five days
before being weighed on an analytical balance. Bark thickness was measured on both ends of

	
  

12	
  

	
  
the stem section and the values averaged. Bark was defined as in Perez-Haguindeguy et al.
(2013) as the tissue external to the wood or xylem, and includes the phloem. Approximate
maximum vessel length was measured using the ‘air method’ as described in Ewers and Fisher
(1989). Air was forced through a stem section that was placed underwater and was cut back
incrementally until bubbles emerged. While an interesting trait in itself, AMVL was also used to
determine the stem length of the samples measured for hydraulic conductance. In all but a few
instances samples used were longer than the AMVL, thus preventing the problem of
overestimation of hydraulic conductance. Samples measured that were not longer than the
AMVL had similar Ks values as other individuals of the same species, and thus were not
excluded from the analyses. Sapwood-specific conductivity was calculated as

Ks = J x L x A-1 x ΔΨ-1,

where J is the rate of water flow through the stem (kg s-1), L is the length of the stem (m), A is
the mean cross-sectional area of the stem (m2) and ΔΨ is the difference in hydrostatic pressure
(MPa). All measurements were taken following the protocol outlined in Perez-Haguindeguy et
al. (2013). In this gravitational method, a known pressure head is used to force a 10 mM KCl
solution through a sample with a known stem diameter and length for a known period of time.
The volume of the liquid forced through the stem was measured using a micro-pipette. All stem
samples were purged with filtered and degassed water for twenty minutes at a pressure of 60 100 kPa before Ks was measured to clear the stem section of potential embolisms.

	
  

13	
  

	
  
Species-Specific Climatic Niche Envelopes

The primary goal of this work was to determine the phylogenetic, trait, and environmental niche
dissimilarity among co-occurring and non co-occurring species in the genus Pittosporum. To
evaluate the degree of environmental niche differentiation I first estimated climatic niches for
each species using an ecological niche modeling approach. Specifically, I used the species
distribution modeling software, Maxent (ver. 3.3.3k; Phillips et al. 2006, Phillips and Dudik
2008), to estimate the optimal environmental envelope occupied by each species. Using habitat
and/or climatic data and species occurrence data as input variables, Maxent employs a maximum
entropy approach to predict species geographic ranges as determined by habitat suitability
(Phillips et al. 2006). Maxent is an ideal program to use in this study as it operates under the
assumption of presence-only data, performs well with relatively small sample sizes (Hernandez
et al. 2006, Pearson et al. 2007), and has proven a robust method in a recent comparison of
multiple habitat modeling methods (see Elith et al. 2006).
The Maxent algorithm requires geo-referenced occurrence data for each species. Georeferenced species occurrence records were obtained via the Global Biodiversity Information
Facility (GBIF) database (http://data.gbif.org) with additional data coming from individual
herbaria throughout New Zealand. Through the use of the GBIF database and herbaria records I
was able to utilize over 5000 geo-referenced occurrence records for 16 Pittosporum species of
the 24 species found in New Zealand. Although the number of occurrences varied for each
species, I only included those species that have greater than 10 occurrence records in our
analyses. The original geo-referenced data come from multiple institutions that have provided
their data to GBIF for re-distribution to the public. The majority of the occurrence data came

	
  

14	
  

	
  
from the National Vegetation Survey (https://nvs.landcareresearch.co.nz). This databank
comprises survey data from over 19000 permanent plots sampled over 50 years. I used the 19
BIOCLIM variables developed by Hijmans et al. (2005, http://www.worldclim.org), in addition
to altitude, as the environmental covariates defining the background data for the Maxent models.
Combining and summarizing temperature and precipitation data, these 19 bioclimatic variables
inform us about the climatic tolerances associated with different species. All environmental
layers in the model were defined using a raster grid size of 30 arc-seconds.

Quantifying Predicted Niche Occupancies

I used the results from the Maxent models to quantify the multi-dimensional climatic niche space
occupied by each species and to compare niche overlap between species. Incorporating the
locality data and environmental background data, the Maxent algorithm generates a predicted
geographic range map. The probability of occurrence of the focal species within any given grid
cell within the specified range is indicated using a heat map color scheme with warmer colors
indicating a higher probability of occurrence in a grid cell. It is possible to extract the “raw
probabilities” from the predicted Maxent range map/model. Following the methods developed
by Evans et al. (2009) I binned the raw probabilities assigned to each grid cell on the range map
to obtain a “predicted niche occupancy” (PNO; Evans et al. 2009) associated with each of the
environmental variables studied for each species. The resulting unit area histograms allow for
the visualization of the expected distribution of individuals of a given species along a specified
environmental axis. The PNOs were subsequently used to determine niche overlap of all
pairwise species comparisons for all environmental variables included in the Maxent model.

	
  

15	
  

	
  

Determining Pairwise Species Differentiation

The phylogenetic distance between each pair of species was calculated as the branch lengths
separating the species and calculated using the cophenetic function in the R package “ape” (v3.1
– 1; Paradis et al. 2004).
I quantified climatic niche overlap for species pairs using the I niche overlap metric
described in Warren et al. (2008). This similarity statistic, which is an adaptation of the
Hellinger distance metric (van der Vaart 1998, page 211), ranges from 0 (no niche overlap) to 1
(complete niche overlap) and is calculated as
	
  
1
𝐼 𝑝! , 𝑝! =   1 −    ∑! ( 𝑝!,! −    𝑝!,!   )!
2

where px and py denote the probability that a species will occur in location i. This metric is often
preferred as it carries no biological assumptions about the population densities of species in a
given habitat unlike other metrics that are commonly used, such as Schoener’s D statistic
(Warren et al. 2008). Additionally, this metric proves especially useful in this study as I can use
the output from the PNO plots to determine the probability of occurrence in each habitat for each
species. I also determined overall niche overlap among species pairwise comparisons by
evaluating niche overlap of entire environmental niche models. All calculations were carried out
using the “phyloclim” package in R (v0.9-4; R Development Core Team 2013).
Functional trait differentiation for all pairwise species comparisons was determined for
each trait individually by calculating the Euclidean distance between two species’ traits using the
	
  

16	
  

	
  
“stats” package in R (v3.0.2; R Development Core Team 2013). I also conducted a principal
components analysis (PCA) to summarize the multivariate associations among all functional
traits and used the first four components in our evaluation of trait differentiation among cooccurring, non co-occurring, and randomly distributed species pairs. Lastly I calculated a
Gower’s dissimilarity to estimate the overall dissimilarity in traits among species pairs. I used
the Gower’s dissimilarity index on the PCA output instead of on all individual traits combined as
the PCA removes the redundancy in traits. The Gower calculation was implemented in the “FD”
package in R (v1.0-11; R Development Core Team 2013).

Determining Co-occurrence vs. Non Co-occurrence

To determine whether species pairs are co-occurring more than expected, non co-occurring more
than expected, or randomly distributed I first binned all geo-referenced occurrence data into 0.25
x 0.25 degree (latitude and longitude) grid cells and calculated Schoener's Co-occurrence index
(Cij) from the present-absence data matrix (Schoener 1968). The observed Cij values were then
compared to a distribution of 999 random Cij values generated by implementing an independent
swap null model. Briefly, the independent swap randomizes the presence-absence matrix of
species while maintaining the species richness observed in a map grid cell and the occupancy
rates of individual species across map grid cells in the study system. I then calculated whether
each pair of species co-occurs more or less than expected given the null expectation using the
ranks of the observed Cij values in the null distribution of 999 random Cij values. Cooccurrence calculations were implemented in the R package “picante” (v1.6-2; R Development
Core Team 2013).

	
  

17	
  

	
  
Statistical Analyses

To test for significant phylogenetic, functional trait, and environmental niche dissimilarity
between co-occurring, random, and non co-occurring species pairs, I implemented separate linear
mixed effects models (LMMs) using all pairwise comparisons as data points (n = 124). LMMs
are useful for examining relationships among data points that exhibit non-independence and
represent multiple comparisons (Faraway 2005). Using this approach, I assigned occurrence
designation as a fixed factor but was able to account for non-independence of data points by
including species 1 and species 2 as random effects. I used Honest Significant Difference Tukey
tests to determine post-hoc differences among occurrence categories. Calculations were carried
out using the “lme4” (v1.0-6; R Development Core Team 2013) and “multcomp” (v1.3-2; R
Development Core Team 2013) packages in R.
Table 1. Results from the LMM analysis of trait dissimilarity differences among co-occurring,
non co-occurring and randomly distributed species pairs.
	
  	
  

	
  	
  

	
  	
  

Co-occurring and
Non Co-occurring

Trait

	
  	
  

	
  	
  

Co-occurring and
Random

	
  	
  

Non Co-occurring
and Random

z - score

p -value

z - score

p -value

z - score

p -value

Chlorophyll Content

0.551

0.845

-0.530

0.856

-1.184

0.46

Leaf Thickness

0.716

0.754

-0.635

0.801

-1.445

0.317

Leaf Density

0.525

0.859

-0.171

0.984

-0.757

0.729

Leaf Dry Matter Content

0.417

0.908

-1.057

0.540

-1.534

0.274

Leaf Mass per Area (LMA)

0.615

0.811

0.777

0.717

0.105

0.994

Leaf Area

1.796

0.170

1.081

0.525

-0.934

0.618

Leaf Carbon Content

1.048

0.546

0.484

0.879

-0.667

0.782

Leaf Nitrogen Content

4.080

< 0.001

3.541

0.001

-0.899

0.640

Approx. Max Vessel Length

8.780

< 0.001

4.275

< 0.001

-5.674

0.001

Hydraulic Conductance

0.272

0.960

1.396

0.342

1.151

0.482

Wood Density

-0.494

0.874

-0.062

0.998

0.487

0.887

Bark Thickness

1.886

0.142

0.806

0.699

-1.275

0.409

	
  

	
  

	
  	
  

	
  

	
  

	
  

18	
  

	
  

	
  

	
  

	
  
RESULTS

I evaluated a total of 120 species pairwise comparisons to determine if species co-occurred more
than expected based on a null distribution, did not co-occur more than expected, or were
distributed randomly with respect to each other. Species were considered co-occurring if they
had a Cij value that fell in the upper tail of the null distribution (0.975 and above) and non cooccurring when they fell in the lower tail of the distribution (0.025 and below). Thirty-three
species pairs were described as co-occurring, 32 were described as non co-occurring, and 55
were described as randomly distributed.

Phylogenetic Relatedness and Co-occurrence

I found no significant difference in mean pairwise phylogenetic distance between co-occurring,
non co-occurring, and randomly distributed species (Figure 2), indicating that phylogenetic
relatedness does not predict whether or not species co-occur in this genus. The same result was
obtained when the clade consisting of P. pimelioides ssp. pimelioides and P. cornifolium (ie: the
Table 2. Results from the LMM analysis of trait dissimilarity of the first four principal
components among co-occurring, non co-occurring and randomly distributed species pairs.
	
  	
  
	
  

	
  	
  

Co-occurring and
Non Co-occurring

Trait

	
  	
  

	
  	
  

Co-occurring and
Random

	
  	
  
Non Co-occurring
and Random

z - score

p -value

z - score

p -value

z - score

p -value

Principal Component 1

1.659

0.221

0.617

0.810

-1.230

0.434

Principal Component 2

-0.149

0.988

-0.901

0.639

-0.747

0.735

Principal Component 3

-0.267

0.961

1.640

0.228

1.956

0.123

Principal Component 4

3.185

0.004

1.834

0.158

-1.752

0.185

Multivariate PCA

2.956

0.009

1.240

0.429

-2.032

0.104

	
  

	
  

	
  	
  

	
  

	
  

	
  

19	
  

	
  

	
  

	
  

2.0

A

A

C

NC

R

B

C

NC

R

1.0
0.5
0.0

	
  

0.2

0.4

0.6

A

0.0

Niche	
  Overlap

A

1.5

	
  

0.8

Phylogenetic	
  Distance

	
  

A

B

AB

2

3

4

5

6

7

	
  

C

NC

R

	
  

1

Trait	
  Dissimilarity

C

	
  
Figure 2. The mean phylogenetic distance,
environmental niche overlap and trait
dissimilarity among co-occurring (C), non
co-occurring (NC) and randomly distributed
(R) species pairs. The environmental niche
overlap and trait dissimilarity are
represented by a multivariate summary of all
variables and traits, respectively.

	
  

20	
  

	
  
second colonization event) was included and excluded. Thus I only show the results with the
clade included.

Niche Overlap and Co-occurrence

There was a significant difference in degree of overall niche overlap between occurrence
classifications, with those that co-occur having very high niche overlap (Figure 2). When
evaluating overlap of individual environmental variables, co-occurring species always had higher
niche overlap than non co-occurring species, but there were some instances in which differences
were not significantly different than those species pairs that were randomly distributed (Figure
3).

	
  

21	
  

	
  

B

C

C

NC

R

0.8
0.6
0.4
0.2

	
  

0.75 0.80 0.85 0.90 0.95 1.00

A

Isothermality
Niche	
  Overlap

	
  

0.0

Niche	
  Overlap

1.0

Mean	
  Annual	
  Temp.
	
  

A

C

1.0

R

0.6
0.4

B

C

C

NC

R

0.8

NC

A

0.6

C

	
  

0.4

C

R

	
  

0.0

	
  

0.2

B

NC

B

Altitude
Niche	
  Overlap

A

0.8

	
  

0.2

Niche	
  Overlap

1.0

Mean	
  Annual	
  Precip.

B

Figure 3. Mean environmental niche overlap among co-occurring (C), non co-occurring (NC)
and randomly distributed (R) species pairs for four independent environmental variables: mean
annual temperature (Bio1), isothermality (Bio3), mean annual precipitation (Bio12), and altitude.

Trait Dissimilarity and Co-occurrence

I evaluated trait dissimilarity for all 12 functional traits independently, in addition to evaluating
dissimilarity along individual principal component axes and Euclidean distance in multivariate
PCA space. For the principal components analysis, I used the first four components. These four
components explained 92 percent of the variation, with each component explaining between 12

	
  

22	
  

	
  

C

NC

R

4
3
4

A

2

3

B

1

Trait	
  Dissimilarity

B

	
  

	
  

0

5
4
3
1

2

A

0

Trait	
  Dissimilarity

A

NC

NC

PC	
  Axis	
  4
A

C

C

2

R

PC	
  Axis	
  3
	
  

A

1

Trait	
  Dissimilarity

1

2

3

4

A

A

0

A

A
5

A

	
  

5

6

PC	
  Axis	
  2

0

Trait	
  Dissimilarity

PC	
  Axis	
  1

C

R

NC

R

Figure 4. Mean trait dissimilarity among co-occurring (C), non co-occurring (NC) and
randomly distributed (R) species pairs for the first four principal components. Components one
through three represent “leaf axes” whereas component four represents a “wood axis.”

and 35 percent of the variation (see Appendix 1, Tables A1 and A2 for summary of PCA).
Principal components one through three can be described as “leaf axes” whereas principal
component four had wood density and approximate maximum vessel length loading most
heavily, and thus I describe this component as a “wood axis.” There was no significant
	
  

23	
  

	
  

R

NC

10
8
6
4

NC

R

1.2
0.8

B

	
  

B

A

R

	
  

	
  
C

Trait	
  Dissimilarity

120
20 40 60 80

A

A

0

Trait	
  Dissimilarity

A

C

Leaf	
  Nitrogen	
  Content

LMA
	
  

	
  

	
  
NC

A

2

0.10
0.00

C

C

0

A

B

	
  

0.4

A

0.0

A

0.20

Trait	
  Dissimilarity

	
  

Trait	
  Dissimilarity

Approx.	
  Max.	
  Vessel	
  Length

Wood	
  Density

C

NC

R

Figure 5. Mean trait dissimilarity among co-occurring (C), non co-occurring (NC) and
randomly distributed (R) species pairs for four independent functional traits: wood density,
approximate maximum vessel length, LMA, and leaf nitrogen content.

difference in mean trait dissimilarity between co-occurring, non co-occurring, and randomly
distributed species pairs for components one through three, the leaf axes (Figure 4, Table 2), but
there was a significant difference between co-occurring and non co-occurring species pairs for
component four, the wood axis (Figure 4, Table 2). Co-occurring species pairs were more
similar than non co-occurring species pairs for this axis. The same result was observed when
evaluating dissimilarity in the multivariate principal component axis (Figure 1, Table 2). When
	
  

24	
  

	
  
evaluating the traits independently, approximate maximum vessel length and leaf nitrogen
content were the only traits that were significantly different between co-occurring and non cooccurring species pairs, again, with these traits being more similar in co-occurring versus non cooccurring species pairs (Figure 5, Table 1).

	
  

25	
  

	
  
DISCUSSION

In this study I assessed mechanisms of co-existence in the New Zealand Pittosporum.
Specifically, I evaluated whether phylogenetic relatedness, environmental niche overlap, and
functional trait dissimilarity could explain patterns of species co-occurrence and non cooccurrence. I found that there was no difference in mean phylogenetic distance between cooccurring and non co-occurring species (Figure 1). Co-occurring species had a higher degree of
niche overlap than non co-occurring species (Fig. 1), but the relative difference in mean niche
overlap between the groups differed based on the environmental axis examined (Figure 2).
Lastly, when evaluating all traits simultaneously, co-occurring species had a lower degree of trait
dissimilarity than non co-occurring species (Figure 1). However, when the traits were evaluated
independently we see that approximate maximum vessel length and leaf nitrogen content are the
only traits that exhibit patterns of low trait dissimilarity in co-occurring species. There was no
difference in trait dissimilarity between co-occurring and non co-occurring species with the other
traits considered (Table 1). As leaf nitrogen content was not included in the principal
components analysis, this suggests that approximate maximum vessel length is a major driver
contributing to the pattern of trait similarity in co-occurring species when evaluating the traits
together. Below I discuss possible explanations behind the observed results in the context of
species co-occurrence mechanisms in the New Zealand Pittosporum.

	
  

26	
  

	
  
Phylogenetic Relatedness and Species Associations

I expected that due to shared ancestry closely related species would be more ecologically similar
to each other than to distantly related species (Darwin 1959), and thus not be able to co-occur.
Instead, the results show no difference in phylogenetic relatedness among co-occurring and non
co-occurring species. One explanation for this result is that the assumption of niche
conservatism is not upheld in this group. That is, the traits that are important for co-existence
may not have phylogenetic signal and thus closely related species are actually less similar than
expected based on phylogenetic relatedness. This lack of trait phylogenetic signal and its
implications for species co-existence and community assembly have been observed and explored
in other study systems including Floridian oak (Cavender-Bares et al. 2004) and Caribbean
Anolis communities (Losos et al. 2003) where the phylogenetic patterns of co-existence can only
be understood in the context of convergent trait and environmental niche evolution. I therefore
will now discuss patterns of environmental niche and functional trait similarity observed in the
genus Pittosporum.

Environmental Niche Similarity and Co-Occurring Species

It is not surprising that I observed high climatic niche overlap among co-occurring species, and
lower climatic niche overlap among non co-occurring and randomly distributed species, given
that the co-occurring species occupy the same environments. It is interesting to note, however,
that for several variables there was still a relatively high degree of niche overlap for non cooccurring species. This suggests that either these variables represent a strong environmental

	
  

27	
  

	
  
filter for the entire genus, in that there is only a certain range of environmental space that these
species can occupy, or that all species occupy the whole range of environmental space and thus
there is no differentiation along these niche axes. In either scenario, this lack of niche
differentiation might be limiting the number of species in the genus through lack of
diversification into different habitats, and also the ability of these species to co-exist across New
Zealand.

Trait Convergence in Co-Occurring Species

I expected that regardless of phylogenetic dissimilarity, co-occurring species would be less
similar to each other with respect to their traits than non co-occurring species as a result of biotic
interactions. Instead, I found that co-occurring species are either more similar in traits, or that
there is no difference in trait dissimilarity between co-occurring and non co-occurring species
pairs. The high degree of trait similarity for approximate maximum vessel length and leaf
nitrogen content is likely due to abiotic filtering where only a small range of trait values are
suitable for the given abiotic environment. High trait similarity among co-occurring plant
species has been previously shown suggesting an important role for abiotic filtering in
community assembly (e.g. Weiher et al. 1998, Grime 2006, Kraft et al. 2008, Swenson and
Enquist 2009).
In studies of plant functional traits, it is recognized that species tend to have trait values
that fall along a ‘leaf economics spectrum’ (Wright et al. 2004) as well as a ‘wood economics
spectrum’ (Chave et al. 2009). The extremes of this spectrum can be described with respect to
different growth and survival strategies. On one end, species are highly conservative in their

	
  

28	
  

	
  
traits, meaning that they have lower growth rates but higher survival. On the other end of the
spectrum, species are more acquisitive, and tend to have higher growth rates but lower survival.
These plant strategies, and the traits associated with each, have been explored in theoretical
models (Sterck et al. 2011) as well as in many empirical studies (eg. Reich et al. 1999, Wright et
al. 2004, Ordoñez et al. 2009, Markesteijn et al. 2011). In the case of leaf nitrogen content and
approximate maximum vessel length in the present study, I believe that species are exhibiting
different growth and survival strategies in differing environments, leading to the observed
pattern of ecological trait convergence among co-occurring species.
Leaf nitrogen content is positively related to photosynthetic capacity (Reich et al. 1999).
In this study, species with high leaf nitrogen content are often found in cooler, relatively wet
habitats, whereas species with low leaf nitrogen content are more often found in warmer, drier
habitats (data not shown). This suggests that species in the cooler, wetter environments are
investing more energy in photosynthetic machinery in order to maximize growth rates. In
contrast, species in the drier, warmer habitats may be restricted in their photosynthetic capacity.
Drought limited species are prone to hydraulic failure if there is too much stomatal conductance,
as this increases the negative pressure in the xylem and can thus lead to cavitation (Tyree and
Sperry 1989, Sperry and Sullivan 1992). It is noted, however, that other studies have shown no
relationship between leaf nitrogen content and climatic variables (Ordoñez et al. 2009), and even
a negative correlation between leaf nitrogen content and precipitation (Cunningham et al. 1999),
suggesting that this is a trait-environment relationship that needs further investigation.
Approximate maximum vessel length represents a tradeoff between water conductance
and safety, and thus an acquisitive and conservative strategy, respectively. Species with long
vessels maximize hydraulic conductance by minimizing resistance, but they are also more

	
  

29	
  

	
  
susceptible to cavitation and mechanical failure (reviewed in Comstock and Sperry 2000,
Markesteijn et al. 2011). Cavitation is caused by both drought and freezing, as drought may
induce air seeding in the conduit, and freeze-thaw events can result in embolisms (Tyree and
Sperry 1989, Sperry and Sullivan 1992). In this study system those species with the shortest
vessels occur in wetter habitats, suggesting that drought is not a factor limiting distributions.
However, they are also found in the cooler habitats, which are subjected to freezing temperatures
during the coldest months, suggesting that temperature could be an abiotic filter causing the
observed convergence in traits in co-occurring species. However, more information about vessel
diameter will be needed to make any conclusive inferences about the hydraulic strategies of each
species.
It has also been suggested that competitive interactions can lead to ecological trait
convergence among co-existing species when performance differences are small (Chesson 2000,
Mayfield and Levine 2010). However, this is likely to occur on very local scales and in a
homogenous environment (HilleRisLambers 2012). The designation of co-occurrence versus
non co-occurrence in this study was defined at a scale of 0.25 X 0.25 longitude and latitude grid
cells. While this scale is useful in describing patterns of co-occurrence, it does not define local
co-existence. Additionally, New Zealand is climatically heterogeneous therefore it is unlikely
that co-occurring species were restricted to identical environments. Lastly, these two traits show
strong correlations with both precipitation and temperature variables (data not shown) and thus I
conclude that it is most likely that abiotic filtering effects are acting to structure species and trait
distributions in these environments.

	
  

30	
  

	
  
Trait Dissimilarity and Species Associations

Interestingly, all individual traits, with the exceptions of approximate maximum vessel length
and leaf nitrogen content, showed no difference in the degree of dissimilarity between cooccurring and non co-occurring species. There are several possible explanations for this result.
First, the level of trait dissimilarity observed among co-occurring species may be enough to
allow for niche differentiation. One of the limitations of this, and many other functional trait
studies, is that it is difficult to identify a minimum threshold of dissimilarity needed for coexistence. Thus, the trait dissimilarity observed here could be great enough to explain cooccurrence. This does not, however, explain why co-occurring and non co-occurring species
have the same mean trait dissimilarity. I suggest that two mechanisms are acting to cause the
observed results. As discussed above, the degree of dissimilarity in co-occurring species could
be driven by biotic interactions and niche differentiation, as has been observed in other plant
(Cavender-Bares et al. 2004) and animal (Tobias et al. 2013) systems. In contrast, the
dissimilarity in traits among non co-occurring species could be driven by environmental
differences in the habitats in which the species occur. Both scenarios could lead to similar trait
dissimilarities between co-occurring and non co-occurring species pairs.
Another possible explanation for the observed results is that many of the traits I measured
can be plastic (reviewed in Berg and Ellers 2010), and thus may differ within a species among
varying environments. Natural populations likely grow in environments different from the
environments in which the study individuals were grown. If I was to take trait measurements
from species in these varying environments, we may see a very different pattern of trait
dissimilarity among co-occurring species. Traits may be more overdispersed if competitive

	
  

31	
  

	
  
interactions are at play (see Burns and Strauss 2012), but traits could also be more convergent
owing to the prevalent occurrence of trait-environment relationships (Ordoñez et al. 2009).
Lastly, it is possible that I am missing some important axis or axes of trait differentiation
that are mediating co-occurrence. This could be especially important in this system, as I am
looking at species in a single genus as opposed to species across a broader taxonomic scale. The
species are recently diverged, and so have had less time to accumulate trait and environmental
niche differentiation. It may be possible that these closely related species are maintaining
continued co-existence by differing along a reproductive trait axis, such as flower type, flower
timing, and/or seed size.

Community Assembly in the New Zealand Pittosporum

Historically researchers have used lists of species, knowledge of their distributions, and a null
model approach to try and understand the processes responsible for structuring species
assemblages in general, and species co-occurrence, specifically (eg. Diamond 1975). In recent
years the inclusion of trait and phylogenetic data has greatly enhanced our power to elucidate the
actual mechanisms acting that lead to observed community assemblages (eg. Weiher et al. 1998,
Gotelli and Ellison 2002, Cavender-Bares et al. 2004, Peres-Neto 2004, Tobias et al. 2013). The
results from this study suggest that neither phylogenetic dissimilarity nor dissimilarity in the
measured traits alone predict patterns of species co-occurrence in the New Zealand Pittosporum.
Although we do see trait clustering in co-occurring species for leaf nitrogen content and
approximate maximum vessel length, we do not observe any pattern of high trait dissimilarity
among co-occurring species. It will be interesting to reevaluate this study system with the

	
  

32	
  

	
  
inclusion of additional traits that may exhibit patterns of trait divergence as has been observed in
other systems (Losos et al. 2003, Cavender-Bares 2004, Tobias et al. 2013).
Additionally, the inclusion of evolutionary history into this and other studies of
mechanisms of co-occurrence will inform us about how species and trait diversification have led
to patterns of diversity and distributions we observe today. Preliminary analysis of the data in
this study suggests that many of the functional traits evaluated are not conserved (data not
shown) suggesting that phenotypic convergence over time is potentially confounding the signal
of phylogenetic dissimilarity among co-occurring species. Information about the evolutionary
history of a radiation can also inform us about how (ie: the magnitude and timing of
diversification) species have diverged along functional trait and environmental niche axes as they
radiated across New Zealand. Lastly, I suggest that additional studies of trait evolution and
species co-existence be conducted across different lineages to evaluate whether the same patterns
are observed, and thus the same mechanisms are acting to structure the spatial distributions of
species.

	
  

33	
  

	
  

APPENDIX

	
  

34	
  

	
  
Table A1. The percent of variation explained by each of the four first components from the
principal components analysis conducted on the 12 functional traits investigated in this study.
Comp 1
0.3522

Comp 2
0.2645

Comp 3
0.1781

Comp 4
0.1222

TOTAL
0.9170

Table A2. The loadings from the principal components analysis conducted on the 12 functional
traits investigated in this study.
Trait
Approx. Max Vessel Length
Bark Thickness
Chlorophyll Content
Hydraulic Conductance
Leaf Density
Leaf Dry Matter Content
Leaf Mass per Area
Leaf Thickness
Wood Density
Leaf Area

	
  

Comp 1
-0.288
-0.458

Comp 2
-0.172

-­‐	
  

0.517
-0.373
-0.336
-0.464
0.103
0.333

-0.267
0.131
0.135
-0.405
-0.431
0.332
-0.37

-­‐	
  

Comp 3
-0.26
0.261
0.289
-0.266
0.58
0.417
0.407

Comp 4
0.599
-­‐	
  

-0.183
-0.336
-­‐	
  
-­‐	
  

-­‐	
  

-­‐	
  

0.167
0.125
0.662

-0.321

-0.167

	
  -­‐	
  

35	
  

-­‐	
  

	
  
Table A3. Results from the LMM analysis of environmental niche overlap differences among cooccurring, non co-occurring and randomly distributed species pairs.
	
  	
  

	
  	
  

	
  	
  

Co-occurring and
Non Co-occurring

Environmental Variable

	
  

	
  	
  

	
  	
  

	
  	
  

Co-occurring and
Random

	
  	
  

Non Co-occurring
and Random

z - score

p -value

z - score

p -value

z - score

p -value

Bio1 - Annual Mean Temp.

-7.687

<0.001

-3.835

<0.001

4.668

<0.001

Bio2 - Mean Diurnal Range

-4.728

<0.001

-1.606

0.242

3.624

<0.001

Bio3 - Isothermality

-2.904

0.010

-3.216

0.004

-0.028

1.000

Bio4 - Temp. Seasonality

-6.494

<0.001

-3.623

0.001

3.528

0.001

Bio5 - Max. Temp. of Warmest Month

-6.218

<0.001

-2.712

0.018

4.160

<0.001

Bio6 - Min. Temp. of Coldest Month

-7.588

<0.001

-4.259

<0.001

4.114

0.001

Bio7 - Temp. Annual Range

-7.366

<0.001

-3.476

0.002

4.682

<0.001

Bio8 - Mean Temp. of Wettest Quarter

-7.865

<0.001

-4.024

0.002

4.669

<0.001

Bio9 - Mean Temp. of Driest Quarter

-7.332

<0.001

-3.971

<0.001

4.109

<0.001

Bio10 - Mean Temp. of Warmest Quarter

-5.994

<0.001

-3.653

<0.001

2.943

0.009

Bio11 - Mean Temp. of Coldest Quarter

-8.022

<0.001

-4.048

<0.001

4.852

<0.001

Bio12 - Annual Precipitation

-7.577

<0.001

-3.733

<0.001

4.652

<0.001

Bio13 - Precip. of Wettest Month

-7.797

<0.001

-3.949

<0.001

4.671

<0.001

Bio14 - Precip. of Driest Month

-4.687

<0.001

-1.984

0.116

3.193

0.004

Bio15 - Precip. Seasonality

-4.249

<0.001

-2.685

0.020

1.996

0.113

Bio16 - Precip. of Wettest Quarter

-4.339

<0.001

-2.185

0.074

2.604

0.025

Bio17 - Precip. of Driest Quarter

-6.104

<0.001

-3.421

0.002

3.319

0.003

Bio18 - Precip. of Warmest Quarter

-4.280

<0.001

-2.833

0.013

1.881

0.144

Bio19 - Precip. of Coldest Quarter

-5.015

<0.001

-1.818

0.163

3.727

0.001

Altitude

-6.300

<0.001

-3.133

0.005

3.835

<0.001

Multivariate Niche Overlap

-8.945

<0.001

-5.922

<0.001

3.957

<0.001

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

	
  

36	
  

	
  

LITERATURE CITED

	
  

37	
  

	
  
LITERATURE CITED

Alvarez-Clare, S., and K. Kitajima. 2007. Physical defence traits enhance seedling survival of
neotropical trees. Functional Ecology 21:1044-1054.
Baralato, C., C. E. T. Paine, L. Poorter, J. Beauchene, D. Bonal, A. Domenach, B. Hérault, S.
Patiño, J. Roggy, and J. Chave. 2010. Decoupled leaf and stem economics in rain forest trees.
Ecology Letters 13:1338-1347.
Berg, M. P., and J. Ellers. 2010. Trait plasticity in species interactions: a driving force of
community dynamics. Evolutionary Ecology 24:617-629.
Burns, J. H., and S. Y. Strauss. 2012. Effects of competition on phylogenetic signal and
phenotypic plasticity in plant functional traits. Ecology 93:S126-S137.
Cavender-Bares, J., D. D. Ackerly, D. A. Baum, and F. A. Bazzaz. 2004. Phylogenetic
overdispersion in Floridian oak communities. American Naturalist 163: 823-843.
Chandler, G. T., G. M. Plunkett, S. M. Pinney, L. W. Cayzer, and C. E. C. Gemmill. 2007.
Molecular and morphological agreement in Pittosporaceae: phylogenetic analysis with
nuclear ITS and plastid trnL-trnF sequence data. Australian Systematic Botany 20: 390-401.
Chave, J., D. Coomes, S. Jansen, S. L. Lewis, N. G. Swenson, and A. E. Zanne. 2009. Towards a
worldwide wood economics spectrum. Ecology Letters 12: 351-366.
Chesson, P. 2000. Mechanisms of maintenance of species diversity. Annual Review in Ecology
and Systematics 31: 343-366.
Comstock, J. P., and J. S. Sperry. 2000. Theoretical considerations of optimal conduit length for
water transport in vascular plants. New Phytologist 148:195-218.
Connor, E. F., and D. Simberloff. 1983. Interspecific competition and species co-occurrence
patterns on islands: null models and the evaluation of evidence. Oikos 41: 455-465.
Cooper, R. C. 1956. The Australian and New Zealand species of Pittosporum. Annual Missouri
Botanical Garden 43:87-187.
Cornelissen, J. H. C., S. Lavorel, E. Garnier, S. Diaz, N. Buchmann, D. E. Gurvich, P. B. Reich,
H. ter Steege, H. D. Morgan, M. G. A. van der Heijden, J. G. Pausas and H. Poorter. 2003. A
handbook of protocols for standardised and easy measurement of plant functional traits
worldwide. Australian Journal of Botany 51: 335-380.
Cunningham, S. A., B. Summerhayes, and M. Westoby. 2009. Evolutionary divergences in leaf
structure and chemistry, comparing rainfall and soil nutrient gradients. Ecology 69:569-588.
	
  

38	
  

	
  

Darwin, C. 1859. The Origin of Species by Means of Natural Selection. London: Murray.
De Lillis, M., and A. Valletta. 1985. Leaf anatomy of evergreen woody plants from Canary
Islands and Mediterranean Basin. A comparison. Annali di Botanica 43:209-225.
Diamond, J. M. 1975. Assembly of species communities. In Ecology and Evolution of
Communities, ed. M. L. Cody and J. M. Diamond, pp. 342-444. Cambridge, MA: Harvard
University Press.
Diamond, J. M., and M. E. Gilpin. 1982. Examination of the “null” model of Connor and
Simberloff for species co-occurrence on islands. Oecologia 52:64-74.
Elith, J., C. H. Graham, R. P. Anderson, M Dud ́ık, S. Ferrier, A. Guisan, R. J., Hijmans, F.
Huettmann, J. R. Leathwick, A. Lehmann, J. Li, L. G. Lohmann, B. A. Loiselle, G. Manion,
C. Moritz, M. Nakamura, Y. Nakazawa, J. McC. Overton, A. T. Peterson, S. J. Phillips, K. S.
Richardson, R. Scachetti-Pereira, R. E. Schapire, J. Soberón, S. Williams, M. S. Wisz, and N.
E. Zimmermann. 2006. Novel methods improve prediction of species’ distributions from
occurrence data. Ecography 29:129-151.
Elser, J. J., W. F. Fagan, A. J. Kerkhoff, N. G. Swenson, and B. J. Enquist. 2010. Biological
stoichiometry of plant production: metabolism, scaling and ecological response to global
change. New Phytologist 186:593-608.
Emerson, B. C., and R. G. Gillespie. 2008. Phylogenetic analysis of community assembly and
structure over space and time. Trends in Ecology and Evolution 23 (11): 619-630.
Evans, M. E. K., S. A. Smith, R. S. Flynn, and M. J. Donoghue. 2009. Climate, niche evolution,
and diversification of the “Bird-Cage” Evening Primroses (Oenothera, sections Anogra and
Kleinia. American Naturalist 173:225-240.
Ewers, F. W., and J. B. Fisher. 1989. Techniques for measuring vessel lengths and diameters in
stems of woody plants. 76:645-656.
Faraway, J. J. 2005. Extending the linear model in R: generalized linear, mixed effects and
nonparametric regression models. Chapman and Hall/CRC, Boca Raton, FL.
Garnier, E., B. Shipley, C. Roumet, and G. Laurent. 2001. A standardized protocol for the
determination of specific leaf area and dry matter content. Functional Ecology 15:668-695.
Gemmill, C. E. C., G. J. Allan, W. L. Wagner, and E. A. Zimmer. Evolution of insular Pacific
Pittosporum (Pittosporaceae): origin of the Hawaiian radiation. 2002. Molecular
Phylogenetics and Evolution 22: 31-42.
Gotelli, N. J. 2000. Null model analysis of species co-occurrence patterns. Ecology 81: 26062621.
	
  

39	
  

	
  

Gotelli, N. J., and A. M. Ellison. 2002. Assembly rules for New England ant assemblages. Oikos
99:591-599.
Grime, J. P. 2006. Trait convergence and trait divergence in herbaceous plant communities:
mechanisms and consequences. Journal of Vegetation Science 17:255-260.
Grime, J. P., K. Thompson, R. Hunt, J. G. Hodgson, J. H. C. Cornelissen, I. H. Rorison, G. A. F.
Hendry, T. W. Ashenden, A. P. Askew, S. R. Band, R. E. Booth, C. C. Bossard, B. D.
Campbell, J. E. L. Cooper, A. W. Davison, P. L. Gupta, W. Hall, D. W. Hand, M. A.
Hannah, S. H. Hiller, D. J. Hodkinson, A. Jalili, Z. Liu, J. M. L. Mackey, N. Matthews, M.
A. Mowforth, A. M. Neal, R. J. Reader, K. Reiling, W. Ross-Fraser, R. E. Spencer, F. Sutton,
D. E. Tasker, P. C. Thorpe, and J. Whitehouse. 1997. Integrated screening validates primary
axes of specialisation in plants. Oikos 79:259-281.
Grinnell, J. 1917. The niche-relationship of the California thrasher. Auk 34: 427-433.
Hacke, U. G., J. S. Sperry, W. T. Pockman, S. D. Davis, and K. A. McCulloch. 2001. Trends in
wood density and structure are linked to prevention of xylem implosion by negative pressure.
Oecologia 126:457-461.
Harmon, M. E. 1984. Survival of trees after low-intensity surface fires in Great Smoky
Mountains National Park. Ecology 65:796-802.
Hare, R. C. 1965. Contribution of bark to fire resistance of southern trees. Journal of Forestry
63:248-251.
Hernandez, P. A., C. H. Graham, L. L. Master, and D. L. Albert. 2006. The effect of sample size
and species characteristics on performance of different species distribution modeling
methods. Ecography 29:773-785.
HilleRisLambers, J., P. B. Adler, W. S. Harpole, J. M. Levine, and M. M. Mayfield. 2012.
Rethinking community assembly through the lens of co-existence theory. Annual Review in
Ecology, Evolution and Systematics 43: 227-248.
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.
Hutchinson, G. E. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals?
American Naturalist 95: 137-145.
John, R., J. W. Dalling, K. E. Harms, J. B. Yavitt, R. F. Stallard, M. Mirabello, S. J. Hubbell, R.
Valencia, H. Navarrete, M. Vallejo, and R. B. Foster. 2007. Soil nutrients influence spatial
distributions of tropical tree species. Proceedings of the National Academy of Sciences
104:864-869.
	
  

40	
  

	
  

Keddy, P. A. 1992. Assembly and response rules: two goals for predictive community ecology.
Journal of Vegetation Science 3:157–164.
Kitajima, K., and L. Poorter. 2010. Tissue-level leaf toughness, but not lamina thickness,
predicts sapling leaf lifespan and shade tolerance of tropical tree species. New Phytologist
186:708-721.
Körner, C., M. Neumayer, S. Pelaez Menendez-Riedl, and A. Smeets-Scheel. 1989. Functional
morphology of mountain plants. Flora 182:353-383.
Kraft, N. J. B., R. Valencia, and D. D. Ackerly. 2008. Functional traits and niche–based tree
community assembly in an Amazonian forest. Science 322:580-582.
Lambers, H., F. S. Chapin III, and T. L. Pons. 1998. Plant physiological ecology. SpringerVerlag New York Inc. New York, NY.
Losos, J. B. 2008. Phylogenetic niche conservatism, phylogenetic signal and the relationship
between phylogenetic relatedness and ecological similarity among species. Ecology Letters
11:995-1007.
Losos, J. B., M. Leal, R. E. Glor, K. de Queiroz, P. E. Hertz, L. Rodriguez Schettino, A. C. Lara,
T. R. Jackman, and A. Larson. 2003. Niche lability in the evolution of a Caribbean lizard
community. Nature 424: 542-545.
MacArthur, R., and R. Levins. 1967. The limiting similarity, convergence, and divergence of
coexisting species. American naturalist 101:377-385.
Markesteijn, L., L. Poorter, H. Paz, L. Sack, and F. Bongers. 2011. Ecological differentiation in
xylem cavitation resistance is associated with stem and leaf structural traits. Plant, Cell and
Environment 34:137-148.
Mayfield, M. M., and J. M. Levine. 2010. Opposing effects of competitive exclusion on the
phylogenetic structure of communities. Ecology Letters 13: 1085-1093.
Mittelbach, G. G. 2012. Community Ecology. Sinauer Associates, Inc. Sunderland, MA.
Moles, A. T., and M. Westoby. 2006. Seed size and plant strategy across the whole life cycle.
Oikos 113:91-105.
Moles, A. T., D. I. Warton, L. Warman, N. G. Swenson, S. W. Laffan, A. E. Zanne, A. Pitman,
F. A. Hemmings, and M. R. Leishman. 2009. Global patterns in plant height. Journal of
Ecology 97:923-932.
Niinemets, Ü. 1999. Research review. Components of leaf dry mass per area-thickness and
density alter leaf photosynthetic capacity in reverse directions in woody plants. New
Phytologist 144:35-57.
	
  

41	
  

	
  

Niinemets, Ü. 2001. Global-scale climatic controls of leaf dry mass per area, density, and
thickness in trees and shrubs. Ecology 82:453-469.
Ordoñez, J. C., P. M. van Bodegom, J. M. Witte, I. J. Wright, P. B. Reich, and R. Aerts. 2009. A
global study of relationships between leaf traits, climate and soil measures of nutrient
fertility. Global Ecology and Biogeography 18:137-149.
Paine, C. E. T., C. Stahl, E. A. Courtois, S. Patiño, C. Sarmiento, and C. Baraloto. 2010.
Functional explanations for variation in bark thickness in tropical rain forest trees. Functional
Ecology 24:1202-1210.
Paradis, E., J. Claude, and K. Strimmer. 2004. APE: analyses of phylogenetics and evolution in
R language. Bioinformatics 20:289-290.
Pearson, R. G., C. J. Raxworthy, M. Nakamura, and A. T. Peterson. 2007. Predicting species
distributions from small numbers of occurrence records: a test case using cryptic geckos in
Madagascar. Journal of Biogeography 34:102-117.
Peres-Neto, P. R. 2004. Patterns in the co-occurrence of fish species in streams: the role of site
suitability, morphology and phylogeny versus species interactions. Oecologia 140:352-360.
Perez-Harguindeguy, N., S. Diaz, E. Garnier, S. Lavorel, H. Poorter, P. Jaureguiberry, M. S.
Bret-Harte, W. K. Cornwell, J. M. Craine, D. E. Gurvich, C. Urcelay, E. J. Vaneklaas, P. B.
Reich, L. Poorter, I. J. Wright, P. Ray, L. Enrico, J. G. Pausas, A. C. de Vos, N. Buchmann,
G. Funes, F. Quetier, J. G. Hodgson, K. Thompson, H. D. Morgan, H. ter Steege, M. G. A
van der Heijden, L. Sack, B. Blonder, P. Poschlod, M. V. Vaieretti, G. Conti, A. C. Staver, S.
Aquino, and J. H. C. Cornelissen. 2013. New handbook for standardized measurement of
plant functional traits worldwide. Australian Journal of Botany 61: 167-234.
Peterson, A. T., J. Soberón, and V. Sánchez-Cordero. 1999. Conservatism of ecological niches in
evolutionary time. Science 285:1265-1267.
Phillips, S. J., and M. Dudík. 2008. Modeling of species distributions with Maxent: new
extensions and a comprehensive evaluation. Ecography 31:161–175.
Phillips, S. J., Anderson, R. P., and R. E. Schapire. 2006. Maximum entropy modeling of species
geographic distributions. Ecological Modelling 190:231–259.
Pieters, G. A. 1974. The growth of sun and shade leaves of Populus euramericana 'Robusta' in
relation to age, light intensity and temperature. Mededelingen van de Landbou- whogeschool
de Wageningen, Nederland 74:1-107.
Poorter, L., I. McDonald, A. Alarcón, E. Fichtler, J. Licona, M. Peña-Claros, F. Sterck, Z.
Villegas, and U. Sass-Klaassen. 2010. The importance of wood traits and hydraulic

	
  

42	
  

	
  
conductance for the performance and life history strategies of 42 rainforest tree species. New
Phytologist 185:481-492.
R Core Team. 2013. R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.
Reich, P. B., D. S. Ellsworth, M. B. Walters, J. M. Vose, C. Gresham, J. C. Volin, and W. D.
Bowman. 1999. Generality of leaf trait relationships: a test across six biomes. Ecology 80:
1955-1969.
Reich, P. B., M B. Walters, and D. S. Ellsworth. 1997. From tropics to tundra: global
convergence in plant functioning. PNAS 94:13730-13734.
Ricklefs, R. E. 1987. Community diversity: relative roles of local and regional processes.
Science 235: 167-171.
Schoener, T. W. 1968. The Anolis lizards of Bimini: resource partitioning in a complex fauna.
Ecology 49: 704-726.
Sperry, J. S., and J. E. M. Sullivan. 1992. Xylem embolism in response to freeze-thaw cycles and
water stress in ring-porous, diffuse-porous, and conifer species. Plant Physiology 100:605613.
Sterck, F., L. Markesteijn, F. Schieving, and L. Poorter. 2011. Functional traits determine tradeoffs and niches in a tropical forest community. Proceedings of the National Academy of
Sciences 108:20627-20632.
Sutherland, W. J., R. P. Freckleton, H. C. J. Godfray, S. R. Beissinger, T. Benton, D. C.
Cameron, Y. Carmel, D. A. Coomes, T. Coulson, M. C. Emmerson, R. S. Hails, G. C. Hays,
D. J. Hodgson, M. J. Hutchings, D. Johnson, J. P. G. Jones, M. J. Keeling, H. Kokko, W. E.
Kunin, X. Lambin, O. T. Lewis, Y. Malhi, N. Mieszkowska, E. J. Milner-Gulland, K. Norris,
A B. Phillimore, D. W. Purves, J. M. Reid, D. C. Reuman, K. Thompson, J. M. J. Travis, L.
A. Turnbull, D. A. Wardle, and T. Wiegand. 2013. Identification of 100 fundamental
ecological questions. Journal of Ecology 101: 58-67.
Swenson, Nathan G. 2011. The role of evolutionary processes in producing biodiversity patterns,
and the interrelationships between taxonomic, functional and phylogenetic diversity.
American Journal of Botany 98: 472-480.
Swenson, Nathan G. 2013. The assembly of tropical tree communities – the advances and
shortcomings of phylogenetic and functional trait analyses. Ecography 36: 001-013.
Swenson, N. G., and B. J. Enquist. 2007. Ecological and evolutionary determinants of a key
plant functional trait: wood density and its community wide variation across latitude and
elevation. American Journal of Botany 94:451-459.

	
  

43	
  

	
  
Swenson, N. G., and B. J. Enquist. 2009. Opposing assembly mechanisms in a Neotropical dry
forest: implications for phylogenetic and functional community ecology. Ecology 90: 21612170.
Syversten, J. P., J. Lloyd, C. McConchie, P. E. Kriedemann, and G. D. Farquhar. 1995 On the
relationship between leaf anatomy and CO2 diffusion through the mesophyll of
hypostomatous leaves. Plant, Cell and Environment 18:149-157.
Tobias, J. A., C. K. Cornwallis, E. P. Derryberry, S. Claramunt, R. T. Brumfield, and N. Seddon.
2013. Species coexistence and the dynamics of phenotypic evolution in adaptive radiation.
Nature 000:1-7.
Tyree, Melvin T., and John S. Sperry. 1989. Vulnerability of xylem to cavitation and embolism.
Annual review of plant biology 40:19-36.
Van der Vaart, A. W. 1998. Asymptotic statistics. Cambridge Univ. Press, Cambridge, U.K.
Wallace, A. R. 1869. The Malay Archipelago: the land of the orang-utan and the bird of
paradise. A narrative of travel, with studies of man and nature. Macmillan. London.
Warren, D. L., R. E. Glor, and M. Turelli. 2008. Environmental niche equivalency versus
conservatism: quantitative approaches to niche evolution. Evolution 62: 2868-2883.
Webb, C. O. 2000. Exploring the phylogenetic structure of ecological communities: an example
for rainforest trees. American Naturalist 156: 145-155.
Weiher, E., G. D. P. Clarke, and P. A. Keddy. 1998. Community assembly rules, morphological
dispersion, and the coexistence of plant species. Oikos 81: 309-322.
Weiher, E., and P. A. Keddy. 1995. Assembly rules, null models and trait dispersion – New
questions from old patterns. Oikos 74:159-164.
Wiens, J. J., D. D. Ackerly, A. P. Allen, B. L. Anacker, L. B. Buckley, H. V. Cornell, E. I.
Damschen, T. J. Davies, J. Grytnes, S. P. Harrison, B. A. Hawkins, R. D. Holt, C. M.
McCain, and P. R. Stephens. 2010. Niche conservatism as an emerging principle in ecology
and conservation biology. Ecology Letters 13:1310-1324.
Wilson, P. J., K. Thompson, and J. G. Hodgson. 1999. Specific leaf area and leaf dry matter
content as alternative predictors of plant strategies. New Phytologist 143:155-162.
Winston, M. R. 1995. Co-occurrence of morphologically similar species of stream fishes.
American Naturalist 145:527-545.
Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-Bares,
T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, J. Gulias, K.
Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgely, M. Navas, Ü Niinemets, J.
	
  

44	
  

	
  
Oleksyn, N. Osada, H. Poorter, P. Poot, L. Prior, V. I. Pyankov, C. Roumet, S. C. Thomas,
M. G. Tjoelker, E. J. Veneklass, and R. Villar. 2004. The worldwide leaf economics
spectrum. Nature 428:821-827.

	
  

45