Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
rspb.royalsocietypublishing.org
The building of a biodiversity hotspot
across a land-bridge in the Mediterranean
Rafael Molina-Venegas1,†, Abelardo Aparicio1, Sébastien Lavergne2
and Juan Arroyo1
Research
Cite this article: Molina-Venegas R, Aparicio
A, Lavergne S, Arroyo J. 2015 The building of a
biodiversity hotspot across a land-bridge in the
Mediterranean. Proc. R. Soc. B 282: 20151116.
http://dx.doi.org/10.1098/rspb.2015.1116
Received: 15 May 2015
Accepted: 30 June 2015
Subject Areas:
evolution, ecology
Keywords:
angiosperm flora, land-bridges, Mediterranean
hotspots, phylogenetic alpha diversity,
phylogenetic beta diversity, speciation
Author for correspondence:
Rafael Molina-Venegas
e-mail: [email protected]
†
Present address: Department of Life Sciences,
Universidad de Alcalá, Alcalá de Henares,
Madrid, Spain.
Electronic supplementary material is available
at http://dx.doi.org/10.1098/rspb.2015.1116 or
via http://rspb.royalsocietypublishing.org.
1
2
Departamento de Biologı́a Vegetal y Ecologı́a, Universidad de Sevilla, Apartado 1095, Sevilla 41080, Spain
Laboratoire d’Écologie Alpine, CNRS - Université Grenoble Alpes, Grenoble, France
Many of the macroevolutionary processes that have shaped present-day phylogenetic patterns were caused by geological events such as plate tectonics and
temporary land-bridges. The study of spatial patterns of phylogenetic diversity
can provide insights into these past events. Here we focus on a western Mediterranean biodiversity hotspot located in the southern Iberian Peninsula and
northwest Africa, two regions that are separated by the Strait of Gibraltar. We
explore the spatial structure of the phylogenetic relationships within and
across large-scale plant assemblages. Significant turnover in terminal lineages
tends to occur between landmasses, whereas turnover in deep lineages tends
to occur within landmasses. Plant assemblages in the western ecoregions of
this hotspot tend to be phylogenetically overdispersed but are phylogenetically
clustered on its eastern margins. We discuss our results in the light of potential
scenarios of niche evolution (or conservatism) and lineage diversification. The
significant turnover between landmasses suggests a common scenario of allopatric speciation that could have been facilitated by the intermittent joining of the
two continents. This may have constituted an important stimulus for diversification and the emergence of this western Mediterranean biodiversity hotspot.
1. Introduction
The field of ecophylogenetics, which can be viewed as the merging of community
ecology, biogeography and macroevolution, has been rapidly expanding over the
past decade [1]. However, most work has been conducted on local-scale communities [2] and less attention has been paid to understanding phylogenetic
structures within or across regional species pools (but see [3–6]). At local
scales, non-random phylogenetic structure patterns have been mostly interpreted
to be the result of ecological sorting processes such as competition [7–9], habitat
filtering [10,11] and facilitation [12,13]. At larger geographical scales, macroevolutionary and historical processes (in addition to ecological sorting) are also thought
to have shaped the phylogenetic structure of regional assemblages. In particular,
past biogeographical events such as orogenies and massive climate changes have
affected many of the macroevolutionary processes, and hence have shaped the
phylogenetic relationships between extant species [14,15]. One type of biogeographical event that has long fascinated biogeographers and evolutionary biologists
is the intermittent connections that have occurred between different regions or
continents (the so-called land-bridges [16–18]), which have probably triggered
species diversification in numerous plant and animal lineages [19–22].
The southern part of the Iberian Peninsula (Andalusia) and northwest
Africa (north Morocco; figure 1) are home to exceptional levels of plant diversity [23] and account for 18% of the Mediterranean Basin’s plant richness [24].
This region is divided between the Iberian and African tectonic plates, whose
intermittent connection over geological time (e.g. the Mediterranean desiccation
during the Messinian salinity crisis [25]) has allowed biotic exchanges between
both landmasses to take place [26,27]. Thus, the Neogene geological history of
this region could have had a profound impact on species distributions and
species diversification in certain lineages, and probably contributed to the
emergence of this western Mediterranean biodiversity hotspot [22].
& 2015 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
Europe
Africa
vir
lqui
er
Riv
da
Gua
tic
Bae
Proc. R. Soc. B 282: 20151116
Strait of
Gibraltar
ge
Ran
Mediterranean Sea
Rif
an
Ran
ge
Sebo
u Riv
er
km
0
100
Figure 1. Map of the southern Iberian Peninsula (Andalusia) and northwest Africa (north Morocco) showing the position of the Baetic and Rifan Ranges, the Strait
of Gibraltar, and the Guadalquivir and Sebou Rivers. The shading represents elevation (maximum value in black). The inset shows the location of Andalusia and
northern Morocco in the western Mediterranean.
Botanists have long been aware of the floristic links
between the southern Iberian Peninsula (Andalusia) and
northwest Africa (north Morocco) [23,28]. More recently,
Molina-Venegas et al. [24] found that plant assemblages in
the western ecoregions of northern Morocco are more similar
to those of Andalusia than to other nearby northern Morocco
ecoregions, thereby suggesting that the Strait of Gibraltar
acted as a route for plant migration. On the other hand,
Lavergne et al. [22] found a significant relationship between
the migration probability across the Strait and certain species’
life-history and dispersal traits, which suggests that the Strait
also acted as an effective biotic filter, only allowing migration
and gene flows in certain lineages, and triggering allopatric
speciation in others. However, this analysis was restricted
to only 566 species and subspecies occurring in two narrow
ecoregions lying on both sides of the Strait of Gibraltar.
This limitation strongly limits the extent to which their conclusions can be extrapolated to the whole hotspot and the
over 4000 species and subspecies it harbours.
Our first goal was to explore the large-scale phylogenetic
structure across plant assemblages (phylogenetic beta diversity,
PBD) of southern Iberian Peninsula (Andalusia) and northwest
Africa (north Morocco), to unravel past biogeographical and
macroevolutionary processes that shaped plant biodiversity in
this western Mediterranean hotspot. To do so, PBD and taxonomic beta diversity (TBD) were considered in tandem. TBD is
the change in species composition across any two areas, while
PBD is defined as the phylogenetic distance (branch lengths)
between species assemblages across the same areas [29]. For
example, a scenario of repeated splits in recently coexisting
ancestors between two areas would result in a high TBD between
assemblages across the areas and a lower PBD (figure 2).
However, widely used indices for PBD incorporate different aspects of assemblages’ evolutionary history that are
rspb.royalsocietypublishing.org
an
Iberi sula
Penin
N
2
often confounded. PBD can be broken down into two components: ‘true’ phylogenetic turnover and phylogenetic
nestedness [30,31]. The nestedness component of PBD is
related to differences in phylogenetic diversity (PD) between
species assemblages (sensu [32]), whereas the ‘true’ turnover
component is the fraction of PBD explained without increasing the PD between assemblages (figure 3). Thus, a low value
of ‘true’ phylogenetic turnover will indicate a trend towards
the segregation of close relatives between assemblages
(figure 3, top-left panel); a high ‘true’ phylogenetic turnover
will indicate a trend towards the segregation of deep lineages
between assemblages (figure 3, bottom-left panel). By contrast, a high nestedness component of PBD will also
indicate a trend towards the segregation of deep lineages,
but also an increase in the PD between plant assemblages
(figure 3, bottom-right panel). Thus, estimating the PBD
without taking into account both components of beta diversity may lead to an overestimation of the ‘true’ spatial
turnover of lineages if the two compared assemblages have
contrasting levels of PD [30].
In addition, we considered phylogenetic structure within
assemblages ( phylogenetic alpha diversity) jointly with PBD,
in order to capture additional processes shaping spatial patterns of species diversity. At the local scale, a non-random
pattern in phylogenetic alpha diversity may arise through a
process known as habitat filtering [33] if locally coexisting
species share adaptations to particular environmental conditions and are phylogenetically clustered compared with a
regional species pool, or if distantly related coexisting species
have converging key adaptations (causing phylogenetic overdispersion). At large geographical scales, ‘habitat filtering’
can be thought of in terms of differential expansion of species’
geographical ranges, according to species’ suitability to the
overall environmental conditions of a study region [3]. This
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
(a)
3
(b)
A
A
A
Mediterranean Sea
B
C
B
NW Africa
B
C
C
A1
C
A1
B1
B1
A2
AFRICA
B2
A2
C
EUROPE
B2
C
Figure 2. Schematic of a hypothetical scenario of allopatric speciation facilitated by the intermittent connections between the southern Iberian Peninsula and
northwest Africa. (a) The shaded areas represent the home ranges of two species with low long-distance dispersal ability (‘A’ and ‘B’) and a third species with
a high long-distance dispersal ability (‘C’). (b) After the formation of the land-bridge (e.g. the Mediterranean desiccation), the three species expand their home
ranges into the other continent. (c) Finally, the rupture of the land-bridge (e.g. the refilling of the Mediterranean Sea) isolates the southern Iberian and northwest
African populations of the species with low long-distance dispersal ability (‘A’ and ‘B’), thereby driving evolutionary divergence and speciation. Before the formation
of the land-bridge (a), taxonomic and phylogenetic beta diversity (TBD and PBD, respectively) across the two landmasses is relatively high (note that the split in the
more recent common ancestor of the divergent lineages ‘A’ and ‘B– C’ occurred between the two landmasses). After the rupture of the land-bridge (c), TBD across
the two landmasses remains high, whereas PBD becomes low (note that species pairs ‘A1– A2’ and ‘B1– B2’ are close relatives).
PBD turnover
component
PBD nestedness
component
low
high
Figure 3. Four hypothetical situations of two plant assemblages (blue and orange circles) showing contrasting levels of ‘true’ turnover and nestedness components
of PBD. Blue and orange colours on the tree branches represent the total branch length that is unique to each assemblage, while the black colour represents a
branch length that is shared by the two assemblages. The two-colour circles represent species shared by the two assemblages. In the two left-hand side diagrams,
all the PBD is due to the ‘true’ turnover component (DPD between assemblages ¼ 0), while in the right-hand side diagrams the PBD is entirely because of the
nestedness component (DPD between assemblages . 0). Note that the ‘true’ turnover component of PBD can vary even when the TBD between assemblages
remains constant, whereas an increase in the nestedness component of PBD necessarily implies an increase in TBD.
large-scale ‘habitat filtering’ is expected to be more determinant for differential expansion of species’ geographical
ranges in regions of high environmental heterogeneity and
more stable climatic conditions [34] such as the southern
Iberian Peninsula and northwest Africa, which are thought
to have acted jointly as refugia during the Neogene and the
Pleistocene [35,36]. Finally, differential diversification rates
among lineages could also produce phylogenetic clustering
Proc. R. Soc. B 282: 20151116
(c)
C
B
rspb.royalsocietypublishing.org
S Iberian
Peninsula
A
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
2. Material and methods
(a) Study area
(b) Floristic dataset
We used the dataset of all the angiosperm plants occurring in Andalusia and northern Morocco in Molina-Venegas et al. [24], which
gave a site-by-species matrix of 48 sites (ecoregions) with 4332
species and subspecies (see Molina-Venegas et al. [24] for full details
on the floristic dataset compilation; see electronic supplementary material, appendix S1 for details on the lineages occurring in
the study area and the species numbers they comprise). An ecoregion can be defined generally as a territory characterized by the
existence of homogeneous ecological systems involving
interrelationships among organisms and their environment [39].
A territorial subdivision of this type, based on natural physiographic boundaries, has been used in a number of recent studies
and is appropriate for different analytical methods (e.g. [6,40]).
(c) Phylogenetic trees
In order to account for the phylogenetic relationships between the
species and subspecies in the dataset, we used the genus-level
time-calibrated phylogeny described by Molina-Venegas &
Roquet [41]. This phylogeny was inferred using a maximumlikehood approach with a multilocus supermatrix of chloroplastic
and nuclear DNA sequences (rbcL, matK, ndhF, trnl-F, ITS1 and
ITS2) available in GenBank (accessed in September 2011) following the pipeline of Roquet et al. [42]. Node support was estimated
using bootstrap values (70% of nodes showed a bootstrap support
greater than 70), and nodes with values less than 50% were collapsed into soft polytomies (see Molina-Venegas and Roquet
[41] for further details on the phylogenetic procedure).
In order to account for the possible influence of intra-genus
phylogenetic uncertainty on phylogenetic structure metrics, we
randomly resolved genus-level polytomies by applying a YuleHarding branching process with constant birth rates (n ¼ 100).
The algorithm assigns to each node the same probability of splitting
in two lineages, resulting in a balanced topology [43]. Subspecies
were constrained to split within their respective species.
(d) Statistical analysis
(i) Phylogenetic beta diversity
In order to test for non-random patterns in PBD among plant
assemblages, we decomposed the observed PBD between the
turnover and the nestedness components of beta diversity
using the betadecompo R function and the PhyloSor index [30].
SES ¼
Xobs mean(Xnull Þ
,
s.d. (Xnull Þ
ð2:1Þ
where Xobs is the observed score between two ecoregions, and
mean (Xnull ) and s.d. (Xnull ) are the mean and standard deviation
of a null distribution of scores generated by shuffling taxa labels of
the site-by-species matrix 999 times. It is important to note that
this shuffling scheme randomizes phylogenetic relationships
among the regional species pool but holds species richness and
TBD constant between assemblages, thereby allowing us to discard the effect of the different number of species in each plant
assemblage [30]. We assessed the statistical significance of the
SES scores between ecoregions by calculating two-tailed p-values
(quantiles) as
p-values ¼
rankobs
,
runs + 1
ð2:2Þ
where rankobs is the rank of the observed scores compared with
those of their null distributions, and ‘runs’ is the number of randomizations [46]. SES scores with p , 0.05 and p . 0.95 were
considered to be significantly lower and higher than expected
for the given TBD, respectively. We conducted the calculations
across 100 phylogenetic trees after randomly resolving intragenus phylogenetic uncertainty (see ‘Phylogenetic trees’ section),
then considered an SES score as significant when the fraction of
p-values (n ¼ 100) below and above the threshold value (0.05
and 0.95, respectively) was . 95%.
We considered three groups of pairwise comparisons in
terms of the geographical location of the ecoregions (i) across
ecoregions of Andalusia, (ii) across ecoregions of northern
Morocco and (iii) across ecoregions between the two landmasses.
We tested for non-random distributions of the significant pairwise comparisons within each group by conducting x 2-tests
(significance threshold value at 0.05).
(ii) Phylogenetic alpha diversity
In order to quantify phylogenetic structure within plant
assemblages, we estimated PD (sensu [32]) and the mean phylogenetic distance (MPD) [33] of the species inhabiting each of the
ecoregions. We tested for non-random patterns in the PD and
MPD by estimating their SES scores (SESpd and SESmpd,
respectively) as in equation (2.1), where Xobs is the observed
score within an ecoregion, and mean (Xnull ) and s.d. (Xnull ) are
the mean and standard deviation of a null distribution of
scores generated by shuffling taxa labels of the site-by-species
matrix 999 times. We conducted the calculations across 100 phylogenetic trees after randomly resolving intra-genus phylogenetic
uncertainty (see ‘Phylogenetic trees’ section), then extracted the
arithmetic mean of the distribution of SESpd and SESmpd scores.
There is some evidence to suggest that estimates of the phylogenetic structure of plant assemblages containing a deep and
unbalanced initial bifurcation in the phylogeny (e.g. angiosperms – gymnosperms, eudicots – monocots) are very sensitive
to the ratio of the species sustained by each bifurcation [47].
Thus, in order to test for the consistency of our phylogenetic
structure estimates, we repeated all the analyses after removing
4
Proc. R. Soc. B 282: 20151116
Andalusia (south Iberian Peninsula) and northern Morocco (northwest Africa) are geomorphologically heterogeneous areas
characterized by high mountain ranges (i.e. the Baetic–Rifan complex, up to 3482 m.a.s.l.) surrounded by extensive lowlands of
alluvial origin that have been shaped by the rivers Guadalquivir
(Andalusia) and Sebou (north Morocco; figure 1). These two landmasses are separated by the Mediterranean Sea, which is around
14 km at its narrowest point (Strait of Gibraltar) and over 150 km
at the easternmost end of the study area. The region’s current
Mediterranean-type seasonal precipitation rhythm has prevailed
since around 3.6 Ma [38], but is clearly affected by the proximity of
the Atlantic Ocean [36].
This procedure allows us to distinguish between differences in
PBD arising from (i) differences in PD (sensu [32], nestedness
component) or from (ii) those that are due to the ‘true’ turnover
of lineages. This function uses the PhyloSor index as a distance
metric as it computes the fraction of non-shared branch length
between assemblages (ranging from 0 to 1). Previous empirical
studies emphasized that TBD and PBD may be highly correlated
[44,45]. Thus, we determined whether the turnover and nestedness components of PBD among plant assemblages were lower
or higher than expected given the observed TBD by calculating
the standardized effect size (SES) scores of both components
(SESturn and SESness) independently as
rspb.royalsocietypublishing.org
if diversifying clades tend towards occupying restricted
geographical regions [37]. Such a diversification pattern is
expected under a scenario of environment-linked diversification processes through particular combinations of adaptive
traits in diversifying lineages [15].
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
all non-eudicot taxa from the dataset (n ¼ 796, 18% of the species
and subspecies in the dataset).
(iv) Regression analyses
First, we quantified whether temperature- and precipitationrelated variables are associated with the phylogenetic structure
within plant assemblages. First, we fitted ordinary least-squares
models (OLS) between the SESpd and SESmpd scores and both
PCA1 subsets of climatic variables, and sought spatial autocorrelation in the phylogentic structure metrics by visually exploring
Moran’s I spatial correlograms of the residuals of the OLS
models (see electronic supplementary material, appendix S2 for
details on the procedure). Because the data showed strong spatial
autocorrelation, we fitted spatial error simultaneous autoregressive models (SARerr) to incorporate the autocorrelation
structure of both phylogenetic structure metrics in the OLS
models [50]. Although changes in coefficients between spatial
and non-spatial methods may be largely idiosyncratic of the
method used [51], SAR models has been empirically demonstrated to be among the best-performing methods for spatial
data analysis in terms of Type I and Type II error rates [52].
Specifically, the SARerr model is preferable to other SAR
models when dealing with spatially autocorrelated species distribution data as it generates more accurate estimates of regression
model coefficients and performs well for all SAR model assumptions [50]. We arbitrarily defined the neighbourhood structure of
each ecoregion (two-dimensional coordinates) using the minimum Euclidean distance required to connect each location to at
least one neighbouring location, and weighted neighbouring
locations using the row-standardization method. We used the
Nagelkerke pseudo R 2 as an estimate of the goodness-of-fit of
the SARerr models.
All analyses were carried out in R v. 3.1.2 [53] with the packages
PICANTE [46], APE [54], VEGAN [55], SPDEP [56], NCF [57] and
3. Results
(a) Phylogenetic beta diversity
The significant pairwise comparisons (i.e. those differing
from expectation due to the TBD) of both components of
PBD were non-randomly distributed across the study area.
Lower-than-expected standardized scores of the turnover
component (SESturn) were found mainly across ecoregions
between landmasses (x 2 ¼ 22.54, n ¼ 26), while higherthan-expected SESturn scores were more numerous within
landmasses (x 2 ¼ 50.19, n ¼ 170; figure 4). By contrast,
lower-than-expected standardized scores of the nestedness
component (SESness) were randomly distributed (x 2 ¼ 1.64,
n ¼ 26), while higher-than-expected scores occurred mainly
across ecoregions between landmasses (x 2 ¼ 172.70, n ¼ 393).
Interestingly, the removal of non-eudicot plants from the analyses resulted in a positive reinforcement of the SESturn pattern
between landmasses, but a negative reinforcement within
landmasses (figure 4).
(b) Climatic variables
The first axis of the PCAs for temperature- and precipitationrelated variables explained 65% and 69% of the variance,
respectively. Positive values were related to high maximum
temperatures values and to high total monthly precipitation
and standard deviation of monthly precipitation, whereas
negative values were related to high intra-annual variability
in maximum temperature values and relatively high precipitations during the driest month. Overall, the PCA1 scores for
temperature-related variables, and particularly those for the
precipitation-related variables, tend to decrease eastward
(see electronic supplementary material, appendix S3). On
the other hand, the PCA1 scores for temperature-related
variables tend to increase from north to south.
(c) Phylogenetic alpha diversity
When considering all angiosperms as a whole, the relationship
between SESmpd and the PCA1 of the precipitation-related
variables was strong and positive (R 2 ¼ 0.50, p , 0.001;
table 1 and figure 5), whereas the effect of the temperaturerelated variables was non-significant. On the other hand, the
relationship between SESpd and the climatic-related variables
was also quite strong, but non-significant. When considering
only eudicots in the analysis, the relationship between either
SESmpd or SESpd and the PCA1 of the precipitation-related
variables was non-significant, whereas the relationship
between either SESmpd or SESpd and the PCA1 of the temperature-related variables was strong and negative (R 2 ¼ 0.53,
p , 0.01 and R 2 ¼ 0.55, p , 0.001 respectively).
4. Discussion
The incorporation of phylogenetic information in the study of
species distributions and community assembly can provide
valuable insights into the biogeographical and ecological
processes that generate biodiversity patterns [3,5], as these processes leave tractable imprints on present-day phylogenetic
Proc. R. Soc. B 282: 20151116
We tested whether climatic conditions are associated with variation
in the phylogenetic structure of the plant assemblages in the study
area. For the climatic characterization of the ecoregions, we used
the same 18 climatic variables as Molina-Venegas et al. [24]. We considered precipitation- and temperature-related variables as two
independent subsets of climatic variables. First, we explored the
correlation coefficients among the climatic variables and
arbitrarily selected those with Spearman’s correlation coefficient
r , 0.90 for precipitation- and temperature-related variables,
respectively. We finally selected (i) total monthly precipitation,
(ii) standard deviation of monthly precipitation, (iii) total precipitation of the driest month, (iv) mean monthly maximum temperature
and (v) standard deviation of monthly maximum temperature.
These subsets of precipitation- and temperature-related variables
are representations of the climatic regime of the study area, which
is characterized by a marked seasonality in the distribution of precipitation and extreme high temperature values during the dry
season [28,48]. We standardized all variables using Gower’s standardization of quantitative variables (division by the range). Given
that climatic variables had a high degree of collinearity, we conducted separate principal components analyses for the temperature- and
precipitation-related variables to reduce their dimensionality, and
only used the first principal component of each analysis (PCA1).
Finally, we used the geographical coordinates of the centroids of
ecoregion polygons as an estimate of their spatial location.
Here, we use current climatic variation across the region
rather than past climatic conditions. Nevertheless, former palaeoclimatic studies show that despite strong historical variation in
absolute values in precipitation and temperature, the climatic
gradient in the study region has been quite similar to that
found nowadays at least since the Pliocene [49].
5
rspb.royalsocietypublishing.org
(iii) Climatic and spatial variables
the betadecompo R function [30]. The extraction of the geographical
coordinates was carried out in ESRI ARCMAP v. 9.3.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
6
rspb.royalsocietypublishing.org
angiosperms
Proc. R. Soc. B 282: 20151116
eudicots
Figure 4. Graphic representation of the ‘true’ turnover component of PBD across plant assemblages. Lines connect pairwise ecoregions of lower-than-expected (red
lines) and higher-than-expected (blue lines) ‘true’ turnover component of PBD for the given TBD. The red colour on the tree branches represents phylogenetic
turnover in terminal lineages (lower-than-expected turnover) while the blue colour represents phylogenetic turnover in deep lineages (higher-than-expected
turnover).
Table 1. Nagelkerke pseudo R 2 coefficients, coefficient estimates (m) and p-values from the simultaneous autoregressive regressions between SESmpd and
SESpd and the PCAs for temperature- and precipitation-related variables. In bold, significant coefficient estimates at p , 0.05.
precipitation
temperature
p-value
R2
20.07
0.825
0.50
0.195
0.823
20.36
20.75
0.134
0.003
0.56
0.53
0.382
20.79
<0.001
0.55
lineage
variable
m
p-value
m
angiosperms
SESmpd
0.99
<0.001
eudicots
SESpd
SESmpd
0.33
0.06
SESpd
0.18
patterns [4,29,58]. Here, we found that phylogenetic structure patterns of plant assemblages across two continental
landmasses (i.e. southern Iberian Peninsula and northern Morocco) and across ecoregions within each land mass are nonrandomly structured. This suggests that the particular geographical and historical settings of this region in the western
Mediterranean may have constituted an important stimulus
for the evolutionary diversification and the emergence of a
major biodiversity hotspot [23].
(a) Phylogenetic structure between landmasses
A primary finding of our study is that both components of
PBD showed a significant spatial structure. Lower-thanexpected phylogenetic ‘true’ turnover mainly occurred across
ecoregions between landmasses, which implies a segregation
of closely related taxa between the two continents (figure 4).
We believe that this pattern could have emerged because of
repeated splits between the two landmasses in recently coexisting ancestors of extant taxa, probably to intermittent
connections and spatial isolation (i.e. land-bridges) between
the two continents (figure 2). The floristic similarity between
the two landmasses is quite high (47% of taxa are shared;
data not shown). This confirms that the Mediterranean Sea
has biased species migration between northern Morocco
and the southern Iberian Peninsula, in such a way that
better dispersers ancestors have been less likely to experience
vicariance through continuous gene flows, thus favouring speciation only in certain lineages with low long-distance
dispersal ability [22]. Molecular evidence further supports
varying levels of the Mediterranean Sea and intermittent
land-bridges as the most plausible explanation for species
differentiation between southern Iberian Peninsula and northern Morocco in disparate angiosperm clades [59–62], and
between other disjunct areas within the Mediterranean
Basin [63–65]. These results illustrate how land-bridges
between Europe and Africa have contributed to the emergence
of this western Mediterranean biodiversity hotspot. Other
land-bridges (e.g. the Sicilian–Tunisian bridge) may have
acted in a similar way [66].
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
EARLY ANG. MONOC.
EUDIC.
eudicots
random
clustering
Figure 5. Graphic representation of the averaged SESpd and SESmpd scores of plant assemblages in the study area. The estimated scores were scaled from 0
(minimum value) to 1 (maximum value). The inset shows the two major climatic gradients in the study area, and the phylogeny schematizes the spatial sorting
of basal angiosperm lineages (i.e. early angiosperms, monocots and eudicots) along the longitudinal axis of the study area.
The predominance of the nestedness component of PBD
across ecoregions between landmasses can be explained by
their marked differences in PD (figure 5). Thus, the pattern
of segregation of closely related taxa between the two
continents would be masked by differences in PD if an undecomposed metric of PBD had been considered. This issue
highlights the importance of disentangling both components
of beta diversity in order to avoid overestimations of the
‘true’ spatial phylogenetic turnover [30].
(b) Phylogenetic structure within landmasses
Higher-than-expected phylogenetic ‘true’ turnover mainly
occurred within landmasses, which implies that the overall
floristic differences observed between the western and eastern ecoregions of the study area [24] are mainly driven by
changes in the distribution of deep lineages within continents
(figure 4). We also found a strong relationship between the
mean phylogenetic distance within plant assemblages
(SESmpd) and the precipitation-related variables that match
the general decreasing precipitation gradient eastward in
the study area (figure 5). This pattern can be explained largely by the differential distribution of deep angiosperm
lineages along the precipitation gradient as the observed pattern weakens when considering only eudicots in the analysis.
Prinzing et al. [67] found that a high amount of variance in
niche-related traits in vascular plants can be explained at
higher taxonomic levels, thereby further supporting the
hypothesis of an ecological sorting of deep angiosperm
lineages along the precipitation gradient in the study
region. This was probably because of a certain tendency of
phylogenetic niche conservatism of higher plants; that is,
the tendency of taxa to retain their ancestral traits adapted
to particular climates [68]. Thus, the ecological restriction
and long-term local persistence (thanks to micro-climatic
refugia) of some relatively old, water-dependent lineages of
non-eudicot plants (e.g. Laurus, Nuphar, Nymphaea, Triglochin,
Butomus, Hydrocharis, Wolffia, Spirodela, Ceratophyllum) could
have caused the observed phylogenetic overdispersion of
plant assemblages in the western ecoregions. By contrast,
plant assemblages in the eastern ecoregions of the study
area tend to be phylogenetically clustered because of the
widespread occurrence of more xeric-adapted lineages that
have experienced in situ diversification, such as Helianthemum,
Thymus, Sideritis, Teucrium, Arenaria and Armeria, and the
absence of the non-eudicot hygrophilous taxa.
The geological materials that shape most of the Baetic and
Rifan Cordilleras accreted at the southeast Iberian Peninsula
and northwest Africa around 10 Ma [69]. The accretion of
such geological materials is roughly coincident with the
oldest records of the Mediterranean climate in the region
[70], and was followed by the rapid uplift of the Baetic and
Rifan mountains that began around 8 Ma [71]. These new combinations of climatic and geological conditions probably
generated unoccupied environmental conditions such as contrasting substrates in eastern areas of the territory [72] that
provided an ecological opportunity for the diversification
of certain lineages, thereby contributing to phylogenetic clustering [37]. Further, Verdú & Pausas [15] found evidence of
in situ diversification of certain lineages in the eastern Iberian
Peninsula due to the onset of the Mediterranean climate
(e.g. Helianthemum, Thymus, Sideritis, Teucrium, Arenaria and
Armeria). Nevertheless, future comparative studies of diversification focusing on lineages that are well represented in the
Baetic–Rifan complex and beyond will help to improve our
understanding of the different ecological drivers of species
diversification, and unravel general mechanisms that have
contributed to the emergence of Mediterranean biodiversity
hotspots [73,74].
Data accessibility. The datasets and the phylogenetic tree supporting this
article are deposited in Dryad repository: http://dx.doi.org/10.
5061/dryad.54137.
Competing interests. We declare we have no competing interests.
Proc. R. Soc. B 282: 20151116
overdispersion
7
rspb.royalsocietypublishing.org
angiosperms
+ precipitation –
MPD
+ temperature –
PD
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
Funding. This work was supported primarily by the Spanish
study, collected the data, carried out the statistical analyses and
drafted the manuscript. A.A. collected the data, helped in
drafting the manuscript and participated in conceiving the study.
S.L. participated in data analyses and helped in drafting the
manuscript. J.A. helped in drafting the manuscript and participated
in conceiving the study. All authors gave final approval for
publication.
Organismo Autónomo de Parques Nacionales through grant no.
296/2011, and secondarily by the Spanish Ministerio de Ciencia e
Innovación and the Regional Government of Andalucı́a through
grant nos. HF-2008-0040 and P09-RNM-5280.
Acknowledgements. We would like to thank the Centre for Supercomputing of Galicia (CESGA) and the Scientific Computation Centre of
Andalusia (CICA) for the computing services they provided.
References
Mouquet N et al. 2012 Ecophylogenetics: advances
and perspectives. Biol. Rev. 87, 769–785. (doi:10.
1111/j.1469-185X.2012.00224.x)
2. Vamosi SM, Heard B, Vamosi JC, Webb CO. 2009
Emerging patterns in the comparative analysis of
phylogenetic community structure. Mol. Ecol. 18,
572–592. (doi:10.1111/j.1365-294X.2008.04001.x)
3. Cardillo M. 2011 Phylogenetic structure of mammal
assemblages at large geographical scales: linking
phylogenetic community ecology with
macroecology. Phil. Trans. R. Soc. B 366,
2545–2553. (doi:10.1098/rstb 2011.0021)
4. Hardy OJ, Couteron P, Munoz F, Ramesh BR,
Pélissier R. 2012 Phylogenetic turnover in tropical
tree communities: impact of environmental filtering,
biogeography and mesoclimatic niche conservatism.
Glob. Ecol. Biogeogr. 21, 1007 –1016. (doi:10.1111/
j.1466-8238.2011.00742.x)
5. Cantalapiedra JL, Fernández MH, Morales J. 2014
The biogeographic history of ruminant faunas
determines the phylogenetic structure of their
assemblages at different scales. Ecography 37, 1–9.
(doi:10.1111/j.1600-0587.2013.00236.x)
6. Mazel F et al. 2014 Multifaceted diversity –area
relationships reveal global hotspots of mammalian
species, trait and lineage diversity. Glob. Ecol.
Biogeogr. 23, 836– 847. (doi:10.1111/geb.12158)
7. Cavender-Bares J, Ackerly DD, Baum DA, Bazzaz FA.
2004 Phylogenetic overdispersion in Floridian oak
communities. American Naturalist 163, 823 –843.
(doi:10.1086/386375)
8. Swenson NG, Enquist BJ, Thompson J, Zimmerman
JK. 2007 The influence of spatial and size scale on
phylogenetic relatedness in tropical forest
communities. Ecology 88, 1770 –1780. (doi:10.
1890/06-1499.1)
9. Violle C, Nemergut DR, Pu Z, Jiang L. 2011
Phylogenetic limiting similarity and competitive
exclusion. Ecol. Lett. 14, 782–787. (doi:10.1111/j.
1461-0248.2011.01644.x)
10. Verdú M, Pausas JG. 2007 Fire drives phylogenetic
clustering in Mediterranean Basin woody plant
communities. J. Ecol. 95, 1316– 1323. (doi:10.1111/
j.1365-2745.2007.01300.x)
11. Vandelook F, Verdú M, Honnay O. 2012 The role of
seed traits in determining the phylogenetic
structure of temperate plant communities. Ann. Bot.
110, 629–636. (doi:10.1093/aob/mcs121.)
12. Valiente-Banuet A, Rumebe AV, Verdú M, Callaway
RM. 2006 Modern Quaternary plant lineages
promote diversity through facilitation of ancient
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Tertiary lineages. Proc. Natl Acad. Sci. USA 103,
16 812– 16 817. (doi:10.1073/pnas.0604933103)
Oviedo R, Faife-Cabrera M, Noa-Monzón A, Arroyo J,
Valiente-Banuet A, Verdú M. 2013 Facilitation
allows plant coexistence in Cuban serpentine soils.
Plant Biol. 16, 711–716. (doi:10.1111/plb.12116)
Wiens JJ, Donoghue MJ. 2004 Historical
biogeography, ecology and species richness. Trends
Ecol. Evol. 19, 639 –644. (doi:10.1016/j.tree.2004.
09.011)
Verdú M, Pausas JG. 2013 Syndrome-driven
diversification in a Mediterranean ecosystem.
Evolution 67, 1756–1766. (doi:10.1111/evo.12049)
Simpson GG. 1940 Mammals and land bridges.
J. Washington Acad. Sci. 30, 137–163.
Elias SA. 1996 Life and times of the Bering Land
Bridge. Nature 382, 60– 63. (doi:10.1038/
382060a0)
Elias SA, Crocker B. 2008 The Bering Land Bridge: a
moisture barrier to the dispersal of steppe –tundra
biota? Quat. Sci. Rev. 27, 2473– 2483. (doi:10.1016/
j.quascirev.2008.09.011)
Xiang Q, Soltis DE. 2001 Dispersal-vicariance
analyses of intercontinental disjuncts: historical
biogeographical implications for angiosperms in
the Northern Hemisphere. Int. J. Plant Sci. 162,
S29– S39. (doi:10.1086/323332)
Milne RI. 2006 Northern Hemisphere plant
disjunctions: a window on Tertiary land bridges and
climate change? Ann. Bot. 98, 465– 472. (doi:10.
1093/aob/mcl148)
Morales-Castilla I, Olalla-Tárraga MÁ, Purvis A,
Hawkins BA, Rodrı́guez MÁ. 2012 The imprint of
Cenozoic migrations and evolutionary history on the
biogeographic gradient of body size in New World
mammals. Am. Nat. 180, 246– 256. (doi:10.1086/
666608)
Lavergne S, Hampe A, Arroyo J. 2013 In and out of
Africa: how did the Strait of Gibraltar affect plant
species migration and local diversification?
J. Biogeogr. 40, 24 –36. (doi:10.1111/j.1365-2699.
2012.02769.x)
Médail F, Quézel P. 1997 Hot-spots analysis for
conservation of plant biodiversity in the
Mediterranean Basin. Ann. Missouri Bot. Garden 84,
112 –127. (doi:10.2307/2399957)
Molina-Venegas R, Aparicio A, Pina FJ, Valdés B,
Arroyo J. 2013 Disentangling environmental
correlates of vascular plant biodiversity in a
Mediterranean hotspot. Ecol. Evol. 3, 3879–3894.
(doi:10.1002/ece3.762)
25. Duggen S, Hoernle K, van den Bogaard P, Rupke L,
Morgan JP. 2003 Deep roots of the Messinian
salinity crisis. Nature 422, 602–606. (doi:10.1038/
nature01553)
26. Agustı́ J, Garcés M, Krijgsman W. 2006 Evidence for
African-Iberian exchanges during the Messinian in
the Spanish mammalian record. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 238, 5–14. (doi:10.
1016/j.palaeo.2006.03.013)
27. Guzmán B, Vargas P. 2005 Systematics, character
evolution, and biogeography of Cistus L. (Cistaceae)
based on ITS, trnl-trnF, and matK sequences. Mol.
Phylogenet. Evol. 37, 644– 660. (doi:10.1016/j.
ympev.2005.04.026)
28. Valdés B. 1991 Andalusia and the Rif. Floristic links
and common flora. Bot. Chronika 10, 117 –124.
29. Graham CH, Fine PVA. 2008 Phylogenetic beta
diversity: linking ecological and evolutionary
processes across space in time. Ecol. Lett. 11,
1265–1277. (doi:10.1111/j.1461-0248.2008.01256.x)
30. Leprieur F, Albouy C, De Bortoli J, Cowman PF,
Bellwood DR, Mouillot D. 2012 Quantifying
phylogenetic beta diversity: distinguishing between
‘true’ turnover of lineages and phylogenetic
diversity gradients. PLoS ONE 7, e42760. (doi:10.
1371/journal.pone.0042760)
31. Baselga A. 2010 Partitioning the turnover and
nestedness components of beta diversity. Glob. Ecol.
Biogeogr. 19, 134–143. (doi:10.1111/j.1466-8238.
2009.00490.x)
32. Faith DP. 1992 Conservation evaluation and
phylogenetic diversity. Biol. Conserv. 61, 1–10.
(doi:10.1016/0006-3207(92)91201-3)
33. Webb CO, Ackerly DD, McPeek MA, Donoghue MJ.
2002 Phylogenies and community ecology. Annu.
Rev. Ecol. System. 33, 475–505. (doi:10.1146/
annurev.ecolsys.33.010802.150448)
34. Svenning J, Skov F. 2004 Limited filling of the
potential range in European tree species. Ecol. Lett. 7,
565–573. (doi:10.1111/j.1461-0248.2004.00614.x)
35. Finlayson C, Carrión JS. 2007 Rapid ecological
turnover and its impact on Neanderthal and other
human populations. Trends Ecol. Evol. 22, 213 –222.
(doi:10.1016/j.tree.2007.02.001)
36. Rodrı́guez-Sánchez F, Pérez-Barrales R, Ojeda F,
Vargas P, Arroyo J. 2008 The Strait of Gibraltar
as a melting pot for plant biodiversity. Quat. Sci.
Rev. 27, 2100 –2117. (doi:10.1016/j.quascirev.
2008.08.006)
37. Johnson MTJ, Stinchcombe JR. 2007 An emerging
synthesis between community ecology and
Proc. R. Soc. B 282: 20151116
1.
8
rspb.royalsocietypublishing.org
Authors’ contributions. R.M.-V. conceived the study, designed the
Downloaded from http://rspb.royalsocietypublishing.org/ on June 16, 2017
39.
41.
42.
43.
44.
45.
46.
47.
48.
49.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63. Mansion G, Rosenbaum G, Schoenenberger N,
Bacchetta G, Rossello JA, Conti E. 2008 Phylogenetic
analysis informed by geological history supports
multiple, sequential invasions of the Mediterranean
Basin by the angiosperm family Araceae. Syst. Biol.
57, 269 –285. (doi:10.1080/10635150802044029.)
64. Yesson C, Toomey NH, Culham A. 2009 Cyclamen:
time, sea and speciation biogeography using a
temporally calibrated phylogeny. J. Biogeogr. 36,
1234–1252. (doi:10.1111/j.1365-2699.2008.01971.x)
65. Salvo G, Ho SYW, Rosenbaum G, Ree R, Conti E. 2010
Tracing the temporal and spatial origins of island
endemics in the Mediterranean region: a case study
from the citrus family (Ruta L., Rutaceae). Syst. Biol.
59, 705–722. (doi:10.1093/sysbio/syq046)
66. Quézel P. 1987 Analysis of the Flora of
Mediterranean and Saharan Africa. Ann. Missouri
Bot. Garden 65, 479–534. (doi:10.2307/2398860)
67. Prinzing A, Durka W, Klotz S, Brandl R. 2001 The
niche of higher plants: evidence for phylogenetic
conservatism. Proc. R. Soc. Lond. B 268,
2383– 2389. (doi:10.1098/rspb.2001.1801)
68. Wiens JJ et al. 2010 Niche conservatism as an
emerging principle in ecology and conservation
biology. Ecol. Lett. 13, 1310 –1324. (doi:10.1111/j.
1461-0248.2010.01515.x)
69. Rosenbaum G, Lister GS, Duboz C. 2002
Reconstruction of the tectonic evolution of the
western Mediterranean since the Oligocene. J. Virtual
Explorer 8, 107–130. (doi:10.3809/jvirtex.2002.00053)
70. Brachert TC, Reuter M, Felis T, Kroeger KF, Lohmann
G, Micheels A, Fassoulas C. 2006 Porites corals from
Crete (Greece) open a window into Late Miocene
(10 Ma) seasonal and interannual climate
variability. Earth Planet. Sci. Lett. 245, 81 –94.
(doi:10.1016/j.epsl.2006.03.005)
71. Braga JC, Martı́n JM, Quesada C. 2003 Patterns and
average rates of late Neogene—recent uplift of the
Betic Cordillera, SE Spain. Geomorphology 50,
3–26. (doi:10.1016/S0169-555X(02)00205-2)
72. Molina-Venegas R, Aparicio A, Slingsby J, Lavergne
S, Arroyo J. 2015 Investigating the evolutionary
assembly of a Mediterranean biodiversity hotspot:
deep phylogenetic signal in the distribution of
eudicots across elevational belts. J. Biogeogr. 42,
507–518. (doi:10.1111/jbi.12398)
73. Fiz-Palacios O, Valcárcel V. 2013 From Messinian
crisis to Mediterranean climate: a temporal gap of
diversification recovered from multiple plant
phylogenies. Perspect. Plant Ecol. Evol. Syst. 15,
130–137. (doi:10.1016/j.ppees.2013.02.002)
74. Lancaster LT, Kay KM. 2013 Origin and
diversification of the California flora: re-examining
classic hypotheses with molecular phylogenies.
Evolution 67, 1041– 1054. (doi:10.1111/
evo.12016)
9
Proc. R. Soc. B 282: 20151116
40.
50.
Biogeogr. 17, 685 –695. (doi:10.1111/j.1466-8238.
2008.00410.x)
Kissling WD, Carl G. 2008 Spatial autocorrelation
and the selection of simultaneous autoregressive
models. Glob. Ecol. Biogeogr. 17, 59 –71. (doi:10.
1111/j.1466-8238.2007.00334.x)
Bini LM et al. 2009 Coefficient shifts in geographical
ecology: an empirical evaluation of spatial and nonspatial regression. Ecography 32, 193 –204. (doi:10.
1111/j.1600-0587.2009.05717.x)
Beale CM, Lennon JL, Yearsley JM, Brewer MJ,
Elston DA. 2010 Regression analysis of spatial data.
Ecol. Lett. 13, 246–264. (doi:10.1111/j.1461-0248.
2009.01422.x)
R Development Core Team. 2014 R: a language and
environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing.
Paradis E, Claude J, Strimmer K. 2004 APE: analyses
of phylogenetics and evolution in R language.
Bioinformatics 20, 289–290. (doi:10.1093/
bioinformatics/btg412)
Oksanen J et al. 2012 vegan: community ecology
package. R package v. 2.0-4. See http://cran.rproject.org/package¼vegan.
Bivand R. 2014 spdep: spatial dependence:
weighting schemes, statistics and models. R
package v. 0.5.71. See http://cran.r-project.org/
package¼spdep.
Bjornstad ON. 2013 ncf: spatial nonparametric
covariance functions. R package v. 1.1-5. See http://
cran.r-project.org/package¼ncf.
Cavender-Bares J, Kozak KH, Fine PVA, Kembel SW.
2009 The merging of community ecology and
phylogenetic biology. Ecol. Lett. 12, 693 –715.
(doi:10.1111/j.1461-0248.2009.01314.x)
Cano-Maqueda J, Talavera S, Arista M, Catalán P.
2008 Speciation and biogeographical history of the
Campanula lusitanica complex (Campanulaceae) in
the Western Mediterranean region. Taxon 57,
1252 –1266.
Escudero M, Vargas P, Valcárcel V, Luceño M. 2008 Strait
of Gibraltar: an effective gene-flow barrier for windpollinated Carex helodes (Cyperaceae) as revealed by
DNA sequences, AFLP, and cytogenetic variation.
Am. J. Bot. 95, 745–755. (doi:10.3732/ajb.2007342.)
Fernández-Mazuecos M, Vargas P. 2011 Historical
isolation versus recent long-distance connections
between Europe and Africa in bifid toadflaxes
(Linaria sect. Versicolores). PLoS ONE 6, e22234.
(doi:10.1371/journal.pone.002223)
Santos-Gally R, Vargas P, Arroyo J. 2012 Insights
into Neogene Mediterranean biogeography based
on phylogenetic relationships of mountain and
lowland lineages of Narcissus (Amaryllidaceae).
J. Biogeogr. 39, 782–798. (doi:10.1111/j.13652699.2011.02526.x)
rspb.royalsocietypublishing.org
38.
evolutionary biology. Trends Ecol. Evol. 22,
250–257. (doi:10.1016/j.tree.2007.01.014)
Tzedakis PC. 2007 Seven ambiguities in the
Mediterranean palaeoenvironmental narrative. Quat.
Sci. Rev. 26, 2042 –2066. (doi:10.1016/j.quascirev.
2007.03.014)
Omernik JM. 1995 Ecoregions: a spatial framework
for environmental management. In Biological
assessment and criteria: tools for water resource
planning and decision making (eds WS Davis,
TP Simon), pp. 49 –62. Boca Raton, FL: Lewis
Publishers.
Veech JA, Crist TO. 2007 Habitat and climate
heterogeneity maintain beta-diversity of birds
among landscapes within ecoregions. Glob. Ecol.
Biogeogr. 16, 650– 656. (doi:10.1111/j.1466-8238.
2007.00315.x)
Molina-Venegas R, Roquet C. 2014 Directional
biases in phylogenetic structure quantification: a
Mediterranean case study. Ecography 37, 572–580.
(doi:10.1111/j.1600-0587.2013.00442.x)
Roquet C, Thuiller W, Lavergne S. 2013 Building
megaphylogenies for macroecology: taking up the
challenge. Ecography 36, 13 –26. (doi:10.1111/j.
1600-0587.2012.07773.x)
Nee S. 2006 Birth–death models in macroevolution.
Annu. Rev. Ecol. Evol. System. 37, 1–17. (doi:10.1146/
annurev.ecolsys.37.091305.110035)
Bryant JA, Lamanna C, Morlon H, Kerkhoff AJ,
Enquist BJ, Green JL. 2008 Microbes on
mountainsides: contrasting elevational patterns of
bacterial and plant diversity. Proc. Natl Acad. Sci.
USA 105, 11 505–11 511. (doi:10.1073/pnas.
0801920105)
Qian H, Swenson NG, Zhang J. 2013 Phylogenetic
beta diversity of angiosperms in North America.
Glob. Ecol. Biogeogr. 22, 1152 –1161. (doi:10.1111/
geb.12076)
Kembel SW, Cowan PD, Helmus MR, Cornwell WK,
Morlon H, Ackerly DD, Blomberg SP, Webb CO. 2010
Picante: R tools for integrating phylogenies and
ecology. Bioinformatics 26, 1463–1464. (doi:10.
1093/bioinformatics/btq166)
Coyle JR, Halliday FW, López BE, Palmquist KA,
Wilfahrt PA, Hurlbert AH. 2014 Using trait and
phylogenetic diversity to evaluate the generality of
the stress-dominance hypothesis in eastern North
American tree communities. Ecography 37,
814–826. (doi:10.1111/ecog.00473)
Rivas-Martı́nez S, Asensi A, Dı́ez-Garretas B, Molero
J, Valle F. 1997 Biogeographical synthesis of
Andalusia (southern Spain). J. Biogeogr. 24,
915–928. (doi:10.1046/j.1365-2699.1997.00149.x)
Rodrı́guez-Sánchez F, Arroyo J. 2008 Reconstructing
the demise of Tethyan plants: climate-driven range
dynamics of Laurus since the Pliocene. Glob. Ecol.