Journal of Biogeography (J. Biogeogr.) (2011) 38, 664–680
ORIGINAL
ARTICLE
Early evolutionary history of the
flowering plant family Annonaceae:
steady diversification and boreotropical
geodispersal
Thomas L. P. Couvreur1*, Michael D. Pirie2, Lars W. Chatrou3, Richard M.
K. Saunders4, Yvonne C. F. Su4, James E. Richardson5,6 and Roy H. J. Erkens7
1
The New York Botanical Garden, 200th St
and Kazimiroff Blvd, Bronx, NY 10458-5126,
USA, 2Department of Biochemistry, University
of Stellenbosch, Private Bag X1, Matieland
7602, Western Cape, South Africa, 3Nationaal
Herbarium Nederland – Wageningen branch,
Biosystematics Group, Wageningen University,
Generaal Foulkesweg 37, 6703 BL,
Wageningen, The Netherlands, 4Division of
Ecology & Biodiversity, School of Biological
Sciences, The University of Hong Kong,
Pokfulam Road, Hong Kong, China, 5Royal
Botanic Garden Edinburgh, 20A Inverleith
Row, Edinburgh EH3 5LR, UK, 6Depto de
Ciencias Biológicas, Universidad de Los Andes,
Cra 1A No. 18A-10, Edificio J – Piso 4, Bogotá,
Colombia, 7Institute of Environmental Biology,
Ecology and Biodiversity group, Utrecht
University, Padualaan 8, 3584 CH Utrecht,
The Netherlands
ABSTRACT
Aim Rain forest-restricted plant families show disjunct distributions between the
three major tropical regions: South America, Africa and Asia. Explaining these
disjunctions has become an important challenge in biogeography. The
pantropical plant family Annonaceae is used to test hypotheses that might
explain diversification and distribution patterns in tropical biota: the museum
hypothesis (low extinction leading to steady accumulation of species); and
dispersal between Africa and Asia via Indian rafting versus boreotropical
geodispersal.
Location Tropics and boreotropics.
Methods Molecular age estimates were calculated using a Bayesian approach
based on 83% generic sampling representing all major lineages within the family,
seven chloroplast markers and two fossil calibrations. An analysis of
diversification was carried out, which included lineage-through-time (LTT)
plots and the calculation of diversification rates for genera and major clades.
Ancestral areas were reconstructed using a maximum likelihood approach that
implements the dispersal–extinction–cladogenesis model.
Results The LTT plots indicated a constant overall rate of diversification with
low extinction rates for the family during the first 80 Ma of its existence. The
highest diversification rates were inferred for several young genera such as
Desmopsis, Uvariopsis and Unonopsis. A boreotropical migration route was
supported over Indian rafting as the best fitting hypothesis to explain present-day
distribution patterns within the family.
*Correspondence: Thomas L. P. Couvreur, The
New York Botanical Garden, 200th St and
Kazimiroff Blvd, Bronx, NY 10458-5126, USA.
E-mails: [email protected];
[email protected]
664
Main conclusions Early diversification within Annonaceae fits the hypothesis
of a museum model of tropical diversification, with an overall steady increase in
lineages possibly due to low extinction rates. The present-day distribution of
species within the two largest clades of Annonaceae is the result of two
contrasting biogeographic histories. The ‘long-branch clade’ has been diversifying
since the beginning of the Cenozoic and underwent numerous geodispersals via
the boreotropics and several more recent long-distance dispersal events. In
contrast, the ‘short-branch clade’ dispersed once into Asia via the boreotropics
during the Early Miocene and further dispersal was limited.
Keywords
Biogeographic hypothesis testing, boreotropical hypothesis, diversification rates,
Indian rafting, K/Pg boundary, LTT plots, molecular dating, museum model.
http://wileyonlinelibrary.com/journal/jbi
doi:10.1111/j.1365-2699.2010.02434.x
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
INTRODUCTION
Tropical rain forests cover c. 7% of all continental land surface
and are the most biodiverse terrestrial ecosystems on the planet
(Whitmore, 1998; Morley, 2000). The South American rain
forests alone hold 30% of the world’s plant diversity (Smith
et al., 2004). Rain forests are currently restricted to a belt
spanning the equator in Southeast Asia, northern Australasia,
Africa, South/Central America, India–Sri Lanka and some
Pacific Islands. Plant families whose species are mainly
restricted to rain forests are frequently distributed across all
three continents, i.e. their distributions are disjunct, separated
by large expanses of ocean.
A number of hypotheses have been advanced to explain how
such intercontinental disjunctions could have originated.
These include vicariance resulting from the break up of the
Gondwanan supercontinent (Raven & Axelrod, 1974), concerted migrations across land bridges (van Steenis, 1962;
Thorne, 1972; Davis et al., 2002) during times when the
climate was favourable (geodispersal), continental rafting
(Conti et al., 2002), and long-distance dispersal (LDD)
(Renner et al., 2001). Providing evidence for any of these
hypotheses will give important insights into the evolutionary
history of tropical rain forest ecosystems. Analyses of diversification through time on the basis of dated molecular
phylogenies have been undertaken in several families within
most major clades of angiosperms (asterids, basal eudicots,
basal core eudicots, monocots and rosids). However, such
studies are lacking within the early diverging Magnoliales
(Angiosperm Phylogeny Group, 2009).
Annonaceae are a large pantropical family of trees and
lianas. Although this family (c. 2440 species; Table 1) is not as
diverse in terms of species richness as some core angiosperm
families (Rubiaceae, c. 13,000; Asteraceae, c. 25,000 species) it
is the most diverse family within the Magnoliales (Keßler,
1993). The family contributes significantly to tree diversity in
rain forests around the world (Gentry, 1993; Tchouto et al.,
2006; Punyasena et al., 2008) and provides one of the best
examples of tropical plant families for which abundance and
species richness, and temperature and rainfall, respectively, are
positively correlated (see Punyasena et al., 2008). This suggests
that the evolution of Annonaceae may be informative on the
historical development of tropical rain forests in general.
Table 1 Total number of species and genera in Annonaceae and
per major clade, and number (percentage) of these sampled for
this study.
Species
Anaxagorea
Ambavioids
Long-branch clade
Short-branch clade
Total
Genera
Total
Sampled
Total
Sampled
26
57
1631
726
2440
1
8
40
51
100
1
8
52
51
112
1
8
39
45
93
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
(3.8)
(14)
(2.5)
(7.2)
(4.8)
(100)
(100)
(75)
(88)
(83)
Molecular phylogenetic analyses within the family have
identified two main sister clades containing the majority of
species richness and initially characterized by their molecular
attributes (Richardson et al., 2004): the long-branch clade
(LBC) and the short-branch clade (SBC). On average, the LBC
has twice the level of chloroplast sequence divergence
compared with the SBC. In addition, the LBC contains more
than twice the species richness when compared with the SBC
(Table 1). Based on fossil data and past molecular date
estimates within the family, intercontinental disjunctions
between genera have been attributed mainly to boreotropical
geodispersal during the Late Eocene (Doyle & Le Thomas,
1997; Richardson et al., 2004). In the light of new age estimates
for major nodes within the family, geodispersal via Indian
rafting (McKenna, 1973; Morley, 2000) was suggested as being
an important factor in determining how this clade reached
Southeast Asia (Su & Saunders, 2009).
The goal of this study is to undertake a detailed analysis of
the diversification and general biogeographic history of
Annonaceae using DNA sequence data from nearly all genera
in the family. It represents the most comprehensive study on
diversification-rate patterns in the magnoliids so far. We tested
models involving shifts in rates of diversification (as opposed
to a more constant rate) and high versus low levels of
extinction (the latter corresponding to the ‘museum’ model, in
which low levels of extinction have been invoked to explain the
richness of tropical biota). Molecular age estimates calibrated
using the fossil record were used to discriminate between
alternative biogeographic hypotheses: boreotropical geodispersal, LDD and Indian rafting. These explanations are tested
explicitly within a maximum likelihood (ML) framework.
MATERIALS AND METHODS
Taxon and character sampling
For this study, we selected all genera with sequence data
available. Most genera were represented by one species,
although for genera that have been shown to be polyphyletic
(e.g. Polyalthia, Oxandra and Orophea; Mols et al., 2004), one
species per known monophyletic group was sampled. In total,
100 ingroup taxa were included, representing 93 out of the 112
(Table 1) currently recognized genera (83%). In addition, four
outgroup taxa were selected: Eupomatia bennettii (Eupomatiaceae are recovered as sister to Annonaceae; Qiu et al., 2000;
Sauquet et al., 2003), Degeneria vitiensis (Degeneriaceae),
Galbulimima belgraveana (Himantandraceae) and Magnolia
kobus (Magnoliaceae).
We adopted a supermatrix approach (de Queiroz & Gatesy,
2007), concatenating all available sequence data for the selected
taxa (mostly sourced from our own published work; Mols
et al., 2004; Richardson et al., 2004; Pirie et al., 2005, 2006;
Scharaschkin & Doyle, 2006; Erkens et al., 2007, 2009;
Couvreur et al., 2008a, 2009; Nakkuntod et al., 2009). Data
from seven chloroplast markers were available: three coding
(rbcL, matK and ndhF) and four non-coding (trnS–trnG,
665
T. L. P. Couvreur et al.
trnL–trnF, psbA–trnH and atpB–rbcL). GenBank numbers are
given in Appendix S1 in the Supporting Information.
The total length of the aligned molecular data matrix was
8823 characters. Gaps were not coded as separate indel
characters because binary characters could not be analysed
under the combined phylogenetic inference and molecular
dating approach (see below). In total, 47% of characters were
missing (including gaps; Appendix S2), with 55 taxa having up
to 50% of characters missing, and 49 with more than 50% of
characters missing. Woodiellantha and Pseudephedranthus had
the greatest proportion of missing characters with 83 and 80%,
respectively. Sequences of rbcL and trnL–trnF were available for
all taxa. In contrast, atpB–rbcL was available for only 40 taxa
(Appendix S2). Taxa with a high proportion of missing
characters can be included successfully in phylogenetic analyses
(Wiens, 2003; Wiens & Moen, 2008), and the supermatrix
approach is often used for family-wide phylogenies (Baker
et al., 2009; Couvreur et al., 2010) as it allows the inclusion of
data from disparate sources.
Fossil calibration
The present study used two fossils as minimum age constraints
for two separate nodes (see Discussion). The first fossil was
Endressinia (Mohr & Bernardes-de-Oliveira, 2004) from the
Aptian of Brazil of c. 115 Ma (Gradstein et al., 2004). The
phylogenetic positions of this and other relevant fossils were
evaluated recently by Doyle & Endress (2010) by optimizing
morphological characters over a molecular phylogeny of ‘basal’
angiosperms. Doyle & Endress (2010) showed that both
Endressinia and the more recent Archaeanthus (100 Ma)
(Dilcher & Crane, 1984) provide a minimum age for the
crown node of Magnoliineae (the clade comprising Magnoliaceae, Degeneria, Galbulimima, Eupomatia and Annonaceae;
Sauquet et al., 2003). Archaeanthus was widely used in
previous molecular dating analyses of Annonaceae (Doyle
et al., 2004; Richardson et al., 2004; Scharaschkin & Doyle,
2005; Pirie et al., 2006; Couvreur et al., 2008a; Erkens et al.,
2009; Su & Saunders, 2009), thus the use of Endressinia results
in a considerable increase in the minimum age for the
Magnoliineae crown node (Doyle & Endress, 2010; Pirie &
Doyle, in press). The second fossil, Futubanthus (Takahashi
et al., 2008), found in Japan and dating from 89 Ma, was
recognized as the earliest probable fossil of Annonaceae.
Futabanthus exhibits characters, such as a trimerous perianth
and numerous stamens and carpels, that are not unique in
Magnoliineae but that, in combination with the small size of
the stamens and their packing into a dome-like structure, may
be synapomorphies with Annonaceae as a whole (Pirie &
Doyle, in press). As it also lacks inner staminodes, a primitive
character found in Anaxagorea (sister to the rest of Annonaceae), Eupomatia, Degeneria and Galbulimima, Futabanthus is
associated with the crown rather than the stem node of the
family (Takahashi et al., 2008; Su & Saunders, 2009; Pirie &
Doyle, in press). Besides Futabanthus, no other fossils are
currently known that are both unequivocally associated with
666
nodes within Annonaceae and of an age that is likely to further
constrain the ages of those nodes (Pirie & Doyle, in press).
To calibrate the molecular tree, we used the older of the two
fossils (Endressinia), which is associated with the deepest node,
effectively to fix the age of the Magnoliineae crown node (the
root node in our analyses) using a uniform prior with minimal
bounds (115.1–114.9 Ma). This means that, assuming the
molecular method (see below) is accurate, all resulting ages
reported here should be interpreted as minima. We applied the
second, more recent fossil Futabanthus as a minimum
constraint at the crown node of Annonaceae by applying an
exponential prior distribution with a mean of 5 and a hard
bound offset of 89 (the age of the fossil). The mean value
allowed the tail of the distribution (soft bound) to reach
115 Ma (the fixed age of the root node), thus assuming that
Annonaceae could have originated at any time between 115
and 89 Ma, but biased towards the younger age.
Molecular dating
Molecular dating was undertaken using the beast package
(Drummond & Rambaut, 2007) version 1.5.3. Seven partitions
were created, one for each marker, using BEAUti 1.5.3. The
best performing evolutionary model for each partition was
determined using the Akaike information criterion (AIC;
Akaike, 1973) as implemented in MrModeltest (Nylander,
2004). The GTR+G model was selected for rbcL and ndhF,
whereas the GTR model + gamma + invariant sites model was
selected for the other five markers.
The dataset was run separately using a strict clock model
and an uncorrelated relaxed clock model (URC) assuming a
lognormal distribution of rates. We used a Bayes Factor (BF)
test as implemented in Tracer 1.5 (Rambaut & Drummond,
2003) to compare the two clock models statistically under the
smoothed marginal likelihood estimate and with 100 bootstrap
replicates (Suchard et al., 2001). The strict clock model was
significantly rejected over the relaxed clock model.
For each analysis, a total of 30 independent runs of 10
million generations, sampling every 1000th generation, were
undertaken on the online cluster of the Computational Biology
Service Unit from Cornell University (http://cbsuapps.tc.cornell.edu/beast.aspx). The starting tree for each independent
run was derived from an ML tree found using RAxML
(Stamatakis, 2006) and rendered ultrametric using the penalized likelihood method implemented in the program r8s
(Sanderson, 2003). The RAxML analysis was performed using
the web-server program version 7.0.4 (Stamatakis et al., 2008)
available at the CIPRES portal in San Diego (http://
www.phylo.org). Tracer 1.5 was used to check for convergence of the model likelihood and parameters between each
run. The resulting log files were combined (each independent
run with the first 10% samples as burn-in) using LogCombiner 1.5.2 (Rambaut & Drummond, 2003). Results were
considered reliable once the effective sampling size (ESS) for all
parameters exceeded 200 (see Results for the total number of
generations). For the tree files, we resampled from each run to
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
achieve a reduced sampling density of one in every 5000
generations. The maximum clade credibility topology with a
posterior probability limit of 0.85 and mean branch lengths
was then summarized from this sample using TreeAnnonator 1.5.2 (Rambaut & Drummond, 2003).
Diversification analyses
To provide an indication of net diversification rates within the
family, we generated a semi-logarithmic lineage-through-time
(LTT) plot of the resulting dated trees using the laser package
version 2.2 (Rabosky, 2006). A mean LTT plot with 95%
confidence intervals was computed from a collection of 1000
posterior trees sampled every 200,000 generations from the
beast output file (which contained a total of 270,000,000 trees
after burn-in).
Our species-level sampling of the family is very incomplete,
with 95.2% of all species missing (Table 1). However, our
sampling strategy was not random and 83% of genera were
represented, covering all the known major lineages of the
family. This has important implications for the interpretations
of the LTT plots: the older ends of the LTT plots should be
accurate, as most of the missing taxa should be attached to the
tree after the stem node of each genus. The missing genera are
mostly monotypic, representing only a fraction of the total
number of species (31). In order to test for a constant rate of
diversification under incomplete taxon sampling, null models
are generally constructed (Pybus & Harvey, 2000) where a
number of phylogenies are generated to a true standing species
richness of the group and then pruned randomly to the
number of taxa sampled. We argue that this method is not
appropriate in our case because of the non-randomness of our
taxon sampling (phylogenetically informed sampling; Cusimano & Renner, 2010). Rather, changes in rates of diversification can be interpreted directly from the resulting LTT plot
because, under this sampling strategy, the deeper nodes of the
phylogeny are well represented. For that, we define a threshold
up to which our LTT plot is considered accurate, that is, the
point at which our LTT plot will not be significantly influenced
by the sampling of extra tips (or species). This threshold is
arbitrarily defined to include 85% of all genera stem nodes.
This threshold was estimated for the whole family, the LBC
and the SBC.
In order to test the null hypothesis of no rate change
(constant diversification) versus a variable-rate change in
diversification, we used an ML approach implemented in the
laser package version 2.2 (Rabosky, 2006). The test statistic
for diversification rate-constancy is calculated as: DAICRC =
AICRC ) AICRV, where AICRC is the AIC score for the best
fitting rate-constant diversification model, and AICRV is the
AIC for the best fitting variable-rate diversification model.
Thus a negative value for DAICRV indicates that the data are
best approximated by a rate-constancy model. For each clade
(Annonaceae, LBC and SBC), we tested five different models,
of which two are rate-constant and three are rate-variable: (1)
the constant-rate birth model (the Yule process; Yule, 1924)
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
with speciation (k) and extinction (l) set to zero; (2) the
constant-rate birth–death model with two parameters, speciation (k) and extinction (l); (3) a pure birth rate-variable
model where the speciation rate k1 shifts to rate k2 at time ts,
with three parameters (k1, k2, ts); (4) an exponential
density-dependent speciation rate ‘DDX’ model; and (5) a
logistic density-dependent speciation rate ‘DDL’ model. The
LTT plot derived from the maximum clade credibility tree was
used for this part. In addition, we used the ‘truncateTree’
function to prune terminals of the LTT, corresponding to the
nodes found after the threshold age set as explained in the
previous paragraph.
Absolute net diversification rates (speciation minus extinction) were calculated for all non-monotypic genera by using
the stem age under a high level of extinction (b = 0.95) and no
extinction (b = 0) following Magallón & Sanderson (2001).
These rates were computed using the laser package version
2.2 (Rabosky, 2006). A standard boxplot was generated in
order to identify 10th and 90th percentile outliers. Diversification rates calculated from stem and crown nodes (when
available) of major clades in Annonaceae as well as for the
family were compared with values for angiosperms as a whole
and Magnoliales (relaxed values, obtained from Magallón &
Castillo, 2009). Genera of uncertain taxonomic status were
excluded from these analyses as their species richness estimation could be inaccurate (see Appendix S3).
Maximum likelihood ancestral area reconstruction
For this analysis, six taxa were pruned from the maximum
clade credibility (MCC) tree in line with our generic level
sampling (insofar as genera can be considered comparable) as
they are known to be nested within broader generic circumscriptions: Enicosanthum fuscum, Friesodielsia sp., Haplostichanthus longirostris, Orophea creaghii, Polyalthia cf. longifolia
and Pseudephedranthus fragrans.
Seven areas were delimited based on distribution data of
genera and continental divisions past and present: A, South
America; B, North/Central America; C, Africa; D, Madagascar;
E, India; F, Southeast Asia (west of Wallace’s Line); G,
Southeast Asia (east of Wallace’s Line), Melanesia and
northern Australia (Fig. 1).
Distribution data, taken from different taxonomic sources
(monographs and revisions), were used to assign each genus to
one or more areas (Fig. 2; Appendix S1). When data were
available, genera occurring in several areas were coded with the
inferred ancestral distribution of the genus (see Fig. 2 for
references to these studies) as recommended by Ronquist
(1996). For example, Guatteria is distributed in South and
Central America and Erkens et al. (2007) inferred that the
ancestral distribution of this genus was located in Central
America. Thus Guatteria was coded as operational unit B
(Central/North American).
Ancestral area reconstruction was performed using an ML
method under the divergence–extinction–cladogenesis model
(DEC; Ree et al., 2005; Ree & Smith, 2008) as implemented in
667
T. L. P. Couvreur et al.
Model 0: Unconstrained
A
B
A
B
C
D
E
F
G
C
1
0-100 Myr
D
E
F
1 0.01 0.01
1 0.01 0.01
1
1
1
G
1
1
1
1
1
1
1
1
1
0.01
1
B
E
F
C
A
G
D
Model 1: Dispersal into India viable
A
A
B
C
D
E
F
G
B
1
C
0.01
0.01
0-5 Myr
D
E
0.01 0.01
0.01 0.01
0.25 0.25
0.25
F
0.01
0.01
0.25
0.25
1
G
0.01
0.01
0.01
0.01
0.01
1
A
A
B
C
D
E
F
G
B
0.25
5-30 Myr
C
D
E
0.01 0.01 0.01
0.01 0.01 0.01
0.5
0.25
0.25
F
0.01
0.01
0.25
0.25
1
G
0.01
0.01
0.01
0.01
0.01
0.75
F
0.01
0.01
0.25
0.25
0.01
G
0.01
0.01
0.01
0.01
0.01
0.75
A
A
B
C
D
E
F
G
B
0.01
30-45 Myr
C
D
E
0.25 0.01 0.01
0.01 0.01 0.01
0.01
0.5
0.01
F
0.01
0.01
0.25
0.25
0.75
G
0.01
0.01
0.01
0.01
0.25
0.25
F
0.01
0.01
0.25
0.25
0.01
G
0.01
0.01
0.01
0.01
0.01
0.25
A
A
B
C
D
E
F
G
B
0.25
45-65 Myr
C
D
E
0.5
0.01 0.01
0.75 0.01 0.01
0.5
0.01
0.01
F
0.01
0.75
0.25
0.25
0.25
G
0.01
0.01
0.01
0.01
0.25
0.01
F
0.01
0.75
0.25
0.25
0.01
G
0.01
0.01
0.01
0.01
0.01
0.01
A
B
0.5
A
B
0.5
A
B
C
D
E
F
G
65-100 Myr
C
D
E
1
0.01 0.01
0.25 0.01 0.01
1
1
1
F
0.01
0.5
0.01
0.01
0.01
G
0.75
0.01
0.5
0.75
0.75
0.01
F
0.01
0.5
0.01
0.01
0.01
G
0.75
0.01
0.5
0.75
0.01
0.01
Model 2: Dispersal into India not viable
A
A
B
C
D
E
F
G
B
1
C
0.01
0.01
0-5 Myr
D
E
0.01 0.01
0.01 0.01
0.25 0.25
0.25
F
0.01
0.01
0.25
0.25
1
G
0.01
0.01
0.01
0.01
0.01
1
A
A
B
C
D
E
F
G
B
0.25
5-30 Myr
C
D
E
0.01 0.01 0.01
0.01 0.01 0.01
0.5
0.01
0.01
A
A
B
C
D
E
F
G
B
0.01
30-45 Myr
C
D
E
0.25 0.01 0.01
0.01 0.01 0.01
0.5
0.01
0.01
A
A
B
C
D
E
F
G
B
0.25
45-65 Myr
C
D
E
0.5
0.01 0.01
0.75 0.01 0.01
0.5
0.01
0.01
A
B
C
D
E
F
G
65-100 Myr
C
D
E
1
0.01 0.01
0.25 0.01 0.01
1
0.01
0.01
Figure 1 Delimitation of the seven areas assigned to the genera included and parameters of the three alternative biogeographic models
tested in this study. A, South America; B, Central/North America; C, Africa; D, Madagascar; E, Indian plate F, Mainland Asia/Sundaland/
west of Wallace’s Line; G, Sahul/Australia/Pacific/east of Wallace’s Line. Probabilities of dispersal: 0.01, none or low; 0.25, medium-low; 0.5,
medium; 0.75, medium-high; 1, high.
the software Lagrange build 20091004 (Ree & Smith, 2008).
Python scripts were all generated using the online helper
(http://www.reelab.net/lagrange). Ancestral areas were limited
to two, under the assumption that past distributions were
about as wide as current ones and to limit the complexity of
the analysis. Given the low dispersal ability of extant Annonaceae, we assume that dispersal was possible only between
adjacent areas (i.e. no dispersal allowed between areas A and D
or E, and between B and D or E). Ancestral areas were
estimated for five nodes relevant to this study (Fig. 2).
The goodness of fit of the data to three alternative
biogeographic models was evaluated in an ML framework:
model 0 was unconstrained with dispersal events between all
adjacent areas possible (probability = 1.0) during the whole
period considered (100–0 Ma). For models 1 and 2, we
incorporated prior information on range evolution as well as
dispersal probabilities between areas given discrete periods of
time. Both models were formulated based on past climatic
data, tectonic history and postulated presence/absence of land
bridges (Morley, 2000, 2003, 2007; Tiffney & Manchester,
2001; Zachos et al., 2001). Five time frames were delimited
(Fig. 1) and dispersal probabilities were assigned between all
adjacent areas. Dispersal probabilities were divided into five
categories: low or no dispersal = 0.01; low dispersal = 0.25;
medium dispersal = 0.5; high dispersal = 0.75; areas adjacent
or very close = 1. For model 1, dispersal into India from the
other areas coded was allowed and dispersal probabilities were
assigned as shown in Fig. 1. Model 2 was identical to model 1,
except that dispersal into India was not allowed from 100 to
5 Ma. We thus forced model 2 to represent the hypothesis that
Annonaceae did not successfully raft with India and that India
played no role in explaining present-day distribution patterns
(see Discussion). Finally, we investigated the influence of the
different number of categories of the dispersal probabilities on
the outcome of the analysis. We did this by analysing the data
under models 2 and 3 using just three probability categories
(instead of five) set to 0.01, 0.5 and 1. Differences between
models were assessed by directly comparing their respective
log-likelihoods. Following Ree et al. (2005), we used the
conventional cut-off value of two log-likelihood units to assess
statistical significance of likelihood differences.
RESULTS
Phylogeny
Weakly supported nodes (posterior probability, PP < 0.90) are
indicated by arrows in Fig. 2 and are mainly concentrated in
Figure 2 Maximum clade credibility tree of the Annonaceae with 95% highest probability density bars. Arrows indicate nodes with <0.9
posterior probability support. Numbers represent references to ancestral area coding for genera found in several areas. 1, Scharaschkin &
Doyle (2005); 2, L.W. Chatrou, pers. comm.; 3, Erkens et al. (2007); 4, Zhou et al. (2009); 5, Couvreur et al. (2008a); 6, Su & Saunders
(2009). Stars represent nodes used for calibrating the tree. Lower-case letters indicate nodes for which the ancestral areas were estimated
(Table 3). Letters next to taxon names represent coding of areas. When no box is present, the taxon was deleted from the Lagrange analysis.
668
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
d
c
b
e
a
Lower Cretaceous
120
110
Upper Cretaceous
100
90
80
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Palaeocene
70
60
Eocene
50
Oligocene
40
30
Miocene
20
E
F
F
E
G
E
E
F
E
B
B
B
B
F
B
F
E
F
E
F
E
F
F
F
F
F
C
F
A
A
A
A
A
A
A
A
A
A
A
A
A
A
F
C
C
C
C
C
F
F
C
E
C
F
C
C
D
F
C
C
C
C
C
C
C
C
C
C
C
C
C
B
F
A
F
C
C
B
C
C
A
A
C
C
A
A
A
A
C
F
D
C
A
F
F
C
C
A
G
F G
F G
F G
F G
F
F G
G
DF
G
F
D
DF
G
G
C
A
Enicosanthum fuscum
Polyalthia cf longifolia
Polyalthia xanthopetala
Neo uvaria paralellivenia
Sageraea lanceolata
Miliusa mollis
Fitzalania heteropetala
Meiogyne virgata
Mitrephora alba
Phaeanthus ebracteolatus
Popowia hirta
Tridimeris sp.
Sapranthus viridiflorus
Stenanona panamensis
Desmopsis microcarpa
Stelechocarpus burahol
Guamia sp.
Orophea creaghii
Orophea kerrii
Woodiellantha sp.
Orophea celebica
Platymitra macrocarpa
Alphonsea kinabaluensis
Marsypopetalum pallidum
Pseuduvaria pamattonis [6]
Haplostichanthus longirostris
Polyalthia suberosa
Trivalvaria macrophylla
Polyalthia stuhlmannii
Monocarpia euneura
Klarobelia inundata
Pseudephedranthus fragrans
Oxandra espintana
Pseudomalmea diclina
Oxandra macrophylla
Ephedranthus sp.
Ruizodendron ovale
Mosannona costaricensis
Pseudoxandra lucida
Cremastosperma brevipes
Malmea dielsiana
Bocageopsis canescens
Onychopetalum periquino
Unonopsis stipitata
Maasia sumatrana
Piptostigma mortehani
Polyceratocarpus pellegrini
Mwasumbia alba
Greenwayodendron oliveri
Annickia affinis
Friesodielsia desmoides
Desmos elegans
Dasymaschalon macrocalyx
Friesodielsia sp.
Monanthotaxis whytei
Melodorum fruticosum
Sphaerocoryne gracilis
Toussaintia orientalis
Mitrella kentii
Fissistigma glaucescens
Uvaria lucida [4]
Dielsiothamnus divaricatus
Hexalobus salicifolius
Uvariastrum insculptum
Asteranthe asterias
Monodora myristica
Isolona campanulata [5]
Uvariopsis vanderystii
Monocyclanthus vegnei
Uvariodendron molundense
Mischogyne michelioides
Sanrafaelia rufonammari
Ophrypetalum odoratum
Asimina triloba
Disepalum platypetalum
Annona muricata
Goniothalamus griffithii
Neostenanthera myristicifolia
Anonidium manii
Guatteria anomala [3]
Pseudartabotrys letestui
Letestudoxa bella
Duguetia staudtii
Fusaea peruviana
Xylopia peruviana
Artabotrys hexapetalus [2]
Hornschuchia citriodora
Trigynaea lanceipetala
Cymbopetalum brasiliense
Porcelia steinbachii
Mkilua fragrans
Mezzettia parviflora
Ambavia gerrardii
Cleistopholis glauca
Tetrameranthus duckei
Cananga odorata
Cyathocalyx martabanicus
Lettowianthus stellatus
Meiocarpidium lepidotum
Anaxagorea silvatica [1]
Eupomatia bennettii
Degeneria vitiensis
Galbulimima belgraveana
Magnolia kobus
miliusoid
clade
Shortbranch
clade
South/Central
American
clade
African
SBC
Uvaria
clade
African
LBC
Longbranch
clade
annonoid
clade
Duguetia
clade
Bocageeae
Ambavioids
Anaxagorea
Outgroups
PL P
10
0
669
T. L. P. Couvreur et al.
Table 2 Differences in age (Ma) between recent Annonaceae family-wide molecular dating studies.
Penalized likelihood method
beast
Age (Ma) (SD in parentheses)
Mean age (95% HPD in parentheses)
Clade
Pirie & Doyle (in press)
Su & Saunders (2009)
This study
Annonaceae stem age
Annonaceae crown age
Ambavioids/LBC/SBC crown
Ambavioids crown age
LBC/SBC crown
LBC crown age
SBC crown age
SAC/miliusoid clade crown
SAC crown age
Miliusoid crown age
90.9
75.5
65.3
–
62.8
57.6
55.3
45.5
44.0
40.3
98.0
89.4
74.6
–
67.3
59.6
39.8
29.1
24.3
23.0
106.29
90.44
76.08
69.40
71.71
65.85
32.768
23.60
21.38
19.72
(1.6)
(1.7)
(1.9)
(1.9)
(2.1)
(2.4)
(2.9)
(2.5)
(3.4)
(101.46–94.91)
(90.41–89) (cal)
(84.4–63.6)
(78.1–55.2)
(70.5–48.1)
(55.1–26.8)
(39.2–20.2)
(37.7–15.7)
(31.1–16)
(110.37–102.02)
(92.98–89) (cal)
(82.5–69)
(60.2–78)
(78.26–64.77)
(72.42–59.16)
(40–25.8)
(28.79–19.16)
(26.09–16.85)
(24.04–15.89)
HPD, highest posterior density; SD, standard deviation; LBC, long-branch clade of Annonaceae; SBC, short-branch clade of Annonaceae; cal, node
used for calibration.
the SBC. In general, the topology and support values agree well
with previous phylogenetic analyses (Mols et al., 2004; Richardson et al., 2004; Pirie et al., 2006; Couvreur et al., 2008b;
Erkens et al., 2009; Su & Saunders, 2009).
Molecular dating
The maximum clade credibility tree is presented in Fig. 2. A
total of 300 million generations (30 runs of 10 million
generations each) were necessary to reach sufficient ESS. All
parameters between independent runs were equal. For all runs,
stationarity was achieved relatively quickly, within 100,000
generations, which was presumably the result of using a nearly
optimal starting tree. In general, most parameters and age
estimations reached ESS values >200, and all reached ESS
values >150.
The dataset was strongly non-clock-like according to a BF
test (ln BF > 1000 in favour of relaxed clock analysis), thus
only the results using the relaxed clock model are presented
here. For all five partitions, the rate of covariance was centred
on zero, which can be interpreted as a lack of evidence for rate
autocorrelation among lineages (Peng et al., 2006; Drummond
& Rambaut, 2007; Ho, 2009). The first 200 sampled trees of
each run were treated as burn-in and deleted, and the
remaining 16,000 in total were combined into a single file.
Ages of major clades obtained for this study are given in
Table 2.
Diversification
The thresholds for the LTT plots were calculated as 25 Ma for
the whole family, 31 Ma for the long-branch clade and 16 Ma
for the short-branch clade (Fig. 3). Following our assumptions, these dates represent the point in time before which our
LTTs are assumed to depict a fairly accurate picture of
diversification rates. Overall, the family level and LBC LTT
plots were fairly linear. At the family and clade level (LBC and
670
SBC), the data better fitted a constant rate model of
diversification with the diversification rate-constancy statistic
DAICRV being )1.05 (Annonaceae), )1.0 (LBC) and )3.4
(SBC) when compared with the second-best model. In all three
cases, the pure birth model was identified as having the lowest
AIC value amongst the other models tested.
The total number of genera analysed for the diversification
rate was 75. Eight genera were outliers with higher rates (>90th
percentile; Appendix S3; Fig. 4a) under no extinction (b = 0).
When extinction was factored in, eight genera were outliers
with higher rates (>90th percentile; Appendix S3; Fig. 4b). In
addition, nine genera (under both b = 0 and b = 0.95) were
outliers with lower rates (>10th percentile; Appendix S3;
Fig. 4a). Diversification rates calculated for the stem and
crown nodes within Annonaceae were more or less the same,
except in the case of the SBC, where a significant difference was
seen (Fig. 5, grey panel). The diversification rate of Annonaceae as a whole (Fig. 5) was higher than that of angiosperms
and the Magnoliales (Magallón & Castillo, 2009). At the level
of the major clades within the family, the LBC and SBC had the
highest diversification rates (Fig. 5), while Anaxagorea and the
ambavioids had the lowest. The SBC had a higher diversification rate when the crown node, as opposed to the stem node,
was used for the calculation.
Biogeographic reconstruction analyses
All models returned similar ancestral areas for the five nodes
(Table 3). Model 2, in which rafting with India was not
allowed, returned a significantly better likelihood score compared with models 0 and 1. Lowering the number of
probabilities to three had no major effects on the results
under model 2 (Table 3), indicating that our choice of
dispersal probability categories did not significantly influence
the results.
We coded widespread genera according to the ancestral
areas inferred in previous studies, with a single exception: in
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
Ln number of lineages
(a)
1000
Clade ages within Annonaceae
10
Ln number of lineages
Ln number of lineages
100
80
60
40
20
0
long-branch clade
1000
100
10
1
80
(c)
DISCUSSION
100
1
120
(b)
Xylopia and, as the overall results for nodes focal to this study
were otherwise unaffected (data not shown), we used this
coding in all further analyses.
Annonaceae
60
1000
40
20
0
20
0
short-branch clade
100
10
1
80
60
40
Millions of years ago (Ma)
Figure 3 Semi-logarithmic lineage-through-time (LTT) plots
within (a) Annonaceae, (b) long-branch and (c) short-branch
clades. Triangles: mean LTT plot from 1000 posterior trees; solid
lines represent 95% confidence limits. Squares on the right represent the extant number of taxa for each clade. The grey column
in (a) represents the Cretaceous–Palaeogene boundary, that in (b,
c) represents the Early Eocene Climatic Optimum event. Vertical
bars represent threshold limit of LTT plots indicating where the
plot is assumed to be accurate (not influenced by missing taxa).
the absence of a detailed phylogeny, this was not possible for
the large pantropical genus Xylopia. We thus undertook three
analyses in order to test the impact of different plausible
ancestral areas for Xylopia: Africa, South America and
Southeast Asia. Of the three reconstructions, the best likelihood score was obtained with Africa as the ancestral area for
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
In general, the minimum ages estimated here using the URC as
implemented in beast are slightly older for deeper parts of the
chronogram, and significantly younger for shallower parts of
the chronogram, than those of Pirie & Doyle (in press),
generated using the same calibrations but with the penalized
likelihood method of Sanderson (2002), which assumes
autocorrelation of rates (Table 2). The discrepancies between
some previous estimates (Richardson et al., 2004; Pirie &
Doyle, in press) and those presented here could be related to
several different factors, such as taxon sampling density
(Linder et al., 2005; Pirie et al., 2005), molecular marker
sampling (Magallón & Sanderson, 2005), alternative dating
methods (Drummond et al., 2006), and the use of different
fossils as calibration points (in particular compared with
studies of Doyle et al., 2004; Richardson et al., 2004; Pirie
et al., 2006 that used the relatively recent fossil Archaeanthus).
Su & Saunders (2009) also used the URC model with a
comparable age calibration and recovered age estimations
similar to those presented here (Table 2) despite having only
approximately one-third of the taxon sampling included here.
The molecular dating method may thus explain at least some
of the inconsistencies, which should therefore be interpreted in
terms of the fit of the underlying assumptions to the data (rate
autocorrelation versus uncorrelated distribution of rates;
Drummond et al., 2006). The URC model is more general: it
neither assumes nor excludes rate autocorrelation, and in fact
provides a means to test for autocorrelated rates as indicated
by the posterior distribution of the covariance parameter
(Drummond et al., 2006; although a possible bias against
discovering rate autocorrelation in this manner has been
suggested, e.g. Moore & Donoghue, 2007). As our results do
not indicate significant rate autocorrelation in Annonaceae, we
interpret the ages derived under the URC model as being more
plausible.
Annonaceae originated at least 110–102 Ma [95% highest
posterior density (HPD), stem node] and started to diversify
by c. 89 Ma. Our mean (minimum) estimates for the three
major lineages are c. 69 Ma (crown node ambavioids),
c. 66 Ma (crown node LBC) and c. 33 Ma (crown node
SBC). Besides the major clades, the age estimations for stem
nodes of genera reported here correspond closely with previous
analyses under higher levels of infra-generic species sampling
(Table 2).
Diversification at family and generic levels
The Cretaceous–Palaeogene (K/Pg) boundary mass extinction
at 65 Ma had profound effects on marine biodiversity and
671
(a)
0.6
Diversification rates rate b=0
T. L. P. Couvreur et al.
0.5
Uvariopsis *
Unonopsis
*
Desmopsis *
Stenanona *
0.4
Sphaerocoryne *
Desmos*, Friesodielsia*
Monanthotaxis *
0.3
Pseuduvaria
Uvaria
Annona
0.2
Goniothalamus
Guatteria
Xylopia
Duguetia
0.1
Anaxagorea
Artabotrys
0
0
Diversification rates with b=0.95
(b)
20
0.16
40
60
80
100
Unonopsis *
0.14
*
Monanthotaxis
0.12
Uvariopsis
0.1
0.08
*
Uvaria *
Desmopsis *
Pseuduvaria*
Desmos*, Friesodielsia*
Fissistigma
Stenanona
Annona
Dasymachalon
Guatteria
Goniothalamus
0.06
Xylopia
Duguetia
0.04
Artabotrys
0.02
Anaxagorea
0
0
20
40
60
80
100
Millions of years since origin
0.25
400
SBC
350
250
0.15
LBC
0.10
Annonaceae
Magnoliales
Anaxagorea
b, Diversification rate
300
Angiosperms
200
150
Ambavioids
100
672
Stem
Crown
Stem
Crown
Stem
Crown
Stem
Stem
Crown
Stem
Crown
Stem
0.00
Crown
50
0
Millions of years ago (Ma)
0.20
0.05
Figure 4 Diversification rates of genera as a
function of time for (a) no extinction (b = 0)
and (b) high extinction (b = 0.95). Circle size
is proportional to species richness within
each genus. Asterisks indicate genera for
which diversification rates were identified as
90th percentile outliers.
Figure 5 Diversification rates for several
plant groups as a function of the estimated
stem and crown nodes. Left axis, diversification rate: no extinction (b = 0; diamonds);
high extinction (b = 0.95; squares). Right
axis, mean age of stem and crown nodes in
Ma (grey triangles). SBC, short-branch clade
of Annonaceae; LBC, long-branch clade of
Annonaceae.
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
0.0039
0.0167
0.0186
0.01392
)196
)191
)187
)190
0
1
2
Five dispersal probabilities
Three dispersal probabilities
0.001546
0.0017
0.0016
0.0020
Extinction
[A|AC] 0.72
[A|C] 0.18
[A|AC] 0.6
[A|C] 0.24
[A|AC] 0.65
[A|C] 0.21
[A|C] 0.38
[A|AC] 0.21
[AC|C] 0.12
[C|C] 0.08
[A|A] 0.06
Crown Annonaceae [a]
[C|AC] 0.41
[AC|C] 0.3
[C|C] 0.25
[C|AC] 0.37
[C|C] 0.37
[AC|C] 0.23
[C|AC] 0.4
[C|C] 0.31
[C|C] 0.64
[C|AC] 0.21
Crown LBC/SBC/ambavioid [b]
[C|C] 0.54
[AC|C] 0.38
[C|C] 0.59
[AC|C] 0.31
[C|C] 0.55
[AC|C] 0.34
[C|C] 0.71
[C|AC] 0.11
Crown LBC/SBC [c]
[C|F] 0.9
[C|F] 0.86
[C|F] 0.86
[C|A] 0.53
[C|F] 0.42
Crown SBC [d]
[C|C] 0.56
[AC|C] 0.39
[C|C] 0.61
[AC|C] 0.33
[C|C] 0.59
[AC|C] 0.36
[C|C] 0.80
[AC|C] 0.14
Crown LBC [e]
Model 0: unconstrained; model 1: dispersal into India possible; model 2: dispersal into India not allowed. Model 2 was further analysed using five and three dispersal probabilities (see text for details). Letters
represent alternative ancestral area reconstructions that fall within two log-likelihood units of the optimal scenario. Following Fig. 2, the vertical bar separates the ancestral area reconstructed for the lower
branch (left letter) from the one reconstructed for the upper branch (right letter) arising from the node. Single-area letters indicate an ancestor restricted to a single area; two-area letters indicate an ancestor
restricted to two areas. Values represent the relative probability of that inference. Bold text represents the model with a significantly better likelihood when compared with the other models tested (more than
two log-likelihoods better).
ln L, log-likelihood of the tested model; LBC, long-branch clade of Annonaceae; SBC, short-branch clade of Annonaceae.
Dispersal
ln L
Model
Table 3 Maximum likelihood reconstruction for major nodes within Annonaceae under three different models of geographic range evolution.
Evolutionary history of the Annonaceae
673
T. L. P. Couvreur et al.
terrestrial animals, such as the extinction of non-avian
dinosaurs. However, the effect of such a mass extinction on
plant terrestrial ecosystems has been controversial (McElwain
& Punyasena, 2007). Some authors argue that, in comparison
with animals or marine ecosystems, plant communities were
little affected by such drastic events (Traverse, 1988). Other
authors indicate, mainly based on the plant fossil record, that
the K/Pg boundary resulted in massive extinction of land
plants (Wolfe & Upchurch, 1986; Wilf & Johnson, 2004;
Nichols & Johnson, 2008). One major difference noted for
plants is that, even though species extinction could have been
relatively high, it did not eradicate entire genera or families
(McElwain & Punyasena, 2007; Nichols & Johnson, 2008). In
our analysis of diversification through time in Annonaceae, the
LTT plot reveals a brief pause in diversification (gap) after the
K/Pg mass extinction (Fig. 3a). However, after a mass extinction event, LTT plots should reveal a clear anti-sigmoid curve
(Harvey et al., 1994; Crisp & Cook, 2009), which is not
observed here. Thus our data do not indicate that Annonaceae
diversification was severely affected by the events of the K/Pg
boundary. It could be that tropical rain forests proved more
resilient to such drastic changes, despite the fact that some
studies suggested that evergreen taxa were more affected than
deciduous taxa during those times (Wolfe & Upchurch, 1986;
McElwain & Punyasena, 2007). Our results, however, pertain
to the origination of major lineages, and the effects of the K/Pg
boundary at the specific level would have to be investigated
more closely using the fossil record. Unfortunately, studies of
fossil flora diversity in tropical regions across the K/Pg
boundary are rare, being more centred in North America
(McElwain & Punyasena, 2007; Nichols & Johnson, 2008).
The overall shapes of the LTT plots for the whole family, the
LBC and the SBC are otherwise fairly linear (Fig. 3). Such
patterns are generally associated with a constant rate of
diversification (Nee et al., 1994; Nee, 2006; Ricklefs, 2007).
This was confirmed by the test statistic for diversification rateconstancy (DAICRC), which was negative at the family and
clade (LBC and SBC) levels. Thus the overall picture indicates
no major shifts of diversification throughout the family’s
evolutionary history (at least up to 25 Ma). Interestingly,
equivalent family or higher-order level LTT plots within
different plant groups do not show an early diversification
scenario consistent with a constant diversification rate. Diversification analyses within clades such as Burmanniaceae
(Merckx et al., 2008), Brassicaceae (Couvreur et al., 2010),
Proteaceae (Sauquet et al., 2009) and the rosid clade (Wang
et al., 2009) all indicate an early and rapid increase in
diversification followed by a decrease. In addition, the order
Malpighiales was shown to have undergone a rapid increase in
diversification shortly after its origin c. 114 Ma (Davis et al.,
2005). Such a contrast could be the result of several factors,
possibly linked to the strictly tropical ecology of Annonaceae.
One hypothesis advanced to explain high levels of diversity in
tropical ecosystems is the ‘museum model’, whereby lineages
accumulated steadily through time due to low background
extinction (Stebbins, 1974). This pattern is supported by the
674
fact that the diversification model that best fitted our data was
the pure birth model, where the speciation rate is constant and
the extinction rate is fixed to 0. Such a pattern was also
recorded for the liverwort family Lejeuneaceae (Wilson et al.,
2007), although this family had representatives in more
temperate as well as tropical climates. More data should be
gathered from other plant families with similar tropical
distributions to confirm this.
Finally, despite a common origin, it is apparent in Fig. 2
that diversification within the LBC and SBC started at
significantly different times. Based on crown node ages,
present-day species richness within the SBC appears to be
the result of higher diversification rates when compared with
the LBC, although diversification rates are similar when the
stem node is used (Fig. 5).
As we indicated above, the estimation of the minimum stem
node age of genera is reliable across the family, even in the
absence of dense taxon sampling within each genus. Thus
diversification rates taken from stem node ages should not be
sensitive to taxon undersampling within genera. Under a zero
extinction model, the seven largest genera of Annonaceae (>95
species: Annona, Artabotrys, Duguetia, Goniothalamus, Guatteria, Uvaria, Xylopia) do not have the highest diversification
rates within the family (Fig. 4; Appendix S3). This was also the
case under a high extinction-rate model, with the exception of
Uvaria, for which the rate estimation increased considerably
(Fig. 4; Appendix S3). The largest genus associated with the
highest levels of diversification with or without extinction was
the African liana genus Monanthotaxis (Fig. 4a; Appendix S3),
with c. 86 species. The highest diversification rates were instead
identified in smaller genera with stem nodes younger than
8 Ma (Fig. 4; Appendix S3). The four fastest-diversifying
genera under no extinction are Uvariopsis, Unonopsis, Desmopsis and Stenanona (Fig. 4; Appendix S3), representing 17, 48,
18 and 13 described species, respectively. Net diversification of
a clade is a function between species and the age of that clade
(Magallón & Sanderson, 2001), and in Annonaceae the larger
genera are not associated with the highest diversification rates
(large circles concentrated to the right in Fig. 4). Such a
pattern was also found at the angiosperm level (Magallón &
Sanderson, 2001). Diversification rates of these large genera
could, however, be underestimated because they were calculated using stem ages. If their crown ages are estimated to be
young, this would lead to higher diversification rates (Magallón & Castillo, 2009). For example, Linder (2008) regarded
Guatteria as a recent and rapid radiation (in which recent
speciation overwhelmed extinction) based on its age and
species richness. In a well sampled species-level molecular
phylogeny of the genus, the crown node age was estimated to
be 11.4 (±1.4) Ma (Erkens et al., 2007). This would provide a
diversification rate of 0.44 under no extinction, placing it
amongst the fastest-diversifying genera in Annonaceae
(Appendix S3; Fig. 4). Goniothalamus, a Southeast Asian genus
with c. 150 species (Nakkuntod et al., 2009), was found to have
the most significant rate increase along its stem branch
(Erkens, 2007). Based on very restricted sampling of species
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
within the genus, the crown node was estimated to be fairly
young, varying from 4.8 to 3.6 Ma (three species sampled;
Richardson et al., 2004) to c. 10 Ma (two species sampled; Su
& Saunders, 2009). This could indicate a rapid radiation within
Goniothalamus, which we would fail to identify here on the
basis of its stem node age alone. Further sampling of taxa is
likely to identify more distantly related species, thus increasing
the estimated age of the crown node of Goniothalamus.
A dated species-level phylogeny is necessary to reconstruct
the diversification history of the family during more recent
times and beyond our arbitrary threshold. Many geological
events have occurred from the Miocene (23 Ma) onwards that
could have influenced diversification rates within the family,
including major uplifts of the Andes, the New Guinea highlands
and the East African rift system; the collision of Sundanian and
Papuasian shelves around Wallace’s Line; and closure of the
Isthmus of Panama and the Tethys Sea. The combination of
these events could have had a profound impact on rates of
diversification and extinction in the tropics along with all other
parts of the world. For example, diversification in the genus Inga
(Fabaceae), which comprises c. 350 species, has been shown to
have been concentrated within the past 10 Myr (Richardson
et al., 2001), over a similar period to that of Guatteria (if not
younger). Substantial diversification of tropical plant lineages
has therefore taken place during comparatively recent times and
could have been caused by some of the events discussed above.
Modern tropical species diversity therefore seems to be the
result of gradual accumulation of lineages through time, as
suggested by the ‘museum hypothesis’ coupled with rapid and
comparatively recent diversification in a few lineages. Similar
patterns are evident in studies of tropical lineages of leaf beetles
(McKenna & Farrell, 2006).
Origins and diversification of the long- and
short-branch clades
Both the LBC and SBC are inferred to have an African
ancestral area. However, our results reveal that these lineages
have different patterns of diversification in both space and
time (Fig. 2; Fig. 3b,c). During the Early Eocene Climatic
Optimum (EECO), numerous speciation events leading to
extant taxa are observed within the LBC, whereas for the SBC
there is a period of apparent stasis (Fig. 2). Speciation in the
SBC is concentrated at more recent nodes of the miliusoid and
Neotropical clades (c. 30 Ma), as is apparent in the SBC LTT
plot (Fig. 3c).
The SBC has more geographic structure than the LBC. In the
former we have four major clades confined to specific regions:
the African short-branch clade (Couvreur et al., 2009)
restricted to Africa; the South/Central American clade (SAC)
restricted to the Neotropics (Pirie et al., 2006); the miliusoid
clade restricted mainly to Southeast Asia (Mols et al., 2004);
and the weakly supported Central American clade (which also
includes the Southeast Asian genus Stelechocarpus) nested
within the miliusoid clade. Finally, the genus Maasia is
restricted to Asia (Mols et al., 2008). The spatial and temporal
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
differences between the LBC and SBC could reflect alternative
diversification responses to the overall change in ecology and
climate that affected rain forest taxa during the Cenozoic.
Disjunct distributions between the tropics of South America,
Africa and Southeast Asia of clades younger than Gondwanan
break-up are currently often assumed to be the result of a
migration through Laurasia during the EECO, when climatic
conditions supported tropical vegetation at those latitudes
(Wolfe, 1975; Lavin & Luckow, 1993). Indeed, the connections
and climatic conditions at that time would have facilitated
migration between the tropical floras of Africa, Indochina and
North America. Molecular dating analyses, for example in
Malpighiaceae (Davis et al., 2002), Meliaceae (Muellner et al.,
2006), Melastomataceae s. str. (Renner et al., 2001) and Rubiaceae (Antonelli et al., 2009), all indicate that these families
have tropical disjunctions with lineages splitting up at a time
consistent with the disruption of boreotropical ranges around
the Eocene–Oligocene boundary. Such a scenario appears to fit
well with the patterns and tempo of diversification found in the
LBC (e.g. Erkens et al., 2009), where numerous splitting events
occurred during the EECO as well as from 45 Ma until 32 Ma
when global temperatures were dropping (Zachos et al., 2001).
Such widespread distribution of tropical rain forests in
northern latitudes would have allowed migration of lineages
amongst Southeast Asia, the New World and Africa, resulting in
a lack of strong geographic clustering within the LBC. Most of
the intercontinental splits found in the LBC are estimated to
have originated during the Eocene (Fig. 2).
Based on a prior age estimation of c. 62–53 Ma for the crown
node of the SBC (Richardson et al., 2004), it was suggested that
distribution patterns within this clade also resulted from
boreotropical geodispersal (Richardson et al., 2004; Pirie et al.,
2006). Such claims could be supported by the presence of seeds
with typical SBC-type spiniform ruminations in the London
Clay, at 50 Ma (Reid & Chandler, 1933; Doyle et al., 2004), but
the dates recovered here for the crown node of the SBC
(32.8 Ma, 95% HPD 40.1–26.1) are more recent. This may
suggest that these seeds represent stem SBC lineages, rather
than the crown group, and that they are therefore not the direct
ancestors of taxa that subsequently dispersed out of Africa. An
alternative explanation, suggested by Su & Saunders (2009), is
that the SBC dispersed ‘out of India’ into Southeast Asia
(McKenna, 1973; Morley, 2000). The out-of-India hypothesis
suggests that plants inhabiting the Deccan plate (India) prior to
its separation from Gondwana ‘rafted’ across the Indian Ocean
during the Cretaceous and then dispersed into the newly
contiguous landmass of Southeast Asia when the plates
collided. Corroborative evidence has been inferred from frogs
(Bossuyt & Milinkovitch, 2001) and plant families such as
Crypteroniaceae (Conti et al., 2002; Rutschmann et al., 2004),
Dipterocarpaceae (Morley, 2003), Melastomataceae (Morley &
Dick, 2003), and Lowiaceae (Kress & Specht, 2006). Evidence
potentially supporting this hypothesis with regard to the SBC
includes putative Annonaceae fossils found in the Maastrichtian of India (Bonde, 1993), indicating the presence of the
family in India before its collision with Asia; and the estimated
675
T. L. P. Couvreur et al.
timing of dispersal of the SBC into Southeast Asia (30–19 Ma;
Fig. 2), which post-dates the collision of the Indian plate with
the Asian continent c. 35 Ma (Ali & Aitchison, 2008).
The results of the ML analysis provide significant support
for the boreotropical dispersal route over the Indian rafting
route. Our model 2, in which India played no role in
explaining the present-day distribution of Annonaceae species,
fitted the data significantly better than the two other models
(M0 and M1), being 4 and 9 log units better, respectively
(Table 3). Thus, given the phylogeny and the biogeographic
models used here, the boreotropics would have been the main
dispersal route of the SBC into Southeast Asia. The fossil taxon
found in India during the Maastrichtian could represent a
lineage that subsequently went extinct during the rapid
movement of India across the Indian Ocean (see below); and
because its rumination is lamelliform, rather than the spiniform type restricted to the SBC, it might even represent a
different lineage altogether. The split between the African
short-branch clade and the rest of the SBC, estimated at
32 Ma, could be explained by the drastic Oligocene global
temperature drop, which affected rain forest vegetation worldwide (Morley, 2000; Zachos et al., 2001), effectively breaking
up any direct tropical rain forest connection between Africa
and Southeast Asia (via Europe). However, during the Late
Miocene, changes towards more seasonal climates led to the
disappearance of rain forests across much of India, restricting
them to western parts of the country and Sri Lanka (Morley,
2000). These changes potentially could have led to numerous
extinctions, ultimately weakening the signal in the data (India
contains no endemic Annonaceae genera today). Future
support for either hypothesis should come from a comparative
approach comparing several species-level dated molecular
phylogenies of genera in both Southeast Asia and India. Under
a boreotropical hypothesis, we would expect most Indian
species to be nested within the Southeast Asian ones, whereas
under the Indian rafting hypothesis, we would expect the
Southeast Asian species to be nested within Indian lineages.
The LBC and SBC thus provide contrasting evolutionary
scenarios. The LBC diversified from the Early Palaeogene and
migrated by geodispersal from Africa through the boreotropics
into other major tropical regions, resulting in lower geographic
clustering. In contrast, the SBC started to diversify only during
the Early Oligocene and mid Miocene, and migration events
were limited, resulting in strong geographic clustering.
ACKNOWLEDGEMENTS
We thank Pauline Ladiges, Jim Doyle, Angela Whittaker and
an anonymous reviewer for their comments and improvements to the article. Jing Wang is also thanked for providing
sequence data for Desmos.
REFERENCES
Akaike, H. (1973) Information theory as an extension of
the maximum likelihood principle. Second International
676
Symposium on Information Theory (ed. by B.N. Petrov and F.
Csaki), pp. 267–281. Akademiai Kiado, Budapest.
Ali, J.R. & Aitchison, J.C. (2008) Gondwana to Asia: plate
tectonics, paleogeography and the biological connectivity of
the Indian sub-continent from the Middle Jurassic through
latest Eocene (166–35 Ma). Earth-Science Reviews, 88,
145–166.
Angiosperm Phylogeny Group (2009) An update of the
Angiosperm Phylogeny Group classification for the orders
and families of flowering plants: APG III. Botanical Journal
of the Linnean Society, 161, 105–121.
Antonelli, A., Nylander, J.A.A., Persson, C. & Sanmartin, I.
(2009) Tracing the impact of the Andean uplift on Neotropical plant evolution. Proceedings of the National Academy of Sciences USA, 106, 9749–9754.
Baker, W.J., Savolainen, V., Asmussen-Lange, C.B., Chase,
M.W., Dransfield, J., Forest, F., Harley, M.M., Uhl, N.W. &
Wilkinson, M. (2009) Complete generic-level phylogenetic
analyses of palms (Arecaceae) with comparisons of supertree
and supermatrix approaches. Systematic Biology, 58,
240–256.
Bonde, S.D. (1993) Unonaspermum corneri gen. et sp. nov., an
annonaceous seed from the Deccan Intertrappean beds of
India. Journal of the Indian Botanical Society, 72, 251–253.
Bossuyt, F. & Milinkovitch, M.C. (2001) Amphibians as indicators of Early Tertiary ‘‘out-of-India’’ dispersal of vertebrates. Science, 292, 93–95.
Conti, E., Eriksson, T., Schonenberger, J., Sytsma, K.J.,
Baum, D.A. & Orti, G. (2002) Early Tertiary out-of-India
dispersal of the Crypteroniaceae: evidence from phylogeny
and molecular dating. Evolution, 56, 1931–1942.
Couvreur, T.L.P., Chatrou, L.W., Sosef, M.S.M. & Richardson,
J.E. (2008a) Molecular phylogenetics reveal multiple tertiary
vicariance origins of the African rain forest trees. BMC
Biology, 6, 54.
Couvreur, T.L.P., Richardson, J.E., Sosef, M.S.M., Erkens,
R.H.J. & Chatrou, L.W. (2008b) Evolution of syncarpy and
other morphological characters in African Annonaceae: a
posterior mapping approach. Molecular Phylogenetics and
Evolution, 47, 302–318.
Couvreur, T.L.P., van der Ham, R.W.J.M., Mbele, Y.M.,
Mbago, F.M. & Johnson, D.M. (2009) Molecular and
morphological characterization of a new monotypic genus
of Annonaceae, Mwasumbia, from Tanzania. Systematic
Botany, 34, 266–276.
Couvreur, T.L.P., Franzke, A., Al-Shehbaz, I.A., Bakker, F.T.,
Koch, M.A. & Mummenhoff, K. (2010) Molecular phylogenetics, temporal diversification, and principles of evolution
in the mustard family (Brassicaceae). Molecular Biology and
Evolution, 27, 55–71.
Crisp, M.D. & Cook, L.G. (2009) Explosive radiation or cryptic
mass extinction? Interpreting signatures in molecular phylogenies. Evolution, 63, 2257–2265.
Cusimano, N. & Renner, S.S. (2010) Slowdowns in diversification rates from real phylogenies may not be real. Systematic Biology, 59, 458–464.
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
Davis, C.C., Bell, C.D., Mathews, S. & Donoghue, M.J. (2002)
Laurasian migration explains Gondwanan disjunctions:
evidence from Malpighiaceae. Proceedings of the National
Academy of Sciences USA, 99, 6833–6837.
Davis, C.C., Webb, C.O., Wurdack, K.J., Jaramillo, C.A. &
Donoghue, M.J. (2005) Explosive radiation of Malpighiales
supports a mid-Cretaceous origin of modern tropical rain
forests. The American Naturalist, 165, E36–E65.
Dilcher, D.L. & Crane, P.R. (1984) Archaeanthus: an early
angiosperm from the Cenomanian of the Western interior of
North America. Annals of the Missouri Botanical Garden, 71,
351–383.
Doyle, J.A. & Endress, P.K. (2010) Integrating Early Cretaceous
fossils into the phylogeny of living angiosperms: Magnoliidae and eudicots. Journal of Systematics and Evolution, 48,
1–35.
Doyle, J.A. & Le Thomas, A. (1997) Phylogeny and geographic
history of Annonaceae. Géographie physique et Quaternaire,
51, 353–361.
Doyle, J.A., Sauquet, H., Scharaschkin, T. & Le Thomas, A.
(2004) Phylogeny, molecular and fossil dating, and biogeographic history of Annonaceae and Myristicaceae (Magnoliales). International Journal of Plant Sciences, 165, S55–S67.
Drummond, A.J. & Rambaut, A. (2007) BEAST: Bayesian
evolutionary analysis by sampling trees. BMC Evolutionary
Biology, 7, 214.
Drummond, A.J., Ho, S.Y.W., Phillips, M.J. & Rambaut, A.
(2006) Relaxed phylogenetics and dating with confidence.
PloS Biology, 4, 699–710.
Erkens, R.H.J. (2007) From morphological nightmare to
molecular conundrum. Phylogenetic, evolutionary and taxonomic studies on Guatteria (Annonaceae). PhD Thesis,
Utrecht University, Utrecht. Available at: http://igiturarchive.library.uu.nl/dissertations/2007-0227-200317/UU
index.html.
Erkens, R.H.J., Chatrou, L.W., Maas, J.W., van der Niet, T. &
Savolainen, V. (2007) A rapid diversification of rainforest
trees (Guatteria; Annonaceae) following dispersal from
Central into South America. Molecular Phylogenetics and
Evolution, 44, 399–411.
Erkens, R.H.J., Maas, J.W. & Couvreur, T.L.P. (2009) From
Africa via Europe to South America: migrational route of a
species-rich genus of Neotropical lowland rain forest trees
(Guatteria, Annonaceae). Journal of Biogeography, 36,
2338–2352.
Gentry, A. (1993) Diversity and floristic composition of lowland tropical forest in Africa and South America. Biological
relationships between Africa and South America (ed. by P.
Goldblatt), pp. 500–547. Yale University Press, New Haven.
Gradstein, F.M., Ogg, J.G. & Smith, A.G. (2004) A geologic time
scale 2004. Cambridge University Press, Cambridge.
Harvey, P.H., May, R.M. & Nee, S. (1994) Phylogenies without
fossils. Evolution, 48, 523–529.
Ho, S.Y.W. (2009) An examination of phylogenetic models of
substitution rate variation among lineages. Biology Letters, 5,
421–424.
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Keßler, P.J.A. (1993) Annonaceae. The families and genera of
vascular plants, Vol. 2. Magnoliid, hamamelid and caryophyllid families (ed. by K. Kubitzki, J.G. Rohwer and V.
Bittrich), pp. 93–129. Springer Verlag, Berlin.
Kress, J.W. & Specht, C.D. (2006) The evolutionary and biogeographic origin of the tropical monocot order Zingiberales. Aliso, 22, 619–630.
Lavin, M. & Luckow, M. (1993) Origins and relationships of
tropical North America in the context of the boreotropics
hypothesis. American Journal of Botany, 80, 1–14.
Linder, H.P. (2008) Plant species radiations: where, when,
why? Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 3097–3105.
Linder, H.P., Hardy, C.R. & Rutschmann, F. (2005) Taxon
sampling effects in molecular clock dating: an example from
the African Restionaceae. Molecular Phylogenetics and Evolution, 35, 569–582.
Magallón, S.A. & Castillo, A. (2009) Angiosperm diversification through time. American Journal of Botany, 96, 349–365.
Magallón, S.A. & Sanderson, M.J. (2001) Absolute diversification rates in angiosperm clades. Evolution, 55, 1762–1780.
Magallón, S.A. & Sanderson, M.J. (2005) Angiosperm divergence times: the effect of genes, codon positions, and time
constraints. Evolution, 59, 1653–1670.
McElwain, J.C. & Punyasena, S.W. (2007) Mass extinction
events and the plant fossil record. Trends in Ecology and
Evolution, 22, 548–557.
McKenna, D.D. & Farrell, B.D. (2006) Tropical forests are both
evolutionary cradles and museums of leaf beetle diversity.
Proceedings of the National Academy of Sciences USA, 103,
10947–10951.
McKenna, M.C.C. (1973) Sweepstakes, filters, corridors,
Noah’s Arks, and beached Viking funeral ships in palaeogeography. Implications of continental drift to the earth sciences (ed. by D.H. Tarling and S.K. Runcorn), pp. 291–304.
Academic Press, London.
Merckx, V., Chatrou, L.W., Lemaire, B., Sainge, M.N., Huysmans, S. & Smets, E.F. (2008) Diversification of mycoheterotrophic angiosperms: evidence from Burmanniaceae.
BMC Evolutionary Biology, 8, 178–194.
Mohr, B.A.R. & Bernardes-de-Oliveira, M.E.C. (2004)
Endressinia brasiliana, a magnolialean angiosperm from the
lower Cretaceous Crato Formation (Brazil). International
Journal of Plant Sciences, 165, 1121–1133.
Mols, J.B., Gravendeel, B., Chatrou, L.W., Pirie, M.D., Bygrave,
P.C., Chase, M.W. & Kessler, P.J.A. (2004) Identifying clades
in Asian Annonaceae: monophyletic genera in the polyphyletic Miliuseae. American Journal of Botany, 91, 590–600.
Mols, J.B., Kessler, P.J.A., Rogstad, S.H. & Saunders, R.M.K.
(2008) Reassignment of six Polyalthia species to the new
genus Maasia (Annonaceae): molecular and morphological
congruence. Systematic Botany, 33, 490–494.
Moore, B.R. & Donoghue, M.J. (2007) Correlates of diversification in the plant clade Dipsacales: geographic movement
and evolutionary innovations. The American Naturalist, 170,
S28–S55.
677
T. L. P. Couvreur et al.
Morley, R.J. (2000) Origin and evolution of tropical rain forests.
John Wiley & Sons, New York.
Morley, R.J. (2003) Interplate dispersal paths for megathermal
angiosperms. Perspectives in Plant Ecology, Evolution and
Systematics, 6, 5–20.
Morley, R.J. (2007) Cretaceous and Tertiary climate change and
the past distribution of megathermal rainforests. Tropical
rainforest responses to climatic changes (ed. by M.B. Bush and
J.R. Flenley), pp. 1–31. Praxis Publishing, Chichester.
Morley, R.J. & Dick, C.W. (2003) Missing fossils, molecular
clocks, and the origin of the Melastomataceae. American
Journal of Botany, 90, 1638–1644.
Muellner, A.N., Savolainen, V., Samuel, R. & Chase, M.W.
(2006) The mahogany family ‘‘out-of-Africa’’: divergence
time estimation, global biogeographic patterns inferred from
plastid rbcL DNA sequences, extant, and fossil distribution of
diversity. Molecular Phylogenetics and Evolution, 40, 236–250.
Nakkuntod, M., Su, Y.C.F., Seelanan, T. & Saunders, R.M.K.
(2009) Molecular phylogenetic and morphological evidence
for the congeneric status of Goniothalamus and Richella
(Annonaceae). Taxon, 58, 127–132.
Nee, S. (2006) Birth–death models in macroevolution. Annual
Review of Ecology, Evolution, and Systematics, 37, 1–17.
Nee, S., May, R.M. & Harvey, P.H. (1994) The reconstructed
evolutionary process. Philosophical Transactions of the Royal
Society B: Biological Sciences, 344, 305–311.
Nichols, D.J. & Johnson, K.R. (2008) Plants and the K–T
boundary. Cambridge University Press, Cambridge.
Nylander, J.A.A. (2004) MrModeltest v2. Program distributed
by the author, Evolutionary Biology Centre, Uppsala University, Uppsala.
Peng, Z.G., Ho, S.Y.W., Zhang, Y.G. & He, S.P. (2006) Uplift
of the Tibetan plateau: evidence from divergence times of
glyptosternoid catfishes. Molecular Phylogenetics and Evolution, 39, 568–572.
Pirie, M.D. & Doyle, J.A. (in press) The age of clades in Annonaceae: progress in fossil and molecular dating. The natural history of Annonaceae (ed. by L.W. Chatrou, R.H.J.
Erkens, R. M. K. Saunders and J.E. Richardson). A.R.G.
Gantner Verlag, Ruggell, Liechtenstein.
Pirie, M.D., Chatrou, L.W., Erkens, R.H.J., Maas, J.W., van der
Niet, T., Mols, J. & Richardson, J.E. (2005) Phylogeny
reconstruction and molecular dating in four Neotropical
genera of Annonaceae: the effect of taxon sampling in age
estimation. Plant species-level systematics: new perspectives on
pattern and process (ed. by F.T. Bakker, L.W. Chatrou, B.
Gravendeel and P.B. Pelser), pp. 149–174. Gantner Verlag,
Ruggell.
Pirie, M.D., Chatrou, L.W., Mols, J.B., Erkens, R.H.J. & Oosterhof, J. (2006) ‘Andean-centred’ genera in the shortbranch clade of Annonaceae: testing biogeographical
hypotheses using phylogeny reconstruction and molecular
dating. Journal of Biogeography, 33, 31–46.
Punyasena, S.W., Eshel, G. & McElwain, J.C. (2008) The
influence of climate on the spatial patterning of Neotropical
plant families. Journal of Biogeography, 35, 117–130.
678
Pybus, O.G. & Harvey, P.H. (2000) Testing macro-evolutionary
models using incomplete molecular phylogenies. Proceedings
of the Royal Society B: Biological Sciences, 267, 2267–2272.
Qiu, Y.L., Lee, J., Bernasconi-Quadroni, F., Soltis, D.E., Soltis,
P.S., Zanis, M., Zimmer, E.A., Chen, Z., Savolainen, V. &
Chase, M.W. (2000) Phylogeny of basal angiosperms: analyses of five genes from three genomes. International Journal
of Plant Sciences, 161, S3–S27.
de Queiroz, A. & Gatesy, J. (2007) The supermatrix approach
to systematics. Trends in Ecology and Evolution, 22, 34–41.
Rabosky, D.L. (2006) LASER: a maximum likelihood toolkit
for detecting temporal shits in diversification rates from
molecular phylogenies. Evolutionary Bioinformatics Online,
2, 257–260.
Rambaut, A. & Drummond, A.J. (2003) Tracer. Version 1.4.
Available at: http://evolve.zoo.ox.ac.uk.
Raven, H.P. & Axelrod, D.I. (1974) Angiosperm biodiversity
and past continental movements. Annals of the Missouri
Botanical Garden, 61, 539–673.
Ree, R.H. & Smith, S.A. (2008) Maximum likelihood inference
of geographic range evolution by dispersal, local extinction,
and cladogenesis. Systematic Biology, 57, 4–14.
Ree, R.H., Moore, B.R., Webb, C.O. & Donoghue, M.J. (2005) A
likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution, 59, 2299–2311.
Reid, M.E. & Chandler, M.E.J. (1933) The London Clay flora.
Trustees of the British Museum, London.
Renner, S.S., Clausing, G. & Meyer, K. (2001) Historical biogeography of Melastomataceae: the roles of Tertiary
migration and long-distance dispersal. American Journal of
Botany, 88, 1290–1300.
Richardson, J.E., Pennington, R.T., Pennington, T.D. & Hollingsworth, P.M. (2001) Rapid diversification of a speciesrich genus of Neotropical rain forest trees. Science, 293,
2242–2245.
Richardson, J.E., Chatrou, L.W., Mols, J.B., Erkens, R.H.J. &
Pirie, M.D. (2004) Historical biogeography of two cosmopolitan families of flowering plants: Annonaceae and
Rhamnaceae. Philosophical Transactions of the Royal Society
B: Biological Sciences, 359, 1495–1508.
Ricklefs, R.E. (2007) Estimating diversification rates from
phylogenetic information. Trends in Ecology and Evolution,
22, 601–610.
Ronquist, F. (1996) DIVA 1.1 user’s manual. Available at: http://
www.ebc.uu.se/systzoo/research/diva/manual/dmanual.html.
Rutschmann, F., Eriksson, T., Schönenberger, J. & Conti, E.
(2004) Did Crypteroniaceae really disperse out of India?
Molecular dating evidence from rbcL, ndhF, and rpl16 intron
sequences. International Journal of Plant Sciences, 165, S69–
S83.
Sanderson, M.J. (2002) Estimating absolute rates of molecular
evolution and divergence times: a penalized likelihood
approach. Molecular Biology and Evolution, 19, 101–109.
Sanderson, M.J. (2003) r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a
molecular clock. Bioinformatics, 19, 301–302.
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Evolutionary history of the Annonaceae
Sauquet, H., Doyle, J.A., Scharaschkin, T., Borsch, T., Hilu,
K.W., Chatrou, L.W. & Le Thomas, A. (2003) Phylogenetic
analysis of Magnoliales and Myristicaceae based on multiple
data sets: implications for character evolution. Botanical
Journal of the Linnean Society, 142, 125–186.
Sauquet, H., Weston, P.H., Anderson, C.L., Barker, N.P.,
Cantrill, D.J., Mast, A.R. & Savolainen, V. (2009) Contrasted
patterns of hyperdiversification in Mediterranean hotspots.
Proceedings of the National Academy of Sciences USA, 106,
221–225.
Scharaschkin, T. & Doyle, J.A. (2005) Phylogeny and historical
biogeography of Anaxagorea (Annonaceae) using morphology and non-coding chloroplast sequence data. Systematic Botany, 30, 712–735.
Scharaschkin, T. & Doyle, J.A. (2006) Character evolution in
Anaxagorea (Annonaceae). American Journal of Botany, 93,
36–54.
Smith, N., Mori, S.A., Henderson, A., Stevenson, D.W. &
Heald, S.V. (2004) Flowering plants of the Neotropics.
Princeton University Press, Princeton, NJ.
Stamatakis, A. (2006) RAxML-VI-HPC: maximum likelihoodbased phylogenetic analyses with thousands of taxa and
mixed models. Bioinformatics, 22, 2688–2690.
Stamatakis, A., Hoover, P. & Rougemont, J. (2008) A rapid
bootstrap algorithm for the RAxML web-servers. Systematic
Biology, 57, 758–771.
Stebbins, G.L. (1974) Flowering plants: evolution above the
species level. Harvard University Press, Cambridge, MA.
van Steenis, C.G.G.J. (1962) The land bridge theory in botany.
Blumea, 11, 235–542.
Su, Y.C.F. & Saunders, R.M.K. (2009) Evolutionary divergence times in the Annonaceae: evidence of a Late Miocene
origin of Pseuduvaria in Sundaland with subsequent
diversification in New Guinea. BMC Evolutionary Biology,
9, 153.
Suchard, M.A., Weiss, R.E. & Sinsheimer, J.S. (2001) Bayesian
selection of continuous-time Markov chain evolutionary
models. Molecular Biology and Evolution, 18, 1001–1013.
Takahashi, M., Friis, E.M., Uesugi, K., Suzuki, Y. & Crane, P.R.
(2008) Floral evidence of Annonaceae from the Late Cretaceous of Japan. International Journal of Plant Sciences, 169,
908–917.
Tchouto, M.G.P., de Boer, W.F., de Wilde, J.J.F.E. & van der
Maesen, L.J.G. (2006) Diversity patterns in the flora of the
Campo-Ma’an rain forest, Cameroon: do tree species tell it
all? Biodiversity and Conservation, 15, 1353–1374.
Thorne, R.F. (1972) Major disjunctions in the geographic
ranges of seed plants. The Quarterly Review of Biology, 47,
365–411.
Tiffney, B.H. & Manchester, S.R. (2001) The use of geological
and paleontological evidence in evaluating plant phylogeographic hypotheses in the Northern Hemisphere Tertiary.
International Journal of Plant Sciences, 162, S3–S17.
Traverse, A. (1988) Plant evolution dances to a different beat.
Plant and animal evolutionary mechanisms compared.
Historical Biology, 1, 277–301.
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd
Wang, H.C., Moore, M.J., Soltis, P.S., Bell, C.D., Brockington,
S.F., Alexandre, R., Davis, C.C., Latvis, M., Manchester, S.R.
& Soltis, D.E. (2009) Rosid radiation and the rapid rise of
angiosperm-dominated forests. Proceedings of the National
Academy of Sciences USA, 106, 3853–3858.
Whitmore, T.C. (1998) An introduction to tropical rain forests.
Clarendon Press, Oxford.
Wiens, J.J. (2003) Missing data, incomplete taxa, and phylogenetic accuracy. Systematic Biology, 52, 528–538.
Wiens, J.J. & Moen, D.S. (2008) Missing data and the accuracy
of Bayesian phylogenetics. Journal of Systematics and Evolution, 46, 307–314.
Wilf, P. & Johnson, K.R. (2004) Land plant extinction at the
end of the Cretaceous: a quantitative analysis of the North
Dakota megafloral record. Paleobiology, 30, 347–368.
Wilson, R., Heinrichs, J., Hentschel, J., Gradstein, S.R. &
Schneider, H. (2007) Steady diversification of derived liverworts under Tertiary climatic fluctuations. Biology Letters,
3, 566–569.
Wolfe, J.A. (1975) Some aspects of plant geography of the
Northern Hemisphere during Late Cretaceous and Tertiary.
Annals of the Missouri Botanical Garden, 62, 264–279.
Wolfe, J.A. & Upchurch, G.R., Jr (1986) Vegetation, climatic
and floral changes at the Cretaceous–Tertiary boundary.
Nature, 324, 148–152.
Yule, G.U. (1924) A mathematical theory of evolution based
on the conclusions of Dr. J.C. Willis. Philosophical Transactions of the Royal Society B: Biological Sciences, 213,
21–87.
Zachos, J., Pagani, M., Sloan, L., Thomas, E. & Billups, K.
(2001) Trends, rhythms, and aberrations in global climate
65 Ma to present. Science, 292, 686–693.
Zhou, L.L., Su, Y.C.F. & Saunders, R.M.K. (2009) Molecular
phylogenetic support for a broader delimitation of Uvaria
(Annonaceae), inclusive of Anomianthus, Cyathostemma,
Ellipeia, Ellipeiopsis and Rauwenhoffia. Systematics and Biodiversity, 7, 249–258.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article:
Appendix S1 GenBank numbers and area coding for all taxa
included in the Lagrange analysis.
Appendix S2 Number of missing characters per taxon, per
marker and in total.
Appendix S3 Diversification rates under no extinction
(b = 0) and high extinction (b = 0.95).
As a service to our authors and readers, this journal
provides supporting information supplied by the authors.
Such materials are peer-reviewed and may be reorganized
for online delivery, but are not copy-edited or typeset.
Technical support issues arising from supporting information (other than missing files) should be addressed to the
authors.
679
T. L. P. Couvreur et al.
BIOSKETCH
Thomas Couvreur completed his doctorate under the auspices of the National Herbarium of the Netherlands in 2008, working on
systematics and biogeography of African Annonaceae, and is currently a postdoctoral researcher at the New York Botanical Garden.
In December 2010 he will start as a researcher at the Institut de Recherche pour le Développement (IRD) in Montpellier, France. His
main research interests focus on evolution, biogeography and systematics of tropical plant families, mainly Annonaceae and palms.
The main research interest of the authors is centred on the systematics, evolution and biogeography of flowering plants, mainly of
tropical lineages.
Author contributions: T.L.P.C., R.H.J.E. and M.D.P. conceived the ideas; T.L.P.C., R.H.J.E., M.D.P., R.M.K.S., Y.C.F.S and L.W.C.
contributed data; T.L.P.C. collected and analysed the data; T.L.P.C., R.H.J.E., J.E.R. and M.D.P. led the writing; all authors made
significant comments on and improvements to the manuscript.
Editor: Pauline Ladiges
680
Journal of Biogeography 38, 664–680
ª 2010 Blackwell Publishing Ltd