Phylogeny, biogeography, and ecology of Ficus section Malvanthera (Moraceae).

Abstract

We conducted the first molecular phylogenetic study of Ficus section Malvanthera (Moraceae; subgenus Urostigma) based on 32 Malvanthera accessions and seven outgroups representing other sections of Ficus subgenus Urostigma. We used DNA sequences from the nuclear ribosomal internal and external transcribed spacers (ITS and ETS), and the glyceraldehyde-3-phosphate dehydrogenase (G3pdh) region. Phylogenetic analysis using maximum parsimony, maximum likelihood and Bayesian methods recovered a monophyletic section Malvanthera to the exclusion of the rubber fig, Ficus elastica. The results of the phylogenetic analyses do not conform to any previously proposed taxonomic subdivision of the section and characters used for previous classification are homoplasious. Geographic distribution, however, is highly conserved and Melanesian Malvanthera are monophyletic. A new subdivision of section Malvanthera reflecting phylogenetic relationships is presented. Section Malvanthera likely diversified during a period of isolation in Australia and subsequently colonized New Guinea. Two Australian series are consistent with a pattern of dispersal out of rainforest habitat into drier habitats accompanied by a reduction in plant height during the transition from hemi-epiphytic trees to lithophytic trees and shrubs. In contradiction with a previous study of Pleistodontes phylogeny suggesting multiple changes in pollination behaviour, reconstruction of changes in pollination behaviour on Malvanthera, suggests only one or a few gains of active pollination within the section.

Full text

Page 1

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Page 2

Author's personal copy
Phylogeny, biogeography, and ecology of Ficus section Malvanthera (Moraceae)
Nina Rønsteda,b,c,*, George D. Weiblenb, V. Savolainena,d, James M. Cookd,e
aJodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK
bBell Museum of Natural History, Department of Plant Biology, University of Minnesota, 1445 Gortner Avenue, Saint Paul, MN 55108, USA
cDepartment of Medicinal Chemistry, University of Copenhagen, Universitetsparken 2, Copenhagen 2100, Denmark
dDepartment of Biological Sciences, Imperial College London, Ascot, UK
eSchool of Biological Sciences, University of Reading, Reading RG6 6AS, UK
a r t i c l ei n f o
Article history:
Received 28 June 2007
Revised 2 April 2008
Accepted 5 April 2008
Available online 14 April 2008
Keywords:
Moraceae
Ficus
Malvanthera
Phylogeny
ITS
ETS
G3pdh
a b s t r a c t
We conducted the first molecular phylogenetic study of Ficus section Malvanthera (Moraceae; subgenus
Urostigma) based on 32 Malvanthera accessions and seven outgroups representing other sections of Ficus
subgenus Urostigma. We used DNA sequences from the nuclear ribosomal internal and external tran-
scribed spacers (ITS and ETS), and the glyceraldehyde-3-phosphate dehydrogenase (G3pdh) region. Phy-
logenetic analysis using maximum parsimony, maximum likelihood and Bayesian methods recovered a
monophyletic section Malvanthera to the exclusion of the rubber fig, Ficus elastica. The results of the phy-
logenetic analyses do not conform to any previously proposed taxonomic subdivision of the section and
characters used for previous classification are homoplasious. Geographic distribution, however, is highly
conserved and Melanesian Malvanthera are monophyletic. A new subdivision of section Malvanthera
reflecting phylogenetic relationships is presented. Section Malvanthera likely diversified during a period
of isolation in Australia and subsequently colonized New Guinea. Two Australian series are consistent
with a pattern of dispersal out of rainforest habitat into drier habitats accompanied by a reduction in
plant height during the transition from hemi-epiphytic trees to lithophytic trees and shrubs. In contra-
diction with a previous study of Pleistodontes phylogeny suggesting multiple changes in pollination
behaviour, reconstruction of changes in pollination behaviour on Malvanthera, suggests only one or a
few gains of active pollination within the section.
? 2008 Elsevier Inc. All rights reserved.
1. Introduction
Figs (Ficus, Moraceae) constitute one of the largest genera of
angiosperms, with almost 750 species of terrestrial trees, shrubs,
hemi-epiphytes, climbers and creepers occurring in the tropics
and subtropics worldwide (Berg and Corner, 2005). All species of
figs share the distinctive fig inflorescence (syconium), which is
the site of an obligate mutualism with pollinating fig wasps of
the family Agaonidae (Cook and Rasplus, 2003). Figs are important
genetic resources with high economic and nutritional value. They
also play an important role in generating biodiversity in the rain-
forest ecosystem by setting fruits throughout the year and provid-
ing an important source of food for most fruit-eating vertebrates in
the tropics (Harrison, 2005). Most insects pollinate passively, but
fig wasps are one of a few cases, where active pollination behav-
iour has evolved (Cook and Rasplus, 2003). While most genera of
pollinating wasps are active pollinators, five out of 20 genera con-
tain both passive and active pollinators (Kjellberg et al., 2001).
Recent classification divided the genus into six subgenera based
primarily on morphology (Berg, 2003). The monoecious subgenus
Urostigma, to which section Malvanthera belongs, is the largest
with about 280 species worldwide, most of them displaying the
characteristic hemi-epiphytic habit (banyans and stranglers). Ficus
section Malvanthera Corner (subg. Urostigma) includes 23 species
of hemi-epiphytes and lithophytes producing aerial, adventitious,
or creeping root systems. The section has its primary centre of
diversity in Australia and a second centre in New Guinea and the
Bismarck Archipelago. A few species extend eastwards into Ocea-
nia (e.g. Lord Howe Island, New Caledonia, the Solomon Islands
and Vanuatu). Section Malvanthera includes species with two very
distinct ecologies: hemi-epiphytic stranglers and free-standing
trees in the rainforests of Eastern Australia and New Guinea, and
lithophytic shrubs and trees occurring in more arid parts of Austra-
lia (Dixon, 2003).
The section was established by Corner (1959) who recognized
19 species characterized by features including: a slit-shaped or tri-
radiate ostiole with all bracts descending; syconia with two or
three basal bracts; reniform unilocular anthers with crescentic or
transverse dehiscence; ovaries attached at the base to the recepta-
cle or imbedded in the receptacle; a red spot at the base or the apex
1055-7903/$ - see front matter ? 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2008.04.005
* Corresponding author. Address: Department of Medicinal Chemistry, University
of Copenhagen, Universitetsparken 2, Copenhagen 2100, Denmark.
E-mail address: nr@farma.ku.dk (N. Rønsted).
Molecular Phylogenetics and Evolution 48 (2008) 12–22
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev

Page 3

Author's personal copy
of the ovary; and a stigma that is either simple or bifid. Corner
(1959) and again Berg and Corner (2005) noted the similarity be-
tween subg. Urostigma sections Malvanthera and Galoglychia in
the inflexed apical bracts of the ostiole and internal bracts. Berg
and Corner (2005) also noted that similarities in venation suggest
a relationship between Malvanthera and the F. benjamina group of
section Urostigma subsection Conosycea. Molecular phylogenetic
studies have shown sections Conosycea and Malvanthera to be sis-
ter taxa (Rønsted et al., 2005).
Corner (1965) subdivided section Malvanthera into two series.
Malvanthereae is distinguished by reniform anthers dehiscing lon-
gitudinally and crescentically. The series includes 18 species in
four loosely defined subseries, Eubracteatae, Hesperidiiformes,
Malvanthereae, and Platypodeae based on characters of the basal
bracts and whether the ovaries are partly embedded in the
receptacle or not. The monotypic series Cyclanthereae includes F.
sterrocarpa from New Guinea, which has depressed-subglobose
anthers (Table 1).
Chew (1989) followed Corner (1965) in treating Australian Fi-
cus, whereas Australian Malvanthera were subsequently revised
by Dixon (2001a–d, 2003), who proposed a new subdivison of
the section into ser. Malvanthereae including species with imbri-
cate basal bracts, and ser. Hesperidiiformes with valvate basal
bracts (Table 1).
In treating Ficus of the Malesian region, Berg and Corner (2005)
united the monotypic section Stilpnophyllum (F. elastica, the rubber
fig) and section Malvanthera as subsections of an expanded section
Stilpnophyllum based on similarities between F. elastica and the
Malvanthera species in the venation of the lamina, the length of
the stipules and cucullate (hood-shaped) caducous basal bracts.
However, F. elastica is distinct from the Malvanthera species in
the shape of the ostiole and in having anthers with separate theca
and connate stipules. The Indo-Chinese origin of F. elastica, which is
widely cultivated, also has little geographical affinity to the Austra-
lian and Melanesian Malvanthera. A global molecular phylogenetic
study (Rønsted et al., 2005) including F. elastica, 12 species of
Conosycea figs and 11 species of Malvanthera figs strongly sug-
gested that F. elastica is a member of section Conosycea with a
derivative morphology.
Within subsection Malvanthera, Berg and Corner (2005) recog-
nised 18 species in three informal groups, based on characters of
the basal bracts and ostiole shape. Berg and Corner (2005) united
most of Corner ´s species from New Guinea (F. augusta, F. hesperidi-
iformis, F. heteromeka, F. mafuluensis, F. sterrocarpa and F. xylosycia)
Table 1
Historical classification of Ficus section Malvanthera
Rønsted et al. (present study, Fig. 1)Corner (1965)Dixon (2003)Berg (2005)h
Subsect. Malvantherae
F. macrophylla
F. pleurocarpa
Subser. Malvanthereae
Subser. Hesperidiiformes
Ser. Malvanthereae
Ser. Hesperidiiformes
Subsect. Malvanthera C
Subsect. Malvanthera C
Subsect. Hesperidiiformes
Ser. Hesperidiiformes
F. hesperidiiformisa
F. sterrocarpa
Subser. Hesperidiiformes
Ser. Cyclanthereae
Ser. Hesperidiiformes
Ser. Hesperidiiformis
Subsect. Malvanthera A
F. hesperidiiformesa
Ser. Glandiferaeb
F. baolac
F. glandifera
F. rhizophoriphylla
F. obliquac
Subser. Malvanthereae
Subser. Malvanthereae
F. obliquac
Ser. Hesperidiiformes
Ser. Malvanthereae
Subsect. Malvanthera B
Subsect. Malvanthera B
Subsect. Malvanthera B
Ser. Xylosyciae
F. augusta
F. heteromeka
F. mafuluensis
F. xylosycia
Subser. Hesperidiiformes
Subser. Hesperidiiformes
Subser. Hesperidiiformes
Subser. Hesperidiiformes
Ser. Hesperidiiformes
Ser. Hesperidiiformes
Ser. Hesperidiiformes
Ser. Hesperidiiformes
F. hesperidiiformesa
F. hesperidiiformesa
F. hesperidiiformesa
F. hesperidiiformisa
Subsect. Platypodeae
Ser. Eubracteatae
F. triradiata
Subser. Eubracteatae
Ser. Hesperidiiformes
Subsect. Malvanthera A
Ser. Obliquae
F. cerasicarpad
F. lilliputianae
F. obliqua
F. platypodad
F. subpuberula
F. leuchotrichad
F. brachypodae
Subser. Platypodeae
Subser. Platypodeaed
Subser. Platypodeae
Ser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Subsect. Malvanthera C
Subsect. Malvanthera C
Subsect. Malvanthera C
Subsect. Malvanthera A
Subsect. Malvanthera C
Ser. Crassipeae
F. crassipes
F. destruens
Subser. Hesperidiiformes
Subser. Platypodeae
Ser. Hesperidiiformes
Ser. Malvanthereae
Subsect. Malvanthera A
Subsect. Malvanthera B
Ser. Rubiginosae
F. atrichaf
F. brachypodaf
F. rubiginosag
F. watkinsiana
F. platypodaf
F. platypodaf
Subser. Platypodeae
Subser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Ser. Malvanthereae
Subsect. Malvanthera C
Subsect. Malvanthera C
Subsect. Malvanthera C
Subsect. Malvanthera C
aF. hesperidiiformis includes F. augusta, F. heteromeka, F. mafuluensis, F. sterrocarpa, and F. xylosycia in Bergs (2005) classification.
bSeries Glandiferae possibly includes F. baola and F. rhizophoriphylla based on morphological affinities such as a slit-shaped ostiole (Berg, 2002).
cF. baola C.C. Berg was raised from F. obliqua (Berg 2002).
dF. cerasicarpa has affinities to F. platypoda (Dixon 2001a). F. cerasicarpa and F. platypoda were included in F. leuchotricha in Corner’s classification.
eF. lilliputiana is a new species with affinities to F. brachypoda (Dixon 2001b).
fF. atricha and F. brachypoda were included in F. platypoda in Corner’s classification.
gF. baileyana was included in F. rubiginosa by Dixon (2001d).
hF. elastica constituted a monotypic section Stilpnophyllum in Corner’s classification.
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22
13

Page 4

Author's personal copy
under a broadly circumscribed F. hesperidiiformis evidently based
on the apparent presence of intermediate forms. At the same time,
Berg accepted the narrow species concept of Dixon (2003) for the
Australian taxa.
The purpose of our study is to (1) infer phylogenetic relation-
ships in section Malvanthera using three molecular markers: the
internal and external transcribed spacers (ITS and ETS) of nuclear
ribosomal DNA, and the glyceraldehyde-3-phosphate dehydroge-
nase (G3pdh) gene, (2) reconstruct the evolution of morphological
characters used to subdivide the section in previous treatments, (3)
trace the biogeographic history of the section, (4) identify evolu-
tionary changes in habitat use, and (5) retrace the evolution of ac-
tive pollination.
2. Materials and methods
2.1. Taxon sampling
Taxon sampling was aimed at testing hypotheses of phyloge-
netic relationship among recognized species. Phylogeographic
analyses of particular complexes are needed to clarify species lim-
its in Malvanthera (Haine et al. 2006), as are population genetic
analyses to identify the role that hybridization may have played,
if any, in the diversification of Malvanthera (Machado et al.
2005). We inferred phylogenetic relationships in section Malvan-
thera using three molecular markers: the internal and external
transcribed spacers (ITS and ETS) of nuclear ribosomal DNA, and
the glyceraldehyde-3-phosphate dehydrogenase (G3pdh) gene.
We did not include chloroplast DNA markers in the study based
on the absence of phylogenetically informative variation in four re-
gions (rps16 intron, trnL intron, trnL-F spacer, and psB–psbF spacer)
within Ficus according to Rønsted et al. (2007).
Contrary to Berg and Corner (2005), we follow the circumscrip-
tion of Malvanthera sensu Corner (1965) treating F. elastica as an
outgroup according to its historical taxonomic and molecular phy-
logenetic position (Rønsted et al., 2005).
Total genomic DNA was extracted from 29 Ficus specimens. In
addition 19 ITS sequences, 16 ETS sequences, and 11 G3pdh se-
quences were retrieved from GenBank/EBI (following papers by
Weiblen, 2000; Jousselin et al., 2003; Rønsted et al., 2005; Rønsted
et al., 2008), resulting in a total of 34 Malvanthera and seven out-
groups representing the other sections of subgenus Urostigma.
Specimen determinations, vouchers, localities and GenBank/EBI
Accession Nos. (EF545651–EF545669, ETS: EF538767–EF538785,
and G3pdh: EF538786–EF538801) are listed in Table 2. Species con-
cepts are primarily according to Dixon (2001a–d, 2003), for the
Australian taxa and Corner (1959) for the remainder of the section.
We included multiple individuals of some species to evaluate spe-
cies limits and the extent of infraspecific variation. We included
samples representing 19 of the 23 species of Ficus section Malvan-
thera. Material of F. cerasicarpa D.J. Dixon from Northern Australia,
F. rhizophoriphylla King and F. mafuluensis Summerh. from New
Guinea, and F. baola C.C. Berg from the Solomon Islands, was not
available for this study.
2.2. DNA extraction, amplification and sequencing
DNA extractions were performed in two ways. Some specimens
were extracted using the Qiagen DNeasy plant extraction kit
(Valencia, CA, USA) from 15–20 mg of dried leaf-fragments or her-
barium material. Other specimens were extracted using 0.2–0.3 g
silica dried leaves and a modified version of the 2 ? CTAB method
of Doyle and Doyle (1987). Before precipitation, an aliquot was
purified using the Qiagen PCR purification kit (Qiagen, Inc., Santa
Clarita, CA, USA) following the manufacturer’s protocols. The
remainder of the DNA was purified using a caesium chloride/ethi-
dium bromide gradient (1.55 g ml?1density) followed by a dialysis
and was deposited in the DNA Bank at the Royal Botanic Gardens,
Kew (www.rbgkew.org.uk).
The internal and external transcribed spacers, ITS and ETS
(Baldwin et al., 1995; Baldwin and Markos, 1998), were amplified
using primers 17SE and 26SE (Sun et al., 1994) or ITS4 and ITS5
(White et al., 1990) and Hel1 and 18S ETS (Baldwin and Markos,
1998), respectively. The glyceraldehyde-3-phosphate dehydroge-
nase (G3pdh) region (Strand et al., 1997) was amplified using prim-
ers 7F and 9R (Strand et al., 1997).
All three regions were amplified and sequenced following pro-
tocols by Rønsted et al. (2008). Both strands were sequenced for
each region for the majority of taxa. For some samples, internal
primers (GCT ACG TTC TTC ATC GAT GC) and (GCA TCG ATG AAG
AAC GTA GC) modified from ITS2 and ITS3, respectively (White
et al., 1990) were used for sequencing of the ITS region, and inter-
nal primers 286F (TGT ATT CTG GTT GGG TTT C, Rønsted et al.,
2008) and 437R (TTC TGA AGC CTG ACA GTG AGG, Rønsted et al.,
2008) for sequencing of the G3pdh region in addition to the primers
used for amplification. Internal primers were used to ensure over-
lap between sequenced fragments, where chromatograms could
not be unambiguously interpreted throughout the whole region.
2.3. Phylogeny reconstruction
Sequences were edited and assembled using Sequencer 4.1.2TM
software (Gene Codes Corp., Ann Arbor, MI, USA), and all sequences
were aligned by eye in PAUP v. 4.0b10 for Macintosh (Swofford,
2002). Phylogenetic analyses were conducted using PAUP v.
4.0b10 (Swofford, 2002). All changes were assessed as unordered
and equally weighed (Fitch parsimony; Fitch, 1971).
We analysed the three gene regions separately to identify phy-
logenetic conflicts among the regions prior to performing a com-
bined analysis. For the separate analyses, we show only the
bootstrap tree (Fig. 1A–C) to establish that there were no cases of
strong conflict, namely clades that are highly supported by boot-
strap analysis in a single-loci analysis (e.g. G3pdh) that are incon-
gruent with highly supported clades in another single-locus
analysis (e.g. ITS). All data were therefore combined into one anal-
ysis, and we consider the combined analysis to provide the best
estimate of phylogeny (Fig. 2).
We attempted to avoid potential problems with non-overlap-
ping datasets in the combined analysis by including only samples
for which at least ITS was sequenced. Taxon sampling was maxi-
mised in two cases by combining different accessions of the same
species. We treated as a single taxon two accessions of F. xylosycia,
which were both from the same site in the Madang Province of Pa-
pua New Guinea, and likewise, two accessions of F. brachypoda,
from the same region of the Australian Northern Territory. In the
case of F. subpuberula, however, only an ETS sequence was
available.
Parsimony analysis of the combined matrix was not exhaustive
due to large numbers of equally parsimonious trees at each step in
the search. Most parsimonious trees (MP) were obtained using: (i)
1000 replicates of random taxon addition sequence and TBR
branch swapping with only 25 trees held at each step to save time
by avoiding swapping on suboptimal islands; (ii) the trees collec-
tively found in these 1000 replicates were then used as starting
trees for a second search using TBR branch swapping until up to
a maximum of 15,000 trees were found. Relative levels of homo-
plasy in all the datasets were assessed using the consistency index
(CI) and the retention index (RI). In all cases RI was at least 0.78,
which suggests that multiple islands of equally parsimonious trees
are unlikely (Maddison, 1991).
14
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22

Page 5

Author's personal copy
The combined dataset was also analysed under maximum like-
lihood (ML). We first selected a best fitting model of molecular
evolution using Modeltest (Posada and Crandall, 1998). Model
parameters thus obtained were then used in 25 heuristic search
replicates under the ML criterion.
Clade support was assessed using non-parametric bootstrap
re-sampling and Bayesian analysis of the combined dataset.
Bootstrap analyses (Felsenstein, 1985) of the three individual
datasets and the combined dataset were carried out using
1000 replicates each consisting of 100 simple addition sequence
replicates, with TBR swapping, and a limit of 10 trees retained
for each replicate. We defined scores between 50 and 74 boot-
strap percentages (BS) as weak support, scores between 75%
and 89% BS as moderate support, and scores of greater than
90% BS as strong support. We consider percentages <50 % to
be unsupported because such clades are not present in the
majority of trees for the re-sampled datasets.
Bayesian analysis of the combined dataset was performed with
MrBayes 3.1.2 (Ronquist and Hulsenbeck, 2003). We used an
HKY85 model of evolution (lset NST = 2 RATES = equal). The analy-
sis was performed with 2,000,000 generations on four Monte Carlo
Markov chains with equal rates and trees sampled every 100 gen-
erations (mcmc NGEN = 2,000,000, PRINTFREQ = 1000, SAMPLEF-
REQ = 100, NCHAINS = 4). The average standard deviation of the
split frequencies was 0.0022 after 1,685,000 and also after
2,000,000 generations indicating that there was no great advantage
to be gained by sampling additional generations. The first 10,000
trees of low posterior probability were deleted and all remaining
trees were imported into PAUP. A majority rule consensus tree
was produced showing the posterior probabilities (pp) of all ob-
served bi-partitions. We consider posterior probabilities of at least
0.95 as strong clade support.
To investigate if the Australian lithophytes (F. atricha, F. brachy-
poda, F. lilliputiana, F. platypoda, and F. subpuberula) could have orig-
inated only once, they were constrained as monophyletic and all
other ingroup and outgroup branches were collapsed. This tree
was then used as a topological constraint in 1000 replicates of MP
analysis in PAUP. We then used the Templeton test (Templeton,
Table 2
Voucher information and Genbank accession numbers of Ficus material included in this study
TaxaVoucherOrigin ITSETS G3PDH
F. atricha D. J. Dixon
F. augusta Corner
F. brachypoda (Miq) Miq.
F. brachypoda
Rønsted 142 (K)
GW597 (MIN)
Collector: John Zammit
Dixon/Kerrigan 1101
(DNA)
Cook 987
GW943 (MIN)
Dixon, PHD 312
Liv. Col. (BG) 2003, Berg 93.220
Madang, PNG 1995
Kulgera, NT, Australia, 1997
NT, Australia, 2004
Rønsted 2005
EF545651
EF545652
—
Rønsted 2005
EF538767
EF538768
—
EF538786
EF538787
—
EF538788
F. crassipes Bailey
F. destruens C.T. White
F. destruens
F. glandifera Summerh.
F. hesperidiiformis King
F. hesperidiiformis
F. hesperidiiformis
F. heteromeka Corner
F. lilliputiana D. J. Dixon
F. macrophylla Desf. ex. Pers. subsp. macrophylla
F. macrophylla
Boonjee, QLD, Australia, 1998
Cow Bay, QLD 1997
Mount Halifax, Bluewater, QLD
Vanuatu
Madang, PNG 1998
Madang, PNG 1997
Madang, PNG 2006
Chimbu, PNG 2004
Kununurra, WA
Rønsted 2005
Weiblen 2000
EF545653
Rønsted 2005
EF545654
Weiblen 2000
EF545655
EF545656
EF545657
Jousselin 2003
EF545658
Rønsted 2005
EF538769
—
Rønsted 2005
EF538770
EF538771
—
EF538772
EF538773
Jousselin 2003
—
EF538789
EF538790
—
Rønsted 2008
—
Rønsted 2008
—
EF538791
—
—
—
B204 (MIN)
GW825 (MIN)
GW2736 (MIN)
GW2125 (MIN)
Dixon, PHD 424 (DNA)
GW679 (MIN) Auckland Museum, cult. NZ
1995
Lord Howe Island, 2003
F. macrophylla subsp. columnaris (C.Moore)
P.S.Green
F. obliqua G. Forst.
F. obliqua
F. platypoda sensu Dixon
F. pleurocarpa F. Muell.
F. pleurocarpa
F. rubiginosa Desf. ex Vent.
F. rubiginosa
F. rubiginosa
F. sterrocarpa Diels
F. sterrocarpa
F. cf. sterrocarpa
F. cf. sterrocarpa
F. subpuberula Corner
Savolainen/Baker (K) Rønsted 2005Rønsted 2005EF538792
Cook 2003–6
GW1220 (MIN)
Jacobs/Wilson 5665 (K)
Port Douglas, QLD, Australia
Liv. col. (NOU) 2001
WT, Australia 1988
EF545659
EF545660
Rønsted 2005
Jousselin 2003
EF545661
EF545662
EF545663
EF545664
EF545665
EF545666
EF545667
EF545668
—
EF538774
EF538775
Rønsted 2005
Jousselin 2003
EF538776
EF538777
EF538778
—
EF538779
EF538780
EF538781
EF538782
EF538783
EF538793
—
EF538794
—
EF538795
—
Rønsted 2008
—
EF538796
EF538797
—
EF538798
—
Cook 9812/CLV441
Cook FRYN3
Rønsted 89 (C)
MW Chase 19870 (K)
GW734 (MIN)
T9414 (MIN)
GW1126 (MIN)
GW1881 (MIN)
Dixon/Kerrigan 1102
(DNA)
Hind/Herscovitch 6409 (K)
Rønsted 83 (C)
Cook (NR256)
Tully, QLD, Australia 1998
Yungaburra, QLD 2005
Liv. col. (C) 2002, E1859-0014
Liv. col. (K) 2003, K1961-35201
Eastern Highlands, PNG 1996
Eastern Highlands, PNG
East Sepik, PNG 2000
East Sepik, PNG 2003
NT, Australia 2004
F. triradiata var. sessilicarpa Corner
F. watkinsiana F.M.Bailey
F. watkinsiana
F. xylosycia Diels
F. xylosycia
QLD, Australia, 1991Rønsted 2005
Rønsted 2005
EF545669
Weiblen 2000
—
Rønsted 2005
Rønsted 2005
EF538784
—
EF538785
EF538799
Rønsted 2008
EF538800
—
EF538801
Atherton, QLD, Australia 2005
Madang, PNG
Madang, PNG 1999 B207 (MIN)
Outgroup
F. americana Aubl.
F. benjamina L.
F. binnendykii Miq.
F. elastica Roxb. ex.
F. ingens (Miq.) Miq.
F. lutea Vahl
F. sundaica Blume
Rønsted 2005
Jousselin 2003
Jousselin 2003
Jousselin 2003
Rønsted 2005
Jousselin 2003
Rønsted 2005
Rønsted 2005
Jousselin 2003
Jousselin 2003
Jousselin 2003
Rønsted 2005
Jousselin 2003
Rønsted 2005
Rønsted 2008
Rønsted 2008
Rønsted 2008
Rønsted 2008
Rønsted 2008
Rønsted 2008
Rønsted 2008
Notes: Abbreviations: Cult, cultivated; Liv. col, Living collections; NT, Northern Territory; NZ, New Zealand; PNG, Papua New Guinea; QLD, Queensland; WT, Western
Australia. Herbarium codes: (BG), University of Bergen, Norway; (C), University of Copenhagen, Denmark; (DNA), Department of Natural Resources, Environment, and the
Arts, Northern Territory, Australia; (K), Royal Botanic Gardens, Kew, UK; (MIN), University of Minnesota, Saint Paul, USA; (NOU), Institut de la Recherche pour le Devel-
oppement, Nouméa, New Caledonia. References: Jousselin 2003, Jousselin et al. (2003); Rønsted 2005, Rønsted et al. (2005); Rønsted 2008, Rønsted et al. (2008).
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22
15

Page 6

Author's personal copy
1983) to compare the length of the constrained and the uncon-
strained trees.
2.4. Morphological characters
Five morphological characters used for delimiting subsections,
series and informal groups of section Malvanthera in previous clas-
sifications were scored for all species (Fig. 3). Characters were ta-
ken from descriptions in the literature (Corner, 1959; Chew,
1989; Dixon, 2001a–d, 2003; Berg and Corner, 2005) and supple-
mented by personal observations of herbarium specimens. The
characters are described in the legend of Fig. 3. Although seed ova-
ries embedded in the receptacle or attached at the base to the
receptacle has been used to distinguish subseries in the past (Cor-
ner, 1965), we found this character difficult to score and we did not
include it Fig. 3.
2.5. Mapping habitat use
Habitat use was scored for all species (Fig. 3). Character states
were taken from descriptions in the literature (Corner, 1959; Chew,
1989; Dixon, 2001a–d, 2003; Berg and Corner, 2005) and supple-
mented by personal observations of herbarium collections and
during fieldwork in Australia and Papua New Guinea. The charac-
ters are described in the legend of Fig. 3.
2.6. Mapping changes in pollination behaviour
Previous studies have shown a strong association between
active pollination, low anther/ovule ratios (floral sex ratio,
FRS), and the presence of coxal combs (a line of setae or dense,
parallel and long hairs on the fore coxa) on the pollinator,
whereas the presence of pollen pockets provide a good, but
imperfect index of active pollination (Cook et al. 2004; Jouss-
elin et al. 2003; Kjellberg et al. 2001). Floral sex ratios, the
presence or absence of coxal combs, pollen pockets, and
pollinationbehaviourwere taken
supplemented by personal observations of F. glandifera, F. het-
eromeka, F. sterrocarpa and F. cf. sterrocarpa as well as the poll-
inators of F. glandifera, F. heteromeka, and F. cf. sterrocarpa (see
Fig. 4).
Following Cook et al. (2004), we estimated floral sex ratios (FSR)
as the proportion of flowers that are male (i.e. anthers/anthers + o-
vules), by counting and sexing all flowers in one syconium for each
of F. glandifera, F. heteromeka, F. sterrocarpa and F. cf. sterrocarpa. All
syconia were collected from trees growing within their natural
ranges.
We mapped pollination behaviour onto the ML tree presented
in Fig. 2, collapsing all duplicate samples of the same species,
and analysed the patterns of changes using parsimony techniques
in MacClade version 4.03 (Maddison and Maddison, 2001). We
fromtheliterature and
100
62
52
69
52
52
63
86
97
56
66
59
93
99
100
99
78
64
83
64
90
89
99
52
73
86
98
56
97
61
53
91
71
52
63
66
63
64
88
60
86
87
73
51
76
ABC
Fig. 1. Bootstrap consensus trees from analysis of (A) G3pdh nrDNA sequences, (B) ITS nrDNA sequences, and (C) ETS nrDNA sequences. Ficus americana, F. benjamina, F.
binnendykii, F. elastica, F. ingens, F. lutea, and F. sundaica were designated as an outgroup.
16
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22

Page 7

Author's personal copy
treated
character and weighted gains and losses of active behaviour
behaviour asan unordered, binary(active/passive)equally in seeking the most parsimonious reconstruction of
changes.
F. atricha
F. brachypoda
F. rubiginosa (Chase19870)
F. rubiginosa (FRYN3)
F. watkinsiana (Cook)
F. watkinsiana (NR83)
F. rubiginosa (NR89)
F. crassipes
F. destruens (GW943)
F. destruens (PHD312)
F. lilliputiana
F. subpuberula
F. platypoda
F. obliqua (Cook2003)
F. obliqua (GW1220)
F. triradiata
F. augusta
F. heteromeka
F. xylosycia
F. glandifera
F. hesperidiiformis (B204)
F. hesperidiiformis (GW825)
F. hesperidiiformis (GW2736)
F. cf.sterrocarpa (GW1881)
F. cf sterrocarpa (GW1126)
F. sterrocarpa (GW734)
F. sterrocarpa (T9414)
F. macrophylla ssp. columnaris (GW679)
F. macrophylla ssp. columnaris (LH)
F. macrophylla ssp. macrophylla
F. pleurocarpa (J)
F. pleurocarpa (Cook9812)
F. americana
F. lutea
F. ingens
F. benjamina
F. binnendykii
F. sundaica
F. elastica
5 changes
Fig. 2. One of 14 maximum likelihood trees with ML branch lengths, and with bootstrap percentages and posterior probabilities below and above branches, respectively.
Arrowheads indicate branches that collapse in the strict consensus of 14 ML trees. Our preferred classification of Malvanthera suggested in the present study is indicated on
the right.
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22
17

Page 8

Author's personal copy
3. Results
3.1. Analyses of the combined dataset
Since there were no strongly supported conflicting clades with-
in the ingroup, we considered it appropriate to directly combine all
data. The aligned combined matrix of all three regions contained
32 ingroup taxa and 7 outgroups, and a total of 2024 base pairs
of which 186 (9%) were parsimony informative. MP analysis with
a limit of 25 trees per replicate produced 4286 trees of 557 steps
with CI = 0.75 and RI = 0.78. These trees were used as starting trees
in a second search with no limit on the number of trees per repli-
cate, but with a maximum of 15,000 trees. This search produced
the maximum number of trees, which were then swapped to
completion.
The strict consensus tree of the combined MP analysis is not
fully resolved and a considerable number of clades do not ob-
tain more than 50% BS. The ingroup is monophyletic (80% BS).
Within the ingroup, we find a soft polytomy with four clades.
One clade includes F. pleurocarpa and F. macrophylla samples
Fig. 3. Maximum likelihood (ML) tree from Fig. 2 showing the distribution of macro-morphological characters used in previous classifications and the distribution of
biogeography and habitat use. White, black, and grey boxes show character states 0, 1, or polymorphic, respectively. 1. Prominence of basal or primary lateral veins on leaves:
basal veins distinct (1) or not (0). 2. Persistence of basal bracts: Basal bracts persistent (1) or cauducous (0). 3. Shape of basal bracts: Bracts imbricate (1) or valvate (0). 4.
Ostiole shape: Triradiate (1) or slit-shaped (bilabiate; 0). 5. Shape of stigma: Stigma bifid (1) or simple (0). LH—Lord Howe Island, NG—New Guinea, NSW—New South Wales,
NT—Northern Territory, PAC—Pacific islands, QLD—Queensland, SA—South Australia, SOL—Solomon Islands, WA—Western Australia.
18
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22

Page 9

Author's personal copy
(80%), another clade includes F. hesperidiiformis and F. sterrocar-
pa (100% BS), a third clade includes F. augusta, F. glandifera, F.
heteromeka and F. xylosycia (55%), and a last clade includes
the remainder of the Australian taxa (84%). The MP consensus
tree is not shown, but MP bootstrap support is indicated on
the ML tree in Fig. 2.
Fig. 4. Parsimony reconstruction of changes (*) between passive (white) and active (black) pollination behaviour showing the distribution of flower sex ratios (FSR) of
Malvanthera species and correlated morphology of their pollinators. White and black boxes show character states 0 and 1, respectively. 1. FSRs from Cook et al. (2004), except
Ficus glandifera, F. heteromeka, F. sterrocarpa and F. cf. sterrocarpa, which were counted for the present study. 2. Coxal combs absent (0) or present (1). 3. Pollen pockets absent
(0), present, but small (grey boxes), or present and large (1). 4. Pollination mode passive (0) or active (1). 5. Pollinating Pleistodontes species. Wasp data from Cook et al. (2004),
Lopez-Vaamonde et al. (2002), and Wiebes (1994), except for Pleistodontes blandus, P. ex. F. heteromeka and P. ex. F. cf. sterrocarpa, which were observed for the present study.
Some Malvanthera species have more than one pollinator, but no differences in pollination behaviour between pollinators of the same Malvanthera species have been
described.
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22
19

Page 10

Author's personal copy
For the ML analysis, Modeltest suggested TrN + I + G as the best
fitting model using both hierarchical likelihood ratio tests (hLRTs)
and the Akaike information criterion (AIC). Based on recommenda-
tions by Posada and Buckley (2004), we used the model selected by
the Akaikecriterion for the
Base = (0.2393, 0.2548, 0.2819) Nst = 6 Rmat = (1.0000, 3.9211,
1.0000, 1.0000,4.9209)Rates = gamma
var = 0.5620). This analysis produced 14 trees and the strict con-
sensus tree provided resolution between and within most of the
clades found in the MP analysis. One of the 14 trees is shown in
Fig. 2 with Bayesian posterior probabilities and MP bootstrap sup-
port indicated above and below the branches, respectively. The in-
group is monophyletic and the first split is between a F. pleurocarpa
plus F. macrophylla clade and the rest of the section. Among the rest
of Malvanthera we find two major clades. In the first major clade, F.
glandifera is sister to the F. augusta, F. heteromeka plus F. xylosycia
clade, and F. hesperidiiformis plus F. sterrocarpa is sister to all. In
the F. hesperidiiformis–F. sterrocarpa clade, there were multiple
well-supported clades that indicate the presence of at least three
species in the complex (see discussion in Section 4.3). The other
major clade includes F. triradiata as sister to the remainder of the
Australian taxa, which form a trichotomy of (1) F. crassipes and F.
destruens; (2) F. lilliputiana F. obliqua, F. platypoda, and F. subpuber-
ula; and (3) F. atricha, F. brachypoda, F. rubigionosa and F.
watkinsiana.
The Bayesian analysis provided a tree that was identical to the
ML consensus tree, except that two more clades were resolved,
but only with pp = 0.51. These weakly supported clades included
F. crassipes–F. destruens as sister to the F. atricha–F. watkinsiana
clade, and within this clade, F. brachypoda as sister to the
remainder.
The analysis constraining Australian lithophytes to be mono-
phyletic produced 247 trees, which were only 8 steps longer
(565 steps) and were not significantly different from the 15,000
trees from the unconstrained combined analysis (p = 0.1276).
ML search(Parameters: Lset
Shape = 0.8798 Pin-
4. Discussion
4.1. Infrasectional relationships of section Malvanthera
The results of our molecular phylogenetic study do not support
any of the previous classifications of section Malvanthera (Corner,
1965; Dixon, 2003; Berg and Corner, 2005, Table 1). Instead we
find correspondence with geography. Two major lineages (subsec-
tions Malvantherae and Platypodeae) are primarily Australian and a
third major lineage (subsection Hesperidiiformes) has its centre of
diversity in New Guinea. Based on these results we propose a
new subdivision of section Malvanthera that reflects phylogenetic
relationships (Table 1, Fig. 2).
4.2. Homoplasy of characters previously used for subdivision of section
Malvanthera
The connection between phylogenetic subdivisions and bioge-
ography was not apparent in any of the previous subdivisions of
section Malvanthera, because all of the macro-morphological char-
acters used for subdivision are homoplasious. Most of the charac-
ters mapped in Fig. 3 vary even between closely related species.
For example, F. macrophylla has imbricate bracts, a triradiate osti-
ole, and a simple stigma, whereas its sister species, F. pleurocarpa,
has valvate bracts, a bilabiate ostiole, and a bifid stigma. However,
the close relationship of F. macrophylla and F. pleurocarpa is sup-
ported by several other lines of evidence. Both species are hemi-
epiphytes or trees occurring in tropic rainforest in Northeast
Queensland and in coastal habitats from Queensland to New South
Wales and both species have distinct basal veins, three ostiolar
bracts, female florets embedded in the receptacle, and a similar
indumentum on the abaxial surface of the leaves.
Dixon (2003) used the shape of the basal bracts to divide ser.
Malvanthereae with imbricate bracts from ser. Hesperidiiformes
with valvate bracts. All species from New Guinea have valvate
bracts, and Dixon’s scheme was the first to unite F. glandifera with
the other species from New Guinea. However, the Australian F.
pleurocarpa (subsect. Malvanthereae), as well as F. crassipes and F.
triradiata (subsect. Platypodeae) also have valvate bracts, and were
therefore placed in ser. Hesperidiiformes with geographically dis-
junct species from New Guinea.
Some of the morphological characters used previously to deli-
mit groups are still useful. For example, within subsection Platypo-
deae, the persistence of basal bracts separates series Eubracteatae
and Crassipeae with persistent basal bracts from series Rubiginosae
and Obliqua with cauducous basal bracts. Some characters appear
to be associated with aspects of ecology, such as the basal or pri-
mary lateral veins of leaves, which are inconspicuous in the Austra-
lian lithophytes, but prominent in the rainforest species. In
summary, phylogenetic relationships were not reflected in previ-
ous subdivisions of Malvanthera because the morphological charac-
ters used to delimit taxa are quite homoplasious.
4.3. Species delimitation within subsection Hesperidiiformes in New
Guinea
Berg and Corner (2005) united most of Corner’s (1965) New
Guinean species under an expanded F. hesperidiiformis. Our results
suggest that Corner ´s species must be maintained. Based on phylog-
eny alone, it is clear that an expanded F. hesperidiiformis concept is
not monophyletic unless it also includes F. glandifera, which was
retained as a species (Berg and Corner, 2005).
Berg and Corner (2005) recognised no clear separation of F. hes-
peridiiformis and F. sterrocarpa and species delimitation in this
complex remains uncertain. Our phylogenetic analyses (Figs. 2
and 3) show a clade of coastal samples from the Wewak and Mad-
ang provinces, which we recognize as F. hesperidiiformis. The F. ster-
rocarpa samples are divided into two clades, a clade with hill forest
samples found at 500–900 m, which we recognize as F. sterrocarpa,
and a second clade with lowland samples from Sepik, which may
represent a separate species here designated F. cf. sterrocarpa. We
think that there are at least three species in this complex, but fur-
ther studies including more extensive sampling and comparison of
morphological characters are needed.
The existence of several species in the F. hesperidiiformis–F. ster-
rocarpa complex is further supported by preliminary tests of pair-
wise distances of ITS among figs and the cox1 gene among their
specific pollinators (Rønsted and Weiblen, unpublished data). Dis-
tances between F. hesperidiiformis and F. sterrocarpa are in the same
range as distances between other New Guinean species or between
closely related Australian species. Likewise, pairwise distances of
Pleistodontes immaturis, P. cf. immaturis, and P. plebejus pollinating
F. sterrocarpa, F. cf. sterrocarpa and F. hesperidiiformis, respectively,
are comparable to distances between pollinators of closely related
Australian Malvanthera species (Lopez–Vaamonde et al. 2001).
4.4. Biogeography and habitat use within section Malvanthera
We suggest that section Malvanthera is of Australian origin
(Figs. 2 and 3) given that the sister lineage to the New Guinean
Malvanthera and most Australian Malvanthera is a F. pleurocarpa–
F. macrophylla clade (subsection Malvanthereae). Ficus pleurocarpa
and F. macrophylla subsp. macrophylla are endemic to the Austra-
lian mainland, while F. macrophylla subsp. columnaris probably
represents a later colonization event to Lord Howe Island, where
20
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22

Page 11

Author's personal copy
this subspecies is endemic (Dixon, 2001c). Our analyses indicate
that subsp. columnaris could be the oldest lineage in the clade,
but more sampling of both subspecies is needed to clarify this.
According to Rønsted et al. (2005), section Malvanthera originated
at least 41 million years ago and radiated gradually from about 35
million years ago. The section probably colonized New Guinea once
(subsection Hesperidiiformes) with F. glandifera and F. xylosycia sub-
sequently colonising other islands in Oceania.
In subsection Platypodeae, (Figs. 2 and 3), F. triradiata (series
Eubracteatae), is a rainforest hemi-epiphyte and represents the first
diverging lineage, although only well supported by the Bayesian
analysis. Series Crassipeae (F. crassipes and F. destruens) are also
hemi-epiphytic rainforest species. In series Obliquae, the first
diverging lineage is F. obliqua, which occurs both as a hemi-epi-
phyte and as a lithophyte throughout its range, whereas the
remaining species in this series are all lithophytes. Our results
are consistent with a pattern of radiation within series Obliquae
from the rainforest into drier habitats in Australia with a transition
from tall hemi-epiphytic trees to small trees and shrubs. Within
series Rubiginosae, F. watkinsiana is restricted to rainforest, F. rubi-
ginosa can both be hemi-epiphytic and lithophytic, whereas F. atri-
cha and F. brachypoda are both lithophytic shrubs or small trees.
Relationships within series Rubiginosae are not well resolved in this
study, but this clade possibly represents a second radiation from
the forest into drier habitats in Australia parallel to series Obliquae.
A Templeton test showed that trees with all five Australian litho-
phytes constrained as monophyletic was not significantly different
from unconstrained trees, indicating that the transition from forest
to arid zones might have happened only once.
However, in addition to the lithophytes, we also find transitional
species,suchasF.rubiginosaandF.obliquainAustraliaandF.glandif-
erainNewGuineaandOceania.FicusglandiferaispartofaNewGuin-
ean rainforest clade and is not closely related to the Australian
lithophytic and transitional species further supporting the possibil-
ity that colonisationof arid environmentsmay have evolved at least
twice.In conclusionthere seemsto be morehomoplasyin morphol-
ogy and ecology within section Malvanthera than previous subdivi-
sions of the section indicated. This agrees with a hypothesis of
multiple gains of active pollination behaviour in Pleistodontes spe-
cies pollinating section Malvanthera (Cook et al., 2004).
4.5. Evolution of pollination behaviour
Cook et al. (2004) mapped changes in pollination behaviour
onto a phylogeny of Pleistodontes. They found that there have been
three to six changes in pollination behaviour within Pleistodontes,
emphasizing that co-evolution does not reach an endpoint and that
selection pressures on mutualists are not constant. Changes in pol-
lination behaviour correlated perfectly with changes in anther/
ovule ratio in the host figs as also found by Kjellberg et al. (2001)
and no evidence was found of phylogenetic restrictions at the spe-
cies level (Cook et al. 2004). Although all actively pollinated Mal-
vanthera, have a simple stigma (Figs. 3 and 4), passively
pollinated species either have a bifid or a simple stigma, implying
limited correlation between this stigma trait and pollination
behaviour.
We used F. ingens (subgenus Urostigma section Urostigma) and
its active pollinator Platyscapa soraria as an outgroup. As in Cook
et al. (2004) the ancestral state for Malvanthera/Pleistodontes was
inferred to be passive pollination, implying one change from active
pollination, although this is wholly dependent on outgroup status.
Within Malvanthera, one change to active pollination is needed for
series Obliquae, and at least one change to active pollination within
the ser. Rubiginosae. With the presence of small pollen pockets, F.
watkinsiana and its pollinator P. nigriventris, could represent a sec-
ondary loss of active pollination, depending on the true relation-
ship within this series. However, the number of pollination mode
changes is not certain and at least one inferred ingroup change de-
pend on weakly supported nodes within subsect. Platypodeae.
Accordingly, mapping traits associated with pollination behaviour
onto our phylogenetic hypothesis for Malvanthera, implies only 1–
3 changes within the section, in contradiction with 3–6 changes
previously inferred by mapping pollination behaviour onto a phy-
logenetic hypothesis for Pleistodontes (Cook et al. 2004).
It is noteworthy, that the pattern of radiation out of the rainfor-
est habitat into drier habitats associated with a transition from tall
hemi-epiphytic trees to small trees and shrubs (Fig. 3), also seems
to be associated with a gain of active pollination behaviour (Fig. 4).
However, both the phylogenies of the figs and the pollinators needs
improved resolution and support before we can really understand
the patterns and processes of co-evolution in the Malvanthera–
Pleistodontes mutualism.
Acknowledgments
N.R. thank M. Thomas (K), H.V. Hansen (C), D. Dixon (DNA) and
C.C. Berg (BG) for providing plant material, Jean-Yves Rasplus for
information on Pleistodontes blandus, Dale Dixon, Finn Kjellberg
and an anonymous reviewer for comments on an earlier version
of this manuscript. This work was partly supported by a Marie Curie
Outgoing International Fellowship within the 6th European Commu-
nity Framework Program (N.R./V.S.) and the NERC (UK) (J.M.C.).
References
Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S.,
Donoghue, M.J., 1995. The ITS region of nuclear ribosomal DNA: a valuable
source of evidence on angiosperm phylogeny. MO Bot. Gard. 82, 247–277.
Baldwin, B.G., Markos, S., 1998. Phylogenetic utility of the external transcribed
spacer (ETS) of 18S–26S rDNA: congruence of ETS and ITS trees of Calycadenia
(Compositae). Mol. Phyl. Evol. 10, 449–463.
Berg, C.C., 2002. Ficus baola, a new species of Ficus subgenus Urostigma section
Malvanthera (Moraceae) from the Solomon Islands. Blumea 47, 315–317.
Berg, C.C., 2003. Flora Malesiana precursor for the treatment of Moraceae 1: the
main subdivision of Ficus: the subgenera. Blumea 48, 167–178.
Berg, C.C., Corner, E.J.H., 2005. Moraceae (Ficus). In: Noteboom, H.P. (Ed.), Flora
Malesiana ser. 1, vol. 17. National Herbarium of Nederland, Leiden, The
Netherlands, pp. 1–730.
Chew, W.L., 1989. Moraceae. In: George, S. (Ed.), Flora of Australia, vol. 3. Australian
Government Printing Service, Canberra, pp. 15–68.
Cook, J.M., Bean, D., Power, S.A., Dixon, D.J., 2004. Evolution of a complex coevolved
trait: active pollination in a genus of fig-wasps. Evol. Biol. 17, 238–246.
Cook, J.M., Rasplus, J.-Y., 2003. Mutualists with attitude: coevolving fig wasps and
figs. TREE 18, 241–248.
Corner, E.J.H., 1959. Taxonomic notes on Ficus Linn., Asia and Australasia. I. Subgen.
Urostigma (Gasp.) Miq. Gard. Bull. Sing. 17, 368–404.
Corner, E.J.H., 1965. Checklist of Ficus in Asia and Australasia with keys to
identification. Gard. Bull. Sing. 21, 1–186.
Dixon, D.J., 2001a. A chequered history: the taxonomy of Ficus platypoda and F.
leuchotricha (Moraceae: Urostigma sect. Malvanthera) unravelled. Austr. Syst.
Bot. 14, 535–563.
Dixon, D.J., 2001b. Ficus lilliputiana (Moraceae), a new species from the Kimberley
region of Western Australia and the Northern Territory. Nuytsia 13, 457–464.
Dixon, D.J., 2001c. Figs, wasps and species concepts: a re-evaluation of the
infraspecifictaxa of
Ficus macrophylla
Malvanthera). Austr. Syst. Bot. 14, 125–132.
Dixon, D.J., 2001d. Figuring out the figs: the Ficus obliqua—Ficus rubiginosa complex
(Moraceae: Urostigma sect. Malvanthera). Austr. Syst. Bot. 14, 133–154.
Dixon, D.J., 2003. A taxonomic revision of the Australian Ficus species in the section
Malvanthera (Ficus subg. Urostigma: Moraceae). Telopia 10, 125–153.
Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure from small quantities of
fresh leaf tissue. Phytochem. Bull. 19, 11–15.
Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the
bootstrap. Evolution 39, 783–791.
Fitch, W.M., 1971. Towards defining the course of evolution: minimum change for a
specific tree typology. Syst. Zool. 20, 406–416.
Haine, E.R., Martin, J., Cook, J.M., 2006. Deep mtDNA divergences indicate cryptic
species in a fig-pollinating wasp. BMC Evol. Biol. 6, 83.
Harrison, R.D., 2005. Figs and the diversity of tropical rainforests. Bioscience 55,
1053–1064.
Jousselin, E., Rasplus, J.-Y., Kjellberg, F., 2003. Convergence and coevolution in a
mutualism: evidence from a molecular phylogeny of Ficus. Evolution 57, 1255–
1269.
(Moraceae:
Urostigma
sect.
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22
21

Page 12

Author's personal copy
Kjellberg, F., Jousselin, E., Bronstein, J.L., Patel, A., Yokoyama, J., Rasplus, J.-Y., 2001.
Pollination mode in fig wasps: the predictive power of correlated traits. Proc.
Roy. Soc. Lond. B 268, 113–1121.
Lopez-Vaamonde, C., Rasplus, J.Y., Weiblen, G.D., Cook, J.M., 2001. Molecular
phylogenies of figwasps: partial cocladogenesis of pollinators and parasites.
Mol. Phyl. Evol. 21, 55–71.
Lopez-Vaamonde, C., Dixon, D.J., Cook, J.M., Rasplus, J.-Y., 2002. Revision of
Australian speciesof
Pleistodontes
pollinating wasps and their host plant associations. Zool. J. Linn. Soc.
136, 637–683.
Machado, C.A., Robbins, N., Gilbert, T.P., Herre, E.A., 2005. Critical review of host
specificity and its coevolutionary implications in the fig/fig-wasp mutualism.
Proc. Natl. Acad. Sci. USA 102, 6558–6565.
Maddison, D.R., 1991. The discovery and importance of multiple islands of most-
parsimonius trees. Syst. Zool. 40, 315–328.
Maddison, D.R., Maddison, W.P., 2001. MacClade. Sinauer Sunderland, MA, USA.
Posada, D.,Buckley, T.,2004.Model
phylogenetics: advantages of Akaike information criterion and Bayesian
approaches over likelihood ration tests. Syst. Biol. 53, 793–808.
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution.
Bioinformatics 14, 817–818.
Ronquist, F., Hulsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference
under mixed models. Bioinformatics 19, 1572–1574.
Rønsted, N., Weiblen, G.D., Cook, J.M., Salamin, N., Machado, C.A., Savolainen, V.,
2005. 60 million years of co-divergence in the fig-wasp symbiosis. Proc. Roy.
Soc. Lond. B 272, 2593–2599.
(Hymenoptera: Agaonidae)fig-
selection andmodelaveraging in
Rønsted,N.,Yektaei-Karin,E.,Turk,K.,Clarkson,J.M.,Chase,M.W.,2007.Species-level
phylogeneticsoflargegenera:prospectsofstudyingco-evolutionandpolyploidy.
In:Hodkinson, T.,Parnell, J.(Eds.),Reconstructingthe TreeofLife:Taxonomy and
Systematics of Species Rich Taxa. Systematic Association Series. CRC Press.
Rønsted, N., Weiblen, G.D., Clement, W., Zerega, N., Savolainen, V., 2008.
Reconstructing the phylogeny of figs (Ficus, Moraceae) to unravel the origin of
fig-wasp mutualisms. Symbiosis 45, 45–56.
Strand, A.E., Leebens-Mack, J., Milligan, B.G., 1997. Nuclear DNA-based markers for
plant evolutionary biology. Mol. Ecol. 6, 113–118.
Sun, Y., Skinner, D.Z., Liang, G.H., Hulbert, S.H., 1994. Phylogenetic analysis of
Sorghum and related taxa using internal transcribed spacers of nuclear
ribosomal DNA. Theor. Appl. Genet. 89, 26–32.
Swofford, D.L. 2002. PAUP*: Phylogenetic Methods Using Parsimony (*and other
methods), version 4, Sinauer, Sunderland, MA.
Templeton, A.R., 1983. Phylogenetic inference from restriction endonuclease
cleavage site maps with particular reference to the evolution of humans and
the apes. Evolution 37, 221–244.
Weiblen, G.D., 2000. Phylogenetic relationships of functionally dioecious Ficus
(Moraceae) based on ribosomal DNA sequences and morphology. Am. J. Bot. 87,
1342–1357.
White, T.J., Bruns, T., Lee, S., Taylor, J., 1990. Amplification and direct sequencing of
fungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H.,
Sninsky, J.J., White, T.J. (Eds.), PCR Protocols. Academic Press, San Diego, CA,
USA, pp. 315–322.
Wiebes, J.T., 1994. The Indo-Australian Agaoninae (pollinators of figs). Proc. K. Ned.
Akad. Wet. 92, 1–208.
22
N. Rønsted et al./Molecular Phylogenetics and Evolution 48 (2008) 12–22