Molecular Phylogenetics and Evolution 49 (2008) 1–16
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
Radiation of southern African daisies: Biogeographic inferences for subtribe
Arctotidinae (Asteraceae, Arctotideae)
Robert J. McKenzie *, Nigel P. Barker
Molecular Ecology and Systematics Group, Department of Botany, Rhodes University, P.O. Box 94, Grahamstown 6140, South Africa
a r t i c l e
i n f o
Article history:
Received 18 April 2007
Revised 29 June 2008
Accepted 9 July 2008
Available online 18 July 2008
Keywords:
Arctotideae
Arctotidineae
Asteraceae
Biogeography
Cape floral region
Compositae
Dispersal–vicariance analysis
Molecular phylogeny
Molecular dating
a b s t r a c t
The majority of the approximately 80–90 species in subtribe Arctotidinae occur in southern Africa with
the centre of diversity in the winter-rainfall region. Three species are restricted to afromontane eastern
Africa and three species are endemic to Australia. To investigate biogeographic and phylogenetic relationships within Arctotidinae, sequence data from four cpDNA regions (psbA-trnH, trnT-trnL and trnL-trnF
spacers and trnL intron) and the ITS nrDNA region for 59 Arctotidinae species were analyzed with parsimony and Bayesian-inference approaches. Eight well-supported major lineages were resolved. The earliest-diverging extant lineages are afromontane or inhabit mesic habitats, whereas almost all sampled taxa
from the winter-rainfall and semi-arid areas have diverged more recently. Molecular dating estimated
that the major clades diverged during the Miocene and Pliocene, which is coincident with the trend of
increasing rainfall seasonality, aridification and vegetation changes in southwestern Africa. Trans-oceanic
dispersal to Australia was estimated to have occurred during the Pliocene.
Ó 2008 Elsevier Inc. All rights reserved.
1. Introduction
Southern Africa is an important centre of floristic biodiversity.
Approximately 24,000 species of vascular plants are recorded from
the region, of which approximately 80% are endemic to southern
Africa (Cowling and Hilton-Taylor, 1997). Almost 10% of this plant
diversity is contributed by the Asteraceae (the daisy family), with
approximately 250 genera and 2250 species recorded from the region (Koekemoer, 1996). Over 80% of these species are endemic to
southern Africa (Cowling and Hilton-Taylor, 1997). The family has
radiated in all of the recognized southern African biomes and is one
of the more species-rich families in most, if not all, of these biomes.
The diversity and endemism of Asteraceae in southern Africa is
focused in eight putative centres (Koekemoer, 1996). These centres
are generally consistent with centres of endemism indicated by the
southern African flora as a whole (e.g. see van Wyk and Smith,
2001). Despite being a major component of the flora, the evolution
and biogeography of southern African Asteraceae is poorly investigated and well-resolved phylogenetic reconstructions are required
to elucidate the origins and radiation of the Asteraceae lineages in
the region (Galley and Linder, 2006). The only major southern African lineage of Asteraceae that has been investigated to any extent
is the tribe Arctotideae, which has been the subject of both molec* Corresponding author. Fax: +27 46 6229550.
E-mail address: r.mckenzie@ru.ac.za (R.J. McKenzie).
1055-7903/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2008.07.007
ular (McKenzie et al., 2006a) and morphological (Karis 2006) phylogenetic investigations. However, the sampling in previous
studies was not comprehensive and here we present the first
well-sampled phylogenetic and biogeographic investigation of
the predominantly southern African subtribe Arctotidinae to help
to fill this gap in our knowledge.
Arctotideae is a small tribe containing about 215 species in 17
genera with an almost exclusively African distribution (Karis,
2007). Two subtribes are recognized with some authors placing
other lineages in the tribe (for a brief review of the taxonomy
and relationships of the tribe, see Funk et al., 2004). The subtribe
Arctotidinae contains approximately 80–90 species currently classified into five genera (Karis, 2007). One species, Haplocarpha scaposa, has a wide distribution in montane and temperate
southeastern Africa (e.g. Hilliard, 1977; Pope, 1992). Three species
are indigenous to the high-altitude East African mountains (Mesfin
Tadesse, 2004) and three species are endemic to Australia (Holland
and Funk, 2006). The remaining species occur in southern Africa,
with the centre of diversity in the Cape and Succulent Karoo Centres of Floristic Endemism (see van Wyk and Smith, 2001). Both
areas occur in the winter-rainfall zone (Cowling, 1992; Dean and
Milton, 1999; see Fig. 1) and are combined into a ‘Greater Capensis’
floristic region by some workers (Born et al., 2007). Because more
than 50% of the species occur in the Cape Floristic Region (CFR), the
Arctotidinae, and especially Arctotis s.str., might represent a true
‘Cape clade’ sensu Linder (2003, 2005).
2
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Fig. 1. A topographical map of southern Africa with the approximate boundaries of different rainfall regimes indicated (based on Chase and Meadows, 2007).
Understanding the factors responsible for the unusually high
levels of floristic richness and endemism in the winter-rainfall region of southern Africa (and the CFR in particular) has attracted
considerable interest (for reviews for the CFR, see Linder, 2003,
2005). Molecular dating has indicated that plant lineages have
been recruited into the Cape flora since at least the Cretaceous
and the onset of radiations in these lineages is indicated to range
from the Oligocene to the Pliocene (Linder, 2005), so the extant
diversity cannot be explained solely by vicariance or explosive speciation in response to one or more concurrent triggers. Furthermore, the southern African flora as a whole is indicated to have
evolved from disparate sources, including descendants of Gondwanan ancestors (Goldblatt, 1978; Anderson et al., 1999), temperate
southern African progenitors (Goldblatt, 1978), groups of Eurasian
or tropical African origin (e.g. Meerow et al., 2003; Hurka et al.,
2005; Mummenhoff et al., 2005), and Australasian ancestors (e.g.
Linder et al., 2003). Therefore a thorough understanding of diversification processes in the southern African flora is likely to be best
achieved by a taxon-focused, rather than a broad floristic, approach
because of the individuality of taxon histories within a flora (Verboom et al., 2003).
The above evidence indicates that transcontinental and transoceanic dispersal has been important in the compilation of the
present-day southern African flora. Indeed, there is increasing evidence for the importance of long-distance dispersal in determining
extant plant biogeographic patterns worldwide (e.g. Price and Clague, 2002; de Queiroz, 2005). Phylogenetic reconstructions based
on molecular data, in combination with molecular calibrations to
estimate the date of divergence events, have provided strong evidence that in the Southern Hemisphere, following the breakup of
Gondwana, transoceanic dispersal has been important in a diversity of animals (e.g. Vences et al., 2003; Schwarz et al., 2006) and
plant groups (e.g. Baum et al., 1998; Mummenhoff et al., 2004;
Cook and Crisp, 2005; Linder and Barker, 2005; Barker et al.,
2007). Vicariance appears to have been much stronger in animals
than plant groups (Sanmartín and Ronquist, 2004). Of particular
significance to the present study, plant taxa with a disjunct southern African–Australian distribution, such as the Arctotidinae, are
rare (Good, 1964; Thorne, 1972; Goldblatt, 1978), and as yet few
of these taxa have been investigated in a phylogenetic context.
A previous molecular phylogenetic study of Arctotidinae examined sequence variation in five cpDNA regions (ndhF, psbA-trnH,
rps16, trnS-trnfM, and trnT-trnF) and the ITS nrDNA region for 18
species representing the major morphological variation in the subtribe (McKenzie et al., 2006a). Strong incongruence between the
prevailing taxonomy and evolutionary relationships were indicated. A clade comprising two Haplocarpha species (H. nervosa
and H. rueppellii, both formerly segregated in Landtia) from the
southern and East African mountain chain was sister to the rest
of the subtribe, and the widely distributed H. scaposa was indicated
to be an early divergence. The placement of the Australian Cymbonotus lawsonianus and East African H. schimperi in a well-supported
clade along with south-eastern African Arctotis species provided
evidence for dispersal from southern Africa to Australia and migration to East Africa, respectively. However, comprehensive sampling
of extant lineages is required for confident phylogenetic reconstructions and to allow more accurate biogeographic inferences
to be made. The primary objective of the present study was to
reconstruct a species-level phylogenetic hypothesis for a comprehensive sample of Arctotidinae based on sequence data from four
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
cpDNA regions (the psbA-trnH, trnT-trnL and trnL-trnF intergenic
spacers and trnL intron) and the ITS nrDNA region. The specific
aims were to use the phylogeny to: (1) investigate biogeographic
patterns at the continental and regional levels; (2) make a first attempt at dating the diversification of Arctotidinae; and (3) further
explore congruence between species relationships and the taxonomic classification, focusing on the polyphyly of Arctotis and
Haplocarpha.
2. Materials and methods
2.1. Taxon sampling
DNA sequences were obtained for 71 accessions from 52 species
of Arctotidinae. These data were supplemented with published sequences for single accessions of 18 Arctotidinae and four species
from the sister subtribe Gorteriinae (McKenzie et al., 2006a). Thus
the ingroup comprised 89 accessions from 59 species of Arctotidinae, representing the five currently accepted genera, and the outgroup consisted of single accessions of four species of the sister
subtribe Gorteriinae (Table 1).
It is essential to avoid nomenclatural and taxonomic confusion,
and we thus adopt the following nomenclatural usage. Generic concepts as elucidated by Karis (2007) are adopted, and Lewin’s (1922)
infrageneric classification of Arctotis and Beauverd’s (1915) infrageneric classification of Haplocarpha are followed. In addition, the
following corrections of recent treatments by Beyers (2000) and
names that appear in Germishuizen and Meyer (2003) are relevant.
Arctotis decurrens is the correct name for A. merxmuelleri and A. scullyi (McKenzie et al., 2006b). Two specimens (McKenzie 797/3 and
McKenzie 1124/1) belong to a newly recognized species, A. debensis
(McKenzie et al., 2006c). Arctotis flaccida is the correct name for Arctotis sp. ‘1’ designated by Beyers (2000). Arctotis semipapposa (listed
as a synonym of A. flaccida by Beyers, 2000) is a very different, morphologically distinct species (McKenzie et al., in press). Arctotis sp.
‘2’ and sp. ‘4’ were designated by Beyers (2000) and this usage is retained here. Five other undetermined Arctotis species herein are
designated spp. ‘A’, ‘B’, ‘C’, ‘D’, and ‘E’ (see Table 1). A full revision
of the taxonomy of southern African Arctotidinae by the first author
is currently in progress.
3
(Bioline, London), 1 ll 0.1 lM solution of each forward and reverse
primer, 0.2 ll BioTaqÒ DNA polymerase (5 U/ll, Bioline, London,
UK) and 0.75–1.5 ll unquantified DNA extract. The volume of
50 mM MgCl2 varied from 0.75–1.5 ll (1.5–3 mM) for trnT-trnF
and 2–3 ll (4–6 mM) for psbA-trnH. Some reaction solutions contained 1.5–2 ll 0.1% bovine serum albumen. For some ITS amplification reactions, the solution differed in containing 5 ll 5
Colorless GoTaqÒ reaction buffer, 0.25 ll GoTaqÒ DNA polymerase
(5 U/ll, Promega, Madison, WI) and no additional MgCl2.
The DNA regions were amplified using a Hybaid PCR Sprint
thermal cycler. The following parameters in the amplification reactions were standard for all regions: in the first reaction cycle, denaturing 95 °C, 45 s; primer extension 72 °C, 3 min; and the final
extension cycle 72 °C, 10 min. The number of amplification cycles
and primer annealing temperature differed for each region as follows; psbA-trnH: 25 cycles, 52 °C, 45 s; trnT-trnL spacer (A and
B): 27–30 cycles, 52 °C, 45 s; trnT-trnL (A and D): 35 cycles, 52 °C,
45 s; trnL-trnF intron and spacer (C and F): 30 cycles, 52 °C, 45 s;
trnL-trnF spacer (E and F): 25–30 cycles, 52 °C, 45 s; ITS: 30–33 cycles, 48–52 °C, 45 s. The PCR products were purified using the WizardÒ SV Gel and PCR purification kit (Promega, Madison,
Wisconsin) and resuspended in 25–30 ll nuclease-free water.
PCR products were sequenced in both forward and reverse
directions. Sequencing reactions of the PCR products were performed using the BigDyeÒ Terminator v. 3.1 cycle sequencing kit
according to the manufacturer’s instructions (Applied Biosystems,
Foster City, California). The cycle-sequencing products were precipitated using the sodium acetate/EDTA protocol, and electrophoresed and resolved on an ABI Prism 3100 Genetic Analyzer at the
Rhodes University DNA sequencing facility.
2.3. Sequence alignment
For each accession contiguous sequences were compiled with
Sequencher 4.2.2 (Gene Codes Corp., Ann Arbor) and edited visually. All sequences were deposited in GenBank (http://
www.ncbi.nlm.nih.gov/; accession numbers listed in Table 1). Sequences were aligned manually in MacClade 4.06 (Maddison and
Maddison, 2000). Gaps were inserted manually based on visual
inspection of the sequences. Alignments are deposited in TreeBASE
(http://www.treebase.org/).
2.2. DNA extraction, amplification, and sequencing
2.4. Phylogenetic analyses
For most samples total genomic DNA was extracted from about
1 cm2 of either fresh leaf material or leaves desiccated in silica gel
using a CTAB extraction protocol (Doyle and Doyle, 1987). For
Cymbonotus maidenii DNA was extracted from a herbarium specimen. Two non-coding cpDNA regions, the psbA-trnH intergenic
spacer and trnT-trnF region (the trnT-trnL and trnL-trnF intergenic
spacers and trnL intron), and one non-coding nrDNA region (ITS)
were amplified and sequenced. The following primers were used
for amplifying and sequencing: the primers A, B, C, D, E, and F of
Taberlet et al. (1991) and Arct-trnL-R (ATT WTA TCR TTT CTG
TAT CSG; previously unpublished) for the trnT-trnF region; psbAF and trnH-R for the psbA-trnH intergenic spacer (Sang et al.,
1997); ITS18 (Käss and Wink, 1997; modified by Beyra-Matos
and Lavin, 1999), ITS26 (Käss and Wink, 1997), Chrysanth-5.8F
and Chromo-5.8R (Barker et al., 2005), ITS1, ITS4, and ITS5 (White
et al., 1990) for the ITS region. Attempts to amplify both cpDNA regions for the C. maidenii accession and the psbA-trnH spacer for one
Arctotis decurrens sample (Mucina 170903/20) were unsuccessful. In
some instances, each species is represented by multiple samples, in
order to assess species monophyly.
Each 25 ll Polymerase Chain Reaction (PCR) solution contained
2.5 ll 10 PCR buffer (Bioline, London, UK), 1 ll 20 mM dNTPs
Because in flowering plants the chloroplast genome is usually
inherited as a single unit without recombination (Soltis and Soltis,
1998), the cpDNA regions were combined into one data set. Three
data sets were thus analyzed: the cpDNA data set; the ITS data set;
and a ‘total evidence’ data set (Kluge, 1989; Nixon and Carpenter,
1996) with the cpDNA and ITS data sets combined. Congruence
in the phylogenetic signal of the cpDNA and ITS data sets was assessed with the partition homogeneity (or ILD) test (Farris et al.,
1994) using PAUP 4.0b10 (Swofford, 2002) with 750 replicates,
a maximum tree limit of 500, tree bisection–reconnection (TBR)
branch swapping, with uninformative characters excluded following Lee (2001). Although conflict between the cpDNA and ITS data
sets was indicated to be significant, the results must be interpreted
with caution as the ILD test is an unreliable measure of data-set
congruence in some circumstances (Ramírez, 2006). Under certain
conditions this test can indicate significant incongruence between
congruent data (Yoder et al., 2001; Hipp et al., 2004) and non-significance in cases of high incongruence (‘‘hypercongruence” sensu
Ramírez, 2006). As an alternative strategy, conflicting nodes were
regarded to be incongruent if they each received parsimony bootstrap support (BS) > 70% (Mason-Gamer and Kellogg, 1996) or
Bayesian posterior probabilities (PP) > 95% (Alfaro et al., 2003).
4
Table 1
Collection, voucher and GenBank accession information for sequences included in this study
Provenance of sample
Collector(s) and collection number
Herbariumb
Arctotheca calendula (L.) Levyns
Arctotheca forbesiana (DC.) K. Lewin
Arctotheca marginata Beyers
Arctotheca populifolia (P.J. Bergius) Norl.
Arctotheca prostrata (Salisb.) Britten
Arctotis acaulis L. 1
Arctotis acaulis L. 2
Arctotis adpressa DC.
Arctotis angustifolia L. var. latifolia Harv.
Arctotis arctotoides (L.f.) O. Hoffm. 1
Arctotis arctotoides (L.f.) O. Hoffm. 2
Arctotis argentea Thunb.
Arctotis aspera L. var. aspera 1
Arctotis aspera L. var. aspera 2
Arctotis aspera L. var. scabra P.J. Bergius
Arctotis auriculata Jacq.
Arctotis bellidifolia P.J. Bergius 1
Arctotis bellidifolia P.J. Bergius 2
Arctotis breviscapa Thunb.
Arctotis campanulata DC. var. puberula DC.
Arctotis canescens DC.
Arctotis debensis R.J. McKenzie 1
Arctotis debensis R.J. McKenzie 2
Arctotis decurrens Jacq. 1
Arctotis decurrens Jacq. 2
Arctotis dregei Turcz. 1
Arctotis dregei Turcz. 2
Arctotis elongata Thunb.
Arctotis erosa Harv. 1
Arctotis erosa Harv. 2
Arctotis erosa Harv. 3
Arctotis erosa Harv. 4
Arctotis fastuosa Jacq. 1
Arctotis fastuosa Jacq. 2
Arctotis flaccida Jacq. 1
Arctotis flaccida Jacq. 2
Arctotis graminea K. Lewin
Arctotis hirsuta (Harv.) Beauverd
Arctotis hispidula (Less.) Beauverd
Arctotis incisa Thunb. 1
Arctotis incisa Thunb. 2
Arctotis laevis Jacq.
Arctotis lanceolata Harv. 1
Arctotis lanceolata Harv. 2
Arctotis leiocarpa Harv.
Arctotis microcephala (DC.) Beauverd 1
Arctotis microcephala (DC.) Beauverd 2
South Africa: Western Cape
South Africa: Western Cape
South Africa: Northern Cape
South Africa: Eastern Cape
South Africa: Eastern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Eastern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Northern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Eastern Cape
South Africa: Eastern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Eastern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Western Cape
South Africa: Northern Cape
Namibia: Karas
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Eastern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Eastern Cape
Namibia: Karas
South Africa: Eastern Cape
South Africa: Eastern Cape
McKenzie 808/3
McKenzie 1215/1
Bosenburg 2
Barker 1766
Barker 1765
McKenzie 823/4
McKenzie 827/1
McKenzie 1365/1
McKenzie 839
Barker 1950
McKenzie 855/1
McKenzie 1250
McKenzie 844
Samuel 41
McKenzie 1372
McKenzie 1285/1
Barker 1856
McKenzie 1060/1
McKenzie 1383/1
McKenzie 1284
McKenzie 1311/1
McKenzie 797/3
McKenzie 1124/1
McKenzie 1302/1
Mucina 170903/20
McKenzie 815/2
McKenzie 834/4
McKenzie 1085/1
McKenzie 1282
McKenzie 1290
Samuel 15
Barker 1872
McKenzie 1351
Mannheimer 2492
McKenzie 1065/3
McKenzie 1363/2
McKenzie 824/1
McKenzie 1070/2
McKenzie 760/1
McKenzie 826/2
McKenzie 1380/1
McKenzie 1355/1
McKenzie 835
Ramdhani 616
Mucina 150704/7
McKenzie 784/3
McKenzie 1409
GRA
GRA
NBG
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA, WIND
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GenBank accession numbers
ITS
psbA–trnH
trnT–trnF
DQ444720
EU846328
EU846329
EU846330
EU846331
DQ444721
EU846344
EU846345
EU846346
EU846334
DQ444722
EU846347
DQ444723
EU846348
EU846349
EU846350
EU846351
EU846352
DQ444724
EU846353
EU846354
EU846335
EU846336
EU846356
EU846355
EU846332
DQ444725
EU846357
EU846358
EU846359
EU846360
EU846361
EU846362
EU846373
EU846363
EU846364
EU846365
EU846366
EU846337
EU846367
EU846368
EU846369
EU846370
EU846371
EU846372
EU846339
EU846340
DQ444764
EU846399
EU846400
EU846401
EU846402
DQ444765
EU846414
EU846415
EU846416
EU846405
DQ444766
EU846417
DQ444767
EU846418
EU846419
EU846420
EU846421
EU846422
DQ444768
EU846423
EU846424
EU846406
EU846407
EU846425
—
EU846403
DQ444769
EU846426
EU846427
EU846428
EU846429
EU846430
EU846431
EU846442
EU846432
EU846433
EU846434
EU846435
EU846408
EU846436
EU846437
EU846438
EU846439
EU846440
EU846441
EU846409
EU846410
DQ444808
EU846468
EU846469
EU846470
EU846471
DQ444809
EU846483
EU846484
EU846485
EU846474
DQ444810
EU846486
DQ444811
EU846487
EU846488
EU846489
EU846490
EU846491
DQ444812
EU846492
EU846493
EU846475
EU846476
EU846495
EU846494
EU846472
DQ444813
EU846496
EU846497
EU846498
EU846499
EU846500
EU846501
EU846512
EU846502
EU846503
EU846504
EU846505
EU846477
EU846506
EU846507
EU846508
EU846509
EU846510
EU846511
EU846478
EU846479
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Speciesa
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: KwaZulu-Natal
South Africa: Eastern Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Northern Cape
South Africa: Free State
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Western Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
South Africa: Northern Cape
Australia: Australian Capital Territory
Australia: New South Wales
Australia: New South Wales
South Africa: Western Cape ex hort.
South Africa: Western Cape
South Africa: Eastern Cape
South Africa: Eastern Cape
South Africa: Eastern Cape
South Africa: Western Cape
Ethiopia: Bale
Kenya: Central Province
South Africa: Eastern Cape
South Africa: KwaZulu-Natal
Ethiopia: Arsi
Kenya: Central Province
McKenzie 1059/1
McKenzie 822
McKenzie 825/1
McKenzie 777/1
McKenzie 1205/2
McKenzie 1071/2
Samuel 42
McKenzie 1389/1
Ramdhani 508
White 41
McKenzie 811/1
McKenzie 843/1
McKenzie 892/1
McKenzie 1074/1
McKenzie 1281/1
McKenzie 875/1
Barker s.n.
McKenzie 1234/1
Barker 1865
McKenzie 1075
McKenzie 1296/1
McKenzie 1300/1
McKenzie 1336/1
Samuel 31
McKenzie 1339/1
Samuel 35
Samuel 20
Bayer ACT-05001
Weston 2486
Radunz s.n., 5.ii.1982
Barker 1780
McKenzie 845/1
Barker 1767
McKenzie 970
McKenzie 1192
Mucina 290805/13
Barker 1906
Namaganda, Abdillahi & Nakamatle 1746
Barker 1772
Ramdhani 519
Barker 1899
Namaganda, Abdillahi & Nakamatle 1748
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
GRA
NSW
NSW
GRA
GRA
GRA
GRA
GRA
GRA
ETH
EA
GRA
GRA
ETH
EA, GRA
EU846374
EU846341
DQ444726
EU846375
EU846376
EU846377
EU846378
EU846379
EU846342
EU846343
EU846380
EU846381
EU846382
EU846383
DQ444728
DQ444729
EU846384
EU846385
DQ444727
EU846386
EU846387
EU846388
EU846389
EU846390
EU846391
EU846392
EU846393
DQ444730
EU846333
EU846338
DQ444731
DQ444732
DQ444733
DQ444734
EU846323
EU846327
DQ444735
EU846324
DQ444736
EU846325
DQ444737
EU846326
EU846443
EU846411
DQ444770
EU846444
EU846445
EU846446
EU846447
EU846448
EU846412
EU846413
EU846449
EU846450
EU846451
EU846452
DQ444772
DQ444773
EU846453
EU846454
DQ444771
EU846455
EU846456
EU846457
EU846458
EU846459
EU846460
EU846461
EU846462
DQ444774
EU846404
—
DQ444775
DQ444776
DQ444777
DQ444778
EU846394
EU846398
DQ444779
EU846395
DQ444780
EU846396
DQ444781
EU846397
EU846513
EU846480
DQ444814
EU846514
EU846515
EU846516
EU846517
EU846518
EU846481
EU846482
EU846519
EU846520
EU846521
EU846522
DQ444816
DQ444817
EU846523
EU846524
DQ444815
EU846525
EU846526
EU846527
EU846528
EU846529
EU846530
EU846531
EU846532
DQ444818
EU846473
—
DQ444819
DQ444820
DQ444821
DQ444822
EU846463
EU846467
DQ444823
EU846464
DQ444824
EU846465
DQ444825
EU846466
Outgroup taxa:
Berkheya carduoides (Less.) Hutch.
Cuspidia cernua (L.f.) B.L. Burtt
Gazania krebsiana Less.
Hirpicium echinus Less.
South
South
South
South
Barker 1924
Barker 1896
Barker s.n.
McKenzie 861
GRA
GRA
GRA
GRA
DQ444716
DQ444717
DQ444718
DQ444719
DQ444760
DQ444761
DQ444762
DQ444763
DQ444804
DQ444805
DQ444806
DQ444807
Africa:
Africa:
Africa:
Africa:
Eastern Cape
Eastern Cape
Eastern Cape
Northern Cape
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Arctotis muricata Thunb.
Arctotis perfoliata (L.f.) O. Hoffm. 1
Arctotis perfoliata (L.f.) O. Hoffm. 2
Arctotis pinnatifida Thunb.
Arctotis reptans Jacq.
Arctotis revoluta Jacq. 1
Arctotis revoluta Jacq. 2
Arctotis rotundifolia K. Lewin
Arctotis scapiformis Thell. 1
Arctotis scapiformis Thell. 2
Arctotis semipapposa (DC.) Beauverd 1
Arctotis semipapposa (DC.) Beauverd 2
Arctotis stoechadifolia P.J. Bergius 1
Arctotis stoechadifolia P.J. Bergius 2
Arctotis sulcocarpa K. Lewin
Arctotis venusta Norl.
Arctotis verbascifolia Harv. 1
Arctotis verbascifolia Harv. 2
Arctotis sp. ‘2’
Arctotis sp. ‘4’
Arctotis sp. ‘A’
Arctotis sp. ‘B’
Arctotis sp. ‘C’ 1
Arctotis sp. ‘C’ 2
Arctotis sp. ‘D’ 1
Arctotis sp. ‘D’ 2
Arctotis sp. ‘E’
Cymbonotus lawsonianus Gaudich. 1
Cymbonotus lawsonianus Gaudich. 2
Cymbonotus maidenii (Beauverd) A.E. Holland & V.A. Funk
Dymondia margaretae Compton
Haplocarpha lanata Less.
Haplocarpha lyrata Harv.
Haplocarpha nervosa (Thunb.) Beauverd 1
Haplocarpha nervosa (Thunb.) Beauverd 2
Haplocarpha parvifolia (Schltr.) Beauverd
Haplocarpha rueppellii (Sch. Bip.) Beauverd 1
Haplocarpha rueppellii (Sch. Bip.) Beauverd 2
Haplocarpha scaposa Harv. 1
Haplocarpha scaposa Harv. 2
Haplocarpha schimperi (Sch. Bip.) Beauverd 1
Haplocarpha schimperi (Sch. Bip.) Beauverd 2
GenBank numbers in bold are new submissions.
a
Duplicate samples of a species are numbered sequentially.
b
EA, East African Herbarium, Nairobi; ETH, National Herbarium of Ethiopia, Addis Ababa; GRA, Selmar Schonland Herbarium, Grahamstown; NBG, Compton Herbarium, Cape Town; NSW, National Herbarium of New South
Wales, Sydney; WIND, National Herbarium of Namibia, Windhoek.
5
6
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Parsimony analyses were performed using PAUP. For each data
set, a heuristic search was conducted with 1000 simple taxonaddition replicates, TBR branch swapping, and the MULTREES option in effect in order to ensure that multiple islands of equally parsimonious trees were found (Maddison, 1991). Further heuristic
searches employed 1000 random taxon-addition replicates, holding one tree at each step, TBR branch swapping, with the MULTREES and STEEPEST DESCENT options in effect, for each
replicate saving at most one tree P the tree length from the simple
heuristic search, followed by a complete heuristic search on the
saved trees with a maximum tree limit of 100,000. In all analyses
no shorter trees were recovered than from the simple-addition
search. Uninformative characters were excluded before all analyses
and all characters were equally weighted and unordered. Parsimony analyses were repeated with unambiguous insertion/deletion events (indels) greater than 2 bp in length recoded as a
binary (presence/absence) character, following Bayer et al.
(2002), using the ‘simple indel coding’ method of Simmons and
Ochoterena (2000). Parsimony bootstrap assessments for each data
set were carried out using 1000 replicates in PAUP. The ITS data
set was analyzed twice. In an initial analysis, all sites were considered, while a second analysis excluded four ambiguously aligned
regions (aligned bp 70–79, 208–232, 256–262, and 479–497). Relative branch support was assessed using nonparametric bootstrap
analysis (Felsenstein, 1985) with 1000 replicates, TBR branch
swapping and a maximum tree limit of 500.
Bayesian inference (BI) analyses using Markov Chain Monte
Carlo methods (Yang and Rannala, 1997) were performed using
MrBayes 3.1.1 (Huelsenbeck and Ronquist, 2001). Prior to analysis,
sequences were partitioned into coding, exon, intron, and intergenic spacer regions based on sequences from the chloroplast genomes of Lactuca sativa, Nicotiana sylvestris, and N. tabacum from
GenBank, and the ITS secondary structure model for the Asteraceae
of Goertzen et al. (2003). Eleven partitions were defined. The psbAtrnH sequence comprised the psbA-trnH spacer, 53 bp at the 30 end
of the psbA gene, and 32 bp at the 50 end of the trnH gene. The trnTtrnF region comprised the trnL intron, the flanking trnL 50 and 30
exons, and the trnT-trnL and trnL-trnF intergenic spacers. The ITS
region comprised the 5.8S gene and the flanking internal transcribed spacers ITS1 and ITS2 (Baldwin, 1992). An optimal nucleotide-substitution model for each partition (Table 2) was selected
for use in the BI analyses using the Akaike information criterion
with MrModeltest 2.2 (J.A.A. Nylander, Uppsala University,
Uppsala).
Two independent BI analyses each with four Markov chains,
three heated and one cold, starting from a random tree were run
simultaneously for 1 106 generations with trees sampled every
100 generations. The trees sampled prior to stabilization of the
log-likelihood value were discarded as ‘burn-in’ samples; the
burn-in varied from 23,000 generations for the ITS data analysis,
30,000 for the cpDNA analysis and 40,000 for the total-evidence
analysis. The remaining trees from the simultaneous runs were
combined and used to generate a 50% majority rule consensus tree
and to determine the PP for each node. The consensus trees from
the Bayesian inference analyses are deposited in TreeBASE
(http://www.treebase.org/; submission No. SN3970).
2.5. Molecular dating
Ideally, dating of molecular phylogenies requires either reliable
calibration points or a known rate of molecular evolution, but
unfortunately obtaining reliable calibration points to date phylogenetic reconstructions within Asteraceae is problematic due to the
family’s poor fossil record. Consequently, divergence dates were
estimated from the ITS data set utilizing ITS mutation rates published for other Asteraceae taxa following the methodology of S
Table 2
Nucleotide-substitution model for each partition specified in the Bayesian inference
analyses
Partition
Modela
Rate variation across sites
psbA
psbA–trnH spacer
trnH
trnT–trnL spacer
trnL 30 exon
trnL
trnL 50 exon
trnL–trnF spacer
ITS1
5.8S
ITS2
JC
GTR
JC
GTR
JC
GTR
JC
GTR
SYM
GTR
SYM
Equal
Gamma
Equal
Gamma
Equal
Equal
Equal
Gamma
Gamma
Proportion invariable + gamma
Gamma
a
GTR, General Time Reversible; JC, Jukes-Cantor; SYM, Symmetrical.
Howis (unpublished data). Three mutation rates (‘slow’, ‘average’,
and ‘fast’) were used. Mutation rates published for Asteraceae
range from 2.51 10 9 substitutions per site per year (s s 1 y 1;
‘slow’) in Eupatorium (Schmidt and Schilling, 2000) to 7.83
10 9 s s 1 y 1 (‘fast’) in Robinsonia (Sang et al., 1995). These rates
span much of the range of rates published for angiosperms as a
whole (see Kay et al., 2006). An ‘average’ rate (4.58 10 9
s s 1 y 1) was calculated based on the rates for four Asteraceae
taxa listed by Kay et al. (2006); this rate was virtually identical
to the average rate (4.59 10 9 s s 1 y 1) for the 10 herbaceous
angiosperm taxa listed by the same authors.
The ITS data set and cladogram were reduced to 50 terminals
to exclude duplicates of species with almost identical sequences.
Log-likelihood ratio tests were performed to compare the likelihoods of obtaining the same topology with or without a clock
assumption. An appropriate evolutionary model was selected
with MrModeltest 2.2. Maximum likelihood analyses were performed with PAUP with and without a clock enforced and the
difference in log-likelihoods was tested by means of a v2 test.
As the assumption of a strict molecular clock was rejected, divergence dates were estimated using a relaxed clock with Bayesian
inference and MCMC procedures implemented in BEAST 1.4.7
(Drummond and Rambaut, 2007). A relaxed clock model allows
the among-branch evolution rate to vary. In addition, BEAST
has the advantage of permitting specification of a number of evolutionary models in a single framework (Drummond and Rambaut, 2007). Consistent with the MrModeltest results, a General
Time
Reversible
nucleotide-substitution
model
with
gamma + invariant sites was used. A relaxed molecular clock
with branch substitution rates drawn from a lognormal distribution, auto-optimization of the operators, no topological constraints and a constant speciation rate per lineage (i.e. Yule
tree prior) with a uniform prior distribution was used. Rates
were sampled every 1000 cycles from 50,00,000 MCMC steps
with a burn-in of 500,000 cycles. Running the analysis for
100,00,000 steps had no impact on the estimated divergence
dates (data not presented). For all analyses two independent runs
were performed, the log files were combined to check for convergence on the same distribution and to ensure adequate sample
sizes, and viewed using Tracer v.1.4 (distributed with BEAST).
Divergence dates estimated when branch substitution rates were
drawn from an exponential distribution were considerably older
and with much wider 95% highest posterior density (HPD) limits
than those derived from a lognormal distribution, and in some
instances the lower HPD bound extended well beyond the probable origin of the Asteraceae of c. 60 Mya. Therefore analyses
utilising a lognormal distribution were the more plausible and
are only considered herein.
7
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
2.6. Dispersal–vicariance analysis
To infer the ancestral areas of the basal and internal nodes, a
dispersal–vicariance analysis was performed with DIVA 1.1 (Ronquist, 1997). A fully dichotomized tree (a requirement of DIVA)
was created based on a neighbor-joining (NJ) tree produced using
PAUP. As can be seen from Figs. 2 and 3, not all clades were fully
resolved, so use of a NJ approach obviated a subjective resolution
of species-level relationships. We acknowledge that there are
many possible permutations of species relationships in poorly resolved clades. However, in this study lack of resolution affected
primarily the ‘core Arctotis’ clade, all members of which are from
the Fynbos or Succulent Karoo regions, and exploration of alternative relationships is unlikely to affect the overall result of the DIVA
analysis. It is emphasised that the basal and internal nodes of the
NJ tree, which were the nodes of interest, were identical to those
in the maximum parsimony and Bayesian total-evidence
phylogenies.
The selection of unit areas of endemism used in DIVA analyses
can be problematic, as different workers may not agree on boundaries of regions of endemism, and some areas might be viewed as
subsets of larger areas. For this analysis, we mostly used areas that
correspond to vegetation units, namely afromontane, Fynbos,
Nama-Karoo, Succulent Karoo, and Savanna biomes, as well as
the widely recognized Albany Centre of Floristic Endemism (Cowling and Hilton-Taylor, 1997). We thus chose to differentiate between the Fynbos and Succulent Karoo, and not consider them as
part of a larger floristic region (as suggested by Born et al., 2007,
for example). Australia was treated as a single area. Analyses were
run without limiting the number of ancestral area optimizations.
The exclusion of some unsampled species may affect the results
of a DIVA analysis, just as they can affect phylogenetic interpretations. In the present study, the majority of the unsampled taxa are
Arctotis species and have CFR distributions and, based on morphological data, it is predicted they will be members of the ‘core Arctotis’ clade. This would not affect the outcome of the DIVA analysis.
3. Results
3.1. Phylogenetic analyses
The sequence characteristics and variability of each DNA region
and the combined datasets are summarized in Table 3. For each
data set parsimony and BI generated almost identical phylogenetic
reconstructions. Consequently, in the following discussion, the
method of analysis is only specified when a phylogenetic arrangement is in conflict. In all analyses, the ‘Landtia’ clade (comprising
Haplocarpha nervosa, H. rueppellii, and H. parvifolia) was sister to
the rest of the Arctotidinae with maximum BS and PP support.
Haplocarpha scaposa and Dymondia margaretae were placed on subsequently diverging, well-supported monotypic branches. The
remaining taxa were placed in four strongly supported clades: a
large ‘core Arctotis’ clade; a ‘Cymbonotus’ clade (composed of
Cymbonotus lawsonianus, H. schimperi and species from Arctotis
sect. Austro-orientales); a clade consisting of the two sampled species from Arctotis sect. Anomalae; and an Arctotheca–‘core Haplo-
A. cpDNA
B. ITS
Berkheya carduoides
Cuspidia cernua
100
1
100
1
97
1
100
1
0.01
Gazania krebsiana
Hirpicium echinus
Dymondia margaretae
Cymbonotus maidenii
Arctotis scapiformis2
Arctotis scapiformis1
78
Arctotis microcephala1
1
Arctotis microcephala2
Haplocarpha schimperi1
76
Haplocarpha schimperi2 *
0.99
‘Cymbonotus’
72
Arctotis hispidula
59
*
Arctotis arctotoides1
1
0.83
Arctotis arctotoides2
Arctotis debensis1
Arctotis debensis2
100
Cymbonotus lawsonianus2
Cymbonotus lawsonianus1
1
92
Haplocarpha lanata
Haplocarpha lyrata
1 53
—
Arctotheca forbesiana
1
Arctotheca marginata
100
0.53
Arctotheca +
Arctotheca prostrata
1
Arctotheca
calendula
‘core Haplocarpha’
Arctotheca populifolia *
Arctotis dregei1
100
Arctotis dregei2
Arctotis sect.
1
Arctotis sulcocarpa
Arctotis breviscapa
—
Anomalae
Arctotis perfoliata1 100
0.67
Arctotis perfoliata2
1
Arctotis campanulata puberula
Arctotis lanceolata2
Arctotis graminea
Arctotis semipapposa1
Arctotis decurrens2
Arctotis decurrens1
Arctotis spE
66
Arctotis acaulis1
1
Arctotis acaulis2
Arctotis adpressa
Arctotis lanceolata1
Arctotis argentea
*
Arctotis aspera scabra
Arctotis verbascifolia1
Arctotis sp2
Arctotis spA
66
Arctotis aspera1
Arctotis incisa2 *
0.67
Arctotis stoechadifolia2
Arctotis angustifolia latifolia
97
Arctotis aspera2
1
Arctotis bellidifolia1
Arctotis bellidifolia2
Arctotis elongata
100
Arctotis reptans
Arctotis revoluta1
1
‘core Arctotis’
Arctotis rotundifolia
Arctotis semipapposa2
Arctotis verbascifolia2
Arctotis pinnatifida
Arctotis incisa1
Arctotis laevis *
Arctotis revoluta2
Arctotis auriculata
92
Arctotis spD2 *
1
Arctotis spD1
Arctotis flaccida1
Arctotis canescens
Arctotis flaccida2
Arctotis spC2
Arctotis erosa1
Arctotis spC 1
Arctotis erosa3
Arctotis muricata
Arctotis fastuosa 2
87
Arctotis stoechadifolia1
Arctotis graminea
1
Arctotis sp4
Arctotis verbascifolia2
Arctotis leiocarpa
Arctotis auriculata
Arctotis erosa1
Arctotis bellidifolia2
87
Arctotis hirsuta
Arctotis laevis
Arctotis erosa3
Arctotis campanulata puberula
1
Arctotis erosa2
Arctotis spA
Arctotis erosa4
Arctotis spB
95
Arctotis flaccida1
Arctotis spD2
Arctotis flaccida2
Arctotis spD1
1
Arctotis fastuosa 1
Arctotis canescens
Arctotis rotundifolia
Arctotis fastuosa 2
75
Arctotis venusta
Arctotis incisa1
0.88
Arctotis spB
Arctotis semipapposa2
100
Haplocarpha scaposa1
Haplocarpha scaposa1
Haplocarpha scaposa
Haplocarpha scaposa2
Haplocarpha scaposa2
73 1
96
Haplocarpha parvifolia
Haplocarpha nervosa1
0.99
Haplocarpha nervosa1
Haplocarpha nervosa2
100
1
Haplocarpha
nervosa2
85
Haplocarpha parvifolia
‘Landtia’
1
97
Haplocarpha rueppellii1
60
Haplocarpha rueppellii1
1
Haplocarpha rueppellii2
Haplocarpha rueppellii2
1
0.62
Gazania krebsiana
Hirpicium echinus
Dymondia margaretae
Haplocarpha schimperi1
Haplocarpha schimperi2
Arctotis arctotoides1
Arctotis arctotoides2
Arctotis debensis1
87 Arctotis debensis2
Arctotis hispidula
0.98
Arctotis microcephala1
Arctotis microcephala2
Arctotis scapiformis2
56
Arctotis scapiformis1
0.96
59
Cymbonotus lawsonianus2
75
Cymbonotus lawsonianus1
1
1
100
Arctotis perfoliata1
1
Arctotis perfoliata2
Haplocarpha lanata
90
Arctotheca marginata
96
1
Arctotheca prostrata
1
Haplocarpha lyrata
97
Arctotheca calendula
1
Arctotheca forbesiana
Arctotheca populifolia
100
Arctotis sulcocarpa
100
Arctotis dregei1
1
Arctotis dregei2
1
69
Arctotis breviscapa
0.99
Arctotis acaulis2
Arctotis adpressa
Arctotis angustifolia latifolia
Arctotis argentea
Arctotis aspera2
Arctotis aspera1
Arctotis aspera scabra
Arctotis bellidifolia1
Arctotis decurrens2
Arctotis decurrens1
Arctotis incisa2
Arctotis muricata
77
Arctotis reptans
Arctotis revoluta2
0.96
Arctotis revoluta1
Arctotis semipapposa1
Arctotis stoechadifolia2
Arctotis sp2
Arctotis sp4
Arctotis spE
Arctotis acaulis1
Arctotis pinnatifida
Arctotis elongata
Arctotis lanceolata2
Arctotis erosa4
Arctotis venusta
96
Arctotis erosa2
1
Arctotis spC2
Arctotis spC1
Arctotis lanceolata1
Arctotis fastuosa 1
Arctotis hirsuta
Arctotis leiocarpa
Arctotis stoechadifolia1
Arctotis verbascifolia1
98
1
100
1
Dymondia
Berkheya carduoides
Cuspidia cernua
100
1
—
1
100
1
0.1
Fig. 2. Phylograms obtained from Bayesian-inference analyses. (A) cpDNA data set; (B) ITS data set. Support values above the branches are bootstrap percentages, and below
the branches are Bayesian posterior probabilities. For terminal clades, support values are only given for nodes receiving bootstrap percentages > 75% and Bayesian posterior
probabilities > 0.95.
8
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Dymondia margaretae
Haplocarpha lanata
Haplocarpha lyrata
62
Arctotheca forbesiana
1
100
95
Arctotheca calendula
—
1
1
Arctotheca populifolia
0.93
Arctotheca marginata
55
0.99
Arctotheca prostrata
Arctotis
sulcocarpa
100
100
Arctotis dregei1
1
1
Arctotis dregei2
Cymbonotus maidenii
54
Arctotis scapiformis2
0.97
Arctotis
scapiformis1
81
Arctotis microcephala1
0.97
Arctotis microcephala2
61
Haplocarpha schimperi1
0.95 76
95
1
Haplocarpha schimperi2
0.88
Arctotis
hispidula
80
1 79
Arctotis arctotoides1
0.97
Arctotis arctotoides2
65
Arctotis debensis1
0.99
Arctotis debensis2
Cymbonotus lawsonianus2
100
1
Cymbonotus lawsonianus1
Arctotis breviscapa
Arctotis perfoliata1
100
1
Arctotis perfoliata2
Arctotis campanulata puberula
Arctotis lanceolata2
Arctotis semipapposa1
Arctotis decurrens2
Arctotis decurrens1
Arctotis spE
Arctotis acaulis2
56
0.94
Arctotis adpressa
Arctotis argentea
Arctotis aspera scabra
Arctotis sp2
61
Arctotis spA
0.94
Arctotis acaulis1
—
0.92
Arctotis lanceolata1
90
0.98
Arctotis verbascifolia1
Arctotis
aspera1
85
Arctotis
incisa2
1
Arctotis stoechadifolia2
Arctotis angustifolia latifolia
Arctotis aspera2
95
Arctotis bellidifolia1
1
Arctotis bellidifolia2
Arctotis reptans
Arctotis revoluta1
—
Arctotis rotundifolia
0.62
Arctotis pinnatifida
—
Arctotis elongata
0.68
77
Arctotis graminea
1
Arctotis verbascifolia2
—
Arctotis
incisa1
0.97
Arctotis semipapposa2
100
Arctotis laevis
67
1
0.99
Arctotis revoluta2
Arctotis auriculata
75
Arctotis
spD2
—
1
Arctotis spD1
0.91
—
Arctotis canescens
0.96
96
Arctotis spC2
1
Arctotis spC1
Arctotis
muricata
—
Arctotis stoechadifolia1
0.82 —
0.83
Arctotis sp4
Arctotis spB
91
Arctotis venusta
1 —
Arctotis fastuosa 2
0.81 83
Arctotis erosa2
1
Arctotis leiocarpa
—
0.79
Arctotis fastuosa 1
—
Arctotis hirsuta
0.72
Arctotis erosa3
— 58
Arctotis erosa1
0.76 1
61
Arctotis erosa4
1
—
Arctotis flaccida1
0.54
Arctotis flaccida2
Haplocarpha scaposa1
Haplocarpha scaposa2
95
Haplocarpha nervosa1
1
Haplocarpha nervosa2
Haplocarpha parvifolia
74
Haplocarpha rueppellii1
95
0.97
1
Haplocarpha rueppellii2
93
1
94
1
86
1
—
0.76
94
1
92
1
57
0.98
—
0.84
100
1
100
1
100
1
100
1
0.01
Dymondia
‘core’
Haplocarpha
Arctotheca
Arctotis sect.
Anomalae
‘Cymbonotus’
perennial
‘core
Arctotis’
annual
‘core
Arctotis’
Haplocarpha
scaposa
‘Landtia’
Fig. 3. Phylogram obtained from Bayesian-inference analysis of the total evidence. Support values are as in Fig. 2.
carpha’ clade (comprising all Arctotheca species, H. lanata and H.
lyrata). Relationships amongst these four clades varied depending
on the data set and method of analysis.
The major clades in the cpDNA phylogeny were strongly supported by both phylogenetic methods, but within-clade relationships were poorly resolved (Fig. 2A). The interior nodes linking
Dymondia margaretae, ‘core Arctotis’, an Arctotis sect. Anomalae–
Arctotheca–‘core Haplocarpha’ clade, and a ‘Cymbonotus’–Arctotis
perfoliata clade were weakly to moderately supported by parsimony bootstrap (BS = 59–77%), but well supported in the BI analysis (PP = 0.96–1). Arctotis perfoliata was placed sister to the
‘Cymbonotus’ clade with weak support (BS = 56%, PP = 0.96). Arcto-
9
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Table 3
Summary of statistics from parsimony analyses of each DNA fragment and the combined datasets
Characteristic
psbA-trnH
trnT-trnF
Combined cpDNA
ITS
Total evidence
Length range (nucleotides)
Aligned length (nucleotides)
Number of indels recoded (number that are parsimony informative)
Number of variable sites
Number of parsimony informative sites
Number of most parsimonious trees
Tree length (steps)
CI
RI
330–495
563
20 (12)
83 (14.7%)
48 (8.5%)
—
—
—
—
1348–1463
1593
31 (20)
195 (12.2%)
94 (5.9%)
—
—
—
—
1716–1950
2155
51 (32)
278 (12.9%)
142 (6.6%)
100 000
268
0.743
0.928
632–653
688
6 (4)
275 (40.0%)
206 (29.9%)
92 524
611
0.622
0.865
–
2843
57 (36)
553 (19.5%)
348 (12.2%)
94 496
872
0.606
0.872
The statistics are from parsimony analyses in which indels were recoded. Recoded indels are excluded from the data for sequence length, variability and parsimony
informativeness.
tis breviscapa was placed sister to the ‘core Arctotis’ clade with
moderate support (BS = 77%, PP = 0.96). Although the rest of the
‘core Arctotis’ clade was strongly supported (BS = 96%, PP = 1), relationships within the clade were poorly resolved. Arctotis sect.
Anomalae, Arctotheca and ‘core Haplocarpha’ formed a strongly
supported clade (BS = 97%, PP = 1). A ‘core Haplocarpha’ clade
was not retrieved. Haplocarpha lanata was sister to the rest of the
Arctotheca–‘core Haplocarpha’ clade, and H. lyrata was nested
within the Arctotheca clade but with little support (BS = 58%,
PP = 0.78).
The ITS phylogeny (Fig. 2B) differed from the cpDNA phylogeny
in the following respects: the internal branches linking the Arctotheca–‘core Haplocarpha’, Arctotis sect. Anomalae, ‘Cymbonotus’ and
‘core Arctotis’ clades were poorly supported or unresolved, but
relationships within the major clades were somewhat better resolved. ‘Core Haplocarpha’ and Arctotheca formed strongly supported sister clades. Cymbonotus lawsonianus was sister to the
rest of the ‘Cymbonotus’ clade (BS = 72%, PP = 1). The two Cymbonotus species formed a basal grade to the southern and eastern African species in this clade. Arctotis perfoliata was placed close to the
base of the ‘core Arctotis’ clade but with poor support (BS = 66%,
PP = 0.67). The ‘core Arctotis’ clade was differentiated into highly
supported annual and perennial lineages (BS = 97%, PP = 1, and
BS = 95%, PP = 1, respectively).
Exclusion of four ambiguously aligned regions reduced the ITS
data set to 632 bp, of which 238 bp (37.7%) were variable and
174 bp (27.5%) were parsimony informative. Parsimony analysis
yielded 1526 MPTs of 487 steps (CI 0.616, RI 0.867). The strict consensus tree topology differed to that obtained from the complete
ITS data analysis in that the Arctotis sect. Anomalae clade was
placed sister to the ‘core Haplocarpha’ clade but with no BS support
(BS < 50% from 100 replicates).
The ILD test indicated significant conflict existed between the
cpDNA and ITS data sets (p = 0.001342). Visual comparison of tree
topologies revealed incongruence in the placement of A. perfoliata
and H. lyrata between the cpDNA and ITS trees (Fig. 2A and B),
but the placement of both species was weakly supported by one
data set. The placement of A. perfoliata in the cpDNA phylogeny
was strongly supported (BS = 100%, PP = 1), but only weakly supported in the ITS tree (BS = 66%, PP = 0.67). In contrast, the placement of H. lyrata was strongly supported (BS = 92%, PP = 1) by
the ITS data, but its placement within the Arctotheca clade from
the cpDNA data received conflicting support (BS = 58%, PP = 1).
Placement of the ‘Cymbonotus’ clade also differed in the cpDNA
and ITS trees, but overall support was low (BS = 59%, PP = 1 in
the cpDNA phlyogeny; BS < 50%, PP = 0.53 in the ITS phylogeny)
and the data sets were thus combined.
Both analytical methods yielded phylogenies with improved
resolution from the total evidence data set (Fig. 3) than was obtained in the analyses of the separate data sets (Fig. 2A and B).
The basal nodes were well supported (BS P 86%, PP = 1). The place-
ment of the ‘Cymbonotus’ clade sister to the Arctotheca–‘core
Haplocarpha’–Arctotis sect. Anomalae clade, and the ‘core Arctotis’
node were poorly supported, but the other internal nodes were
well supported. The placement of A. perfoliata within the ‘core
Arctotis’ clade received little support (BS < 50%, PP = 0.84). Highly
supported annual and perennial clades were resolved within ‘core
Arctotis’ (BS = 99%, PP = 1).
Exclusion of A. perfoliata and H. lyrata from total-evidence analyses had no impact on the remaining tree topology, but support for
the ‘core Arctotis’ clade improved considerably (BS = 83% from 200
bootstrap replicates), whereas the ‘Cymbonotus’–Arctotis sect.
Anomalae–Arctotheca–‘core Haplocarpha’ clade remained weakly
supported (BS = 56%).
3.2. Molecular dating
Divergence dates estimated with the ‘slow’ and ‘average’ mutation rates resulted in considerably older mean divergence dates
with wider 95% HPD ranges than dates estimated with the ‘fast’
mutation rate (Table 4). Based on the mean estimated dates, the
Landtia clade (node 1) was estimated to have diverged some time
during the early Oligocene to late Miocene epochs, but the crown
of extant species (node 15) is of much more recent origin (Fig. 4
and Table 4). The three major extant clades—‘core Arctotis’, ‘Cymbonotus’, and Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca
(nodes 4, 5, and 6, respectively)—were indicated to have diverged
in a rapid radiation during the mid or late Miocene. Cymbonotus
lawsonianus (node 9) was indicated to have diverged within the
‘Cymbonotus’ clade during the late Miocene to late Pliocene. Radiation of the Arctotis arctotoides species complex (node 10) was dated to the late Pliocene at the earliest. Within this complex, Arctotis
microcephala and H. schimperi (node 11) were estimated to have diverged during the late Pliocene or Pleistocene.
3.3. Biogeography and dispersal–vicariance analysis
The two basalmost lineages in Arctotidinae have a wide geographic distribution (Fig. 4). The ‘Landtia’ clade grows largely at
higher altitudes, and is distributed from the CFR (H. parvifolia) to
southeastern Africa (H. nervosa) and afromontane regions of the
East African mountain chain (H. nervosa and H. rueppellii). Haplocarpha scaposa is widely distributed in submontane and montane
southeastern Africa and tropical eastern Africa. DIVA analysis indicated that Arctotidinae might have an afromontane origin, but the
possible ancestral area of the widespread H. scaposa was equivocal.
Furthermore, the accurate assessment of the ancestral area requires a resolved phylogeny of the sister group and use of a more
distant outgroup. The ancestral area of all other extant Arctotidinae
was indicated to be in the fynbos (i.e., in the CFR). The Arctotheca–
‘core Haplocarpha’ clade was indicated to have originated in fynbos, consistent with the centre of diversity in the CFR, with subse-
10
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Table 4
Estimated divergence dates for nodes of interest from BEAST analysis using ITS mutation rates published for other Asteraceae taxa
Node
Slow
ESS
Average
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
32.6 ± 0.17 (21.7–44.0)
18.6 ± 0.14 (13.1–24.7)
17.3 ± 0.11 (12.7–22.5)
16.1 ± 0.1 (11.7–20.7)
16.1 ± 0.1 (11.8–20.9)
13.8 ± 0.1 (9.4–18.8)
9.4 ± 0.07 (5.7–13.6)
3.9 ± 0.03 (1.4–6.9)
9.6 ± 0.07 (5.5–14.0)
3.7 ± 0.03 (1.9–6.0)
2.3 ± 0.02 (0.7–4.4)
10.6 ± 0.06 (7.4–13.9)
9.4 ± 0.05 (6.5–12.5)
6.8 ± 0.04 (4.6–9.1)
9.7 ± 0.07 (5.4–14.1)
1128.463
479.403
565.099
545.097
556.745
658.419
947.15
3391.419
1051.392
1120.797
2013.626
801.202
918.283
1055.801
1125.109
27.2 ± 0.11
11.8 ± 0.06
10.7 ± 0.05
9.3 ± 0.04
9.3 ± 0.04
8.0 ± 0.05
5.3 ± 0.03
2.0 ± 0.01
5.4 ± 0.03
1.7 ± 0.01
1.1 ± 0.01
6.0 ± 0.03
5.0 ± 0.02
3.4 ± 0.01
5.6 ± 0.03
(17.6–38.6)
(8.2–15.9)
(7.5–14.0)
(6.9–12.2)
(6.7–12.2)
(5.6–10.6)
(3.1–7.5)
(0.8–3.5)
(3.0–8.0)
(0.9–2.7)
(0.3–2.2)
(4.2–8.1)
(3.5–6.9)
(2.3–4.6)
(3.5–8.1)
ESS
Fast
2894.552
1157.966
1087.701
1035.193
1077.755
823.279
1827.21
4844.536
1957.358
1258.726
2001.685
1102.589
1533.579
1800.349
1423.365
10.2 ± 0.05
5.9 ± 0.03
5.5 ± 0.03
5.1 ± 0.03
5.1 ± 0.03
4.4 ± 0.03
3.0 ± 0.02
1.2 ± 0.01
3.0 ± 0.02
1.2 ± 0.01
0.73 ± 0.01
3.4 ± 0.02
3.0 ± 0.02
2.2 ± 0.01
3.0 ± 0.02
ESS
(6.8 13.9)
(4.2–7.8)
(4.0–7.1)
(3.8–6.7)
(3.6–6.6)
(3.0–6.0)
(1.8–4.3)
(0.44–2.2)
(1.7–4.4)
(0.59–1.9)
(0.20–1.4)
(2.4–4.4)
(2.0–4.0)
(1.5–2.9)
(1.7–4.4)
1128.886
1027.545
966.825
836.345
859.889
749.412
1453.303
2792.901
1478.561
1229.467
1606.189
987.303
1040.264
1043.336
974.539
Results are from two independent analyses combined. For the node numbers see Fig. 3. Values (in My) are the mean divergence date ± SD and the 95% highest posterior
density range. Slow = 2.51 10 9 s s 1 y 1 (Eupatorium; Schmidt and Schilling, 2000); average = mean of rates for four Asteraceae taxa listed by Kay et al. (2006); fast = 7.83 10 9 s s 1 y 1 (Robinsonia; Sang et al. 1995). Effective sample size (ESS) = the number of effectively independent samples from the posterior distribution.
quent extension into Namaqualand (A. calendula and A. marginata),
the Nama-Karoo and Succulent Karoo (A. calendula) and the Albany
hotspot in southeastern South Africa (A. populifolia, A. prostrata and
H. lyrata). The Arctotis sect. Anomalae lineage is centred in the CFR
and Succulent Karoo, and extends into the Nama-Karoo (an undescribed species) and the Albany region (A. dregei). The ‘Cymbonotus’ clade was indicated to have migrated from fynbos to the
Albany hotspot, savanna and afromontane southeastern Africa
(Arctotis sect. Austro-orientales p.p.) and to have dispersed to Australia (C. lawsonianus and C. maidenii). An origin in the fynbos is
implicated for the ‘core Arctotis’ clade, with subsequent independent radiations in the Succulent Karoo in both the perennial and
annual clades.
4. Discussion
The eight major lineages retrieved in the present investigation
are congruent with those obtained in a previous study (McKenzie
et al., 2006a), which was based on a smaller sample of taxa. Our results confirm the polyphyly of Arctotis and Haplocarpha as the genera are presently circumscribed. Haplocarpha species are
segregated between four clades and Arctotis species are placed in
three well-resolved lineages. A molecular phylogenetic study by
Funk et al. (2007) included sequence data for two species not included in the present study, namely Haplocarpha oocephala (the
sample was incorrectly named H. lanata) and Cymbonotus preissianus, as well as both cpDNA (ndhF and trnL-F) and ITS data for C.
maidenii. Haplocarpha oocephala was placed on a monotypic lineage diverging between the ‘Landtia’ and H. scaposa lineages, and
C. preissianus and C. maidenii were in the same clade as C. lawsonianus. The inclusion of a comprehensive sample of taxa in the present study has clarified the phylogenetic placement of several
species whose placement was previously considered equivocal
(McKenzie et al., 2006a). Arctotheca calendula was shown to belong
unequivocally in the Arctotheca clade. In all analyses Arctotis breviscapa was placed sister to the rest of the ‘core Arctotis’ clade. The
sample of H. lyrata was indicated to possess a chloroplast haplotype shared with Arctotheca, whereas the ITS data resolved a
monophyletic ‘core Haplocarpha’. The conflicting placement of
two accessions of Arctotis perfoliata sister to the ‘Cymbonotus’
clade in the cpDNA phylogeny and near-basal within the ‘core
Arctotis’ clade from ITS data confirms the findings of McKenzie
et al. (2006a).
Using the criteria of Mason-Gamer and Kellogg (1996) and Alfaro et al. (2003), ‘hard’ incongruence (the conflicting placement
of a taxon or clade is strongly supported in phylogenies derived
from different data sets) was not apparent for any taxon or clade
in the present study. Although the placement of Arctotis perfoliata
and Haplocarpha lyrata conflicted between the cpDNA and ITS phylogenies, support was high for one data set only (cpDNA for A. perfoliata, ITS for H. lyrata). Exclusion of both species from analyses
did not affect the resulting tree topology (data not presented), indicating the overall topologies reflect strong phylogenetic signal in
the data. Both A. perfoliata samples possessed a unique 93 bp deletion in the psbA-trnH intergenic spacer and an additional seven unique 1 bp substitutions (three each in the trnT-trnF and ITS1
spacers and one in the psbA-trnH spacer). In morphology A. perfoliata possesses a combination of distinctive traits of the ‘Cymbonotus’ and ‘core Arctotis’ clades (McKenzie et al., 2005; R.J.
McKenzie, unpublished data), which might reflect an ancestral
hybridization event occurring during the evolution of A. perfoliata.
However, incongruence between phylogenetic hypotheses might
also reflect lineage sorting of ancestral polymorphisms, paralogy,
lateral gene transfer, or erroneous phylogenetic reconstruction
(Sang and Zhong, 2000). Sampling of additional accessions and
DNA regions is needed to further resolve the phylogenetic relationships and evolutionary histories of A. perfoliata and H. lyrata and to
establish what factors might account for the phylogenetic
uncertainty.
The ITS data resolved highly supported annual and perennial
clades within the ‘core Arctotis’ clade, in contrast to the cpDNA
data, but this might reflect the fewer informative cpDNA characters
or a need for additional cpDNA data, rather than hard incongruence
in the phylogenetic signal. The annual and perennial Arctotis clades
were well-supported in the total-evidence analyses, indicating any
conflicting signal in the cpDNA data set was weak.
4.1. Phylogenetic relationships and taxonomic implications
Correspondence between relationships suggested by morphological and molecular data in Arctotidinae has been discussed previously (McKenzie et al., 2005, 2006a). Therefore the following
discussion is focused on findings from the present study that build
on those of previous studies.
The present study confirms the finding of McKenzie et al.
(2006a) that H. nervosa and H. rueppellii are closely related and belong to the earliest-diverging extant clade (‘Landtia’), and provides
evidence that H. parvifolia, which was not included in the previous
study, is a member of the same clade. This supports Beauverd’s
(1915) observation that H. parvifolia possesses similarities in mor-
11
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
FYN
Dymondia margaretae
Haplocarpha lanata
FYN
ALB
Haplocarpha lyrata
FYN
Arctotheca forbesiana
Arctotheca calendula FYN, KAR, SK
Arctotheca populifolia
ALB, FYN
Arctotheca marginata
SK
ALB, FYN
Arctotheca prostrata
Arctotis sulcocarpa
SK
Arctotis dregei
ALB, FYN, KAR
Cymbonotus maidenii
AUS
FYN
7
8
FYN
6
FYN
FYN
3
5
Arctotis scapiformis
ALB
11
10
Arctotis arctotoides
Arctotis hispidula
9
FYN
4
SK
FYN
12
SK
FYN
AFR
FYN
ALB
SAV
13
SK
FYN
2
SK
FYN
14
AFR
1
SK
AFR
15
AFR
AFR
Haplocarpha schimperi
Arctotis microcephala AFR, ALB, SAV
ALB
ALB
ALB
Arctotis debensis
AUS
Cymbonotus lawsonianus
Arctotis breviscapa
FYN
FYN
Arctotis perfoliata
Arctotis campanulata puberula SK
Arctotis lanceolata
FYN
Arctotis semipapposa
FYN
Arctotis decurrens
SK
SK
Arctotis spE
FYN
Arctotis acaulis
FYN
Arctotis adpressa
FYN
Arctotis aspera scabra
FYN
Arctotis sp2
SK
Arctotis spA
Arctotis verbascifolia
FYN
FYN
Arctotis angustifolia latifolia
FYN
Arctotis aspera
FYN
Arctotis bellidifolia
Arctotis reptans
FYN
Arctotis revoluta
FYN, SK
FYN
Arctotis rotundifolia
ALB, FYN
Arctotis pinnatifida
FYN
Arctotis elongata
FYN
Arctotis graminea
FYN
Arctotis incisa
FYN
Arctotis laevis
SK
Arctotis auriculata
SK
Arctotis spD
Arctotis canescens
SK
SK
Arctotis spC
FYN
Arctotis muricata
FYN
Arctotis stoechadifolia
FYN
Arctotis sp4
SK
Arctotis spB
KAR, SAV, SK
Arctotis venusta
SK
Arctotis fastuosa 2
SK
Arctotis fastuosa 1
KAR, SK
Arctotis leiocarpa
SK
Arctotis erosa
FYN
Arctotis hirsuta
SK
Arctotis flaccida
Haplocarpha scaposa AFR, ALB, SAV
Haplocarpha nervosa AFR, ALB, FYN
AFR
Haplocarpha rueppellii
FYN
Haplocarpha parvifolia
Fig. 4. Ancestral areas in Arctotidinae indicated by DIVA analysis, mapped on a simplified cladogram from Bayesian-inference analysis of the total evidence data set. Key to
area codes: AFR, afromontane biome; ALB, Albany centre of endemism; AUS, Australia; FYN, fynbos biome; KAR, Nama-Karoo biome, SAV, savanna biome; SK, Succulent Karoo
biome. These regions in black boxes represent extant distributions, those in grey represent ancestral areas. Numbers in circles and squares (calibration points) represent node
numbers associated with node ages listed in Table 4.
phology to H. subg. Landtia, and specifically H. rueppellii. Haplocarpha nervosa, H. rueppellii and species now placed in their synonymy
have been segregated as Landtia in the past (e.g. Lessing, 1831,
1832; Harvey, 1865; Bentham, 1873; Phillips, 1951). The morphological characters previously used to diagnose Landtia have been
deemed untenable (e.g. Hedberg, 1957) and, especially focusing
12
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
on fruit morphology (McKenzie et al., 2005), the placement of H.
parvifolia in the same clade further complicates morphological
characterization of the clade. Haplocarpha rueppellii and H. scaposa
possess a cypsela anatomy distinct from the other Arctotidinae
sampled by Reese (1989), which thus might offer a possible synapomorphy for the ‘Landtia’ clade. Given the similarity in external
vegetative morphology in the ‘Landtia’ clade, stem or leaf anatomy
might also yield informative characters. Similarities in the morphology of H. parvifolia and H. oocephala have been noted previously (McKenzie et al., 2005, 2006a), but Funk et al. (2007) found
that a sample of H. oocephala (misidentified as H. lanata in their paper) was placed on a monotypic lineage diverging after the ‘Landtia’ clade.
Placement of H. scaposa and Dymondia margaretae on monotypic branches in all analyses confirmed previous findings (McKenzie et al., 2006a). Segregation of H. scaposa into a monotypic genus
is strongly supported by the molecular data, but the species shares
certain distinctive morphological features (e.g. papillose filaments,
pappus scales with long acute apices) with species in the ‘core
Haplocarpha’ clade.
Sampling of all five Arctotheca species in the present study resolved a well-supported Arctotheca clade and confirmed a sister
relationship with ‘core Haplocarpha’ (McKenzie et al., 2006a), but
with the inclusion of H. lyrata in the Arctotheca clade in cpDNA-derived phylogenies. In the total-evidence analyses, ‘core Haplocarpha’ formed a basal grade sister to Arctotheca. As discussed
earlier, sampling of additional DNA regions and multiple accessions is needed to resolve the evolutionary history of H. lyrata.
The two species of Arctotis sect. Anomalae sampled (A. dregei
and A. sulcocarpa) formed a highly supported clade distinct from
‘core Arctotis’ in all analyses, in agreement with previous results
(McKenzie et al., 2006a). Lewin (1922) distinguished sect. Anomalae from other sections of Arctotis by the possession of neuter ray
florets (female in all other Arctotis sections), a trait also found in
Arctotheca and H. parvifolia. The species of A. sect. Anomalae share
some distinctive characteristics with ‘core Arctotis’, such as possession of cypselae with well-developed abaxial wings (McKenzie
et al., 2005), but other features (e.g. possession of floral scent,
papillose filaments) supports a sister relationship with ‘core Haplocarpha’–Arctotheca.
Holland and Funk (2006) broadened the concept of Cymbonotus
to encompass the three Australian-endemic species of Arctotidinae
by formally transferring Arctotis maidenii. Our results demonstrate
that the ‘Cymbonotus’ clade has an Afro-Australian distribution
and also includes the East African Haplocarpha schimperi and
southern African species currently placed in Arctotis sect. Austroorientales (McKenzie et al., 2006a; this study). Our data resolved
a well-supported ‘Cymbonotus’ clade, but whether it is sister to
‘core Arctotis’ or the Arctotis sect. Anomalae —‘core Haplocarpha’—
Arctotheca lineage remains uncertain, seemingly reflecting a rapid
divergence of the three lineages. Cymbonotus maidenii was sister
to C. lawsonianus and C. preissianus in the study of Funk et al.
(2007), which utilized both cpDNA and ITS data, and was not
monophyletic with C. lawsonianus in the present study, a finding
consistent with disparities in the morphology of these species
(McKenzie et al., 2005; R.J. McKenzie, unpublished data). A relationship between C. maidenii and Arctotis perfoliata, as indicated
by cypsela morphology (McKenzie et al., 2005), is not supported
by the molecular data.
A sister relationship between Haplocarpha schimperi and Arctotis
microcephala within the ‘Cymbonotus clade’ is indicated. Despite
the limited divergence in non-coding DNA sequences between
the two species, H. schimperi has undergone greater evolution in
phenotypic traits, especially in cypsela morphology (McKenzie
et al., 2005), and possesses at least four autapomorphic character
states absent in other members of the ‘Cymbonotus’ clade (geotro-
pic capitula, woolly tomentum on the ray limb abaxial surface, a
coronate pappus, and cypsela abaxial ribs not developed into
wings).
The present study, in which 45 of the approximately 60 species
of Arctotis were sampled, reinforces the conclusion of McKenzie
et al. (2006a) that the sections Anomalae and Austro-orientales
should be excluded from Arctotis in order to render the genus
monophyletic. Arctotis sect. Austro-orientales (excluding A. erosa
and A. perfoliata) has a close phylogenetic relationship with Cymbonotus, rather than with ‘core Arctotis’. Arctotis erosa, an annual
species placed in sect. Austro-orientales by Lewin’s (1922), belongs
in the ‘annual Arctotis’ clade in ‘core Arctotis’. As discussed earlier,
further investigations are needed to resolve the phylogenetic affinities of Arctotis perfoliata. In all analyses Arctotis breviscapa, an annual species endemic to sandveld in the southwestern Cape, was
placed sister to the rest of the ‘core Arctotis’ clade and is indicated
to be an early divergence from the ‘core Arctotis’ lineage. This
agrees with Lewin’s (1922) hypothesis of evolutionary relationships among his sections in Arctotis. Arctotis breviscapa is anomalous in possessing cypselae with two well-developed adaxial
wings in addition to three abaxial wings (McKenzie et al., 2005),
but is otherwise similar in morphology to ‘core Arctotis’. The resolution of annual and perennial ‘core Arctotis’ clades by the ITS data
conflicts with Lewin’s (1922) infrageneric classification of Arctotis.
These clades were not retrieved in the cpDNA phylogenies, which
might reflect ancestral polymorphisms, hybridization and introgression, or insufficient informative sites in the cpDNA data.
4.2. Biogeography and timing of diversification of Arctotidinae
The estimated divergence dates presented in this paper are a
first attempt at dating Arctotidinae diversification and must be
interpreted with caution. Resolution of the timing of divergence
events preferentially requires either a complete and reliably dated
fossil record or reliable molecular mutation rates, neither of which
are presently available for Arctotideae. Owing to the poor fossil record, obtaining reliable calibration points to date phylogenetic
reconstructions within Asteraceae is a widespread problem. Rather
than using an indirect calibration point derived from a higher-level
molecular-dating analysis, which may potentially compound the
error in estimated divergence dates (Graur and Martin, 2004), we
utilized ITS mutation rates, representing the fastest and slowest
extremes and an average rate, published for other Asteraceae taxa.
It is acknowledged that the divergence dates estimate the evolution of the ITS region and not necessarily speciation or cladogenetic
events. Natural selection is likely to act differentially on mutations
(and thus evolution rates will vary) between loci and lineages. The
strength of concerted evolution (Hillis et al., 1991) on the ITS region may vary between lineages and in lineages of ancestral hybrid
origin. In addition, the similarity of the rates used to the overall ITS
mutation rate in Arctotidinae is presently unknown.
Nevertheless, the results obtained permit the formulation of
hypotheses regarding the biogeography and timing of diversification of Arctotidinae. Bearing in mind the likely Calyceraceae–Asteraceae divergence within the last 60 Mya based on the oldest fossil
pollen evidence (Zavada and de Villiers, 2000), and that Arctotidinae comprise primarily annuals and fast-growing perennials that
may flower in their first year after germination, divergence dates
estimated with the ‘fast’ ITS mutation rate are considered the most
plausible of the three mutation rates used in this study. Thus the
following discussion pertains to the dates estimated with a ‘fast’
mutation rate.
Furthermore, the dating of Arctotidinae diversification with a
‘fast’ ITS mutation rate is compatible with the results of a recent
molecular-dating study across all major clades of Asteraceae by
Kim et al. (2005). Their study focused on the basal nodes and in-
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
cluded only one representative of Arctotideae (Gazania krebsiana),
which was indicated to have diverged within the Liabeae–Arctotideae–Vernonieae clade approximately 25 Mya. However, the estimated divergence dates reported by Kim et al. (2005) may be
slight underestimates, as the oldest microfossil evidence (Zavada
and de Villiers, 2000) currently available for Asteraceae was not
considered. Nevertheless, the divergence and early radiation in
Arctotidinae may have followed expansion of the Antarctic ice
cap, which covered most of that continent by about 17 Mya and
had far-reaching global climatic impacts. In southwestern Africa
a variety of new arid and semi-arid environments arose at this time
(Pickford, 2004).
The ancestral area of Arctotidinae was indicated to be in our afromontane area. The present distribution of the two earliest-diverging extant lineages, the ‘Landtia’ clade and Haplocarpha oocephala
(the latter based on Funk et al., 2007), comprises a chain of disjunct
populations along the southern and eastern African mountains
extending from the northern Cederberg area in the Western Cape
province, South Africa to the Ethiopian highlands (Hedberg,
1957; Hilliard, 1977; Pope, 1992; Beyers, 2000; Mesfin Tadesse,
2004). The next lineage to diverge (Haplocarpha scaposa) is widespread in montane parts of southeastern Africa and the Zambezi
River catchment area (Hilliard, 1977; Pope, 1992). These distributions roughly coincide with what has been termed the Afrotemperate Phytogeographical Region, which comprises the Cape Floral
Region, the greater Drakensberg mountains, and the afromontane
Centre (Linder, 1990; Galley et al., 2007). The more recent derivation of Arctotidinae lineages in presently semi-arid and arid areas
is consistent with trends in rainfall regimes in southwestern Africa
since the Oligocene (e.g. see Linder, 2003) and phylogenetic analyses of a diverse range of plant lineages endemic to the winter-rainfall region (Verboom et al., in press).
The estimated divergence of the ‘Landtia’ clade in the late Miocene follows the change to a more humid climate in the early midMiocene about 16 Mya (Dingle and Hendey, 1984). Mesic conditions persisted in southern Africa for about 7–8 Mya until the late
Miocene (Partridge, 1997). The extant taxa in the ‘Landtia’ clade, as
well as H. oocephala and H. scaposa, prefer mesic or perennially/
seasonally wet habitats, including bogs, streambanks and seepages. Dymondia margaretae, the next-diverging extant lineage
within Arctotidinae, is a component of the seasonally inundated
vlei vegetation that is a significant component of the Agulhas Plain
flora (Cowling et al., 1988). Thus a preference for mesic or hydric
habitats might be plesiomorphic in Arctotidinae and reflect an
ancestral niche. This pattern exhibits similarities to the radiations
of Ehrharta (Verboom et al., 2003) and Thamnocortus (Linder and
Hardy, 2005), in that the earliest-diverging extant species are
found in mesic habitats and that xerophytic adaptations or invasion of semi-arid regions are indicated to be derived.
Most of the extant diversity in Arctotidinae was resolved into
three well-supported lineages—‘core Arctotis’, ‘Cymbonotus’ and
Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca clades—that
were estimated to have diverged during a rapid radiation centred
in southern Africa during the late Miocene or around the Miocene–Pliocene boundary. The present-day centre of Arctotidinae
diversity is in the winter-rainfall region, but almost all of the species in this region were indicated to be derived. Based on our dating estimates, the extensive radiation in the winter-rainfall region
coincided with the trend towards increasing rainfall seasonality
and intensified aridification following increased glaciation in Antarctica from 14 Mya (Zachos et al., 2001), associated with strengthening of the South Atlantic high-pressure cell (Linder, 2005). In
addition, the development of the Benguela Current about 11–
14 Mya and summer drought in southwestern Africa are closely
associated (Linder, 2003). The lower species diversity in the summer-rainfall region might reflect either a shorter time period avail-
13
able for diversification, less environmental heterogeneity
compared to the winter-rainfall region, or the extant diversity
might be relictual and that many species radiated but have gone
extinct.
Our estimated dates for the radiation of Arctotidinae in the winter-rainfall region are consistent with estimates obtained for
numerous other plant clades endemic or near-endemic to the region (see Linder, 2003, 2005; Verboom et al., in press). However,
molecular-dating analyses indicate that the onset of radiations in
the winter-rainfall region in different families has not been coordinated but has occurred over the course of at least the last 20 Mya
(e.g. Linder et al., 2003, 2005; Schrire et al., 2003; Klak et al., 2003;
Bakker et al., 2004; Mummenhoff et al., 2005; Verboom et al. in
press).
The ‘core Arctotis’ clade, in which most of the extant diversity
occurs in the Fynbos and Succulent Karoo biomes, was estimated
to have diverged around the Miocene–Pliocene boundary. This is
consistent with the appearance of fynbos vegetation, and a change
to a cooler and drier climate in southwestern Africa, around the
Miocene–Pliocene boundary (Linder, 2003). Divergence of the annual and perennial clades within ‘core Arctotis’ was estimated to
have occurred during the late Pliocene. This followed closure of
the Panamanian seaway and abrupt strengthening of the Benguela
Current about 3.2 Mya, probably related to increased Antarctic glaciation, which further enhanced the seasonal aridity in the northwestern Cape during the past 3 Mya (Marlow et al., 2000). A
hypothesized Pliocene origin of the annual life history in Arctotis
is coincident with estimates for development of annualness in
southern African Nemesia (Datson et al., 2008).
The ‘annual Arctotis’ clade is centreed in the semi-arid Succulent Karoo. The ‘perennial Arctotis’ clade is centred in the Fynbos
with outlying species occurring in Namaqualand and the Albany
hotspot, and DIVA resolved the ancestral area as being either or
both the Fynbos or Succulent Karoo. It is evident that migration
in both directions between the Fynbos and Succulent Karoo has occurred in both clades. This might reflect range expansion, contraction and refugial phases during Pleistocene climatic oscillations,
which Midgley et al. (2001, 2005) suggested have promoted vicariance and allopatric speciation in the contiguous Fynbos and Succulent Karoo biomes. The Pleistocene climate is characterized by
alternating cycles of short interglacial periods (10,000–20,000
years) and longer glacial periods (about 100,000 years; Petit
et al., 1999). Expansion and contraction of the winter-rainfall zone
during glacial and interglacial periods, respectively, is indicated
(Chase and Meadows, 2007). During glacial cycles, much of the
western coast of southern Africa and its adjacent interior probably
experienced an increase in rainfall concentrated in the winter
months (Tankard and Rogers, 1978). This would have facilitated
expansion of the mesic and fire-adapted CFR taxa. Conversely,
expansion of the semi-arid and drought-adapted Succulent Karoo
taxa is hypothesized during the drier interglacial periods.
In contrast to ‘core Arctotis’, the extant diversity in the ‘Cymbonotus’ and Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca
clades is more limited, particularly in the winter-rainfall region,
yet they have undergone considerable range expansion in southern
Africa and, in the case of ‘Cymbonotus’, successfully crossed the Indian Ocean to Australia. An ancestral area for the Arctotis sect.
Anomalae–‘core Haplocarpha’–Arctotheca lineage in the CFR was
indicated, implying range expansion followed by speciation in
southeastern Africa (Arctotheca populifolia, A. prostrata and H. lyrata), the Nama-Karoo and Succulent Karoo (Arctotheca calendula, A.
marginata, Arctotis dregei and A. sulcocarpa). Divergence of the winter-annual Arctotis sect. Anomalae clade possibly during the early
Pliocene suggests adaptation to summer-drought conditions.
An origin for the ‘Cymbonotus’ clade in southeastern Africa is
indicated. The dispersal from southern Africa to Australia, possibly
14
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
during the Pliocene, occurred long after the severing of a land connection between Africa and Antarctica during the Cretaceous
(McLoughlin, 2001). The geologic history of the Indian Ocean Basin
since the breakup of Gondwana is well documented. There is no
geological evidence for a land mass that might have acted as a ‘raft’
or continuous bridge between Africa and Australia during the last
approximately 20 My. Dispersal could have been either direct or
via the few small isolated islands, such as on the Kerguelen Plateau,
which could have acted as terrestrial ‘stepping stones’ for some
plants and animals (e.g. Linder et al., 2003; Schwarz et al., 2006).
However, the cypselae of all species in the ‘Cymbonotus’ clade lack
adaptations for long-distance dispersal (see McKenzie et al., 2005),
so the mode of dispersal is difficult to envisage.
Increasing evidence indicates that many other Southern Hemisphere plant biogeographic patterns are best explained by hypotheses incorporating dispersal rather than Gondwanan vicariance
alone (e.g. Baum et al., 1998; Mummenhoff et al., 2004; Cook
and Crisp, 2005; Linder and Barker, 2005; Barker et al., 2007). Bidirectional dispersal of plants is implicated between most of the formerly Gondwanan land masses (Sanmartín and Ronquist, 2004).
Molecular dating estimates indicate that trans-Indian-Ocean dispersal has occurred within the last 5 Mya in a number of diverse
plant groups (e.g. Bakker et al., 1998; Meerow et al., 2003; Linder
and Barker, 2005), and trans-oceanic dispersal from Malesia is
implicated for elements of Indian Ocean island floras (Renvoize,
1979), indicating that the Indian Ocean is not an insuperable barrier to angiosperm dispersal.
The nesting of H. schimperi within the ‘Cymbonotus’ clade
presents another biogeographic puzzle. Haplocarpha schimperi is
distributed from Eritrea to Tanzania at moderate to high altitudes (Mesfin Tadesse, 2004). Its sister species, Arctotis microcephala, is distributed in southern Namibia, southern Botswana,
northern South Africa and Lesotho, and occurs at high altitudes
in the Eastern Cape and Drakensberg mountains, which thus provide a possible afromontane link between the two species.
Northward migration of A. microcephala along the eastern African
mountain corridor during more mesic Quaternary interglacial
periods, followed by extinction in the intervening region in response to increasing aridity, or long-distance migration followed
by speciation, are alternative explanations for the present-day
distribution. Regardless, divergence of A. microcephala and H.
schimperi is indicated to have been recent, possibly during the
Pleistocene.
In conclusion, this study presents a comprehensive molecular
phylogenetic investigation into a group of southern African Asteraceae. Our findings correspond with previous studies on the Cape
flora in indicating that radiation might be linked to climate
changes around the late Miocene and Pliocene. Only one of the
clades identified here (the ‘core Arctotis’ clade) corresponds to a
‘Cape clade’ sensu Linder (2003), and the origins, biogeography,
taxonomy and morphology of the subtribe as a whole are rather
more complex. The phylogenetic reconstructions resolved wellsupported clades and provide the framework for a revised taxonomic classification of Arctotidinae. Future studies focusing on
the basalmost lineages and sister taxa are needed to further elucidate the origins of Arctotidinae.
Acknowledgments
This study was supported by the Rhodes University Joint Research Committee and the National Research Foundation of South
Africa (grant numbers 2042600, 2046932 and 2053645 to N.P.B.,
and a Postdoctoral Fellowship to R.J.McK.). The authors gratefully
thank Coleen Mannheimer, Siro Masinde, Ladislav Mucina, Syd
Ramdhani, James Samuel, Peter Weston and Andrew White for collection of material for DNA extraction; the late Josephine Beyers for
providing material of Arctotheca marginata; the NSW herbarium for
permission to extract DNA from loaned specimens; Sandra Mitchell
and Elizabeth Muller for experimental data; Seranne Howis,
Rhodes University DNA Sequencing Unit for technical assistance;
the late Anne-Marie Brutsch, Ralph Clark, Este Coetzee, Nicolas Devos, Graeme Ellis, Per Ola Karis, Caryl Logie, Roy Lubke, Ladislav
Mucina, Syd Ramdhani, Frida Stångberg, Peter Teske and Gerardo
Zardi for field-work assistance; Per Ola Karis for many interesting
discussions; two anonymous reviewers for comments on the
paper.
References
Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A simulations study
comparing the performance of Bayesian Markov Chain Monte Carlo sampling
and bootstrapping in assessing phylogenetic confidence. Mol. Biol. Evol. 20,
255–266.
Anderson, J.M., Anderson, H.M., Archangelsky, S., Bamford, M., Chandra, S.,
Dettmann, M.E., Hill, R.S., McLoughlin, S., Rösler, O., 1999. Patterns of
Gondwana plant colonisation and diversification. J. Afr. Earth Sci. 28, 145–167.
Bakker, F.T., Hellbrügge, D., Culham, A., Gibby, M., 1998. Phylogenetic relationships
within Pelargonium sect. Peristera (Geraniaceae) inferred from nrDNA and
cpDNA sequence comparisons. Plant Syst. Evol. 211, 273–287.
Bakker, F.T., Culham, A., Hettiarachi, P., Touloumenidou, T., Gibby, M., 2004.
Phylogeny of Pelargonium (Geraniaceae) based on DNA sequences from three
genomes. Taxon 53, 17–28.
Baldwin, B.G., 1992. Phylogenetic utility of the internal transcribed spacers of
nuclear ribosomal DNA in plants: an example from the Compositae. Mol.
Phylogenet. Evol. 1, 3–16.
Barker, N.P., von Senger, I., Howis, S., Zachariades, C., Ripley, B.S., 2005. The use of
the Internal Transcribed Spacer (ITS) sequence data for plant phylogeography:
two examples from the Asteraceae. In: Bakker, F.T., Chatrou, L.W., Gravendeel,
B., Pelser, P.B. (Eds.), Plant Species-level Systematics: New Perspectives on
Pattern and Process. Regnum Veg. 143. Koeltz, Königstein, pp. 217–244.
Barker, N.P., Weston, P.H., Rutschmann, F., Sauquet, H., 2007. Molecular dating of
the ‘‘Gondwanan” plant family Proteaceae is only partially congruent with the
timing of Gondwanan break-up. J. Biogeog. 34, 2012–2027.
Baum, D.A., Small, R.L., Wendel, J.F., 1998. Biogeography and floral evolution of
baobabs (Adansonia, Bombacaceae) as inferred from multiple data sets. Syst.
Biol. 47, 181–207.
Bayer, R.J., Greber, D.G., Bagnall, N.H., 2002. Phylogeny of Australian Gnaphalieae
(Asteraceae) based on chloroplast and nuclear sequences, the trnL intron, trnLtrnF intergenic spacer, matk, and ETS. Syst. Bot. 27, 801–814.
Beauverd, G., 1915. Contribution à l’étude des Composées (suite X). Bull. Soc. Bot.
Genève (sér. 2) 7, 21–56.
Bentham, G., 1873. Compositae. In: Bentham, G., Hooker, J.D. (Eds.), Genera
plantarum ad exemplaria imprimis in herbariis kewensibus servata definata,
vol. 2 (1). Lovell Reeve, London, pp. 163–533.
Beyers, J.B.P., 2000. Arctotis and Haplocarpha. In: Goldblatt, P., Manning, J.C. (Eds.),
Cape Plants: a Conspectus of the Cape Flora of South Africa. Strelitzia 9. National
Botanical Institute, Pretoria, and Missouri Botanical Garden, St. Louis, pp. 303–
306. 329.
Beyra-Matos, A., Lavin, M., 1999. A monograph of Pictetia (Papilionoideae;
Leguminosae) and review of the Aeschynomeneae. Syst. Bot. Monogr. 56, 1–93.
Born, J., Linder, H.P., Desmet, P., 2007. The Greater Cape floristic region. J. Biogeog.
34, 147–162.
Chase, B.M., Meadows, M.E., 2007. Late Quaternary dynamics of southern Africa’s
winter rainfall zone. Earth-Sci. Rev. 84, 103–138.
Cook, L.G., Crisp, M.D., 2005. Not so ancient: the extant crown group of Nothofagus
represents a post-Gondwanan radiation. Proc. R. Soc. Lond. B 272, 2535–2544.
Cowling, R.M. (Ed.), 1992. The Ecology of Fynbos: Nutrients, Fire and Diversity.
Oxford University Press, Cape Town.
Cowling, R.M., Campbell, B.M., Mustart, P., McDonald, D.J., Jarman, M.L., Moll, E.J.,
1988. Vegetation classification in a floristically complex area: the Agulhas Plain.
South Afr. J. Bot. 54, 290–300.
Cowling, R.M., Hilton-Taylor, C., 1997. Phytogeography, flora and endemism. In:
Cowling, R.M., Richardson, D.M., Pierce, S.M. (Eds.), Vegetation of Southern
Africa. Cambridge University Press, Cambridge, pp. 43–61.
Datson, P.M., Murray, B.G., Steiner, K.E., 2008. Climate and the evolution of annual/
perennial life-histories in Nemesia (Scrophulariaceae). Plant Syst. Evol. 270, 39–
57.
Dean, W.R.J., Milton, S.J. (Eds.), 1999. The Karoo: Ecological Patterns and Processes.
Cambridge University Press, Cambridge.
de Queiroz, A., 2005. The resurrection of oceanic dispersal in historical
biogeography. Trends Ecol. Evol. 20, 68–73.
Dingle, R.V., Hendey, Q.B., 1984. Late Mesozoic and Tertiary sediment input into the
eastern Cape basin (S.E. Atlantic) and palaeodrainage systems in south western
Africa. Marine Geol. 56, 13–26.
Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of
fresh leaf material. Phytochem. Bull. 19, 11–15.
Drummond, A., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by
sampling trees. BMC Evol. Biol. 7, 214.
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1994. Testing significance of
congruence. Cladistics 10, 315–319.
Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the
bootstrap. Evolution 39, 783–791.
Funk, V.A., Chan, R., Holland, A., 2007. Cymbonotus (Compositae: Arctotideae,
Arctotidinae): an endemic Australian genus embedded in a southern African
clade. Bot. J. Linn. Soc. 153, 1–8.
Funk, V.A., Chan, R., Keeley, S.C., 2004. Insights into the evolution of the tribe
Arctoteae (Compositae: subfamily Cichoroideae s.s.) using trnL-F, ndhF and ITS.
Taxon 53, 637–655.
Galley, C., Bytebier, B., Bellstedt, D.U., Linder, H.P., 2007. The Cape element in the
Afrotemperate flora: from Cape to Cairo? Proc. R. Soc. Lond. B 274, 535–543.
Galley, C., Linder, H.P., 2006. Geographical affinities of the Cape flora, South Africa. J.
Biogeog. 33, 236–250.
Germishuizen, G., Meyer, N.L. (Eds.), 2003. Plants of Southern Africa: an Annotated
Checklist. Strelitzia 14. National Botanical Institute, Pretoria.
Goertzen, L.R., Cannone, J.J., Gutell, R.R., Jansen, R.K., 2003. ITS secondary
structure derived from comparative analysis: implications for sequence
alignment and phylogeny of the Asteraceae. Mol. Phylogenet. Evol. 29,
216–234.
Goldblatt, P., 1978. An analysis of the flora of southern Africa: its characteristics,
relationships, and origins. Ann. Missouri Bot. Gard. 65, 369–436.
Good, R., 1964. The Geography of the Flowering Plants. Longmans Green and Co.,
London. Third edition.
Graur, D., Martin, W., 2004. Reading the entrails of chickens: molecular timescales
and the illusion of precision. Trends Genet. 20, 80–86.
Harvey, W.H., 1865. Compositae. In: Harvey, W.H., Sonder, O.W. (Eds.), Flora
Capensis, Being a Systematic Description of the Plants of the Cape Colony,
Caffraria and Port Natal, vol. 3. L. Reeve, London, pp. 44–530.
Hedberg, O., 1957. Afroalpine vascular plants: a taxonomic revision. Symbol. Bot.
Upsal. 15, 1–412.
Hilliard, O.M., 1977. Compositae in Natal. University of Natal Press,
Pietermaritzburg.
Hillis, D.M., Moritz, C., Porter, C.A., Baker, R.J., 1991. Evidence for biased gene
conversion in concerted evolution of ribosomal DNA. Science 251, 308–310.
Hipp, A.L., Hall, J.C., Sytsma, K.J., 2004. Congruence versus phylogenetic accuracy:
revisiting the incongruence length difference test. Syst. Biol. 53, 81–89.
Holland, A.E., Funk, V.A., 2006. A revision of Cymbonotus (Compositae: Arctotideae,
Arctotidinae). Telopea 11, 266–275.
Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference of phylogenetic
trees. Bioinformatics 17, 754–755.
Hurka, H., Paetsch, M., Bleeker, W., Neuffer, B., 2005. Evolution within the
Brassicaceae. Nova Acta Leopold. N. F. 92, 113–127.
Karis, P.O., 2006. Morphological data indicates two major clades of the subtribe
Gorteriinae (Asteraceae–Arctotideae). Cladistics 22, 199–221.
Karis, P.O., 2007. Tribe Arctotideae Cass. (1819). In: Kadereit, J.W., Jeffrey, C. (Eds.),
The Families and Genera of Vascular Plants, vol. 8. Springer-Verlag, Berlin, pp.
200–206.
Käss, E., Wink, M., 1997. Phylogenetic relationships in the Papilionideae (family
Leguminosae) based on nucleotide sequences of cpDNA (rbcL) and nrDNA (ITS 1
and 2). Mol. Phylogenet. Evol. 8, 65–88.
Kay, K.M., Whittall, J.B., Hodges, S.A., 2006. A survey of ribosomal internal
transcribed spacer substitution rates across angiosperms: an approximate
molecular clock with life history rates. BMC Evol. Biol. 6, 36.
Kim, K.J., Choi, K.S., Jansen, R.K., 2005. Two chloroplast DNA inversions originated
simultaneously during the early evolution of the sunflower family (Asteraceae).
Mol. Biol. Evol. 22, 1783–1792.
Klak, C., Reeves, G., Hedderson, T., 2003. Unmatched tempo of evolution in southern
African semi-desert ice plants. Nature 427, 63–65.
Kluge, A.G., 1989. A concern for evidence and a phylogenetic hypothesis of
relationships among Epicrates (Boidae, Serpentes). Syst. Zool. 38, 7–25.
Koekemoer, M., 1996. An overview of the Asteraceae of southern Africa. In: Hind,
D.J.N., Beentje, H.J. (Eds.), Compositae: Systematics. Royal Botanic Gardens,
Kew, pp. 95–110.
Lee, M.S.Y., 2001. Uninformative characters and apparent conflict between
molecules and morphology. Mol. Biol. Evol. 18, 676–680.
Lessing, C.F., 1831. Synanthereae Rich. In: Chamisso, L.C.A. von, Schlechtendal, D.F.L.
von (Eds.), De plantis in expeditione speculatoria Romanzoffiana observatis
rationem dicunt, pp. 83-170. Linnaea 6, 76–170.
Lessing, C.F., 1832. Synopsis generum compositarum earumque dispositionis novae
tentamen monographiis multarum capensium interjectis. Duncker and
Humblot, Berlin.
Lewin, K., 1922. Systematische Gliederung und geographische Verbreitung der
Arctotideae–Arctotidinae. Repert. spec. nov. regni veg., Beih. 11, 1–75.
Linder, H.P., 1990. On the relationship between the vegetation and floras of the
afromontane and the Cape regions of Africa. Mitt. Inst. Allg. Bot. Hamburg 23b,
777–790.
Linder, H.P., 2003. The radiation of the Cape flora, southern Africa. Biol. Rev. 78,
597–638.
Linder, H.P., 2005. Evolution of diversity: The Cape flora. Trends Plant Sci. 10, 536–
541.
Linder, H.P., Barker, N.P., 2005. From Nees to now—changing questions in the
systematics of the grass subfamily Danthonioideae. Nova Acta Leopold. N. F. 92,
29–44.
Linder, H.P., Eldenäs, P., Briggs, B.G., 2003. Contrasting patterns of radiation in
African and Australian Restionaceae. Evolution 57, 2688–2702.
15
Linder, H.P., Hardy, C.R., 2005. Species richness in the Cape flora: a
macroevolutionary and macroecological perspective. In: Bakker, F.T., Chatrou,
L.W., Gravendeel, B., Pelser, P.B. (Eds.), Plant Species-level Systematics: New
Perspectives on Patterns and Process. Koeltz, Königstein, pp. 47–73.
Linder, H.P., Hardy, C.R., Rutschmann, F., 2005. Taxon sampling effects in molecular
clock dating: an example from the African Restionaceae. Mol. Phylogenet. Evol.
35, 569–582.
Maddison, D.R., 1991. The discovery and importance of multiple islands of mostparsimonious trees. Syst. Biol. 40, 315–328.
Maddison, W.P., Maddison, D.R., 2000. MacClade 4: Analysis of Phylogeny and
Character Evolution. Sinauer Associates, Sunderland.
Marlow, J.R., Lange, C.B., Wefer, G., Rosell-Melé, A., 2000. Upwelling intensification
as part of the Pliocene-Pleistocene climate transition. Science 290, 2288–2291.
Mason-Gamer, R.J., Kellogg, E.A., 1996. Testing for phylogenetic conflict among
molecular data sets in the tribe Triticeae (Gramineae). Syst. Biol. 45, 524–545.
McKenzie, R.J., Herman, P.P.J., Barker, N.P., 2006b. Arctotis decurrens (Arctotideae),
the correct name for A. merxmuelleri and A. scullyi. Bothalia 36, 171–173.
McKenzie, R.J., Hjertson, M., Barker, N.P. in press. Typification of Arctotis plantaginea
and names in the Arctotis semipapposa species complex (Asteraceae,
Arctotideae). Taxon..
McKenzie, R.J., Mitchell, S.D., Barker, N.P., 2006c. A new species of Arctotis
(Compositae, Arctotideae) from kommetjie grassland in Eastern Cape
Province, South Africa. Bot. J. Linn. Soc. 151, 581–588.
McKenzie, R.J., Muller, E.M., Skinner, A.K.W., Karis, P.O., Barker, N.P., 2006a.
Phylogenetic relationships and generic delimitation in subtribe Arctotidinae
(Asteraceae: Arctotideae) inferred by DNA sequence data from ITS and five
chloroplast regions. Am. J. Bot. 93, 1222–1235.
McKenzie, R.J., Samuel, J., Muller, E.M., Skinner, A.K.W., Barker, N.P., 2005.
Morphology of cypselae in subtribe Arctotidinae (Compositae–Arctotideae)
and its taxonomic implications. Ann. Missouri Bot. Gard. 92, 569–594.
McLoughlin, S., 2001. The breakup history of Gondwana and its impact on preCenozoic floristic provincialism. Austral. J. Bot. 49, 271–300.
Meerow, A.W., Lehmiller, D.J., Clayton, J.R., 2003. Phylogeny and biogeography of
Crinum L. (Amaryllidaceae) inferred from nuclear and limited plastid noncoding DNA sequences. Bot. J. Linn. Soc 141, 349–363.
Mesfin Tadesse, 2004. Asteraceae (Compositae). In: Hedberg, I., Friis, I., Edwards, S.
(Eds.), Flora of Ethiopia and Eritrea, vol. 4 (2). National Herbarium, Addis Ababa
University, Addis Ababa and Department of Systematic Botany, Uppsala
University, Uppsala, pp. 1–408.
Midgley, G.F., Hannah, L., Roberts, R., Macdonald, D.J., Allsopp, J., 2001. Have
Pleistocene climatic cycles influenced species richness patterns in the Greater
Cape Mediterranean Region? J. Mediterr. Ecol. 2, 137–144.
Midgley, G.F., Reeves, G., Klak, C., 2005. Late Tertiary and Quaternary climate change
and centres of endemism in the southern African flora. In: Purvis, A., Gittleman,
J., Brooks, T. (Eds.), Phylogeny and Conservation. Cambridge University Press,
Cambridge, pp. 230–242.
Mummenhoff, K., Al-Shehbaz, I.A., Bakker, F.T., Linder, H.P., Mühlhausene, A., 2005.
Phylogeny, morphological evolution, and speciation of endemic Brassicaceae
genera in the Cape flora of southern Africa. Ann. Missouri Bot. Gard. 92, 400–
424.
Mummenhoff, K., Linder, H.P., Friesen, N., Bowman, J.L., Lee, J.Y., Franzke, A., 2004.
Molecular evidence for bicontinental hybridogenous genomic constitution in
Lepidium sensu stricto (Brassicaceae) species from Australia and New Zealand.
Am. J. Bot. 91, 254–261.
Nixon, K.C., Carpenter, J.M., 1996. On simultaneous analysis. Cladistics 12, 221–241.
Partridge, T.C., 1997. Evolution of landscapes. In: Cowling, R.M., Richardson, D.M.,
Pierce, S.M. (Eds.), Vegetation of Southern Africa. Cambridge University Press,
Cambridge, pp. 5–20.
Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M.,
Chappellaz, J., Davis, M., Delayque, G., Delmotte, M., Kotlyakov, V.M., Legrand,
M., Lipenkov, V.Y., Lorius, C., Pépin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999.
Climate and atmospheric history of the past 420 000 years from the Vostok ice
core, Antarctica. Nature 399, 429–436.
Phillips, E.P., 1951. The Genera of South African Flowering Plants. Department of
Agriculture, Pretoria.
Pickford, M., 2004. Southern Africa: a cradle of evolution. South Afr. J. Sci. 100, 205–
214.
Pope, G.V., 1992. Compositae. In: Pope, G.V. (Ed.), Flora Zambesiaca: Mozambique,
Malawi, Zambia, Zimbabwe & Botswana, vol. 6 (1). Royal Botanic Gardens, Kew,
pp. 1–264.
Price, J.P., Clague, D.A., 2002. How old is the Hawaiian biota? Geology and
phylogeny suggest recent exchange. Proc. R. Soc. Lond. B 269, 2429–2435.
Ramírez, M.J., 2006. Further problems with the incongruence length difference test:
‘‘hypercongruence” effect and multiple comparisons. Cladistics 22, 289–295.
Reese, H., 1989. Die Entwicklung von Pericarp und Testa bei Calenduleae und
Arctotideae (Asteraceae)–ein Beitrag zur Systematik. Bot. Jahrb. Syst. 110, 325–
419.
Renvoize, S.A., 1979. The origins of Indian Ocean island floras. In: Bramwell, D. (Ed.),
Plants and Islands. Academic Press, London, pp. 107–129.
Ronquist, F., 1997. Dispersal–vicariance analysis: a new approach to the
quantification of historical biogeography. Syst. Biol. 46, 193–201.
Sang, T., Crawford, D.J., Stuessy, T.F., 1997. Chloroplast DNA phylogeny, reticulate
evolution and biogeography of Paeonia (Paeoniaceae). Am. J. Bot. 84, 1120–
1136.
Sang, T., Crawford, D.J., Stuessy, T.F., Solva, O.M., 1995. ITS sequences and the
phylogeny of the genus Robinsonia (Asteraceae). Syst. Bot. 20, 55–64.
16
R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16
Sang, T., Zhong, Y., 2000. Testing hybridization hypotheses based on incongruent
gene trees. Syst. Biol. 49, 422–434.
Sanmartín, I., Ronquist, F., 2004. Southern Hemisphere biogeography inferred by
event-based models: plant versus animal patterns. Syst. Biol. 53, 216–243.
Schmidt, G.J., Schilling, E.E., 2000. Phylogeny and biogeography of Eupatorium
(Asteraceae: Eupatorieae) based on nuclear ITS sequence data. Am. J. Bot. 87,
716–726.
Schwarz, M.P., Fuller, S., Tierney, S.M., Cooper, S.J.B., 2006. Molecular phylogenetics
of the exoneurine allodapine bees reveal an ancient and puzzling dispersal from
Africa to Australia. Syst. Biol. 55, 31–45.
Schrire, B.D., Lavin, M., Barker, N.P., Cortes-Burns, H., von Senger, I., Kim, J.H., 2003.
Towards a phylogeny of Indigofera (Leguminosae—Papilionoideae):
identification of major clades and relative ages. Adv. Legume Syst. 10, 269–302.
Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-based
phylogenetic analyses. Syst. Biol. 49, 369–381.
Soltis, D.E., Soltis, P.S., 1998. Choosing an approach and an appropriate gene for
phylogenetic analysis. In: Soltis, D.E., Soltis, P.S., Doyle, J.J. (Eds.), Molecular
Systematics of Plants II: DNA Sequencing. Kluwer, Boston, pp. 1–42.
Swofford, D.L., 2002. PAUP. Phylogenetic Analysis Using Parsimony (* and Other
Methods), v. 4.0 beta 10. Sinauer Associates, Sunderland.
Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for
amplification of three non-coding regions of chloroplast DNA. Plant Mol. Biol.
17, 1105–1109.
Tankard, A.J., Rogers, J., 1978. Late Cenozoic palaeoenvironments and
palaeoclimates of southern Africa. South Afr. J. Sci. 95, 194–201.
Thorne, R.F., 1972. Major disjunctions in the geographic ranges of seed plants. Q.
Rev. Biol. 47, 365–411.
van Wyk, A.E., Smith, G.F., 2001. Regions of Floristic Endemism in Southern Africa.
Umdaus Press, Hatfield.
Vences, M., Vieites, D.R., Glaw, F., Brinkmann, H., Kosuch, J., Veith, M., Meyer, A.,
2003. Multiple overseas dispersal in amphibians. Proc. R. Soc. Lond. B 270,
2435–2442.
Verboom, G.A., Archibald, J.K., Bakker, F.K., Bellstedt, D.U., Conrad, F., Dreyer, L.L.,
Forest, F., Galley, C., Goldblatt, P., Henning, J.F., Mummenhoff, K., Linder, H.P.,
Muasya, A.M., Oberlander, K.C., Savolainen, V., Snijman, D.A., van der Niet, T.,
Nowell, T.L., 2008. Origin and diversification of the Greater Cape flora: Ancient
species repository, hot-bed of recent radiation, or both? Mol. Phylogenet. Evol.,
in press, doi: 10.1016/j.ympev.2008.01.037..
Verboom, G.A., Linder, H.P., Stock, W.D., 2003. Phylogenetics of the grass genus
Ehrharta: evidence for radiation in the summer-arid zone of the South African
Cape. Evolution 57, 1008–1021.
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: a Guide to Methods and
Applications. Academic Press, San Diego, pp. 315–324.
Yang, Z., Rannala, B., 1997. Bayesian phylogenetic inference using DNA
sequences: a Markov Chain Monte Carlo method. Mol. Biol. Evol. 14,
717–724.
Yoder, A.D., Irwin, J.A., Payseur, B.A., 2001. Failure of the ILD to determine data
combinability for slow loris phylogeny. Syst. Biol. 50, 408–424.
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.
Zavada, M.S., de Villiers, S.E., 2000. Pollen of the Asteraceae from the PaleoceneEocene of South Africa. Grana 39, 39–45.