Biome conservatism prevailed in repeated long-distance colonization of Madagascar’s mountains by Helichrysum (Compositae, Gnaphalieae)

Full text

Biome conservatism prevailed in repeated long-distance colonization of
Madagascar’s mountains by Helichrysum (Compositae, Gnaphalieae)
Carme Blanco-Gavald`
a
a,*
, Cristina Roquet
a
, Genís Puig-Surroca
a
, Santiago Andr´
es-S´
anchez
b
,
Sylvain G. Razamandimbison
c
, Rokiman Letsara
d
, Nicola Bergh
e,f
, Glynis V. Cron
g
,
Lucía D. Moreyra
h
, Juan Antonio Calleja
i
, `
Oscar Castillo
h
, Randall J. Bayer
j
, Frederik Leliaert
k
,
Alfonso Susanna
h
, Merc`
e Galbany-Casals
a
a
Systematics and Evolution of Vascular Plants (UAB), Associated Unit to CSIC by IBB, Departament de Biologia Animal, Biologia Vegetal i Ecologia, Facultat de
Bioci`
encies, Universitat Aut`
onoma de Barcelona, 08193 Bellaterra, Spain
b
University of Salamanca, Department of Botany and Plant Physiology, Pharmacy Faculty, C/Licenciado M´
endez Nieto s/n 37007, Salamanca, Spain
c
Swedish Museum of Natural History, BOX 50007, SE-10405 Stockholm, Sweden
d
Herbarium of the Parc Botanique et Zoologique of Tsimbazaza (PBZT), 3G9G+V6C, Antananarivo, Madagascar
e
South African National Biodiversity Institute, Kirstenbosch NBG, Rhodes Drive, Newlands, Cape Town, South Africa
f
Gothenburg Botanical Gardens, Carl Skottsbergs Gata 22A, 413 19 Gothenburg, Sweden
g
School of Animal, Plant & Environmental Sciences, University of Witwatersrand, 1 Jan Smuts Avenue, Braamfontein2000, Johannesburg, South Africa
h
Botanic Institute of Barcelona (IBB, CSIC- Ajuntament de Barcelona), Pg. del Migdia s.n., 08038 Barcelona, Spain
i
Autonomous University of Madrid, 28049 Madrid, Spain
j
University of Memphis, Ellington Hall, 3700 Walker Avenue, Memphis, TN 38152-3540, USA
k
Meise Botanic Garden, Nieuwelaan 38, 1860 Meise, Belgium
ARTICLE INFO
Keywords:
Afromontane ora
Asteraceae
Biogeography
Dispersal
Helichrysum
Madagascar
Target-enrichment
ABSTRACT
Colonization and diversication processes are responsible for the distinctiveness of island biotas, with
Madagascar standing out as a biodiversity hotspot exceptionally rich in species and endemism. Regardless of its
signicance, the evolutionary history and diversication drivers of Madagascar’s ora remain understudied.
Here we focus on Helichrysum (Compositae, Gnaphalieae) to investigate the evolutionary and biogeographic
origins of the Malagasy ora. We inferred a highly resolved phylogeny based on target-enrichment data from 327
species (including 51 % of Malagasy endemics) and conducted ancestral range estimation analyses. Our results
revealed at least six trans-oceanic dispersal events from different African regions to Madagascar during the
Pliocene. In this process, biome conservatism prevailed, as evidenced by similarities between Malagasy lineages
and their African relatives. The southern African grasslands, known to be the center of diversication and the
main source of African Helichrysum lineages, played a key role in the colonization of Madagascar as the ancestral
source area of at least three clades. The Tropical Afromontane region was revealed as the source of at least two
montane Malagasy lineages that substantially radiated in-situ. Finally, a dispersal event from southwestern Africa
led to a lineage represented by a single species adapted to the island’s southwestern arid conditions. The main
radiations of Helichrysum in Madagascar’s mountains occurred within the last 2 My, coinciding with a transition
towards cooler and drier conditions and the expansion of open habitats, likely driven by a combination of
geographic and ecological speciation. Overall, our ndings highlight the afnities between the montane oras of
continental Africa and Madagascar.
1. Introduction
Unraveling the origins and evolution of island biotas has been of
interest to scientists since Darwin’s times (Darwin 1859). Due to their
isolation, comparatively small size and varying distances from main-
lands, islands are considered natural laboratories for the study of
* Corresponding author: Systematics and Evolution of Vascular Plants (UAB), Associated Unit to CSIC by IBB, Departament de Biologia Animal, Biologia Vegetal i
Ecologia, Facultat de Bioci`
encies, Universitat Aut`
onoma de Barcelona, 08193 Bellaterra, Spain.
E-mail address: carme.blanco@uab.cat (C. Blanco-Gavald`
a).
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
https://doi.org/10.1016/j.ympev.2024.108283
Received 22 October 2024; Received in revised form 5 December 2024; Accepted 30 December 2024

ecological and evolutionary processes. Species richness of islands can
generally be explained through features such as island area and isolation
as well as the age and formation history (Whittaker et al. 2008, Graham
et al. 2017). Distinguishing between oceanic islands and continental
islands is important as it dictates the mechanisms underlying species
colonization. Unlike the relatively younger oceanic islands, which
emerge devoid of any resident species and gradually accumulate their
diversity through dispersal followed by speciation, older continental
islands inherit a baseline biota upon isolation from the mainland
(Matthews & Triantis 2021).
Madagascar, once part of Gondwana, is an ancient continental is-
land. It rst broke apart from Africa 155 to 165 million years ago (mya)
(Yoder & Nowak 2006 and references therein) and later from India 84 to
91 mya (Wells 2003) and has remained separated since then.
Madagascar hosts a hyperdiverse biota in a wide range of ecosystems,
being home to an estimated total ora of ca. 14,900 vascular plant
species, of which 87 % are endemics (Callmander et al. 2011, Lowry
et al. 2018, Antonelli et al. 2022), with 310 endemic plant genera and
ve endemic families (Buerki et al. 2013). Such exceptional species
richness and endemicity are the result of a complex geological, climatic
and evolutionary history. Despite hosting vicariant groups predating
Gondwana’s breakup (e.g., Takhtajania, Winteraceae, Thomas et al.
2014), most of the present-day diversity likely established on
Madagascar through long-distance dispersal (Yoder & Novak 2006,
Buerki et al. 2013), probably assisted by wind and oceanic currents. The
existence of land bridges and stepping-stones between Africa and
Madagascar has been hypothesized, but not supported by any direct
evidence, so their potential role and relevance in facilitating dispersal
remains under debate (Warren et al. 2010, Masters et al. 2021, Ali &
Hedges 2023, Aslanian et al. 2023). Geographic proximity likely ac-
counts for the Malagasy biota having the strongest taxonomic links to
African lineages (Yoder & Novak 2006, Buerki et al. 2013).
Madagascar’s complex topography plays a key role in shaping the
main climatic and vegetation zones of the island and has probably had a
signicant impact on in-situ diversication of many groups. The main
massifs of the island, reaching 1800–2900 m a.s.l., are placed in a
north–south axis and are connected through the Central Highlands, a
plateau ranging 800–1300 m. The orography combined with the pre-
vailing easterly trade winds results in a rainfall gradient from the humid
tropical east-northeast to the sub-arid southwest. The Central Highlands
have a subhumid climate and are dominated by a grassland-woodland
mosaic, heavily modied and degraded by human activity (Yoder
et al. 2016; Antonelli et al. 2022). The highest parts of the mountains,
above 1800–1900 m, are mainly occupied by sclerophyllous shrublands
dominated by Ericaceae and Compositae, in addition to open grasslands
and rupicolous plant communities (Burgoyne et al. 2005; Yoder et al.
2016; Antonelli et al. 2022, and pers. obs.). Outside the mountains,
landscapes are dominated by deciduous forests, sclerophyllous and
succulent plants westwards and evergreen rainforests eastwards. This
ecological and biological heterogeneity provides the perfect ground for
various post-colonization scenarios, such as repeated colonization
without subsequent diversication (e.g. the Inulea-Pulcheinae group,
Nylinder et al. 2016) as well as colonization followed by in-situ diver-
sication (e.g. Psychotrieae alliance, Razamandimbison et al. 2017;
Coffeeae alliance, Kainulainen et al. 2017; C3 grasses, Hackel et al.
2018). Nevertheless, there is a knowledge gap on the relative impor-
tance of these scenarios in the evolutionary assembly of the Malagasy
biota, given that the biogeographic history of many highly diverse
Malagasy plant groups remains largely unexplored (Antonelli et al.
2022).
The relative contribution of adaptive evolution to the generation of
biodiversity is still under debate. Niche conservatism -i.e., the tendency
of species to retain their ancestral niches- has been suggested to be
predominant in allopatric (also known as geographic) speciation (Wiens
& Graham 2005, Wiens et al. 2010) and has been assumed to prevail in
angiosperm evolution (Wiens & Graham 2005, Crisp. et al. 2009, Wiens
et al. 2010) since migrating might be easier and quicker to achieve than
adapting to new environmental conditions (Donoghue 2008). However,
multiple studies support the idea that niche shifts within plant genera
are not as rare as previously thought (e.g. Lonicera L., Smith & Donoghue
2010; Coccinia Wight & Arn., Holstein & Renner 2011; Ranunculus L.,
H¨
orandl & Emadzade 2011; Hakea Schrad & J.C.Wendl., Cardillo et al.
2017) and indeed older and/or more extensive biomes (e.g. scle-
rophyllous shrubland, tropical rainforest) have often been the source of
lineages found in younger and/or less extensive biomes (e.g. grasslands,
alpine, Mediterranean shrublands; Crisp et al. 2009, Donoghue &
Edwards 2014). To date, few studies have explored the role that niche
evolution may have played in the diversication of Malagasy plant
groups (but see the case of Bulbophyllum Thouars in Gamisch et al. 2016,
2021).
In terms of species richness, Compositae are one of the ve dominant
plant families in the Malagasy ora (520 spp. and 83 % endemism;
Callmander et al. 2011, Catalogue of the Vascular Plants of Madagascar
2024). Within Malagasy Compositae, the most species-rich genus is
Helichrysum Mill. In the most recent oristic treatment of Malagasy
Helichrysum, Humbert (1962) recognized 111 species −all but one
endemic- and numerous infraspecic taxa at the subspecic and varietal
levels. He did not propose a formal infrageneric classication but
organized the species into eleven informal taxonomic groups based on
morphological afnities. Helichrysum occurs in almost all Malagasy bi-
omes, from the extremely arid south-western coastal dunes (e.g.,
H. mahafaly Humbert) to the humid eastern rainforests (e.g. H. geayi
Humbert), adopting a wide range of growth forms (lianas, herbs, sub-
shrubs, shrubs and small trees). However, most of the Malagasy diversity
of Helichrysum is found in the Central Highlands and on the highest
mountains, distributed across their full elevational and latitudinal gra-
dients. The diversity of Helichrysum in Madagascar reects the extraor-
dinary variation and adaptability of the genus, already evident in the
case of continental African lineages (Blanco-Gavald`
a et al. 2023).
Previous studies place the origin of Helichrysum in southern Africa
(Galbany-Casals et al. 2014, Andr´
es-S´
anchez et al. 2019, Blanco-Gavald`
a
et al. 2023) and suggest that several independent and asynchronous
dispersals occurred from the African continent to Madagascar. However,
these works treated Madagascar as a single biogeographic area and the
sampling of both mainland and Malagasy species was limited. Here, we
aim to infer the biogeographic history of Helichrysum in Madagascar as a
study-case to elucidate the relative contribution of in-situ speciation vs.
colonization to the generation and maintenance of Malagasy plant di-
versity. For this purpose, we rst generated a highly resolved time-
calibrated phylogeny based on target-enrichment sequences, substan-
tially increasing the sampling of African and Malagasy species compared
to previous studies. We used the resulting phylogeny to estimate
ancestral ranges and infer the source, number and age of colonization
events of the genus on Madagascar. Finally, we evaluate whether the
dominant mode of diversication within Helichrysum has been adaptive
speciation associated with biome shifts or non-adaptive speciation
associated with allopatry and biome conservatism.
2. Materials and methods
2.1. Taxon sampling
We sampled 327 Helichrysum species (ca. 60 % of the genus, see
Supplementary Table S1) including 57 (ca. 51 %) of the Malagasy spe-
cies with representatives of all Humbert’s (1962) taxonomic groups, the
latest taxonomic treatment of Malagasy Helichrysum. For some of these
Malagasy species, we included several infraspecic taxa (29 additional
samples). While polyploids are known in Helichrysum and related
genera, previous studies have shown that clades of polyploid origin are
conned to specic geographical regions and lineages (the Mediterra-
nean, Macaronesian and Asian members of Helichrysum, and the smaller
genera Achyrocline (Less.) DC., Anaphalis DC. and Pseudognaphalium
C. Blanco-Gavald`
a et al.

Kirp. in America and Asia; Galbany-Casals et al. 2009, 2014). None of
these is closely related to Malagasy lineages (Blanco-Gavald`
a et al.
2023). Taking this into account, we excluded well-delimited polyploid
clades to optimize read-mapping processes and minimize potential
sources of phylogenetic discordance (Tiley et al. 2024a). With this, we
also eliminated non-African biogeographic regions, allowing for the
recognition of more precise Malagasy areas while keeping the analyses
computationally feasible.
We also included 34 specimens of other genera representing the main
lineages of the tribe (Nie et al. 2016, Smissen et al. 2020) so that we
could implement secondary age calibrations. The complete dataset
comprises a total of 386 samples, 289 from previous studies (273 from
Blanco-Gavald`
a et al. 2023 BioProject PRJNA936872; six from Mandel
et al. 2019, BioProject PRJNAS40287; and ten from Schmidt-Lebuhn &
Bovill 2021, BioProject PRJNA665592). Here we sequenced for the rst
time 97 samples using the same baits set, representing mainly Heli-
chrysum species from continental Africa and Madagascar.
2.2. Next-generation sequencing
DNA was extracted from 10–30 mg of dried leaf material obtained
from herbarium specimens and from our eld expeditions (see Supple-
mentary Table S1) using the E.N.Z.A® SP Plant DNA Kit (Omega Bio-Tek
Inc., Norcross, GA, USA) following manufacturer instructions. We
measured DNA concentration using Qubit
TM
Flex Fluorometer (Thermo
Scientic, Waltham, MA, USA). Then, we used a Qsonica Q800R3 Son-
icator (Qsonica LLC, Newton, CT, USA) at 20 % amplitude for 45 sec to 8
min to shear 0.2–1 ug in 50 uL of DNA into 300–400 bp fragments. We
performed gel electrophoresis (1.2 % agarose) to check fragment length.
We prepared target-enrichment libraries from 25ul of the sonicated
DNA using the NEBNext Ultra II DNA Library Prep Kit for Illumina ®
(New England Biolabs, Ipswich, MA, USA) employing half of the vol-
umes specied by the manufacturer and 15 cycles of PCR amplication.
We barcoded the libraries using NEBNext Multiplex Oligos sets with
unique single or dual index combinations. Afterward, we pooled the
indexed libraries with more than 17 ng of DNA in groups of up to 10
samples and around 250 ng of DNA per library. We evaporated or lled
with water the arranged pools to 7 ul of total volume to perform target-
enrichment (protocol from Mandel et al. 2014) using the Microarray
MyBaits COS kit (Daicel Arbor Biosciences, Ann Arbor, MI, USA), spe-
cically developed for the Compositae family. The nal sequencing
pools were prepared by pooling enriched libraries with unenriched li-
braries at a 60:40 ratio and the samples were sequenced (PE 150 bp) on
Illumina HiSeq2500 and HiSeqX platforms.
We deposited newly generated raw sequence reads in the NCBI Short
Read Archive database (SRA; access: https://www.ncbi.nlm.nih.
gov/sra) under the BioProject accession number PRJNA1121119.
2.3. Molecular data processing
We used HybPhyloMaker, a bioinformatic workow developed to
process Hyb-Seq data (F´
er & Schmickl 2018, available at https://github.
com/tomas-fer/HybPhyloMaker, indicated hereafter as HPM, followed
by the number of the corresponding script) in combination with Paral-
ogWizard (Umov et al. 2022, available at https://github.com/ru
mov/ParalogWizard), which detects and separates paralogs of a
given locus based on sequence similarity to generate orthologous
alignments.
Specically, we used Trimmomatic v.0.39 (Bolger et al. 2014) to
remove adaptors and low-quality reads, and BBMap v.38.42 (Bushnell
2014) to remove duplicates, both implemented in HPM1 (read statistics
in Supplementary Table S2). We generated a reference le for initial
read mapping with BWA (Li & Durbin 2009) and SPAdes (Bankevich
et al. 2012) based on sunower genome sequences from the Composi-
tae1061 probe set (Mandel et al. 2014). Then, we generated a custom-
ized reference based on our ingroup samples to increase mapping
specicity. Pairwise exonic sequence divergence was calculated to
identify paralogs. The rst resulting peak represents putative allelic
variation, while the second peak represents highly divergent sequences
corresponding to putative paralogs. We used the value of the latter as the
threshold to retrieve putative paralogous sequences. We then aligned
orthologous matrices using MAFFT v.7.475 (Katoh & Toh 2008) to
nally concatenate exons into putative loci. To reduce missing data, we
excluded sequences missing more than 70 % of the data and removed
loci for which less than 75 % of all samples were represented (HPM5).
2.4. Phylogenetic analyses
We applied concatenation as well as coalescence summary ap-
proaches to infer phylogenetic relationships. We concatenated all nu-
clear loci into a single supermatrix and conducted maximum likelihood
(ML) partitioned analyses using RAxML-NG v.1.1.0 (modied HPM8f;
Kozlov et al. 2019). Specically, we rst estimated the best nucleotide
substitution model for each locus with ModelTest-NG (Darriba et al.
2020) and then performed 20 independent ML tree searches. We
assessed branch support with 1000 bootstrap replicates and annotated
the best-scoring ML tree with Felsenstein’s Bootstrap (BS, Felsenstein
1985) and Transfer Bootstrap Expectation (TBE, Lemoine et al. 2018)
proportions considering branches with BS ≥70 % and TBE ≥0.7 to be
statistically supported (Hillis & Bull 1993, Lemoine et al. 2018).
We performed summary-coalescence inference with ASTRAL-III
v.5.7.8 (Zhang et al. 2018). To do so, we rst inferred individual gene
trees for each retrieved locus generated with RAxML v.8.2.12
(Stamatakis 2014) and performed 100 bootstrap replicates (HPM6a,
HPM7 and HPM8a). For the ASTRAL tree, we calculated branch support
values as local posterior probabilities (LPP), considering well-supported
branches those with LPP ≥0.95 (Sayyari & Mirarab 2016). In addition,
we conducted a second summary-coalescence analysis with the same
parameters but using only the 25 most informative loci, which we
selected with SortaDate (Smith et al. 2018) based on three criteria:
bipartition support, clock-likeness, and tree length (which is propor-
tional to information content).
2.5. Divergence time estimation
We applied the RelTime method (Tamura et al. 2012, 2018) imple-
mented in MEGA 11 (Tamura et al. 2021) to time-calibrate the most
likely ML tree, and computed condence intervals using the method of
Tao et al. (2020). This methodology is especially suited for large
empirical genomic datasets and allows the use of calibration densities
(Costa et al. 2022). The tribe Gnaphalieae lacks old enough fossils to be
useful as primary calibration points. Therefore, we used four secondary
calibration points (CP, Supplementary Figure S1) from previously
inferred divergence time estimates (Nie et al. 2016) as age constraints.
We applied a normal density to each calibration point providing
appropriate mean and standard deviation values to reect the 95 %
condence interval values reported by Nie et al. (2016): the tribe crown
node (mean age 25 mya, std ±2.55, CP1), the “crown radiation” node
(mean age 20.7 mya, std ±2.55, CP2), the “HAP clade” crown node
(mean age 15.39 mya, std ±1.95, CP3) and the “FLAG clade” crown
node (mean age 12.78 mya, std ±1.85, CP4).
2.6. Ancestral geographic range reconstruction
To infer the most probable ancestral geographic ranges of Heli-
chrysum lineages, we dened 17 areas. Regarding continental Africa and
Arabia, we applied the same criteria as in Blanco-Gavald`
a et al. (2023)
resulting in the following areas: (A) the tropical Afroalpine area, (C) the
Indian Ocean Coastal Forest Belt, (D) the high Drakensberg area, (F) the
Fynbos Biome, (G) the southern African grasslands, (L) the tropical Af-
rican lowlands, (N) the arid to semi-arid southern African area, (P) the
Arabian Peninsula in Asia, (S) the southern African savannah, and (T)
C. Blanco-Gavald`
a et al.

the tropical Afromontane area.
For our delimitation of Malagasy regions (Fig. 1) we consulted many
bibliographic sources: the denitions of the main vegetation types
described in Jenkins (1987), Burgess et al. (2004), Goodman & Benstead
(2003) −which in turn integrates previous classications by Faramalala
(1995), Du Puy & Moat (1996) and Lowry et al. (1997)- and Gautier
et al. (2018); the delimitation of Madagascar’s six principal biomes by
Yoder & Nowak (2006), based on Goodman & Benstead (2003); the map
of predominant vegetation types shown in Antonelli et al. (2022)
simplied from Moat & Smith (2007); and the bioclimatic zones by
Rakotoarivelo et al. (2019) modied in Razamandimbison et al.
(2022). However, our nal proposal is not completely coincident with
any of the cited sources, as it also integrates distribution patterns of the
genus and our eld experience. We used the following seven distribution
areas: (B) southern and southwestern arid lowlands, an area that ranges
from 0 to 800 m and mainly includes Madagascar’s arid spiny thicket,
with drought-resistant species adapted to extreme aridity and secondary
grasslands; (W) dry western lowlands, an area that ranges approxi-
mately from 0 to 800 m and mainly includes dry and highly seasonal
forests, succulent woodlands and secondary grasslands; (O) humid
eastern lowlands, an area that ranges approximately from 0 to 800 m
and mainly includes very humid lowland rainforests with some herba-
ceous clearings; (H) subhumid central highlands, an area that ranges
from 800 to 1300 m and is mainly constituted by grassland formations
and small patches of Tapia forests; (M) subhumid central montane belt,
an area that ranges from 1300 to 1800 m and includes a mosaic of Tapia
forest, woodlands, grasslands and rocky habitats; (R) humid eastern and
northern montane belt, an area that ranges from 800 to 1800 m and
mainly constituted by moist, broadleaf forests that, because they grow at
certain altitude, are not as tall as lowland rainforests; and (E) highest
areas, which are disjoint small patches comprising all areas above 1800
m and up to the summit of Maromokotro (2876 m) and include a mosaic
of ericoid thickets, mountain grasslands and rocky habitats (these
correspond, broadly speaking, to the massifs of Tsaratanana [and its
surrounding mountain chains], Marojejy, Anjanaharibe, Ankaratra,
Vavavato, Ibity [and its neighbor mountains], Itremo, Andringitra [and
its satellite peaks] and Beampingaratra). We excluded the Lowland
Sambirano Rainforest biome (grey area in Fig. 1) represented in some
maps at the north-western tip of the island, and coastal mangroves (not
shown in Fig. 1), because none of the included Helichrysum species in-
habits these regions.
Based on all these criteria, our proposed biogeographical areas can
be viewed as macroecological units, which despite not being
geographically continuous in some cases, share bioclimatic conditions
determined by factors such as temperature and precipitation that likely
ltered out lineages according to their general bioclimatic preferences.
Species occurrence in the geographical areas was assigned consid-
ering information on distribution and elevational range (based on
Humbert 1962, Catalogue of the Vascular Plants of Madagascar 2024,
herbarium records and own eld observations). Some species are only
occasionally or marginally present in other areas outside of their main
distribution range. In these cases, we only considered the core area/s or
Fig. 1. Maps illustrating the 17 geographic areas dened in this study. The colors and labels correspond to those in the biogeographical reconstruction analyses. (a)
General map at scale, including continental Africa, Arabia and Madagascar. The dark green spots within the tropical Afromontane area are an overrepresentation of
the tropical Afroalpine area. (b) Closeup of Madagascar’s map to facilitate area distinction.
C. Blanco-Gavald`
a et al.

elevational belts in which there are a signicant number of occurrences
of the species, or that can be interpreted as the central distribution of the
species.
We carried out biogeographic range evolution analyses in R with the
package BioGeoBEARS (Matzke 2013), using the ML time-calibrated
tree as input. The maximum number of areas for any node was set to
three, which is the highest number of areas occupied by the most
widespread extant taxon in our study. We tested the t of three
biogeographical models: Dispersal-Extinction-Cladogenesis (DEC; Ree
et al. 2005; Ree & Smith 2008), a likelihood implementation of the
Fig. 2. Ancestral range estimation of Helichrysum using the best-tting model DEC +j. It is based on a time-calibrated phylogeny generated under the concatenation
approach using target-enrichment data (Compositae1061 probe set). Pie charts at nodes show the relative probability of the possible states (areas in primary colors,
combinations of areas in grey). Relevant node numbers are to the right of the node. Support values are to the left of the node. The rst numerical value corresponds to
the BS metric and the second to the TBE metric. Asterisks indicate nodes supported with a BS and TBE of 100 and 1, respectively. Names of species corresponding to
genera other than Helichrysum are in bold. Malagasy clades highlighted in grey and labeled MAD1 to MAD6 are shown in pairs across panels: (a) includes clades
MAD5 and MAD6; (b) includes clades MAD3 and MAD4; (c) includes clades MAD1 and MAD2. Humbert’s (1962) taxonomic groups are indicated in blue. Unrelated
African clades are collapsed and their geographical distribution is indicated to the right. The complete tree obtained from the ancestral range estimation analyses is
shown in Supplementary Figure S6.
C. Blanco-Gavald`
a et al.

Fig. 2. (continued).
C. Blanco-Gavald`
a et al.

Dispersal-Vicariance model (DIVAlike; Ronquist 1997), and the BayArea
model (BAYAREAlike; Landis et al. 2013). We also tested a more com-
plex version of each model that accounts for founder-event speciation by
adding the jump-dispersal parameter (+j). We compared and chose the
best tting model based on the Akaike Information Criterion (AIC) and
AIC weight (AICw), considering that the comparison of DEC and DEC +j
Fig. 2. (continued).
C. Blanco-Gavald`
a et al.

models is statistically valid based on Matzke’s (2022) reply to Ree &
Sanmartín (2018). We also performed Biogeographic Stochastic Map-
ping (BSM; Dupin et al. 2017) on the best tting model (DEC +j) to
estimate the frequency and types of biogeographical events, taking the
mean and standard deviation of event counts from 100 BSM replicates.
3. Results
3.1. Molecular data processing
Out of the 1061 loci targeted, we initially recovered 929 loci. Based
on pairwise sequence divergence histograms (Supplementary Figure
S2), we estimated that values of 7.0 to 19.0 % of divergence indicated
paralogy. Using the Helichrysum-customized reference tailored to our
dataset, 228 (±58) paralogous loci were detected on average
(Supplementary Table S3). After ltering for missing data and splitting
alignments containing paralogs, which resulted in 322 new alignments,
we performed phylogenetic inference with a total of 971 loci. The
aligned length of each locus averaged 278 bp (ranging from 37 to 735
bp, see Supplementary Table S4). Each locus had on average 77 (ranging
from 4 to 264) parsimony informative sites and 113 (ranging from 5 to
383) variable sites. The average proportion of missing data was 2.8 %
(ranging from 0 to 65 %). Concatenation of all loci resulted in a super-
matrix with 295610 bp and 384 taxa.
3.2. Phylogenetic analyses
The topologies of the phylogenetic trees inferred using the concate-
nation approach (hereafter, ML tree, collapsed version Fig. 2 and full
version Supplementary Figure S3) and the summary-coalescence
approach (hereafter, ASTRAL tree, Supplementary Figure S4) based on
971 loci were congruent except for a few intermediate nodes, which are
supported in both trees but show different relationships between spe-
cies. In particular, two montane Malagasy clades are recovered as close
relatives in the ML tree, but distant in the ASTRAL tree (ML tree node
367). In both trees, there are six lineages constituted by Malagasy spe-
cies (MAD1-MAD6, as shown in Fig. 2). Three of them are represented by
a single species −Helichrysum mahalafy (MAD1), H. plantago DC. (MAD3)
and H. madagascariense DC.(MAD5)-, each of them sister to an African
species; a fourth one is constituted by four Malagasy species and a South
African species; and the rest of Malagasy species constitute two speciose
lineages, MAD2 and MAD6, the latter including some African species.
Lower overall clade support was obtained with the summary-
coalescence approach: 61 % of the nodes received signicant support
(LPP ≥0.95), whereas the percentage of signicantly supported nodes in
the ML tree was 85 % and 96 % according to BS and TBE metrics (≥70
%), respectively. The summary-coalescence analysis based on the 25
most informative loci resulted in a tree (Supplementary Figure S5) that
is congruent with the former ASTRAL tree, but even overall lower sup-
ports were obtained (only 25 % of the nodes received signicant sup-
port). Because of that, we will not further discuss the results of this tree.
3.3. Biogeographic reconstruction
The best-tting biogeographical model was DEC +j according to AIC
values (Table 1). Regarding dispersal types, cladogenetic founder-event
processes (j =0.0085) had a slightly larger contribution than anagenetic
range expansions (d =0.0076).
Ancestral range reconstruction analyses (probabilities in Supple-
mentary Table S5, full tree in Supplementary Figure S6) suggest that the
genus most probably originated and initially diversied in the winter
rainfall regions of western South Africa (Fynbos Biome, F and arid to
semi-arid southern Africa, N) around 11.6 mya (11.4 – 15.0 95 % CI),
although the ancestral range may have also included some of the sur-
rounding areas (the southern African grasslands, G; and/or the high
Drakensberg area, D). The largest lineage in our tree notably diversied
in the southern African grasslands and dispersals northwards and east-
wards occurred from the late Miocene onwards (c. 8 mya, 6.5 – 11.7 95
% CI, node 356). Repeated dispersals gave rise to the current global
distribution of Helichrysum.
Within Madagascar, Biogeographic Stochastic Mapping (BSM) ana-
lyses reveal at least six founder dispersal events from continental Africa
to Madagascar (Fig. 3; BSM summaries in Supplementary Table S6), all
of them giving rise to endemic lineages. The main geographic source of
Madagascar’s colonization events was the southern African grasslands
(G, with a mean of 3 colonization events) followed by the Tropical
Afromontane area (T, with a mean of 2 colonization events).
According to our results, the oldest Malagasy clade (Fig. 2, clade
MAD6) is sister to Helichrysum galpinii N.E.Br., likely descending from an
ancestor occurring in the southern African grasslands that dispersed at
an inferred age of 3.7 mya (2.2 – 6.2 95 % CI, node 366, BS =100, TBE =
0.99). This clade comprises mainly Malagasy montane species found in
the highest areas (area E), followed by species found in the subhumid
northern and eastern montane belt (area R), but also includes two
Tropical Afromontane species from mainland Africa. The probabilities of
the ancestral range of the whole clade involve both Afromontane areas
(tropical continental Africa and the highest areas of Madagascar). This
clade is made up of two large and highly diversied Malagasy montane
lineages (one inferred to have originated around 3.5 mya, 2.0 – 5.9 95 %
CI, node 367, BS =99, TBE =0.99 and the other at about 2.3 mya, 1.2 –
4.5 95 % CI, node 396, BS =100, TBE =1). However, these two lineages
are not closely related to each other in the ASTRAL phylogeny. There-
fore, these results should be taken with caution since concatenation and
coalescence-based phylogenetic inferences provide conicting re-
lationships around these nodes. The two Malagasy montane clades
notably diversied during the Pleistocene, with some species colonizing
or extending their area into the subhumid central montane belt (area M,
e.g. H. abbayesii Humbert), the grasslands of Madagascar’s central
highlands (area H, e.g. H. lecomtei R. Vig. & Humbert and H. xylocladum
Baker) or the humid eastern lowlands (area O) in the case of H. geayi
(node 381).
The second oldest colonization of Madagascar has an inferred age of
3.6 mya (2.2 – 5.9 95 % CI, node 425, BS =100, TBE =1, MAD4), also
from a southern African grassland ancestor. The ancestral range of the
Malagasy descendant is highly uncertain due to the wide range of the
extant species belonging to this clade, which include four areas in
Table 1
Summary statistics of the biogeographic models tested in BioGeoBEARS. The best-tting model used to infer the most likely area occupied by the ancestors of Heli-
chrysum is highlighted in bold (DEC +j). Values of parameters for dispersal (d), extinction (e), founder effect (j), likelihood scores (lnL) and Akaike Information
Criterion (AIC) are provided.
Biogeographic models lnL No. free parameters d e j AIC AIC weight
DEC −1077 2 0.012 0.011 0 2157 2.6E-33
DEC þJ¡1001 3 0.0082 1.00E-12 0.0084 2007 1
DIVALIKE −1099 2 0.014 3.4E-09 0 2201 7.4E-43
DIVALIKE +J−1028 3 0.0091 1.00E-12 0.0087 2062 1.3E-12
BAYAREALIKE −1088 2 0.010 0.23 0 2181 2.10E-38
BAYAREALIKE +J−1012 3 0.0068 0.0077 0.011 2031 7.6E-06
C. Blanco-Gavald`
a et al.

Madagascar (southern and southwestern arid lowlands, area B; subhu-
mid central montane belt, area M; subhumid central highlands, area H
and dry western lowlands, area W) and one area from continental Africa
(Indian Ocean Coastal Forest Belt, area C). These results suggest a back-
colonization to the southern African coast, inferred to have occurred 2.3
mya (1.2 – 4.5 95 % CI, node 429, BS =100, TBE =1) giving rise to
H. silvaticum Hilliard.
Our results suggest that a recent dispersal from the tropical Afro-
montane area took place at an inferred age of 1.7 mya (0.9 – 3.3 95 % CI,
node 583, BS =100, TBE =1, MAD2) giving rise to a recent radiation in
the subhumid central montane belt of Madagascar (area M; ca. 0.9 mya,
0.4 – 2.0 95 % CI, node 586, BS =95, TBE =1). Again, this was followed
by several independent dispersals within Madagascar, mainly consisting
of populations establishing in lower-elevation habitats. The distribution
range of some species extended beyond the Central Highlands (e.g.
Helichrysum leucocladum Humbert into the humid eastern lowlands and
H. aphelexioides DC. into the western arid lowlands). Nevertheless, there
were also a few instances of upward dispersal to higher elevations, such
as the case of H. cremnophilum Humbert, whose distribution range
reached the highest areas.
Three other independent dispersal events to Madagascar took place,
each one giving rise to a single Malagasy species. Helichrysum mahafaly
(MAD1) from the southern and south-western arid lowlands (area B) of
Madagascar is nested within a clade of species from the geographically
distant arid to semi-arid southwestern African area (area N). This
dispersal event is inferred to have taken place 2.6 mya (1.4 – 4.7 95 %
CI, node 664, BS =100, TBE =1). The Malagasy widespread
H. madagascariense (MAD5) diverged from its sister from the southern
African grasslands around 1.8 mya (0.8 – 3.7 95 % CI, node 435, BS =
100, TBE =1). Lastly, the ancestor of H. plantago (MAD3) is also inferred
to have dispersed from the southern African grasslands at about 1.1 mya
(0.5 – 2.6 95 % CI, node 542, BS =93, TBE =0.93).
4. Discussion
Here, we explore the evolutionary dynamics of colonization and
diversication on Madagascar using the species-rich genus Helichrysum.
This is the rst time that the origin and geographic diversication of a
Malagasy plant group have been explored using a large phylogeny based
on NGS data from over 300 taxa, including more than half of the species
endemic to the island. We report at least six independent colonizations
from continental Africa from the Pliocene onwards, most of these from
the southern African grasslands and the tropical Afromontane areas. We
also recover a single back-colonization to the southeastern African
Fig. 3. Summary of mean dispersal events estimated from 100 Biogeographic Stochastic Mappings (BSM) in Helichrysum (see all event counts in Supplementary Table
S6). Arrow tips indicate the directionality of the dispersal. Numbers on the arrows are the mean of dispersal event counts (only given for dispersals related to
Madagascar). Arrow thickness is proportional to the mean number of dispersals. For events involving Madagascar, either within Madagascar or between continental
Africa and Madagascar: blue arrows represent dispersals with mean counts below 0.5; black arrows represent dispersals with mean counts between 0.5 and 4.9; red
arrows represent dispersals with mean counts ≥5. Gray arrows represent dispersal events within continental Africa and the Arabian Peninsula. For readability,
dispersals below 0.7 involving only these regions have not been represented.
C. Blanco-Gavald`
a et al.

coastal region. Our results suggest a clear tendency to retain bioclimatic
preferences and morphological characters in the colonization of
Madagascar by Helichrysum. However, our ndings indicate that once
lineages established on the island, their diversication was partly
associated with biome shifts, especially shifts from high-elevation to
lower-elevation habitats. We also identify several simultaneous and
recent in-situ radiations in the montane areas of the island, including the
subhumid central montane belt and the highest areas, which enriched
the diversity of their ora. Although some lineages are morphologically
and functionally very uniform, others include a great variety of life
forms and morphological characters.
4.1. Helichrysum colonized Madagascar multiple times from Africa since
the late Pliocene and likely returned once to the continent
Given the geological history of Madagascar, the two possible mech-
anisms by which Madagascar’s biota was established are Gondwanan
vicariance and long-distance dispersal. We infer a recent colonization of
Madagascar by Helichrysum, with the earliest dispersal event estimated
at around 3.7 mya (2.2 – 6.2 95 % CI), followed by multiple independent
colonizations until at least 1 mya (0.5 – 2.6 95 % CI). Considering that
Madagascar separated from Africa more than 150 mya, Helichrysum
must therefore have colonized the island via long-distance dispersal
across the Mozambique Channel. Short-lived land bridges across this
ancient biogeographical barrier at different geological times have been
proposed (Masters et al. 2021), but their existence has not been
demonstrated, keeping their true biogeographical impact under debate
(Ali & Hedges 2023, Aslanian et al. 2023). In any case, the last poten-
tially available land bridges would have probably disappeared before
the rst colonization of Madagascar by Helichrysum. Most Compositae
species produce anemochorous fruits, and Helichrysum cypselae are
extremely small (<1 mm long), which enables wind dispersal (Nathan
et al. 2009). Since the Miocene, dominant sea currents and winds in the
southwestern Indian Ocean region have predominantly owed west-
ward, reducing the likelihood of dispersal from Africa to Madagascar,
except through cyclones crossing the Mozambique Channel in the
opposite direction. At the same time, this shift in ocean currents and
wind directionality increased the probability of dispersal from Asia to
Madagascar, as reected in the strong oristic afnities between south-
eastern Asia, India and Madagascar, especially in the humid northern
and eastern regions (Schatz 1996, Ali & Huber 2010, Warren et al. 2010,
Buerki et al. 2013). Another plausible mechanism that could have acted
in parallel is avian dispersal, as there is evidence of direct seed dispersal
by birds, either stuck in the plumage or contained in the digestive tract
of the consumed prey (Padilla et al. 2012).
According to Buerki et al. (2013), during the Miocene there was an
increase in colonizations of Madagascar by non-endemic genera, coin-
ciding also with the emergence of most Malagasy endemic genera. Our
dating indicates a much more recent colonization of Madagascar by
Helichrysum, occurring during the Pliocene and Pleistocene. Buerki et al.
(2013) identied a few genera with similarly recent inferred coloniza-
tion dates, such as Hibiscus L. (Malvaceae), Paracorynanthe Capuron
(Rubiaceae), Colvillea Bojer and Lemuropisum H.Perrier (Fabaceae).
Another notable example within the tribe Gnaphalieae with a similar
age of colonization is the genus Stoebe L. (Bergh & Linder 2009).
Nevertheless, unlike the diversication patterns observed in Heli-
chrysum, all these genera are represented in Madagascar by few species.
The already mentioned prevailing westward currents and winds,
which have dominated since the Miocene, should have increased the
likelihood of dispersal from Madagascar to Africa, thus the potential of
back-colonizations to the mainland. However, documented examples of
such events are scarcely found in the literature. Some notable exceptions
include Dombeyoideae (Malvaceae s.l.) with at least ve migrations
back to the continent (Skema et al. 2023), the Coffeeae and Psychotrieae
alliances (Rubiaceae, Kainulainen et al. 2017; Razamandimbison et al.
2017), Croton (Euphorbiacee, Haber et al. 2017) and grammitid ferns
(Bauret et al. 2017). Here, we inferred at least one potential back-
colonization from Madagascar to continental Africa in Helichrysum,
involving H. silvaticum, which occurs in the Indian Ocean Coastal Forest
Belt (area C, in Mozambique and South Africa). This biome, separated
from Madagascar by the 415 km wide Mozambique Channel (Masters
et al. 2021), is the geographically closest continental biome to the is-
land. This species is sister to H. leucosphaerum Baker which inhabits
Madagascar’s subhumid central highlands (area H) and the dry western
lowlands (area W). Both species belong to the MAD4 clade (Fig. 2),
which originated from a dispersal from the southern African grassland
around 3.6 mya (2.2 – 5.9 95 % CI), giving rise to ve species, four of
which are endemic to Madagascar. The most likely scenario in our
reconstruction is a single dispersal to Madagascar, followed by diversi-
cation and a back-colonization of the ancestor of H. silvaticum. How-
ever, uncertainty at nodes 426 and 429 suggests an alternative
possibility in which the ancestor of both species remained in the conti-
nent, in which case H. leucosphaerum would be the result of a second,
independent colonization of Madagascar within the MAD4 clade. Given
the clade’s high morphological diversity, increased taxonomic sampling
–particularly of continental African species, since Madagascar mor-
photypes and taxonomic groups are all represented– could reduce un-
certainty regarding the biogeographic and evolutionary history of this
lineage.
4.2. Biome conservatism prevails in the colonization of Madagascar by
Helichrysum: The continental Afromontane region reveals as the main
source of Malagasy lineages
Our results for Helichrysum are consistent with previous ndings of
biome conservatism following transoceanic dispersal (Crisp et al. 2009,
Vences et al. 2009) as well as the maintenance of biome-related adap-
tations (Wiens & Graham 2005, Wiens et al. 2010). Specically, we infer
that ve out of the six hypothesized independent colonization events of
Madagascar by Helichrysum happened in the island’s highlands and
mountains by ancestors from the continental Afromontane region s.l.,
which apart from the tropical Afromontane area also includes the high
southern African grasslands in the Drakensberg mountains. This sug-
gests that Helichrysum dispersed from areas with similar broad-scale
climatic conditions in continental Africa, reinforcing the idea that
long-distance dispersal to Madagascar by Helichrysum was associated
with biome conservatism. In greater detail, we identify two colonization
events from the southern African montane grasslands to the Malagasy
montane grasslands. In these cases, the colonizers preserved not only
broad-scale bioclimatic preferences but also morphological traits. For
example, H. plantago (MAD3) shares traits with its sister species
H. nudifolium (L.) Less (Fig. 4, more details in Table S7) and other closely
related species. Likewise, H. madagascariense (MAD5) is strikingly
similar to the species comprising its sister group, H. dasycephalum O.
Hoffm. and H. rutilans D.Don, and to H. callicomum Harv., the sister to all
three (Fig. 4, more details in Table S7).
Malagasy grasslands, now extensive, have a complex history, and
whereas some grassland formations are ancient, some others are much
more recent and anthropogenic (Vorontsova et al. 2016, Hackel et al.
2018, Joseph & Seymour 2020, Solofondranohatra et al. 2020, Joseph
et al. 2021, Lehmann et al. 2022, Bond et al. 2023, Tiley et al. 2024b).
Thus, some form of grassland environment was undoubtedly already
available for Helichrysum with relevant adaptations to “dispersify”
(Donoghue 2008) from similar African environments. Dispersals from
southern Africa to Madagascar have been reported for multiple grasses
and sedge clades (e.g. Linder et al. 2014, Larridon et al. 2021). However,
little is known about non-graminoid grassland species, although
Alchemilla L. (Gehrke et al. 2016) and Stoebe (Bergh & Linder 2009),
currently found in the highest areas of Malagasy mountains, probably
dispersed from the montane southern African grasslands.
Two ancestors from the tropical Afromontane area successfully
established and later radiated in the mountains of Madagascar, one
C. Blanco-Gavald`
a et al.

radiation resulting in most of the current species growing in the sub-
humid central montane belt (area M, Fig. 2, clade MAD2) and the other
resulting in most of the species occurring in the highest areas (area E,
Fig. 2, clade MAD6). The Tapia woodland mosaics in the subhumid
central montane belt are structurally equivalent to the miombo wood-
land savannas of continental Africa (Alvarado et al. 2014) and the iconic
montane ericoid shrublands found in Africa display similar physiog-
nomy to those growing on the summits of the highest Malagasy moun-
tains (Silander et al. 2024). Examples of Afromontane migrations from
tropical Africa to Madagascar involve several grass clades (Vorontsova
et al. 2016, Hackel et al. 2018), sedge lineages (Larridon et al. 2021),
Kniphoa Moench (Ramdhani et al. 2009), and the diverse Erica L. (Pirie
et al. 2019). In the case of clade MAD2, morphological traits have been
conserved in this jump between montane biomes in Africa and
Madagascar by Helichrysum (Fig. 4, detailed information in Table S7).
Lastly, we propose an alternative interpretation of the ancestral
range reconstruction for the clade MAD6 (Fig. 2). According to our ML
phylogeny, Helichrysum galpinii (inhabiting the southern African grass-
lands, Fig. 2, node 366) is sister to two Malagasy lineages, suggesting
that the entire clade likely originated from an ancestor in the southern
African grasslands. This clade also includes a couple of tropical Afro-
montane species, introducing uncertainty about the biogeographic
range of the ancestor of these two Malagasy montane lineages. However,
our ASTRAL phylogeny (Supplementary Figure S4) presents a different
topology, where the two Malagasy montane clades are recovered as
unrelated: one is sister to H. galpinii and the other is sister to
H. whyteanum Britten and H. brunioides Moeser. This topology thus
suggests two independent colonizations of Madagascar’s montane area,
one from the southern African grasslands and the other from the tropical
Afromontane region. Indeed, the morphological distinctness of these
two Malagasy montane lineages is consistent with the hypothesis of two
independent colonizations.
Finally, we would like to note a striking case of conservatism in
bioclimatic preferences in the colonization of Madagascar by the
ancestor of Helichrysum mahafaly (Fig. 2, clade MAD1). This species,
which is endemic to the extremely hot and arid southwestern part of the
island, is nested within a southern African clade from the western arid to
semi-arid regions. In other words, H. mahafaly occupies a habitat char-
acterized by environmental stressors strikingly similar to those experi-
enced by its sister species in southern Africa (H. argyrosphaerum, Fig. 4).
Other examples of Malagasy plants adapted to arid environments and
likely originating from African ancestors include baobabs (Adansonia L.,
Leong Pock Tsy et al. 2009), Portulacaceae members (Eggli 1997,
Hershkovitz & Zimmer 2000) and Neoapaloxylon Rauschert (Legumi-
noseae, Choo et al. 2020).
4.3. Diversication within Madagascar
Over time, lineages that colonize new territories, such as islands, can
lead to various evolutionary outcomes. These may include limited or
extensive speciation, remaining relatively unchanged or adapting to
new conditions. In Helichrysum, we observed instances of different
diversication patterns. Some lineages have speciated considerably. The
timing of colonization and speciation on Madagascar by the ancestors of
these lineages coincided with the transition towards the so-called
“Icehouse Climate State” (Westerhold et al. 2020). The cooling trend
resulted in increased aridity, pronounced seasonality and lower atmo-
spheric CO
2
concentrations, which have been linked to several plant
radiations (especially in the two most species-rich grassland families
Poaceae and Compositae, Palazzesi et al. 2022) in the open grassland
biomes that had proliferated worldwide during the Miocene (Spriggs
et al. 2014). In parallel, ocean circulation patterns changed around 3 – 4
mya, resulting in a decrease in rainfall and increased aridity in both East
Africa and Madagascar (De Wit 2003). Most speciation events leading to
extant Helichrysum species are inferred to have occurred during the
Pleistocene, as reported for other Malagasy groups such as orchid genera
Fig. 4. Species of Helichrysum illustrating the similarity between closely related
continental Africa and Malagasy taxa. First row: (a) Helichrysum argyrosphaerum
from South Africa, which is the inferred sister to (b) Helichrysum mahafaly
(MAD1) from Madagascar. Both are prostrate herbs with solitary, subglobose
capitula, inhabiting arid sandy habitats. Second row: (c) Helichrysum forskahlii
var. compactum a tropical Afromontane species closely related to taxa from the
Malagasy clade MAD2; (d) Helichrysum fulvescens, a Malagasy species from
clade MAD2, mostly subshrubs with numerous small cylindrical capitula ar-
ranged in corymbs and mbrilliferous receptacles. Third row: (e) Helichrysum
nudifolium var. pilosellum from South Africa, inferred sister species of Heli-
chrysum plantago (f) and (g) from Madagascar (MAD3). Both are perennial herbs
with basal leaf rosette and dense corymbs of capitula with yellow involucral
bracts. Fourth row: (h) Helichrysum rutilans, a southern African species closely
related to (i) Helichrysum madagascariense (MAD5) from Madagascar, both
subshrubs with dense corymbs of narrow cylindrical capitula. (Photos: Merc`
e
Galbany-Casals, except (a) and (e): Marinda Koekemoer).
C. Blanco-Gavald`
a et al.

Fig. 5. Morphological diversity of clade MAD6. (a) Helichrysum mutisiifolium (GIX), a lianioid species; (b) Helichrysum chamaeyucca, (e) Helichrysum marojejyense, and
(f) Helichrysum danguyanum, the three from GIII, all with long white radiant involucral bracts; (c) Helichrysum ibityense var. ibityense (GII) and (d) Helichrysum
hypnoides (GIV), both with ericoid leaves; (g) Helichrysum gymnocephalum, (h) Helichrysum xylocladum and (i) Helichrysum geniorum, the three from GI, shrubs or
treelets with tiny capitula grouped in glomerules surrounded by leaves and gathered in big corymbose synorescences. (Photos: Merc`
e Galbany-Casals).
C. Blanco-Gavald`
a et al.

(Aeranthes Lindl., Angraecum Bory and Jumellea Schltr., Andriananja-
manantsoa et al. 2016) and scaly tree ferns (Cynthaeaceae, Janssen et al.
2008).
As is common for many groups that have diversied in archipelagos
(Rundell & Price 2009), Helichrysum likely experienced a combination of
adaptive and non-adaptive radiations (summarized in Table S7). Some
clades such as the clade MAD4 (Figure S7) and the youngest subclade in
MAD6 (Fig. 5), comprise ecologically and morphologically well-
differentiated species. This heterogeneity suggests ecological specia-
tion (i.e. reproductive isolation arises through divergent natural selec-
tion in populations adapted to different ecological environments;
Rundle & Nosil 2005) as the main evolutionary force. Similar evolu-
tionary patterns have been reported for other groups, such as the radi-
ation of the small Afromontane genus Arrowsmithia DC. (formerly
Macowania Oliv., also belonging to the tribe Gnaphalieae) in the Dra-
kensberg mountains (Bentley et al. 2014). Nonetheless, members of the
more species-rich clades within Helichrysum tend to be morphologically
uniform, making them easily identiable as a cohesive group. One of
such examples is a subclade in MAD6, which comprises all sampled
species from Humbert’s taxonomic group I (GI, Humbert 1962). Despite
their morphological similarities (Fig. 5), species in this group show
notable variation in geographic distribution and ecological preferences.
This suggests that diversication in GI likely resulted from a combina-
tion of geographic (or allopatric) and ecological speciation events.
However, despite the occurrence of several shifts in broad-scale biocli-
matic preferences, the overall pattern within this clade points towards
biome conservatism. Each of the two major subclades within GI has a
distinct primary distribution area (see Fig. 2): one diversied in the
northern and eastern montane belt, while the other diversied on
mountain summits. A similar pattern is observed in MAD2, which in-
cludes all sampled species of taxonomic group XI (GXI, Humbert 1962).
Although the group is morphologically well-dened (more details in
Table S7), it exhibits considerable variation in growth form, capitula
size, bract color, and ecological preferences.
Our ndings indicate that biome shifts likely occurred in
Madagascar, following speciation within the colonized area and/or
expansion of the distribution range across Madagascar’s biomes. Most
shifts in bioclimatic preferences in Malagasy Helichrysum are associated
with independent dispersals from higher to lower elevation environ-
ments. The colder glacial periods of the Pleistocene may have promoted
this transition to mid-low elevations. For instance, the elevational range
of many species in group XI (GXI, clade MAD2) appears to have
expanded downward to the highland grasslands (e.g. H. fulvescens DC.
and H. viguieri Humbert), while multiple species that are currently
restricted to the highlands’ grasslands have montane ancestors. An
example is the clade comprising H. hirtum Humbert, H. tenue Humbert
and H. heterotrichum Humbert (GXI), in which biome conservatism
prevailed after an initial shift to lower elevations. Exceptionally, a few
montane species have colonized the dry western lowlands (e.g.
H. triplinerve DC) or the humid eastern lowlands (e.g. H. geayi), two bi-
omes with few Helichrysum species. Such repeated downward migrations
are infrequently reported in the literature (e.g. Dendrosenecio (Hauman
ex Hedberg) B.Nord., Knox & Palmer 1995), the general evolutionary
trend being migration from lower to higher elevations (see Gamisch
et al. 2016 and references therein).
Lastly, two of the three Malagasy lineages represented by a single
species, and descending from African ancestors, are morphologically
distinct from all other Malagasy Helichrysum species. This was recog-
nized in the most recent taxonomic treatment (Humbert 1962), which
placed these species into monospecic taxonomic groups: H. mahafaly in
group VI (Fig. 4, GVI, clade MAD1) and H. plantago in group X (Fig. 4,
GX, clade MAD3). This is not surprising as these species closely resemble
their relatives in continental Africa, both morphologically and ecologi-
cally. Even if our sampling is not complete, we hypothesize that these
two species are the only extant descendants of their respective colo-
nizing ancestors. The lack of diversication in the lineage represented by
H. mahafaly could be attributed to its highly specialized niche, which
results in geographical isolation to the southwestern coast. Conversely,
H. plantago is a widespread and opportunistic species that thrives in the
heavily degraded highlands. Its recent origin and its ability to exploit
this transformed landscape, which lacks signicant geographical bar-
riers, may have allowed it to spread widely, and all these factors may
explain lack of further diversication in this clade. The sole member of
the third lineage, H. madagascariense (Fig. 2, GVII, clade MAD5, see
Fig. 4) was classied by Humbert (1962) in group VII together with
three other species. However, our tree shows that group VII is not
monophyletic, as H. leucosphaerum (node 429) is placed in a separate
Malagasy clade containing species from different taxonomic groups
(clade MAD4 in Fig. 2, see Figure S7). Without sampling the other two
members of group VII, we cannot conrm whether H. madagascariense
truly constitutes a monospecic lineage.
5. Conclusions
This study has focused on the Malagasy radiations of the highly
diverse plant genus Helichrysum to unravel its biogeographic and
evolutionary history within this long-isolated fragment of Gondwana.
Our ndings reveal a scenario of repeated recent transoceanic dispersal
events, in which biome conservatism played a critical role in shaping
speciation and diversication of Malagasy lineages. Some colonization
events were followed by in-situ diversication, likely driven by a com-
bination of allopatric and ecological speciation processes. The primary
source of Malagasy lineages can be traced to the Afromontane region s.l.,
including the highest mountains of tropical Africa and southern Africa.
The ancestors were probably preadapted to high-elevation environ-
ments, which facilitated the colonization of the island’s montane areas,
harboring now most of the genus’ diversity. Once established, the de-
scendants of these African ancestors appear to have radiated in-situ,
possibly through geographic speciation, giving rise to several high-
elevation Malagasy endemics distributed across different massifs. In
addition, our inferences suggest that some montane species shifted their
ranges to lower-elevation environments, such as the highlands’ grass-
lands, western dry deciduous forests or eastern evergreen forests. In
some cases, these range shifts resulted in expanded distribution ranges,
while in others, they promoted further speciation. In contrast, Malagasy
clades coming from southern African grasslands ancestors consist of one
or few widespread species. Notably, we identied a rare example of
conservatism in bioclimatic preferences after long-distance dispersal
from the arid southwestern Africa to Madagascar’s similarly arid
southwestern coastal dunes. Our ndings also reveal a potential case of
back-dispersal by Helichrysum to the southeastern coast of continental
Africa, supporting the idea of sporadic biogeographical dispersals from
Madagascar to mainland Africa.
CRediT authorship contribution statement
Carme Blanco-Gavald`
a: Writing – original draft, Methodology,
Investigation, Formal analysis, Data curation, Conceptualization. Cris-
tina Roquet: Writing – review & editing, Methodology, Investigation,
Formal analysis, Data curation, Conceptualization. Genís Puig-Surroca:
Writing – review & editing, Resources, Investigation, Formal analysis,
Data curation. Santiago Andr´
es-S´
anchez: Writing – review & editing,
Resources, Methodology, Investigation. Sylvain G. Razamandimbi-
son: Writing – review & editing, Resources, Methodology, Investigation.
Rokiman Letsara: Writing – review & editing, Resources. Nicola
Bergh: Writing – review & editing, Resources, Investigation. Glynis V.
Cron: Writing – review & editing, Resources. Lucía D. Moreyra: Vali-
dation, Software. Juan Antonio Calleja: Writing – review & editing,
Resources, Investigation. `
Oscar Castillo: Writing – review & editing,
Investigation, Data curation. Randall J. Bayer: Resources. Frederik
Leliaert: Writing – review & editing, Resources. Alfonso Susanna:
Writing – review & editing, Resources, Methodology, Investigation,
C. Blanco-Gavald`
a et al.

Conceptualization. Merc`
e Galbany-Casals: Writing – review & editing,
Resources, Project administration, Methodology, Investigation, Funding
acquisition, Formal analysis, Data curation, Conceptualization.
Funding
This work received nancial support from the Spanish Ministry of
Science, Innovation and Universities (PD2019-105583GB-C22/AEI/
10.13039/501100011033) and the Catalan government (“Ajuts a grups
consolidats” 2021SGR00315 and Ph.D. FI-AGAUR grant to C.B.-G.
2022FI_B00150).
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgements
The authors thank María Luisa Guti´
errez for providing technical
support during the laboratory process and curators and technicians of
the herbaria who provided material for the study as well as all collab-
orators who provided material from their own collections. We thank
Sebasti´
an Arrabal for his huge help and fantastic company in the eld-
work in Madagascar. We thank the Direction des Aires Prot´
eg´
ees, des
Ressources Naturelles renouvellables et des Ecosyst`
emes, and the
Madagascar National Parks in Madagascar for issuing collecting permits
for M.G.-C. and S.G.R.; Parc Botanique et Zoologique de Tsimabazaza;
Missouri Botanical Garden, Madagascar; and especially Faranirina
Lantoarisoa for logistical support and arranging collecting permits for
M.G.-C and S.G.R. Collection in Rwanda was done under Research
permit No NCST/482/304/2022 and with the permission of Rwanda
Development Board (RDB) Tourism and Conservation to collect in the
Volcanoes National Park. We thank Prof. Elias Bizuru for allowing
afliation of M.G.-C. and J.A.C. to University of Rwanda, Dr. Richard
Muvunyi from RDB for his support, the staff of the Volcanoes National
Park for eld assistance; and Beth Kaplin and the staff from the National
Herbarium of Rwanda, as well as the staff of the Ellen DeGeneres
Campus in Musanze, for logistic support. Collection in South Africa was
done under research permits CN35-28-23663 (Western Cape Nature
Conservation Board) and HO/RSH/47/2021 (Province of Eastern Cape,
Economic Development, Environmental Affairs & Tourism). We also
thank Arne Anderberg for the identication of some specimens from
Madagascar and Marinda Koekemoer for providing some pictures of
South African Helichrysum. We appreciate the suggestions of an anony-
mous reviewer. Finally, we profusely thank Porter P. Lowry for his
detailed revision of the manuscript that substantially contributed to
improving the text.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.ympev.2024.108283.
Data availability
The Sequence Read Archive (SRA) identier for each sample can be
found in the Supplementary Table S1 as well as the link to the repository.
References
Ali, J.R., Hedges, S.B., 2023. The colonisation of Madagascar by land-bound vertebrates.
Biol. Rev. 98, 1583–1606. https://doi.org/10.1111/brv.12966.
Alvarado, S.T., Buisson, E., Rabarison, H., Rajariarisonm, C., Birkinsha, C., Lowry II, P.P.,
2014. Comparision of plant communities on the Ibity and Itremo massifs,
Madagascar, with contrasting conservation histories and current status. Plant Ecol.
Divers. 7 (4), 497–508. https://www.doi.org/10.1080/17550874.2013.804131.
Andr´
es-S´
anchez, S., Verboom, G.A., Galbany-Casals, M., Bergh, N.G., 2019. Evolutionary
history of the arid climate-adapted Helichrysum (Asteraceae: Gnaphalieae): Cape
origin and association between annual life-history and low chromosome numbers.
J. Syst. Evol. 57 (5), 468–487. https://doi.org/10.1111/jse.12472.
Andriananjamanantsoa, H.N., Engberg, S., Louis, E.E., Jr., Brouillet, L., 2016.
Diversication of Angraecum (Orchidaceae, Vandeae) in Madagascar: revised
phylogeny reveals species accumulation through time rather than rapid radiation.
PloS one, 11(9), e0163194. doi: 10.1371/journal.pone.0163194.
Antonelli, A., Smith, R.J., Perrigo, A.L., Crottini, A., et al., 2022. Madagascar’s
extraordinary biodiversity: evolution, distribution and use. Science 378 (6623),
eabf0869. https://doi.org/10.1126/science.abf0869.
Aslanian, D., Pellen, R., Rabineau, M., Moulin, M., Leroux, E., Delaunay, A., Baby, G.,
Courgeon, S., Linol, B., Suc, J.P., Popescu, S., Fauquette, S., Mazza, P.P.A.,
Couette, S., et al., 2023. The postulation of intermittent land bridges as an
explanation for reiterated colonization events of Madagascar by African vertebrates:
an in-depth review and novel insights in honour of the late Judith Masters and
Fabien G´
enin. Earth-Sci. Rev. 246, 104585. https://doi.org/10.1016/j.
earscirev.2023.104585.
Bankevich, A., Nurk, S., Antipov, D., Gurevich, A.A., Dvorkin, M., Kulikov, A.S., Lesin, V.
M., Nikolenko, S.I., Pham, S., Prjibelski, A.D., Pyshkin, A.V., Sirotkin, A.V.,
Vyahhi, N., Tesler, G., Alekseyev, M.A., Pevzner, P.A., 2012. SPAdes: a new genome
assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19
(5), 455–477. https://doi.org/10.1089/cmb.2012.0021.
Bauret, L., Gaudeul, M., Sundue, M.A., Parris, B.S., Ranker, T.A., Rakotondrainibe, F.,
Hennequin, S., Ranaivo, J., Selosse, M.A., Rouhan, G., 2017. Madagascar sheds new
light on the molecular systematics and biogeography of grammitid ferns: new
unexpected lineages and numerous long-distance dispersal events. Mol. Phylogenet.
Evol. 111, 1–17. https://doi.org/10.1016/j.ympev.2017.03.005.
Bentley, J., Verboom, G.A., Bergh, N.G., 2014. Erosive processes after tectonic uplift
stimulate vicariant and adaptive speciation: evolution in an Afrotemperate-endemic
paper daisy genus. BMC Evol. Biol. 14, 27. https://doi.org/10.1186/1471-2148-14-
27.
Bergh, N.G., Linder, H.P., 2009. Cape diversication and repeated out-of-southern-Africa
dispersal in paper daisies (Asteraceae-Gnaphalieae). Mol. Phylogenet. Evol. 51 (1),
5–18. https://doi.org/10.1016/j.ympev.2008.09.001.
Blanco-Gavald`
a, C., Galbany-Casals, M., Susanna, A., Andr´
es-S´
anchez, S., Bayer, R.J.,
Brochmann, C., Cron, G.V., Bergh, N.G., Garcia-Jacas, N., Gizaw, A., Kandziora, M.,
Kol´
aˇ
r, F., L´
opez-Alvarado, J., Leliaert, F., Letsara, R., Moreyra, L.D.,
Razamandimbison, S.G., Schmickl, R., Roquet, C., 2023. Repeatedly northwards
and upwards: southern African grasslands fuel the colonization of the African sky
islands in Helichrysum (Compositae). Plants 12 (11), 2213. https://doi.org/
10.3390/plants12112213.
Bolger, A.M., Lohse, M., Usadel, B., 2014. Trimmomatic: a exible trimmer for Illumina
sequence data. Bioinformatics 30 (15), 2114–2120. https://doi.org/10.1093/
bioinformatics/btu170.
Bond, W.J., Silander, J.A., Ratsirarson, J., 2023. Madagascar’s grassy biomes are ancient
and there is much to learn about their ecology and evolution. J. Biogeogr. 50,
614–621. https://doi.org/10.1111/jbi.14494.
Buerki, S., Devey, D.S., Callmander, M.W., Phillipson, P.B., Forest, F., 2013. Spatio-
temproal history of the endemic genera of Madagascar. Bot. J. Linn 171 (2),
304–329. https://doi.org/10.1111/boj.12008.
Burgess, N., 2004. Terrestrial ecoregions of Africa and Madagascar: a conservation
assessment. Island Press, Washington, DC, USA.
Burgoyne, P.M., van Wyk, A.E., Anderson, J.M., Schrire, B.D., 2005. Phanerozoic
evolution of plants on the African plate. J. Afr. Earth Sci. 43 (1–3), 13–52. https://
doi.org/10.1016/j.jafrearsci.2005.07.015.
Bushnell, B., 2014. BBMap: A fast, accurate, splice-aware aligner. www. ourceforge.net/
projects/bbmap/.
Callmander, M.W., Phillipson, P.B., Schatz, G.E., Andriambololonera, S.,
Rabarimanarivo, M., Rakotonirina, N., Raharimampionona, J., Chatelain, C.,
Gautier, L., Lowry, P.P., 2011. The endemic and non-endemic vascular ora of
Madagascar updated. Plant Ecol. Evol. 144 (2), 121–125. https://doi.org/10.5091/
plecevo.2011.513.
Cardillo, M., Weston, P.H., Reynolds, Z.K.M., Olde, P.M., Mast, A.R., Lemmon, E.M.,
Lemmon, A.R., Bromham, L., 2017. The phylogeny and biogeography of Hakea
(Proteaceae) reveals the role of biome shifts in a continental plant radiation.
Evolution 71, 1928–1943. https://doi.org/10.1111/evo.13276.
Choo, L.M., Forest, F., Wieringa, J.J., Bruneau, A., de la Estrella, M., 2020. Phylogeny
and biogeography of the Daniellia clade (Leguminosae: Detarioideae), a tropical tree
lineage largely threatened in Africa and Madagascar. Mol. Phylogenet. Evol. 146,
106752. https://doi.org/10.1016/j.ympev.2020.106752.
Costa, F.P., Schrago, C.G., Mello, B., 2022. Assessing the relative performance of fast
molecular dating methods for phylogenomic data. BMC Genomics 23, 798. https://
doi.org/10.1186/s12864-022-09030-5.
Crisp, M., Arroyo, M., Cook, L., Gandolfo, M.A., Jordan, G.J., McGlone, M.S., Weston, P.
H., Westoby, M., Wilf, P., Linder, H.P., 2009. Phylogenetic biome conservatism on a
global scale. Nature 458, 754–756. doi: 10.1038/nature07764.
Darriba, D., Posada, D., Kozlov, A.M., Stamatakis, A., Morel, B., Flouri, T., 2020.
ModelTest-NG: a new and scalable tool for the selection of DNA and protein
evolutionary models. Mol. Biol. Evol. 37 (1), 291–294. https://doi.org/10.1093/
molbev/msz189.
Darwin, C., 1859. On the origin of species by means of natural selection or preservation
of favoured races in the struggle for life. John Murray, London.
C. Blanco-Gavald`
a et al.

De Wit, M.J., 2003. Madagascar: heads it’s a continent, tails it’s an island. Annu. Rev.
Earth Planet. Sci. 31 (1), 213–248.
Donoghue, M.J., Edwards, E.J., 2014. Biome shifts and niche evolution in plants. Annu.
Rev. Ecol. Evol. Syst. 45, 547–572. https://doi.org/10.1146/annurev-ecolsys-
120213-091905.
Donoghue, M.J. (2008). A phylogenetic perspective on the distribution of plant diversity.
Proc. Natl. Acad. Sci. U.S.A., 105, 11549–11555. doi: 10.1073/Proc. Natl. Acad. Sci.
U.S.A..0801962105.
Du Puy, D.J., Moat, J., 1996. A rened classication of the primary vegetation of
Madagascar based on the underlying geology: using GIS to map its distribution and
to assess its conservation status. In: Lourenço, W.R. (ed.). Biogeography of
Madagascar. ORSTROM, Paris.
Dupin, J., Matzke, N.J., S¨
arkinen, T., Knapp, S., Olmstead, R.G., Bohs, L., Smith, S.D.,
2017. Bayesian estimation of the global biogeographical history of the Solanaceae.
J. Biogeogr. 44, 887–899. https://doi.org/10.1111/jbi.12898.
Eggli, U., 1997. A synopsis of woody Portulacaceae in Madagascar. Adansonia S´
er. 3
(19), 45–59.
Faramalala, M.H., Conservation International (1995). Formations V´
eg´
etales et Domaine
Forestier National de Madagascar.
Felsenstein, J., 1985. Condence limits on phylogenies: an approach using the bootstrap.
Evolution 39 (4), 783–791. https://doi.org/10.1111/j.1558-5646.1985.tb00420.x.
F´
er, T., Schmickl, R.E., 2018. HybPhyloMaker: target enrichment data analysis from raw
reads to species trees. Evol. Bioinform. Online 14, 1176934317742613. https://doi.
org/10.1177/1176934317742613.
Galbany-Casals, M., Garcia-Jacas, N., S´
aez, L., Benedí, C., Susanna, A., 2009. Phylogeny,
biogeography and character evolution in Mediterranean, Asiatic and Macaronesian
Helichrysum (Asteraceae, Gnaphalieae) inferred from nuclear phylogenetic analyses.
Int. J. Plant Sci. 170 (3), 365–380. https://doi.org/10.1086/596332.
Galbany-Casals, M., Unwin, M., Garcia-Jacas, N., Smissen, R.D., Susanna, A., Bayer, R.J.,
2014. Phylogenetic relationships in Helichrysum (Compositae: Gnaphalieae) and
related genera: Incongruence between nuclear and plastid phylogenies,
biogeographic and morphological patterns, and implications for generic
delimitation. Taxon 63, 608–624. https://doi.org/10.12705/633.8.
Gamisch, A., Fischer, G.A., Comes, H.P., 2016. Frequent but asymmetric niche shifts
in Bulbophyllum orchids support environmental and climatic instability in
Madagascar over Quaternary time scales. BMC Evol. Biol. 16, 14. https://doi.org/
10.1186/s12862-016-0586-3.
Gamisch, A., Winter, K., Fischer, G.A., Comes, H.P., 2021. Evolution of crassulacean acid
metabolism (CAM) as an escape from ecological niche conservatism in
Malagasy Bulbophyllum (Orchidaceae). New Phytol. 231, 1236–1248. https://doi.
org/10.1111/nph.17437.
Gautier, L.G., Tahinarivony, J.A., Ranirison, P., Hohlhauser, S., 2018. Vegetation. In:
Goodman, S.M., M.J. Raharilalao & S. Wohlhauser (eds). The Terrestrial Protected
Areas of Madagascar: Their History, Description, and Biota. Vol. 1., pp. 207–242
Assn. Vahatra, Antananarivo, Madagascar.
Gehrke, B., Kandziora, M., Pirie, M.D., 2016. The evolution of dwarf shrubs in alpine
environments: a case study of Alchemilla in Africa. Ann. Bot. 117 (1), 121–131.
https://doi.org/10.1093/aob/mcv159.
Goodman, S.M., Benstead, J.P. (eds.), 2003. The Natural History of Madagascar.
University of Chicago Press, Chicago and London.
Graham, N., Gruner, D., Lim, J., Gillespie, R., 2017. Island ecology and evolution:
challenges in the anthropocene. Environ. Conserv. 44 (4), 323–335. https://doi.org/
10.1017/S0376892917000315.
Haber, E.A., Kainulainen, K., Van Ee, B.W., Oyserman, B.O., Berry, P.E., 2017.
Phylogenetic relationships of a major diversication of Croton (Euphorbiaceae) in
the western Indian Ocean Region. Bot. J. Linn 183 (4), 532–544. https://doi.org/
10.1093/botlinnean/box004.
Hackel, J., Vorontsova, M.S., Nanjarisoa, O.P., Hall, R.C., Razanatsoa, J., Malakasi, P.,
Besnard, G., 2018. Grass diversication in Madagascar: In situ radiation of two large
C
3
shade clades and support for a Miocene to Pliocene origin of C
4
grassy biomes.
J. Biogeogr. 45, 750–761. https://doi.org/10.1111/jbi.13147.
Hershkovitz, M.A., Zimmer, E.A., 2000. Ribosomal DNA evidence and disjunctions of
western American Portulacaceae. Mol. Phylogenet. Evol. 15 (3), 419–439. https://
doi.org/10.1006/mpev.1999.0720.
Hillis, D.M., Bull, J.J., 1993. An empirical test of bootstrapping as a method for assessing
condence in phylogenetic analyses. Syst. Biol. 42, 182–192.
Holstein, N., Renner, S.S., 2011. A dated phylogeny and collection records reveal
repeated biome shifts in the African genus Coccinia (Cucurbitaceae). BMC Evol. Biol.
11, 28. https://doi.org/10.1186/1471-2148-11-28.
H¨
orandl, E., Emadzade, K., 2011. The evolution and biogeography of alpine species
in Ranunculus (Ranunculaceae): a global comparison. Taxon 60, 415–426. https://
doi.org/10.1002/tax.602011.
Humbert, H., 1962. Flore de Madagascar et des Comores (Plantes Vasculaires); 189e
familia. Compos´
ees, part 2. Leroy, J.F. (ed.). Mus´
eum National d’Histoire Naturelle,
Paris, France.
IUCN/UNEP/WWF, 1987. Madagascar, an environmental prole. Jenkins M.D. ed.,
IUCN, Gland, Switzerland and Cambridge, U.K.
Janssen, T., Bystriakova, N., Rakotondrainibe, F., Coomes, D., Labat, J.N., Schneider, H.,
2008. Neoendemism in Malagasy scaly tree ferns results from recent, coincident
diversication bursts. Evolution 62 (8), 1876–1889. https://doi.org/10.1111/
j.1558-5646.2008.00408.x.
Joseph, G.S., Rakotoarivelo, A.R., Seymour, C.L., 2021. How expansive were Malagasy
Central Highland forests, ericoid, woodlands and grasslands? A multidisciplinary
approach to a conservation conundrum. Biol. Conserv. 261, 109282. https://doi.
org/10.1016/j.biocon.2021.109282.
Joseph, G.S., Seymour, C.L., 2020. Malagasy highlands: originally woodland and forest
containing endemic grasses, not grazing-adapted grassland. Proc. Biol. Sci. 287
(1937), 20201956. https://doi.org/10.1098/rspb.2020.1956.
Kainulainen, K., Razamandimbison, S.G., Wikstr¨
om, N., Bremer, B., 2017. Island
hopping, long-distance dispersal and species radiation in the Western Indian Ocean:
historical biogeography of the Coffeeae alliance (Rubiaceae). J. Biogeogr. 44,
1966–1979. https://doi.org/10.1111/jbi.12981.
Katoh, K., Toh, H., 2008. Recent developments in the MAFFT multiple sequence
alignment program. Brief. Bioinform. 9 (4), 286–298. https://doi.org/10.1093/bib/
bbn013.
Knox, E.B., Palmer, J.D., 1995. Chloroplast DNA variation and the recent radiation of the
giant senecios (Asteraceae) on the tall mountains of eastern Africa. Proc. Natl. Acad.
Sci. U.S.A., 92(22), 10349–10353. doi: 10.1073/Proc. Natl. Acad. Sci. U.S.
A..92.22.10349.
Kozlov, A.M., Darriba, D., Flouri, T., Morel, B., Stamatakis, A., 2019. RAxML-NG: a fast,
scalable, and user-friendly tool for maximum likelihood phylogenetic inference.
Bioinformatics 35 (21), 4453–4455. https://doi.org/10.1093/bioinformatics/
btz305.
Landis, M.J., Matzke, N.J., Moore, B.R., Huelsenbeck, J.P., 2013. Bayesian analysis of
biogeography when the number of areas is large. Syst. Biol. 62 (6), 789–804. https://
doi.org/10.10933/sysbio/syt040.
Larridon, I., Spalink, D., Jim´
enez-Mejías, P., M´
arquez-Corro, J.I., Martín-Bravo, S.,
Muasya, M., Escudero, M., 2021. The evolutionary history of sedges (Cyperaceae) in
Madagascar. J. Biogeogr. 48, 917–932. https://doi.org/10.1111/jbi.14048.
Lehmann, C.E.R., Solofondranohatra, C.L., Morton, J.A., Phelps, L.N., Ralimanana, H.,
Razanatsoa, J., Rakotoarimanana, V., Vorontsova, M.S., 2022. The grassy ecosystems
of Madagascar. In: The New Natural History of Madagascar (Goodman, S.M.,
Andrianarimisa, A., Armstrong, A.H., Cooke, A., De Wit, M., Ganzhorn, J.U., Gautier,
L., Goodman, S.M., Jones, J.P.G., Jungers, W.L., Krause, D.W., Langrand, O., Lowry,
P.P., Racey, P.A., Raselimanana, A.P., Safford, R.J., Sparks, J.S., Stiassny, M.L.J.,
Tortosa, P. and Vences, M., eds., pp. 152–168. Princeton University Press. Oxford-
Princeton. doi: 10.2307/j.ctv2ks6tbb.27.
Lemoine, F., Domelevo Entfellner, J.B., Wilkinson, E., Correia, D., D´
avila Felipe, M., De
Oliveira, T., Gascuel, O., 2018. Renewing Felsenstein’s phylogenetic bootstrap in the
era of big data. Nature 556 (7702), 452–456. https://doi.org/10.1038/s41586-018-
0043-0.
Li, H., Durbin, R., 2009. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25 (14), 1754–1760. https://doi.org/10.1093/
bioinformatics/btp324.
Linder, H.P., Rabosky, D.L., Antonelli, A., Wüest, R.O., Ohlemüller, R., 2014.
Disentangling the inuence of climatic and geological changes on species radiations.
J. Biogeogr. 41, 1313–1325. https://doi.org/10.1111/jbi.12312.
Lowry, P.P.I., Schatz, G.E., Phillipson, P.B., 1997. The classication of natural and
anthropogenic vegetation in Madagascar. In: Goodman, S.M., Patterson, B.D. (eds.),
Natural change and human impact. Smithsonian Institution Press, Washington, D.C.
Lowry, P.P., Phillipson, II, P.B., Andriamahefarivo, L., Schatz, G.E., Rajaonary, F.,
Andriambololonera, S., 2018. Flora, in S.M. Goodman, M.J. Raherilalao,
Wohlhauser, S. (eds.). The terrestrial protected areas of Madagascar: their history,
description, and biota. pp. 243–255. Assn. Vahatra, Antananarivo, Madagascar.
Madagascar Catalogue, Catalogue of the Vascular Plants of Madagascar. Missouri
Botanical Garden 2024 http://legacy.tropicos.org/Project/Madagascar [Accessed:
28 February 2024].
Mandel, J.R., Dikow, R.B., Funk, V.A., Masalia, R.R., Staton, S.E., Kozik, A.,
Michelmore, R.W., Rieseberg, L.H., Burke, J.M., 2014. A target enrichment method
for gathering phylogenetic information from hundreds of loci: an example from the
Compositae. Appl. Plant. Sci. 2 (2), apps.1300085. https://doi.org/10.3732/
apps.1300085.
Mandel, J.R., Dikow, R.B., Siniscalchi, C.M., Thapa, R., Watson, L.E., Funk. V.A., 2019. A
fully resolved backbone phylogeny reveals numerous dispersals and explosive
diversications throughout the history of Asteraceae. Proc. Natl. Acad. Sci. U.S.A.,
116(28) 14083–14088 doi: 10.1073/Proc. Natl. Acad. Sci. U.S.A..190387111.
Masters, J.C., G´
enin, F., Zhang, Y., Pellen, R., Huck, T., Mazza, P.P.A., Rabineau, M.,
Doucour´
e, M., Aslanian, D., 2021. Biogeographic mechanisms involved in the
colonization of Madagascar by African vertebrates: Rifting, rafting and runways.
J. Biogeogr. 48, 492–510. https://doi.org/10.1111/jbi.14032.
Matthews, T.J., Triantis, K., 2021. Island biogeography. Curr. Biol., 31(19),
R1201–R1207. doi: 10.1016/j.cub.2021.07.033.
Matzke, N.J., 2022. Statistical comparison of DEC and DEC+J is identical to comparison
of two ClaSSE submodels, and is therefore valid. J. Biogeogr. 49, 1805–1824.
https://doi.org/10.1111/jbi.14346.
Matzke, N.J., 2013. Probabilistic historical biogeography: new models for founder-event
speciation, imperfect detection, and fossils allow improved accuracy and model-
testing. Front. Biogeogr. 5(4). http://dx.doi.org/10.21425/F5FBG19694 Retrieved
from https://escholarship.org/uc/item/44j7n141.
Moat, J., Smith, P., 2007. Atlas of the vegetation of Madagascar. Royal Botanic Gardens,
Kew Richmond, UK.
Nathan, R., Bullock, J.M., Ronce, O., Schurr, F.M., 2009. Seed Dispersal. In: Encyclopedia
of Life Sciences (eLS), (Ed.). doi: 10.1002/9780470015902.a0021225.
Nie, Z.-L., Funk, V.A., Meng, Y., Deng, T., Sun, H., Wen, J., 2016. Recent assembly of the
global herbaceous ora: evidence from the paper daisies (Asteraceae: Gnaphalieae).
New Phytol. 209, 1795–1806. https://doi.org/10.1111/nph.13740.
Nylinder, S., Razamandimbison, S.G., Anderberg, A.A., 2016. From the Namib around
the world: biogeography of the Inuleae–Plucheinae (Asteraceae). J. Biogeogr. 43,
1705–1716. https://doi.org/10.1111/jbi.12764.
Padilla, D.P., Gonz´
alez-Castro, A., Nogales, M., 2012. Signicance and extent of
secondary seed dispersal by predatory birds on oceanic islands: the case of the
C. Blanco-Gavald`
a et al.

Canary archipelago. J. Ecol. 100, 416–427. https://doi.org/10.1111/j.1365-
2745.2011.01924.x.
Palazzesi, L., Hidalgo, O., Barreda, V.D., Forest, F., H¨
ohna, S., 2022. The rise of
grasslands is linked to atmospheric CO
2
decline in the late Palaeogene. Nat.
Commun. 13 (1), 293. https://doi.org/10.1038/s41467-021-27897-y.
Pirie, M.D., Kandziora, M., Nürk, N.M., Le Maitre, N.C., de Kuppler, A.M., Gehrke, B.,
Oliver, E.G.H., Bellstedt, D.U., 2019. Leaps and bounds: geographical and ecological
distance constrained the colonisation of the Afrotemperate by Erica. BMC Evol. Biol.
19, 222. https://doi.org/10.1186/s12862-019-1545-6.
Rakotoarivelo, A.R., Goodman, S.M., Schoeman, M.C., Willows-Munro, S., 2019.
Phylogeography and population genetics of the endemic Malagasy bat,
Macronycteris commersoni s.s. (Chroptera: Hipposideridae). PeerJ, 7, e5866. doi:
10.7717/peerj.5866.
Ramdhani, S., Barker, N.P., Baijnath, H., 2009. Rampant nonmonophyly of species
in Kniphoa Moench (Asphodelaceae) suggests a recent Afromontane radiation.
Taxon 58, 1141–1152. https://doi.org/10.1002/tax.584008.
Razamandimbison, S.G., Kainulainen, K., Wikstr¨
om, N., Bremer, B., 2017. Historical
biogeography and phylogeny of the pantropical Psychotrieae alliance (Rubiaceae),
with particular emphasis on the Western Indian Ocean Region. Am. J. Bot. 104,
1407–1423. https://doi.org/10.3732/ajb.1700116.
Razamandimbison, S.G., Wikstr¨
om, N., Khodabandeh, A., Rydin, C., 2022. Phylogeny of
the Madagascar-centred triba Danaideae (Rubiaceae) as a precursor to taxonomic
revision: insights into its generic and species limits, afnities and distribution. Ann.
Bot. 130 (6), 849–867. https://doi.org/10.1093/aob/mcac121.
Ree, R.H., Sanmartín, I., 2018. Conceptual and statistical problems with the DEC+J
model of founder-event speciation and its comparison with DEC via model selection.
J. Biogeogr. 45, 741–749. https://doi.org/10.1111/jbi.13173.
Ree, R.H., Moore, B.R., Webb, C.O., Donoghue, M.J., 2005. A likelihood framework for
inferring the evolution of geographic range on phylogenetic trees. Evolution 45 (11),
2299–2311. https://doi.org/10.1111/j.0014-3820.2005.tb00940.x.
Ree, R.H., Smith, S.A., 2008. Maximum likelihood inference of geographic range
evolution by dispersal, local extinction and cladogenesis. Syst. Biol. 57 (1), 4–14.
https://doi.org/10.1080/10635150701883881.
Ronquist, F. (1997). Dispersal-Vicariance Analysis: A New Approach to the
Quantication of Historical Biogeography. Syst. Biol., 46(1), 195–203. doi: 10.2307/
2413643Rundell, R.J., & Price, T.D. (2009). Adaptive radiation, nonadaptive
radiation, ecological speciation and nonecological speciation. Trends Ecol. Evol., 24
(7), 394–399. doi: 10.1016/j.tree.2009.02.007.
Rundle, H.D., Nosil, P., 2005. Ecological speciation. Ecol. Lett. 8, 336–352. https://doi.
org/10.1111/j.1461-0248.2004.00715.x.
Sayyari, E., Mirarab, S., 2016. Fast Coalescent-Based computation of local branch
support from quartet frequencies. Mol. Biol. Evol. 33 (7), 1654–1668. https://doi.
org/10.1093/molbev/msw079.
Schatz, G.E., 1996. Malagasy/Indo-australo-malesian phytogeographic connections. In:
W. R. Lourenço (ed.), Biog´
eographie de Madagascar. Pp. 73–84. Editions ORSTOM,
Paris, France.
Schmidt-Lebuhn, A.N., Bovill, J., 2021. Phylogenomic data reveal four major clades of
Australian Gnaphalieae (Asteraceae). Taxon 70, 1020–1034. https://doi.org/
10.1002/tax.12510.
Silander Jr., J.A., Bond, W.J., Ratsirarson, J., 2024. The grassy ecosystems of Madagascar
in context: ecology, evolution, and conservation. Plants People Planet 6 (1), 94–115.
https://doi.org/10.1002/ppp3.10417.
Skema, C., Jourdain-Fievet, L., Dubuisson, J.Y., Le P´
echon, T., 2023. Out of Madagascar,
repeatedly: The phylogenetics and biogeography of Dombeyoideae (Malvaceae s.l.).
Mol. Phylogenet. Evol. 182, 107687. https://doi.org/10.1016/j.
ympev.2022.107687.
Smissen, R.D., Bayer, R.J., Bergh, N.G., Breitwieser, I., Freire, S.E., Galbany-Casals, M.,
Schmidt-Lebuhn, A.N., Ward, J.M., 2020. A revised subtribal classication of
Gnaphalieae (Asteraceae). Taxon 69, 778–806. https://doi.org/10.1002/tax.12294.
Smith, S.A., Donoghue, M.J., 2010. Combining historical biogeography with niche
modeling in the Caprifolium clade of Lonicera (Caprifoliaceae, Dipsacales). Syst.
Biol. 59 (3), 322–341. https://doi.org/10.1093/sysbio/syq011.
Smith, S.A., Brown, J.W., Walker, J.F., 2018. So many genes, so little time: A practical
approach to divergence-time estimation in the genomic era. PLoS ONE, 13(5),
e0197433. doi: 10.1371/journal.pone.0197433.
Solofondranohatra, C.L., Vorontsova, M.S., Hempson, G.P., Hackel, J., Cable, S.,
Vololoniaina, J., Lehmann, C.E.R., 2020. Fire and grazing determined grasslands of
central Madagascar represent ancient assemblages. Proc. Biol. Sci. 287 (1927),
20200598. https://doi.org/10.1098/rspb.2020.0598.
Spriggs, E.L., Christin, P.A., Edwards, E.J., 2014. C4 photosynthesis promoted species
diversication during the Miocene grassland expansion. PloS one, 9(8), e105923.
doi: 10.1371/journal.pone.0097722.
Stamatakis, A., 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis
of large phylogenies. Bioinformatics 30 (9), 1312–1313. https://doi.org/10.1093/
bioinformatics/btu033.
Tamura, K., Battistuzzi, F.U., Billing-Ross, P., Murillo, O., Filipski, A., Kumar, S., 2012.
Estimating divergence times in large molecular phylogenies. Proc. Natl. Acad. Sci. U.
S.A., 109(47), 19333–19338. doi: 10.1073/Proc. Natl. Acad. Sci. U.S.
A..1213199109.
Tamura, K., Tao, Q., Kumar, S., 2018. Theoretical foundation of the RelTime methods for
estimating divergence times from variable evolutionary rates. Mol. Biol. Evol. 35 (7),
1770–1782. https://doi.org/10.1093/molbev/msy044.
Tamura, K., Stecher, G., Kumar, S., 2021. MEGA11: molecular evolutionary genetics
analysis version 11. Mol. Biol. Evol. 38 (7), 3022–3027. https://doi.org/10.1093/
molbev/msab120.
Tao, Q., Tamura, K., Mello, B., Kumar, S., 2020. Reliable condence intervals for
RelTime estimates of evolutionary divergence times. Mol. Biol. Evol. 37 (1),
280–290. https://doi.org/10.1093/molbev/msz236.
Thomas, N., Bruhl, J.J., Ford, A., Weston, P.H., 2014. Molecular dating of Winteraceae
reveals a complex biogeographical history involving both ancient Gondwanan
vicariance and long-distance dispersal. J. Biogeogr. 41, 894–904. https://doi.org/
10.1111/jbi.12265.
Tiley, G.P., Crowl, A.A., Manos, P.S., Sessa, E.B., Solís-Lemus, C., Yoder, A.D.,
Burleigh, G., 2024a. Benets and limits of phasing alleles for network inference of
allopolyploid complexes. Syst. Biol. syea024. https://doi.org/10.1093/sysbio/
syae024.
Tiley, G.P., Crowl, A.A., Almary, T.O.M., Luke, W.R.Q., Solofondranohatra, C.L.,
Besnard, G., Lehmann, C.E.R., Yoder, A.D., Vorontsova, M.S., 2024b. Genetic
variation in Loudetia simplex supports the presence of ancient grasslands in
Madagascar. Plants People Planet 6 (2), 315–329. https://doi.org/10.1002/
ppp3.10437.
Leong Pock Tsy, J. M., Lumaret, R., Mayne, D., Vall, A. O., Abutaba, Y. I., Sagna, M.,
Rakotondralambo Raoseta, S. O., Danthu, P., 2009. Chloroplast DNA
phylogeography suggests a West African centre of origin for the baobab, Adansonia
digitata L. (Bombacoideae, Malvaceae). Mol. Ecol., 18(8), 1707–1715. doi: 10.1111/
j.1365-294X.2009.04144.x.
Umov, R., Gorospe, J.M., F´
er, T., Kandziora, M., Salomon, L., van Loo, M., Schmickl, R.,
2022. Utilizing paralogues for phylogenetic reconstruction has the potential to
increase species tree support and reduce gene tree discordance in target enrichment
data. Mol. Ecol. Resources 22 (8), 3018–3034. https://doi.org/10.1111/1755-
0998.13684.
Vences, M., Wollenberg, K.C., Vieites, D.R., Lees, D.C., 2009. Madagascar as a model
region of species diversication. Trends Ecol. Evol. 24 (8), 456–465. https://doi.org/
10.1016/j.tree.2009.03.011.
Vorontsova, M.S., Besnard, G., Forest, F., Malakasi, P., Moat, J., Clayton, W.D.,
Ficinski, P., Savva, G.M., Nanjarisoa, O.P., Razanatsoa, J., Randriatsara, F.O.,
Kimeu, J.M., Luke, W.R., Kayombo, C., Linder, H.P., 2016. Madagascar’s grasses and
grasslands: anthropogenic or natural? Proc. Biol. Sci. 283 (1823), 20152262.
https://doi.org/10.1098/rspb.2015.2262.
Warren, B.H., Strasberg, D., Bruggemann, J.H., Prys-Jonces, R.P., Th´
ebaud, C., 2010.
Why does the biota of the Madagascar region have such a strong Asiatic avour?
Cladistics 26, 526–538. https://doi.org/10.1111/j.1096-0031.2009.00300.x.
Wells, N.A., 2003. Some hypotheses on the Mesozoic and Cenozoic paleoenvironmental
history of Madagascar. In Natural History of Madagascar (Goodmand S.M., and
Bestead, J.P., eds). The University of Chicago Press, Chicago. USA.
Westerhold, W., Marwan, N., Drury, A.J., Liebrand, D., Agnini, C., Anagnostou, E.,
Barnet, J.S.K., Bothaty, S.M., De Vleeschouwer, D., Florindo, F., Frederichs, T.,
Hodell, D.A., Holbourn, A.E., Kroon, D., Lauretano, V., Littler, K., Lourens, L.J.,
Lyle, M., P¨
alike, H., R¨
ohl, U., Tian, J., Wilkens, R.H., Wilson, P.A., Zachos, J.C.,
2020. An astronomically dated record of Earth’s climate and its predictability over
the last 66 milion years. Science 369, 1383–1387. https://doi.org/10.1126/science.
aba6853.
Whittaker, R.J., Triantis, K.A., Ladle, R.J., 2008. A general dyanimc theory of oceanic
island biogeography. J. Biogeogr. 35, 977–994. https://doi.org/10.1111/j.1365-
2699.2008.01892.x.
Wiens, J.J., Graham, C.H., 2005. Niche conservatism: integrating evolution, ecology and
conservation biology. Annu. Rev. Ecol. Evol. Syst. 36 (1), 519–539. https://doi.org/
10.1146/annurev.ecolsys.36.102803.095431.
Wiens, J.J., Ackerly, D.D., Allen, A.P., Anacker, B.L., Buckley, L.B., Cornell, H.V.,
Damschen, E.I., Jonathan Davies, T., Grytnes, J.A., Harrison, S.P., Hawkins, B.A.,
Holt, R.D., McCain, C.M., Stephens, P.R., 2010. Niche conservatism as an emerging
principle in ecology and conservation biology. Ecol. Lett. 13, 1310–1324. https://
doi.org/10.1111/j.1461-0248.2010.01515.x.
Yoder, A.D., Campbell, C.R., Blanco, M.B., Dos Reis, M., Ganzhorn, J.U., Goodman, S.M.,
Hunnicutt, K.E., Larsen, P.A., Kappeler, P.M., Rasoloarison, R.M., Ralison, J.M.,
Swofford, D.L., Weisrock, D.W., 2016. Geogenetic patterns in mouse lemurs (genus
Microcebus) reveal the ghosts of Madagascar’s forests past. Proc. Natl. Acad. Sci. U.
S.A., 113(29), 8049–8056. doi: 10.1073/Proc. Natl. Acad. Sci. U.S.A..1601081113.
Yoder, A.D., Nowak, M.D., 2006. Has vicariance or dispersal been the predominant
biogeographic force in Madagascar? Only time will tell. Annu. Rev. Ecol. Evol. Syst.
37 (1), 405–431. https://doi.org/10.1146/annurev.ecolsys.37091305.110239.
Zhang, C., Rabiee, M., Sayyari, E., Mirarab, S., 2018. ASTRAL-III: polynomial time
species tree reconstruction from partially resolved gene trees. BMC Bioinformatics
19 (Suppl 6), 153. https://doi.org/10.1186/s12859-018-2129-y.
C. Blanco-Gavald`
a et al.