Molecular systematics and historical biogeography of Araceae at a worldwide scale and in Southeast Asia

This dissertation addresses the biogeographic history of the
Araceae family and of one of its largest genera, Alocasia. With
>3300 species, Araceae are among the largest families of
flowering plants. It is the monocot lineage with the deepest fossil
record, reaching back to the Early Cretaceous. Araceae are
distributed worldwide, but >3100 species occur in the tropical
regions of the Americas, Asia, Africa, and Australia; most fossils
from the Late Cretaceous and many younger ones come from the
temperate zone in the northern hemisphere, implying much extinction
and range expansion. Most subfamilies are pantropically
distributed, and almost all genera are restricted to one continent.
Alocasia comprises 113 species, many as yet undescribed, making it
the 7 th -largest genus of the Araceae. Many species are
ornamentals, and two species are of interest for man, either for
food (giant taro) or in local cultures (Chinese taro). The origin
of these species was not known. Alocasia is distributed in
Southeast Asia from India to Australia, with species occurring on
all islands of the Malay Archipelago. This region has a complex
geologic history shaped by the collision of the Eurasian, the
Pacific, and the Indo-Australian plate. The Malesian flora and fauna
comprises Laurasian and Gondwanan elements, reflecting the influence
of changing sea levels, uplift and submergence of islands, and
other tectonic movement. In this thesis, I used molecular
phylogenetics, Bayesian divergence dating, ancestral area
reconstruction to understand the past distribution of the Araceae
family and the Alocasia clade in the context of past continent
movements and climate history. For the family analysis, existing
chloroplast DNA matrices were augmented so that all Araceae genera
were represented by one or more species, with a focus on covering
geographic disjunctions, especially between continents. Divergence
dating relied on seven confidently assigned fossil constraints,
comparing uniform and gamma-shaped prior distributions on fossil
ages, as well as several molecular clock models. Biogeographic
analyses were performed in a model-based likelihood framework that
took into account past dispersal routes based on continent
connectivity and climate. I also integrated fossils into the
ancestral area reconstruction, either simulating extinct or still
existing ranges, and then compared results to those obtained from
analyses without fossils. To study the morphology and ecology of
Alocasia, fieldwork was conducted in Malaysia and herbarium work in
Germany, the Netherlands, Indonesia, Malaysia, and Singapore.
Maximum likelihood phylogenies were inferred based on chloroplast
and nuclear loci, sequenced for 71 species of Alocasia plus 25
outgroup species from 16 genera. Bayesian divergence dating of the
nuclear phylogeny relied on one fossil constraint and ancestral
areas were reconstructed using parsimony- and likelihood-based
methods. The Araceae diverged from the remaining Alismatales in the
Early Cretaceous (ca. 135 Ma ago), and all eight subfamilies
originated before the Cenozoic. The earliest lineages are inferred
to have occurred in Laurasia (based on fossils and tree topology),
and most lineages reached Africa, South America, Southeast Asia,
and Australia during the Paleogene and Neogene. Many clades
experienced extinction in the temperate regions of the northern
hemisphere during the Oligocene climate cooling. Two continentally
disjunct genera (Nephthytis and Philodendron) are polyphyletic and
need taxonomic rearrangement. Plastid substitution rates are
exceptionally high in free-floating and water-associated Araceae.
Ancestral area reconstructions obtained when fossil (no longer
occupied) ranges where included in the analyses were more plausible
than those without fossil ranges. This is not a trivial result
because only in a quantitative (computer-based) analysis is it
possible for fossil ranges to influence results (here areas) at
distant nodes in the phylogenetic tree. The nuclear and plastid
phylogenies of Alocasia revealed the polyphyly of the two genera
Alocasia and Colocasia; to achieve monophyly, two species (Alocasia
hypnosa and Colocasia gigantea) have to be moved to other genera.
There were strong incongruencies between phylogenies from the two
partitions: The chloroplast data reflect geographical proximity, the
nuclear morphological similarity. This may indicate hybridization
events followed by chloroplast capture. Based on the nuclear tree,
Alocasia split from its sister group by the end of the Oligocene
(ca. 24 Ma) and colonized the Malay Archipelago from the Asian
mainland. Borneo played a central role, with 11–13 of 18–19
inferred dispersal events originating there. The Philippines were
reached from Borneo 4–5 times in the late Miocene and early
Pliocene, and the Asian mainland 6–7 times during the Pliocene. The
geographic origin of two domesticated species could be resolved:
Giant taro originated on the Philippines and Chinese taro on the
Asian mainland.