Babblers, Biogeography and Bayesian Reasoning M G Department of Zoology

Babblers, Biogeography and Bayesian Reasoning
MAGNUS GELANG
Department of Zoology
Stockholm University
2012
Babblers, Biogeography and Bayesian Reasoning
Doctoral dissertation 2012
Magnus Gelang
Department of Zoology
Stockholm University
SE-106 91 Stockholm
Sweden
Department of Vertebrate Zoology
Swedish Museum of Natural History
PO Box 50007
SE-104 05 Stockholm
Sweden
[email protected]
© Magnus Gelang, Stockholm 2012
ISBN 978-91-7447-438-1
Cover illustration: Large wren babbler Turdinus macrodactyla, Bodogol, Java, Indonesia.
Photo: Magnus Gelang.
Printed in Sweden by US-AB, Stockholm 2012
Distributor: Department of Zoology, Stockholm University
ABSTRACT
In this thesis, I try to proceed one step further towards an understanding of the
biogeographic processes forming the distribution patterns of organisms that we see today. Babblers
and warblers are diverse groups of passerines that are phylogenetically intermixed with other
groups in the superfamily Sylvioidea.
First, the gross phylogeny of the babblers and associated groups was estimated. Five major
lineages of a well-supported monophyletic babbler radiation were recovered, and we proposed a
new classification at family and subfamily level. Further, the genus Pnoepyga was excluded from
Timaliidae, and we proposed the new family Pnoepygidae fam. nov.
Second, the systematic position was investigated for the Albertine Rift taxon Hemitesia
neumanni, which was found to be nested within the almost entirely Asian family Cettidae, and
possible biogeographical scenarios were discussed. We concluded that the most plausible
explanation involved late Miocene vicariance in combination with local extinctions.
Third, the historical biogeography of a Leiothrichinae subclade, the Turdoides babblers and
allies, was inferred. We concluded that the Middle East region probably played an important role in
the early history of this clade, followed by local extinctions in this region.
Fourth, a Bayesian method to reconstruct the historical biogeography under an event-based
model was proposed, where the total biogeographic histories are sampled from its posterior
probability distribution using Markov chains.
In conclusion, I believe that, especially with more sophisticated methods available, we will
see an increasing number of studies inferring biogeographic histories that lead to distribution
patterns built up by a combination of dispersals and vicariance, but where these distributions have
been extensively reshaped, or litterally demolished, by local extinctions. Therefore, my answer to
the frequently asked question dispersal or vicariance? is both, but not the least: extinctions.
Keywords: Africa, Asia, Bayesian inference, biogeography, Cettidae, dispersal, extinction, Middle
East, persistence, Sylvioidea, Sylviidae, Timaliidae, vicariance.
LIST OF PAPERS
I
Gelang, M., Cibois, A., Pasquet, E., Olsson, U., Alström, P. & Ericson, P. G. P.
(2009) Phylogeny of babblers (Aves, Passeriformes): major lineages, family limits
and classification. Zoologica Scripta. 38(3): 225–236.
II
Irestedt, M., Gelang, M., Sangster, G., Olsson, U., Ericson, P. G. P. & Alström, P.
(2011) Neumann´s warbler Hemitesia neumanni (Sylvioidea): the sole African
member of a Palaeotropic Miocene avifauna. Ibis. 153: 78–86.
III
Gelang, M., Pasquet, E., Cibois, A., Alström, P. & Ericson, P. G. P. (submitted)
Ancestral ranges concealed by local extinctions: the historical biogeography of the
African and Asian Turdoides babblers and allies (Aves: Passeriformes). m.s.
IV
Gelang, M. & Bohlin, A. (submitted) Bayesian inference of total biogeographic
history under an event-based model. m.s.
The published papers are reprinted with permission from the publishers. No part of this thesis must be reproduced
without permission.
CONTENTS
INTRODUCTION
1
The Study Group
1
The Old World in the Neogene Time
2
Event-Based Biogeography
3
Conditional Probabilities and Bayesian Methods
4
RESULTS AND DISCUSSION
5
Phylogenetic Relationships
5
Phylogenetic Biogeography
7
A New Method in Historical Biogeography
8
CONCLUSIONS
9
REFERENCES
10
ACKNOWLEDGEMENTS
14
INTRODUCTION
To explain the origin and history of an organism is tricky, but it is crucial for the
understanding of the organisms biology. Biogeography is an interdisciplinary field containing both
biotic, abiotic and mathematical components, which aims to reconstruct the history in terms of
distribution ranges, diversity patterns etc. To make such reconstructions we may need to investigate
both phylogenetic relationships, geological and climatological history and ecology and general
biology of the study group, as well as other factors that might have influenced the distributions. To
me, this complexity is what makes biogeography such a stimulating field.
In my work, I have studied a group of birds that inhabits large parts of the Old World, and
whose phylogenetic relationships were, at the beginning of my work, poorly known. During the
course of my PhD project, many questions have been answered, but even more questions have
arised. For me personally, it has been a journey streching over many horizons and which has taken
me both into and out of tangled bushes, both litterally and metaphorically. The project started with
much focus on babblers in Southeast Asia, i.e. birds and a region well familiar to me, but ended in
relatively new environments, such as mathematical and computational, and with a focus on the
Middle East. This has been a fantastic journey with quite a few tough but stimulating challenges.
The Study Group
Babblers and warblers constitute a large portion of the passerine superfamily Sylvioidea,
which comprises about 25% of the passerine species diversity (Dickinson, 2003). While babblers
are typically stocky, small to medium-sized birds showing extraordinary social behaviours, warblers
are generally small neatly built birds without the sociality shown by babblers (Collar & Robson,
2007). However, the two groups, as traditionally treated, can almost be considered ecological
groups and both are represented in a number of sylvioid clades where often phylogenetically
intermixed (e.g. Beresford et al., 2005; Alström et al., 2006; Johanssson et al., 2008).
Traditionally, both Timaliidae (babblers) and Sylviidae (warblers) were included in the “Old
world insectivorous group” (Hartert, 1910; Mayr & Amadon, 1951; Deignan, 1964). Beecher
(1953) proposed a phylogeny of oscines based on jaw musculature, dividing oscines in the two
major groups Sylvioidea and Timalioidea. Together with a large number of passerine groups most
babblers were placed in Timalioidea, but only fragments of his research have been corraborated in
later studies. Among the morphological characters which have been proposed to be diagnostic for
Timaliidae are unspotted juvenile plumage, presence of rictal bristles and scutellated tarsus (Sibley
& Ahlquist, 1990 and references therein). Jean Delacour made major contributions to the
knowledge of babblers (e.g. Delacour, 1946, 1950) by subdividig the babblers into major groups.
Sibley & Ahlquist (1990) proposed a revolutionary phylogeny of birds based on DNA-DNA
hybridization. In this extensive study, they excluded the Australian babblers (i.e. genera Garritornis
and Pomatostomus) and placed typical babblers together with warblers and white-eyes. In recent
studies (Barker et al., 2002, 2004; Ericson & Johansson, 2003; Beresford et al., 2005; Alström et
al., 2006; Johansson et al., 2008), a major pattern of distinct clades has been proposed, but with
either poorly supported, unresolved or conflicting relationships among the families of Sylvioidea.
1
Regarding babblers, Alice Cibois and Eric Pasquet with colleagues have made several
important studies (Cibois et al., 1999, 2001, 2010; Cibois, 2003a, 2003b; Pasquet et al., 2006).
Among many findings, they excluded a number of taxa, included the genera Zosterops and Sylvia,
and found several paraphyletic genera of babblers. Recently, other studies have resolved important
parts of the babbler tree, such as white-eyes and allies (Moyle et al., 2009), laughingthrushes and
allies (Lou et al., 2009), barwings (Dong et al., 2010a) and scimitar babblers (Dong et al., 2010b;
Reddy & Moyle, 2011). The phylogenetic relationships of the clade containing Cettidae etc. have
been extensively studied by Per Alström and Urban Olsson with colleagues (e.g. Alström et al.,
2007, 2008, 2011a, 2011b; Olsson et al., 2004, 2005, 2006).
The Old World in the Neogene Time
Paleogene
Neogene
Quaternary
The Neogene period follows the Paleogene
Period
Epoch
Age (MY)
period and is succeeded by the Quatenary period (Fig.
Holocene
0–0.0117
1). The Neogene period is subdivided into the Miocene
epoch (ca. 23–5.3 million years ago (MYA)) and the
Pleistocene 0.0117–2.588
Pliocene epoch (ca. 5.3–2.6 MYA), and Pliocene is
Pliocene
2.588–5.332
followed by the Pleistocene epoch (ca. 2.6 MYA–12
000 years before present) in the Quaternary period. The
geographical region covered in this thesis was affected
by some important changes during the Neogene period.
After the warm later part of Oligocene, this warm
Miocene
5.332–23.03
climate remained during Miocene and a peak in
temperature occurred around 15 MYA. After this, the
climate started to change towards the cooler Pliocene
(Zachos, 2001). Further, India continued to collide with
Asia, resulting in the uplift of the Himalayas (Hall et
Oligocene
23.03–33.9
al., 2008). Africa collided with Eurasia, causing the
closure of the Tethys sea (Vrielnyck et al., 1997; Rögl,
Eocene
33.9–55.8
1998), which together with global fall of sea levels
caused land bridges between Africa and Asia. In the
Paleocene
55.8–65.5
mid-Miocene, extensive laurel forests covered much of
Fig. 1. Overview of the Cenozoic era.
the area from northern Africa to eastern Asia
Only the Neogene epochs are shown to
(Mandaville, 1977; Utescher et al., 2007; Fernándezscale.
Palacios et al., 2011). As a result of aridification and
cooling of the climate, these forests were shrinking in the end of Miocene, while areas of grassland
expanded throughout the region (Mandaville, 1977; Retallack, 1992; Vrba, 1993; Flower &
Kennett, 1994; Jacobs et al., 1999). During Pliocene, the concerned geographic area looked much
as we see it today (Vrielnyck et al., 1997; Hall et al., 2008; Hall, 2009).
2
Event-Based Biogeography
Evolutionary processes linked to geology and to geological processes is a corner stone in
this thesis. These processes, the biogeographic events, are often both components of the main
question of a biogeographic study and key components in the inference of the biogeographic
history. Below, I pay some attention to these events, mainly following Cox & Moore (2006) and
references therein.
Dispersal is when an organism, or a population, colonizes one place from another. Dispersal
can be subdivided into jump dispersal and range extension, where the former refers to an across
barrier dispersal, and the latter refers to an increase of range area.
Extinction is when an organism, or a population, disappears from either a range unit or
globally. In biogeographic analyses, extinction can be divided into global extinction, local
extinction and range contraction, but global extinction is typically not used in this kind of analyses.
Obviously, extinctions are not straight-forward to infer as they normally lead to incomplete data
rather than available information.
Vicariance is when an organism, or a population, is split by an abiotic process such as when
a land mass is subdivided geologically and its inhabitants survives in both of its new units. The
object lesson of vicariance is the break-up of Gondwana into Antarctica, Australasia and South
America (e.g. Barker et al., 2002, 2004; Ericson et al., 2002a, 2002b). Vicariance has often been the
concurrent explanation of distribution patterns to that of explanations involving dispersal.
Persistence is when an organism, or a population, survives in its range. The term duplication
is often used almost synonymously albeit with the meaning of including a radiation. Parallell to this
term is co-speciation among host and parasite (Ronquist, 1998).
For decades, the debate on the relative importance of dispersal and vicariance (e.g. Nelson &
Platnick, 1980; Wiley, 1988; Chesser & Zink, 1994; Zink et al., 2000; Via, 2001) has sometimes
been nearly bitter, but fortunately this debate has also contributed to a more nuanced knowledge and
has resulted in a consensus where a combination of events should be taken into account. Among the
most prominent contribution is the parsimony-based Dispersal-Vicariance Analysis (DIVA;
Ronquist, 1997, 1998), which was the first available event-based approach to biogeography. DIVA
applies a simple three-dimensional cost matrix where the costs of dispersal and extinction are both
set to 1, and the costs of vicariance and duplication are both set to 0. Therefore, a DIVA analysis
favours duplications and vicariance events over dispersals and extinctions, which in turn leads to
the typical scenario with widespread ancestors and underestimation of dispersals and extinctions
(Ronquist, 1996; Sanmartín, 2003; Clarke et al., 2008; Buerki et al., 2010). While an analysis
performed by DIVA needs a fully bifurcate tree (Ronquist, 1996), Nylander et al. (2008) proposed a
way of parse a set of sampled trees (e.g. from the target distribution from an Bayesian estimation of
phylogeny) in DIVA. From these, the marginal probabilities of ancestral ranges can be calculated,
which enable us to account for phylogenetic uncertainties. During the last decade, a maximum
likelihood approach has started to play an increasingly important role in the field of historical
biogeography. The maximum likelihood is inferred under the Dispersal-Extinction-Cladogenesis
(DEC; Ree & Smith, 2005; Ree et al., 2008) model. The DEC model has proven to infer plausible
reconstructions (Clarke et al., 2008; Buerki et al., 2010), especially as additional data from geology
3
etc. can be included in the analysis. Also Bayesian approaches in biogeography have been proposed
(Moore and Donoghue, 2009; Sanmartín et al., 2008).
Conditional Probabilities and Bayesian Methods
Especially during the last decades, Bayesian methods have revolutionized biological
sciences in general, and phylogenetic research in particular (Huelsenbeck et al., 2001b; Beaumont
& Rannala, 2004). Bayesian methods rely on the relations of conditional probabilities explained in
Bayes´ theorem,
P  H ∣D =
P  D∣H  P  H 
P  D
,
(1)
which, in this case, states that the posterior probability of a hypothesis given some data, P(H|D),
equals the product of the likelihood of the data given the hypothesis, P(D|H), and the prior
probability of the hypothesis, P(H), normalized by the marginal probability of the data, P(D). The
marginal probability is the probability of the data integrated over all possible hypothesis, as shown
in equation 3, Paper IV. A very simple metaphore, or comparison, of the proportionality between
the posterior probability and the numerator in Bayes´ theorem is that our knowledge (cf. posterior
probability) is proportional to some observation (cf. likelihood) and our assumption (cf. prior
probability).
Bayes´ theorem relies on conditional probabilities. This subject needs some further
explanation. We take the likelihood as an example, where the probability of the data is conditioned
on the hypothesis. This can also be expressed as the probability of the data to fit the hypothesis, or
the probability of the data to be true in the light of the hypothesis. In practice, the likelihood may be
calculated as the joint probability of the conjunction of the data and hypothesis. If we look at the
prior probability of the hypothesis, this is the probability for the hypothesis to be valid regardless of
the data, thus if we are unaware of the data.
The denominator in Bayes´ theorem is typically impossible to calculate in phylogenetic
tasks, and therefore we choose to pay interest to the probability distributions of the components in
Bayes theorem. We solve the problems of calculating the marginal probability by using the
proportionality of the posterior probability and the product of the likelihood and the prior, and use
Markov chains to approximate these distributions. Practically, we seek a local optimum in a
parameter space, equivalent to the desired posterior probability distribution. The use of Markov
methods to perform statistical sampling to get the expectation of a function is widely used,
particularly by using the “Metropolis-Hastings-Green algorithm” (Metropolis et al., 1953; Hastings,
1970; Green, 1995) on Metropolis-coupled Markov chain Monte Carlo (MC3). If we think of the
parametric space as a landscape where mountains represent local optima, we want the chains to
climb uphill but still not to get stuck on just any hill (local suboptimum). Instead, we want the
chains to be able to traverse over valleys and so leave a smaller hill to reach higher summits. This is
performed in basically three ways. We let a chain take a random step. If this step is uphill the step is
4
accepted, if downhill the step is accepted with a probability proportional to the step. The second
way to facilitate a chain to cross a valley is to “flatten” the landscape. This is made by letting
different chains be affiliated to different “temperatures”, where a chain affiliated with a high
temperature traverses a landscape which appears to be “melted” and therefore the summits and
valleys appear flattened. The third way is to apply incremental heating schemes on a set of chains,
where the states can be swapped between chains and where we pay attention to the state of the
“cold” chain (Geyer, 1991; Geyer and Thompson, 1995). In all, methods applying incrementally
heated MC3 have proven efficient in phylogenetic research (Yang & Rannala, 1997; Larget &
Simon, 1999; Huelsenbeck & Ronquist, 2001; Drummond & Rambaut, 2007). Once the MC3 have
found the desired local optimum (the target posterior probability distribution), the states of the cold
chain is sampled and the posterior probability is calculated as the frequency of a state among these
samples.
RESULTS AND DISCUSSION
Phylogenetic Relationships
This thesis relies on molecular data in one way or the other. In paper I, the main source of
sequence data was obtained from fresh material (blood- or tissue samples), while in paper II and
paper III the main source was from museum study skins. Extraction, amplification and sequencing
of DNA were performed using standard protocols (QiaGene), which were modified when
appropriate such as for the study skins. In these cases, we mainly followed Irestedt et al. (2006),
and extra effort was put on design of primers. In general, fragment lengths of about 200–350 base
pairs were amplified for the skins.
Phylogenies were reconstructed using both Bayesian and maximum likelihood methods, but
in all cases we have chosen to rely on Bayesian inference as the main approach. Phylogenetic
estimations were performed using MRBAYES (Huelsenbeck & Ronquist, 2001a), and molecular
dating was performed using BEAST (Drummond & Rambaut, 2007). Further, inferences were made
on the datasets partitioned on loci following the substitution models proposed by the Akaike
criterion (Akaike, 1974).
5
*
*
* **
*
*
*
*
*
*
*
*
*
* *
*
*
*
*
* *
*
* *
*
*
*
*
*
*
*
*
*
Timaliidae
*
Zosteropinae
*
*
Timaliinae
*
Sylviidae
*
*
*
Pellorneinae
* *
**
*
Muscicapa
Alauda
Panurus
Hirundo
Acrocephalus
Donacobius
Thamnornis
Megalurus
Pnoepyga albiventer
Pnoepyga pusilla
Prinia
Pycnonotus
Aegithalos
Phylloscopus
Myzornis pyrrhoura
Parophasma galinieri
Sylvia atricapilla
Lioparus chrysotis
Chrysomma sinense
Fulvetta vinipectus
Rhopophilus pekinensis
Chamaea fasciata
Paradoxornis gularis
Paradoxornis nipalensis
Paradoxornis verreauxi
Yuhina diademata
Yuhina flavicollis
Yuhina gularis
Stachyris whiteheadi
Speirops lugubris
Zosterops japonica
Lophozosterops javanicus
Heleia crassirostris
Lophozosterops superciliaris
Stachyris chrysaea
Macronous gularis
Dumetia hyperythra
Timalia pileata
Spelaeornis chocolatinus
Stachyris nigriceps
Stachyris striolata
Pomatorhinus ochraceiceps
Pomatorhinus schisticeps
Xiphirhynchus superciliaris
Alcippe poioicephala
Graminicola bengalensis
Turdinus macrodactyla
Gampsorhynchus rufulus
Schoeniparus rufogularis
Malacocincla abbotti
Kenopia striata
Pellorneum ruficeps
Illadopsis cleaveri
Ptyrticus turdinus
Napothera epilepidota
Jabouilleia danjoui
Rimator pasquieri
Babax lanceolatus
Garrulax sannio
Garrulax leucolophus
Turdoides jardinei
Kupeornis gilberti
Phyllanthus atripennis
Garrulax erythrocephalus
Cutia nipalensis
Leiothrix argentauris
Heterophasia melanoleuca
Liocichla steerii
Actinodura souliei
Minla cyanouroptera
Minla ignotincta
* *
* *
**
Leiothrichinae
©MrEnt
*
*
Pnoepygidae
Lanius
Corvus
Fig. 2. The five loci phylogeny of babblers modified from paper
I (fig. 1) estimated by Bayesian inference. Posterior
probability over 95 % are shown as asterisk, and main clades
concerned in the proposed taxonomy in paper I are visualized
in coloured boxes.
Although Cibois (2003) recognised a pattern of main clades of babblers, the support was
relatively low in parts of the phylogeny. Based upon this work, we extended the analysis both
regarding taxonomic sampling and number of loci to better resolve the babbler tree and establish the
family limits for the group. In paper I, we recognise five main clades of babblers (fig. 2) and we
also exclude the wren babbler genus Pnoepyga from the babbler radiation. A novel classification is
proposed for these five main clades. The parrotbills (Paradoxornis), Sylvia warblers, the enigmatic
and monotypic genus Myzornis and a few others form a well supported clade which we refer to as
Sylviidae. Sylviidae is sister to a large clade, which consists of four major clades. The entire clade
is referred to as Timaliidae, and we propose a subfamily classification following the four main
6
clades. The white-eyes, formerly Zosteropidae, and the yuhinas form a clade which we refer to as
Zosteropinae. A second clade comprising “tree babblers” (Stachyris, Macronous etc.), scimitar
babblers (Pomatorhinus) and the genus Timalia is referred to as Timaliinae. The third clade, which
consists mostly of “jungle babblers” (e.g. Malacocincla and Illadopsis) and wren babblers (e.g.
Napothera and Turdinus) is referred to as Pellorneinae, while the last clade, which comprises e.g.
laughingthrushes (Garrulax and Babax), “song babblers” (e.g. Leiothrix and Heterophasia), and the
genus Turdoides, is referred to as Leiothrichinae.
The excluded genus Pnoepyga is proposed to form the new family Pnoepygidae based on
both morphological and genetic characteristics. No other genus is represented in this family (fig. 2),
which therefore only comprises three or four species, depending on which taxonomy is followed.
The family Pnoepygidae is distributed from the Himalayas through China and Southeast Asia to the
Indonesian archipelago, reaching all the way to Timor. All members inhabit mountain and hill
forests and are shy and secretive (Collar & Robson, 2007).
In paper II, we find a surprising phylogenetic affiliation of Neumann´s warbler from the
Albertine Rift valley in eastern Africa. This warbler is nested within the Old World family Cettidae,
which consists mainly of the genera Cettia and Tesia, with some other additions. An attempt to
explain this pattern from a biogeographical point of view is further discussed below.
In paper III, we estimate the phylogenetic relationships of the clade containing the genera
Turdoides, Kupeornis and Phyllanthus (paper I ) and Garrulax cinereifrons (based on unpublished
data). Turdoides is revealed to be paraphyletic. This large genus forms two major clades, one
comprises most of the African taxa, whereas the other contains most of the Asian-Arabian taxa.
Interestingly, Garrulax cinereifrons is inferred to be sister to T. malcolmi, and T. nipalensis is
inferred to be sister to the Kupeornis-Phyllanthus clade.
Phylogenetic Biogeography
The surprising phylogenetic position of Hemitesia neumanni inferred in paper II poses two
intriguing biogeographical questions, how did this aberrant pattern arise, and which underlying
processes have played the most important role in this case? We propose two competing
explanations, where either a long-distance dispersal has occurred from Asia to Africa which resulted
in this sole African member of the clade, or that the historical geographic distribution of the
ancestor to Hemitesia and its closely related Cettidae taxa covered an area comprising both Africa
and Asia and where habitat fragmentations caused a vicariance scenario in combination with local
extinctions. We suggested that the latter explanation is most plausible, although long-distance
migration is found in closely related species, such as Urosphena squameiceps which migrates from
Japan, north-east China and adjacent countries to Southeast Asia (Collar & Robson, 2007).
In paper III, we investigate a similar pattern, but where the investigated clade is more
diverse, and the African portion is of roughly equal size as the Asian part. Further, a small number
of taxa in this clade are found in the Middle East. This, in combination with the African and Indian
representatives makes this an interesting group for studying the role of the Middle East and local
extinctions connected to the late Miocene aridifications, something also adressed in paper II. The
7
debate on the relative importance of dispersal or vicariance adressed above has put much light on
biogeographic events. In paper II and III, we discuss the role of extinctions connected to the
cooling and drying during the end of the Miocene. We find no firm support for this in paper II,
although we find indications of these processes in paper III. Many recent papers on biogeography
adress the question dispersal or vicariance? (e.g. Sanmartín, 2003; Gorog et al., 2004; Bartish et
al., 2010). Our paper suggests that much attention should also be paid on extinctions. With respect
to the Turdoides clade in paper III, we find that the Middle East seems to have played an important
role initially, but as a result of extinctions there is low diversity in this area today. We discuss the
role of the mid-Miocene forest cover over Africa and Asia and land-bridges in the Middle East as
prerequisites for past range extensions and dispersals. I think it is relevant to think of the Middle
East and adjacent areas as an important region, which role for the present pattern of African and
Asian diversity may be strongly underestimated. Further, the large portion of open habitat taxa in
the Turdoides clade, compared to the other clades in Timaliidae, and the timing of its initial
radiations resembles those of some other groups. Among others, these are the tyrant flycatchers of
the New World (Ohlson et al., 2008), and the chameleons of southern Africa (Tolley et al., 2008),
which radiated and colonized open habitats during the global shift from C3 to C4 plants during the
mid to late Miocene.
A New Method in Historical Biogeography
In paper IV, a Bayesian approach to infer the historical biogeography under an event-based
model, given a phylogeny and taxon distributions, is proposed. Among the main differences from
available methods, this approach handles total histories instead of stepping through the nodes of a
phylogeny or “pruning” a phylogenetic tree. This enables details of the histories to be less likely to
have happened but which instead can lead to a history which is overall more likely, as the
probability of the total history is considered.
To implement the biogeographic events in the analysis, we need a model on which the
biogeographic properties of a history depend. Such a model is proposed as a collection of arguments
(Φ; equation 2 in paper IV) from which the events for each node can be coded, given a history and
the taxon distributions. These arguments concord with the definitions of biogeographical events,
and the model Φ is shown as pseudo-code, to facilitates the usage of the model in other applications
by other analysts. Compared to the DEC model (Ree & Smith, 2005; Ree et al., 2008) and DIVA
(Ronquist, 1997, 1998) our arguments accept polytomies which is, according to my opinion, an
important advantage although phylogenetic uncertainties are not taken into account. Further, the Φmatrix can either invoke the four events used in DIVA, or extend these with the division of dispersal
to jump dispersal and range extension, and extinction to local extinction and range contraction, as
invoked in the DEC model.
To use Bayesian inference to estimate the posterior probability of an history given the data
and the model, we include the model in the likelihood function and we make two assumptions: we
assume the ranges in each node of a history to be multinomially distributed, and we assume the
events in a history to be independent. If we accept these assumptions, we can use the proportionality
in Bayes´ theorem to explore the probability distributions. If we denote any target distribution π(•),
8
then the posterior probability distribution of a history (h) given some data (D) and the arguments is
proportional to the likelihood function of the data conditioned on the history, the arguments and
some stochastic parameters and the target prior probability distribution as
h∣ D ,  ∝ f  D∣h ,  , 
,
(2)
where f(•) is the likelihood function.
The advantages, or disadvantages, of Bayesian reasoning is the prior. Typically, a prior
distribution which is conjugate to the posterior distribution is chosen. The conjugate to the
multinomial distribution is the Dirichlet distribution (Fergusson, 1973), which is frequently used in
many field of Bayesian inference, not at least in biological applications such as in phylogenetic
analyses (e.g. Huelsenbeck & Ronquist, 2001; Drummond & Rambaut, 2007).
©MrEnt
To be able to implement variables from a Dirichlet distribution in our method, we need to
sample a number of variables equal in numbers to the number of event types. This can be made in
various ways, and we useNicatoridae
a stick-breaking process (fig 3), where randomly chosen breaking points
Alaudidae
Panurus
"Sphenoeacus
group"
Pycnonotidae
Cisticolidae
Hirundinidae
Acrocephalidae
on an imagined stick of the
length
1
(one
breaking point less than the desired number of variables)
Pnoepygidae
Locustellidae
Donacobidae
Bernieridae
Aegithalidae
Cettidae
Phylloscopidae
generates some variablesHylia
(the pieces of the broken stick). In our case these breaking points are
Erythrocercus
Sylviidae
Zosteropinae
Timaliinae
Pellorneinae
Leiothrichinae
drawn from an equal distribution. The pieces of the stick represent the variables, and those are
ordered in concordance with the assumed order of event types applied in our method (pdispersal <
pextinction < ppersistence < pvicariance).
0
θ3
θ4
θ1
K1
K2
θ2
1
K3
Fig. 3. Visualization of a stick-breaking process on an
imagined stick of the length 1, where four variables
ordered θ1–θ2 are generated on a 3-simplex (K1–K3).
CONCLUSIONS
One of the main findings in this thesis is the existence of five well supported major lineages
of babblers, each one proposed to be referred to at family or subfamily level. The exclusion of the
genus Pnoepyga and proposal of the new family Pnoepygidae, as well as the inferred phylogenetic
positions of the genera Myzornis, Kupeornis, Phyllanthus and Hemitesia and the species Turdoides
nipalensis and Garrulax cinereifrons add significantly to the general knowledge of babblers. Of the
biogeographical findings presented in this thesis, the hypothetical scenario that includes vicariance
in combination with local extinctions as an explanation for the origin of Hemitesia is of obvious
interest, especially when a similar history is inferred for the Turdoides clade. The importance of the
Middle East for a group that is today mostly found outside this region is intriguing, and this
illustrates the importance of considering extinctions when reconstructing historical biogeography.
9
One additional contribution of this thesis to the field of biogeography is the methodological
improvement of event-based biogeography by describing a method that takes the total histories into
account and by using Bayesian inference in a such context.
The relative role of different biogeographical events may vary extensively between clades
and areas. In my opinion persistence, vicariance and dispersal are the general underlying processes
of present distribution patterns, but these patterns might have been extensively deformed or
delimited by local extinctions.
REFERENCES
Akaike, H. (1974) Information theory as an extension of
U. (2011) Vicariance or long-distance dispersal:
the maximum likelihood principle. In B. N.
historical biogeography of the pan-tropical
Petrov & F. Csaki (Eds) Second International
subfamily Chrysophylloideae (Sapotaceae).
Symposium on Information Theory. Budapest:
Journal of Biogeography, 38(1): 177–190.
Akademiai Kiado.
Beaumont, M. A. & Rannala, B. (2004) The Bayesian
Alström, P., Fjeldså, J., Fregin, S. & Olsson, U. (2011a)
revolution in genetics. Nature Reviews Genetics,
Gross morphology betrays phylogeny - the
5: 251–261.
Scrub Warbler Scotocerca inquieta is not a
Beecher, W. J. (1953) A phylogeny of the Oscines. Auk,
cisticolid. Ibis, 153: 87–97.
70: 270–333.
Alström, P., Höhna, S., Gelang, M., Ericson, P. G. P. & Beresford, P., Barker, F. K., Ryan, P. G. & Crowe, T. M.
Olsson, U. (2011b) Non-monophyly and
(2005) African endemics span the tree of
intricate morphological evolution within the
songbirds (Passeri): molecular systematics of
avian family Cettidae revealed by multilocus
several evolutionary ´enigmas´. Proceedings of
analysis of a taxonomically denselly sampled
the Royal Society of London. B, 272: 849–858.
dataset. BMC Evolutionary Biology, 11: 352.
Buerki, S., Forest, F., Alvarez, N., Nylander, J. A. A.,
Alström, P, Rasmussen, P.C., Olsson, U., & Sundberg, P.
Arrigo, N. & Sanmartín, I. (2011) An evaluation
(2008) Species delimitation based on multiple
of new parsimony-based versus parametric
criteria: the Spotted Bush Warbler Bradypterus
inference methods in biogeography: a case study
thoracicus complex (Aves, Megaluridae).
using the globally distributed plant family
Zoological Journal of the Linnean Society, 154:
Sapindaceae. Journal of Biogeography, 38(3):
291–307.
531–550.
Alström, P, Olsson, U., Rasmussen, P.C., Yao, C-t,
Chesser, R. T. & Zink, R. M. (1994) Modes of speciation
Ericson, P.G.P. & Sundberg, P. (2007)
in birds: A test of Lynch´s method. Evolution.
Morphological, vocal and genetic divergence in
48: 490–497.
the Cettia acanthizoides complex (Aves:
Cibois, A. (2003) Mitochondrial DNA phylogeny of
Cettiidae). Zoological Journal of the Linnean
babblers (Timaliidae). Auk, 120: 35–54.
Society, 149: 437–452.
Cibois, A. (2003b) Sylvia is a babbler: taxonomic
Alström, P., Ericson, P. G. P., Olsson, U. & Sundberg, P.
implications for the families Sylviidae and
(2006) Phylogeny and classification of the avian
Timaliidae. Bulletin of the British Ornithologist
superfamily Sylvioidea. Molecular
´s Club, 123: 257–261.
Phylogenetics and Evolution, 38: 381–397.
Cibois, A., Gelang, M. & Pasquet, E. (2010) An
Barker, F. K., Barrowclough, G. F. & Groth, J. G. (2002)
overview of the babblers and associated groups.
A phylogenetic hypothesis for passerine birds:
Systematic notes on Asian birds: 68.
taxonomic and biogeographic implications of an Cibois, A., Pasquet, E. & Schulenberg, T. (1999)
analysis of nuclear DNA sequence data.
Molecular systematics of Malagasy babblers
Proceedings of the Royal Society of London. B,
(Passeriformes: Timaliidae) and warblers
269: 295–308.
(Passeriformes: Sylviidae), based on
Barker, F. K., Cibois, A., Schikler, P., Feinstein, J. &
cytochrome b and 16S rna sequences. Molecular
Cracraft, J. (2004) Phylogeny and
Phylogenetics and Evolution, 13: 581–595.
diversification of the largest avian radiation.
Cibois, A., Slikas, B., Schulenberg, T. S. & Pasquet, E.
Proceedings of the National Academy of
(2001) An endemic radiation of Malagasy
Sciences, USA, 101:11040–11045.
songbirds is revealed by mitochondrial DNA
Bartish, S., Antonelli, A., Richardson, J. E. & Swensson,
sequence data. Evolution, 55: 1198–1206.
10
Clarke, J. R., Ree, R. H., Alfaro, M. E., King, M. G.,
126–138.
Wagner, W. L. & Roalson, E. H. (2008) A
Fergusson, T. S. (1973) A Bayesian analysis of some
comparative study in ancestral range
nonparametric problems. The Annals of
reconstruction methods: retracing the uncertain
statistics, 1(2): 209–230.
histories of insular lineages. Systematic Biology. Fernández-Palacios, J. M., Nascimento, L. de, Otto, R.,
57(5): 693–707.
Delgado, J. D., García-del-Rey, E., Arévalo, J.
Collar, N. J. & Robson, C. (2007) Family Timaliidae
D. & Whittaker, R. J. (2011) A reconstruction of
(Babblers). In J. del Hoyo, A. Elliott & D. A.
Palaeo-Macronesia, with particular reference to
Christie (Eds.) Handbook of the Birds of the
the long-term biogeography of the Atlantic
World Vol. 12. Picathartes to tits and
laurel forest. Journal of Biogeography, 38: 226–
chickadees. (pp. 70–291). Barcelona: Lynx
246.
Edicions.
Flower, B. P. & Kennett, J. P. (1994) The middle
Cox, C. B. & Moore, P. D. (2006) Biogeography: an
Miocene climatic transition: East Antarctic ice
ecological and evolutionary approach, 7th
sheet development, deep ocean circulation and
edition. Blackwell publishing Ltd.UK.
global carbon cycling. Palaeogeography
Delacour, J. (1946) Les timaliinés. L`Oiseau, 16: 7–36.
Palaeoclimatology Palaeoecology, 108: 537–
Delacour, J. (1950) Les timaliinés: additions et
555.
modifications. L`Oiseau, 20: 186–191.
Geyer, C. J. (1991) Markov chain Monte Carlo
Deignan, H. G. (1964) Timaliinae, Panurinae,
maximum likelihood. In Keramidas (ed.),
Picathartinae. In: E. Mayr & R. A. Paynter Jr.
Computing Science and Statistics: Proceedings
(Eds.) A check-list of birds of the World: a
of the 23rd Symposium on the Interface. Interface
continuation of the works of James L. Peters.
Foundation, Fairfax Station, pp. 156–163.
(pp. 240–442). Massachusetts: Museum of
Geyer, C. J. & E. A. Thompson. (1995) Annealing
Comparative Zoology, Cambridge.
Markov chain Monte Carlo with applications to
Dickinsson, E. C (Ed.) (2003) The Howard and Moore
ancestral inference. Journal of the Amer.ican
complete check-list of the birds of the World.
Statistical Association, 90: 909–920.
London: Christopher Helm.
Gorog, A. J., Sinaga, M. H., & Engstrom, M. D. (2004)
Dong, F., Wu, F., Liu, L-M., Yang, X-J. (2010a)
Vicariance or dispersal? Historical biogeography
Molecular phylogeny of the barwings (Aves:
of three Sunda shelf murine rodents (Maxomys
Timaliidae: Actinodura), a paraphyletic group,
surifer, Leopoldamy sabanus and Maxomys
and its taxonomic implications. Zoological
whiteheadi). Biological Journal of the Linnean
Studies. 49(5): 703–709.
Society, 81(1): 91–109.
Dong, F., Li, S-H. & Yang, X-J. (2010b) Molecular
Green, P. (1995) Reversible jump Markov chain Monte
systematics and diversification of the Asian
Carlo computation and Bayesian model
scimitar babblers (Timaliidae, Aves) based on
determination. Biometrica, 82: 711–732.
mitochondral and nucear DNA sequences.
Hall, R. (2009) SE Asia's changing palaeogeography.
Molecular Phylogenetics and Evolution, 57(3):
Blumea, 54: 148–161.
1268–1275.
Hall, R., van Hattum, M.W.A, Spakman, W. (2008) The
Drummond, A. J. & Rambaut, A. (2007) BEAST:
impact of India-Asia Collision on SE Asia: the
Bayesian evolutionary analysis by sampling
record in Borneo. Tectonophysics, 451: 366–
trees. BMC Evolutionary Biology, 7: 214.
389.
Ericson, P. G. P., Christidis, L., Cooper, A., Irestedt, M., Hartert, E (1910) Die Vögel der Paläarktischen Fauna.
Jackson, J. & Norman, J. A. (2002a) A
Berlin: R. Friedlander.
Gondwanan origin of the passerine birds
Hastings, W. K. (1970) Monte Carlo sampling methods
supported by DNA sequences of the endemic
using Markov chains and their applications.
New Zealand wrens. Proceedings of the Royal
Biometrika 57: 97–109.
Society of London, B., 269: 235–241.
Huelsenbeck, J.P., Ronquist, F., (2001) MrBayes:
Ericson, P. G. P., Christidis, L., Irestedt, M. & Norman,
Bayesian inference of phylogeny.
J. A. (2002b) Systematic affinities of the
Bioinformatics 17, 754–755.
lyrebird (Passeriformes: Menura), with a novel Huelsenbeck, J. P., Ronquist, F., Nielsen, R. & Bollback,
classification of the major groups of passerine
J. P. (2001) Bayesian inference of phylogeny
birds. Molecular Phylogenetics and Evolution,
and its impact on evolutionary biology. Science,
25: 53–62.
294: 2310–2314.
Eicson, P. G. P. & Johansson, U. S. (2003) Phylogeny Irestedt, M., Ohlson, J. I., Zuccon, D., Kjällersjö, M. &
of Passerida (Aves: Passeriformes) based on
Ericson, P. G. P. (2006) Nuclear DNA from old
nuclear and mitochondrial sequence data.
collections of avian study skins reveals the
Molecular Phylogenetics and Evolution, 29:
evolutionary history of the Old World
11
suboscines (Aves, Passeriformes). Zoologica
Phylogenetics and Evolution, 41: 556–565.
Scripta, 35(6): 567–580.
Olsson, U., Alström, P., Ericson, P.G.P. & Sundberg, P.
Jacobs, B. I., Kingston, J. D. & Jacobs, L. L. (1999) The
(2005) Non-monophyletic taxa and cryptic
origin of grass-dominated ecosystems. Annals of
species - evidence from a molecular phylogeny
the Missouri Botanical Garden, 86: 590–643.
of leaf-warblers (Phylloscopus, Aves).
Johansson, U. S., Fjeldså, J. & Bowie, R. C. K. (2008)
Molecular Phylogenetics and Evolution, 36:
Phylogenetic relationships within Passerida
261–276.
(Aves: Passeriformes): a review and a new
Olsson, U., Alström, P. & Sundberg, P. (2004) Nonmolecular phylogeny based on three nuclear
monophyly of the avian genus Seicercus (Aves:
intron markers. Molecular Phylogenetics and
Sylviidae) revealed by mitochondrial DNA.
Evolution, 48: 858–876.
Zoologica Scripta, 33: 501–510.
Larget, B. & Simon, D. L. 1999. Markov chain Monte Pasquet, E., Bourdon, E., Kalyakin, M. V. & Cibois, A.
Carlo algorithms for the Bayesian analysis of
(2006) The fulvettas (Alcippe, Timaliidae,
phylogenetic trees. Molecular Biology and
Aves): a polyphyletic group. Zoologica
Evolution, 16: 750–759.
Scripta, 35(6): 559–566.
Luo, X., Qu, Y. H., Han, L. X., Li, S. H. & Lei, F. M.
Reddy, S. & Moyle, R. G. (2011) Systematics of the
(2008) A phylogenetic analysis of
scimitar babblers (Pomatorhinus: Timaliidae):
laughingtrushes (Timaliidae: Garrulax) and
phylogeny, biogeography, and species limits of
allies based on mitochondrial and nuclear DNA
four species complexes. Biological Journal of
sequences. Zoologica Scripta, 38: 9–22.
the Linnean Society, 102(4): 846–869.
Mandaville, J. P. (1997) Plants. In Harrison, D. L. (Ed.) Ree, R. H., Moore, B. R., Webb, C. O. & Donoghue, M.
The scientific results of the Oman flora and
J. (2005) A likelihood framework for inferring
fauna survey 1975. Journal of Oman Studies.
the evolution of geographic range on
Special reports, 1: 229–267.
phylogenetic trees. Evolution, 59(11): 2299–
Mayr, E. & Amadon, D. (1951) A classification of recent
2311.
birds. American Museum Novitates, no. 1496. Ree, R. H. & Smith, S. A. (2008) Maximum likelihood
Metropolis, N. Rosenbluth, A. W., Teller, A. H. and
inference of geographic range evolution by
Teller, E. (1953) Equation of state calculations
dispersal, local extinction, and cladogenesis.
by fast computing machines. Journal of
Systematic Biology, 57(1): 4–14.
Chemical Physics, 21: 1087–1092.
Retallack, G. J. (1992) Middle Miocene fossil plants
Moyle, R. G., Filardi, C. E., Smith, C. E. & Diamond, J.
from Fort Terna (Kenya) and evolution of
(2009) Explosive Pleistocene divirsification and
African grasslands. Paleobiology, 18: 383–400.
hemispheric expansion of a “great speciator”. Ronqusit, F. (1996) DIVA version 1.2. Computer
Proceedings of the National Academy of
program and manual available via
Sciences, 106(6): 1863–1868.
ftp.systbot.uu.se.
Moore, B. R. & Donoghue, M. J. (2009) A Bayesian
Ronquist, F. (1997) Dispersal-vicariance analysis: a new
approach for evaluating the impact of historical
approach to the quantification of historical
events on rates of diversification. Proceedings
biogeography. Systematic Biology, 46(1): 193–
of the National Academy of Sciences, 106(11):
203.
4307–4312.
Ronquist, F. (1998) Three-dimensional cost-matrix
Nelson, G. & Platnick, N. I. (1980) A vicariance
optimization and maximum cospeciation.
approach to historical biogeography.
Cladistics, 14: 167–172.
BioScience, 30(5): 339–343.
Rögl, F. (1998) Palaeogeographic considerations for
Nylander, J. A. A., Olsson, U., Alström, P. & Sanmartin,
Mediterranean and Paratethys seaways
I. (2008a) Accounting for phylogenetic
(Oligocene to Miocene). Annalen des
uncertainty in biogeography: A bayesian
Naturhistorischen. Museums in Wien, 99A: 279–
approach to dispersal-vicariance analysis of the
310.
thrushes (Aves: Turdus). Systematic Biology.
Sanmartín, I. (2003) Dispersal vs. vicariance in the
57(2): 257–268.
Mediterranean: historical biogeography of the
Ohlson, J. I., Fjeldså, J. & Ericson, P. G. P. (2008) Tyrant
Palearctic Pachydeminae (Coleoptera,
flycatchers coming out int the open: phylogeny
Scarabaeoidea). Journal of Biogeography, 30:
and ecological radiation of Tyrannidae (Aves,
1883–1897.
Passeriformes). Zoologica Scripta, 37: 315–335. Sanmartín, I. Van der Mark, P. & Ronquist, F. (2008)
Olsson, U., Alström, P., Gelang, M., Ericson, P. G. P. &
Inferring dispersal: A Baysian approach to
Sundberg, P. (2006) Phylogeography of
phylogeny-based island biogeography, with
Indonesian and Sino-Himalayan region bush
spcial reference to the Canary islands. Journal
warblers (Cettia, Aves). Molecular
of Biogeography, 35(3): 428-449.
12
Sibley, C. G. & Ahlquist, J. E. (1990) Phylogeny and
classification of birds: a study in molecular
evolution. New Haven, CT: Yale University
Press.
Tolley, K. A., Chase, B. M. & Forest, F. (2008)
Speciation and radiations track climate
transitions since the Miocene Climatic
Optimum: a case study of southern African
chameleons. Journal of biogeography, 35:
1402–1414.
Utescher, T., Djordjevic-Milutinovic, D., Bruch, A. &
Mosbrugger, V. (2007) Palaeoclimate and
vegetation change in Serbia during the last 30
Ma. Palaeogeography, Palaeoclimatology,
Palaeoecology, 253: 226–250.
Via, S. (2001) Sympatric speciation in animals: the ugly
duckling grows up. Trends in Ecology and
Evolution. 16(7): 381–390.
Vrba, E. S. (1993) Mammal evolution in the African
Neogene and a new look at the great American
interchange. In Goldblatt, P. (ed.) Relationships
between Africa and South America: 393–434.
New Haven, CT: Yale University Press.
Vrielnyck, B., Odin, G. S. & Dercourt, J. (1997)
Miocene palaeogeography of the Tethys ocean;
potential global correlations in the
Mediterranean. In Montanari, A., Odin, G. S. &
Cocconi, R (Eds.) Miocene Stratigraphy: an
integrated approach, 157–165. Amsterdam,
Elsevier Science.
Wiley, E. O. (1988) Vicariance biogeography. Annual
Review of Ecology and Systematics. 19: 513–
542.
Yang, Z. and Rannala, B. 1997. Bayesian phylogenetic
inference using DNA sequences: A Markov
chain Monte Carlo method. Molecular Biology
and Evolution, 14: 717–724.
Zachos, J., Pagani, M., Sloan, L., Thomas, E. & Billups,
K. (2001) Trends, rythms, and aberrations in
global climate 65 Ma to present. Science.
292(27): 686–693.
Zink, R. M., Blackwell-Rago, R. C. & Ronquist, F.
(2000) The shifting roles of dispersal and
vicariance in biogeography. Proceedings of the
Royal Society of London, B. 267: 497–503.
13
ACKNOWLEDGEMENTS
I am deeply grateful to my two supervisors, Per Ericson and Per Alström, who have guided me,
supported me and over all embellished my time during the PhD project. Your contribution to the
positive atmosphere in the research group, to the outlines and content of this thesis, and to our
friendly relationships is huge. Your two quite different personalities supplemented each others in a
really good way, and brought a great width to the supervision. Thank you!!!
I want to pass the warmest thanks to the coauthors of the papers in this thesis. Your contributions
have been, and continue to be crucial. Your patience with my extreme lack of speed has also been
extraordinary! Thank you Alexis Bohlin, Alice Cibois, Eric Pasquet, George Sangster, Martin
Irestedt, Per Alström, Per Ericson and Urban Olsson! I am also grateful to Sören Nylin, Mari
Källersjö and Kjell-Arne Johanson, the members of my review commitee. You came up with a lot of
good advices, and you made the annual reports to both friendly and concise meetings.
My work and my time on the museum would have been nothing at all without the wonderful
hospitality from my collegues. Thanks of you, every hour in the lab, in field, on conferences,
birding, at the pizzeria or at the pub have been a pure pleasure! Thank you Jan Ohlson, Martin
Irestedt, Ulf Johansson, George Sangster, Johan Dalsätt, Te-Yu Liao and Dario Zuccon! I wish that
I have had the possibility to spend more time with the staff at the Vertebrate Department. The good
atmosphere at Verte, with all the various discussions, jokes and comments at the lunch room, and
the kind-hearted friendship among the entire staff is something I look back on with particular
delight. Thank you Peter Nilsson, Peter Mortensson, Ingrid Cederholm, Nisse Jacobsson, Bosse
Delling, Ulf Johansson, Ann-Katrin Gustafsson, Sven Kullander, Erik Åhlander, Bodil Kajrup,
Thord Fransson, as well as the former staff Göran Frisk, Jonas Nordin, Jonathan Ready among
others! I have spent days and nights in the Molecular Systematics Laboratory at the museum. Your
hospitality and support has been enormous! Thank you Pia Eldenäs, Ewa Sjödin, Love Dahlen,
Bodil Cronholm, Keyvan Mirbakhsh, Martin Irestedt and others, including all of you Post Docs and
PhD´s who passed through the lab! I also want to thank the friends in Copenhagen, Jon Fjeldså,
Knud Jønsson, Pierre-Henri Fabre for great discussions and nice birding moments etc. Knud is
especially thanked for keeping an eye open for treacherous topez!
This thesis would not have been possible to write without all samples obtained from museums and
other institutions. I am indebted to Göran Frisk and Ulf Johansson (Swedish Museum of Natural
History), Eric Pasquet (Muséum National d´Histoire Naturelle, Paris), Alice Cibois (Natural History
Museum of Geneva), Jon Fjeldså and Jan Bolding Kristensen (Zoological Museum, Copenhagen),
Silke Fregin (Ernst-Moritz-Arndt-Universität Greifswald), Mikhail Kalyakin (Zoological Museum
of Moscow), David Willard (Field Museum, Chicago), Sharon Birks (University of Washington
Burke Museum), Kristof Zyskowski (Peabody Museum of Natural History and Yale University),
Michel Louette (Royal Museum for Central Africa, Tervuren), Storrs Olson (Museum of Natural
History, Smithsonian Institution) and Urban Olsson (Göteborg University). I also wish to express
14
my warmst regards to my Indonesian friends Irham Mohammad and Dewi Praviradilaga, for nice
times in field and a great support regarding permissions etc.
A number of friends have made serious contributions to this thesis, in a wider sense. First of all, I
like to thank Ola Marklund with whom I started my scientific career by playing in our gardens as
kids and collecting insects, various kinds of feces etc., which later turned to intensive birding
around the globe. I am pleased to thank Johannes Persson and Joakim Fagerström who I spent a lot
of time with, both in field, while studying and not the least by sharing uncountable hours of good
music! I also thank my parents and my sister, Pelle, Gittan and Johanna. You have supported me
throughout my life, and we have shared many moments together that have lead to my PhD in one
way or the other. An extra thanks to Pelle for your 24-7-support on computer questions and
programming problems!
Jan Ohlson and Jenni Andersson – your hospitality is beyond any comparison! I cannot even figure
out how much you have contributed to this thesis. First of all by letting me stay in your house for
hundreds of nights, and also for beeing such good friends in any situation and by inducing a
considerable increase of my egg-and-bacon consumption!
Last, but far from least, I thank my beloved family. You have had an endless patience, and you have
played an almost unvisible but still the most prominent part of this thesis. Above all, I would like to
thank you for reminding me about the most important things in life – far beyond scientific papers
and impact factors! Bettina, Hanna-Lina, Leo och Maja, TACK för att ni finns och för att ni är så
himla goa!!!
15