Molecular phylogeny and biogeography of ranoid frogs

Dissertation zur Erlangung des akademischen Grades des Doktors der Naturwissenschaften (Dr. rer. nat.)

an der

Universität Konstanz, Mathematisch- Naturwissenschaftliche Sektion

Fachbereich Biologie

vorgelegt von

Arie van der Meijden, M.Sc.

Konstanz, März 2006

Prüfungskommission:

Prof. Iwona Adamska Prof. Axel Meyer Prof. Miguel Vences

..love for all living creatures [is] the most noble attribute of man..

Charles Darwin (1809-1882), in “The Descent of Man and Selection in Relation to Sex” (1872).

2

Acknowledgements

First and foremost I would like to thank Prof. Axel Meyer for hiring me as a technician, and for subsequently showing his confidence in me by allowing me to start my PhD work in his lab. Working in his lab, I have acquired knowledge and skills indispensable to further pursue a career at the high level of biology that Prof. Meyer so adamantly exemplifies, and for this I am very grateful. No less essential to the realization of this thesis was the guidance of Prof. Miguel Vences, who installed in me a purposeful mode of research, and who showed me the field of amphibian biology as more vibrant, dynamic and enticing as I could have thought. I am greatly indebted to my collaborators, without whom much of my research would not have been possible, or the results greatly diminished. Apart from Axel Meyer and Miguel Vences, I am therefore thankful to (alphabetical oder) Renaud Boistel, Allan Channing, Ylenia Chiari, Justin Gerlach, Simone Hoegg, Annemarie Ohler, Meike Thomas and David Vieites. Since my first arrival in Konstanz on October 15th 2002, I have had the privilege of learning lab and analysis techniques from the skilled people who make up the Meyer-lab. In particular I would like to mention Simone Hoegg, Dirk Steinke and Elke Hespeler for their patience with my initial ignorance. Outside the realm directly related to my work, I have many people to thank for making the time I spent in Konstanz a lot of fun. The people of the Meyer-lab form a tightly knit community with whom I greatly enjoyed talking and partying with. I would like to thank Ylenia Chiari here in particular, for discussions of work and many other things, and for weathering all the fun and not-so-fun times together with me.

This thesis is a direct result of my incurable affinity to biology, and herpetology in particular, since an early age. Many people have suffered at my hands because of this and I would like to take this space to thank them: Most of all, I could not have achieved this without the unwavering and full support from my parents. Despite initially poor school results, astronomical electricity bills, lizards in the curtains, snakes in the kitchen, and numerous other escapees, my parents supported

3

my interests. My sister too patiently endured her share of unexpected encounters around the house. My travels would certainly have never been without my family supporting me. For the support, understanding and guidance of my family I am therefore especially grateful. To all my friends who, despite my traveling, kept close contact and who made my every visit to the Netherlands a joy, and who showed or feigned interest all those times I told them something “interesting” about animals, I am very grateful. Special thanks to Lennart Pors for printing and binding this thesis. I would like to thank Arendo Flipse for putting a steady stream of herpetological biodiversity into my very hands for over twelve years, and for teaching me a lot about herpetology.

To all these people and many more I am greatly indebted, and to all I dedicate this thesis.

4

Table of Contents

General Introduction ______7

Amphibian declines______7 Continuous discovery ______8 Phylogenetic theory as a framework for evolutionary discovery ______9 An introduction to frogs in the superfamily Ranoidea ______10 Biogeography of the Ranoidea ______12 Taxonomic difficulties owing to homoplasious characters______13 Structure of this thesis______13

1. Novel phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear rag-1 gene____ 16

1.1 Abstract______16 1.2 Introduction______17 1.3 Materials and methods ______18 1.4 Results______19 1.5 Discussion______20 Acknowledgements ______23

2. Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians______25

2.1 Abstract______25 2.2 Introduction______26 2.3 Materials and methods ______27 2.4 Results______29 2.5 Discussion______37 Acknowledgements ______41

3. A previously unrecognised radiation of ranid frogs in Southern Africa revealed by nuclear and mitochondrial DNA sequences ______43

3.1 Abstract______43 3.2 Introduction______44 3.3 Materials and methods ______47

5

3.4 Results______51 3.5 Discussion______54 3.6 Conclusion ______59 Acknowledgements ______60

4. Molecular phylogenetic evidence for paraphyly of the genus Sooglossus, with the description of a new genus of Seychellean frogs ______62

4.1 Abstract______62 4.2 Introduction______63 4.3 Materials and methods ______65 4.4 Results______69 4.5 Discussion______72 Acknowledgements ______79

5. Nuclear gene phylogeny of narrow-mouthed toads, family Microhylidae, and a discussion of competing hypotheses for their origin ______81

5.1 Abstract______81 5.2 Introduction______81 5.3 Materials and methods ______84 5.4 Results______88 5.5 Discussion______91 Acknowledgements ______98

Summary ______99

Zusammenfassung______102

General references ______106

Appendix 1______116

Appendix 2______117

Results produced by collaborators ______119

6 General Introduction

General Introduction

In this thesis I provide the first broad molecular phylogenetic hypotheses for well known ranoid frog groups, in particular the Brevicipitinae (chapter 1), the Ranidae (chapter 3) and the Microhylidae (chapter 5). Also included in this thesis is a phylogenetic study of the enigmatic Sooglossidae, basal neobatrachians endemic to the Seychelles (chapter 4). I discuss the taxonomic and biogeographic consequences of the recovered phylogenies. Furthermore the relative merit of the genes 16S and CO1 for genetic barcoding of amphibians is discussed (chapter 2).

Amphibian declines Since the initial identification of worldwide enigmatic amphibian declines at the world congress of herpetology in 1989, more reports on the disappearance of amphibian populations from all parts of the world have been published (Wake, 1991). The global amphibian assessment (GAA) published their findings on this phenomenon in the journal Science (Stuart et al., 2004), showing that a sobering 32.5% of amphibian species are suffering declines. Moreover, although many are declining due to well understood factors such as introduced species (Vredenburg, 2004), overexploitation (Lannoo et al., 1994) and habitat loss, alteration and fragmentation due to human encroachment (Fisher and Shaffer, 1996; Marsh and Pearman, 1997), 48% of the declining populations are doing so due to unknown reasons. Many factors have been proposed as possible causes of these enigmatic declines; climate change (Pounds et al., 1999; Pounds et al., 2006), chemical contaminants (Hayes et al., 2002), UV radiation (Blaustein et al., 2003), infectious diseases (Daszak, 1998) or a combination of these factors. Four families contribute most to the total number of declining species; Bufonidae, Leptodactylidae, Hylidae and Ranidae (figure 1). Most ranoid frogs (Arthroleptidae, Ranidae, Microhylidae among others) suffer very little or not at all from “enigmatic” declines so far, and most of the affected species are suffering from habitat loss and overexploitation (Stuart et al., 2004). Despite the overwhelming attention that has rightfully been given to the declines of

7 General Introduction unknown cause, most species are in decline directly due to human activities and population expansion. There is therefore a pressing need for conservation actions. Preservation of taxa selected from the whole tree of life, designed to preserve as much evolutionary history as possible, has proven to be unnecessary (Nee and May, 1997), but for phylogenetically biased extinctions like those of the amphibians (Purvis et al., 2000), such a strategy might be necessary. To identify lineages rich in unique evolutionary history for conservation, a phylogenetic theory of the clade to be pruned by extinction is necessary.

Figure 1 Percentages and numbers of rapidly declining species in amphibian families (with at least one rapidly declining species), broken into groups reflecting the dominant cause of rapid decline: overexploitation; habitat loss; or enigmatic decline. Taken from Stuart et al. (2004).

Continuous discovery As instances of amphibian decline and possible extinctions are discovered throughout the world, new species are being discovered at the high rate of 25% in the last 11 years

8 General Introduction

(Köhler et al., 2005). The use of molecular techniques to identify genetically distinct lineages facilitates the distinction of sibling species (e.g. Ohler and Delorme, 2006), and the reassignment of known taxa to different taxonomic units (see chapter 4). Despite this, most recent discoveries are due to newly collected and described specimens (Bossuyt et al., 2004). The discovery of a representative of a new anuran family (Biju and Bossuyt, 2003) and a plethodontid salamander from Korea (Min et al., 2005), a high endemic diversity on Sri Lanka (Meegaskumbura et al., 2002), as well as the recent discovery of at least 20 possible new frog species in New Guinea (unpublished at time of writing) show that amphibian biodiversity is far from being fully described. It also illustrates our lack of knowledge of relatively well-studied groups as amphibians. As awareness of both the high undescribed amphibian biodiversity and the high rate of uncurbed decline grows, the need to describe and chart “what is out there now” seems to become more urgent (Wilson, 2002). A promising technique that will allow the identification of genetically distinct and possibly unknown lineages is DNA barcoding (Hebert and Gregory, 2005). If the previsions of the proponents of this technique come to pass, handheld devices will allow workers with relatively little training in taxonomy and species identification to identify known species, and thus also identify unknown lineages. This will certainly accelerate the discovery of new amphibian species.

Phylogenetic theory as a framework for evolutionary discovery The description of evolution as descent with modification from a common ancestor (Darwin, 1859), leads to a dichoto- or polytomous unidirectional branching pattern of evolution. Although this is insufficient to describe the evolutionary process of organisms that engage in horizontal transfer of characters, for most multicellular organisms a bi- or polyfurcating phylogeny describes their evolutionary history well. Knowledge of evolutionary relationships is indispensable to the inference of the direction of acquisition and loss of characters in the course of evolution. As molecular characters, which are largely independent from morphology and biology and therefore a vast source of relatively independent characters, were increasingly used for phylogenetic inference surprising patterns of evolution emerged that were hitherto obscured by ambiguity in morphological datasets (e.g. Zardoya et al., 2003; Zardoya and Meyer, 1996). One group

9 General Introduction of thus far notoriously instable taxonomy due to ambiguity in morphological datasets are the ranoid frogs, particularly the families Ranidae and Microhylidae (Frost, 2004). Although some recent work based on molecular datasets has somewhat alleviated this situation for a small subset of taxa (e.g. Roelants et al., 2004; Vences and Glaw, 2001), broader studies like the ones presented in this thesis (chapters 3 and 5) are increasingly necessary to provide a broad-scale phylogenetic theory.

An introduction to frogs in the superfamily Ranoidea Here follows a brief introduction to the frogs of the superfamily Ranoidea, as their relationships were understood prior to the studies in this thesis. Although a detailed introduction to the issues of taxonomy and biology of the specific groups can be found at the beginning of each chapter, a broad but brief overview is given here. For the only non ranoid neobatrachian family in this thesis, the Sooglossidae, an introduction is given in chapter 4. The Ranoidea are a subclade of the more inclusive monophyletic crown group of frogs, informally known as the Neobatrachia (see figure 2). Other neobatrachian frogs belong to the superfamily Hyloidea, or to several families of uncertain and possibly basal position; the Heleophrynidae, Myobatrachidae, Nasikabatrachidae and Sooglossidae. Ranoidea are classically defined by the presence of a firmisternal pectoral girdle. The Ranoidea superfamily can be divided into three epifamilies:

1. Ranoidae, containing the Ranidae, Rhacophoridae and Mantellidae Ranid frogs are “typical” frogs that can be found leaping from the water’s edge anywhere in Europe, and any lay person asked to describe a frog will probably describe a likeness to a typical ranid. They are therefore known as the “True frogs” and are present on every continent except Antarctica. The family Ranidae has many species (773 species, AmphibiaWeb, accessed February 2006) divided into several highly disputed subgroupings (Frost, 2004). The Ranidae have been suggested to be paraphyletic or polyphyletic relative to the Mantellidae and Rhacophoridae (Ford and Cannatella, 1993). Recent molecular studies have started to elucidate the phylogenetic relationships among the Ranidae (Hoegg et al., 2004a; Roelants et al., 2004), but the taxonomy of this group is still far from truly reflecting the relationships of its constituents.

10 General Introduction

Figure 2 Taxonomic classification of the Neobatrachian groups relevant to this thesis. Evidence that was produced in this thesis is not incorporated in this figure, showing large unresolved polytomies in the studied groups.

The Rhacophoridae (273 species) are mostly arboreal, with a high diversity in Asia, but also represented in Africa and on Madagascar. The Mantellidae (157 species) are endemic to Madagascar and have radiated into both arboreal and terrestrial forms.

2. Arthroleptoidae, containing Arthroleptidae, Astylosternidae, Hemisotidae and Hyperoliidae. All members of this epifamily are restricted to sub-Saharan Africa.

11 General Introduction

Arthroleptids (32 species) and astylosternids (29 species) are both associated with mesic environments. The arthroleptids possess a horizontal pupil and expanded toe pads, which the astylosternids lack. Hemisotids (9 species) are burrowing ant- and termite specialists which care for their tadpoles in an underground chamber before they are led to open water. Many hyperoliids (260 species) are arboreal but some genera are terrestrial. They are often associated with more xeric environments. The leptopelines are traditionally considered to be part of the Hyperoliidae, but new evidence places this group closer to the arthroleptids (Emerson et al., 2000).

3. Microhyloidae, containing only the Microhylidae The Microhylidae (422 species), or narrow mouthed frogs, are mostly terrestrial or fossorial forms, although arboreality has evolved several times in this family. Several subfamilies have direct development, or non-feeding larvae. The larvae have a suite characters unique among frogs. They are distributed circumtropically.

Biogeography of the Ranoidea The Ranoidea superfamily is circumglobally distributed. They are represented on every continent except Antarctica. Of the constituent epifamilies, the Arthroleptoidae (Brevicipitidae) are restricted to sub-Saharan Africa, the Microhyloidae are circumtropically distributed, whereas the Ranoidea enjoy an even wider distribution; they are distributed in the holarctic up to the polar circle, and in the southern hemisphere they are only absent from the southern half of South America, the southern part of Australia, and from New Zealand. Ranoids were initially thought to have originated on Gondwana approximately 140 million years ago (Mya) (Duellman and Trueb, 1986) but more recent molecular clock based divergence time estimates place their initial divergence between 99 Mya (San Mauro et al., 2005) and 69 Mya (Vences et al., 2003b). Current knowledge does not allow for the testing of the various biogeographic scenarios put forth (Duellman and Trueb, 1986; Roelants et al., 2004; Savage, 1973), Although these theories stipulate vicariance due to the breakup of Gondwana and terrestrial dispersal, trans-oceanic dispersal of amphibians cannot be ruled out (Vences et al., 2003b). In chapter 3 and chapter 5, novel biogeographic scenarios based on new data are discussed.

12 General Introduction

Taxonomic difficulties owing to homoplasious characters The Anura show a highly conserved body plan (Shubin and Jenkins, 1995). Although the initial divergence within the Anura took place approximately 230 Mya (Graur and Martin, 2004), their basic body plan has not been altered much. This in contrast to the Mammalia, which since the end of the Cretaceous, 65 Mya, have diverged widely. If the anuran body plan is highly constrained, then variability in morphology is probably due to functional adaptations. Since members of different amphibian groups occupy similar ecological niches in various parts of the world, it is likely that homoplasy due to convergence in some instances has occurred (e.g. Bossuyt and Milinkovitch, 2000). In all investigated groups within the Ranoidea, the here presented molecular studies show relationships that were thus far not suspected or difficult to corroborate based on morphological datasets. This high incidence of disagreement between the molecular and morphological data suggests that ranoid taxonomy has been confounded by a prevalence of homoplasy in the morphological characters. This has been suggested to repeatedly complicate the elucidation of phylogenetic relationships among the Microhylidae (Wild, 1995; Zweifel, 1986). Homoplasy has also led to the placement of the arthroleptoid/brevicipitid Brevicipitinae with the Microhylidae (see chapter 1), and has obscured ranid relationships (chapter 3). Alternative to homoplasy, the mosaic retention of ancestral characters in different taxa due to a rapid radiation can also confound the use of these characters for inference of relationships. Although the morphological characters involved in the spurious placements were not studied further, in some instances inferences about the evolution of homoplasious characters can be made (chapter 5). A more detailed and basally better resolved phylogenetic hypothesis will be necessary to map more of these misleading characters.

Structure of this thesis If, in the light of the current declines and extinctions, we wish to conserve evolutionary history as it is contained in currently living lineages, then knowledge of their phylogeny is a first requirement. Also for the study of evolutionary processes, phylogenetic

13 General Introduction knowledge is the basic framework. In this thesis I have used molecular phylogenetic techniques to alleviate the current lack of knowledge on the relationships within important crown group frog taxa. I also have contributed to the selection of a suitable marker for “DNA barcoding”, which promises to aid in further discovering thus far unidentified lineages. The chapters are arranged in chronological order of publication or submission. Chapter one deals with the microhylid subfamily Brevicipitinae. In our study we show that the rag-1 gene unambiguously places this subfamily outside the Microhylidae, where it had been placed based on morphological characters without discussion. Chapter two shows the superior performance of the 16S gene as opposed to that of CO1 for the novel application of DNA barcoding, at least for amphibians. Chapter three shows the potential of molecular phylogeny for discovering relationships that may not be identifiable using morphological markers. Based on a molecular dataset containing three nuclear and two mitochondrial genes, we recovered a unexpected clade of South African frogs with a possibly basal position in the Ranoidae epifamily, and discuss the biogeographic consequences. Chapter four concentrates on an non-ranoid basal neobatrachian family; the Sooglossidae. These enigmatic Seychellean frogs are organized into two genera. We show that the genus Sooglossus is paraphyletic relative to the genus Nesomantis using two nuclear and one mitochondrial gene and we propose an alternative nomenclature to reflect the new evidence. Chapter five for the first time shows a molecular phylogenetic theory for the circumtropic family Microhylidae. The molecular dataset consisting of four nuclear genes shows novel relationships, and I discuss the taxonomic and biogeographic implications.

14 Chapter 1

Novel phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear rag-1 gene

Breviceps acutirostris

Chapter 1

1. Novel phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear rag-1 gene

Published: Van der Meijden, A., M. Vences and A. Meyer. 2004. Novel Phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear rag-1 gene. Proceeding of the Royal Society B 271: S378-S381.

1.1 Abstract Due to a general paucity of characters and an apparently general high level of homoplasy, the systematics of frogs have remained disputed. A phylogeny based on the single-copy nuclear rag-1 gene revealed unexpected placements of scaphiophrynine and brevicipitine toads. The former have usually been considered as sister group to all other extant microhylids or are even classified as their own family. Their basal position among microhylids was weakly indicated in our analysis; but they clearly were part of a strongly supported clade composed of representatives from five other microhylid subfamilies. In contrast, the brevicipitines, a group that hitherto was unanimously considered to belong to the Microhylidae, were highly divergent and placed as a sister group to the arthroleptoid clade. These novel phylogenetic placements are best reflected by a classificatory status of the Scaphiophryninae as subfamily of the Microhylidae, whereas the brevicipitines may merit recognition as distinct family. Our findings seem to corroborate a high degree of morphological homoplasy in frogs and suggest that even highly derived morphological states, such as the hydrostatic tongue of microhylids, hemisotids and brevicipitines, may be subject to convergent evolution, parallelism or character reversal.

16 Chapter 1

1.2 Introduction Due to their pre-Gondwanan age and cosmopolitan distribution, amphibians are a good model system for the study of biogeography (Duellman and Trueb, 1986; Feller and Hedges, 1998). Their tolerance of salt water is limited; although they are capable of transoceanic dispersal (Vences et al., 2003b) their distribution is likely to have been shaped in great part by vicariance (Duellman and Trueb, 1986). Application of molecular methods to the elucidation of amphibian phylogeny has revealed surprising instances of morphological homoplasy among regional radiations, e.g. of Madagascan and Indian tree frogs (Bossuyt and Milinkovitch, 2000), or Indian and African burrowing frogs (Biju and Bossuyt, 2003). These taxa belong to the Neobatrachia, a monophyletic group that contains the wide majority of the recent frogs (Feller and Hedges, 1998; Hoegg et al., 2004a). Despite the renewed interest in the biogeographic and evolutionary history of anurans, one circumtropic neobatrachian family, the Microhylidae, has so far not been studied through molecular phylogenetic analyses. Although single representatives of this family were included in some works (Biju and Bossuyt, 2003; Feller and Hedges, 1998), the intrafamilial relationships remain unstudied from a molecular perspective. The Microhylidae contains 349 species in 67 genera (excluding scaphiophrynines), occurring in the Americas, sub-Saharan Africa, Madagascar, India and most of southeast Asia to New Guinea and northernmost Australia (AmphibiaWeb.org, as of 2003). Microhylids are defined by a uniquely derived tadpole morpholog (type II of Orton, 1952), by an osteological trend towards reduction of shoulder girdle elements, and by a specialized microphagous feeding behaviour with hydrostatic tongues (Meyers et al., 2004). Among microhylids, the phylogenetic position of the eight species in the subfamily Scaphiophryninae from Madagascar is especially enigmatic. Scaphiophrynine tadpoles are intermediate between Orton's tadpole types II and IV (Orton, 1952), the latter being the generalized neobatrachian type (Wassersug, 1984). Scaphiophrynines were placed within the Ranidae until Guibé (1956) placed them into the Microhylidae. Savage (1973) suggested their inclusion in yet another family, the Hyperoliidae. Dubois (Dubois, 1992) raised them to family rank as Scaphiophrynidae. Another microhylid subfamily of

17 Chapter 1

uncertain affinities is the African Brevicipitinae, or rain frogs, composed of 18 species in five genera. Interestingly, these are the only microhylids in which direct development occurs, posing difficulties for an assessment of their larval features. Here we present the first data on the phylogenetic position of scaphiophrynine and brevicipitine toads using DNA sequences of a single copy nuclear gene, rag-1, which is known to provide an adequate resolution in the analysis of anuran relationships (Hoegg et al., 2004a). Surprisingly our results indicate that brevicipitines might not belong into the microhylid lineage while scaphiophrynines do, contrary to current classification and morphological evidence.

1.3 Materials and methods Taxa were selected to cover major clades among ranoid neobatrachians to which microhylids are known to belong (Biju and Bossuyt, 2003; Hoegg et al., 2004a). We included taxa of six out of the nine microhylid subfamilies accepted by Duellman & Trueb (1986), i.e., all except the Asterophryinae, Genyophryninae and Melanobatrachinae. The archaeobatrachian Xenopus, and several hyloid neobatrachians, were used as hierarchical outgroups. A list of taxa and GenBank accession numbers are given in table 1.1. DNA was extracted from muscle tissue stored at -80°C or fixed in 70% ethanol. Tissue samples were digested using Proteinase K (final concentration 1 mg/mL), homogenised and subsequently purified following a standard salt extraction protocol. We used primers as in Hoegg et al. (2004). PCR was performed in 25 µL reactions containing 0.5-1.0 units of REDTaq DNA Polymerase (Sigma, Taufkirchen, Germany), 0.01 units of Pwo DNA polymerase (Roche, Mannheim, Germany), 50 ng genomic DNA, 10 pmol of each

primer, 15 nmol of each dNTP, 50 nmol additional MgCl2 and the REDTaq PCR reaction

buffer (one fold concentrated: 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.1 mM MgCl2 and 0.01% gelatine). Cycle conditions were adapted from a long range PCR protocol (Barnes, 1994), with an initial denaturation step at 94°C for 5 minutes, followed by ten cycles with 94°C for 30 seconds, annealing temperatures increasing by 0.5°C per cycle from 52 to 57°C and extending for 3 minutes at 68°C. Additional 20 cycles were performed with

18 Chapter 1

94°C for 10 seconds, 57°C for 40 seconds and 68°C for 3 minutes. The final extension was done at 68°C for 5 minutes. PCR products were purified directly via spin columns (QIAGEN). Sequencing was performed directly using the corresponding PCR primers (forward and reverse). DNA sequences of both strands were obtained using the BigDye Terminator cycle- sequencing ready reaction kit (Applied Biosystems Inc.) on an ABI 3100 capillary sequencer using the manufacturer’s instructions. Maximum Parsimony (MP) and Maximum Likelihood (ML) phylogenies were calculated using PAUP* (Swofford, 2002). The best fitting models of sequence evolution for ML analyses were obtained by Modeltest 3.06 (Posada and Crandall, 1998). Heuristic searches were performed using 10 replicates of a stepwise addition of taxa. Robustness of the MP tree topology was tested by bootstrap analysis with 2000 replicates; 500 ML bootstrap replicates were calculated. Bayesian Inference was conducted with MrBayes 2.0 (Huelsenbeck and Ronquist, 2001) using the GTR model with 100,000 generations, sampling trees every 10th generation, (and calculating a consensus tree after omitting the first 5,000 trees) ("burn-in" set at 1000). We tested alternative phylogenetic hypotheses using Shimodaira-Hasegawa tests as implemented in PAUP*, with RELL optimization and 1000 bootstrap replicates. To avoid biases by the previous selection of alternative topologies we applied the SH test simultaneously to all possible unrooted trees in a reduced set of six taxa, containing Breviceps, Scaphiophryne, Bufo regularis as outgroup, and the microhylid, arthroleptoid and ranoid taxa with the shortest branch length each (Plethodontohyla, Kassina, Lankanectes), assuming that short branch lengths indicate a low number of autapomorphies that could mask phylogenetic affinities

1.4 Results The dataset consisted of 1566 DNA positions in 33 species. The trees obtained through MP, ML and Bayesian methods (figure 1.1) subdivide the ranoids into three well supported major clades, corresponding to the epifamilies Ranoidae, Microhyloidae and the Arthroleptoidae as defined by Vences and Glaw (2001). The Ranoidae contained the families Rhacophoridae and Mantellidae, and the paraphyletic Ranidae. Within the

19 Chapter 1

Arthroleptoidae, the hyperoliid Leptopelis is a sister taxon of Arthroleptis, rendering Hyperoliidae paraphyletic with respect to Arthroleptis. Microhylids formed a highly supported clade that contained Scaphiophryne but not Breviceps. Within this clade, relationships were poorly resolved due to very short basal branch lengths, suggesting the possibility of a rapid lineage formation early in the evolution of this group. The ML and Bayesian analyses placed Scaphiophryne as a sister group to the remaining microhylids. However this placement did not receive strong support, and the differentiation of Scaphiophryne within the clade of the remaining microhylids was small as indicated by the short branch lengths between splits in this clade. Breviceps did not group with other microhylids but instead was the sister group of the Arthroleptoidae. All alternative tree topologies reflecting current classification, i.e., Breviceps is part of the Microhylidae while Scaphiophryne is not, were significantly rejected by the SH tests: Plethodontohyla was the sister group of Scaphiophryne and not of Breviceps in the reduced set of taxa analysed (P<0.001). However, maintaining a sister-group relationship of Scaphiophryne and Plethodontohyla, alternative positions of Breviceps could not be significantly excluded; this applied also to its placement as the sister group to the (Scaphiophryne, Plethodontohyla) clade.

1.5 Discussion Larvae of Scaphiophryne are characterized by morphological characters that are considered to be plesiomorphic relative to the highly specialized, suspension-feeding microhylid type (Haas, 2003; Wassersug, 1984). Their derived larval traits define microhylids as monophyletic group to the exclusion of Scaphiophryne. The most parsimonious phylogeny based on this character complex therefore would expect this genus to occupy a distinctly basal position relative to other microhylids. However, what seems clear from the tree shape (figure 1.1) is that scaphiophrynines did not diverge particularly early in microhylid evolution but were one of the major lineages in the initial radiation of these frogs. This indicates a fast evolutionary transition from the Scaphiophryne-like tadpole morphology to a derived microhylid tadpole.

20 Chapter 1

Figure 1.1 Maximum Likelihood tree based on the analysis of 1566 base pairs of the rag-1 gene, highlighting the phylogenetic position of scaphiophrynines (Scaphiophryne) and brevicipitines (Breviceps) among ranoid neobatrachians. The numbers indicate on the branches are bootstrap support values in percent of ML (100 replicates) and MP (2000 replicates) searches. Asterisks placed to the right of nodes indicate Bayesian posterior probabilities >95%. The tree was rooted with Xenopus laevis (not shown). The insert pictures show representatives of the genera Scaphiophryne and Breviceps .

21 Chapter 1

Also, the tree presented by Biju and Bossuyt (2003) was unambiguous in suggesting a sister-group relationship between scaphiophrynines and other microhylids. From a classificatory point of view, this phylogenetic pattern would best be reflected by a status as subfamily of the Microhylidae rather than as separate family. The family Microhylidae, according to our analysis, contains endemic genera from South America (Dermatonotus), Asia (Kaloula), Africa (Phrynomantis) and Madagascar (Dyscophus, Scaphiophryne, Plethodontohyla). The Madagascan taxa were not a monophyletic clade. Instead, Dyscophus grouped with the Asian Kaloula, indicating possible intercontinental relationships parallel to those of the rhacophorid (Asia) and mantellid (Madagascar) tree frogs (Biju and Bossuyt, 2003; Bossuyt and Milinkovitch, 2000). Surprisingly, the African Breviceps (Brevicipitinae) was resolved not as being part of the microhylid clade, but they were grouped with the Arthroleptoidae. This placement differs from conclusions based on morphological and mitochondrial characters (e.g., Emerson et al., 2000). Our data were not sufficient to significantly exclude alternative phylogenetic hypotheses, but the SH tests did significantly exclude the classical hypothesis, in which the brevicipitines are part of the Microhylidae to the exclusion of Scaphiophryne. If confirmed by further data sets, the grouping favoured by our analysis would suggest that the specialized hydrostatic tongue that is characteristic for microhylids including brevicipitines (Meyers et al., 2004) was reversed back to a more generalized state in the Arthroleptoidae, or, possibly even more interestingly, evolved convergently or in parallel at least twice (in brevicipitines and in microhylids). This hypothesis is further supported by the finding that Hemisus, another taxon characterized by a hydrostatic tongue, groups with arthroleptoids rather than with microhylids as usually thought (Biju and Bossuyt, 2003). The separate phylogenetic placement of brevicipitines from other microhylids, together with the possession of several striking morphological specialisations shared only with the Microhylidae and Hemisotidae, might justify a change in their classificatory assignment, i.e., their inclusion in an own family. Microhylids are characterized by a high variability in their osteological characters due to the repeated evolution of fossoriality and the effects of miniaturization (Wild, 1995). Osteological characters are usually more conservative in anurans and are therefore considered to be informative features for higher level taxonomy (Duellman and Trueb,

22 Chapter 1

1986). The unexpected molecular phylogenetic placement of scaphiophrynine and brevicipitine toads, if further confirmed, could provide a stark example for the high level of homoplasy in morphological characters in anurans and indicates that, in these organisms, convergent evolution and reversals may be possible even in seemingly highly derived morphological traits.

Acknowledgements We are grateful to Marius Burger, Alan Channing, Frank Glaw and Stefan Wanke for their help during sample collection and to Simone Hoegg and Dirk Steinke for valuable comments and technical assistance. We thank three anonymous reviewers for their helpful comments on the manuscript. Financial support was provided through grants of the Deutsche Forschungsgemeinschaft to M.V. and A.M.

23 Chapter 2

Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians

Chapter 2

2. Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians

Published: Vences, M., M. Thomas, A. van der Meijden, Y. Chiari, D. R. Vieites. 2005. Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians. Frontiers in Zoology 2: article 5

2.1 Abstract Identifying species of organisms by short sequences of DNA has been in the center of ongoing discussions under the terms DNA barcoding or DNA taxonomy. A C-terminal fragment of the mitochondrial gene for cytochrome oxidase subunit I (COI) has been proposed as universal marker for this purpose among animals. Herein we present experimental evidence that the mitochondrial 16S rRNA gene fulfills the requirements for a universal DNA barcoding marker in amphibians. In terms of universality of priming sites and identification of major vertebrate clades the studied 16S fragment is superior to COI. Amplification success was 100% for 16S in a subset of fresh and well-preserved samples of Madagascan frogs, while various combinations of COI primers had lower success rates. COI priming sites showed high variability among amphibians both at the level of groups and closely related species, whereas 16S priming sites were highly conserved among vertebrates. Interspecific pairwise 16S divergences in a test group of Madagascan frogs were at a level suitable for assignment of larval stages to species (1- 17%), with low degrees of pairwise haplotype divergence within populations (0-1%). We strongly advocate the use of 16S rRNA as standard DNA barcoding marker for vertebrates to complement COI, especially if samples a priori could belong to various phylogenetically distant taxa and false negatives would constitute a major problem.

25 Chapter 2

2.2 Introduction The use of short DNA sequences for the standardized identification of organisms has recently gained attention under the terms DNA barcoding or DNA taxonomy (Floyd et al., 2002; Hebert et al., 2003b; Tautz et al., 2003). Among the promising applications of this method are the assignments of unknown life-history stages to adult organisms (Hebert et al., 2004c; Thomas et al., 2005), the large-scale identification of organisms in ecological or genomic studies (Blaxter, 2004; Floyd et al., 2002) and, most controversially, explorative studies to discover potentially undescribed "candidate" species (Hebert et al., 2004c; Hebert et al., 2004d; Venter et al., 2004). Although it is not a fundamentally new technique (Moritz and Cicero, 2004), DNA barcoding is promising because technical progress has made its large-scale, automated application feasible (Blaxter, 2004; Tautz et al., 2003), which may accelerate taxonomic progress (Wilson, 2004). DNA barcoding and DNA taxonomy are realities in many fields (Blaxter, 2004). Consensus by the scientific community is essential with respect to the most suitable genes that allow robust and repeatable amplification and sequencing and provide unequivocal resolution to identify a broad spectrum of organisms. While D. Tautz and co-workers (Tautz et al., 2003) proposed the nuclear ribosomal RNA genes for this purpose, P. D. N. Hebert and colleagues have strongly argued in favor of a 5' fragment of the mitochondrial gene for cytochrome oxidase, subunit I (COI or COXI) (Hebert et al., 2003b; Hebert et al., 2003c). This gene fragment has been shown to provide a sufficient resolution and robustness in some groups of organisms, such as arthropods and, more recently, birds (Hebert et al., 2003b; Hebert et al., 2004c; Hebert et al., 2003c; Hebert et al., 2004d). A genetic marker suitable for DNA barcoding needs to meet a number of criteria (Hebert et al., 2003a). First, in the study group, it needs to be sufficiently variable to be able to discriminate among most species but sufficiently conserved to be less variable within than between species. Second, priming sites need to be sufficiently conserved to permit a reliable amplification without the risk of false negatives when pooled samples or environmental DNA is analyzed. Third, the gene should convey sufficient phylogenetic information to assign species to major taxa using simple phenetic approaches. Fourth, its

26 Chapter 2 amplification and sequencing should be as robust as possible, also under variable lab conditions and protocols. Fifth, sequence alignment should be possible also among distantly related taxa. Here we explore the performance of a fragment of the 16S ribosomal RNA gene (16S) in DNA barcoding of amphibians. As a contribution to the discussion about suitable standard markers we provide experimental data on comparative amplification success of 16S and COI in amphibians, empirical data on conservedness of priming sites, and an example from the 16S-based identification of amphibian larval stages.

2.3 Materials and methods To test for universality of primers and cycling conditions, we performed parallel experiments in three different laboratories (Berkeley, Cologne, Konstanz) using the same primers but different biochemical products and thermocyclers, and slightly different protocols. The selected primers for 16S (Palumbi et al., 1991) amplify a fragment of ca. 550 bp (in amphibians) that has been used in many phylogenetic and phylogeographic studies in this and other vertebrate classes: 16SA-L, 5' - CGC CTG TTT ATC AAA AAC AT - 3'; 16SB-H, 5' - CCG GTC TGA ACT CAG ATC ACG T - 3'. For COI we tested (1) three primers designed for birds (Hebert et al., 2004b) that amplify a 749 bp region near the 5'-terminus of this gene: BirdF1, 5' - TTC TCC AAC CAC AAA GAC ATT GGC AC - 3', BirdR1, 5' - ACG TGG GAG ATA ATT CCA AAT CCT G - 3', and BirdR2, 5' - ACT ACA TGT GAG ATG ATT CCG AAT CCA G - 3'; and (2) one pair of primers designed for arthropods (Hebert et al., 2003a) that amplify a 658 bp fragment in the same region: LCO1490, 5' - GGT CAA CAA ATC ATA AAG ATA TTG G - 3', and HCO2198, 5'-TAA ACT TCA GGG TGA CCA AAA AAT CA-3'. Sequences of additional primers for COI that had performed well in mammals and fishes were kindly made available by P. D. N. Hebert (personal communication in 2004) and these primers yielded similar results (not shown). The optimal annealing temperatures for the COI primers were determined using a gradient thermocycler and were found to be 49-50°C; the 16S annealing temperature was 55°C. Successfully amplified fragments were sequenced using various automated

27 Chapter 2

sequencers and deposited in GenBank. Accession numbers for the complete data set of adult mantellid sequences used for the assessment of intra- and interspecific divergences (e.g., figure 2.5) are AY847959-AY848683. Accession numbers of the obtained COI sequences are AY883978-AY883995. Nucleotide variability was scored using the software DnaSP (Rozas et al., 2003) at COI and 16S priming sites of the following complete mitochondrial genomes of nine amphibians and 59 other vertebrates: Cephalochordata: AF098298, Branchiostoma. Myxiniformes: AJ404477, Myxine. Petromyzontiformes: U11880, Petromyzon. Chondrichthyes: AJ310140, Chimaera; AF106038, Raja; Y16067, Scyliorhinus; Y18134, Squalus. Actinopterygii: AY442347, Amia; AB038556, Anguilla; AB034824, Coregonus; M91245, Crossostoma; AP002944, Gasterosteus; AB047553, Plecoglossus; U62532, Polypterus; U12143, Salmo. Dipnoi: L42813, Protopterus. Coelacanthiformes: U82228, Latimeria. Amphibia, Gymnophiona: AF154051, Typhlonectes. Amphibia, Urodela: AJ584639, Ambystoma; AJ492192, Andrias; AF154053, Mertensiella; AJ419960, Ranodon. Amphibia, Anura: AB127977, Buergeria; NC_005794, Bufo; AY158705; Fejervarya; AB043889, Rana; M10217, Xenopus. Testudines: AF069423, NC_000886, Chelonia; Chrysemys; AF366350, Dogania; AY687385, Pelodiscus; AF039066, Pelomedusa. Squamata: NC_005958, Abronia; AB079613, Cordylus; AB008539, Dinodon; AJ278511, Iguana; AB079597, Leptotyphlops; AB079242, Sceloporus; AB080274, Shinisaurus. Crocodilia: AJ404872, Caiman. Aves: AF363031, Anser; AY074885, Arenaria; AF090337, Aythya; AF380305, Buteo; AB026818, Ciconia; AF362763, Eudyptula; AF090338, Falco; AY235571, Gallus; AY074886, Haematopus; AF090339, Rhea; Y12025, Struthio. Mammalia: X83427, Ornithorhynchus; Y10524, Macropus; AJ304826, Vombatus; AF061340, Artibeus; U96639, Canis; AJ222767, Cavia ; AY075116, Dugong; AB099484, Echinops; Y19184, Lama; AJ224821, Loxodonta; AB042432, Mus; AJ001562, Myoxus; AJ001588, Oryctolagus; AF321050, Pteropus; AB061527, Sorex; AF348159, Tarsius; AF217811, Tupaia; AF303111, Ursus (for species names, see GenBank under the respective accession numbers). 16S sequences of a large sample of Madagascan frogs were used to build a database in BioEdit (North Carolina State University). Tadpole sequences were compared with this database using local BLAST searches (Altschul et al., 1990) as implemented in BioEdit.

28 Chapter 2

The performance of COI and 16S in assigning taxa to inclusive major clades was tested based on gene fragments homologous to those amplified by the primers used herein (see above), extracted from the complete mitochondrial sequences of 68 vertebrate taxa. Sequences were aligned in Sequence Navigator (Applied Biosystems) by a Clustal algorithm with a gap penalty of 50, a gap extend penalty of 10 and a setting of the ktup parameter at 2. PAUP* (Swofford, 2002) was used with the neighbor-joining algorithm and LogDet distances and excluding pairwise comparisons for gapped sites. We chose these simple phenetic methods instead of maximum likelihood or maximum parsimony approaches because they are computationally more demanding and because the aim of DNA barcoding is a robust and fast identification of taxa rather than an accurate determination of their phylogenetic relationships.

2.4 Results Amplification experiments We performed independent amplification experiments with one set of 16S primers and three published sets of COI primers (Hebert et al., 2003a; Hebert et al., 2004b) focusing on representatives of different frog, salamander and caecilian genera. The experiments were concordant in yielding more general results for 16S than COI. In a set of of fresh and well-preserved samples from relatively closely related mantellid frogs from Madagascar (appendix 1) the 16S amplification success was complete, whereas the three sets of COI primers yielded success rates of only 50-70%. Considering all three primer combinations, there were two species of frogs (10%) that did not amplify for COI (Boophis septentrionalis and B. tephraeomystax) at all.

Priming sites The variability of priming sites was surveyed using nine complete amphibian mitochondrial sequences from GenBank (figure 2.1), and 59 mt genomes of fishes, reptiles, birds and mammals (figure 2.2).

29 Chapter 2

Figure 2.1 Variability of priming sites for 16S rRNA and COI in amphibians.

30 Chapter 2

Figure 2.2 Variation in priming sites of 16S rRNA (a, F-primer; b, R-primer) and COI (c, Bird-F1, LCO1490; d, HCO2198; e, Bird-R1, Bird-R2) fragments studied herein. Values are nucleotide variability as calculated using the DnaSP program. Grey bars show the values for nine amphibians, black bars the values for a set of 59 other vertebrates (see Materials and methods, and figures 2.3 and 2.4).

31 Chapter 2

A high variability was encountered for COI. The sequences of some species were largely consistent with the primers: Xenopus had two mutations only at each of the priming regions. However, other sequences were strongly different, with up to seven mutations, all at third codon positions. No particular pattern was recognizable for any major group that would facilitate designing COI primers specific for frogs, salamanders or caecilians. Interestingly the variability among the amphibian sequences available was as large as or larger than among the complete set of vertebrates at many nucleotide positions of COI priming sites (figure 2.2), indicating a possible higher than random variability of this gene in amphibians. In contrast, the 16S priming sites were remarkably constant both among amphibians and among other vertebrates (figures 2.1 and 2.2). A wider survey of priming sites, i.e., the alternative reverse priming sites used in arthropod and bird studies (Hebert et al., 2003a; Hebert et al., 2004b) confirmed the high variability of COI in amphibians, and in vertebrates in general (figure 2.2). A screening of the first 800 bp of the C-terminal part of the gene in nine amphibians of which complete mitochondrial genes were available did not reveal a single fragment of 20 bp where all nine species would agree in 80% or more of their nucleotides.

Recovery of major groups The phenetic neighbor-joining analysis using the 16S fragment produced a tree that contained eight major groupings that conform to or are congruent with the current classification and phylogeny (figure 2.3): cartilaginous fishes, salamanders, frogs, turtles, eutherian mammals, mammals, squamates, birds. Of these, the COI tree (figure 2.4) recovered only the lineages of cartilaginuous fishes and birds. No additional relevant major lineage was found in the COI analysis that had not been recovered also by the 16S analysis.

32 Chapter 2

Figure 2.3 Neighbor-joining tree of selected vertebrate taxa based on the fragment of the 16S rRNA gene amplified by primers 16SaL and 16SbH. Numbers in black circles indicate major clades that were recovered by this analysis: (1) cartilaginous fishes, (2) salamanders, (3) frogs, (4) turtles, (5) eutherian mammals, (6) mammals, (7) squamates, (8) birds.

33 Chapter 2

Figure 2.4 Neighbor-joining tree of selected vertebrate taxa based on the fragment of the COI gene amplified by primers LCO1490 and HCO2198. Numbers in black circles indicate major clades that were recovered by this analysis. Only two of the clades recovered by the 16S analysis are also monophyletic here: (1) cartilaginous fishes, (8) birds.

16S rRNA barcoding of tadpoles From an ongoing project involving the large-scale identification of tadpoles of Madagascan frogs (Thomas et al., 2005) we here provide data from larval and adult frog

34 Chapter 2 species from two sites of high anuran diversity in eastern Madagascar, Andasibe and Ranomafana. These two localities are separated by a geographical distance of ca. 250 km. The results will be presented in more detail elsewhere. We selected target species for which morphological and bioacoustic uniformity suggests that populations from Ranomafana and Andasibe are conspecific. All these species belong to the family Mantellidae. We then analysed haplotypes within and between these populations. In addition we assessed divergences among sibling species of mantellid frogs. These were defined as morphologically similar species that are phylogenetically sister to each other, or are in well-defined but phylogenetically poorly resolved clades of 3-5 species. Results revealed a low intrapopulational variation of 0-3% uncorrected pairwise distances in the 16S gene, a surprisingly large differentiation among conspecific populations of 0-5.1%, and a wide range of differentiation among species, ranging from 1-16.5% with a mode at 7-9% (figure 2.5). The few species separated by low genetic distances were allopatrically distributed. The interspecific divergence was higher in those species pairs in which syntopic occurrence has been recorded or is likely (2.7-16.5% divergence, mean 8.5%) as compared to those that so far only have been found in allopatry (1.0-12.9%, mean 6.9%). Phylogenetic and phenetic analyses (Bayesian and Neighbor-joining) of these and many additional sequences (to be published elsewhere) mostly grouped sequences of those specimens from Ranomafana and Andasibe that a priori had been considered to be conspecific (exceptions were Mantidactylus boulengeri, not considered in the intraspecific comparisons here, and M. blommersae). This indicates that cases, in which haplotypes of a species are more similar to those of another species than to those of other conspecific individuals or populations, are rare in these frogs. Sharing of identical haplotypes among individuals belonging to different species, in our dataset, was limited to three closely related species pairs of low genetic divergences: Boophis doulioti and B. tephraeomystax, B. goudoti and B. cf. periegetes, Mantella aurantiaca and M. crocea. Depending on the taxonomic scheme employed, our complete data set contains 200-300 species of Madagascan frogs. Hence, haplotype sharing was demonstrated in 2-3% of the total number of species only.

35 Chapter 2

Figure 2.5 Variation in the fragment of the 16S rRNA gene (ca. 550 bp) studied herein, (a) within populations, (b) among conspecific populations and (c) among sibling species of frogs in the family Mantellidae from Madagascar. The values are uncorrected p-distances from pairwise comparisons in the respective category. Only one (mean) value per species was used in (a) and (b), even when multiple individuals were compared.

To explore the reliability of tadpole identification using the 16S gene we used local BLAST searches against a database containing about 1000 sequences of adult frogs from a wide sampling all over Madagascar. 138 tadpoles from the Andasibe region and 84 tadpoles from the Ranomafana region were compared with adult sequences in the database. In 77% of the cases the highest scores were those from comparisons to adults from the same site as the tadpoles. In most of the unsuccessful comparisons, adult sequences of the corresponding species were not available from the tadpole site (21%). In

36 Chapter 2 only 5 cases (2%) conspecific adults collected from a different site than the tadpoles yielded higher BLAST scores although adult sequences from the same site were in the database.

2.5 Discussion DNA barcoding in amphibians DNA barcoding has great appeal as a universally applicable tool for identification of species and variants of organisms, possibly even in automated handheld devices (Janzen, 2004). However, doubtless severe restrictions exist to its universal applicability (Moritz and Cicero, 2004). Some taxa, e.g., cichlid fishes of Lake Victoria, have radiated so rapidly that the speciation events have not left any traces in their mitochondrial genomes (Verheyen et al., 2003); identifying these species genetically will only be possible through the examination of multiple nuclear markers, as it has been done to assess their phylogeny (Albertson et al., 1999). Some snails are characterized by a high intraspecific haplotype diversity, which could disable attempts to identify and distinguish among species using such markers (Thomaz et al., 1996). Haplotype sharing due to incomplete lineage sorting or introgression is also known in amphibians (Funk and Omland, 2003) although it was not common in mantellid frogs in our data set. However, a number of species showed haplotype sharing with other species, or non-monophyletic haplotypes, warranting a more extensive discussion. In Mantidactylus boulengeri, specimens from Andasibe and Ranomafana have similar advertisement calls and (at least superficially) similar morphologies, but their 16S haplotypes were not a monophyletic group (unpublished data). This species belongs to a group of direct-developing frogs that, like the Neotropical Eleutherodactylus (Dubois, 2004) may be characterized by a high rate of cryptic speciation. Further data are necessary to decide whether the populations from Ranomafana and Andasibe are indeed conspecific. In contrast, there is little doubt that the populations of Mantidactylus blommersae from these two sites are conspecific, yet the Ranomafana haplotypes are closer to those of the clearly distinct species M. domerguei. The species pairs where haplotype sharing has been observed (see Results) all appear to be allopatrically to parapatrically distributed and show no or only low differences in advertisement calls,

37 Chapter 2 indicating that occasional hybridization along contact zones may be possible (e.g., Chiari et al., 2004). Haplotypes of each of these species pairs always formed highly supported clusters or clades, and had divergences below 3%, indicating that haplotype sharing in mantellids may only constitute a problem when individuals are to be assigned to such closely related sister species. Although our data show that DNA barcoding in mantellids is a largely valid approach when both reference and test sequences come from the same site, the occurrence of non- monophyletic and highly divergent haplotypes within species characterizes these and other amphibians as a challenging group for this technique. Certainly, DNA barcoding is unable to provide a fully reliable species identification in amphibians, especially if reference sequences do not cover the entire genetic variability and geographic distribution of a species. However, the same is true for any other morphological or bioacoustic identification method. Case studies are needed to estimate more precisely the margin of error of molecular identification of amphibian species. For many approaches, such as the molecular survey of the trade in frog legs for human consumption (Veith et al., 2000), the error margins might be acceptable. In contrast, the broad overlap of intraspecific and interspecific divergences (figure 2.5) cautions against simplistic diagnoses of presumably new amphibian species by DNA divergences alone. A large proportion of biological and evolutionary species will be missed by inventories that characterize candidate species by DNA divergences above a previously defined threshold.

Comparative performance of DNA barcoding markers in amphibians Phenomena of haplotype sharing or non-monophyletic conspecific haplotypes will affect any DNA barcoding approach that uses mitochondrial genes, and are also to be expected with nuclear genes (e.g., Machado and Hey, 2003). Nevertheless, some genes certainly outperform others in terms of discriminatory power and universal applicability, and these characteristics may also vary among organism groups. The mitochondria of plants are characterized by very different evolutionary patterns than those of animals, including frequent translocation of genetic material into and from the nucleus (Palmer et al., 2004), which limits their use for DNA barcoding purposes. Nuclear ribosomal DNA (18S and

38 Chapter 2

28S), proposed as standard marker (Tautz et al., 2003), has a high potential in invertebrate DNA barcoding but its high-throughput amplification encounters difficulties in vertebrates. As a consequence, despite the need of consensus on markers for universal applicability of DNA barcoding, the use of different genes in different groups of organisms seems reasonable. It has been hypothesized that universal COI primers may enable amplification of a 5' terminal fragment from representatives of most animal phyla due to their robustness (Hebert et al., 2003a). The success in DNA barcoding of lepidopterans and birds suggests that this gene fragment can indeed be used as a standard for many higher animal taxa (Hebert et al., 2003a; Hebert et al., 2004a; Hebert et al., 2004b). In our experiments we compared 16S primers commonly used in amphibians to COI primers that had been developed for other vertebrates (Hebert et al., 2004b) or invertebrates (Hebert et al., 2003a). It may well be possible, with some effort, to design primers that are more successful and consistent in amplifying COI from amphibians. However, our results from mantellid frogs (appendix 1) indicate a very good amplification success of the primers for some species, but failure for other, related species. This and our results on variability of priming sites predict enormous difficulties in designing one pair of primers that will reliably amplify this gene fragment in all vertebrates, all amphibians, or even all representatives of any amphibian order. A set of one forward and three reverse COI primers have been successfully used to amplify and sequence a large number of bird species (Hebert et al., 2004b), but birds are a much younger clade than amphibians with a probably lower mitochondrial variability. A further optimization of COI amplification may also be achieved regarding the PCR protocol. Herein we used standard protocols that optimized annealing temperature only, whereas more complex touchdown protocols have been used for birds and butterflies (Hebert et al., 2004a; Hebert et al., 2004b). However, one major requirement for a DNA barcoding marker is its robustness to variable lab conditions. If DNA barcoding is to be applied as a standard in many different labs, primers and genes need to be chosen that amplify reliably under very different conditions and under standard protocols. This clearly applies to 16S, which we have amplified with very different annealing temperatures and PCR conditions in previous exploratory studies (results not shown).

39 Chapter 2

Alignment of 16S sequences is complicated by the prevalence of insertions and deletions, and this gene is less variable than COI (Hebert et al., 2003a). Nevertheless, our results indicate that even using an uncritical automated alignment this gene has a higher potential than COI to assign vertebrate sequences to the level of classes and orders. The 16S gene is a highly conserved mitochondrial marker but mutations are common in some variable regions, corresponding to loops in the ribosomal RNA structure. In amphibians, where many species are relatively old entities (Wilson et al., 1974), this ensures a sufficient amount of mutations among species. Our results for amphibians, and previous experience with fishes, reptiles and mammals, indicate that 16S is sufficiently variable to unambiguously identify most species. A further mitochondrial gene that has been widely used in amphibian phylogenetic and phylogeographic studies is cytochrome b. This gene can easily be amplified in salamanders and archaeobatrachian frogs using primers that anneal with adjacent tRNA genes. However, neobatrachian frogs (the wide majority of amphibian species) are characterized by rearrangements of the mitochondrial genome (Macey et al., 1997; Sumida et al., 2001), and cytochrome b in these species borders directly to the control region. Although cytochrome b primers are available that work in many neobatrachians (Bossuyt and Milinkovitch, 2000; Vences et al., 2003b), they sometimes fail in closely related forms, similar to the COI primers used herein, presumably because of mutations at the priming sites (pers. obs. M. Vences in mantellid frogs). In contrast, the 16S primer pair used here can be considered as truly universal not only for amphibians but even for vertebrates. This is also reflected by the high number of amphibian 16S sequences in GenBank (2620 hits for 16S vs. 483 hits for COI, as of September 2004). Moreover, the 16S and 12S rRNA genes have been selected as standard markers for phylogeny reconstruction in amphibians (AmphibiaTree consortium), which will lead to a near-complete global dataset of amphibian 16S sequences in the near future. If the development of handheld devices (Janzen, 2004) is envisaged as a realistic goal, then the universality and robustness of primers should be among the most relevant characteristics of a gene for DNA barcoding. When pooled samples containing representatives of various higher vertebrate taxa are to be analysed, the risk of false negatives strongly increases with decreasing universality of primers. As a consequence

40 Chapter 2 we recommend the use of 16S as additional standard DNA barcoding marker for vertebrates, especially for but not limited to applications that involve pooled samples.

Acknowledgements For comments, technical help and/or discussions we are grateful to Paul D. N. Hebert, Axel Meyer, Dirk Steinke, Diethard Tautz and David B. Wake. We are further indebted to Simone Hoegg, Pablo Orozco and Mario Vargas who provided help in the lab, and to the Madagascan authorities for research permits. The DNA barcoding project of Madagascan tadpoles was supported by a grant of the Volkswagen foundation to M Vences and to F Glaw. DR Vieites was supported by the AmphibiaTree project (NSF grant EF –O334939).

41 Chapter 3

A previously unrecognised radiation of ranid frogs in Southern Africa revealed by nuclear and mitochondrial DNA sequences

Pyxicephalus adspersus

Chapter 3

3. A previously unrecognised radiation of ranid frogs in Southern

Africa revealed by nuclear and mitochondrial DNA sequences

Published:

Van der Meijden, A., M. Vences, S. Hoegg and A. Meyer. 2005. A previously unrecognized radiation of ranid frogs in Southern Africa revealed by nuclear and mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 37: 674–685.

3.1 Abstract In sub-Saharan Africa amphibians are represented by a large number of endemic frog genera and species of incompletely clarified phylogenetic relationships. This applies especially to African frogs of the family Ranidae. We provide a molecular phylogenetic hypothesis for ranids, including 11 of the 12 African endemic genera. Analysis of nuclear (rag-1, rag-2 and rhodopsin genes) and mitochondrial markers (12S and 16S ribosomal RNA genes) provide evidence for an endemic clade of African genera of high morphological and ecological diversity thus far assigned to up to five different subfamilies: Afrana, Cacosternum, Natalobatrachus, Pyxicephalus, Strongylopus, and Tomopterna. This clade has its highest species diversity in southern Africa, suggesting a possible biogeographic connection with the Cape Floral Region. Bayesian estimates of divergence times place the initial diversification of the southern African ranid clade at ~ 62-85 million years ago, concurrent with the onset of the radiation of afrotherian mammals. These and other African ranids (Conraua, Petropedetes, Phrynobatrachus, Ptychadena) are placed basally within the Ranoidae with respect to the Eurasian groups, which suggests an African origin for this whole epifamily.

43 Chapter 3

Abbreviations: MYA, Million Years Ago; MP, Maximum Parsimony; ML, Maximum Likelihood; BI, Bayesian Inference; NJ, Neighbour Joining

3.2 Introduction The recent report of the Global Amphibian Assessment project (Stuart et al., 2004) shows that at least a disturbing 42 percent of amphibian species are experiencing declines, in large part due to still unknown processes. In some cases entire diverse clades of frogs are heavily declining (Lötters et al., 2004). Such non-random extinctions can lead to a severe loss of evolutionary history (Purvis et al., 2000) and a reliable phylogeny of all amphibians is needed to identify them. In several very species-rich cosmopolitan groups of frogs the phylogenetic relationships are still insufficiently known. This lack of a robust phylogenetic hypothesis is especially true for the family Ranidae or True Frogs that contains over 700 species, which are distributed throughout the world. A single genus (Rana) is thought to occur on all continents except Antarctica. Yet the phylogenetic relationships among Rana, and ranids in general, are largely uncharted (Emerson et al., 2000). Recent molecular studies have provided important progress in the understanding of ranids and its related groups (Bossuyt and Milinkovitch, 2000; Hoegg et al., 2004a; Van der Meijden et al., 2004; Vences et al., 2003b). Some studies have identified India as a reservoir of ancient ranid lineages, and proposed these animals as a model for "Out of India" dispersal of vertebrates (Bossuyt and Milinkovitch, 2001; Roelants et al., 2004). These works demonstrated the potential of ranids to decipher general patterns of biogeography and diversification although only a part of the currently recognized ranid diversity has been studied so far. Because most of the endemic African ranid genera are still unstudied from a molecular perspective the biogeographical insights remain incomplete. Despite recent compelling evidence for the ability for transoceanic dispersal in amphibians (Hedges et al., 1992; Vences et al., 2004; Vences et al., 2003b), there is little doubt that continental drift has had a major influence in shaping their current distribution and phylogeny. The close relationships of the recently discovered Nasikabatrachus from India with Nesomantis from the Seychelles strikingly demonstrated the importance of the Gondwanan breakup for the vicariance biogeography and hence phylogeny of these basal

44 Chapter 3

Neobatrachian frogs (Ranoidei sensu Sokol, 1977). Africa is generally seen as the place of origin for the current distribution of frogs in the superfamily Ranoidea (Biju and Bossuyt, 2003; Feller and Hedges, 1998; Savage, 1973), and one of its subclades, the Arthroleptoidae (figure 3.1), is endemic to this continent (with a few species in Madagascar and on the Seychelles). Africa is renowned for several endemic radiations such as the afrotherian mammals (Springer et al., 1997) and the haplochromine cichlid fishes (Verheyen et al., 2003). Africa was united with South America, Australia, Antarctica, India and Madagascar in the supercontinent Gondwanaland until the end of the late Jurassic. After the breakup of Gondwanaland, Africa remained isolated until it connected with Eurasia. The India-Seychelles-Madagascar plate broke off from Africa 158-160 million years ago (mya), and Greater India started to drift northwards across the Indian Ocean about 96-84 mya (Briggs, 2003). The India-Madagascar plate has been suggested as possible biogeographic origin of Asian ranoid subclades (Bossuyt and Milinkovitch, 2001; Duellman and Trueb, 1986), i.e., the Rhacophoridae and at least part of the Ranidae. Land bridges that connected Africa with Eurasia after its long isolation from other continents allowed Eurasian faunal elements to disperse into Africa, including several ranoid representatives. Species of the dicroglossine genus Hoplobatrachus, of the ranine lineage containing the genera/subgenera Rana, Amnirana, and the rhacophorid genus Chiromantis have dispersed into Africa from Eurasia (Kosuch et al., 2001; Vences et al., 2003b). Currently 21 ranid genera are restricted in their distribution to Africa, most of which are limited to sub-Saharan Africa. By analyzing nuclear and mitochondrial DNA sequences of representatives of all but one subfamily of ranids we here provide the first inclusive molecular phylogeny of ranid relationships. Our data provide compelling evidence for a deep evolutionary history of many African endemic ranid groups and, unexpectedly, uncover an endemic radiation that includes taxa that had so far been classified into up to five different subfamilies.

45 Chapter 3

Figure 3.1 Schematic representation of the classification of ranids and their phylogenetic position among frogs following Vences and Glaw (2001), with some additions from Dubois (1992) and Blommers-Schlösser (1993), and with modifications from the trees of Biju and Bossuyt (2003), Hoegg et al. (2004), Roelants et al. (2004), Van der Meijden et al. (2004) and own unpublished data: the family Ranidae is a paraphyletic assemblage that together with the Mantellidae and Rhacophoridae forms the epifamily Ranoidae. Together with two other epifamilies (the Arthroleptoidae and Microhyloidae) they form the superfamily Ranoidea in the Neobatrachia. The familial scheme used here includes Bombinatoridae in Discoglossidae, and Limnonastidae and Rheobatrachidae in Myobatrachidae.

46 Chapter 3

3.3 Materials and methods Taxonomy and selection of taxa Duellman and Trueb's (1986) characterization of ranid systematics being ‘in a state of chaos’ has been heavily quoted but the situation has not much improved since. As a convention we here follow the taxonomic scheme of Vences and Glaw (2001), with some modifications from more recent research as outlined in figure 3.1. At present, accepting the proposals of subfamilial arrangement by Dubois and Ohler (Dubois and Ohler, 2001) and Roelants et al. (2004) to the classifications of Dubois (Dubois, 1992) and Blommers- Schlösser (1993), the Ranidae consists of about twelve subfamilies (figure 3.1), of which five are endemic to Africa. Ranids are a paraphyletic group that, together with the Mantellidae and Rhacophoridae, forms the epifamily Ranoidae. These are hierarchically a fraction of the superfamily Ranoidea and the suborder Neobatrachia which both probably are monophyletic (Hoegg et al., 2004a). Sequences were obtained from taxa representing all ranid subfamilies except the Micrixalinae (table 3.1), as well as from the families Mantellidae and Rhacophoridae. We furthermore included taxa belonging to the Arthroleptoidae and Microhyloidae. Latimeria, Homo, Gallus, the salamander Lyciasalamandra, two archaeobatrachians of the genus Alytes, and two hylid neobatrachians, genera Agalychnis and Litoria, as hierarchical outgroups (not shown in figures).

DNA Sequencing DNA was extracted from muscle or skin tissue fixed in 99% ethanol. Tissue samples were digested using Proteinase K (final concentration 1 mg/mL), homogenised and subsequently purified following a standard salt extraction protocol. Primers for rag-1 and rag-2 were from Hoegg et al. (2004) as reported in Chiari et al. (2004). Primers for one fragment of the 12S rRNA gene and one fragment of the 16S rRNA gene were 12SA-L and 12SB-H and 16SA-L and 16SB-H of Palumbi et al. (1991) respectively (see Vences et al., 2003a). Primers for a fragment of rhodopsin exon (Rhod1A and Rhod1D) were from Bossuyt and Milinkovitch (2000).

47 Chapter 3

Table 3.1 Taxa included in this study, voucher specimens, and GenBank accession numbers of rag-1 sequences. Collection acronyms are: MNHN, Muséum National d'Histoire Naturelle, Paris, France; UADBA, Université d'Antananarivo, Département de Biologie Animale, Madagascar; ZFMK, Zoologisches Forschungsinstitut und Museum A. Koenig, Bonn, Germany; ZSM, Zoologische Staatssammlung München, Germany. SIH and MV refer to frozen tissue collections of S. Hoegg and M. Vences. Name General distribution Voucher specimen GenBank Family; microhylid subfamily accession Xenopus laevis Archaeobatrachia: Pipidae Africa Voucher not collected L19324 Arthroleptis variabilis Arthroleptidae Sub-Saharan Africa ZFMK 68794 AY571642 Bufo bufo Hyloidea: Bufonidae Europe Voucher not collected AY323762 Bufo regularis Hyloidea: Bufonidae Africa SIH-04 AY323763 Hyla cinerea Hyloidea: Hylidae North America SIH-06 AY323766 Hyla meridionalis Hyloidea: Hylidae Europe Voucher not collected AY571662 Hetrixalus tricolor Hyperoliidae Madagascar ZSM 700/2001 AY323768 Hyperolius viridiflavus Hyperoliidae Sub-Saharan Africa ZFMK 66726 AY323769 Kassina maculata Hyperoliidae Sub-Saharan Africa ZFMK 66445 AY571651 Leptopelis natalensis Hyperoliidae South Africa ZFMK 68785 AY571654 Aglyptodactylus madagascariensis Mantellidae Madagascar ZSM 183/2002 AY571640 Boophis doulioti Mantellidae Madagascar ZSM 185/2002 AY571643 Laliostoma labrosum Mantellidae Madagascar UADBA-MV2001.1466 AY571652 Dermatotonus muelleri Microhylidae: Microhylinae South America ZFMK uncatalogued AY571647 Kaloula pulchra Microhylidae; Microhylinae Southeast Asia SIH-09 AY323772 Scaphiophryne calcarata Microhylidae: Scaphiophryninae Madagascar ZSM 115/2002 AY571660 Phrynomantis annectens Microhylidae; Phrynomerinae Sub-Saharan Africa ZFMK 66771 AY571657 Breviceps fuscus Microhylidae; Brevicipitinae South Africa ZFMK 66716 AY571644 Plethodontohyla aluaudi Microhylidae; Cophylinae Madagascar ZSM 3/2002 AY571661 Dyscophus antongilii Microhylidae; Dyscophinae Madagascar Voucher not collected, AY571648 Cacosternum boettgeri Ranidae Sub-Saharan Africa ZFMK 66727 AY571645 ZFMK uncatalogued (MV- Fejervarya sp. Ranidae Southeast Asia PBl1) AY571649 Hoplobatrachus occipitalis Ranidae Sub-Saharan Africa ZFMK 65186 AY571650 Lankanectes corrugatus Ranidae Sri Lanka MNHN 2000.616 AY571653 Nyctibatrachus major Ranidae India ZFMK uncatalogued AY571655 Petropedetes cf. parkeri Ranidae Sub-Saharan Africa ZFMK uncatalogued AY571656 Ptychadena mascareniensis Ranidae Sub-Saharan Africa ZSM 190/2002 AY571658 Rana (Amnirana) lepus Ranidae Sub-Saharan Africa MV-Cam1 AY571641 Rana (Rana) temporaria Ranidae Europe Voucher not collected AY323776 Chirixalus sp. Rhacophoridae Southeast Asia to India ZFMK 65463 AY571646 Bangladesh, Nepal, Sri Polypedates maculatus Rhacophoridae Lanka, India Voucher not collected AY323777 Rhacophorus [Polypedates] dennysii Rhacophoridae Southeast Asia ZFMK 65461 AY571659

PCR was performed in 25 µl reactions containing 0.5-1.0 units of REDTaq DNA Polymerase (Sigma, Taufkirchen, Germany), 0.01 units of Pwo DNA polymerase (Roche, Mannheim, Germany), 50 ng genomic DNA, 10 pmol of each primer, 15 nmol of each dNTP, 50 nmol additional MgCl2 and the REDTaq PCR reaction buffer (in final reaction

solution: 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.1 mM MgCl2 and 0.01% gelatine). For

48 Chapter 3 rag-1 and rag-2 cycle conditions were adapted from a long range PCR protocol (Barnes, 1994), with an initial denaturation step at 94°C for 5 minutes, followed by ten cycles with 94°C for 30 seconds, annealing temperatures increasing by 0.5°C per cycle from 52 to 57°C and extending for 3 minutes at 68°C. Additionally, 20 cycles were performed with 94°C for 10 seconds, 57°C for 40 seconds and 68°C for 3 minutes. The final extension was done at 68°C for 5 minutes. For 12S and 16S the denaturation step was followed by 35 cycles of denaturation at 94° for 30 seconds, annealing at 50° for 30 seconds and extension at 72° for 90 seconds. PCR products were purified via spin columns (QIAGEN). Sequencing was performed directly using the corresponding PCR primers (forward and reverse). DNA sequences of both strands were obtained using the BigDye Terminator cycle-sequencing ready reaction kit (Applied Biosystems Inc.) on an ABI 3100 capillary sequencer using the manufacturer’s instructions. New sequences for 37 species were combined with existing sequences taken from GenBank in the final dataset. These sequences were deposited in GenBank (for accession numbers see table 3.1).

Data Analysis DNA sequences were aligned using ClustalW (Thompson et al., 1994). Gapped and hypervariable sites, totalling 729 characters, were excluded from the analyses. A homogeneity partition test (Farris et al., 1994) as implemented in PAUP* (Swofford, 2002) rejected homogeneity of the different markers. Besides a combined analysis of the combined data set we therefore also performed separate analyses of the various genes. The combined dataset was used to calculate neighbor-joining (NJ), maximum parsimony (MP) and maximum likelihood (ML) phylogenies using PAUP* (Swofford, 2002). Heuristic searches were performed using 10 replicates of a stepwise addition of taxa. The best fitting models of sequence evolution for ML analyses (table 3.2) were determined by hierarchical likelihood ratio tests and by the AIC criterion in Modeltest 3.06 (Posada and Crandall, 1998). Bootstrap branch support values were calculated with 500 MP replicates and 100 ML replicates.

49 Chapter 3

Bayesian Inference (BI) of the combined and of separate datasets was conducted with MrBayes 2.0 (Huelsenbeck and Ronquist, 2001), using models estimated with Modeltest under the AIC criterion (table 3.2), with 250,000 generations, sampling trees every 10th generation (and calculating a consensus tree after omitting the first 3000 trees). For the combined dataset, 1,000,000 generations were computed, with a burn-in of 10,000). These BI phylogeny reconstructions were repeated five times each, resulting in only very minor differences in the resulting trees, all referring to unsupported branches without relevance for the present study.

Table 3.2 Models and parameter values used in the maximum likelihood (ML) analysis estimated with Modeltest (Posada and Crandall 1998), and in the Bayesian (BI) analyses estimated with MrModeltest (Nylander, 2002). The upper two models were estimated using hierarchical likelihood ratio tests (HLRT), the lower ones are according to the AIC criterion.

Gene and Model Base frequencies Rate parameters Shape Proportion analysis parameter invariable A C G T A-C A-G A-T C-G C-T G-T sites

Combined TrN+I+G 0.3160 0.2403 0.1927 0.2510 1.0000 3.8766 1.0000 1.0000 5.1032 1.0000 0.9517 0.3287 (ML - HLRT

Combined GTR+I+G 0.3278 0.2251 0.1952 0.2520 1.7919 4.4416 1.4886 1.2223 7.3845 1.0000 0.7962 0.2860 (BI - HLRT)

Combined GTR+I+G 0.3021 0.2286 0.2079 0.2614 1.7245 4.9953 1.3206 1.2157 6.9370 1.0000 0.9443 0.3254 (ML and BI - AIC)

Rag-1 (BI - GTR+I+G 0.3023 0.2279 0.2140 0.2557 1.4603 4.3679 1.0199 1.0124 5.7222 1.0000 1.0672 0.3688 AIC)

Rag-2 (BI - GTR+I+G 0.2915 0.2299 0.2025 0.2760 1.3504 4.7911 0.9256 1.3892 4.8758 1.0000 1.2424 0.2409 AIC)

Rhodopsin HKY+I+G 0.2261 0.3095 0.1804 0.2841 ------1.2011 0.3778 (BI - AIC)

12S + 16S GTR+I+G 0.3527 0.2287 0.2056 0.2130 3.5116 8.4158 5.1886 0.4111 28.236 1.0000 0.5347 0.2952 (BI - AIC) 0

Divergence time estimation We used the MultiDivTime package (Thorne and Kishino, 2002; Thorne et al., 1998) to estimate the divergence times, based on nulcear sequences only. Calibration points were applied as follows: (1) minimum age of the frogs-salamander split at 230 mya (fossil record of frog ancestor Triadobatrachus (Sanchiz, 1998)); (2) minimum age of the split

50 Chapter 3 between Agalychnis and Litoria at 42 mya (last connection between Australia and South America (Seddon et al., 1998)); (3) maximum age of the split between Mantidactylus wittei and Mantidactylus sp. from the Comoro islands at 15 mya (volcanic origin of the oldest Comoro island Mayotte (Vences et al., 2003b)); (4) minimum age of the Alytes muletensis-Alytes dickhillenii split at 5 mya (Mediterranean salinity crisis: Fromhage et al., 2004); and (5) age interval of the split between diapsids and synapsids at 338-288 mya (Graur and Martin, 2004).

3.4 Results After exclusion of highly variable regions of 12S and 16S rRNA, the concatenated dataset consisted of 2995 nucleotides from nuclear genes (rag-1, rag-2 and rhodopsin) and mitochondrial genes (12S and 16S). Of these, 318 nucleotides were uninformative and 1212 base pairs were parsimony informative. For rag-1 606 sites were parsimony informative and 589 were constant of a total of 1330 sequenced nucleotides. For rag-2 755 base pairs were sequenced and contained 472 informative and 219 constant sites. Rhodopsin had 127 informative and 134 constant of a total of 289 characters. The fragments of the mitochondrial genes 12S and 16S had 119 and 141 informative, and 82 and 188 constant of 253 and 368 characters respectively. None of the nuclear gene fragments showed saturation when transitions and transversions were plotted against sequence divergence. All the phylogenies based on the combined dataset resolved the hierarchical outgroups in the expected relationships. The phylogenies based on the combined dataset obtained through MP, NJ and ML consistently show the superfamily Ranoidea as a clade with high bootstrap support (figure 3.2), containing the included representatives of Ranoidae, Microhyloidae and Arthroleptoidae (support values from MP bootstrap and ML bootstrap; 100 percent and BI analysis 98 percent). The clustering of Breviceps with the hyperoliids (82, 62, 100) provides further support for the exclusion of this genus from the Microhylidae, as already indicated in an earlier study using only rag-1 (Van der Meijden et al., 2004). A more inclusive arthroleptoid sampling including the Hemisotidae is necessary to determine the position of Breviceps

51 Chapter 3

and related genera, since the Hemisotidae were found to be closely related to the hyperoliids (Biju and Bossuyt, 2003). The clade Ranoidae receives high support (100, 97, 100), as do several subclades within this epifamily that agree with current classification. This pertains to the Rhacophoridae, Mantellidae, Dicroglossinae and Raninae. Basal resolution within the Ranoidae epifamily is low, however. The nested position of the Rhacophoridae and Mantellidae renders the Ranidae paraphyletic. The position of the platymantine Ceratobatrachus is inconsistent between the different methods of analysis and remains only weakly resolved, indicating that the subfamily Platymantinae is distinct from other ranoids. The relationships of the African genera Phrynobatrachus and Ptychadena to the remaining Ranoidae could not be resolved unambiguously. Excluding these two genera and a further clade of African species (see below), the remaining Eurasian and American Ranoidae form a monophyletic clade with some, albeit low, support (<50, 60, 100). The African species Hoplobatrachus occipitalis and Amnirana lepus are nested within the largely Asian Dicroglossinae and Raninae, supporting the hypothesis of their Asian origin (Kosuch et al., 2001). Most remarkable is the presence of a highly supported clade (100, 94, 100) containing representatives of six sub-Saharan genera, most of which have so far not been considered to be related (figure 3.2): Afrana, Cacosternum, Natalobatrachus, Pyxicephalus, Strongylopus and Tomopterna. Biogeographically, the center of diversity and endemism of this divergent set of taxa is in southernmost Africa (figure 3.4). This endemic southern African clade is highly distinct and supported irrespective of the type of phylogenetic analysis. Further African genera such as Petropedetes and Conraua may be among its basal representatives (figure 3.2) but support for this placement is weak and they are thus not considered further here. The clade is resolved, at least partially, also in separate Bayesian analyses of the gene fragments used (figure 3.3). Rag-1 and rag-2 were congruent in strongly supporting a monophyletic group of all six taxa, whereas rhodopsin supported a group of only five of them (excluding Afrana), and 12S+16S included, in addition, Petropedetes and Phrynobatrachus in this clade (although with conspicuously long branches, indicative of a possibly spurious placement).

52 Chapter 3

Figure 3.2 Maximum likelihood phylogram of the superfamily Ranoidea, rooted with hierarchical outgroups Latimeria, Homo, Gallus, Lyciasalamandra, Alytes, Agalychnis and Litoria (not shown). Support values above branches are ML/MP bootstrap values, a single asterisk below a branch indicates a Bayesian posterior probability above 95%, and two asterisks indicate a Bayesian posterior probability of 100%. When only one value is shown it refers to the ML support (the MP support in these cases was below 50%). This tree was obtained using a substitution model suggested by hierarchical likelihood ratio tests; a tree calculated under the substitution model sugegsted by the AIC criterion in Modeltest (Posada and Crandall, 1998) was identical except in placing Ceratobatrachus sister to rhacophorids. African Ranoidae are marked in bold. Inset shows the diversity of the representatives of the endemic African clade, in order from top to bottom Pyxicephalus adspersus, Tomopterna sp. (Khorixas), Natalobatrachus bonebergi, Afrana angolensis, Cacosternum boettgeri, Strongylopus fasciatus.

Bayesian analysis of divergence times (table 3.3) provided a posterior age estimate of the epifamily Ranoidae of 91.9 mya (95% confidence intervals 65.9-124.4 mya), of the endemic ranid clade with Pyxicephalus as the most basal taxon of 69.9 (48.9-96.5) mya

53 Chapter 3 and of the more inclusive clade with also Petropedetes and Conraua of 85.6 (61.1-116.4) mya.

Table 3.3 Posterior time estimates of most relevant splits within the Ranoidea from a Bayesian analysis using the MultiDivTime program (Thorne and Kishino, 2002) with calibrations and settings given in Materials and methods. Age Standard 95% Confidence Clade (MY) Deviation Interval Ranoidea 133.6 19.8 99.2 - 176.7 Arthroleptoidae-Microhylidae 127.1 19.4 93.2 - 169.4 Ranoidae 91.9 14.9 65.9 - 124.4 African Endemic Clade + Conraua + Petropedetes 85.6 14.1 61.1 - 116.4 Non-African Ranoidae including Amnirana, Hoplobatrachus and Chirixalus 83.3 13.7 59.4 - 113.2 African Endemic Clade + Petropedetes 81.7 13.7 58.0 - 111.6 Mantellidae + Rhacophoridae 73.1 12.4 51.6 - 100.1 African Endemic Clade 69.9 12.3 48.9 - 96.5 African Endemic Clade excluding Pyxicephalus 61.7 11.3 42.7 - 86.0 Mantellidae 58.2 10.2 40.7 - 80.5 Natalobatrachus, Afrana, Cacosternum, Strongylopus 50.4 9.7 34.1 - 71.8 Afrana, Cacosternum, Strongylopus 47.7 9.3 31.9 - 68.0 Cacosternum, Strongylopus 40.2 8.4 25.9 - 58.9 Rhacophoridae 36.0 7.5 23.3 - 52.8 Raninae (excluding Afrana) 33.9 7.5 21.4 - 50.6

3.5 Discussion Endemic ranids from Southern Africa form an unexpected novel and divergent clade With over 200 species of ranids found only in Africa, it is, after Asia, the continent with the second highest species diversity of this family. Most of these species have never been included in global phylogenetic studies. Several thorough osteological studies on African taxa (e.g., Clarke, 1981; Deckert, 1938) included no or very few Asian taxa whereas most of the recent molecular studies on ranids (Bossuyt and Milinkovitch, 2000; Emerson et al., 2000; Marmayou et al., 2000; Roelants et al., 2004) focused on Asian taxa.

54 Chapter 3

Figure 3.3 Separate Bayesian trees of the gene fragments. Mitochondrial markers (12S and 16S rRNA fragments) were combined in a single dataset. Numbers are posterior probabilities.

Despite this relative lack of knowledge, the finding of a highly supported Southern African ranid clade in our analysis was still most surprising. The phylogenetic relationships implied by figure 3.2 have not been previously hypothesized based on morphological data (e.g., Clarke, 1981). This clade, with Pyxicephalus basal to the other five genera, is strongly supported by the combined analysis and by separate analyses of rag-1 and rag-2. Phylogenies based on rhodopsin, and on the mitochondrial 12S and 16S rRNA, provide some additional support for close relationships of taxa in this clade in the combined analyses but are less unequivocal when analysed individually (figure 3.3). However, the low phylogenetic resolution of the latter three genes is not surprising because (a) for rhodopsin, we only included a short fragment (289 bp), and (b) and for mitochondrial genes such as 12S and 16S it is known that they are less informative in resolving deep phylogenetic relationships compared to single-copy protein coding nuclear genes (e.g., Springer et al., 1997). One of the six genera unambiguously included in the endemic southern African clade, Natalobatrachus, is grouped in the Petropedetinae (Blommers-Schlösser, 1993; considered as the family Phrynobatrachidae by Dubois, 1992) mainly based on

55 Chapter 3 osteological and dental characters (Laurent, 1986). Cacosternum has been placed in the Cacosterninae (Blommers-Schlösser, 1993). Afrana and Strongylopus have been classified in the Raninae, Tomopterna and Pyxicephalus in the Tomopterninae and Pyxicephalinae respectively (Blommers-Schlösser, 1993; Dubois, 1992). While the Cacosterninae, Pyxicephalinae and Tomopterninae may belong to the endemic African clade in their totality (except for the Asian Nannophrys that was placed in the Cacosterninae by Blommers-Schlösser, 1993), this is not the case for the Petropedetinae and Raninae. Indeed, the Raninae included here form a well-defined clade when Strongylopus and Afrana, considered subgenera of Rana by Dubois (1992), are excluded (figure 3.2). The elusion of this molecularly well distinguishable endemic southern African clade of ranids to morphological analyses suggests a high incidence of homoplasy in morphological characters used for their classification. Other regional radiations, such as the Madagascan and Indian tree frogs (Bossuyt and Milinkovitch, 2000), or Indian and African burrowing frogs (Biju and Bossuyt, 2003) show a similar pattern of morphological homoplasy. In other ranoid frogs such as microhylids (Wild, 1995) and brevicipitines (Blommers-Schlösser, 1993; Van der Meijden et al., 2004) homoplasy occurs in morphological characters as well. Alternatively, the placement of these taxa into separate subfamilies could have been be due to the lesser amount of attention that this large and highly diverse African ranid fauna has received relative to the other ranids, and therefore an artefact of observation. The genera in the endemic southern African clade were not only considered to belong to five different subfamilies or families, they also are morphologically and ecologically most distinct. Cacosternum are small frogs of generalized ecology and reproductive biology, many Tomopterna are burrowing savanna-dwelling frogs, Afrana are generalized semi-aquatic frogs, Natalobatrachus bonebergi is a semi-arboreal species living along rainforest streams, and Pyxicephalus are giant bullfrogs possessing fang-like projections of the lower jaw and a complex parental care behaviour. The genus Anhydrophryne is also likely to belong to this clade based on previously published mitochondrial data (Vences et al., 2000). These hogsback frogs live in humid South African forests and have direct development (Channing, 2001). Other South African genera of the Cacosterninae probably belong to the endemic southern

56 Chapter 3

African clade as well, although molecular data are lacking so far (see caption to figure 3.4) which would further increase the ecological diversity in this lineage The trend of species-richness of the southern African clade (figure 3.4) does not match the distribution of the total amphibian diversity, which tends to be highest in the humid region around Cameroon and an area covering Zambia, Mozambique, Tanzania and the Southeast of the Democratic Republic of the Congo (Stuart et al., 2004). Some species and genera of the African endemic clade (especially Pyxicephalus and Tomopterna) have succeeded in colonizing vast savanna areas of Africa, but other genera in this clade are restricted to Southern Africa, such as Natalobatrachus and Anhydrophryne. On the contrary, there is no genus in the clade restricted to any other region of Africa. This suggests that these frogs originated in this region and some lineages subsequently radiated across sub-Saharan Africa. The large range in biology and ecology suggest an adaptive radiation in these frogs similar to that of the Leptodactylidae in South America. South Africa has a high endemism for flora, as it has the entire Cape Floristic Region within its borders (Mittmeier et al., 1998). The degree of endemism for amphibians is also spectacularly high; 54% of the 118 frogs that occur in South Africa are only found there. Of the 51 ranids that occur there, 27 (53%) are endemic (calculated using data from AmphibiaWeb.org). Conservation International has marked the Cape Floristic Region as one of the world’s biodiversity hotspots with 5,682 endemic plant species and 53 endemic vertebrates (Myers et al., 2000). Future data will allow testing possible biogeographic correlates between the high floral endemism of the Cape Floristic Region and the southern African diversity hotspot of the endemic ranid clade identified herein. An entire radiation at the family level associated to some degree with the Cape Floral region will further strengthen its status as a biodiversity hotspot, and can possibly serve as a flagship example of endemic biodiversity.

57 Chapter 3

Figure 3.4 Left map shows percentages of all ranid species as part of the total frog species numbers per country, with absolute number of ranids in each country (Data from AmphibiaWeb.org). Percentage of the ranid species belonging to the endemic Southern African clade of the total ranid species count is shown on the right. Numbers represent absolute species numbers of frogs from the endemic Southern African clade. Frogs occurring in Lesotho and Swaziland were included in the South African species counts. The analysis considered all representatives demonstrated to belong to the Southern African clade herein (figure 3.2) and the genus Anhydrophryne that was closely related to Cacosternum in the molecular study of Vences et al. (2000). Inclusion of the other African cacosternine genera (Arthroleptella, Microbatrachella and Nothophryne (Blommers-Schlösser, 1993)) would lead to an even stronger diversity hotspot in South Africa.

Endemic African ranids are phylogenetically basal Despite the inclusion of almost 3 kbp sequence data into our analysis, basal relationships among major ranoid clades remained largely unsolved. The lack of resolution basal within the Ranoidae, in contrast to the good resolution at levels below and above, could be a ‘hard’ polytomy caused by a relatively rapid radiation of the Ranidae. Alternatively, this pattern could be alleviated by the inclusion of more species. Independent from the uncertainty of their precise phylogenetic position, various species had high genetic divergences from all other taxa included. This applies to the African Petropedetes, Ptychadena, Phrynobatrachus and Conraua, but also to the Asian Ceratobatrachus, a Solomon and Bourgainville island representative of platymantines that are endemic to the

58 Chapter 3

Philippines, Papua New Guinea, the Moluccas, New Britain, Admiralty, Solomon and Fiji islands. The long branches of these taxa (figure 3.2) are indicative of long evolutionary histories. The age of their splits from their closest relatives were estimated between 92-86 mya (table 3.3). As predicted by Roelants et al. (Roelants et al., 2004) the inclusion of these taxa leads to the identification of areas of endemism for deep evolutionary ranid lineages in addition to South Asia, namely the Philippine and Pacific region, and southern and central Africa. The long isolation of Africa subsequent to Gondwana fragmentation has, similar to the faunal histories of South America and Australia, allowed for the development of a unique endemic mammalian radiation: the Afrotheria (Murphy et al., 2001). This radiation includes a range of animals as dissimilar as elephants, aardvarks and golden moles. The time of the onset of the Afrotheria radiation has been estimated to be 79.9 mya, with 95% confidence intervals of 73.0-85.8 mya, by Springer et al. (2003). Bayesian analysis of the divergence time of the endemic ranid clade with Pyxicephalus as the most basal taxon provided an estimate of 69.9 (48.9-96.5) mya. Divergence time of the clade including Petropedetes and Conraua is 85.6 (61.1-116.4) mya (table 3.3). This indicates that the radiation of the endemic African ranid clade may have occurred roughly in the same period as the afrotherian radiation, although the large confidence intervals of our estimates inhibit more precise interpretations. This endemic clade and the other deep African lineages identified in our study, therefore contain a large amount of evolutionary history. This should be taken into consideration when outlining conservation strategies. In fact, in South Africa, Lesotho and Swaziland (Minter et al., 2004), 30% (11 out of 37) species belonging to the endemic African clade (if cacosternines are considered as belonging to it entirely) are in a threatened red list category, while this applies to only 18% (14 out of 78) of the remaining frog species. This might be seen as indication for nonrandom extinction processes.

3.6 Conclusion The amphibian decline problem and the persistent elusiveness of its causes highlight the limits of our knowledge of amphibian biodiversity and emphasize the urgency of the need for further studies in order to design informed conservation measures. The discovery of a

59 Chapter 3 clade that is supported by several independent nuclear as well as mitochondrial markers, but which has eluded workers using morphological characters, is indicative of the need of a well-resolved molecular phylogeny of amphibians. Possibly, high levels of phenotypic homoplasy so far hindered the discovery of reliable morphology-based phylogenetic relationships, especially of the ranoids. Clearly, further studies are necessary to investigate the characters that are particularly homoplasious. The uncovered endemic clade may be affected more strongly than others by declines, thereby stressing the importance of a phylogenetic framework for effective conservation priority assessment.

Acknowledgements We are grateful to Carla Cicero, curator of the collections at the Museum of Vertebrate Zoology in Berkeley for granting tissue loans of several species (Tissue loan # 6205). Alan Channing and Marius Burger assisted during fieldwork. We thank Ylenia Chiari, Dirk Steinke, David B. Wake and two anonymous reviewers for comments on the manuscript. This work was supported by grants of the Deutsche Forschungsgemeinschaft to AM and MV.

60 Chapter 4

Molecular phylogenetic evidence for paraphyly of the genus Sooglossus, with the description of a new genus of Seychellean frogs

Silhouette island of the Seychelles archipelago. Inset: Sooglossus sechellensis. Photos by Renaud Boistel

Chapter 4

4. Molecular phylogenetic evidence for paraphyly of the genus

Sooglossus, with the description of a new genus of Seychellean frogs

Submitted: Van der Meijden, A., Boistel, R., Gerlach, J., Ohler, A., Vences, M., Meyer, A. Molecular phylogenetic evidence for paraphyly of the genus Sooglossus, with the description of a new genus of Seychellean frogs. Submitted to the Biological Journal of the Linnean Society, 27 October 2005

4.1 Abstract The Seychelles harbor an endemic frog family, the Sooglossidae, currently containing two genera: Sooglossus, with three species, and Nesomantis, with one species. These unique frogs are generally considered to be basal neobatrachians, although their relationships to other neobatrachian taxa, except the Nasikabatrachidae, remain unresolved. Our molecular phylogeny based on a dataset consisting of fragments of the nuclear rag-1 and rag-2 genes as well as mitochondrial 16S rRNA in representatives of the major neobatrachian lineages confirmed the previously postulated Sooglossidae + Nasikabatrachidae clade and the placement of the South American Caudiverbera with the Australian Myobatrachidae, but did not further resolve the position of sooglossids. Our results do, however, unambiguously show sooglossids to be monophyletic but the genus Sooglossus to be paraphyletic, with the type species Sooglossus sechellensis being more closely related to Nesomantis thomasseti than to Sooglossus gardineri and S. pipilodryas, in agreement with morphological, karyological and bioacoustic data. As a taxonomic consequence we propose to consider the genus name Nesomantis as junior synonym of Sooglossus, and to transfer the species thomasseti to Sooglossus. For the clade composed of the species gardineri and pipilodryas, we here propose the new generic name Leptosooglossus. A significant genetic differentiation of 3% was found between specimens of Sooglossus thomasseti from the Mahé and Silhouette Islands, highlighting the need for further studies on their possible taxonomic distinctness.

62 Chapter 4

4.2 Introduction The frogs of the basal neobatrachid family Sooglossidae are restricted to mossy mountain forests on two islands of the Seychelles archipelago: Mahé and Silhouette. These granitic islands on the Mascarene plateau are highly isolated Gondwanan fragments in the Indian Ocean, 1000 kilometers south of India (Briggs, 2003). The Seychelles have been isolated from other landmasses for about 67-47 My, although the islands may have been connected to each other during the last glaciation (Badyukov et al., 1989). The family Sooglossidae consists of four species divided into two genera; Sooglossus with three small-sized species, and Nesomantis with a single medium-sized species. All four species are listed as “vulnerable” on the IUCN Red List (www.globalamphibians.org, accessed 9 July 2005) due to their restricted ranges, being associated with moist upland rainforests. Sooglossus sechellensis (Boettger, 1896) and Sooglossus gardineri (Boulenger, 1911) are leaf-litter inhabiting species, whilst Sooglossus pipilodryas Gerlach and Willi, 2003 is arboreal and usually lives in the axils of endemic palms and banana trees. Nesomantis thomasseti Boulenger, 1909 is associated with rock overhangs. S. gardineri is one of world’s smallest tetrapods with a snout-vent length of only 9.8 mm in adult males. The relationships of the Sooglossidae to other frogs have been a subject of continuing debate. Except for their placement with the Pelobatidae (Griffiths, 1963), they usually were classified with modern lineages of frogs that several authors (Feller and Hedges, 1998; Hay et al., 1995; Van der Meijden et al., 2005) defined as Neobatrachia. Partly following the classification of Dubois (2005) who proposed to discontinue the use of this name as formal name of a suborder of frogs, we here use only the name "neobatrachians" informally to refer to this clade. Within neobatrachians, the Sooglossidae have been grouped with the ranoids (Savage, 1973), the Microhyloidea (Blommers-Schlösser, 1993), and in the hyloid superfamily sister to the Myobatrachinae (Duellman and Trueb, 1986; Ford and Cannatella, 1993; Lynch, 1973). The last comprehensive review of the taxonomic position of this family was provided more than twenty years ago by Nussbaum (1984). Recent molecular studies using different markers have also failed to compellingly place the Sooglossidae with any other neobatrachian group, with the notable exception of their strongly supported sister group relationship to the newly discovered Nasikabatrachidae

63 Chapter 4 from India (Biju and Bossuyt, 2003). Phylogenies using DNA sequences coding for 12S and 16S rRNA (Hay et al., 1995), 12S, 16S and the valine tRNA, together with rhodopsin and single exon fragments of rag-1 and CXCR-4 (Biju and Bossuyt, 2003), rag-1 and rag-2 (Hoegg et al., 2004a) and a larger fragment of rag-1 (San Mauro et al., 2005) placed the sooglossids, or sooglossids plus nasikabatrachids (Biju and Bossuyt, 2003), either as an isolated clade basal in the neobatrachians, or related to other neobatrachian clades with very low phylogenetic support. A phylogenetic pattern that became obvious from recent molecular studies on deep anuran relationships (Biju and Bossuyt, 2003; Hoegg et al., 2004a; Roelants and Bossuyt, 2005; San Mauro et al., 2005) is the existence of two well-defined and very species-rich neobatrachian clades, the ranoids and the hyloids, and of several other, more basal lineages of largely unclarified relationships. Besides the sooglossids and nasikabatrachids, these basal assemblages include the heleophrynids from South Africa, the myobatrachids from Australia, and Caudiverbera from southern South America which previously was believed to belong to the Leptodactylidae. Firmly established relationships among these taxa would be highly informative for the reconstruction of anuran biogeography and the effect of the fragmentation of Gondwana (Biju and Bossuyt, 2003). However, recent phylogenetic studies did not include sooglossids together with representatives of all other basal lineages, and all used only Nesomantis to represent the Sooglossidae. A molecular test for the monophyly of sooglossids is so far missing, and only a single morphological character unique to this family (os sesamoides tarsale present) is known (Nussbaum, 1982). At the level of intrafamilial relationships in the Sooglossidae, it has long been suspected that the genus Sooglossus might be paraphyletic with respect to Nesomantis (Noble, 1931). Additional morphological evidence for paraphyly of Sooglossus has since been accumulated (Gerlach and Willi, 2002; Griffiths, 1963). We here provide the first intrafamilial molecular phylogenetic hypothesis of the Sooglossidae, based on two nuclear genes and one mitochondrial ribosomal RNA gene. Our study also includes representatives of all deep neobatrachian lineages identified so far, in order to test for monophyly and relationships of the Sooglossidae.

64 Chapter 4

4.3 Materials and methods Selection of taxa All four currently described species of the Sooglossidae were included in this study. We used the following specimens; Sooglossus gardineri (4-2003, R. Boistel, collection of the Muséum National d'Histoire Naturelle, Paris, MNHN 2003.3410, Silhouette, Jardin marron, 400 m a.s.l.), Sooglossus pipilodryas (4-2003, R. Boistel, MNHN 2003.3411, Silhouette, Jardin marron 350 m a.s.l.), two specimens of Sooglossus sechellensis (4- 2003, R. Boistel, MNHN 2003.3412-3413, Silhouette, Jardin marron 450 m a.s.l.), and two specimens of Nesomantis thomasseti (5-7-2001, collected by L. Chong Seng and A. Ohler, MNHN 2001.0269, Mahé; 4-2003, R. Boistel, MNHN 2003.3414, Silhouette, Jardin marron 350 m a.s.l.). We also selected representatives of several hyloid and ranoid families available from GenBank (see table 4.1) in order to place the clade formed by the Sooglossidae and Nasikabatrachus with its closest taxon within the neobatrachians. We included Nasikabatrachus sahyadrensis, although no rag-2 sequence is available for this species. Of the ranoid superfamily we selected Rana temporaria as a representative ranid, Kaloula pulchra representing the Microhylidae and the hyperoliid Hyperolius viridiflavus. The superfamily Hyloidea was represented by the bufonid Bufo regularis, the hylid Agalychnis callidryas and the leptodactylids Leptodactylus fuscus and Caudiverbera caudiverbera. Also included were the basal neobatrachians Heleophryne regis of the Heleophrynidae, and Lechriodus melanopyga representing the Myobatrachidae. The archaeobatrachid Pipa parva was used as outgroup.

Sequencing and alignment DNA was extracted from toe clips fixed in 99% ethanol. Tissue samples were digested using proteinase K (final concentration 1 mg/mL), homogenised and subsequently purified following a high-salt extraction protocol (Bruford et al., 1992). Primers for rag-1 and rag-2 were from Hoegg et al. (Hoegg et al., 2004a) as reported in Chiari et al.(2004). Primers for the fragment of the 16S rRNA gene were 16SA-L and 16SB-H of Palumbi et al. (1991). PCR was performed in 25 µl reactions containing 0.5-1.0 units of REDTaq DNA Polymerase (Sigma, Taufkirchen, Germany), 50 ng genomic DNA, 10 pmol of each

primer, 15 nmol of each dNTP, 50 nmol additional MgCl2 and the REDTaq PCR reaction

65 Chapter 4

buffer (in final reaction solution: 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.1 mM MgCl2 and 0.01% gelatine). For rag-1 and rag-2 cycle conditions were adapted from a long range PCR protocol (Barnes, 1994), with an initial denaturation step at 94°C for 5 minutes, followed by ten cycles with 94°C for 30 seconds, annealing temperatures increasing by 0.5°C per cycle from 52 to 57°C and extending for 3 minutes at 68°C. Additionally, 20 cycles were performed with 94°C for 10 seconds, 57°C for 40 seconds and 68°C for 3 minutes. The final extension was done at 68°C for 5 minutes. For 16S the denaturation step was followed by 35 cycles of denaturation at 94° for 30 seconds, annealing at 50° for 30 seconds and extension at 72° for 90 seconds. PCR products were purified via spin columns (Qiagen). Sequencing was performed directly using the corresponding PCR primers (forward and reverse). DNA sequences of both strands were obtained using the BigDye Terminator cycle-sequencing ready reaction kit (Applied Biosystems Inc.) on an ABI 3100 capillary sequencer using the manufacturer’s instructions. New sequences were combined with existing sequences taken from GenBank in the final dataset. These sequences were deposited in GenBank (for accession numbers see table 4.1). Chromatograms were checked by eye using Sequencher (Gene Codes Corp., Ann Arbor, USA) or Chromas v.1.45 (Technelysium Pty Ltd, Tewantin, Australia) and the sequences were subsequently aligned. Rag-1 and rag-2 sequences were aligned by hand using the Mega3 alignment editor (Kumar et al., 2004). The 16S sequences were aligned using ClustalW (Thompson et al., 1994) and subsequently edited by hand. Gapped and hypervariable sections of the 16S alignment were removed from the full alignment. Hypervariable regions were included in the dataset containing only Nasikabatrachus and the sooglossids.

66 Chapter 4

Table 4.1 Taxa included in this study and their GenBank accession numbers (missing numbers to be added upon manuscript acceptance). *sequence of Heleophryne purcelli used. Species Family 16S rag-1 rag-2 Agalychnis callidryas Hylidae AY330890 AY323765 AY323780 Bufo regularis Bufonidae AY330891 AY323763 AY323784 Heleophryne regis Heleophrynidae AF432230* AY323764 AY323786 Hyperolius viridiflavus Hyperoliidae AF215441 AY323769 AY323789 Kaloula pulchra Microhylidae AY330893 AY323772 AY323790 Lechriodus melanopyga Myobatrachidae ####### AY583341 ####### Caudiverbera caudiverbera Leptodactylidae (?) ####### AY583337 ####### Leptodactylus fuscus Leptodactylidae AY263226 AY323770 AY323791 Nasikabatrachus sahyadrensis Nasikabatrachidae AY364381 AY364225 n.a. Pipa parva Pipidae AY333690 AY323761 AY323799 Rana temporaria Ranidae AF249048 AY323776 AY323803 Nesomantis thomasseti “Mahé” Sooglossidae AY330889 AY323778 AY323798 Nesomantis thomasseti “Mahé” Sooglossidae ####### ------Nesomantis thomasseti Sooglossidae AY364373 ------Nesomantis thomasseti “Silhouette” Sooglossidae ####### ------Nesomantis thomasseti “Silhouette” Sooglossidae X86288 ------Sooglossus sechellensis Sooglossidae ####### ####### ####### Sooglossus sechellensis Sooglossidae ####### ------Sooglossus gardineri Sooglossidae ####### ####### ####### Sooglossus pipilodryas Sooglossidae ####### ####### #######

Data analysis A homogeneity partition test (Farris et al., 1994) as implemented in PAUP* (Swofford, 2002) rejected homogeneity of the different markers (P=0.06). Besides a combined analysis of the combined data set we therefore also performed separate analyses of each of the various genes. Transitions and transversions were plotted against F84 distances (Felsenstein, 1984) for the separate gene alignments. None of the datasets showed signs of saturation. Phylogeny reconstruction based on the separate and combined datasets was performed using Maximum Likelihood (ML) and Bayesian Inference (BI) methods. The best fitting models of sequence evolution were determined by the AIC criterion in Modeltest 3.06 (Posada and Crandall, 1998). ML tree searches were performed using PhyML, version 2.4.4 (Guindon and Gascuel, 2003). Bootstrap branch support values were calculated with 500 replicates. The Bayesian analyses of the combined and separate datasets was conducted with MrBayes 2.0 (Huelsenbeck and Ronquist, 2001), using models estimated

67 Chapter 4

with Modeltest under the AIC criterion, with 250,000 generations, sampling trees every 10th generation (and calculating a consensus tree after omitting the first 3000 trees). To get further indications whether the paraphyly of the genus Sooglossus in our analysis may be caused by introgression phenomena or sample contamination regarding the clade containing Sooglossus sechellensis and Nesomantis, and to test for the differentiation among individuals of Nesomantis from different populations, we sequenced the 16S rRNA gene from a second specimen of S. sechellensis and from a Nesomantis specimen from Silhouette, added three Nesomantis sequences from GenBank, and submitted this 16S data set to a separate ML analysis with Nasikabatrachus as outgroup.

Microtomography We used the following adult specimens: four Nesomantis thomasseti, three Sooglossus sechellensis, four Sooglossus gardineri and three Sooglossus pipilodryas. The animals were deposited in a small tube of polypropylene. Microtomography in absorption-based form (amplitude contrast), consists in recording several hundreds of radiographs, with the sample slightly rotated between exposures, and it uses a standard filtered back-projection algorithm to perform the 3D reconstruction. The detector uses a FReLoN 1024x1024 and 2048x2048 camera (Labiche et al., 1996), and involves an optical microscope assembly between the X-ray sensitive converter and the CCD. The effective pixel size is varied by altering the visible-light part of the assembly. Three different sets of experiments were performed on the imaging beamline ID19 of the ESRF, with pixel sizes respectively 7.5 µm at a sample-detector distance of 40 mm, 10 µm at 40 mm specimen-to-detector distance. and 30 µm at 40 mm specimen-to-detector distance. The experiment was performed to obtain a hight resolution 3D image of the skull. It was performed with the synchrotron radiation monochromatized to 17, 20 and 20.5 keV by a double-crystal silicon monochromator operating in the vertical plane. The flux at the level of the sample for a beam current of 60-180 mA and a wiggler as X-ray source is about 8 109 photons/s/mm2. Image visualization was performed using Amira 3.1 software from TGS and the public domain ImageJ program developed at the United States National Institute of Health.

68 Chapter 4

4.4 Results The final combined dataset consisted of 1447 base pairs of rag-1, 810 base pairs of rag-2, and 435 base pairs of 16S rRNA resulting in a combined alignment of 2692 base pairs. The 16S rRNA alignment had 118 parsimony informative sites, and 256 conserved sites. Rag-1 and rag-2 had 445 and 325 parsimony informative, and 815 and 337 conserved sites respectively. Basal neobatrachian relationships were poorly resolved with all datasets. In both the ML and BI analyses, the rag-1 and rag-2 datasets resolved the hyloid and ranoid clades and a separate myobatrachid clade, whereas the 16S rRNA dataset provided poor resolution at this level. All separate and combined analyses were congruent in the resolution of a clade representing the Sooglossidae with high support. All datasets using both phylogeny reconstruction methods resolved identical relationships within the Sooglossidae with high support. Sooglossids were split in two clades: Sooglossus gardineri and Sooglossus pipilodryas were sister taxa, as were Nesomantis thomasseti and Sooglossus sechellensis. Sooglossids were placed sister to Nasikabatrachus in all analyses, and Caudiverbera clustered with the myobatrachid Lechriodus based on both nuclear datasets. This sister relationship of Caudiverbera and Lechriodus was not supported by the 16S dataset, which grouped Caudiverbera with Heleophryne, albeit with only low support values (68% bayesian posterior probability, and 66% ML bootstrap support). Heleophryne was weakly associated with the clade formed by Caudiverbera and Lechriodus in the combined analyses (figure 4.1). In order to be able to include published data on additional sooglossid individuals, we performed a second analysis based on the 16S rRNA gene only. This dataset included the two N. thomasseti samples (from Mahé and Silhouette) available to us, as well as three further sequences of this species from GenBank, two individuals of S. sechellensis, single individuals of S. gardineri and S. pipilodryas, and Nasikabatrachus as the outgroup.

69 Chapter 4

Figure 4.1 Maximum likelihood phylogram of the combined dataset of rag-1, rag-2 and 16S rRNA. Numbers indicate bootstrap support percentages of 500 replicates. A single asterisk indicates a bayesian posterior probability of over 0.97. Two asterisces indicate a bayesian posterior probability of 1.0. Sizes of insets not to scale.

The obtained tree (figure 4.2) provided the same arrangement of taxa as the combined analysis and furthermore placed sequences of N. thomasseti from Mahé and Silhouette, respectively, in two separate subclades. Genetic differentiation between sooglossid taxa was of similar levels as known for other amphibians. Pairwise Jukes-Cantor corrected distance (Jukes and Cantor, 1969) between

70 Chapter 4

S. gardineri and S. pipilodryas based on 16S was 5.7%, whereas the distance between S. sechellensis and N. thomasseti was 4.4%. The rag-1 and rag-2 fragments showed a similar pattern; the JC corrected pairwise distances between S. gardineri and S. pipilodryas were 4.0% for both rag-1 and rag-2, and 2.3% and 2.5% between S. sechellensis and N. thomasseti, respectively. The mean 16S rRNA-based Jukes-Cantor distance between the Silhouette and Mahé specimens of N. thomasseti was 3.0%. The mean 16S rRNA based distance between the S. gardineri/S. pipilodryas clade and the N. thomasseti/S. sechellensis clade was 16.6% .

Figure 4.2 Maximum likelihood phylogram based on a reduced dataset (16S rRNA only). Numbers indicate bootstrap percentages of 500 replicates. Different 16S rRNA sequences of Nesomantis available through GenBank were included: X86288 (Hay et al., 1995), AY364373 (Biju and Bossuyt, 2003), AY330889 (Hoegg et al., 2004a).

71 Chapter 4

4.5 Discussion Relationships among basal neobatrachians The phylogenetic relationship of the clade formed by the Sooglossidae and Nasikabatrachus to other neobatrachians could not be resolved in our analyses. Although correct assignment of most species to the larger superfamilies Ranoidea and Hyloidea is usually unproblematic, the current work, just as previous studies using molecular characters (Biju and Bossuyt, 2003; Hay et al., 1995; Hoegg et al., 2004a; San Mauro et al., 2005), failed to provide resolution among the more basal neobatrachian taxa. Like the sooglossids, most of these basal taxa are species-poor, and have only relictal distributions. The global scattering of their small distributions, with the Seychellean sooglossids, Nasikabatrachus highly localized in India, South African heleophrynids, Australian myobatrachids and the possibly closely related Caudiverbera in South America, suggests that these taxa might be remnants of an ancient neobatrachian radiation predating the breakup of Gondwana (San Mauro et al., 2005). Lack of basal resolution among neobatrachian groups might be because inadequate markers (rag-1, 16S rRNA) were used in these studies. Alternatively the initial radiation of neobatrachians could have occurred too fast to be reconstructable from present DNA sequences. Nevertheless, the data published previously and those that are included in the present work, allow for three conclusions regarding the relationships of these frogs: (1) the Sooglossidae are a monophyletic group which (2) is confirmed to be sister to the Indian Nasikabatrachus, thereby validating the biogeographic scenario of Biju and Bossuyt (2003); (3) the placement of Caudiverbera, which is typically included with the Leptodactylidae, sister to the myobatrachid Lechriodus rather than with the leptodactylid Leptodactylus, receives further support as compared to the phylogeny of San Mauro et al. (2005) that was based on only rag-1 sequences. In contrast to the other lineages of basal neobatrachians, the Myobatrachidae are a species-rich taxon with 124 species (including Rheobatrachidae and Limnodynastidae; AmphibiaWeb.org, as of August 2005), and a more comprehensive sampling of these Australian taxa in the future will be crucial to fully understand the phylogenetic and biogeographic pattern of the various clades in this initial neobatrachian radiation.

72 Chapter 4

A striking character of sooglossids is their lack of extensible external vocal sacks (R. Boistel, pers. obs.) and the absence of middle ear ossicles (Parker, 1934). One internal vocal sac with very small vocal slits (1 mm length, Tyler, 1985) is present. If the middle ear is an adaptation to hearing in air (Allen, 1985) and the external vocal sacks are a common solution for the need for communication over greater distances in anurans, the frogs of this family would appear to be effectively deaf and voiceless, which is contradicted by the fact that at least some sooglossids emit advertisement calls (Gerlach and Willi, 2002). The closely related Nasikabatrachus shares the absence of the tympanum (see absence of bone columella in figure 4.3 e, f., Biju and Bossuyt, 2003; Dutta et al., 2004; Parker, 1934). Of the 5157 described species of anurans placed in 32 families (AmphibiaWeb.org), about 6% are earless (Boistel, 2003). Although the Sooglossidae and Nasikabatrachidae form one of the basal clades of the neobatrachians, the majority of neobatrachians do have a middle ear and its loss is most probably secondary (R. Boistel, unpublished data).

Intrafamilial phylogeny and the classification of sooglossids The paraphyly of the genus Sooglossus as shown by our data corroborates the earlier findings based on morphology (Gerlach and Willi, 2002; Griffiths, 1963; Noble, 1931), vocalisations (Nussbaum et al., 1982), genetic distance data (Green et al., 1988) and karyology (Nussbaum, 1979). The high support that this placement receives (based on the separate and combined molecular datasets irrespective of phylogeny reconstruction method) and the low genetic distance between Nesomantis thomasseti and Sooglossus sechellensis, suggests a need for further taxonomic reconsideration of the genus Nesomantis. This robust arrangement may also provide a basis for the study of the evolution of reproductive modes and other features of the biology of these genetically highly distinctive frogs. Originally, Sooglossus was described by Boulenger (1906) and assigned to the family Ranidae in order to separate Arthroleptis sechellensis Boettger, 1896 from the African Arthroleptis on base of an entire, elliptical tongue as the only character. When he discovered a second species (Nesomantis thomasseti) from the Seychelles Islands, Boulenger (1909) described it as a member of a new ranid genus mainly distinguished by

73 Chapter 4 the presence of vomerine teeth and the shape of the digits. The characters used by Boulenger (1882) to redefine Nectophryne, in particular the presence of a fleshy web, allow us to understand his generic allocation of a new species Nectophryne gardineri Boulenger, 1911 as a member of the Bufonidae. Thus the frogs from the Seychelles were historically classified as members of two distinct families, grouping S. sechellensis with N. thomasseti as corroborated by the morphological study of Gerlach and Willi (2002). Available information on reproductive modes also shows some differences within the genus Sooglossus: S. sechellensis deposits eggs in a terrestrial nest. These hatch into non- feeding tadpoles that are transported on the back of the male until metamorphosis (Nussbaum, 1984). Also eggs of N. thomasseti are deposited in a terrestrial nest, and hatch into non-feeding tadpoles (R. Boistel, unpublished data). In contrast, S. gardineri lays terrestrial eggs that hatch into froglets without a free tadpole stage. The breeding habits of S. pipilodryas are unknown. Summarizing, except for body size, there is no convincing morphological difference supporting the recognition of a separate genus Nesomantis to place the species thomasseti separate from sechellensis. The genetic distance between these species is relatively low for frogs (4.4% in the 16S rRNA gene). In contrast, the species gardineri and pipilodryas form a genetically highly distinct clade, supported by molecular data and morphological and behavioural evidence. We suggest that these findings should be reflected in the taxonomy by including both sechellensis and thomasseti in a single genus Sooglossus, with the generic name Nesomantis being a junior synonym, and by describing a new genus Leptosooglossus to accommodate the two remaining sooglossid species, gardineri and pipilodryas.

74 Chapter 4

SOOGLOSSIDAE NOBLE, 1931

Sooglossus Boulenger, 1906

Type species by monotypy: Arthroleptis sechellensis Boettger, 1896 Nesomantis Boulenger, 1909: Type species by monotypy: Nesomantis thomasseti Boulenger, 1909

Species included: Sooglossus sechellensis (Boettger, 1896); Sooglossus thomasseti (Boulenger, 1909)

Diagnosis: Small to medium sized sooglossids (>16 mm SVL) with protruding nostrils, widely separated metacarpal tubercles, pointed finger and toe pads, and tubercular skin; toes free, without web. Osteology (figure 4.3): frontoparietals in contact, at least in their posterior half fused, and fused with occipital; exotosis producing a pair of spines (process) on otoccipital; nasals close to each other and close to preorbital process of maxilla; premaxilla and maxilla bearing densely set teeth; quadratojugal and ventral ramus of squamosal fused or articulated to maxilla; braincase entirely ossified; parasphenoid alae oriented at right angles fused to otic capsules; anterior ramus, well developed, in contact with sphenoid; vomers well developed, bearing or not bearing teeth; palatines well developed and articulating with maxilla; pterygoids well developed, anterior rami in contact with maxillae and medial rami not in contact with otic capsule and not overlapping the parasphenoid alea; denticulate serration on dentary present; a sesamoid at maxillo-mandibular articulation present; operculum entirely ossified; anomocaelous with a single condyle to coccyx, sacral diapophyses greatly dilated, arciferal with a cartilaginous unforked omosternum and sternum, terminal phalanges simple, sharply pointed, ending in a very small knob, no intercalary cartilage between penultimate and ultimate phalanges of fingers and toes.

Phylogenetic definition: The clade stemming from the most recent common ancestor of Sooglossus sechellensis (Boettger, 1896) and Sooglossus thomasseti (Boulenger, 1909).

75 Chapter 4

Reproductive behaviour: In S. sechellensis, eggs are laid on the ground; after hatching, non trophic tadpoles will climb on the back of adult and are carried until metamorphosis. Females of S. thomasseti are known to have large and pigmentless ovarian eggs, and they hath into non-feeding tadpoles (R. Boistel, pers. obs.).

Etymology: Composed by the Classical Greek terms soos, safe, sound, unscathed, unwounded; glossa, tongue.

Leptosooglossus new genus

Type species by present designation: Nectophryne gardineri Boulenger, 1911.

Species included: Leptosooglossus gardineri (Boulenger, 1911); Leptosooglossus pipilodryas (Gerlach and Willi, 2002).

Diagnosis: Small sized sooglossids (< 16 mm SVL) with reduced metacarpal tubercles, reduced toe pads (pointed on feet only or on digit III only), and a smooth skin except for rows of well-defined tubercles; toes with fleshy webs. Osteology (figure 4.3): frontoparietals not in contact, but fused with occipital; exotosis on otoccipitals absent; nasals widely separated and not in contact with preorbital process of maxilla; premaxilla and maxilla bearing distantly set teeth in their anterior half only; quadratojugal and ventral ramus of squamosal fused, but widely separated from maxilla; braincase ossified in its dorsal and posterior part only; parasphenoid alae oriented at right angles fused with bony otic capsules or in contact with cartilaginous otic capsules; anterior ramus, poorly developed, not in contact with sphenoid; vomers reduced, not bearing teeth; palatines reduced, not articulating with maxilla; pterygoids small, anterior rami short, not in contact with maxillae and medial rami not in contact with otic capsule and not overlapping the parasphenoid alea; denticulate serration on dentary absent, but a single toothlike process on each dentary; sesamoid at maxillo-mandibular articulation absent;

76 Chapter 4

operculum partly ossified; anomocaelous with a single condyle to coccyx, sacral diapophyses greatly dilated, arciferal with a cartilaginous unforked omosternum and sternum. terminal phalanges simple, sharply pointed, ending in a very small knob, no intercalary cartilage between penultimate and ultimate phalanges of fingers and toes.

Phylogenetic definition: The clade stemming from the most recent common ancestor of Leptosooglossus gardineri (Boulenger, 1911) and Leptosooglossus pipilodryas (Gerlach and Willi, 2002).

Reproductive behaviour: Females of S. gardineri sit on top of eggs laid in hidden terrestrial sites. Fully metamorphosed froglets of 3-4 mm will hatch out of these eggs. No tadpole carrying. Reproduction of S. pipilodryas is not yet known.

Etymology: Composed by the Classical Greek terms lepton, small, fine; soos, safe, sound, unscatted, unwounded; glossa, tongue.

77 Chapter 4

Figure 4.3 Volume rendering of X-ray microtomography of skulls of Sooglossidae species.:A. Nesomantis thomasseti, MNHN 2001-0269, Mahé, resolution 30 µm, B. Sooglossus sechellensis, MNHN 1984-2371, Mahé, resolution 7.5 µm, C. Sooglosuss gardineri, MNHN 1984-2369, Mahé, resolution 7.5 µm, D. Sooglossus pipilodryas, RBSS 2003-0007, Silhouette, resolution 7.5 µm, E. Detailed latero-posterior view of otoccipital region with exotosis producing a pair of processes in N. thomasseti, F. Detail of lateral view of sesamoid bone of articulation of maxillary mandibular of S. sechellensis. Left - lateral view from left; middle - dorsal view; right - ventral view.

78 Chapter 4

Acknowledgements We are very grateful to the Seychelles Bureau of Standards (SBS) for permission to work at the Silhouette Island, Republic of Seychelles, and the A. Gerlach family (Trust of Conservation of Seychelles) for assisting during fieldwork. Also many thanks to M. Géze and M. Dellinger, for facilitating and accommodating the use of the graphic workstation at CEMIM (MNHN, Paris). Many thanks to the Ministry of environment of Seychelles and the Directory General of Conservation Mr. D. Dogley for delivering collecting permits used for the specimens mentioned in this paper. Finally, thanks to A. Dubois, J. M. Boistel , M. Boistel, M. Goyon , T. Aubin, A. Mazabraud, P. Cloetens , E. Boller, P. Tafforeau, J.-F. Aubry, J.-Y. Tinevez, and the teams of id19 in ESRF and Laboratory of Reptilia and Amphibia (MNHN) for help. The ESRF is acknowledged for providing beamtime in the framework of proposal SC-1593 and MD179. This work was financially supported by the Deutsche Forschungsgemeinschaft (grant ME1725/10-1).

79 Chapter 5

Nuclear gene phylogeny of narrow-mouthed toads, family Microhylidae, and a discussion of competing hypotheses for their origin

Gastrophryne carolinensis

Chapter 5

5. Nuclear gene phylogeny of narrow-mouthed toads, family

Microhylidae, and a discussion of competing hypotheses for their origin

5.1 Abstract The family Microhylidae has a large circumtropic distribution and contains about 400 species in a highly subdivided taxonomy. Relationships among its constituent taxa remained controversial due to homoplasy in morphological characters, resulting in conflicting phylogenetic hypotheses. A phylogeny based on four nuclear genes (rag-1, rag-2, tyrosinase, BDNF) and one mitochondrial gene (CO1) of representatives of all currently recognized subfamilies uncovers a basal polytomy between the several subfamilial clades. A sister group relationship between the cophylines and scaphiophrynines is resolved with moderate support, which unites these endemic Malagasy taxa for the first time. The American members of the subfamily Microhylinae are resolved to form a clade entirely separate from the Asian members of that subfamily. Otophryne is excluded from the subfamily Microhylinae, and resolved as a basal taxon. The Asian dyscophine Calluella is placed nested within the Asian Microhyline clade, rather than with the genus Dyscophus as had been suggested before Bayesian estimates of the divergence time between extant Microhylidae (47-90 Mya) and its subgroups are discussed in frameworks of alternative possible biogeographic scenarios.

5.2 Introduction Amphibians have long been regarded as model organisms for the study of biogeography due to their alleged inability to disperse across salt water. Several recent molecular studies have shed new light on the biogeography of anurans (Biju and Bossuyt, 2003; Bossuyt and Milinkovitch, 2001; San Mauro et al., 2005; Van der Meijden et al., 2005). Although continental drift and terrestrial dispersal are generally assumed to be the dominant forces shaping large scale patterns of anuran biogeography, transoceanic dispersal has been shown to have occurred in several instances (Hedges et al., 1992; Vences et al., 2004; Vences et al., 2003b).

81 Chapter 5

The frogs of the family Microhylidae occur in the Americas, sub-Saharan Africa, Madagascar, India and most of Southeast Asia to New Guinea and northernmost Australia, whilehe highest numbers of species are found in Southeast Asia and Madagascar. This wide distribution makes the Microhylidae a model case for biogeographical inference. A typical Gondwanan distribution suggests that their current distribution is primarily due to large-scale vicariance events. Savage (1973) proposed that the early microhylids were present on Gondwana, and their current distribution on Gondwanan continents is due to the break-up of the supercontinent. India and Australia subsequently were the origin of the Southeast Asianmicrohylids through dispersal from there, while some of the South American taxa spread northward to North America. More recently, Feller and Hedges (1998) proposed an "Out of Africa" pattern of dispersal for this family. They suggested a dispersal from Africa to Asia and to North and South America. This pattern, however, cannot explain the presence of microhylids on Madagascar due to vicariance, and hence these authors suggest that Microhylids have reached Madagascar through dispersal from Africa. These biogeographic scenarios have remained untestable due to a lack of a robust phylogenetic hypothesis for this family. Approximately one in every five frog genera belongs to the Microhylidae. While this family encompasses 20% (64) of all frog genera, it includes only 8% (approximately 400) of the world’s frog species (AmphibiaWeb, 2006), a low average number of species per genus. There are on average only 5.2 species per genus, while families with a comparably wide distribution like the Bufonidae and the Hyperoliidae have 13.8 and 13.4 species per genus respectively. Of the 64 genera, 22 are monotypic. Although they are generally characterized as stout bodied mostly fossorial ant- and termite specialists, microhylids are actually rather diverse in their biology. Most are fossorial or terrestrial, but also arboreal forms exist. The Asterophryinae and Genyophryninae of Southeast Asia and Australia undergo direct development, whereas all cophylines have endotrophic, non-free living tadpoles (McDiarmid and Altig, 1999). Both the Microhylinae and Hoplophryninae subfamilies (see table 5.1) contain species that have endotrophic larvae and species that have feeding and free swimming larvae. Although free exotrophic larvae are considered to be plesiomorphic for ranoid frogs, only about a third of the microhylid species retained this characteristic. A robust phylogenetic

82 Chapter 5 hypothesis necessary to study the evolution of endotrophic larvae, direct development and its relation to the unique microhylid tadpole, is lacking thus far. Microhylids also display a high level of morphological diversity. This high variability in morphological characters, especially of the cranium and pectoral girdle (Blommers- Schlösser, 1993; Parker, 1934), and high levels of homoplasy due to loss of pectoral girdle elements in members of all subfamilies has complicated the use of these characters for microhylid taxonomy (Zweifel, 1986). The high variability of these elements might be a by-product of miniaturization (as in Hanken and Wake, 1993), independently recurring in several clades. Wild (1995) attributed the high variability in their osteological characters to the effects of miniaturization and the inferred repeated evolution of fossoriality. This high diversity and frequent homoplasy of morphological characters (Blommers-Schlösser, 1993; Loader et al., 2004; Zweifel, 1986) is probably the cause of the severe taxonomic subdivision of the Microhylidae in eight subfamilies and the high genus to species ratio.Its biological and morphological diversity organized in a highly subdivided taxonomy, in combination with their puzzling distribution, has gained the Microhylidae a reputation of being systematically difficult (Duellman, 1979; Ford and Cannatella, 1993). In this work we will use the terms Microhylidae and microhylids sensu Dubois (2005) to refer to the monophyletic microhylid group excluding the African subfamily Brevicipitinae, which has been shown to be closer related to hyperoliids and arthroleptids than to the non-brevicipitine microhylids in previous studies (Darst and Cannatella, 2004; Loader et al., 2004; van der Meijden et al., 2004). Recent molecular phylogenetic studies (Andreone et al., 2005; Hoskin, 2004; Loader et al., 2004; Sumida et al., 2000; Van der Meijden et al., 2004) focussed on a comparatively small part of the circumtropic distribution of the Microhylidae, or concentrated on one or more subtaxa. In this paper, we present the first inclusive molecular phylogeny of the Microhylidae, sampling all ten subfamilies. We also included a range of ranoid and arthroleptoid frogs as outgroups, to further establish the position of the Microhylidae relative to these taxa. We discuss the systematic and biogeographic implications of our findings.

83 Chapter 5

Table 5.1 Listing of currently recognized subfamilies and genera of the Microhylidae. Compiled using AmphibiaWeb.org (AmphibiaWeb, 2006) and the Amphibian Species of the World Database (Frost, 2004), and altered to partly reflect changes proposed by Dubois (2005). Taxa written in bold font were sampled in this study. Genera previously in the subfamily Brevicipitinae are here considered as belonging to a separate family Brevicipitidae. Asterophryinae Scaphiophryninae Cophylinae Asterophrys Scaphiophryne Anodonthyla Barygenys Paradoxophyla Cophyla Callulops Madecassophryne Hylophorbus Hoplophryninae Platypelis Mantophryne Hoplophryne Plethodontohyla Pherohapsis Parhoplophryne Rhombophryne Xenobatrachus Stumpffia Xenorhina Microhylinae Microhylinae cont’ Genyophryninae Adelastes Kalophrynus Albericus Altigius Kaloula Aphantophryne Arcovomer Melanobatrachus Austrochaperina Chaperina Metaphrynella Choerophryne Chiasmocleis Microhyla Cophixalus Ctenophryne Micryletta Copiula Dasypops Myersiella Genyophryne Dermatonotus Nelsonophryne Liophryne Elachistocleis Otophryne Oreophryne Gastrophryne Phrynella Oxydactyla Gastrophrynoides Ramanella Sphenophryne Glyphoglossus Relictivomer Hamptophryne Stereocyclops Hyophryne Synapturanus Hypopachus Syncope Uperodon Dyscophinae Calluellinae Phrynomerinae Dyscophus Calluella Phrynomantis

5.3 Materials and methods Higher-level classification Numerous classificatory schemes for anurans have been proposed but, although tending to incorporate novel phylogenetic information as it became available, they disagreed in the ranks assigned to family-group taxa. Van der Meijden et al. (2005) distinguished three monophyletic lineages within a superfamily Ranoidea: Arthroleptoidae, Microhyloidae, Ranoidae. These lineages were named "epifamilies" following Dubois (1992) but in a subsequent paper, Dubois (2005) suggested a different epifamily definition in which this rank applies to units above the superfamilial level. We disagree with the classification proposed by Dubois (2005) which sunk a number of well-

84 Chapter 5 established and largely well-defined families such as Arthroleptidae, Astylosternidae, Hemisotidae, Hyperoliidae as subfamilies into a new family Brevicipitidae. Likewise we consider the new taxonomic assignments of ranid subfamilies by Scott (2006) as premature and probably in many cases not reflecting the correct phylogeny. The high rates of discovery of new species as well as active work on the phylogenetic relationships of amphibians (Köhler et al., 2005) also requiremodifications of their classification and, in our opinion, warrant the recognition of additional new families rather than a reduction of the number of family group taxa. For the purpose of this paper we follow the scheme used in Van der Meijden et al. (2005) without referring to "epifamilial" taxa but rather will describe the clades, and accept the Brevicipitidae as a separate family (including the genera Balebreviceps, Breviceps, Callulina, Probreviceps, Spelaeophryne) based on the results reported in Van der Meijden et al. (2004) and herein.

Selection of taxa Our sampling consisted of 34 microhylid species representing all currently recognized subfamilies (see table 5.1 and appendix 2). We included several outgroup taxa to resolve the relationships of the Microhylidae to its closest relatives. To represent the two major clades in the Ranoidea besides the Microhylidae we included (1) the arthroleptid Arthroleptis, the astylosternid Trichobatrachus, the brevicipitids Breviceps and Callulina, the hemisotid Hemisus, and the hyperoliids Hyperolius and Leptopelis, and (2) the ranid Rana and two mantellids of the genus Mantidactylus. The hylids Litoria and Agalychnis, and the pipids Xenopus and Pipa as well as two species of Alytes were included to further represent major groups in the anuran phylogeny. We further included the salamander Lyciasalamandra, a bird and a mammal to account for the synapsid-diapsid split for molecular clock calibration, and the lungfish Protopterus as outgroup.

Sequencing and alignment DNA was extracted from toe clips fixed in 99% ethanol. Tissue samples were digested using proteinase K (final concentration 1 mg/mL), homogenized and subsequently

85 Chapter 5

purified following a high-salt extraction protocol (Bruford et al., 1992). Primers for rag-1 and rag-2 were from Hoegg et al. (2004b) as reported in Chiari et al. (2004). Primers for tyrosinase from Bossuyt and Milinkovitch (2000) were used as in Vences et al. (2003b). Primers were designed for an amplification of a 700 bp fragment of BDNF (Brain- derived neurotrophic factor) (BDNF.Amp.F1 ACCATCCTTTTCCTTACTATGG, BDNF.Amp.R1 CTA TCT TCC CCT TTT AAT GGT C). CO1 primers were from Hebert et al. (2003a; 2004b) and used as in Vences et al. (2005). PCR was performed in 25 µl reactions containing 0.5-1.0 units of REDTaq DNA Polymerase (Sigma, Taufkirchen, Germany), 50 ng genomic DNA, 10 pmol of each primer, 15 nmol of each dNTP, 50 nmol additional MgCl2 and the REDTaq PCR reaction buffer (in final reaction

solution: 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.1 mM MgCl2 and 0.01% gelatin). For rag-1 and rag-2 cycle conditions were adapted from a long range PCR protocol (Barnes, 1994), with an initial denaturation step at 94°C for 5 minutes, followed by ten cycles with 94°C for 30 seconds, annealing temperatures increasing by 0.5°C per cycle from 52 to 57°C and extending for 3 minutes at 68°C. Additionally, 20 cycles were performed with 94°C for 10 seconds, 57°C for 40 seconds and 68°C for 3 minutes. The final extension was done at 68°C for 5 minutes. PCR products were purified via spin columns (Qiagen). Sequencing was performed directly using the corresponding PCR primers (forward and reverse). DNA sequences of both strands were obtained using the BigDye Terminator cycle-sequencing ready reaction kit (Applied Biosystems Inc.) on an ABI 3100 capillary sequencer using the manufacturer’s instructions. New sequences were combined with existing sequences taken from GenBank in the final dataset. New sequences were deposited in GenBank (for accession numbers see appendix 2). Chromatograms were checked by eye using Sequencher (Gene Codes Corp., Ann Arbor, USA) or Chromas v.1.45 (Technelysium Pty Ltd, Tewantin, Australia) and the sequences were subsequently aligned using the Mega3 alignment editor (Kumar et al., 2004). The sequences were aligned using ClustalW (Thompson et al., 1994) and edited by hand when necessary. Despite attempts to design specific primers, for a few taxa, not all gene sequences could be obtained (appendix 2). Despite a few outgroup taxa and two species where complete data sets were available from other, clearly congeneric species (Stumpffia and Microhyla), this mainly regards Hoplophryne (only partial rag-1 sequence, COI

86 Chapter 5

missing) and Hamptophryne (tyrosinase and COI missing). Missing genes were coded as “missing” (?) in the concatenated dataset. In BI and MP analyses of large multi-gene data sets of hylid frogs, no relationship has been found between completeness of the sequence data of a taxon, and the support values the taxon receives (Wiens et al., 2005), suggesting that the limited amount of missing data in our concatenated alignment is unlikely to distort the phylogenetic results. Furthermore, the placement of taxa with missing data in our analysis was congruent between the single gene analyses in which the taxon was represented by a complete sequence, and the combined dataset, suggesting that omission of one or more genes from the combined set did not influence the placement of particular taxa.

Data analysis A homogeneity partition test (Farris et al., 1994) as implemented in PAUP* (Swofford, 2002) rejected homogeneity of the different markers (P=0.01). Besides a combined analysis of the combined data set we therefore also performed separate analyses of each of the various genes. Transitions and transversions were plotted against F84 distances (Felsenstein, 1984) for the separate gene alignments. Of the nuclear genes, only tyrosinase showed slight deviance from a linear relationship between the number of transitions and the genetic distance for the most distantly related outgroup taxa. This gene was therefore excluded from the divergence time estimates. Phylogeny reconstruction based on the separate and combined datasets was performed using Maximum Likelihood (ML) and Bayesian Inference (BI) methods. The best fitting models of sequence evolution were determined by the AIC criterion in Modeltest 3.7 (Posada and Crandall, 1998). ML tree searches were performed using PhyML, version 2.4.4 (Guindon and Gascuel, 2003). Bootstrap branch support values were calculated with 500 replicates. The Bayesian analyses of the combined and separate datasets were conducted with MrBayes 3.1.1 (Huelsenbeck and Ronquist, 2001), using models estimated with Modeltest under the AIC criterion, with 250,000 generations, sampling trees every 10th generation (and calculating a consensus tree after omitting the first 3000 trees).

87 Chapter 5

Divergence time estimation Bayesian divergence time estimates were conducted using the software packages Paml (Yang, 1997) and Multidivtime (Thorne and Kishino, 2002) using a dataset consisting of only the nuclear genes rag-1, rag-2 and BDNF (2919 bp). The tyrosinase dataset was excluded due to the possibility of saturation among the most distantly related taxa. The following calibration points were used: (1) minimum age of the frog-salamander split at 230 Mya (fossil record of frog ancestor Triadobatrachus, (Sanchiz, 1998)); (2) minimum age of the split between Agalychnis and Litoria at 42 Mya (last connection between Australia and South America, (Seddon et al., 1998)); (3) maximum age of the split between Mantidactylus wittei and Mantidactylus sp. from the Comoro islands at 15 Mya (volcanic origin of the oldest Comoro island Mayotte, (Vences et al., 2003b)); (4) minimum age of the Alytes muletensis-Alytes dickhilleni split at 5 Mya (Mediterranean salinity crisis: Fromhage et al., 2004); (5) age interval of the split between diapsids and synapsids at 338-288 Mya (Graur and Martin, 2004); (6) a minimum age of 338 Mya for the divergence between Lissamphibia and Amniota based on the aïstopod fossil, Lethiscus stocki (Ruta et al., 2003); (7) minimum age for the divergence of the South American Pipa and the African Xenopus of 110 Mya, corresponding to the final separation of South America from Africa (Sanmartin and Ronquist, 2004).

5.4 Results The concatenated alignment with a total of 4143 bp consisted of rag-1, rag-2, BDNF, tyrosinase and CO1 with 1380, 819, 720, 651 and 573 bp respectively. Rag-1 showed 773 variable, and 637 parsimony-informative sites, whereas rag-2 had 593 variable and 516 parsimony-informative positions. BDNF appeared to have a relatively slow mutation rate, with only 221 variable and 114 parsimony-informative sites. Tyrosinase had 399 variable sites of which 333 were parsimony informative. The mitochondrial CO1 alignment had 267 variable sites and 241 were parsimony-informative positions. No well supported incongruencies among the results of the separate gene analyses and the combined analysis were found. The ML phylogeny based on the complete dataset (figure 5.1), rooted with Protopterus, Homo, Gallus, Alytes and the pipids (excluded from

88 Chapter 5

figure 5.1) shows a highly supported clade composed of the Hylidae as represented by Agalychnis and Litoria sister to the Ranoidea clade. Resolution within base of the Ranoidea clade is low, but the three major clades within the Ranoidea (see Van der Meijden et al., 2005) are highly supported. The hyperoliid Leptopelis is placed as sister to the astylosternid Trichobatrachus and separated from the second included hyperoliid, Hyperolius. Sister to the clade formed by the arthroleptids, astylosternids and hyperoliids , the brevicipitids form a clade as sister group to Hemisus (Hemisotidae), providing further evidence that they are not part of the Microhylidae as shown earlier using only rag-1 (Van der Meijden et al., 2004). Within the Microhylidae, basal resolution is low, but there are several well resolved clades. All phylogenetic hypotheses from the separate and combined analyses resolve the cophylines as a clade irrespective of the phylogenetic analyses used. Another highly resolved clade contains the included representatives of the South American Microhylinae (tribus Gastrophrynini), which was also supported to being a clade in the separate analyses of all nuclear genes except thosed based on the tyrosinase gene alone. Notable in this clade is the highly nested position of the North American taxa Hypopachus and Gastrophryne. The Asian microhylines included form a separate clade. The Malagasy dyscophines are placed as sister to this Asian microhyline clade with low support, but this placement was consistent in all single gene analyses. Asterophrys and Cophixalus, represent the mainly New Guinean subfamilies Asterophryinae and Genyophryninae, are sister groups with high support. The scaphiophrynines Scaphiophryne and Paradoxophyla were resolved as sister taxa, forming a clade sister to the also Malagasy cophylines. Although this placement has only moderate support (83% ML bootstrap support, bayesian posterior probability 1.00), it was supported as well by the separate analyses of both rag-1 and rag-2. The Asian genus Calluella is placed nested well among the Asian Microhylinae (tribus Microhylini), and not with the Malagasy Dyscophinae. The positions of both Otophryne and Hoplophryne are basal within the Microhylidae.

89 Chapter 5

Figure 5.1 ML tree, hierarchical outgroups omitted. Values are bootstrap support percentages of 500 replicates. Values under 50% not shown. A single asterisk indicates a bayesian posterior probability of over 0.95, two asterices indicate a bayesian posterior probability of 1.00. Drawings show representative species of major clades.

90 Chapter 5

Although the placement of Hoplophryne is inconsistent between the different single gene analyses, Otophryne is placed consistently basal within the microhylid clade, and not with the remaining Gastrophrynini. The Phrynomerinae are also placed basally in the Microhylidae clade.

Divergence time estimates Different methods were used to estimate the divergence times of the major groups included based on the rag-1, rag-2 and BDNF datasets only (2919 characters). The results are shown in table 5.2.

5.5 Discussion Novel phylogenetic relationships among the Microhylidae No comprehensive systematic study of the Microhylidae has been undertaken since the work of Parker (1934) (but see an unpublished Ph.D. thesis by S-H. Wu from the University of Michigan, 1994). Our study provides strong support for a number of clades within this family. Many of these agree with currently recognized taxonomic units, but several novel phylogenetic relationships were discovered and some which had only been poorly supported appear resolved by our data. One of the more surprising results is the moderately supported placement of the Madagascar endemic taxon Scaphiophryninae as sister to the endemic Malagasy Cophylinae. Such a sister group relationship of these Malagasy subfamilies has, to our knowledge, hitherto not been proposed. Although the often proposed position of the Scaphiophryninae as sister to the remaining Microhylidae (Scoptanura) (Duellman and Trueb, 1986; Ford and Cannatella, 1993; Haas, 2003; Wassersug, 1984; Wassersug, 1989) cannot yet be unambiguously ruled out, its placement as sister to the Cophylinae is rather well supported by our data and biogeographically more parsimonious, and therefore demands closer scrutiny. Based on the great differences between the tadpole of Scaphiophryne, which possess a horny beak and oral denticles and resembles a ranoid tadpole, and Paradoxophyla, which lacks all keratinized oral structures and which is a

91 Chapter 5 typical microhylid tadpole, it has been suggested that the scaphiophrynines might not be monophyletic (Ford and Cannatella, 1993).

Table 5.2 Divergence time estimates (Age), 95% Confidence Interval (CI), and corresponding standard deviation (SD). Ages of higher taxa are the divergence times of the most basal branches in that clade, based on the taxa included in this study. Constraints imposed by calibration points are given, and the ages of these nodes are indicated in bold face. Clade Age (Mya) ± SD (My) 95% CI (My) Lissamphibia - Amniota (min. 338 Mya) 356±46 281-446 Anura - Caudata (min. 230 Mya) 326±40 255-412 Diapsids - Synapsids (338-288 Mya) 305±41 232-392 Anura 215±28 164-276 Ranoidea – Hyloidea 151±32 116-197 Xenopus - Pipa (min. 110 Mya) 142±23 102-193 Ranoidea 119±17 89-157 Microhylidae - "arthroleptoids" 116±17 87-153 "Arthroleptoids" 103±16 76-137 Hyperolius - (Arthroleptis - (Trichobatrachus - Leptopelis)) 77±13 55-106 Brevicipitidae – Hemisus 75±13 53-104 Mantellidae – Ranidae 61±11 43-84 Litoria - Agalychnis (min. 42 Mya) 53±9 42-75 Mantidactylus wittei - M. sp. "Comoros" (max. 15 Mya) 11±2 6.7-15 Alytes muletensis - A. dickhilleni (min. 5 Mya) 10±5 2.3-23 Microhylidae Microhylidae 66±11 47-90 Microhylini – Asterophryinae 57±10 40-79 Microhylini – Dyscophinae 55±10 39-76 Scaphiophryninae – Cophylinae 53±9 38-74 Gastrophrynini 53±10 37-74 Microhylini 52±9 37-72 Phrynomerinae 47±9 32-66 Scaphiophryninae 45±8 31-64 Cophylinae 32±7 22-47 Cophixalus – Asterophrys 20±5 12-30 Dyscophinae 19±5 11-29 Hypopachus – Gastrophryne 17±4 10-27

Haas (2003) suggested based on larval characters that Paradoxophyla was more closely related to the Phrynomerinae. Our molecular data provide strong evidence for a sister- group relationship of Paradoxophyla and Scaphiophryne. Parsimony would argue that the morphology of the Scaphiophryne tadpole, intermediate between a ranoid and microhylid larval morphology, should therefore be interpreted to be a reversal from the microhylid

92 Chapter 5

type. The alternative, considering the Scaphiophryne tadpole as the ancestral form for microhylids, would instead imply the unlikely hypothesis that Scaphiophryne has evolved its filter-feeding tadpole morphology fully in parallel to other microhylids. The isolated position of Otophryne corroborates the suggestion of Wassersug and Pyburn (1987) that this species is not a microhyline. They suggested its unique characters might merit the erection of a new subfamily. The phylogenetic hypothesis by Wild (1995), based on the combined morphological data from Zweifel (1986) and Donnelly et al. (1990), instead indicated that Otophryne is nested in the American microhyline clade (tribus Gastrophrynini). Our molecular data resolve Otophryne as a distinct taxon in the microhylid clade, not closely associated with any other microhylid subclade, based both on the combined as well as the single gene analyses. Although basal resolution in the current phylogeny is too low to interpret the position of Otophryne as being basal, the possibility of such a placement allows us to consider some of the unique morphological traits of the psammonic larva, especially the keratinized mouthparts, as a possible plesiomorphic character rather than an adaptation related to its burrowing lifestyle. Excluding Otophryne, the remaining American Microhylinae (tribus Gastrophrynini) form one of the most distinct and well-supported groups within the microhylid clade, clearly unrelated to the Asian Microhylinae (tribus Microhylini). This is congruent with the findings of Haas (2003) using larval morphological characters. Resolution within this group is high and the two North American taxa, Hypopachus and Gastrophryne, are firmly nested within the remaining South American clade. The Asian Microhyline taxa (tribus Microhylini) are here resolved as a highly supported clade. Basal to this clade are the Malagasy dyscophines and the representatives of the mainly New Guinean subfamies Asterophryinae (Asterophrys) and Genyophryninae (Cophixalus). Although the two representatives of these subfamilies studied here were resolved as sister taxa, allozyme data have shown that Cophixalus is more closely related to the asterophryine genus Barygenys than to the genyophrynine Sphenophryne (Sumida et al., 2000), and Cophixalus might therefore not be a proper representative of the Genyophryninae. Despite the inclusion of over 4kbp of sequence data, the basal relationships among the well resolved microhylid clades remain somewhat unresolved. The inclusion of more

93 Chapter 5 sequence data, and a more comprehensive taxon sampling might improve the resolution in this part of the tree. The contrast between the lack of resolution basally within the Microhylidae with the otherwise well resolved clades could point to a “biological” polytomy, due to a fast initial radiation. This pattern is similar to that found basally in the Ranoidae (Van der Meijden et al., 2005), and similarly complicates the selection among alternative biogeographic scenarios.

Relationships among "arthroleptoids" The position of the hyperoliid Leptopelis as sister to the astylosternine Trichobatrachus, corroborates the findings of Emerson et al. (2000) that Leptopelis is distinct from the hyperoliids, and those of Vences et al. (2003a), that the leptopelines are closely related to the astylosternids. Since the remaining hyperoliids are likely to form a well-supported clade, a consequence would be to either include Leptopelis in the Astylosternidae or erect a new family Leptopelidae. However, we suggest postponing this decision until more molecular data on additional astylosternid genera become available to provide evidence whether Leptopelis is the sister group of, or nested within, the Astylosternidae.Our study supplies the first robust evidence of the close relation of Hemisus to the brevicipitines. This relationship has previously been hypothesized by Blommers-Schlösser (1993) based on morphological arguments but Channing (1995) showed that only a single synapomorphy (a single median thyroid gland) supports the sister relationship between the Brevicipitinae and Hemisus. Molecular datasets so far included only representatives of one of these two taxa (e.g., Van der Meijden et al., 2005: only Breviceps; Biju and Bossuyt, 2003: only Hemisus), or did not include sufficient outgroup taxa (Loader et al., 2004).

Alternate biogeographic scenarios for the Microhylidae Divergence time estimates place the initial divergence within the Microhylidae at 66 Mya (with a 95% confidence interval of 47-90 Mya), but the last common ancestor of the Microhylidae and their closest ranoid relatives at 116 Mya (47-90 Mya). The early

94 Chapter 5 microhylids found themselves thus on Gondwana, around or after the disconnection of Madagascar and India from the mainland 121 Mya (Sanmartin and Ronquist, 2004). Several scenarios have been proposed to explain their current distribution.

Feller and Hedges (1998) proposed that the origin of the Microhylidae lies in Africa, and subsequent dispersal to Asia, North- and South America led to the current distribution. This scenario requires no vicariance events. Our data show the North American microhylines as nested within the South American clade. This contradicts the hypothesis of the colonization of South America from North America, unless one postulates extinction and recolonization for North America. Feller and Hedges propose a dispersal route for the hyloid frogs to North America across the proto-Antilles in the late Cretaceous. It is unlikely that the microhylids also used this route. Divergence time estimates place the most recent common ancestor of the North American taxa Hypopachus and Gastrophryne at 17 Mya (10-27 Mya), closer to the connection of North- and South America through the Panama isthmus approximately 5-3.5 Mya. The North American microhyline taxa might therefore have dispersed from South America across the Panama isthmus or across the sea prior to the great faunal interchange. The dispersal of the Microhylidae to Asia after the reconnection of Africa to Eurasia through the Arabian peninsula, latest 10 Mya (McQuarrie et al., 2003) as proposed by Feller and Hedges, is in disagreement with our divergence time estimates. The initial radiation of the Microhylini is estimated at 52 Mya (37-72 Mya, well before the reconnection of Africa to Eurasia. Additionally, the divergence between the Asian taxa and the Malagasy Dyscophus is estimated at 55 Mya (39-76 Mya). This figure provides evidence against a direct dispersal from Africa to Eurasia, but is also somewhat incongruent with the separation of Madagascar and India by the Mascarene basin, 84 Mya, and therefore provides little support for the “biotic ferry” hypothesis (Bossuyt and Milinkovitch, 2001; Duellman and Trueb, 1986). Conversely, Savage (1973) proposed a scenario based almost entirely on vicariance due to the fragmentation of Gondwana. Savage suggested that vicariance events led to the presence of microhylids in South America, Africa, Madagascar, India and Australia. Madagascar and India, however, had broken off from Africa (121 Mya) before the initial

95 Chapter 5 divergence of microhylids at 66 Mya (47-90 Mya). It is therefore unlikely that the presence of cophylines and scaphiophrynines on Madagascar is due to vicariance. Similarly, Africa and South America had separated at 110 Mya, before the initial radiation of the extant Microhylidae, and therefore vicariance is unlikely in light of this divergence time estimate. The split between the Microhylidae and the remaining ranoids, however, is estimated at 116 Mya (87-153 Mya), and therefore still allows for a possible vicariant event between early microhylids on Africa and South America. If a vicariance event between the African and South American microhylids is assumed, microhylids could have dispersed from South America to Australia/New Guinea through Antarctica and to Madagascar/India through Antarctica and a land bridge formed by the Kerguelen plateau. The connection to India/Madagascar remained until at least 80 Mya (Sampson et al., 1998), whereas South America remained connected to Australia and New Guinea through Antarctica until 35 Mya. This places these dispersal routes within the 95% confidence interval for the initial radiation of the Microhylidae. The African continent remained in isolation during this period, making South America a more likely source of dispersal. A basal position of the Asterophryinae and Genyophryninae to the Microhylini allows for a possible route through Australia and New Guinea for these taxa. The initial divergence between the Australo-New Guinean taxa is estimated at 20 Mya, but only two possibly closely related taxa (Sumida et al., 2000) are included in our study. Asterophryines and genyophrynines diverged from the Microhylini at 57 Mya (40-79 Mya), allowing for the possibility that the ancestors of these taxa were present on the Australia-New Guinea landmass before it lost contact with South America and Antarctica. The fossil asterophryine Australobatrachus, from the Oligocene to Miocene of Australia provides further evidence for this possibility. Counter to this scenario is the low species diversity presently found in Australia compared to that on New Guinea. The origins of the Microhylini in Asia have been proposed to be due to their presence in Madagascar-India (Duellman and Trueb, 1986). This hypothesis does provide for the presence of the dyscophines in Madagascar, if the placement of the dyscophines basal to the Microhylini resolved here is correct. This scenario also assumes that the bayesian divergence time estimate between the Dyscophinae and Microhylini of 55 Mya is an underestimation, as Madagascar and India were severed from each other around 88 Mya.

96 Chapter 5

The alternative, dispersal of Dyscophus ancestors from Asia via India to Madagascar, as it has been proposed for boas and iguanas (Rage, 1996), requires at least one event of overseas dispersal and extinction of ancestral dyscophines in India and Asia. Another possibility is the dispersal of the Microhylini through the Arab peninsula after its connection to Eurasia (latest 10 Mya: McQuarrie et al., 2003) but this would require a large error in the molecular divergence time estimations. Alternatively, in addition to India/Madagascar and Australia/New Guinea, both South America and Africa can have been colonized from Antarctica.

Figure 5.2 Reconstruction of the positions of the Southern continents using data from Hay et al. (1999) contemporary with the initial radiation of the Microhylidae, 66 Mya. The most likely possible dispersal and/or vicariance events leading to the current distribution are indicated with arrows.

From these potential biogeographic hypotheses (figure 5.2) it becomes clear that a still more comprehensive taxon sampling will be necessary to unambiguously resolve the biogeographic history of the Microhylidae. The inclusion of more New Guinean taxa of both subfamilies, as well s inclusion of Australian taxa might possibly resolve the order of colonization of these two areas. Increased taxon sampling, as well as a larger dataset might better resolve the relationships between the Microhylini, the Dyscophinae and the Genyophryninae and Asterophryninae and thus resolve the dispersal route of all Asian taxa. Inclusion of the enigmatic Indian taxon Melanobatrachus, usually condsidered to be

97 Chapter 5 related to the African Hoplophryne but included in the Microhylinae by Dubois (2005) is potentially informative as well. An increase in the number of genes in the dataset may also additionally improve the resolution of the placement of all basal taxa, and especially that of the more enigmatic taxa such as Otophryne and Hoplophryne, and resolve on which Gondwanan fragment(s) the Microhylidae arose. Eventually, it is not unlikely that the biogeographic history of this and other frog families will turn out to be highly complex, shaped by a mixture of vicariance and dispersal events rendering unidimensional hypotheses as too simplistic and optimistic.

Acknowledgements We thank Jens V. Vindum at the California Academy of Sciences, San Francisco, Steve Richards at the James Cook University, Frank Glaw at the Zoologische Staatssammlung München and Carla Cicero at the Museum for Vertebrate Zoology, Berkeley (loan number 6205), for generously providing tissue samples crucial to this work. Thanks to Dirk Steinke and Frank Glaw for valuable discussions. We thank our fellow microhylophilic colleagues with whom we have enjoyed constructive correspondence under the auspice of AmphibiaTree.

98 Summary

Summary

Knowledge of evolutionary relationships form the basis of every comparative study in biology. Advancement of the knowledge of amphibian relationships is especially imperative because of the global decline of amphibian populations. Due to a general paucity of characters and an apparently general high level of homoplasy, the systematics of ranoid frogs have remained disputed. In this thesis we address the systematics of several groups of ranoid frogs and related taxa using the molecular characters provided by DNA sequences. In the first chapter, a phylogeny based on the single-copy nuclear rag-1 gene revealed unexpected placements of scaphiophrynine and brevicipitine toads. The former have usually been considered as sister group to all other extant microhylids or are even classified as their own family. Their basal position among microhylids was weakly indicated in our analysis; but they clearly were part of a strongly supported clade composed of representatives from five other microhylid subfamilies. In contrast, the brevicipitines, a group that hitherto was unanimously considered to belong to the Microhylidae, were highly divergent and placed as a sister group to the arthroleptoid clade. These novel phylogenetic placements are best reflected by a classificatory status of the Scaphiophryninae as subfamily of the Microhylidae, whereas the brevicipitines may merit recognition as distinct family. Our findings seem to corroborate a high degree of morphological homoplasy in frogs and suggest that even highly derived morphological states, such as the hydrostatic tongue of microhylids, hemisotids and brevicipitines, may be subject to convergent evolution, parallelism or character reversal. Chapter two deals with the identification of lineages through short mitochondrial sequences. Identifying species of organisms by short sequences of DNA has been in the center of ongoing discussions under the terms DNA barcoding or DNA taxonomy. A C- terminal fragment of the mitochondrial gene for cytochrome oxidase subunit I (COI) has been proposed as universal marker for this purpose among animals. Herein we present experimental evidence that the mitochondrial 16S rRNA gene fulfills the requirements for a universal DNA barcoding marker in amphibians. In terms of universality of priming sites and identification of major vertebrate clades the studied 16S fragment is superior to

99 Summary

COI. Amplification success was 100% for 16S in a subset of fresh and well-preserved samples of Madagascan frogs, while various combinations of COI primers had lower success rates. COI priming sites showed high variability among amphibians both at the level of groups and closely related species, whereas 16S priming sites were highly conserved among vertebrates. Interspecific pairwise 16S divergences in a test group of Madagascan frogs were at a level suitable for assignment of larval stages to species (1- 17%), with low degrees of pairwise haplotype divergence within populations (0-1%). We strongly advocate the use of 16S rRNA as standard DNA barcoding marker for vertebrates to complement COI, especially if samples a priori could belong to various phylogenetically distant taxa and false negatives would constitute a major problem. In chapter three we provide systematic resolution to another highly disputed Ranoid taxon; the Ranidae. We show a molecular phylogenetic hypothesis for ranids, including 11 of the 12 African endemic genera. Analysis of nuclear (rag-1, rag-2 and rhodopsin genes) and mitochondrial markers (12S and 16S ribosomal RNA genes) provide evidence for an endemic clade of African genera of high morphological and ecological diversity thus far assigned to up to five different subfamilies: Afrana, Cacosternum, Natalobatrachus, Pyxicephalus, Strongylopus, and Tomopterna. This clade has its highest species diversity in southern Africa, suggesting a possible biogeographic connection with the Cape Floral Region. Bayesian estimates of divergence times place the initial diversification of the southern African ranid clade at ~ 62-85 million years ago, concurrent with the onset of the radiation of afrotherian mammals. These and other African ranids (Conraua, Petropedetes, Phrynobatrachus, Ptychadena) are placed basally within the Ranoidae with respect to the Eurasian groups, which suggests an African origin for this whole epifamily. Chapter four deals with a non-ranoid endemic family from the Seychelles, the Sooglossidae, currently containing two genera: Sooglossus, with three species, and Nesomantis, with one species. These unique frogs are generally considered to be basal neobatrachians, although their relationships to other neobatrachian taxa, except the Nasikabatrachidae, remain unresolved. Our molecular phylogeny based on a dataset consisting of fragments of the nuclear rag-1 and rag-2 genes as well as mitochondrial 16S rRNA in representatives of the major neobatrachian lineages confirmed the

100 Summary previously postulated Sooglossidae + Nasikabatrachidae clade and the placement of the South American Caudiverbera with the Australian Myobatrachidae, but did not further resolve the position of sooglossids. Our results do, however, unambiguously show sooglossids to be monophyletic but the genus Sooglossus to be paraphyletic, with the type species Sooglossus sechellensis being more closely related to Nesomantis thomasseti than to Sooglossus gardineri and S. pipilodryas, in agreement with morphological, karyological and bioacoustic data. As a taxonomic consequence we propose to consider the genus name Nesomantis as junior synonym of Sooglossus, and to transfer the species thomasseti to Sooglossus. For the clade composed of the species gardineri and pipilodryas, we here propose the new generic name Leptosooglossus. A significant genetic differentiation of 3% was found between specimens of Sooglossus thomasseti from the Mahé and Silhouette Islands, highlighting the need for further studies on their possible taxonomic distinctness. In chapter five, the relationships among the subfamilies of the narrow-mouthed frogs (family Microhylidae) are scrutinized. The family Microhylidae has a large circumtropic distribution and high number of species in a highly subdivided taxonomy. Relationships among its constituent taxa remain controversial due to homoplasy in morphological characters, resulting in conflicting datasets. A phylogeny based on four nuclear genes (rag-1, rag-2, tyrosinase, BDNF) and one mitochondrial gene (CO1) of representatives of all currently recognized subfamilies shows a basal polytomy between the several subfamilial clades. A sister group relationship between the cophylines and scaphiophrynines is resolved with moderate support, uniting these Malagasy endemic taxa for the first time. The American members of the subfamily Microhylinae are resolved to form a clade entirely separate from the Asian members of that subfamily. Otophryne is excluded from the Microhylinae, and resolved as a basal taxon unrelated to any other. The Asian dyscophine Calluella is placed nested in the Asian Microhylinae clade, rather than with the genus Dyscophus. Bayesian estimates of the divergence time between extant Microhylidae (47-90 Mya) and its subgroups are discussed in relation to possible biogeographic scenarios.

101 Zusammenfassung

Zusammenfassung

Kenntnisse der evolutiven Verwandtschaftsverhältnisse bilden die Grundlage jeder vergleichenden Studie in der Biologie. Eine Erweiterung des Wissens über die Verwandtschaftsverhältnisse der Amphibien ist besonders erforderlich aufgrund der weltweit schrumpfenden Amphibienpopulationen. Durch einen generellen Mangel an Merkmalen und einem offensichtlich hohen Grad an Homoplasie bleibt die Systematik ranoider Frösche umstritten. In dieser Arbeit behandle ich die Systematik verschiedener Gruppen ranoider Frösche und verwandter Taxa unter Verwendung molekularer Merkmale, die aus der Analyse von DNA Sequenzen hervorgingen. Im ersten Kapitel zeigt die Phylogenie, die auf dem nukleären Einzelkopiegen rag-1 basiert, eine unerwartete Stellung der Scaphiophryninae and Brevicepitinae. Bisher wurden Scaphiophryninae als Schwesterngruppe aller anderen lebenden Microhylidae oder als eigene Familie bezeichnet. Ihre basale Stellung unter den Microhylidae wird durch die Ergebnisse meiner Studie nur schwach unterstützt. Jedoch sind sie eindeutig Teil eines Stammes, der sich aus fünf anderen Vertretern aus Unterfamilien der Microhylidae zusammensetzt. Die Gruppe der Brevicipitinae, welche bisher einstimmig zu den Microhylidae gezählt wurde, ist hoch divergent und bildet eine Schwestergruppe des Stammes der Arthroleptoidae. Aufgrung dieser neuen phylogenetischen Daten sollten die Scaphiophryninae als Unterfamilie der Microhylidae gesehen werden, während die Breviticipinae eine eigene Familie bilden sollten. Unsere Ergebnisse scheinen die Annahme zu stützen, dass bei Fröschen ein hoher Grat an Homoplasie auf morphologischer Ebene vorliegt. Weiterhin weisen sie darauf hin, dass sogar hoch entwickelte morphologische Merkmale, wie die hydrostatische Zunge der Microhylidae, Hemisotidae und der Brevicipitinae konvergenter Evolution, Parallelismus und Merkmal- Umkehrung unterliegen. Das zweite Kapitel behandelt die Identifikation von Abstammungen mit Hilfe von kurzen Abschnitten mitochondrieller Sequenzen. Die Artbestimmung eines Organismus mittels kurzen DNA-Sequenzen befand sich stets im Zentrum von Diskussionen im Zusammenhang mit DNA-barcoding und DNA-Taxonomie. Das c-terminal gelegene Fragment der Untereinheit I des mitochondrialen Gens Cytochrom-Oxidase (COI) wurde

102 Zusammenfassung als universeller Marker für diese Fragestellungen bei Tieren vorgeschlagen. In dieser Arbeit geben wir den experimentellen Beweis dafür, dass das mitochondrielle 16S rRNA Gen die Anforderungen eines universellen Markers für DNA-barcoding in Amphibien erfüllt. Bezüglich der Allgemeingültigkeit von Primerbindestellen und der Identifikation der Hauptstämme der Vertebraten erweist sich das untersuchte 16S Fragment dem COI als überlegen. Die Erfolgsrate der Amplifizierung von 16S Fragmenten lag bei 100% bei einer Reihe frischer sowie gut konservierter Proben madagassischer Frösche, während verschiedene Kombinationen von COI Primern eine geringere Erfolgsrate hatten. COI Primerbindestellen weisen eine hohe Variabilität sowohl zwischen verschiedenen Amphibienruppen, als auch zwischen eng verwandten Arten auf. 16S Primerbindestellen sind hingegen in Vertebraten stark konserviert. Interartliche, paarweise Distanzen im 16S Gen einer Testgruppe madagassischer Frösche waren brauchbar für die Einteilung larvaler Stadien in Arten (1-17%), mit einem niedrigen Grad paarweiser Abweichungen von Haplotypen innerhalb von Populationen (0-1%). Wir befürworten sehr den ergänzenden Gebrauch von 16S rRNA als Standart-Marker für DNA-barcoding in Vertebraten, besonders wenn Proben schon a priori zu verschiedenen phylogenetisch entfernten Taxa gehören und falsch Negative ein grundlegendes Problem darstellen könnten. In Kapitel drei tragen wir zu einer besseren Auflösung der Systematik eines weiteren stark umstrittenen ranoiden Taxon bei, den Ranidae. Wir zeigen eine molekulare Phylogenie der raniden, die 11 der 12 in Afrika endemischen Gattungen einschließt. Die Analyse nukleärer (rag-1, rag-2 und rhodopsin Gene) und mitochondrialer (Gene für 12S und 16S ribosomale RNA) Marker ergibt starke Hinweise für einen endemischen Stamm afrikanischer Gattungen mit hoher morphologischer und ökologischer Diversität, zu dem bisher fünf verschiedene Unterfamilien geteilt werden: Afrana, Cacosternum, Natalobatrachus, Pyxicephalus, Strongylopus und Tomopterna. Da sich die größte Artenvielfalt dieses Stammes in Südafrika findet, scheint eine mögliche biogeographische Verbindung mit der Cape Floral Region vorzuliegen. Bayes'sche Analysen zeigen, dass die Aufspaltung der südafrikanischen Raniden vor ~62-85 Millionen Jahren statt fand, gleichzeitig mit dem Beginn der Radiation der Afrotherischen Säuger. Diese und andere afrikanische Ranidae (Conraua, Petropedetes, Phrynobatrachus, Ptychadena) werden in

103 Zusammenfassung

Bezug zu den eurasischen Gruppen basal innerhalb der Ranoidae eingeordnet. Dies weist auf einen afrikanischen Ursprung der ganzen Epifamilie hin. Das vierte Kapitel beschäftigt sich mit einer nicht-ranoiden, endemischen Familie von den Seychellen, den Sooglossidae. Diese enthält zur Zeit zwei Gattungen: Sooglossus mit drei Arten und Nesomantis mit einer Art. Diese einzigartigen Frösche werden grundlegend zum basale Neobatrachia bezeichnet, obwohl ihre Verwandtschaft zu anderen Neobatrachia, abgesehen von den Nasikabatrachidae, noch ungeklärt ist. Die hier vorliegende molekulare phylogenetische Analyse basiert auf einem Datensatz, welcher aus Fragmenten der nuklearen rag-1 und rag-2 Gene und des mitochondrialen 16SrRNA Gens besteht. Diese Analysen wurden an Vertretern der Hauptgruppen der Neobatrachia durchgeführt und konnten die Existenz der bereits postulierten Stamm der Sooglossidae + Nasikabatrachidae sowie die Zugehörigkeit der südamerikanischen Caudiverbera zu den australischen Myobatrachidae bestätigen. Jedoch konnte die Stellung der Sooglossidae nicht weiter geklärt werden. Unsere Ergebnisse zeigen jedoch deutlich, dass es sich bei den Sooglossidae um eine monophyletische Gruppe handelt, wobei die Gattung Sooglossus paraphyletisch ist. Weiterhin zeigte sich, dass die Art Sooglossus sechellensis näher verwandt ist zu Nesomantis thomasseti als zu Sooglossus gardineri und S. pipilodryas, übereinstimmend mit morphologischen, karyologischen und bioakustischen Daten. Als taxonomische Konsequenz daraus schlagen wir vor, den Gattungsnamen Nesomantis als Juniorsynonym für Sooglossus zu betrachten und die Art thomasseti zu Sooglossus zu zählen. Für den Stamm, der sich aus den Arten gardineri und pipilodryas zusammensetzt, schlagen wir den Gattungsnamen Leptosooglossus vor. Zwischen Exemplaren der Art Sooglossus thomasseti der beiden Inseln Mahé und Silhouette liegt eine eindeutige genetische Differenz von 3% vor. Dies hebt die Bedeutung weiterer Analysen über ihre mögliche taxonomische Verschiedenheit hervor. In Kapitel fünf werden die Verwandtschaften der Unterfamilien der engmäuligen Frösche der Familie Microhylidae untersucht. Die Familie Microhylidae weist eine grosse circumtropische Verbreitung mit einer hohen Anzahl an Arten mit einer stark unterteilten Taxonomie auf. Die Verwandtschaften zwischen den zu dieser Familie gehörenden Taxa bleiben aufgrund von vorliegender Homoplasie auf morphologischer Ebene strittig. Dies resultiert in widersprüchlichen phylogenetischen Theorien. Die

104 Zusammenfassung

Phylogenie von Vertretern aller bis heute bekannten Unterfamilien, welche auf vier nukleären Genen (rag-1, rag-2, tyrosinase, BDNF) und einem mitochondrialen Gen (COI) basiert, zeigt eine basale Polytomie zwischen den Stämmen der Unterfamilien. Die endemischen Taxa Cophlines und Scaphiophrynes aus Madagaskar wurden zum ersten Mal als Schwesterngruppe in dieser Arbeit identifiziert. Die amerikanischen Vertreter der Unterfamilie Microhylinae bilden einen abgegrenzten Stamm von den asiatischen Vertretern dieser Unterfamilie. Othophryne werden nicht zu den Microhylinae gezählt sondern findet sich als basales Taxon, unabhängig zu allen anderen. Die asiatische discophine Calluella wird dem asiatischen Stamm der Microhylinae zugeordnet und nicht zusammen mit der Gattung Dyscophus zur Subfamilie Dyscophinae gezählt. Bayes'sche Analysen der Aufspaltung zwischen lebenden Microhylidae (vor 47-90 Milliarden Jahren) und ihren Untergruppen werden in Bezug zu möglichen biogeographischen Szenarios diskutiert.

105 General References

General references

Albertson, R., J. Markert, P. Danley, and T. Kocher. 1999. Phylogeny of a rapidly evolving clade: The cichlid fishes of Lake Malawi, East Africa. Proceedings of the National Academy of Sciences of the United States of America 96:5107 - 5110.

Allen, J. 1985. Cochlear modeling. IEEE ASSP 2:2-3.

Altschul, S., W. Gish, W. Miller, E. Myers, and D. Lipman. 1990. Basic local alignment search tool. J Mol Biol 215:403 - 10.

AmphibiaWeb. 2006. Information on amphibian biology and conservation. [web application].Berkeley, California: AmphibiaWeb.

Andreone, F., M. Vences, D. R. Vieites, F. Glaw, and A. Meyer. 2005. Recurrent ecological adaptations revealed through a molecular analysis of the secretive cophyline frogs of Madagascar. Molecular Phylogenetics and Evolution 34:315-22.

Badyukov, D. D., E. L. Demidenko, and P. A. Kaplin. 1989. Palaeogeography of the Seychelles Bank and the northwest Madagascar Shelf during the last glacio-eustatic regression (18,000 B.P.). Chinese Journal of Oceanology and Limnology 7.

Barnes, W. M. 1994. PCR amplification of up to 35 kb DNA with high fidelity and high yield from lambda bacteriophage templates. Proceedings of the National Academy of Sciences of the United States of America 91:2216-2220.

Biju, S. D., and F. Bossuyt. 2003. New frog family from India reveals an ancient biogeographical link with the Seychelles. Nature 425:711-713.

Blaustein, A. R., J. M. Romansic, J. M. Kiesecker, and A. C. Hatch. 2003. Ultraviolet radiation, toxic chemicals and amphibian population declines. Diversity & Distributions 9:123-140.

Blaxter, M. L. 2004. The promise of a DNA taxonomy. Philosophical Transactions of the Royal Society of London. Series B 359:669-79.

Blommers-Schlösser, R. M. A. 1993. Systematic relationships of the Mantellinae Laurent 1946 (Anura Ranoidea). Ethology Ecology & Evolution 5:199-218.

Boistel, R. 2003. Convergence évolutive de l'appareil auditif des Anoures. Une communication acoustique aérienne sans oreille moyenne : une spécialisation unique chez les tétrapodes. Actes du Congrès, GTEM, Paris 6:11.

Bossuyt, F., M. Meegaskumbura, N. Beenaerts, D. J. Gower, R. Pethiyagoda, K. Roelants, A. Mannaert, M. Wilkinson, M. M. Bahir, K. Manamendra-Arachchi, P. K. Ng, C. J. Schneider, O. V. Oommen, and M. C. Milinkovitch. 2004. Local endemism within the Western Ghats-sri Lanka biodiversity hotspot. Science 306:479-81.

Bossuyt, F., and M. C. Milinkovitch. 2000. Convergent adaptive radiations in Madagascaran and Asian ranid frogs reveal covariation between larval and adult traits. Proceedings of the National Academy of Sciences of the United States of America 97:6585-6590.

Bossuyt, F., and M. C. Milinkovitch. 2001. Amphibians as indicators of Early Tertiary "out-of-India" dispersal of vertebrates. Science 292:93-95.

106 General References

Boulenger, G. A. 1882. Catalogue of the Batrachia Gradentia s. Caudata in the collection of the British Museum. Taylor & Francis, London.

Boulenger, G. A. 1906. Descriptions of new Batrachians discovered by Mr. G. L. Bates in South Cameroon. Annales and Magazine of Natural History 7:317-323.

Boulenger, G. A. 1909. A list of freshwater fishes, batrachians, and reptiles obtained by Mr. J. Stanley Gardiner's expedition to the Indian Ocean. Transactions of the Linnean Society London 2:291- 300.

Boulenger, G. A. 1911. List of the batrachians, and reptiles obtained by Prof. Stanley Gardiner on his second expedition to the Seychelles and Aldrabra. Transactions of the Linnean Society London 2:375-378.

Briggs, J. C. 2003. The biogeographic and tectonic history of India. Journal of Biogeography 30:381-388.

Bruford, M. W., O. Hanotte, J. F. Y. Brookfield, and T. Burke. 1992. Single-locus and multilocus DNA fingerprinting. Pages 225-270 in The South American herpetofauna: its origin, evolution, and dispersal. Molecular genetic analysis in conservation. (A. R. Hoezel, ed.) IRL Press, Oxford.

Channing, A. 1995. The relationship between Breviceps (Anura: Microhylidae) and Hemisus (Hemisotidae) remains equivocal. Journal of the Herpetological Association of Africa 44:55-57.

Channing, A. 2001. Amphibians of Central and Southern Africa. Cornell University Press, Ithaca.

Chiari, Y., M. Vences, D. R. Vieites, F. Rabemananjara, P. Bora, O. Ramilijaona Ravoahangimalala, and A. Meyer. 2004. New evidence for parallel evolution of colour patterns in Malagasy poison frogs (Mantella). Molecular Ecology 13:3763-3774.

Clarke, B. T. 1981. Comparative osteology and evolutionary relationships in the African Raninea (Anura Ranidae). Monitore zoologico italiano 14:285-331.

Darst, C. R., and D. C. Cannatella. 2004. Novel relationships among hyloid frogs inferred from 12S and 16S mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 31:462-75.

Darwin, C. 1859. On the Origin of Species by Natural Selection, or the Preservation of favoured races in the struggle for life. Appleton & Co., New York.

Daszak, P. 1998. A new fungal disease associated with amphibian population declines: recent research put into perspective. British Herpetological Society Bulletin:38-41.

Deckert, K. 1938. Beiträge zur Osteologie und Systematik ranider Froschlurche. Sber. Ges. Naturf. Freunde Berlin 1938:127-184.

Donnelly, M. A., R. O. de Sá, and C. Guyer. 1990. Description of the tadpoles of Gastrophryne pictiventris and Nelsonophryne aterrima (Anura: Microhylidae), with a review of morphological variation in free-swimming Microhylid larvae. American Museum Novitates:1-19.

Dubois, A. 1992. Notes sur la classification des Ranidae (amphibiens anoures). Bulletin Mensuel de la Societe Linneenne de Lyon 61:305-352.

Dubois, A. 2004. Developmental pathway, speciation and supraspecific taxonomy in amphibians 1. Why are there so many frog species in Sri Lanka? Alytes 22:19 - 37.

Dubois, A. 2005. Amphibia Mundi. 1.1.; An ergotaxonomy of recent amphibians. Alytes 23:1-24.

107 General References

Dubois, A., and A. Ohler. 2001. A new genus for an aquatic ranid (Amphibia, Anura) from Sri Lanka. Alytes 19:81-106.

Duellman, W. E. (ed) 1979. The South American Herpetofauna: Its Orgin evolution and Dispersal. Museum of Natural History The University of Kansas, Lawrence, Kansas.

Duellman, W. E., and L. Trueb. 1986. Biology of Amphibians, Hightstown, N.J., USA; London, England.

Dutta, S. K., K. Vasudevan, M. S. Chaitra, K. Shanker, and R. K. Aggerwal. 2004. Jurassic frogs and the evolution of amphibian endemism in the Western Ghats. Current Science 86:211-216.

Emerson, S. B., C. Richards, R. C. Drewes, and K. M. Kjer. 2000. On the relationships among ranoid frogs: a review of the evidence. Herpetologica 56:209-230.

Farris, J. S., M. Källersjö, A. G. Kluge, and C. Bult. 1994. Testing significance of incongruence. Cladistics 10:315-319.

Feller, A. E., and S. B. Hedges. 1998. Molecular evidence for the early history of living amphibians. Mol Phylogenet Evol 9:509-516.

Felsenstein, J. 1984. Distance methods for inferring phylogenies: a justification. Evolution 38:16-24.

Fisher, R. N., and H. B. Shaffer. 1996. The decline of amphibians in California's Great Central Valley. Conservation Biology 10:1387–1397.

Floyd, R., E. Abebe, A. Papert, and M. Blaxter. 2002. Molecular barcodes for soil nematode identification. Molecular Ecology 11:839-50.

Ford, L. S., and D. C. Cannatella. 1993. The major clades of frogs. Herpetological Monographs 7:94-117.

Fromhage, L., M. Vences, and M. Veith. 2004. Testing alternative vicariance scenarios in Western Mediterranean discoglossid frogs. Molecular Phylogenetics and Evolution 31:308-22.

Frost, D. R. 2004. Amphibian Species of the World: an Online Reference. Version 3.0 (22 August, 2004).American Museum of Natural History, New York, USA.

Funk, D., and K. Omland. 2003. Species-level paraphyly and polyphyly: frequency, causes and consequences, with insights from animal mitochondrial DNA. Annual Review of Ecology, Evolution and Systematics 34:397 - 423.

Gerlach, J., and J. Willi. 2002. A new species of frog, genus Sooglossus (Anura, Sooglossidae) from Silhouette Island, Seychelles. Amphibia- Reptilia 23:445-458.

Graur, D., and W. Martin. 2004. Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Trends in Genetics 20:80-86.

Green, D. M., R. A. Nussbaum, and Y. Datong. 1988. Genetic divergence and heterozygosity among frogs of the family Sooglossidae. Herpetologica 44:113-119.

Griffiths, I. 1963. The phylogeny of the Salientia. Biological Reviews of the Cambridge Philosophical Society 38:241-292.

Guibé, J. 1956. La position systematique des genres Pseudohemisus et Scaphiophryne (Batraciens). Bulletin du Muséum national d'Histoire naturelle, Paris 28:180-182.

108 General References

Guindon, S., and O. Gascuel. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52:696–704.

Haas, A. 2003. Phylogeny of frogs as inferred from primarily larval characters (Amphibia: Anura). Cladistics 19:23-89.

Hanken, J., and D. B. Wake. 1993. Miniaturization of body size: organismal consequences and evolutionary significance. Annual Review of Ecology and Systematics 24:501-519.

Hay, J. M., I. Ruvinsky, S. B. Hedges, and L. R. Maxson. 1995. Phylogenetic relationships of amphibian families inferred from DNA sequences of mitochondrial 12S and 16S ribosomal RNA genes. Molecular Biology and Evolution 12:928-937.

Hay, W. W., R. DeConto, C. N. Wold, K. M. Wilson, S. Voigt, M. Schulz, A. Wold-Rossby, W.-C. Dullo, A. B. Ronov, A. N. Balukhovsky, and E. Soeding. 1999. Alternative global cretaceous paleogeography. Pages 1-47 in The Evolution of Cretaceous Ocean/Climate Systems, Geological Society of America Special Paper 332 (E. Barrera, and C. Johnson, eds.).

Hayes, T. B., A. Collins, M. Lee, M. Mendoza, N. Noriega, A. A. Stuart, and A. Vonk. 2002. Hermaphroditic, demasculinized frogs after exposure to the herbicide atrazine at low ecologically relevant doses. Proceedings of the National Academy of Sciences of the United States of America 99:5476-80.

Hebert, P., A. Cywinska, S. Ball, and J. deWaard. 2003a. Biological identification through DNA barcodes. Proceedings of the Royal Society of London, Series B 270:313 - 321.

Hebert, P., E. Penton, J. Burns, D. Janzen, and W. Hallwachs. 2004a. Ten species in one: DNA barcoding reveals cryptic species in the Neotropical skipper butterfly Astraptes fulgerator. Proceedings of the National Academy of Sciences of the United States of America 101:14812 - 14817.

Hebert, P., M. Stoeckle, T. Zemlak, and C. Francis. 2004b. Identification of birds through COI DNA barcodes. PLoS Biology 2:1 - 7.

Hebert, P. D., A. Cywinska, S. L. Ball, and J. R. deWaard. 2003b. Biological identifications through DNA barcodes. Proc Biol Sci 270:313-21.

Hebert, P. D., and T. R. Gregory. 2005. The promise of DNA barcoding for taxonomy. Syst Biol 54:852-9.

Hebert, P. D., E. H. Penton, J. M. Burns, D. H. Janzen, and W. Hallwachs. 2004c. Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proc Natl Acad Sci U S A 101:14812-7.

Hebert, P. D., S. Ratnasingham, and J. R. deWaard. 2003c. Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society, Series B 270 Suppl 1:S96-9.

Hebert, P. D., M. Y. Stoeckle, T. S. Zemlak, and C. M. Francis. 2004d. Identification of Birds through DNA Barcodes. PLoS Biol 2:e312.

Hedges, S. B., C. A. Hass, and L. R. Maxson. 1992. Caribbean biogeography: molecular evidence for dispersal in West Indian terrestrial vertebrates. Proceedings of the National Academy of Sciences of the United States of America 89:1909-13.

109 General References

Hoegg, S., M. Vences, H. Brinkmann, and A. Meyer. 2004a. Phylogeny and comparative substitution rates of frogs inferred from sequences of three nuclear genes. Molecular Biology and Evolution 21:1188-200.

Hoegg, S. I., M. Vences, H. Brinkmann, and A. Meyer. 2004b. Phylogeny and comparative substitution rates of frogs inferred from sequences of three nuclear genes. Molecular Biology and Evolution 21:1188-1200.

Hoskin, C. J. 2004. Australian microhylid frogs (Cophixalus and Austrochaperina): phylogeny, taxonomy, calls, distributions and breeding biology. Australian Journal of Zoology 52:237-269.

Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

Janzen, D. 2004. Now is the time. Philosophical Transactions of the Royal Society of London B 359:731 - 732.

Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein moleculesin Mammalian protein metabolism (H. N. Munro, ed.) Academic Press, New York.

Köhler, J., D. R. Vieites, R. M. Bonett, F. H. García, F. Glaw, D. Steinke, and M. Vences. 2005. New amphibians and global conservation: A boost in species discoveries in a highly endangered vertebrate group. BioScience 55:693-696.

Kosuch, J., M. Vences, A. Dubois, A. Ohler, and W. Boehme. 2001. Out of Asia: Mitochondrial DNA evidence for an Oriental origin of tiger frogs, genus Hoplobatrachus. Molecular Phylogenetics & Evolution 21:398-407.

Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated software for molecular evolutionary genetics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163.

Labiche, J. C., J. Segura-Puchades, D. Van Brussel, and J. P. Moy. 1996. FRELON camera: Fast REadout LOw Noise. ESRF Newsletter 25:41-43.

Lannoo, M. J., K. Lang, T. Waltz, and G. S. Phillips. 1994. An Altered Amphibian Assemblage - Dickinson County, Iowa, 70-Years After Blanchard,Frank Survey. American Midland Naturalist 131:311- 319.

Laurent, R. F. 1986. Sous classe des lissamphibiens (Lissamphibia). Systematique. Grasse, P-P; Delsol, M [Eds]. Traite de zoologie. Anatomie, systematique, biologie. Tome 14. Batraciens. Apareil uro- genital:1-828 Chapter pagination 594-797.

Loader, S. P., D. J. Gower, K. M. Howell, D. N., M.-O. Rödel, B. T. Clarke, R. de Sá, B. L. Cohen, and M. Wilkinson. 2004. Phylogenetic relationships of African microhylid frogs inferred from DNA sequences of mitochondrial 12S and 16S rRNA genes. Organisms Diversity & Evolution 4:227- 235.

Lötters, S., E. La Marca, S. Stuart, R. Gagliardo, and M. Veith. 2004. A new dimension of current biodiversity loss? Herpetotropicos 1:29-31.

Lynch, J. D. 1973. The transition from archaic to advanced frogs. Pages 133-182 in Evolutionary biology of the Anurans: Comtemporary research on major problems (J. L. Vial, ed.) University of Missouri Press, Columbia.

110 General References

Macey, J., A. Larson, N. Ananjeva, Z. Fang, and T. Papenfuss. 1997. Two novel gene orders and the role of light-strand replication in rearrangement of the vertebrate mitochondrial genome. Mol Biol Evol 14:91 - 104.

Machado, C., and J. Hey. 2003. The causes of phylogenetic conflict in a classic Drosophila species group. Proceedings of the Royal Society, Series B 270:1193 - 1202.

Marmayou, J., A. Dubois, A. Ohler, E. Pasquet, and A. Tillier. 2000. Phylogenetic relationships in the Ranidae. Independent origin of direct development in the genera Philautus and Taylorana. Comptes Rendus de l'Académie des Sciences III 323:287-97.

Marsh, D. M., and P. B. Pearman. 1997. Effects of habitat fragmentation on the abundance of two species of leptodactylid frogs in an Andean montane forest. Conservation Biology 11:1323-1328.

McDiarmid, R. W., and R. Altig (eds) 1999. Tadpoles: The Biology of Anuran Larvae. University of Chicago Press, Chicago.

McQuarrie, N., J. M. Stock, C. Verdel, and B. P. Wernicke. 2003. Cenozoic evolution of Neotethys and implications for the causes of plate motions. Geophysical Research Letters 30:6-1.

Meegaskumbura, M., F. Bossuyt, R. Pethiyagoda, K. Manamendra-Arachchi, M. Bahir, M. C. Milinkovitch, and C. J. Schneider. 2002. Sri Lanka: an amphibian hot spot. Science 298:379.

Meyers, J. J., J. C. O'Reilly, J. A. Monroy, and K. C. Nishikawa. 2004. Mechanism of tongue protraction in microhylid frogs. Journal of Experimental Biology 207:21-31.

Min, M. S., S. Y. Yang, R. M. Bonett, D. R. Vieites, R. A. Brandon, and D. B. Wake. 2005. Discovery of the first Asian plethodontid salamander. Nature 435:87-90.

Mittmeier, R. A., N. Myers, J. B. Thomsen, G. A. B. da Fonseca, and S. Olivieri. 1998. Biodiversity hotspots and major tropical wilderness areas: approaches to setting conservation priorities. Conservation Biology 12:516-520.

Moritz, C., and C. Cicero. 2004. DNA barcoding: promise and pitfalls. PLoS Biol 2:e354.

Murphy, W. J., E. Eizirik, S. J. O'Brien, O. Madsen, M. Scally, C. J. Douady, E. Teeling, O. A. Ryder, M. J. Stanhope, W. W. de Jong, and M. S. Springer. 2001. Resolution of the early placental mammal radiation using Bayesian phylogenetics. Science 294:2348-51.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853-8.

Nee, S., and R. M. May. 1997. Extinction and the loss of evolutionary history. Science 278:692-4.

Noble, G. K. 1931. The Biology of the Amphibia. McGraw-Hill, New York.

Nussbaum, R. A. 1979. Mitotic chromosomes of Sooglossidae (Amphibia: Anura). Caryologia 32:279-298.

Nussbaum, R. A. 1982. Heterotopic bones in the hindlimbs of frogs of the families Pipidae, Ranidae, and Sooglossidae. Herpetologica 38:312-320.

Nussbaum, R. A. 1984. Amphibians of the Seychelles. Pages 379-415 in Biogeography and ecology of the Seychelles Islands (D. R. Stoddart, ed.).

111 General References

Nussbaum, R. A., A. Jaslow, and J. Watson. 1982. Vocalization in frogs of the family Sooglossidae. Journal of Herpetology 16:198-204.

Ohler, A., and M. Delorme. 2006. Well known does not mean well studied: Morphological and molecular support for existence of sibling species in the Javanese gliding frog Rhacophorus reinwardtii (Amphibia, Anura). Comptes rendus Biologies 329:86-97.

Orton, G. L. 1952. Key to the genera of tadpoles in the United States and Canada. American Midland Naturalist 47:382-395.

Palmer, J., K. Adams, Y. Cho, C. Parkinson, Y. Qiu, and K. Song. 2004. Dynamic evolution of plant mitochondrial genomes: Mobile genes and introns and highly variable mutation rates. Proc Natl Acad Sci USA 97:6960 - 6966.

Palumbi, S., A. Martin, S. Romano, W. McMillan, L. Stice, and G. Grabowski. 1991. The Simple Fool's Guide to PCR, Version 2.0.

Parker, B. A. 1934. A monograph of the frogs of the family Microhylidae. The British Museum, London.

Posada, D., and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14:817-818.

Pounds, A., M. Fogden, and J. Campbell. 1999. Biological response to climate change on a tropical mountain. Nature 398:611-614.

Pounds, J. A., M. R. Bustamante, L. A. Coloma, J. A. Consuegra, M. P. Fogden, P. N. Foster, E. La Marca, K. L. Masters, A. Merino-Viteri, R. Puschendorf, S. R. Ron, G. A. Sanchez-Azofeifa, C. J. Still, and B. E. Young. 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439:161-7.

Purvis, A., P. M. Agapow, J. L. Gittleman, and G. M. Mace. 2000. Nonrandom extinction and the loss of evolutionary history. Science 288:328-30.

Rage, J.-C. 1996. Le peuplement animal de Madagascar: une composante venue de Laurasie est-elle envisageable? Pages 27-35 in Actes du Colloque International Biogéographie de Madagascar (W. R. Lourenco, ed.) Société de Biogéographie-Muséum-ORSTOM, Paris.

Roelants, K., and F. Bossuyt. 2005. Archaeobatrachian paraphyly and Pangaean diversification of crown- group frogs. Systematic Biology 54:111-126.

Roelants, K., J. Jiang, and F. Bossuyt. 2004. Endemic ranid (Amphibia: Anura) genera in southern mountain ranges of the Indian subcontinent represent ancient frog lineages: evidence from molecular data. Molecular Phylogenetics and Evolution 31:730-740.

Rozas, J., J. Sanchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19:2496 - 2497.

Ruta, M., M. I. Coates, and D. L. Quicke. 2003. Early tetrapod relationships revisited. Biological Reviews of the Cambridge Philosophical Society 78:251-345.

Sampson, S. D., L. M. Witmer, C. A. Forster, D. W. Krause, P. M. O'Connor, P. Dodson, and F. Ravoavy. 1998. Predatory dinosaur remains from madagascar: implications for the cretaceous biogeography of gondwana. Science 280:1048-51.

112 General References

San Mauro, D., M. Vences, M. Alcobendas, R. Zardoya, and A. Meyer. 2005. Initial diversification of living amphibians predated the breakup of Pangaea. The American Naturalist 165:590-599.

Sanchiz, B. 1998. Encyclopedia of Palaeoherpetology, part 4. Salienta. Pfeil, München.

Sanmartin, I., and F. Ronquist. 2004. Southern hemisphere biogeography inferred by event-based models: plant versus animal patterns. Systematic Biology 53:216-43.

Savage, J. M. 1973. The geographic distribution of frogs: Patterns and predictions. Pages 351-445 in Evolutionary Biology of Anurans: Contemporary Research on Major Problems (J. L. Vial, ed.) University of Missouri Press, Colombia, Missouri.

Scott, E. 2006. A phylogeny of ranid frogs (Anura: Ranoidea: Ranidae), based on a simultaneous analysis of morphological and molecular data. Cladistics 21:507- 574.

Seddon, J. M., P. R. Baverstock, and A. Georges. 1998. The rate of mitochondrial 12S rRNA gene evolution is similar in freshwater turtles and marsupials. Journal of Molecular Evolution 46:460-4.

Shubin, N. H., and F. A. Jenkins. 1995. An early Jurassic jumping frog. Nature 377:49-52.

Sokol, O. M. 1977. A subordinal classification of frogs (Amphibia: Anura). Journal of Zoology, London 182:505-508.

Springer, M. S., G. C. Cleven, O. Madsen, W. W. de Jong, V. G. Waddell, H. M. Amrine, and M. J. Stanhope. 1997. Endemic African mammals shake the phylogenetic tree. Nature 388:61-4.

Springer, M. S., W. J. Murphy, E. Eizirik, and S. J. O'Brien. 2003. Placental mammal diversification and the Cretaceous-Tertiary boundary. Proceedings of the National Academy of Sciences of the United States of America 100:1056-61.

Stuart, S. N., J. S. Chanson, N. A. Cox, B. E. Young, A. S. Rodrigues, D. L. Fischman, and R. W. Waller. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783-6.

Sumida, M., A. Allison, and M. Nishioka. 2000. Evolutionary relationships among 12 species belonging to three genera of the family Microhylidae in Papua New Guinea revealed by allozyme analysis. Biochemical Systematics and Ecology 28:721-736.

Sumida, M., Y. Kanamori, H. Kaneda, Y. Kato, M. Nishioka, M. Hasegawa, and H. Yonekawa. 2001. Complete nucleotide sequence and gene rearrangement of the mitochondrial genome of the Japanese pond frog Rana nigromaculata. Genes & Genetic Systems 76:311 - 325.

Swofford, D. L. 2002. PAUP *. Phylogenetic analysis using parsimony (*and other methods), version 4beta10. Sinauer Associates.

Tautz, D., P. Arctander, A. Minelli, R. H. Thomas, and A. P. Vogler. 2003. A plea for DNA taxonomy. Trends in Ecology & Evolution 18:70-74.

Thomas, M., L. Rhaharivololoniaina, F. Glaw, M. Vences, and D. R. Vieites. 2005. Montane tadpoles in Madagascar: molecular identification and description of the larval stages of Mantidactylus elegans, M. madecassus and Boophis laurenti from the Andringita Massif. Copeia 2005:174-183.

Thomaz, D., A. Guiller, and B. Clarke. 1996. Extreme divergence of mitochondrial DNA within species of pulmonate land snails. Proceedings of the Royal Society, Series B 263:363 - 368.

113 General References

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. Clustal W: improving the sensitivity of the progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Research 22:4673–4680.

Thorne, J. L., and H. Kishino. 2002. Divergence time and evolutionary rate estimation with multilocus data. Systematic Biology 51:689-702.

Thorne, J. L., H. Kishino, and I. S. Painter. 1998. Estimating the rate of evolution of the rate of molecular evolution. Molecular Biology and Evolution 15:1647-57.

Tyler, M. J. 1985. Phylogenetic significance of the superficial mandibular musculature and vocal sac structure of sooglossid frogs. Herpetologica 4:173-176.

Van der Meijden, A., M. Vences, S. I. Hoegg, and A. Meyer. 2005. A previously unrecognized radiation of ranid frogs in Southern Africa revealed by nuclear and mitochondrial DNA sequences. Molecular Phylogenetics & Evolution 36:674-685.

Van der Meijden, A., M. Vences, and A. Meyer. 2004. Novel phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear Rag-1 gene. Proceedings of the Royal Society, Series B 271 Suppl 5:S378-81. van der Meijden, A., M. Vences, and A. Meyer. 2004. Novel phylogenetic relationships of the enigmatic brevicipitine and scaphiophrynine toads as revealed by sequences from the nuclear Rag-1 gene. Proc Biol Sci 271 Suppl 5:S378-81.

Veith, M., J. Kosuch, and R. Feldmann. 2000. A test for correct species declaration of frogs legs imports from Indonesia into the European Union. Biodiversity and Conservation 9:333 - 341.

Vences, M., and F. Glaw. 2001. When molecules claim for taxonomic changes: New proposals on the classification of Old World treefrogs (Amphibia, Anura, Ranoidea). Spixiana 24:85-92.

Vences, M., F. Glaw, J. Kosuch, I. Das, and M. Veith. 2000. Polyphyly of Tomopterna (Amphibia: Ranidae) based on sequences of the mitochondrial 16S and 12S rRNA genes, and ecological biogeography of Malagasy relict amphibian groups. Pages 229-242 in Diversité et Endémisme de Madagascar (W. R. Lourenço, and S. M. Goodman, eds.).

Vences, M., J. Kosuch, F. Glaw, W. Böhme, and M. Veith. 2003a. Molecular phylogeny of hyperoliid treefrogs: biogeographic origin of Malagasy and Seychellean taxa and re-analysis of familial paraphyly. Journal of Zoological Systematics and Evolutionary Research 41:205-215.

Vences, M., J. Kosuch, M.-O. Rödel, Lötters, S., A. Channing, F. Glaw, and W. Böhme. 2004. Phylogeography of Ptychadena mascareniensis suggests transoceanic dispersal in a widespread African-Malagasy frog lineage. Journal of Biogeography 31:593-601.

Vences, M., M. Thomas, A. van der Meijden, Y. Chiari, and D. R. Vieites. 2005. Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians. Frontiers in Zoology 2:5.

Vences, M., D. R. Vieites, F. Glaw, H. Brinkmann, J. Kosuch, M. Veith, and A. Meyer. 2003b. Multiple overseas dispersal in amphibians. Proceedings of the Royal Society Biological Sciences Series B 270:2435-2442.

Venter, J. C., K. Remington, J. F. Heidelberg, A. L. Halpern, D. Rusch, J. A. Eisen, D. Wu, I. Paulsen, K. E. Nelson, W. Nelson, D. E. Fouts, S. Levy, A. H. Knap, M. W. Lomas, K. Nealson, O. White, J. Peterson, J. Hoffman, R. Parsons, H. Baden-Tillson, C. Pfannkoch, Y. H. Rogers, and H. O. Smith. 2004. Environmental genome shotgun sequencing of the Sargasso Sea. Science 304:66-74.

114 General References

Verheyen, E., W. Salzburger, J. Snoeks, and A. Meyer. 2003. Origin of the superflock of cichlid fishes from Lake Victoria, East Africa. Science 300:325-9.

Vredenburg, V. T. 2004. Reversing introduced species effects: Experimental removal of introduced fish leads to rapid recovery of a declining frog. Proceedings of the National Academy of Sciences of the United States of America 101:7646-50.

Wake, D. B. 1991. Declining amphibian populations. Science 253:860.

Wassersug, R. 1984. The Pseudohemisus tadpole: a morphological link between microhylid (Orton type 2) and ranoid (Orton type 4) larvae. Herpetologica 40:138-149.

Wassersug, R. 1989. What, if anything is a microhylid (Orton type II) tadpole? Pages 532-538 in Trends in vertebrate morphology (H. Splechtna, ed.) G. Fischer, Stuttgart, New York.

Wassersug, R. J., and W. F. Pyburn. 1987. The biology of the Pe-ret' toad, Otophryne robusta (Microhylidae), with special consideration of its fossorial larva and systematic relationships. Zoological Journal of the Linnean Society 91:137-169.

Wiens, J. J., J. W. Fetzner, C. L. Parkinson, and T. W. Reeder. 2005. Hylid frog phylogeny and sampling strategies for speciose clades. Systematic Biology 54:719-48.

Wild, E. R. 1995. New genus and species of Amazonian microhylid frog with a phylogenetic analysis of New World genera. Copeia 1995:837-849.

Wilson, A., L. Maxson, and V. Sarich. 1974. Two types of molecular evolution. Evidence from studies of interspecific hybridization. Proceedings of the National Academy of Sciences of the United States of America 71:2843 - 2847.

Wilson, E. O. 2002. The future of life. Abacus, London.

Wilson, E. O. 2004. Taxonomy as a fundamental discipline. Philosophical Transactions of the Royal Society of London B 359:739.

Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood. Computer Applications in Biosciences 13:555-6.

Zardoya, R., E. Malaga-Trillo, M. Veith, and A. Meyer. 2003. Complete nucleotide sequence of the mitochondrial genome of a salamander, Mertensiella luschani. Gene 317:17-27.

Zardoya, R., and A. Meyer. 1996. Evolutionary relationships of the coelacanth, lungfishes, and tetrapods based on the 28S ribosomal RNA gene. Proceedings of the National Academy of Sciences of the United States of America 93:5449-54.

Zweifel, R. G. 1986. A new genus and species of microhylid frog from the Cerro de la Neblina region of Venezuela and a discussion of relationships among New World microhylid genera. American Museum Novitates:1-24.

115 Appendices

Appendix 1

Taxa of the frog family Mantellidae successfully amplified for COI and 16S. + successful amplification (strong bands on agarose gel), (+) moderate success (faint bands), - no success of amplification. Three primer combinations were simultaneously tested for COI. C C C O O O I I H

I ( ( LCO1490 - B B C i i O 1 r r 6 d d 2

Family Vouchers S F F Species 1

9 1 1 8

- - )

R R 1 2 ) )

Anura Aglyptodactylus Mantellidae FGMV 2002.860, 858 - - (+) + madagascariensis Boophis blommersae Mantellidae FGMV 2002.925, 2001.E52 + - - + B. goudoti Mantellidae ZMA 19545, 19544 - - + + B. marojezensis Mantellidae FAZC 11467, 11483 - - + + B. microtympanum Mantellidae UADBA 21178, ZMA 19555 - + - + B.aff. ankaratra Mantellidae FAZC 11462, 11480 + + - + B. tephraeomystax Mantellidae UADBA 20564, 20623 - - - + B. septentrionalis Mantellidae FGMV 2002.915, 914 - - - + Mantella betsileo Mantellidae FGMV 2000F34, 2000D62 + + + + Mantidactylus asper Mantellidae UADBA 20665, ZMA 19484 + + + + M. brevipalmatus Mantellidae UADBA 21184, 21181 + + + + M.enki Mantellidae UADBA 20680, ZMA 19436 - - + + M. cf. femoralis Mantellidae UADBA 21096, ZMA 19470 (+) + + + M. grandidieri Mantellidae UADBA 20612, 20584 - - + + M. melanopleura Mantellidae no voucher - - + + M. tschenki Mantellidae UADBA 20641, ZMA 19392 + (+) - + M. ulcerosus Mantellidae FGMV 2002.449, 2002.546 + + + + M. sculpturatus Mantellidae UADBA 20605, 20617 + + + + M. aff. wittei Mantellidae FGMV 2002.880. 2002.879 - - + + Ptychadena mascariensis Ranidae no voucher (+) + + + Other vertebrates + Furcifer pardalis [Squamata] no voucher + + + + Mus musculus [Mammalia] no voucher - - (+) +

Appendices

Appendix 2

Voucher specimens and accession numbers of taxa studied. Localities and voucher specimens refer to sequences obtained in this study; some other sequences from GenBank refer to other conspecific individuals. Collection acronyms are as follows: Museum of Vertebrate Zoology, University of California at Berkeley, USA; UADBA, Université d'Antananarivo, Département de Biologie Animale, Madagascar, numbers being fieldnumbers of M. Vences of specimens deposited in UADBA; ZFMK, Zoologisches Forschungsinstitut und Museum A. Koenig, Bonn, Germany; ZSM, Zoologische Staatssammlung München, Germany. Accession numbers marked with an asterisk indicate sequences of congeneric species, except for Lyciasalamandra which we combined with a tyrosinase sequence of a different salamander genus, Hynobius and a BDNF sequence of Tylototriton, and Agalychnis, which was combined with a tyrosinase sequence of Phyllomedusa hypochondrionalis. Accession numbers to be added upon acceptance of manuscript are indicated by ‘#######’.

Species Family Locality, Voucher rag-1 rag-2 tyrosinase bdnf CO1 Anodonthyla boulengerii Microhylidae 2002.a46 ######## ######## ######## ######## ######## Anodonthyla montana Microhylidae 2001.c60 ######## ######## ######## ######## ######## Asterophrys turpicola Microhylidae Steve Richards, 5195 ######## ######## ######## ######## ######## Calluella guttulata Microhylidae Ubon province, Thailand, ZSM434/2002 ######## ######## ######## ######## ######## Chiasmocleis hudsoni Microhylidae st Eugene Ruan? Franc Guy? ######## ######## ######## ######## ######## Chiasmocleis shudikarensis Microhylidae 119MC ######## ######## ######## ######## ######## Cophixalus sp. Microhylidae Steve Richards, 2078 ######## ######## ######## ######## ######## Dermatonotus muelleri Microhylidae Paraguay ######## ######## ######## ######## ######## Dyscophus antongilii Microhylidae Maroantsetra, Madagascar (no voucher) ######## ######## ######## ######## ######## Dyscophus insularis Microhylidae 2001f39 ######## ######## ######## ######## ######## Elachistocleis ovalis Microhylidae 146MC ######## ######## ######## ######## ######## Gastrophryne carolinensis Microhylidae Pet trade (no voucher) ######## ######## ######## ######## ######## Glyphoglossus molossus Microhylidae Sagaing, Myanmar, CAS 210056 ######## ######## ######## ######## ######## Hamptophryne boliviana Microhylidae 42BM ######## ######## missing ######## missing Hoplophryne robustus Microhylidae kmh 12546 22-12-2000 missing ######## ######## ######## missing Hypopachus variolosus Microhylidae Nayarit, Mexico, MVZ 144018 ######## ######## ######## ######## ######## Kaloula pulchra Microhylidae Pet trade (no voucher) ######## ######## ######## ######## ######## Microhyla butleri Microhylidae Vinh Phu Province, Vietnam, MVZ 223728 ######## ######## missing ######## ######## Microhyla heymonsi Microhylidae Hainan province, China, MVZ 236751 ######## ######## ######## ######## ######## Microhyla pulchra Microhylidae Vinh Phu Province, Vietnam, MVZ 223797 ######## ######## ######## ######## ######## Micryletta sp. Microhylidae E14.4.E21.12 ######## ######## ######## ######## ######## Otophryne pyburni Microhylidae 1116MC ######## ######## ######## ######## ######## Paradoxophyla Microhylidae 2002.634 ######## ######## ######## ######## ######## Phrynomantis annectens Microhylidae Namibia ######## ######## ######## ######## ######## Phrynomantis bifasciatus Microhylidae Coast province, Kenya, MVZ 234047 ######## ######## ######## ######## ######## Platypelis grandis Microhylidae 2002f20/2002.1093 ######## ######## ######## ######## ######## Plethodontohyla alluaudi Microhylidae 2002g62/2002.1271 ######## ######## ######## ######## ######## Plethodontohyla brevipes Microhylidae 2002.193 ######## ######## ######## ######## ######## Ramanella sp. Microhylidae E26.1 ######## ######## ######## ######## ######## Rhombophryne testudo Microhylidae 2000b59 ######## ######## ######## ######## ######## Scaphiophryne calcarata Microhylidae Madagascar, ZSM 115/2002 ######## ######## ######## ######## ######## Stumpffia psologlossa Microhylidae 2001f42 ######## missing ######## ######## ######## Stumpffia pygmaea Microhylidae 2001f21 ######## ######## ######## ######## ######## Callulina kreffti Brevicipitidae Tanga region, Tanzania, MVZ 234046 ######## ######## ######## ######## ######## Breviceps fuscus Brevicipitidae Big Tree, South Africa, ZFMK 66716 ######## ######## ######## ######## ######## Breviceps mossambicus Brevicipitidae E2.3 ######## ######## ######## ######## ######## Hemisus marmoratus Hemisotidae Coast province, Kenya, MVZ 233793 AY364216 ######## ######## ######## Missing Hyperolius viridiflavus Hyperoliidae Barberton, South Africa, ZFMK 66726 AY323769 AY323789 AF249161 ######## Missing Arthroleptis variolosus Arthroleptidae Cameroon, Ennenbach ######## ######## AY341756 missing Missing Trichobatrachus robustus Astylosternidae Cameroon, Ennenbach ######## ######## AY844192 ######## Missing Leptopelis natalensis Hyperoliidae Mtunzini, South Africa, ZFMK 68785 ######## ######## AY341755 ######## Missing Mantidactylus wittei Mantellidae Madagascar, ZSM 405/2000 AY323774 AY323795 AY341751 ######## Missing Mantidactylus "comoros" Mantellidae Mayotte, ZSM 652/2000 AY323775 AY323794 AY341750 ######## Missing

Appendices

Species Family Locality, Voucher rag-1 rag-2 tyrosinase bdnf CO1 Rana (temporaria) Ranidae Voucher not collected AY323776 AY323803 AF249182 ######## Missing Agalychnis callidryas Hylidae Pet trade (no voucher) AY323765 AY323780 AY844153 ######## Missing Litoria caerulea Hylidae Pet trade (no voucher) AY323767 AY323793 AY844131 ######## Missing Alytes muletensis Bombinatoridae Mallorca, Spain (no voucher) AY323755 AY323781 AY341747 ######## Missing Alytes dickhilleni Bombinatoridae Parejo, Spain (no voucher) DQ019494 DQ019517 ######## ######## Missing Pipa parva Pipidae Pet trade (no voucher) AY874304 AY323799 AY341762 ######## Missing Xenopus sp. Pipidae GenBank AY874355 L19325 AY333967 BC082887 Missing Lyciasalamandra Salamandridae GenBank AY456261 AY323797 AY341765 AF497712 Missing Gallus gallus Aves: Phasianidae GenBank XM 421090 AY443150 D88349 XM 419645 Missing Mammalia: Homo sapiens GenBank AY130302 NM 000536 U01873 AF411339 Missing Hominidae Protopterus Dipnoi: Protopteridae GenBank AY442928 AF369086 missing U93369 Missing

Results produced by collaborators

Chapter 1: Parts of discussion by Miguel Vences (University of Braunschweig).

Chapter 2: Meike Thomas (University of Köln) carried out the molecular identification of tadpoles. Ylenia Chiari (University of Konstanz) and David Vieites (University of California, Berkeley) contributed part of the data. Figures produced by Miguel Vences (University of Braunschweig).

Chapter 3: Figure 3.1, table 3.3 and divergence time estimates produced by Miguel Vences (University of Braunschweig). Simone Hoegg (University of Konstanz) produced part of the data.

Chapter 4: Microtomography carried out and figure 4.3 produced by Renaud Boistel (University of Paris-Sud). Systematic revision by Annemarie Ohler (Museum National d’Histoire Naturelle, Paris). Parts of the introduction by Justin Gerlach (University Museum of Zoology, Cambridge).

Chapter 5: Simone Hoegg (University of Konstanz) produced part of the data.