MODERN MORPHOLOGICAL TECHNIQUES AND THE EVOLUTIONARY BIOLOGY AND TAXONOMY OF SEPSIDAE (DIPTERA) 1

19 CHAPTER 2: Phylogenetic origins and morphological evolution of the sepsid sternite appendage .... 8 List of Figures Figure 1.1: Phylogenetic hypothesis of Sepsidae used in this study

MODERN MORPHOLOGICAL TECHNIQUES AND

THE EVOLUTIONARY BIOLOGY AND

TAXONOMY OF SEPSIDAE (DIPTERA)

MODERN MORPHOLOGICAL TECHNIQUES AND

THE EVOLUTIONARY BIOLOGY AND

TAXONOMY OF SEPSIDAE (DIPTERA)

D ECLARATION

I hereby declare that this thesis is my original work and it has been written by

me in its entirety I have duly acknowledged all the sources of information

which have been used in the thesis

This thesis has also not been submitted for any degree in any university

previously

_

Yuchen Ang

16 August 2013

2

A CKNOWLEDGEMENTS

The first person I must thank would have to be my supervisor and long-time mentor, Prof Dr Rudolf Meier I have known The Teutonic Terror of S2 for close to nine years now, having been under his wing for both my B.Sc and current Ph.D In this time, I have benefitted so much, learning from him the ‘tools of the trade’, and his frank, critically constructive feedback has only helped to toughen my already-thick hide and more importantly develop my own set of critical thinking skills Through

numerous ‘coffee breaks’ and a lot of cajoling and nudging, I managed to get through

and tackle the great academic Ph.D beast Sometimes I feel undeserving of all the trust,

opportunities and help that he has given me, and ll this I am grateful for – especially his

contagious passion in research and sepsids that I have contracted as well Summarily I would liken him to the Gordon Ramsay of Science: doing research under him is tough and a little frightening because he demands so much, but ultimately, much more rewarding and inspiring as he imparts his knowledge and attitude that helps you achieve your own potential Thank you Prof, this time I’m writing it for you

Next of course are my labmates whom I’ve come a long way with: Amrita, coding sleuth and occasional stats-tutor whom Sepsidnet and I are so in debt with for all the prompt and late-night help – I owe you so much tea now! Denise ‘Little-Red’ Tan, old-bird Mammafly with all the crazy long work-fevered nights we’ve shared, I’m glad you’ve found a way to permanently damage your head in Florida ☺ (sorry I couldn’t make it to your mega-sendoff all because of this) Kathy Su, I’m glad you’ve come back to evolab – it’s good to have a Science-first author doctor around in the consultation room pantry that I can rely on to critically examine my work, and also listen to my yammerings Diego, for all the innuendo jokes and bouncing off interesting

ideas – it really energizes me to explore more in my work Ahlek Wong, I really appreciate all the help in the Chiro project rearing and stuff, sorry if I’m a little slow with the imaging Jay, always with cookies and chips, please make more of that yummy dessert you being to lab Lei, Mindy, Youguang, Nesibe, Ivy, Darren, plus all the others who have left the lab, all you guys and girls have made the lab the quirky and fun place

to work in!

Of course, I cannot forget my loved ones, who have made so many sacrifices; I really would not have made it here without your support Ma and Pa, thank you for all the support in so many different ways you have given me; all your expressed worries I

do consider carefully, so try not to worry Thank you for being so understanding about

my regular late nights and not expecting me to pay the bills even when I’m already 30 I’m going to get an income soon A job With CPF Promise! Mei, thanks for always offering to give me a lift home, both you and Di do brighten my day with your silly

antics and stupid 9GAG posts, specifically, “HEHEUHEHAHEAHUEH” Grace, I’m

sorry you’ve had to tolerate my crazy work-life schedule and thank you for all the care you’ve showered, all the little welfare packages of food and more snacks, coming all the way from the East to deliver hot food And especially for that final push where you just pushed me on so much more I’m sorry it had to come to this

4

Table of Contents

THESIS SUMMARY 7

List of Figures 8

List of tables, videos and appendices 10

List of publications used in this thesis 11

List of species used in this thesis 12

CHAPTER 1: General Introduction 16

A plea for morphology in biology 16

Evolutionary biology and taxonomy of Sepsidae 19

CHAPTER 2: Phylogenetic origins and morphological evolution of the sepsid sternite appendage 23

Abstract 23

Introduction 25

Part I: The evolutionary origins of the sternite appendages in the Sepsidae 27

Materials and Methods 27

Results and Discussion 29

Part II: Co-option of muscles and sternites for the formation of the sepsid sternite appendage 37

Materials and Methods 38

Results 39

Discussion 48

Appendix 1A 50

Appendix 1B 54

CHAPTER 3: Male sepsid legs evolve rapidly when in direct contact with

female structures 55

Abstract 55

Introduction 56

Materials and Methods 62

Results 66

Discussion 73

Appendix 2A 77

Appendix 2B 81

CHAPTER 4: A plea for digital reference collections and other science ‐based digitization initiatives in taxonomy: Sepsidnet as exemplar 82

Abstract 82

Introduction 83

Digital reference collections 86

Building the digital reference collection for Sepsidae 93

Conclusion 99

CHAPTER 5: Using seemingly unnecessary illustrations to improve the diagnostic usefulness of descriptions in taxonomy 101

Abstract 101

Introduction 103

Materials and Methods 106

Results 108

Appendix 3 127

6

CHAPTER 6: Using multiple data-sources to address taxonomic uncertainties

128

Abstract 128

Part I: From ‘cryptic species’ to integrative taxonomy – an iterative process involving DNA sequences, morphology, and behaviour leads to the resurrection of Sepsis pyrrhosoma (Sepsidae: Diptera) 129

Materials and Methods 132

Results 135

Discussion 145

Conclusion 148

Part II: Morphology and DNA sequences confirm the first Neotropical record for the Holarctic sepsid species Themira leachi (Meigen) (Diptera: Sepsidae)149 Materials and Methods 150

Results and Discussion 150

CHAPTER 7: Increasing taxonomic data accessibility by linking wiki-entries to species descriptions 155

Abstract 155

Introduction 156

Materials and Methods 159

Results 160

Key to species of the genus Perochaeta Duda, 1926 (males) 177

Discussion 178

THESIS CONCLUSIONS 181

References 185

THESIS SUMMARY

In my thesis, I pursue my research interests in morphology by conducting a series of studies on the evolutionary biology and taxonomy of Sepsidae (Diptera), using various bioimaging techniques such as microcomputered tomography,

photomicrography and scanning electron-microscopy First, I document how de-novo

moveable appendages evolved from male sternites and show that it evolved twice in sepsids Second, I demonstrate that sexual selection highly increases morphological divergence by quantifying and comparing the rates of evolution between sexually dimorphic structures that are likely to be under sexual selection and monomorphic structures that are mostly under viability selection Third, I explore the benefits of coupling taxonomy with information technology by creating a digital reference collection featuring 140 sepsid species and a web-tool for online species identification,

as well as publishing species descriptions that are simultaneously featured as taxonomic entries in a wiki-based platform Lastly, I show how morphological data-richness and iterative taxonomy can address inadequately diagnostic species descriptions as well as resolve 'cryptic' species proposed based on DNA sequences as well as disjunctive distributions in species

8

List of Figures

Figure 1.1: Phylogenetic hypothesis of Sepsidae used in this study for Chapter 2 31

Figure 1.2: Phylogenetic hypothesis of Sepsidae parsimony and maximum likelihood support that appendages evolved twice and were lost three times 32

Figure 1.3A: Summary of female sternites across Sepsidae 33

Figure 1.3B: Summary of male sternites across Sepsidae 34

Figure 1.4 Photomicrographs of ventral abdomen with expanded view of sternites 4 and 5, for Saltella sphondylli male, Themira lucida male, Themira superba male and Themira superba female, as well as the habitus of T superba male 41

Figures 1.5 – 1.7 (Part 1): Three dimensional models showing various views for Saltella sphondylli male, Themira lucida male and Themira superba male (Views A-B) 44

Figures 1.5 – 1.7 (Part 2): Three dimensional models showing various views for Saltella sphondylli male, Themira lucida male and Themira superba male (Views C-E) 45

Figure 1.8: Three dimensional models showing various views for Themira superba female 46

Figure 1.9: 3D model schematic of relevant structures in Themira superba explaining how VM5-6 powers the arm 'raise' 46

Figure 2.1: ‘Wing clasp’ behavior in various sepsid species 60

Figure 2.2: Various ‘basitarsal thumbs’ in Coelopidae 61

Figure 2.3: Obtaining complexity scores from leg structures 65

Figure 2.4 (part 1): Forefemora and –tibiae for male and female sepsids, mapped onto the sepsid phylogenetic hypothesis 67

Figure 2.4 (part 2): Midfemora and –tibiae for male and female sepsids, mapped onto the sepsid phylogenetic hypothesis 68

Figure 2.5: Femora and tibiae for fore- and midlegs of male and female coelopids, mapped onto the coelopid phylogenetic hypothesis 69

Figure 2.6: graphical representation of tree length for various leg structures of and regions of sepsid males and females for quantitative and qualitative measures 71

Figure 2.7: graphical representation of tree lengthfor various leg structures of and regions of coelopid males and females for quantitative and qualitative measures 71

Figure 3.1: Image of Nemopoda speiseri as seen in Zoomify™ viewer, showing the

habitus, lateral view (a); abdomen, ventral view (b); and dissected hypopygial capsule,

dorsal view (c) … 94

Figure 3.2: Image of male Sepsis cynipsea fore femur and tibia (partial), anterior view 95

Figure 3.3: A visual guide to using Sepsidnet 97

Figure 4.1: Key views and structures of Perochaeta orientalis, Male 109

Figure 4.2: Key views and structures of Perochaeta orientalis, Female 110

Figure 4.3: Additional views for Perochaeta orientalis, Male and Female 111

Figure 4.4: Images of holotype and drawing from description for Perochaeta orientalis, male 112

Figure 4.5: Copulatory profile for Perochaeta orientalis 119

Figure 4.6: Illustration of Archisepsis phallus, as reproduced from Eberhard and Huber (1998) 123

Figure 5.1: Consensus tree of Sepsis flavimana group 136

Figure 5.2: Male and female Sepsis pyrrhosoma 137

Figure 5.3: Morphology of Themira leachi from Cuba 152

Figures 6.1 - 6.14: Various Perochaeta structures 162

Figures 6.15 – 6.23: Various Sepsis forelegs and hypopygia 170

Figures 6.24 - 6.31: Male Sepsis spura 173

10

List of tables, videos and appendices

bootstrap trees (remaining columns) 35

Table 2.1: Tree lengths indicating amount of evolutionary change for mid and forelegs of sepsids and coelopids………… 70

Table 4.1: A summary of the pairwise distances between the COI of P orientalis with that of P cuirassa (KF199839), P dikowi (KF199840) and P lobo (KF199841) 122

Table 5.1: Uncorrected pairwise genetic distances between and within and between Sepsis flavimana and S pyrrhosoma morphotypes 139

Table 5.2: Qualitative comparison of behavioural elements observed in Sepsis flavimana and Sepsis pyrrhosoma (virgin) mating trials 140

Table 5.3: Results of the hybridization experiments in Sepsis flavimana and Sepsis pyrrhosoma 140

Video 1: Video montage for the various behaviours in Perochaeta orientalis 118

Appendix 1A 50

Appendix 1B 54

Appendix 2A 77

Appendix 2B 81

Appendix 3 127

List of publications used in this thesis

(Listed in order of appearance)

Bowsher JH, Ang YC, Ferderer T, Meier R (2012) Deciphering the evolutionary

history and developmental mechanisms of a complex sexual ornament: the abdominal

appendages of Sepsidae (Diptera) Evolution 67(4), 1069-1080… Chapter 2.1

Puniamoorthy N, Silva VC, Munari L, Meier R (2013) A plea for digital reference collections and other science‐based digitization initiatives in taxonomy: Sepsidnet as

exemplar Systematic Entomology 38(3), 637-644……… … Chapter 4

Perochaeta cuirassa sp n., Perochaeta lobo sp n., Sepsis spura sp n., Sepsis sepsi Ozerov, 2003 and Sepsis monostigma Thompson, 1869 ZooKeys, 70, 41-

56……… Chapter 5.1

Tan DS, Ang YC, Lim GS, Ismail MRB, Meier R (2010) From ‘cryptic species’ to

integrative taxonomy: an iterative process involving DNA sequences, morphology, and

behaviour leads to the resurrection of Sepsis pyrrhosoma (Sepsidae: Diptera) Zoologica Scripta 39, 51-61……… Chapter 7.1

Neotropical record for the Holarctic sepsid species Themira leachi (Meigen) (Diptera:

Sepsidae) Zootaxa 1933, 63-6……… Chapter 7.2

12

List of species used in this thesis

Sepsidae (Diptera, Acalyptrate)

41 Leptomerosepsis simplicicrus (Duda 1926) 3,4

14

133 Themira minor (Haliday 1833) 2,4

Outgroups (Coelopidae, Helcomyzidae, Heterocheilidae, Ropalomeridae)

Coelopidae

16

CHAPTER 1

General Introduction

A plea for morphology in biology

Morphology, the study of form and structure of an organism, is a fundamental subject of biology Changes in morphological structures are an expression of the biological and evolutionary processes that act upon organisms, and the comparative study of morphology allows biologists to investigate these processes Morphological data are also significant for taxonomy and systematics because the variation within such traits can be used to delimit organsimal groups; i.e., morphology is an important source of data for the diagnosis and circumscription of species and monophyletic taxa Arguably, even molecular biologists are morphologists studying the structure and qualities of biological molecules such as DNA and proteins As such, many biological sciences have their origins in morphology, which remains important for many disciplines

The study of morphology has become particularly exciting in recent years This

is mainly due to advancements in bioimaging techniques such as high-resolution photomicrography, scanning electron microscopy and micro-computered tomography These techniques have become cheaper and hence more prevalent in use, allowing present-day morphologists to access both the minute and internal details of organisms

much more easily than in the 20th and early 21st centuries (Faulwetter et al 2013) These bioimaging techniques can capture a great deal of information in just a few

‘snapshots’ (Cranston 2005), which greatly speeds up the traditionally time-consuming process of morphological documentation that is based on text and drawings At the same time, some techniques allow for non-invasive study of internal morphology (Heethoff et al 2008, Schneeberg et al 2012), which can even be applied to living specimens (Lowe et al 2013)

The dissemination of morphological information via the internet has also been a boon to morphological studies (Godfray 2002) Because it is mostly image-based, information in morphology can be lost when converted into the kind of text-based or line-drawn formats traditionally used by printed scientific journals In the past, scientific journals had tight page limitations and high costs for printing images but these issues have largely disappeared given that electronic journals have fewer restrictions Combined with the internet’s efficacy in storing and transmitting images, advancements in information technologies have made it much easier to publish and share morphological data, facilitating the research of morphologists and taxonomists alike

Another important development relevant to morphology is the recent advancements in DNA sequencing, which have provided biologists with a relatively cheap and bountiful source of information (Vogler and Monaghan 2007, Meier 2008) While this has led to some researchers calling for the de-emphasizing of morphology in systematics trees (Scotland et al 2003) and taxonomy (Tautz et al 2003, Blaxter 2004, Cook et al 2010), many others have been exploring the integration of these two types

18

of data (Dayrat 2005, Padial et al 2009, Schlick-Steiner et al 2010) There are many on-going discussions over how to properly integrate morphology and molecular data, especially when they yield conflicting results (Wiens and Hollingsworth 2000, Wahlberg and Nylin 2003, Su et al 2008), but such conflict is not necessarily harmful because molecular data can bring fresh perspectives to morphology and vice-versa For example, DNA sequence data can highlight ‘cryptic’ species that morphology may have overlooked initially (Bickford et al 2007) This allows taxonomists to re-evaluate and improve their understanding of the species limits in their study-taxa (Tan et al

2010, Neusser et al 2011) Such iterative methods in taxonomy (more commonly called ‘integrative’ taxonomy) have now become quite established in the literature (Tan

et al 2010, Jansen et al 2011, Ceccarelli et al 2012, Morrow et al 2013) However, arguably the most interesting development for morphology is the use of next generation sequencing (NGS) data because it will lead to the inference of robust phylogenies (Guschanski et al 2013, McCormack et al 2013, Yi and Jin 2013) This allows for the study of morphological change without the need to use morphology for tree reconstruction Thus, rather than viewing this 'golden age' of molecular biology as a threat to the morphological discipline, DNA data may become a boon to morphologists because it allows them to concentrate on the study of morphology and morphological evolution

Of course, even in the context of molecular studies, the morphological perspective remains important as a guard against inaccuracies in molecular inferences which can arise from ‘long branch attraction’ (Felsenstein 1978, Huelsenbeck 1997), divergent evolution between genes and species (Maddison 1997) or even mundane problems such as contamination and misidentification Obvious mistakes can be caught

by morphologists who can highlight unusual conflict between the data Morphology also remains important for specimens for which DNA is not available [e.g., in fossil taxa (Gauthier et al 1988, Smith and Turner 2005)], or degraded [e.g., in old and improperly preserved museum specimens (De Giorgi et al 1994, Vink 2005, Zimmermann et al 2008, Miller et al 2013, Osmundson et al 2013)] One must also remember that many other biological fields such as sexual selection studies cannot be approached based on DNA data alone, because selection acts directly on morphology and only indirectly on the underlying genetic bases of the morphological features that are under sexual selection Questions in evolutionary biology such as how structures

respond, change or are even created de-novo must thus have a morphological

component Such studies simply do not deal with DNA data alone, and require morphology as a basis for study Thus, my thesis will thus utilize morphology to address the questions and studies evolutionary biology

Evolutionary biology and taxonomy of Sepsidae

In my thesis, I use the different morphological techniques described above (i.e., high-resolution photomicrography, scanning electron microscopy and micro-computered tomography) to document the morphology of the dung-fly family Sepsidae

(Diptera, Acalyptratae) Sepsidae are a moderately sized (ca 350 described species)

family of saprophagous flies that have been recorded across all zoogeographic regions (Ozerov 2005) They have been a model in different fields, ranging from sexual selection studies (Blanckenhorn et al 2000, Kraushaar and Blanckenhorn 2002, Teuschl et al 2010b, Puniamoorthy et al 2012) to developmental biology (Bowsher and Nijhout 2007, Bowsher and Nijhout 2009, Bowsher et al 2013), and genomics

In Chapter 2, I investigate the evolutionary origins of an articulated, moveable appendage that has been derived from the immovable, "boring" sternite plates of male

sepsids This appendage is a nascent de-novo structure that is found exclusively in the

Sepsidae I first imaged the external morphology for 70 species using photomicrography and used the data to show that sternite appendages evolved once, was lost three times and then reacquired once again across the sepsid phylogeny These

results have been published as part of a research article in Evolution Next, I used

synchrotron radiation-based microcomputered tomography (SRµCT) and slicing to reconstruct the internal morphology of males and females for thirteen species representing eight genera where males have sternite appendages However, due to time constraints I selected four specimens representing the different transitionary states of the sternite appendage From these specimens I generated three-dimensional (3D) models to examine the internal structures The focus is on how existing structures (sclerites and muscles) have been co-opted to form a new appendage

Chapter 3 investigates the role of sexual selection in the evolution of morphological structures by quantifying the rate of evolution in male sepsid forelegs, which are under sexual selection, and comparing them with homologous structures that are only under viability selection; i.e., legs from female sepsids and an outgroup family

In this study, I coded discrete morphological characters based on photoimages of leg structures to obtain a qualitative dataset, but I also employed a quantitative method to measure shape complexity, which is not easily captured in discrete characters

Chapters 4 and 5 demonstrate the benefits of coupling taxonomy with information technology In Chapter 4, I propose the establishment of digital reference collections based on large amounts of image data Here I use photomicrography to document the morphology of >140 species of Sepsidae (= 40% specific, 80% generic diversity coverage) and making the images available online on the taxonomic database,

Sepsidnet The images are presented in a ‘zoomable’ format that allows the user to both look at an entire specimen and zoom in to examine finer details The websites also provides a species comparator feature that allows for the inspection of up to six species

simultaneously This study is published in Systematic Entomology and the webtools

have been used to generate similar digital reference collections for (1) the immature stages of Singapore's Odonata (23 species), (2) adult male Coelopidae (Diptera) primarily from the Oceanic region (20 species) and (3) >50 species of Chironomidae (Diptera) from Singapore's reservoirs (unpublished) In Chapter 5, I describe three new

sepsid species P cuirassa, P lobo and S spura along with presenting two new geographic records for S monostigma and S sepsi in Vietnam The new species were published in Zookeys and also presented as individual wiki entries in the taxonomic wiki platform, Species-ID

22

In the remaining chapters, I show how new morphological techniques and an abundance of morphological data can address some issues in taxonomy Here, I use three sepsid species as examples: Chapter 6 explores how including a large number of images and illustrations in a species description can help protect it against future

diagnostic inadequacies I redescribe Perochaeta orientalis, a species that has

previously suffered from two descriptions that are unreliable because they lacked a sufficiently large number of images In Chapter 7, I use morphology in conjunction with other data sources such as DNA sequences, mating behaviour and reproductive

isolation experiments to resurrect and re-describe a putatively ‘cryptic’ species Sepsis pyrrhosoma and to investigate a peculiar disjunctive distribution of a Neotropical

specimen of the otherwise Holarctic species, Themira leachi These two studies are published in Zoologica Scripta and Zootaxa respectively

Finally, I summarize in the concluding chapter how advancements in morphology and taxonomy can be made based on available technologies in bioimaging and information technology I believe that these advancements will help with tackling some of the issues that are faced by morphologists and taxonomists

CHAPTER 2

Phylogenetic history and morphological evolution of

the sepsid sternite appendage

Abstract

In this chapter, I investigated two aspects of the evolutionary origins of a novel like appendage that is derived from male sternites in Sepsidae First, I determined where the sternite appendage evolved in sepsid phylogeny: I coded the presence of the sternite appendage based on photomicrography to document sternite morphology, and optimized the trait onto the sepsid phylogeny I demonstrated that they are likely to have evolved once, were lost three times, and then secondarily gained in one lineage I

leg-contributed these data to a multi-authored paper that has been published in Evolution

Second, I documented the morphological origins of the appendage by investigating the internal morphology of the appendage across representative species using 3D models generated based on SRµCT and microtome-sections Deciphering the origin of this complex appendage is here facilitated through the existence of species with intermediate structures that range from simple flat plates to articulated, moveable appendages I demonstrated that the appendage is directly derived from sternite 4 with the associated abdominal musculature having been co-opted for moving the appendage Sternite appendage evolution is a two-stage process: first, moveable lobes bud from the

24 posterior corner of the sternite Then the sternite lengthens and protrudes from the abdomen as paired, moveable lobes, resulting in the sepsid sternite appendage

Introduction

As mentioned earlier in Chapter 1, males in many species of Sepsidae exhibit strong, sexually dimorphic ornamentation, which is known to be associated with mating behaviour (Eberhard 2001b, 2002b, Puniamoorthy et al 2008, Puniamoorthy et al 2009) One body part with such dimorphism in some species is the abdomen, and in particular sternite 4 While female sternites 2-4 are fairly similar, male sternites 4 are highly variable within the family, often being modified in size, shape and chaetotaxy (Eberhard 2001d) In some species such modifications have been elaborated into a pair

of articulated, moveable appendages (hereafter, sternite appendages) (Hennig 1949, Eberhard 2001d), which resemble typical arthropod legs because of their capacity for

controlled movement and rotation of over 180° (Bowsher and Nijhout 2007, Bowsher

and Nijhout 2009) These particular appendages are unique to Sepsidae and absent in closely related taxa (Eberhard 2001d) Within the Sepsidae, the presence of sternite appendages are known in a number of genera that are not closely related based on the current understanding of sepsid phylogeny (see Fig 1.2) This chapter is split into two parts: Part I investigates the phylogenetic history of the appendage, and when did it arise in the Sepsidae Part II shows that sternite appendages can come in various forms and intermediacies, and will address the transition from a flat sternite plate to a fully fledged sternite appendage

Based on behavioural observations, sternite appendages are primarily used in interacting with the female abdomen during mating (Eberhard 2003, Puniamoorthy et

al 2009), which suggests that sexual selection is ultimately responsible for the evolution of these novel and complex sexual ornaments Ultimately, the origin of the sternite appendages is likely related to the evolution of a new mounting pattern and

26

mating behaviour in Sepsidae, where the male climbs/jumps onto the female and uses his (modified) forelegs to clasp onto her wing (Puniamoorthy et al 2008) This mounting position also brings his sternite 4 in close contact to the female, which allows him to use it to interact with the female, and thus becoming subject to sexual selection

In this chapter, I decipher the evolutionary history of the sternite appendages in Sepsidae First, I perform a phylogenetic analysis to determine where the sternite appendage evolved on the sepsid tree Second, I document the morphological origins of the appendage by investigating the internal morphology of the appendage across representative species

Part I: The evolutionary history of the sternite appendages in

II of this chapter

Materials and Methods

Taxon sampling and phylogenetic reconstruction. The phylogenetic reconstruction was a result of collaboration with the Bowsher laboratory (North Dakota State

University) Here, molecular data for nine genes (mitochondrial: 12S, 16S, COI, COII,

evolution This study was contributed to the phylogenetic portion in the paper “Bowsher JH, Ang YC, Ferderer T, Meier R (2012) Deciphering the evolutionary history and developmental mechanisms of a

complex sexual ornament: the abdominal appendages of Sepsidae (Diptera)” in Evolution 67(4),

1069-1080

28

Cytb and nuclear: 18S, 28S, AATS, Ef1a) for 70 sepsid species and six outgroup species

from three families were used to reconstruct the tree Many of these data were derived

from Su et al (Su et al 2008), with the addition of one new species M fasciculata,

based on molecular protocols as specified in Su et al (2008)

Documenting external morphology and scoring of the appendage character The external morphology of the male sternite 4 was imaged using the Visionary Digital BK Lab Plus System (at CF4P3 magnification) for 67 sepsid species and three outgroup species from three families Females from representative taxa were also imaged I scored all sternites as appendages if they consisted of a pair of articulated, moveable lobes on the sternite 4 that terminate in a fascicle of bristles While this constitutes a composite character, all moveable lobes inclusively bear bristles, and the latter feature

is useful for determining the exact location for lobes that are not fully sclerotized (e.g

see sternite of Microsepsis armillata in Fig 1.3B) The presence or absence of sternite appendages was scored for all 70 species Three sepsids (Saltella sp., Saltella nigripes and Susanomira caucasica) and two outgroup species from Coelopidae (Lopa convexa and Icaridion debile) lack photoimages as specimens were not available for imaging,

but I checked specimens and they lack sternite appendages

However, based on Pont and Meier (2002), the sternites in Saltella are not known

to bear appendages, while S cuacasica has the sternite 4 clearly modified into sternite

appendages Coelopids are also not documented to possess sternite appendages

(McAlpine 1991), and personal observation of specimens of L convexa and I debile

did not reveal the presence of any sternite appendages

Mapping the evolution of sternite appendages In order to reconstruct the evolution

of sternite appendages, all species were scored according the presence or absence and mapped onto the optimal tree and optimised using parsimony and likelihood (Mk1 model) as implemented in Mesquite (Maddison and Maddison 2000) For the parsimony optimization, I also used step matrices to test whether higher costs for gains versus losses yield different results For the likelihood reconstructions, I used proportional likelihoods (Mesquite’s default decision threshold =2.0) to test whether the state assignments with lower likelihoods could be rejected In addition to mapping the character onto the optimal ML tree, I also tested whether the reconstruction of character history is robust to changes in tree topology For this purpose I mapped the appendage character onto all 250 bootstrap trees that were derived from 250 resampled character matrices (Mesquite: “Trace Character Over Trees” “Count Trees with Uniquely Best States” for parsimony; Mk1 model for maximum likelihood) For each node, the frequency of character state assignments was determined for all nodes on the

250 trees that were congruent between the optimal tree and the bootstrap trees

Results and Discussion

The phylogenetic hypothesis for Sepsidae used in this study is shown with bootstrap support values in Fig 1.1 All nodes on this tree have bootstrap support >50 and a large proportion of the nodes have high support of > 90 (74%) The mapping of the sternite appendage trait is shown in Fig 1.2 Images of the male sternites 4 for all available species are shown on the tree with the sternite appendage trait mapped on in Appendix 1A; a summary of female and male sternite 4 (and associated sternite 5) diversity is

30

shown on a collapsed tree form (sternite appendage trait also mapped on) in Figs 1.3A and 1.3B respectively: Female sepsids generally have simple, rounded rectangular to oval-shaped sternites with no evidence for a sternite appendage (Fig 1.3A) However,

in males, sternite 4, which bears the appendage, is highly variable in size, structure and/or shape (Fig 1.3B) The associated sternite 5 is also highly variable, and in many cases, reduced to a crescent-shaped sclerite or even missing entirely

Referring to Fig 1.2 (see also Appendix 1A), Mesquite’s parsimony mapping of the abdominal appendage character reveals a single most parsimonious optimization in which the appendages have evolved once, were lost three times, and reacquired once

This is illustrated in Fig 1.3B: Susanomira (E), Zuskamira (F), Nemopoda (I), Themira (J), Microsepsis (H) and Meroplius (K) have homologous abdominal appendages, while they were lost independently in Saltella (D), Archisepsis (G) and a clade branching off after Meroplius The appendage was subsequently regained in Perochaeta (O) The

same optimization implying one origin, three losses and one reacquisition is also found

in a maximum likelihood reconstruction Based on proportional likelihoods the state assignments for 67 of the 74 internal nodes of the tree have significantly higher likelihoods than the alternative state For the remaining seven nodes the proportional likelihood of the preferred state is at least three times higher than for the alternative state (see Table 1.1) To further test the robustness of the results, step matrices were used to determine at which cost assignment for gains versus losses the results of the parsimony reconstruction changes: The reconstruction remains identical as long as the

cost for gains is <2 times the cost of losses At a twofold cost, there is a second

optimization (one origin and five losses) and at higher costs for gains this new optimization is preferred

Figure 1.1: Phylogenetic hypothesis of Sepsidae used in this study Numbers denote bootstrap support

values at nodes

32

Figure 1.2: Parsimony and maximum likelihood support that appendages evolved twice (black rectangles) and were

lost three times (white rectangles) Node circles indicate the proportion of bootstrap tree supporting the assignment of the presence (black) or absence (white) states based on parsimony (grey= proportion of bootstrap trees without node); Node numbers refer to the numbers in Table 1.1

Table 1.1: State optimizations for nodes on ML tree (Proportional likelihoods) and bootstrap trees

(remaining columns) Only those nodes are listed where the proportional likelihoods are not significant and/or character state assignments differ between bootstrap trees; * = preferred state has significantly higher likelihood

abdominal brushes in Perochaeta is on the vast majority of the bootstrap trees

optimized to lack appendages (see Table 1.1), which implies that the appendages have

been reacquired in Perochaeta

Sepsid flies present a rare opportunity to investigate the evolution of the same complex structure in different lineages Phylogenetic analysis reveals that the following scenario for the evolution of abdominal appendages is currently best supported by different optimization techniques and on different trees: the appendages initially

36

evolved near the base of the sepsid tree and were then lost three times independently

before being reacquired in Perochaeta This secondary gain of abdominal appendages

had not been detected previously because this genus had not been included in the cladistic analyses by Su et al (2008)

Part II: Co-option of muscles and sternites for the formation of the sepsid sternite appendage

Deciphering the origins of evolutionary novelties is often challenging due to a lack

of understanding for the events that led to the new structures; this is frequently due to a dearth of transitional forms available for study Fortunately, this is not the case for the sepsid sternite appendage for which intermediate forms are known I here focus

on Themira and Saltella to investigate the evolution of the sternite from an immobile

plate to an articulated, moveable appendage Three distinct morphologies can be identified: the ‘unmodified’ state – a simple, immobile flat plate (Fig 1.4A), the

‘intermediate’ state – a flat plate with paired moveable lobes on its lateral periphery (Fig 1.4B) and the ‘developed’ state – a pair of protruding, moveable arms with

moveable lobes (‘brushes’) on its apices (Fig 1.4C) Themira species possess a diverse

array of ‘intermediate’ and ‘developed’ sternite appendages (See Appendix 1B; Figure 1.2, ‘J’)

For an in-depth study with SR-µCT, Themira lucida was selected to represent the

‘intermediate’ and Themira superba, the ‘developed’ state Saltella sphondylii, an group species to the Themira clade, was selected to represent the ‘unmodified’ state

out-Because female sternites remain simple and largely unchanged throughout sepsid

evolution, a female T superba was also used as an additional representation of the

‘unmodified’ state Using data from synchrotron-based tomography SR-µCT and microtome slices, I generated three-dimensional (3D) models of the sternite structures

38

for the four specimens and documented their external and internal structures to determine the morphological changes that led to the origin of the sternite appendage

Materials and Methods

Taxon Sampling Four specimens from three species were sampled: A male and

female Themira superba, a male T lucida and a male S sphondylii All specimens

were taken from laboratory cultures established from wild-caught females They were then killed in 50% EtOH to prevent dehydration shrinkage and the abdomen was removed The sternite appendage is directly derived from sternite 4, but sternites 5 and

6 are also involved in the movement of the appendage Thus, I used SR-µCT and

microtome-sections to analyze sternites 4–6 and their musculature

SR-µCT. Specimens were dehydrated with an alcohol series, critical-point dried with CO2 (EmiTech K850 Critical Point Dryer) and mounted on a metal rod Scanning was performed at Beamline BW2 of the German Electron Synchrotron Facility (DESY, Hamburg) using a stable low photon energy beam (8kV) and absorption contrast The resulting image stacks had a very high density resolution which allowed for distinguishing sclerites, membranes, and muscles [see Beckmann et al (2006)]

Microtome-sectioning The abdomens were fixed in FAE (Formaldehyde-Acetic Acid-Ethanol; 3:1:6) and embedded in Araldit CY 212® (Agar Scientific, Stansted/Essex, England) for sectioning A 1µm cross-section series was carried out with a HM 360 (Microm, Walldorf, Germany) microtome, stained with Toluidine blue and Pyronin G (Waldeck GmbH & Co KG/Division Chroma, Münster, Germany), and