Effect of Host-Plant Use on Speciation and Parasitoid Community Structure

dissertations

Sanna Leppänen

Effect of host-plant use on | Leppänen | 154 | Sanna speciation and parasitoid community structure in internal-feeding sawflies

Effect of host-plant use on speciation and parasitoid community structure in internal-feeding sawflies in internal-feeding structure community parasitoid and of host-plant on speciation use Effect Sanna Leppänen Effect of host-plant use on speciation and parasitoid community structure in internal-feeding sawflies

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences isbn: 978-952-61-1560-3 (printed) issnl: 1798-5668 issn: 1798-5668 isbn 978-952-61-1561-0 (pdf) issnl: 1798-5668 issn: 1798-5676

SANNA LEPPÄNEN

Effect of host-plant use on speciation and parasitoid community structure in internal-feeding sawflies

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences 154

Academic Dissertation To be presented by permission of the Faculty of Science and Forestry for public examination in the Auditorium N100 in Natura Building at the University of Eastern Finland, Joensuu, on September, 19, 2014, at 12 o’clock.

Department of Biology

Author’s address: University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

Supervisors: Associate professor Tommi Nyman, Ph.D. University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

Professor Heikki Roininen, Ph.D. University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

Reviewers: Docent Niklas Janz, Ph.D Stockholm University Department of Zoology 106 91 STOCKHOLM SWEDEN Grano email: [email protected] Joensuu, 2014 Research scientist Carlos Lopez-Vaamonde, Ph.D Editors: Prof. Pertti Pasanen, Institut National de la Recherche Aqronomique (INRA) Prof. Pekka Kilpeläinen, Prof. Kai Peiponen, Prof. Matti Vornanen Zoologie Forestière 45075 ORLÉANS Distribution: FRANCE Eastern Finland University Library / Sales of publications email: [email protected] P.O.Box 107, FI-80101 Joensuu, Finland Opponent: Professor Niklas Wahlberg, Ph.D tel. +358-50-3058396 University of Turku http://www.uef.fi/kirjasto Department of Biology 20014 TURKU FINLAND ISBN: 978-952-61-1560-3 (printed) email: [email protected] ISSNL: 1798-5668 ISSN: 1798-5668

ISBN 978-952-61-1561-0 (PDF) ISSNL: 1798-5668 ISSN: 1798-5676

Author’s address: University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

Supervisors: Associate professor Tommi Nyman, Ph.D. University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

Professor Heikki Roininen, Ph.D. University of Eastern Finland Department of Biology P.O.Box 111 80101 JOENSUU FINLAND email: [email protected]

ISBN 978-952-61-1561-0 (PDF) ISSNL: 1798-5668 ISSN: 1798-5676

ABSTRACT differentiation has a huge impact on galler speciation, but also that host-associated differentiation is not completely repeatable:

Plants, herbivorous insects, and parasitoids form a major part of the two galler genera have similar host plant repertoires, but ΦST all described species. Still, the evolutionary processes that have estimates among population samples collected from specific led to the formation of these diverse food webs with extremely willow species differ markedly. complex multitrophic interactions are not fully understood. Parasitoids of both leafminers and gallers represent three Insect–plant interactions are taxonomically conserved in the hymenopteran families that are phylogenetically distant from sense that herbivores tend to use closely related host plants, but each other. In the case of miner parasitoids, closely related host–plant shifts are common in most insect taxa. There is parasitoids tend to use the same miner species. However, evidence that shifts among host plants can influence insect phylogeny-based analyses show that the host-plant phylogeny speciation, but the effects of niche shifts on higher trophic levels is a more important factor structuring parasitoid–miner remain a mystery. In addition, the third trophic level (especially associations than is the miner phylogeny. The diet breadth of parasitoids) could be the driving force behind insect individual parasitoid species varies, and both niche- and diversification by causing insects to shift host plants. Using species-specialist enemies can be found. In combination, the molecular-genetic methods, this thesis tries to answer how host tritrophic patterns suggest that both bottom–up and top–down plants influence speciation in leaf-mining and gall-inducing forces have influenced diversification in the plant–leafminer– sawflies and in their associated parasitoids. My main purpose parasitoid system. was to investigate the determinants of tritrophic associations, DNA barcoding proved to be a suitable tool for studying the and the role of parasitoids in herbivore speciation. structure of leaf-galler parasitoid communities. In galler A molecular-phylogenetic analysis of leaf-mining sawflies in parasitoids, habitat (boreal–subarctic vs. arctic–alpine) had the the subfamily Heterarthrinae indicates that the subfamily strongest effect on parasitoid community structure. This originated c. 100 million years ago, which means that the indicates that, in addition to host shifts, colonization of novel leafminers are younger than their main host taxa. habitats may provide shelter from enemies and accelerate Reconstructions of larval feeding modes show that transitions herbivore speciation. Furthermore, at least two parasitoid from external to internal feeding have occurred multiple times species apparently originated by tracking galler hosts that had within the superfamily Tenthredinoidea, but also that leaf- shifted to using dwarf willows that grow in treeless arctic– mining has arisen only twice. On long time scales, host–plant alpine habitats. use is taxonomically unstable, and multiple convergent host In combination, the results of the leafminer and galler studies shifts have happened during the evolutionary history of support the view that parasitoids could be a major driving force Heterarthrinae. in herbivore speciation, but also that diversification can cascade Host-plant shifts also seem to promote genetic differentiation up in the food web, when parasitoids form host races attacking and speciation in gall-inducing sawflies belonging to the genera herbivores on novel plant species and niches. Pontania and Euura, as clear host-based clustering of individuals is observable in phylogenetic trees. Indeed, hierarchical Universal Decimal Classification: 575.858, 575.86, 591.522, 591.531.1, AMOVAs indicated that most of the genetic variation in both 591.557.8, 595.793 galler genera is explained by willow host species rather than by geographic location. The results show that host-associated CAB Thesaurus: speciation; coevolution; host plants; herbivores; insects; Hymenoptera; leaf miners; Pontania; Euura; Pseudodineura; differentiation; Preface parasitoids; food webs; tritrophic levels; diversification; habitats; phylogeny; genetic variation; community ecology; DNA; nucleotide sequences; molecular genetics techniques Firstly, the deepest imaginable gratitude goes to my main Yleinen suomalainen asiasanasto: lajiutuminen; evoluutioekologia; supervisor Tommi Nyman for guiding me on every step and isäntäkasvit; kasvinsyöjät; hyönteiset; sahapistiäiset; parasitoidit; showing me how marvelous creatures sawflies and their erilaistuminen; ravintoverkot; habitaatti; fylogenia; geneettinen muuntelu; parasitoids are. I also thank professor Heikki Roininen for DNA-anlyysi; molekyyligenetiikka encouraging me to begin this project, and all of my coauthors for their patience and the help they provided. I am very grateful for my support-group members Tomas Roslin and Marko Mutanen for their comments and help. I also want to show gratitude for all the sample collectors: without you my work would not have been possible. To our small research group: Mia, you always seemed to have an answer to all of my problems, and Tobias, you helped me in the lab and with analysis whenever I needed. I wish to thank Inna for all coffee breaks we had. Big thanks also belong to the staff and faculty of the Department of Biology for their help and for providing resources for my research. I wish to thank the institutions funding this research project: Academy of Finland (Project No. 124695 and 14868), the Finnish Cultural Foundation North Karelia Regional Fund, and the Department of Biology. Warmest thanks to all my friends outside the university life, especially the climbing community in Joensuu, without you I would not have survived this project. You offered me a break from the academic world, and for once I had to focus on just one thing. Finally, I would like to thank my family: mom, dad, and my sister for their support and having faith in me. Last but not least, to the love of my life Eero and my son, thanks for bringing meaning to my life. There are not enough words to say how grateful I am. You have always stood by me, supported and pushed me on when I wanted to quit.

Joensuu, August 2014 Sanna Leppänen CAB Thesaurus: speciation; coevolution; host plants; herbivores; insects; Hymenoptera; leaf miners; Pontania; Euura; Pseudodineura; differentiation; Preface parasitoids; food webs; tritrophic levels; diversification; habitats; phylogeny; genetic variation; community ecology; DNA; nucleotide sequences; molecular genetics techniques Firstly, the deepest imaginable gratitude goes to my main Yleinen suomalainen asiasanasto: lajiutuminen; evoluutioekologia; supervisor Tommi Nyman for guiding me on every step and isäntäkasvit; kasvinsyöjät; hyönteiset; sahapistiäiset; parasitoidit; showing me how marvelous creatures sawflies and their erilaistuminen; ravintoverkot; habitaatti; fylogenia; geneettinen muuntelu; parasitoids are. I also thank professor Heikki Roininen for DNA-anlyysi; molekyyligenetiikka encouraging me to begin this project, and all of my coauthors for their patience and the help they provided. I am very grateful for my support-group members Tomas Roslin and Marko Mutanen for their comments and help. I also want to show gratitude for all the sample collectors: without you my work would not have been possible. To our small research group: Mia, you always seemed to have an answer to all of my problems, and Tobias, you helped me in the lab and with analysis whenever I needed. I wish to thank Inna for all coffee breaks we had. Big thanks also belong to the staff and faculty of the Department of Biology for their help and for providing resources for my research. I wish to thank the institutions funding this research project: Academy of Finland (Project No. 124695 and 14868), the Finnish Cultural Foundation North Karelia Regional Fund, and the Department of Biology. Warmest thanks to all my friends outside the university life, especially the climbing community in Joensuu, without you I would not have survived this project. You offered me a break from the academic world, and for once I had to focus on just one thing. Finally, I would like to thank my family: mom, dad, and my sister for their support and having faith in me. Last but not least, to the love of my life Eero and my son, thanks for bringing meaning to my life. There are not enough words to say how grateful I am. You have always stood by me, supported and pushed me on when I wanted to quit.

Joensuu, August 2014 Sanna Leppänen

LIST OF ABBREVIATIONS LIST OF ORIGINAL PUBLICATIONS

HAD Host-associated differentiation This thesis is based on data presented in the following articles, MP Maximum parsimony referred to by the Roman numerals I-IV. ML Maximum likelihood NJ Neighbor-joining AMOVA Analysis of molecular variance I Leppänen S A, Altenhofer E, Liston A D, and Nyman T. COI Cytochrome oxidase I Phylogenetics and evolution of host-plant use in leaf-mining Cytb Cytochrome b sawflies (Hymenoptera: Tenthredinidae: Heterarthrinae). EF-1α Elongation factor-1α Molecular Phylogenetics and Evolution 64: 331–341, 2012. NAK Sodium-potassium adenosine triphosphatase II Leppänen S A, Malm T, Värri K, and Nyman T. A comparative analysis of genetic differentiation across six shared willow host species in leaf- and bud-galling sawflies. Submitted.

III Leppänen S A, Altenhofer E, Liston A D, and Nyman T. Ecological versus phylogenetic determinants of trophic associations in a plant–leafminer–parasitoid food web. Evolution 67:1493–1502, 2013.

IV Nyman T, Leppänen S A, Várkonyi G, Shaw M R, Koivisto R, Vikberg V, and Roininen H. Determinants of parasitoid communities of willow-galling sawflies: habitat overrides phylogeny, host plant, and space. Manuscript.

Some unpublished results are also presented.

Publications are reprinted with the kind permission of John wiley and sons and Elsevier.

LIST OF ABBREVIATIONS LIST OF ORIGINAL PUBLICATIONS

Some unpublished results are also presented.

Publications are reprinted with the kind permission of John wiley and sons and Elsevier. AUTHOR’S CONTRIBUTION

The present author together with her supervisor T. Nyman Contents planned all the studies. The present author was responsible for most of the laboratory work and data analyses, and was the main writer in articles I, II and III. The present author collected the data and participated in data analysis and writing of article IV.

1 Introduction ...... 13 1.1 Plant–insect interactions ...... 14 1.2 Insect–parasitoid interactions...... 16 1.3 Aims of this study...... 18

2 Study system ...... 19 2.1 Sawflies...... 19 2.1.1 Leafminers...... 19 2.1.2 Gall-inducers ...... 21 2.2 Parasitoids ...... 22

3 Materials and Methods ...... 25 3.1 Material sampling ...... 25 3.2 Genetic material...... 26 3.3 Phylogeny reconstruction ...... 27 3.4 Insect–plant analyses ...... 29 3.5 Analyses of tritrophic interactions ...... 31

4 Results and Discussion...... 33 4.1 Role of plants in insect speciation...... 33 4.1.1 Leaf-mining sawflies ...... 33 4.2.1 Gall-inducing sawflies ...... 35 4.2 Tritrophic interactions...... 37 4.2.2 The parasitoid community of leaf-mining sawflies...... 37 4.2.2 The parasitoid community of gall-inducing sawflies...... 39

5 Conclusions...... 43

6 References...... 45

AUTHOR’S CONTRIBUTION

1 Introduction ...... 13 1.1 Plant–insect interactions ...... 14 1.2 Insect–parasitoid interactions...... 16 1.3 Aims of this study...... 18

2 Study system ...... 19 2.1 Sawflies...... 19 2.1.1 Leafminers...... 19 2.1.2 Gall-inducers ...... 21 2.2 Parasitoids ...... 22

5 Conclusions...... 43

6 References...... 45

1 Introduction

Insects comprise over half of the described species on earth, and estimates of the number of existing arthropod species range from 5 to 10 million (Strong et al., 1984; Ødegaard et al., 2005). The diversity of insects has puzzled researchers for decades (Ehrlich & Raven, 1964; Strong et al., 1984; Farrell, 1998; Novotny et al., 2006; Mayhew, 2007). Various factors may have played role in the insects’ success in colonizing the earth, e.g., small size, ability to fly, phytophagy, and old age (McPeek & Brown, 2007; Mayhew, 2007). A large proportion of insects feed on plants, and these herbivorous lineages have managed to spread across the entire angiosperm phylogeny, and have developed several distinct feeding modes at the same time (Janz, 2011). Mitter et al. (1988) showed that phytophagous insect groups are on average more diverse than their non-herbivorous sister taxa. Like phytophagous insects, parasitoids form a diverse and species-rich group. Parasitoids have a close connection with their hosts, so they are usually highly specialized, and cryptic species are often found (Smith et al., 2006; 2007; Kaartinen et al., 2010; Gebiola et al., 2012). However, because the degree of specialization varies at all three trophic levels, plants, herbivores, and parasitoids normally form very complex tritrophic interaction networks (Morris et al., 2004; Kitching, 2006; Nyman et al., 2007). Interactions with other insects or organisms like plants may enhance “eat or be eaten” situations (coevolution), which may speed up diversification in one or both of the interacting lineages (Ehrlich & Raven, 1964; Mitter et al., 1988; Farrell, 1998). Traditionally, plant-feeding insects have been thought to coevolve with their host plants (Ehrlich & Raven, 1964; Farrell et al., 1991). In this coevolutionary model, plants develop new defensive chemicals allowing them to multiply and diversify, but later herbivores invade these new plant species, which could

13 1 Introduction

13 lead to adaptive radiation in the herbivores (Ehrlich & Raven, traits, so that the plants can act as a source of divergent natural 1964; Farrell, 1998). Cospeciation, where two species speciate in selection (Funk et al., 2002; Nyman, 2010). parallel, can be the result of coevolution, but it is not essential Host shifts are found in a wide variety of insect taxa, but the (Agosta, 2006; Janz, 2011; Althoff et al. 2014). However, role of host shifts in insect speciation is still unclear (Lopez- cospeciation is rarely seen between herbivore insects and plants Vaamonde et al., 2003; Percy et al., 2004; Agosta, 2006; Nyman et (Roderick, 1997), and of the few examples that have been found, al., 2006a; Winkler & Mitter, 2008). However, host shifts are maybe the best known is the cophylogeny between figs and fig thought to be an important driver of herbivore diversification wasps (Jackson, 2004). Generally, insects and plants form loose (Winkler & Mitter, 2008). Ecological shifts do not occur only phylogenetic associations, so that closely related insects are between plant species, because evolutionary changes can associated with closely related plants, but with little evidence involve also, for example, shifts from external to internal feeding for strict, long-term cospeciation (e.g. Janz & Nylin, 1998; (Nyman et al., 2006a). Recent studies have shown that Nyman, 2010; Janz, 2011; Althoff et al. 2014). A similar situation diversification of herbivorous insects generally has occurred can be found in the third trophic level, as host phylogeny has later than in their host plants, meaning that present host-use little influence on host–parasitoid associations (Lopez- patterns in insects mainly result from shifting among pre- Vaamonde et al., 2005; Ives & Godfray, 2006; III; IV). existing plant lineages (Gómez-Zurita et al., 2007; Hunt et al., 2007; McKenna et al., 2009; Ohshima et al., 2010; Stireman et al., 2010). Generalism increases the likelihood of host shifts not only 1.1 PLANT–INSECT INTERACTIONS to closely related plant species, but also to novel, distantly related plants (Agosta et al., 2010; Janz, 2011). In general, plant–insect associations are taxonomically Host-associated differentiation (HAD) occurs when host- conservative, so that phylogenetically related insects feed on plant species has a stronger effect on differentiation in insect phylogenetically related plants (Ehrlich & Raven, 1964; Janz & populations than does geographic isolation, so that the insects Nylin, 1998; Farrell, 2001; Novotny et al., 2002; Kergoat et al., form genetically distinct host-associated populations on 2007; Winkler & Mitter, 2008). Thus, one would think that host alternative hosts (Berlocher & Feder, 2002; Drès & Mallet, 2002; or niche shifts are rare. “Host shift” refers to a situation where a Stireman et al., 2005; Scheffer & Hawthorne, 2007; Dickey & population of herbivores has started to use a novel host plant Medina, 2012). If HAD is a common process, ecological and (Agosta, 2006). Every host shift begins with colonization, so evolutionary divergence according to host plants would be initially the herbivore retains a capacity to use both plant found in most insects (Stireman et al., 2005). Indeed, existence of species. Later, some herbivores may differentiate to use only the HAD has been demonstrated in numerous insect taxa (Via et al., ancestral or novel host plant, and the populations therefore form 2000; Stireman et al., 2005; Dorchin et al., 2009; Hernández-Vera reproductively partially isolated host races, which may et al., 2010; Dickey & Medina, 2010), but few studies have tested eventually evolve into new, reproductively isolated species HAD in more than one herbivore–host plant pair (Stireman et associated with different host plants (Drès & Mallet, 2002; al., 2005; Dickey & Medina, 2010; Dickey & Medina, 2012; Egan Berlocher & Feder, 2002; Stireman et al., 2005; Peccoud et al., et al. 2013) (II). HAD and resultant ecological speciation 2009). This is possible if there are differences among plant taxa (Schluter, 2001; Rundle & Nosil, 2005) has been used to explain in chemical, ecological, morphological, and/or phenological the extraordinary diversity of phytophagous insects. Ecological speciation results from divergent selection on traits in different

15 lead to adaptive radiation in the herbivores (Ehrlich & Raven, traits, so that the plants can act as a source of divergent natural 1964; Farrell, 1998). Cospeciation, where two species speciate in selection (Funk et al., 2002; Nyman, 2010). parallel, can be the result of coevolution, but it is not essential Host shifts are found in a wide variety of insect taxa, but the (Agosta, 2006; Janz, 2011; Althoff et al. 2014). However, role of host shifts in insect speciation is still unclear (Lopez- cospeciation is rarely seen between herbivore insects and plants Vaamonde et al., 2003; Percy et al., 2004; Agosta, 2006; Nyman et (Roderick, 1997), and of the few examples that have been found, al., 2006a; Winkler & Mitter, 2008). However, host shifts are maybe the best known is the cophylogeny between figs and fig thought to be an important driver of herbivore diversification wasps (Jackson, 2004). Generally, insects and plants form loose (Winkler & Mitter, 2008). Ecological shifts do not occur only phylogenetic associations, so that closely related insects are between plant species, because evolutionary changes can associated with closely related plants, but with little evidence involve also, for example, shifts from external to internal feeding for strict, long-term cospeciation (e.g. Janz & Nylin, 1998; (Nyman et al., 2006a). Recent studies have shown that Nyman, 2010; Janz, 2011; Althoff et al. 2014). A similar situation diversification of herbivorous insects generally has occurred can be found in the third trophic level, as host phylogeny has later than in their host plants, meaning that present host-use little influence on host–parasitoid associations (Lopez- patterns in insects mainly result from shifting among pre- Vaamonde et al., 2005; Ives & Godfray, 2006; III; IV). existing plant lineages (Gómez-Zurita et al., 2007; Hunt et al., 2007; McKenna et al., 2009; Ohshima et al., 2010; Stireman et al., 2010). Generalism increases the likelihood of host shifts not only 1.1 PLANT–INSECT INTERACTIONS to closely related plant species, but also to novel, distantly related plants (Agosta et al., 2010; Janz, 2011). In general, plant–insect associations are taxonomically Host-associated differentiation (HAD) occurs when host- conservative, so that phylogenetically related insects feed on plant species has a stronger effect on differentiation in insect phylogenetically related plants (Ehrlich & Raven, 1964; Janz & populations than does geographic isolation, so that the insects Nylin, 1998; Farrell, 2001; Novotny et al., 2002; Kergoat et al., form genetically distinct host-associated populations on 2007; Winkler & Mitter, 2008). Thus, one would think that host alternative hosts (Berlocher & Feder, 2002; Drès & Mallet, 2002; or niche shifts are rare. “Host shift” refers to a situation where a Stireman et al., 2005; Scheffer & Hawthorne, 2007; Dickey & population of herbivores has started to use a novel host plant Medina, 2012). If HAD is a common process, ecological and (Agosta, 2006). Every host shift begins with colonization, so evolutionary divergence according to host plants would be initially the herbivore retains a capacity to use both plant found in most insects (Stireman et al., 2005). Indeed, existence of species. Later, some herbivores may differentiate to use only the HAD has been demonstrated in numerous insect taxa (Via et al., ancestral or novel host plant, and the populations therefore form 2000; Stireman et al., 2005; Dorchin et al., 2009; Hernández-Vera reproductively partially isolated host races, which may et al., 2010; Dickey & Medina, 2010), but few studies have tested eventually evolve into new, reproductively isolated species HAD in more than one herbivore–host plant pair (Stireman et associated with different host plants (Drès & Mallet, 2002; al., 2005; Dickey & Medina, 2010; Dickey & Medina, 2012; Egan Berlocher & Feder, 2002; Stireman et al., 2005; Peccoud et al., et al. 2013) (II). HAD and resultant ecological speciation 2009). This is possible if there are differences among plant taxa (Schluter, 2001; Rundle & Nosil, 2005) has been used to explain in chemical, ecological, morphological, and/or phenological the extraordinary diversity of phytophagous insects. Ecological speciation results from divergent selection on traits in different

15 ecological environments, which leads to reproductive isolation specialization provides the raw material for evolutionary shifts between populations either directly or indirectly (Schluter & in associations (Janz, 2011), and many studies have indicated Conte, 2009). In addition to phytophagous insects, ecological that niche shifts are an important driving force in the speciation has been studied in a wide variety of groups, diversification of plant-feeding insects as well as of their including parasitoids (Stireman et al., 2006), fishes (Schluter & enemies (Nyman et al., 2007; Fordyce, 2010; Segar et al., 2012). It Conte, 2009), seed-eating birds (Hendry et al., 2009), and lizards has also been suggested that parasitoid pressure could cause (Rosenblum & Harmon, 2011). However, it still remains adaptive shifts along other niche dimensions and thereby, for unknown how important HAD really is for insect example, lead to the evolution of gall induction or other internal diversification, so comparative studies focusing on several feeding habits (Price & Clancy, 1986; Connor & Taverner, 1997). insects that use the same hosts (Stireman et al., 2005; II) provide Diversifying effects can theoretically operate in both “top– good evidence on the commonness and repeatability of the down” and “bottom–up” directions. In the plant-driven process. “bottom–up” model, diversification of plants leads to diversification of herbivores, which, in turn, leads to increased diversity in parasitoid species (Abrahamson et al., 2003; 1.2 INSECT–PARASITOID INTERACTIONS Stireman et al., 2006; Forbes et al., 2009). More research has focused on parasitoid -driven “top–down” diversification, as it Many factors can influence how enemies find the host species has become evident that natural enemies can influence host– that they attack. Plants and herbivores vary in their occurrence plant use of herbivores (Bernays & Graham, 1988; Nosil & in time and space and, thus, only some of the potential hosts of a Crespi, 2006). This happens if parasitoids preferentially attack particular parasitoid species may be available at a given time. herbivores on specific plant species (Rott & Godfray, 2000; Host phylogeny (Desneux et al., 2012) as well as host niche Murphy, 2004), in which case switches to novel plant species (Hawkins, 1994) can also influence host acceptance and could provide the herbivores with “enemy-free space,” and suitability. At least some parasitoids use host plants as cues thereby facilitate speciation by niche shifts (Lill et al., 2002; when searching for their hosts (McCall et al., 1993; De Moraes et Singer & Stireman, 2005). al., 1998; Powell et al., 1998; De Moraes & Mescher, 2004) and, in While an extensive literature on the phylogenetic history of gall-inducing insects gall morphology has been shown to affect insect–plant interactions exists, only few studies involve a parasitism (Price & Clancy 1986; Price et al., 1987; Kopelke, 1999; tritrophic phylogenetic comparison of plants, herbivores, and Craig et al., 2007; Nyman et al., 2007) parasitoids (Lopez-Vaamonde et al., 2005; Ives & Godfray, 2006; In tritrophic interaction networks, the level of specialization Nyman et al., 2007; III; IV). Molecular-phylogenetic methods typically varies at all levels, because diet breadth in both and DNA barcoding have become important tools for resolving herbivores and parasitoids ranges from extreme specialization the history and composition of parasitoid complexes (Smith et to broad generalism (Godfray, 1994; Nosil & Mooers, 2005; al., 2006; 2007; 2008; 2011; Kaartinen et al., 2010; Hrcek et al., Stireman, 2005). Nevertheless, most herbivores use only a small 2011). Identification of parasitoids based on morphology alone proportion of the available plant species (Novotny et al., 2010), can be difficult and time-consuming, and rearing can also be and the situation is similar in parasitoids, which mostly attack unreliable if parasitoid species differ in rearing success. only some of the available herbivores (Hawkins, 1994; Cagnolo Therefore, the use of molecular methods can lead to more et al., 2011). Over longer time scales, such variation in

17 ecological environments, which leads to reproductive isolation specialization provides the raw material for evolutionary shifts between populations either directly or indirectly (Schluter & in associations (Janz, 2011), and many studies have indicated Conte, 2009). In addition to phytophagous insects, ecological that niche shifts are an important driving force in the speciation has been studied in a wide variety of groups, diversification of plant-feeding insects as well as of their including parasitoids (Stireman et al., 2006), fishes (Schluter & enemies (Nyman et al., 2007; Fordyce, 2010; Segar et al., 2012). It Conte, 2009), seed-eating birds (Hendry et al., 2009), and lizards has also been suggested that parasitoid pressure could cause (Rosenblum & Harmon, 2011). However, it still remains adaptive shifts along other niche dimensions and thereby, for unknown how important HAD really is for insect example, lead to the evolution of gall induction or other internal diversification, so comparative studies focusing on several feeding habits (Price & Clancy, 1986; Connor & Taverner, 1997). insects that use the same hosts (Stireman et al., 2005; II) provide Diversifying effects can theoretically operate in both “top– good evidence on the commonness and repeatability of the down” and “bottom–up” directions. In the plant-driven process. “bottom–up” model, diversification of plants leads to diversification of herbivores, which, in turn, leads to increased diversity in parasitoid species (Abrahamson et al., 2003; 1.2 INSECT–PARASITOID INTERACTIONS Stireman et al., 2006; Forbes et al., 2009). More research has focused on parasitoid -driven “top–down” diversification, as it Many factors can influence how enemies find the host species has become evident that natural enemies can influence host– that they attack. Plants and herbivores vary in their occurrence plant use of herbivores (Bernays & Graham, 1988; Nosil & in time and space and, thus, only some of the potential hosts of a Crespi, 2006). This happens if parasitoids preferentially attack particular parasitoid species may be available at a given time. herbivores on specific plant species (Rott & Godfray, 2000; Host phylogeny (Desneux et al., 2012) as well as host niche Murphy, 2004), in which case switches to novel plant species (Hawkins, 1994) can also influence host acceptance and could provide the herbivores with “enemy-free space,” and suitability. At least some parasitoids use host plants as cues thereby facilitate speciation by niche shifts (Lill et al., 2002; when searching for their hosts (McCall et al., 1993; De Moraes et Singer & Stireman, 2005). al., 1998; Powell et al., 1998; De Moraes & Mescher, 2004) and, in While an extensive literature on the phylogenetic history of gall-inducing insects gall morphology has been shown to affect insect–plant interactions exists, only few studies involve a parasitism (Price & Clancy 1986; Price et al., 1987; Kopelke, 1999; tritrophic phylogenetic comparison of plants, herbivores, and Craig et al., 2007; Nyman et al., 2007) parasitoids (Lopez-Vaamonde et al., 2005; Ives & Godfray, 2006; In tritrophic interaction networks, the level of specialization Nyman et al., 2007; III; IV). Molecular-phylogenetic methods typically varies at all levels, because diet breadth in both and DNA barcoding have become important tools for resolving herbivores and parasitoids ranges from extreme specialization the history and composition of parasitoid complexes (Smith et to broad generalism (Godfray, 1994; Nosil & Mooers, 2005; al., 2006; 2007; 2008; 2011; Kaartinen et al., 2010; Hrcek et al., Stireman, 2005). Nevertheless, most herbivores use only a small 2011). Identification of parasitoids based on morphology alone proportion of the available plant species (Novotny et al., 2010), can be difficult and time-consuming, and rearing can also be and the situation is similar in parasitoids, which mostly attack unreliable if parasitoid species differ in rearing success. only some of the available herbivores (Hawkins, 1994; Cagnolo Therefore, the use of molecular methods can lead to more et al., 2011). Over longer time scales, such variation in

17 accurate estimates of interaction strengths (Kaartinen et al., 2010; Smith et al., 2011; Wirta et al., 2014). 2 Study System

1.3 AIMS OF THIS STUDY 2.1 SAWFLIES The aims of this thesis were to explore insect speciation using leaf-mining and gall-inducing sawflies and their parasitoids as 2.1.1 Leafminers model systems. In particular, I wanted to resolve the role of host Leafminer larvae live between the upper and lower epidermal plants and parasitoids in herbivore speciation, and to explore layers of leaves (Connor & Taverner, 1997; Sinclair & Hughes, speciation in parasitoids. Very few studies have compared more 2010). The habit of leaf mining has evolved independently than one insect–plant pair, meaning that little is known about multiple times in four insect orders: the Diptera, Lepidoptera, the repeatability of HAD. The extreme complexity of tritrophic Coleoptera and Hymenoptera (Hespenheide, 1991; Connor & interactions between plants, herbivores, and parasitoids make Taverner, 1997; Sinclair & Hughes, 2010). The adaptive them demanding to study, so little is known about how significance of leaf-mining may be that the mine shelters the tritrophic interactions are assembled and maintained. Even less insect larvae from environmental stress and natural enemies, is known about how herbivore host-plant shifts influence and/or helps in avoiding plant defenses (Connor & Taverner, parasitoid speciation, and whether enemy communities of 1997). Leaf-mining lineages are often less species rich than their phytophagous insects are determined mainly by the phylogeny external-feeding sister taxa, and miners also tend to specialize of the herbivores, by geographic region, or by ecological factors on single plant species or on a few closely related plants (Smith, such as host-plant use or preferred habitat. 1979; Hespenheide, 1991; Connor & Taverner, 1997). Most hymenopteran leafminers belong to the Tenthredinidae, which is the largest family of herbivorous symphytans The specific aims of this study were: (Viitasaari, 2002; Davis et al., 2010) (Fig. 1A,B). Within Tenthredinidae, the leaf-mining habit has evolved twice, in the 1. To study the role of host-plant shifts in insect speciation. subfamily Heterarthrinae and in the nematine tribe (I & II) Pseudodineurini (Pschorn-Walcher & Altenhofer, 1989) (I). 2. To investigate repeatability of host-associated genetic Heterarthrinae includes 29 genera and over 150 species, and differentiation. (II) heterarthrines are found all over the world except for Africa, 3. To investigate the evolutionary assembly of complex Australia, and Antarctica (Goulet, 1992; Taeger et al., 2010). tritrophic food webs. (III & IV) Most heterarthrines are quite specialized, using one or a few (in 4. To explore how parasitoids influence herbivore most cases, woody) plant species as hosts (Smith, 1979; Pschorn- speciation, and to study the role of niche and habitat Walcher & Altenhofer, 1989; Taeger & Altenhofer, 1998), but shifts in parasitoid speciation. (III & IV) their collective host range includes over 20 plant genera in ten families (Smith, 1971; Altenhofer, 2003). The tribe Pseudodineurini belongs to the subfamily Nematinae and includes only 12 known species (Taeger et al., 2010). Pseudodineurini miners are almost exclusively specialized on

19 accurate estimates of interaction strengths (Kaartinen et al., 2010; Smith et al., 2011; Wirta et al., 2014). 2 Study System

19 the plant family Ranunculaceae (Altenhofer, 2003). Larvae of this tribe emit a typical odor of citral, which is lacking in Heterarthrinae leafminers (Boevé et al., 2009).

2.1.2 Gallers Gallers represent a special case of insect–plant interactions, as the insects have to alter plant morphology to produce a gall in which their larvae live. Plant galls are generally defined as pathologically developed cells, tissues, or organs that are formed by either increasing cell number or size (Meyer, 1987). The habit of galling has evolved several times in many different insect orders (Meyer, 1987; Crespi et al., 1997; Nyman et al., 2000; Cook & Gullan, 2004), producing species-rich groups with a long history, as the first galls were present already 300 million years ago (Labandeira & Phillips, 1996). Galls are believed to provide the gall-inducers with better nutrition, shelter from the environment, and protection from natural enemies (Price et al., 1987; Stone & Schönrogge, 2003). Most gallers are very host and tissue specific (Roininen et al., 2005; Hardy & Cook, 2010). Thus, in many galler genera, species attack different plants and plant parts, indicating that there have been shifts and adaptive radiations into novel ecological niches (Price, 2005). Most hymenopteran gallers belong to the tenthredinid subfamily Nematinae (Goulet, 1992; Viitasaari, 2002). The most species-rich group of nematine gallers is the subtribe Euurina, which includes c. 400 species that induce leaf folds or rolls, or various types of closed galls on Salix (a few North American leaf-folding and -rolling species live on Populus) (Kopelke, 1999; Roininen et al., 2005). Euurina gallers can be classified by their gall morphology into three genera: Phyllocolpa species induce open leaf folds or rolls, Pontania species induce closed leaf galls Figure 1. Examples of sawflies studied in this thesis. Leafminers: A) Scolioneura betuleti on Betula pubescens ssp. czerepanovii, B) Fenusa ulmi on Ulmus sp. Leaf (Fig. 1C,D), and Euura species induce bud (Fig. 1E,F), petiole, gallers: C) Pontania aquilonis on Salix herbacea, D) Pontania nivalis on Salix glauca. shoot, or midrib galls (Smith, 1970; Price & Roininen, 1993; Bud gallers: E) Euura mucronata on Salix glauca, F) Euura lanatae on Salix lanata. Photos A, B, D and F by T. Nyman, and C and E by S. Leppänen. Kopelke, 1999; Kopelke, 2003). The genus Pontania is divided into six species groups (dolichura-, herbaceae-, polaris-, vesicator-, proxima-, and viminalis-group), and Euura is divided into the

21 the plant family Ranunculaceae (Altenhofer, 2003). Larvae of this tribe emit a typical odor of citral, which is lacking in Heterarthrinae leafminers (Boevé et al., 2009).

21 mucronata-, laeta-, testaceipes-, venusta-, atra-, and amerinae-groups communities of small insects (such as most leafminers) tend to (Kopelke, 1999; 2003). Euurina gallers live in the Holarctic be dominated by chalcids (especially Eulophidae) (Askew & region (Smith, 1979; Price & Roininen, 1993; Roininen et al., Shaw, 1974). Most parasitoids are niche specialist that search for 2005), as do their species-rich host group Salix (c. 300–500 hosts that have a particular feeding habit or that utilize a specific species) (Argus, 1997; Skvortsov, 1999). plant part (Hawkins, 1994). Small eulophids that attack miners Phylogenetic studies on Euurina have revealed that gall-type are thought to use visual cues like mine shape or color for shifts have been less frequent than host-plant shifts during the finding their host (Salvo & Valladares, 2004). External-feeding evolutionary history of the group (Nyman et al., 2000; Nyman et herbivores on average have fewer parasitoids than do internal- al., 2007). The Euurina gallers’ pronounced host specificity and feeding species. However, parasitoid species richness varies close connection with their host plants makes the group an widely, and leafminers tend to be attacked by comparatively excellent subject for investigations on ecological speciation and high numbers of parasitoids (Hawkins, 1994). diversification. Like other insect herbivores, sawflies are attacked by a diverse community of parasitoids, and parasitoids may compose a significant mortality factor for sawfly larvae 2.2 PARASITOIDS (Hawkins, 1994; Connor & Taverner, 1997; Kopelke, 1999). Heterarthrine larvae are attacked by parasitoids from three Parasitoids are arthropods that live in or on another arthropod hymenopteran families (Ichneumonidae, Braconidae, and species, eventually killing the host individual (Godfray, 1994). Eulophidae). The varying preferences and diet breadths of these Parasitoids differ from predators in that they use only one enemy species (Pschorn-Walcher & Altenhofer, 1989) make individual as a host. Several different insect orders contain them an excellent study group for food-web studies. A very parasitoid lineages, but most parasitoids belong to either similar situation occurs in the Euurina gallers: parasitoids and Hymenoptera or Diptera (Godfray, 1994; Hawkins, 1994). Most parasitic inquilines constitute the main source of mortality in hymenopteran parasitoids belong to the paraphyletic group many galler species (Kopelke, 1999; Roininen et al., 2002; Craig “Parasitica” in the suborder Apocrita (Goulet and Huber, 1993). et al., 2007), and the gall inducers collectively support a complex Hymenopteran parasitoids have specialized ovipositors that are of over 80 parasitoid and inquiline species from 16 families in used for depositing eggs. The parasitoid sting causes paralysis four insect orders (Roininen & Danell, 1997; Roininen et al., in a host, which may be permanent, or the host may continue 2002; Kopelke, 1999). living (Godfray, 1994). Dipteran parasitoids do not have ovipositors capable of penetrating plant tissues, so leafminers and gallers usually lack these parasitoids (Hawkins, 1994). Parasitoids are usually highly specialized because of their close connection with their hosts. Parasitoids are not distributed randomly among host groups, as host feeding type, body size, etc. varies, and some parasitoids are more specialized than others. For example, since ichneumonids are on average larger than chalcids, only few ichneumonids can develop on very small hosts (Hawkins, 1994). Consequently, parasitoid

23 mucronata-, laeta-, testaceipes-, venusta-, atra-, and amerinae-groups communities of small insects (such as most leafminers) tend to (Kopelke, 1999; 2003). Euurina gallers live in the Holarctic be dominated by chalcids (especially Eulophidae) (Askew & region (Smith, 1979; Price & Roininen, 1993; Roininen et al., Shaw, 1974). Most parasitoids are niche specialist that search for 2005), as do their species-rich host group Salix (c. 300–500 hosts that have a particular feeding habit or that utilize a specific species) (Argus, 1997; Skvortsov, 1999). plant part (Hawkins, 1994). Small eulophids that attack miners Phylogenetic studies on Euurina have revealed that gall-type are thought to use visual cues like mine shape or color for shifts have been less frequent than host-plant shifts during the finding their host (Salvo & Valladares, 2004). External-feeding evolutionary history of the group (Nyman et al., 2000; Nyman et herbivores on average have fewer parasitoids than do internal- al., 2007). The Euurina gallers’ pronounced host specificity and feeding species. However, parasitoid species richness varies close connection with their host plants makes the group an widely, and leafminers tend to be attacked by comparatively excellent subject for investigations on ecological speciation and high numbers of parasitoids (Hawkins, 1994). diversification. Like other insect herbivores, sawflies are attacked by a diverse community of parasitoids, and parasitoids may compose a significant mortality factor for sawfly larvae 2.2 PARASITOIDS (Hawkins, 1994; Connor & Taverner, 1997; Kopelke, 1999). Heterarthrine larvae are attacked by parasitoids from three Parasitoids are arthropods that live in or on another arthropod hymenopteran families (Ichneumonidae, Braconidae, and species, eventually killing the host individual (Godfray, 1994). Eulophidae). The varying preferences and diet breadths of these Parasitoids differ from predators in that they use only one enemy species (Pschorn-Walcher & Altenhofer, 1989) make individual as a host. Several different insect orders contain them an excellent study group for food-web studies. A very parasitoid lineages, but most parasitoids belong to either similar situation occurs in the Euurina gallers: parasitoids and Hymenoptera or Diptera (Godfray, 1994; Hawkins, 1994). Most parasitic inquilines constitute the main source of mortality in hymenopteran parasitoids belong to the paraphyletic group many galler species (Kopelke, 1999; Roininen et al., 2002; Craig “Parasitica” in the suborder Apocrita (Goulet and Huber, 1993). et al., 2007), and the gall inducers collectively support a complex Hymenopteran parasitoids have specialized ovipositors that are of over 80 parasitoid and inquiline species from 16 families in used for depositing eggs. The parasitoid sting causes paralysis four insect orders (Roininen & Danell, 1997; Roininen et al., in a host, which may be permanent, or the host may continue 2002; Kopelke, 1999). living (Godfray, 1994). Dipteran parasitoids do not have ovipositors capable of penetrating plant tissues, so leafminers and gallers usually lack these parasitoids (Hawkins, 1994). Parasitoids are usually highly specialized because of their close connection with their hosts. Parasitoids are not distributed randomly among host groups, as host feeding type, body size, etc. varies, and some parasitoids are more specialized than others. For example, since ichneumonids are on average larger than chalcids, only few ichneumonids can develop on very small hosts (Hawkins, 1994). Consequently, parasitoid

3. Materials and Methods

3.1 MATERIAL SAMPLING

Sawfly material for this thesis was provided by many researchers, but mainly collected by T. Nyman. The leafminer study (I) contained 39 heterarthrine species and 20 outgroup species. Galler samples (II) were collected from two locations in northern Fennoscandia (Kilpisjärvi in Finland and Abisko in Sweden). Euura bud galls were collected in the year 1998 from six willow species: Salix lanata, S. glauca, S. lapponum, S. phylicifolia, S. myrsinifolia, and S. hastata. Leaf-galling Pontania were collected in the years 1997 and 1998 from same willow species, but also from S. borealis, S. reticulata, and S. myrsinites. Parasitoid larvae (IV) were likewise collected from Kilpisjärvi and Abisko, but also from Tromsø in Norway. Galls were randomly collected in years 2009 and 2010 into plastic bags from six boreal–subarctic willow species (Salix lanata, S. glauca, S. lapponum, S. myrsinifolia, and S. phylicifolia) and from three arctic–alpine dwarf willow species (S. reticulata, S. polaris, and S. herbacea). In the laboratory, some of the galls were dissected under a microscope in order to find out the fate of the sawfly larva (live galler larva / galler killed by parasitoid / galler’s cause of death unknown), and the rest of the galls were put into glass jars in order to rear galler and parasitoid adults (Fig. 2).

3. Materials and Methods

3.1 MATERIAL SAMPLING

25 variability for studying closely related species (Coleman, 2003; Yao et al., 2010). All sequences were read, and edited, and most of them also aligned, using Sequencher 4.8 (Gene Codes Corp., Ann Arbor, MI). Because the lengths of Cytb and ITS2 sequences differed among species, they were aligned using ClustalW (Larkin et al., 2007). In the galler parasitoid study (IV), parasitoid larvae were identified based on standard DNA barcode methods (Hebert et al., 2003). DNA barcoding rests on the idea that all organisms Figure 2. Galler and parasitoid rearing jars with collection tubes. Leaves with can be identified using one or a few short sequence fragments. Pontania galls were put into glass jars having a sand layer of a few centimeters at the In the case of animals, the standard barcode is a portion of the bottom, and a thin layer of Sphagnum moss on top of the sand. In April, when the mitochondrial COI gene (Hebert et al., 2003), which is relatively rearing jars were taken to room temperature, they were covered with black plastic. The easy to amplify and sequence, and usually has enough lids of the jars were pierced with a silicone tube that led to a transparent plastic differences for reliable identification of species (e.g., Hebert et collection tube. Emerging sawflies and parasitoids were attracted to the collection al., 2003; 2004; Hajibabaei et al., 2006; Dinca et al., 2010). Reared heads by light entering the rearing jar via the silicone tube. parasitoid adults were identified by expert taxonomists, and

then sequenced in order to build a reference library, which could be used for identifying larval samples. With the help of the barcode database, we were able to use larval samples 3.2 GENETIC MATERIAL without a need to rear parasitoids to adults. This facilitates the identification process and reduces errors caused by Details of DNA extraction, PCR amplification, and sequencing taxonomically variable mortality during rearing (Tilmon et al., can be found in Nyman et al. (2000), (2006b) and studies I, III, 2000; Agustí et al., 2005; Smith et al., 2006; 2007; 2008). and IV. In the leafminer study (I), we used two mitochondrial genes, Cytochrome oxidase I (COI) and Cytochrome b (Cytb), and two nuclear genes, Elongation factor-1α (EF-1α) and 3.3 PHYLOGENY RECONSTRUCTION Sodium-potassium adenosine triphosphatase (NAK). The relatively fast-evolving mitochondrial COI and Cytb genes are Phylogenetic trees present evolutionary relationships among suitable for studying relatively recent speciation events, whereas organisms, illustrating how closely they are related based on the “slower” exons of the nuclear EF-1α and NAK genes are their genetic or physical characters (Lemey et al., 2009). There better suited for investigating older divergences (Lin & are various methods to construct a phylogeny and every method Danforth, 2004; Whitfield & Kjer, 2008). The genus-level time- has its pros and cons, but none of them guarantee that the calibrated plant phylogeny used in the miner study (I) was obtained tree is the “true” phylogenetic tree. Basically, methods reconstructed using matK, 26S, 18S, rbcL, and atpB sequences can be divided into character-state and distance-matrix collected from the literature and GenBank. For gall-inducing methods, based on what kind of approach they use to analyze sawflies (II), we sequenced the mitochondrial COI gene and the the data. Simple distance-matrix methods such as UPGMA and nuclear, fast-evolving ribosomal Internal transcribed spacer 2 Neighbor-joining (NJ) are computationally fast and, thus, (ITS2) region, which is easy to amplify and has enough

27 variability for studying closely related species (Coleman, 2003; Yao et al., 2010). All sequences were read, and edited, and most of them also aligned, using Sequencher 4.8 (Gene Codes Corp., Ann Arbor, MI). Because the lengths of Cytb and ITS2 sequences differed among species, they were aligned using ClustalW (Larkin et al., 2007). In the galler parasitoid study (IV), parasitoid larvae were identified based on standard DNA barcode methods (Hebert et al., 2003). DNA barcoding rests on the idea that all organisms Figure 2. Galler and parasitoid rearing jars with collection tubes. Leaves with can be identified using one or a few short sequence fragments. Pontania galls were put into glass jars having a sand layer of a few centimeters at the In the case of animals, the standard barcode is a portion of the bottom, and a thin layer of Sphagnum moss on top of the sand. In April, when the mitochondrial COI gene (Hebert et al., 2003), which is relatively rearing jars were taken to room temperature, they were covered with black plastic. The easy to amplify and sequence, and usually has enough lids of the jars were pierced with a silicone tube that led to a transparent plastic differences for reliable identification of species (e.g., Hebert et collection tube. Emerging sawflies and parasitoids were attracted to the collection al., 2003; 2004; Hajibabaei et al., 2006; Dinca et al., 2010). Reared heads by light entering the rearing jar via the silicone tube. parasitoid adults were identified by expert taxonomists, and

27 suitable for large data sets (Lemey et al., 2009). An NJ tree was with highest likelihood is chosen to be the best tree. ML trees calculated for gall-inducing sawflies based on combined data were calculated in RAxML v. 7.2.8 (Stamatakis et al., 2008; I; IV) (COI + ITS2), after removing specimens that did not have and PhyML v. 3.0 (Guindon et al., 2010; II). sequences of both genes, in PAUP v. 4.0b10 (Swofford, 2003; II). Bayesian methods differ from MP and ML, as Bayesian An NJ tree was also calculated for adult parasitoid sequences in inference approaches do not attempt to find only one best tree MEGA v. 5.1 (Tamura et al., 2011). Later, larval samples were (Lemey et al., 2009). Bayesian methods search for plausible trees added to the dataset for larval identification, and a final NJ tree using likelihood, with the researcher specifying prior beliefs and was then calculated with only larval samples (IV). an evolutionary model. Like for other tree construction methods, Other methods commonly used for phylogeny reconstruction there are multiple programs for Bayesian analyses, with are Maximum parsimony (MP), Maximum likelihood (ML), and different focuses. Leafminer sequences (I) were analyzed using Bayesian methods. The idea behind MP is simple: the aim is to MrBayes v. 3.1.2 (Ronquist & Huelsenbeck, 2003), which focuses find the tree or trees that require the least evolutionary change in phylogenetic inference, and in BEAST v. 1.5.2 (Drummond & (= smallest number of changes in characters) in the observed Rambaut, 2007), which aims at producing trees with a time data (Kluge & Farris, 1969; Fitch, 1971). In sequence analyses, scale. BEAST was especially used to estimate diversification gaps can be treated as a fifth character, which is why MP was times of leaf-mining sawflies and their host plants using a used for analyzing ITS2 sequence data of gall-inducing sawflies Bayesian relaxed molecular-clock method. In the study on gall- (II). Basic parsimony methods work well for small data sets inducing sawflies (II), MrBayes v. 3.2.2 (Ronquist et al. 2012) was containing fewer than 100 taxa, but for larger data sets better used for CoI, ITS2 and the combinined data set (CoI + ITS2). options are available, like Nixon’s (1999) parsimony ratchet Before the analysis, gaps in ITS2 sequence had to be coded method, which was implemented in PAUPrat (Sikes & Lewis, present or absent using the method of Simmons and Ochoterena 2001) in conjunction with PAUP (I; II). Ratchet searches are more (2000) in the program FastGap v. 1.2 (Borchsenius 2009). efficient in finding the most parsimonious tree than are heuristic The Pseudodineura phylogeny presented in this thesis (see searches, by sampling more tree islands and collecting fewer below) was calculated using COI, Cytb, EF-1α and NAK trees from each island than traditional methods (Nixon, 1999). sequences. RAxML was run at the CIPRES web portal (Miller et In model-based analyses, the best-fitting substitution model al., 2010) with a separate GTR+G model for each data partition has to be found for each gene before the analysis. Evolutionary (gene), and support for nodes were evaluated based on 100 models are assumptions of nucleotide or amino acid bootstrap replicates. substitutions, and they describe the probabilities of change from one nucleotide or amino acid to another in a phylogenetic tree (Liò & Goldman, 1998; Lemey et al., 2009). Model selection was 3.4 INSECT–PLANT ANALYSES performed in jModelTest v. 0.1.1 (Posada, 2008; I) and v. 2.1.3 (Darriba et al., 2012; II). After constructing phylogenies for sawflies and plants, it was ML searches for the tree that has the highest likelihood of possible to compare them with each other. Contrasting the producing the observed data given the tree topology, branch phylogenetic trees of insects with those of their host plants can lengths, and the selected model of evolution (Lemey et al., 2009). show how these associations have evolved, for example, have In likelihood calculations, all possible nucleotide states are there been host shifts or has there been cospeciation (Jackson, summed in the ancestral nodes to obtain the tree, and the tree 2004; Lopez-Vaamonde et al., 2003; Nyman, 2010; Wilson et al.,

29 suitable for large data sets (Lemey et al., 2009). An NJ tree was with highest likelihood is chosen to be the best tree. ML trees calculated for gall-inducing sawflies based on combined data were calculated in RAxML v. 7.2.8 (Stamatakis et al., 2008; I; IV) (COI + ITS2), after removing specimens that did not have and PhyML v. 3.0 (Guindon et al., 2010; II). sequences of both genes, in PAUP v. 4.0b10 (Swofford, 2003; II). Bayesian methods differ from MP and ML, as Bayesian An NJ tree was also calculated for adult parasitoid sequences in inference approaches do not attempt to find only one best tree MEGA v. 5.1 (Tamura et al., 2011). Later, larval samples were (Lemey et al., 2009). Bayesian methods search for plausible trees added to the dataset for larval identification, and a final NJ tree using likelihood, with the researcher specifying prior beliefs and was then calculated with only larval samples (IV). an evolutionary model. Like for other tree construction methods, Other methods commonly used for phylogeny reconstruction there are multiple programs for Bayesian analyses, with are Maximum parsimony (MP), Maximum likelihood (ML), and different focuses. Leafminer sequences (I) were analyzed using Bayesian methods. The idea behind MP is simple: the aim is to MrBayes v. 3.1.2 (Ronquist & Huelsenbeck, 2003), which focuses find the tree or trees that require the least evolutionary change in phylogenetic inference, and in BEAST v. 1.5.2 (Drummond & (= smallest number of changes in characters) in the observed Rambaut, 2007), which aims at producing trees with a time data (Kluge & Farris, 1969; Fitch, 1971). In sequence analyses, scale. BEAST was especially used to estimate diversification gaps can be treated as a fifth character, which is why MP was times of leaf-mining sawflies and their host plants using a used for analyzing ITS2 sequence data of gall-inducing sawflies Bayesian relaxed molecular-clock method. In the study on gall- (II). Basic parsimony methods work well for small data sets inducing sawflies (II), MrBayes v. 3.2.2 (Ronquist et al. 2012) was containing fewer than 100 taxa, but for larger data sets better used for CoI, ITS2 and the combinined data set (CoI + ITS2). options are available, like Nixon’s (1999) parsimony ratchet Before the analysis, gaps in ITS2 sequence had to be coded method, which was implemented in PAUPrat (Sikes & Lewis, present or absent using the method of Simmons and Ochoterena 2001) in conjunction with PAUP (I; II). Ratchet searches are more (2000) in the program FastGap v. 1.2 (Borchsenius 2009). efficient in finding the most parsimonious tree than are heuristic The Pseudodineura phylogeny presented in this thesis (see searches, by sampling more tree islands and collecting fewer below) was calculated using COI, Cytb, EF-1α and NAK trees from each island than traditional methods (Nixon, 1999). sequences. RAxML was run at the CIPRES web portal (Miller et In model-based analyses, the best-fitting substitution model al., 2010) with a separate GTR+G model for each data partition has to be found for each gene before the analysis. Evolutionary (gene), and support for nodes were evaluated based on 100 models are assumptions of nucleotide or amino acid bootstrap replicates. substitutions, and they describe the probabilities of change from one nucleotide or amino acid to another in a phylogenetic tree (Liò & Goldman, 1998; Lemey et al., 2009). Model selection was 3.4 INSECT–PLANT ANALYSES performed in jModelTest v. 0.1.1 (Posada, 2008; I) and v. 2.1.3 (Darriba et al., 2012; II). After constructing phylogenies for sawflies and plants, it was ML searches for the tree that has the highest likelihood of possible to compare them with each other. Contrasting the producing the observed data given the tree topology, branch phylogenetic trees of insects with those of their host plants can lengths, and the selected model of evolution (Lemey et al., 2009). show how these associations have evolved, for example, have In likelihood calculations, all possible nucleotide states are there been host shifts or has there been cospeciation (Jackson, summed in the ancestral nodes to obtain the tree, and the tree 2004; Lopez-Vaamonde et al., 2003; Nyman, 2010; Wilson et al.,

29 2012; I). If plants and herbivores have cospeciated, then their 3.5 ANALYSES OF TRITROPHIC INTERACTIONS phylogenies should be mirror images of each other, and their diversification times should be concurrent. Comparing One method to analyze tritrophic interactions is to use Ives and diversification times of insects and hosts provides valuable Godfray's (2006) phylogenetic bipartite linear model (pblm), information on insect radiation (Lopez-Vaamonde et al., 2006; which can be found in the picante v. 1.2 package (Kembel et al., Gómez-Zurita et al., 2007; Ohshima et al., 2010; I). Insect 2010) in R v. 2.10 (R Development Core Team 2009). Pblm was phylogenies can also be used to illustrate, for example, used in studies III and IV, because the model can be used to evolution of larval feeding habits (I). Larval feeding and host- estimate how well ecological versus phylogenetic factors explain plant use were reconstructed in Mesquite using the “Trace parasitoid–insect associations. For this analysis, one has to character over trees” option, with either parsimony or construct a phylogeny for all studied levels. A binary (parasitoid maximum-likelihood optimization (Maddison & Maddison, present / absent) association matrix was constructed for 25 2010). Before the reconstruction, each miner was coded heterarthrine species for which information on parasitoids was according to its host-plant genera or larval feeding habit. An ML available (59 species) (Fig 2; III) (Pschorn-Walcher & Altenhofer, tree was used in larval feeding reconstruction in the tree 1989; Digweed et al., 2009). A similar matrix was constructed for including leaf-mining tribe Pseudodineura, which was seven galler species with 14 barcode-identified parasitoids, but constructed using maximum-likelihood optimization (Fig. 3). this time with quantitative data (IV). In the leafminer study (III), Genetic variation among and within populations can be we also created a “pseudophylogeny” for the miners based on analyzed in ARLEQUIN v. 3.5 (Excoffier & Lischer, 2010). In the the host-plant phylogeny. First, plant species not used by miners

galler study (II) we used ARLEQUIN to estimate pairwise ΦST were removed from the full plant phylogeneny. Then miners values among population samples collected from different hosts were added onto the tree according to their host genera, and locations, and also performed a hierarchical analysis of maeaning thst miners using same plant genera as a host created molecular variance (AMOVA) (Excoffier et al., 1992). These a polytomy. The signal-strength parameter (d) measures the analyses were performed separately for mitochondrial and how well the phylogenetic trees explain associations: a value of nuclear genes. Only the six willow species that were shared by 0 indicates absence of phylogenetic signal, values from 0 to 1 both galler genera were used. In the hierarchical AMOVAs, we represent stabilizing selection, a value 1 conforms to the used two different grouping hierarchies: collection locations Brownian-motion assumption, and values above 1 indicate that within host species, and host species within locations. Statistical the phylogenetic signal exceeds the Brownian-motion significance of the variance components calculated from the assumption (Ives & Godfray, 2006). By comparing mean square AMOVAs were determined with 10,000 permutations. errors, one can evaluate the overall fit of the full model, in

datasets were analyzed using a Mantel test in PC-ORD v. 5.0 (MSEstar) and to a model that assumes Brownian-motion

(McCune and Mefford, 2006). evolution (MSEb). Lower MSE values indicate a better fit of the specific model. In both studies, 95% confidence intervals were obtained by bootstrapping either 500 (III) or 1,000 (IV) times. Several additional statistical analyses were performed in study IV. Effects of host and location on rates of survival and

31 2012; I). If plants and herbivores have cospeciated, then their 3.5 ANALYSES OF TRITROPHIC INTERACTIONS phylogenies should be mirror images of each other, and their diversification times should be concurrent. Comparing One method to analyze tritrophic interactions is to use Ives and diversification times of insects and hosts provides valuable Godfray's (2006) phylogenetic bipartite linear model (pblm), information on insect radiation (Lopez-Vaamonde et al., 2006; which can be found in the picante v. 1.2 package (Kembel et al., Gómez-Zurita et al., 2007; Ohshima et al., 2010; I). Insect 2010) in R v. 2.10 (R Development Core Team 2009). Pblm was phylogenies can also be used to illustrate, for example, used in studies III and IV, because the model can be used to evolution of larval feeding habits (I). Larval feeding and host- estimate how well ecological versus phylogenetic factors explain plant use were reconstructed in Mesquite using the “Trace parasitoid–insect associations. For this analysis, one has to character over trees” option, with either parsimony or construct a phylogeny for all studied levels. A binary (parasitoid maximum-likelihood optimization (Maddison & Maddison, present / absent) association matrix was constructed for 25 2010). Before the reconstruction, each miner was coded heterarthrine species for which information on parasitoids was according to its host-plant genera or larval feeding habit. An ML available (59 species) (Fig 2; III) (Pschorn-Walcher & Altenhofer, tree was used in larval feeding reconstruction in the tree 1989; Digweed et al., 2009). A similar matrix was constructed for including leaf-mining tribe Pseudodineura, which was seven galler species with 14 barcode-identified parasitoids, but constructed using maximum-likelihood optimization (Fig. 3). this time with quantitative data (IV). In the leafminer study (III), Genetic variation among and within populations can be we also created a “pseudophylogeny” for the miners based on analyzed in ARLEQUIN v. 3.5 (Excoffier & Lischer, 2010). In the the host-plant phylogeny. First, plant species not used by miners galler study (II) we used ARLEQUIN to estimate pairwise ΦST were removed from the full plant phylogeneny. Then miners values among population samples collected from different hosts were added onto the tree according to their host genera, and locations, and also performed a hierarchical analysis of maeaning thst miners using same plant genera as a host created molecular variance (AMOVA) (Excoffier et al., 1992). These a polytomy. The signal-strength parameter (d) measures the analyses were performed separately for mitochondrial and how well the phylogenetic trees explain associations: a value of nuclear genes. Only the six willow species that were shared by 0 indicates absence of phylogenetic signal, values from 0 to 1 both galler genera were used. In the hierarchical AMOVAs, we represent stabilizing selection, a value 1 conforms to the used two different grouping hierarchies: collection locations Brownian-motion assumption, and values above 1 indicate that within host species, and host species within locations. Statistical the phylogenetic signal exceeds the Brownian-motion significance of the variance components calculated from the assumption (Ives & Godfray, 2006). By comparing mean square AMOVAs were determined with 10,000 permutations. errors, one can evaluate the overall fit of the full model, in

Comparisons of ΦST values across host-species pairs, and the which d parameters are estimated from the data (MSEd), in existence of a correlation between the Pontania and Euura relation to a model that assumes no phylogenetic covariances datasets were analyzed using a Mantel test in PC-ORD v. 5.0 (MSEstar) and to a model that assumes Brownian-motion

31 parasitism were tested with two-way ANOVA in IBM SPSS v. 21 (IBM SPSS statistics Inc. Chicago, IL, USA). In comparisons of 4. Results and Discussion enemy species richness across community samples (galler × willow × location) and hosts (galler × willow), sampling effort was standardized by constructing species-accumulation curves 4.1 ROLE OF PLANTS IN INSECT SPECIATION in EstimateS v. 9.0.0 (Colwell, 2013). The effect of host species on enemy-community richness was tested with a nonparametric Kruskal–Wallis one-way ANOVA, and the effect of habitat with 4.1.1 Leaf-mining sawflies a Mann–Whitney U-test in SPSS. Non-metric multidimensional Leaf-mining has been thought to have evolved at least three scaling (NMDS) ordination in PC-Ord v. 5.33 (McCune and times in the sawfly family Tenthredinidae, i.e., in the tribes Mefford, 2006) was used to visualize parasitoid community Pseudodineurini (Nematinae), Fenusini (Blennocampinae), and structure. Multi-response permutation procedures (MRPP) in Heterarthrini (Heterarthrinae). However, there has been a PC-Ord were used to test the effects of location, willow species, continuous debate over whether Fenusini in fact belongs to the and habitat (boreal–subarctic / arctic–alpine) on parasitoid Heterarthrinae or not (Goulet, 1992; Pschorn-Walcher & communities. Permutational multivariate analysis of variance Altenhofer, 1989; Smith, 1971; Taeger et al., 2010). Some authors (PERMANOVA) in PRIMER v. 6.1.15 with the PERMANOVA+ have considered that the subfamily Heterarthrinae includes the v. 1.0.5 add-on package (Clarke & Gorley, 2006), was also used tribes Heterarthrini, Fenusini, and Caliroini, of which Caliroini to test the effects of host and location. contains the leaf-skeletonizer genera Caliroa and Endelomyia (Smith, 1971; Goulet, 1992; Viitasaari, 2002). By contrast, others have proposed that heterarthrine leafminers are polyphyletic, so that the habit of leaf-mining would have evolved independently in the genus Heterarthrus and in the tribe Fenusini, of which the latter would belong to the ecologically diverse subfamily Blennocampinae (Pschorn-Walcher & Altenhofer, 1989). Our phylogenetic analyses show that the heterarthrine leafminers form a monophyletic group that evolved c. 100 million years ago (I). Therefore, the traditional division to Fenusini and Heterarthrini is clearly not correct, as Heterarthrus is grouped inside the Fenusini (Fig. 3; I). The phylogenetic trees suggest that Heterarthrinae as currently defined is a polyphyletic group, because the leaf-skeletonizing genera Caliroa and Endelomyia are not closely related to each other or to the leafminers. However, a recent study with more genes shows that Caliroa and Endelomyia are closely related, but are grouped together with external- feeding blennocampines (Malm & Nyman, 2014). Reconstructions of ancestral larval feeding modes indicate that shifts from external to internal feeding (leaf-mining, gall

33 parasitism were tested with two-way ANOVA in IBM SPSS v. 21 (IBM SPSS statistics Inc. Chicago, IL, USA). In comparisons of 4. Results and Discussion enemy species richness across community samples (galler × willow × location) and hosts (galler × willow), sampling effort was standardized by constructing species-accumulation curves 4.1 ROLE OF PLANTS IN INSECT SPECIATION in EstimateS v. 9.0.0 (Colwell, 2013). The effect of host species on enemy-community richness was tested with a nonparametric Kruskal–Wallis one-way ANOVA, and the effect of habitat with 4.1.1 Leaf-mining sawflies a Mann–Whitney U-test in SPSS. Non-metric multidimensional Leaf-mining has been thought to have evolved at least three scaling (NMDS) ordination in PC-Ord v. 5.33 (McCune and times in the sawfly family Tenthredinidae, i.e., in the tribes Mefford, 2006) was used to visualize parasitoid community Pseudodineurini (Nematinae), Fenusini (Blennocampinae), and structure. Multi-response permutation procedures (MRPP) in Heterarthrini (Heterarthrinae). However, there has been a PC-Ord were used to test the effects of location, willow species, continuous debate over whether Fenusini in fact belongs to the and habitat (boreal–subarctic / arctic–alpine) on parasitoid Heterarthrinae or not (Goulet, 1992; Pschorn-Walcher & communities. Permutational multivariate analysis of variance Altenhofer, 1989; Smith, 1971; Taeger et al., 2010). Some authors (PERMANOVA) in PRIMER v. 6.1.15 with the PERMANOVA+ have considered that the subfamily Heterarthrinae includes the v. 1.0.5 add-on package (Clarke & Gorley, 2006), was also used tribes Heterarthrini, Fenusini, and Caliroini, of which Caliroini to test the effects of host and location. contains the leaf-skeletonizer genera Caliroa and Endelomyia (Smith, 1971; Goulet, 1992; Viitasaari, 2002). By contrast, others have proposed that heterarthrine leafminers are polyphyletic, so that the habit of leaf-mining would have evolved independently in the genus Heterarthrus and in the tribe Fenusini, of which the latter would belong to the ecologically diverse subfamily Blennocampinae (Pschorn-Walcher & Altenhofer, 1989). Our phylogenetic analyses show that the heterarthrine leafminers form a monophyletic group that evolved c. 100 million years ago (I). Therefore, the traditional division to Fenusini and Heterarthrini is clearly not correct, as Heterarthrus is grouped inside the Fenusini (Fig. 3; I). The phylogenetic trees suggest that Heterarthrinae as currently defined is a polyphyletic group, because the leaf-skeletonizing genera Caliroa and Endelomyia are not closely related to each other or to the leafminers. However, a recent study with more genes shows that Caliroa and Endelomyia are closely related, but are grouped together with external- feeding blennocampines (Malm & Nyman, 2014). Reconstructions of ancestral larval feeding modes indicate that shifts from external to internal feeding (leaf-mining, gall

33 100 Scolioneura betuleti (6i) 94 100 Scolioneura vicina (S066) 53 Scolioneura tirolensis (2k) Larval feeding habit Scolioneura sp. (S075) confirms that the leaf-mining habit arose independently in 100 Fenusella nana (S036) External feeder 100 Fenusella septentrionalis (3i) 100 Fenusella wuestneii (1i) Leaf miner 100 Fenusella hortulana (S033) Pseudodineura. 46 Fenusella glaucopis (S034) 83 Fenella nigrita (1j) Gall inducer 100 63 Fenusella sp. (5i) 100 Fenusa dohrnii (S038) Shoot borer 39 Fenusa pumila (S037) Heterarthrine species are said to be very specific in their use Fenusa ulmi (4j) Leaf roller 54 Profenusa japonica (S073) 70 Parna reseri (9j) Fruit miner 100 Parna kamijoi (Xr) 44 99 Parna tenella (8i) of host plants (Altenhofer, 2003; Smith, 1979; Taeger & Petiole miner 100 Hinatara sp. cf. excisa (2i) Hinatara recta (9i) Heterarthrinae 96 Profenusa pygmaea (Xj) 100 Profenusa thomsoni (4k) Altenhofer, 1998). However, our phylogeny-based analyses Profenusa lucifex (S064) 100 Heterathrus imbrosensis (S077) 45 100 Heterarthrus healyi (8k) 100 Heterarthrus aceris (S021) 78 74 Heterarthrus cuneifrons (S061) indicate that host-plant use in the group has not been stable on 49 Heterarthrus leucomela (9k) Heterarthus nemoratus (P7) 100 Heterarthrus vagans (S022) 24 Heterarthrus microcephalus (S024) longer time scales, because multiple shifts to distantly related 79 100 Heterarthrus ochropoda (Xk) 100 Metallus lanceolatus (S063) 86 100 Metallus pumilus (S035) 45 Metallus rohweri (S074) Metallus albipes (4i) plant lineages have occurred (I). Comparing the miner and plant 100 8 Notofenusa sp. (Xi) Notofenusa surosa (3k) 14 Phymatocera aterrima (P4) Silliana lhommei (S082) 18 24 Caliroa sp. (N7) phylogenies shows that the miners have radiated after their host Hoplocampoides xylostei (S039) 51 84 Ardis brunniventris (N8) 94 22 Periclista sp. (S080) Blennocampa phyllocolpa (S026) plants, because most of the miner nodes are younger than 49 Endelomyia aethiops (MX) Siobla ruficornis (S040) 18 93 Tenthredo notha (L1) Pachyprotasis rapae (S030) 16 73 Nesoselandria morio (S079) corresponding nodes in the phylogenetic tree of their host Strongylogaster tacita (P8) Athalia circularis (D2) 59 Priophorus pallipes (K6) 100 42 Trichiocampus aeneus (J7) Cladius comari (JX) plants. Our miner results join an increasing group of studies that 79 Susana annulata (H8) 48 100 Hoplocampa marlatti (J5) Hoplocampa oregonensis (G3) Craterocercus fraternalis (G7) have shown that diversification times of herbivore insects 35 97 Eitelius gregarius (E4) 100 Pontania pustulator (AX) Euurina 92 Amauronematus amplus (A5) 100 Mesoneura opaca (G6) Pristiphora erichsonii (K7) postdate those of their host plants (Lopez-Vaamonde et al., 2006; 71 100 Stauronematus platycerus (7s) 14 Stauronematus compressicornis (E5) 53 100 Hemichroa crocea (J8) 56 Hemichroa australis (1s) 39 Platycampus luridiventris (K2) Gómez-Zurita et al., 2007; Hunt et al., 2007; McKenna et al., 100 Hemicroa militaris cf. militaris (LX) 23 Hemichroa militaris cf. thoracicus (P1) 95 Dineura pullior (4s) 100 100 Nematinus steini (2s) Nematinus acuminatus (H3) 2009; Ohshima et al., 2010). Heterarthrine host shifts often occur 31 Fallocampus americanus (P2) Anoplonyx apicalis (H1) 66 Pseudodineura parvula (S225) 88 Pseudodineura heringi (6s) among few plant genera (mainly Betula, Alnus, Salix, and 100 Pseudodineura parvula (S227) 69 Pseudodineura parvula (S226) 85 Pseudodineura enslini (5s) 100 100 24 Pseudodineura enslini (S220) 35 Pseudodineura enslini (S219) Populus) that are not closely related, and this pattern likewise 100 Pseudodineura mentiens (S224) Pseudodineura mentiens (6r) 0.2 35 100 Pseudodineura parva (N5) 52 Pseudodineura parva (S228) shows similarities with other herbivore insect studies (Lopez- 100 Pseudodineura clematidis (S216) Pseudodineurini 100 Pseudodineura clematidis (S215) 100 100 Pseudodineura clematidisrectae (S218) Pseudodineura clematidisrectae (S217) 91 100 Pseudodineura fuscula (S222) Vaamonde et al., 2003; Nyman et al., 2006a; 2010; Scheffer et al., 87 100 100 Pseudodineura fuscula (S221) Pseudodineura fuscula (S223) 100 Pseudodineura fuscula (J2) 100 Endophytus anemones (S229) Endophytus anemones (N4) 2007; Mardulyn et al., 2011). The frequent shifts among plant 32 67 Kerita fidala (4r) 100 Caulocampus acericaulis (F7) Caulocampus matthewsi (J6) 55 Eriocampa ovata (N9) taxa provide evidence that host plants have a major role in Empria fletcheri (S069) 100 Abia candens (L2) Trichiosoma aenescens (S090) Diprion similis (L4) 81 Arge sp. (S086) diversification of herbivorous insects, and thus can drive 93 Sterictiphora sp. (M8) Lophyrotoma analis (S084) Blasticotoma filiceti (S085) herbivore specialization onto alternative host plants (Ehrlich & 0.2 Figure 3. Phylogeny of leaf-mining sawflies from the subfamily Heterarthrinae and the Raven, 1964; Winkler & Mitter, 2008; Fordyce, 2010; Slove & nematine tribe Pseudodineurini. Selected outgroup taxa are from Tenthredinidae and Janz, 2011). five other tenthredinoid families (see inset for full branch lengths). Numbers above branches are bootstrap proportions (%) from a maximum-likelihood analysis in RAxML. Branch colors show ancestral larval feeding habits reconstructed using ML 4.1.2 Gall-inducing sawflies optimization. Nodes and branches were colored with a particular state (see legend) when the likelihood of the most likely larval feeding habit for the node exceeded 99 %, Our phylogenetic analyses revealed clear clustering of otherwise a pie chart is given for the node. An arrow shows the place of Euurina individuals based on willow host species in both Pontania and species, which were used in studies II and IV. Euura (II), but there were no cases in which all specimens from a

single willow species would have formed a monophyletic induction, shoot-boring, and leaf-rolling) have occurred several group. However, individuals from the same host were more times in the Tenthredinoidea (Fig. 3; I). Within Tenthredinidae, likely to cluster together than with specimens from other host two shifts from external feeding to leaf-mining have taken place, species. There are conflicts between the COI and ITS2 trees, in the Heterarthrinae and in the tribe Pseudodineurini. All which may result from of interbreeding if the species are not Pseudodineura samples cluster with representatives of the completely reproductively isolated from each other. subfamily Nematinae, and the ancestral-state reconstruction

35 100 Scolioneura betuleti (6i) 94 100 Scolioneura vicina (S066) 53 Scolioneura tirolensis (2k) Larval feeding habit Scolioneura sp. (S075) confirms that the leaf-mining habit arose independently in 100 Fenusella nana (S036) External feeder 100 Fenusella septentrionalis (3i) 100 Fenusella wuestneii (1i) Leaf miner 100 Fenusella hortulana (S033) Pseudodineura. 46 Fenusella glaucopis (S034) 83 Fenella nigrita (1j) Gall inducer 100 63 Fenusella sp. (5i) 100 Fenusa dohrnii (S038) Shoot borer 39 Fenusa pumila (S037) Heterarthrine species are said to be very specific in their use Fenusa ulmi (4j) Leaf roller 54 Profenusa japonica (S073) 70 Parna reseri (9j) Fruit miner 100 Parna kamijoi (Xr) 44 99 Parna tenella (8i) of host plants (Altenhofer, 2003; Smith, 1979; Taeger & Petiole miner 100 Hinatara sp. cf. excisa (2i) Hinatara recta (9i) Heterarthrinae 96 Profenusa pygmaea (Xj) 100 Profenusa thomsoni (4k) Altenhofer, 1998). However, our phylogeny-based analyses Profenusa lucifex (S064) 100 Heterathrus imbrosensis (S077) 45 100 Heterarthrus healyi (8k) 100 Heterarthrus aceris (S021) 78 74 Heterarthrus cuneifrons (S061) indicate that host-plant use in the group has not been stable on 49 Heterarthrus leucomela (9k) Heterarthus nemoratus (P7) 100 Heterarthrus vagans (S022) 24 Heterarthrus microcephalus (S024) longer time scales, because multiple shifts to distantly related 79 100 Heterarthrus ochropoda (Xk) 100 Metallus lanceolatus (S063) 86 100 Metallus pumilus (S035) 45 Metallus rohweri (S074) Metallus albipes (4i) plant lineages have occurred (I). Comparing the miner and plant 100 8 Notofenusa sp. (Xi) Notofenusa surosa (3k) 14 Phymatocera aterrima (P4) Silliana lhommei (S082) 18 24 Caliroa sp. (N7) phylogenies shows that the miners have radiated after their host Hoplocampoides xylostei (S039) 51 84 Ardis brunniventris (N8) 94 22 Periclista sp. (S080) Blennocampa phyllocolpa (S026) plants, because most of the miner nodes are younger than 49 Endelomyia aethiops (MX) Siobla ruficornis (S040) 18 93 Tenthredo notha (L1) Pachyprotasis rapae (S030) 16 73 Nesoselandria morio (S079) corresponding nodes in the phylogenetic tree of their host Strongylogaster tacita (P8) Athalia circularis (D2) 59 Priophorus pallipes (K6) 100 42 Trichiocampus aeneus (J7) Cladius comari (JX) plants. Our miner results join an increasing group of studies that 79 Susana annulata (H8) 48 100 Hoplocampa marlatti (J5) Hoplocampa oregonensis (G3) Craterocercus fraternalis (G7) have shown that diversification times of herbivore insects 35 97 Eitelius gregarius (E4) 100 Pontania pustulator (AX) Euurina 92 Amauronematus amplus (A5) 100 Mesoneura opaca (G6) Pristiphora erichsonii (K7) postdate those of their host plants (Lopez-Vaamonde et al., 2006; 71 100 Stauronematus platycerus (7s) 14 Stauronematus compressicornis (E5) 53 100 Hemichroa crocea (J8) 56 Hemichroa australis (1s) 39 Platycampus luridiventris (K2) Gómez-Zurita et al., 2007; Hunt et al., 2007; McKenna et al., 100 Hemicroa militaris cf. militaris (LX) 23 Hemichroa militaris cf. thoracicus (P1) 95 Dineura pullior (4s) 100 100 Nematinus steini (2s) Nematinus acuminatus (H3) 2009; Ohshima et al., 2010). Heterarthrine host shifts often occur 31 Fallocampus americanus (P2) Anoplonyx apicalis (H1) 66 Pseudodineura parvula (S225) 88 Pseudodineura heringi (6s) among few plant genera (mainly Betula, Alnus, Salix, and 100 Pseudodineura parvula (S227) 69 Pseudodineura parvula (S226) 85 Pseudodineura enslini (5s) 100 100 24 Pseudodineura enslini (S220) 35 Pseudodineura enslini (S219) Populus) that are not closely related, and this pattern likewise 100 Pseudodineura mentiens (S224) Pseudodineura mentiens (6r) 0.2 35 100 Pseudodineura parva (N5) 52 Pseudodineura parva (S228) shows similarities with other herbivore insect studies (Lopez- 100 Pseudodineura clematidis (S216) Pseudodineurini 100 Pseudodineura clematidis (S215) 100 100 Pseudodineura clematidisrectae (S218) Pseudodineura clematidisrectae (S217) 91 100 Pseudodineura fuscula (S222) Vaamonde et al., 2003; Nyman et al., 2006a; 2010; Scheffer et al., 87 100 100 Pseudodineura fuscula (S221) Pseudodineura fuscula (S223) 100 Pseudodineura fuscula (J2) 100 Endophytus anemones (S229) Endophytus anemones (N4) 2007; Mardulyn et al., 2011). The frequent shifts among plant 32 67 Kerita fidala (4r) 100 Caulocampus acericaulis (F7) Caulocampus matthewsi (J6) 55 Eriocampa ovata (N9) taxa provide evidence that host plants have a major role in Empria fletcheri (S069) 100 Abia candens (L2) Trichiosoma aenescens (S090) Diprion similis (L4) 81 Arge sp. (S086) diversification of herbivorous insects, and thus can drive 93 Sterictiphora sp. (M8) Lophyrotoma analis (S084) Blasticotoma filiceti (S085) herbivore specialization onto alternative host plants (Ehrlich & 0.2 Figure 3. Phylogeny of leaf-mining sawflies from the subfamily Heterarthrinae and the Raven, 1964; Winkler & Mitter, 2008; Fordyce, 2010; Slove & nematine tribe Pseudodineurini. Selected outgroup taxa are from Tenthredinidae and Janz, 2011). five other tenthredinoid families (see inset for full branch lengths). Numbers above branches are bootstrap proportions (%) from a maximum-likelihood analysis in RAxML. Branch colors show ancestral larval feeding habits reconstructed using ML 4.1.2 Gall-inducing sawflies optimization. Nodes and branches were colored with a particular state (see legend) when the likelihood of the most likely larval feeding habit for the node exceeded 99 %, Our phylogenetic analyses revealed clear clustering of otherwise a pie chart is given for the node. An arrow shows the place of Euurina individuals based on willow host species in both Pontania and species, which were used in studies II and IV. Euura (II), but there were no cases in which all specimens from a

35 Hybridization and introgression is often found in insect opportunities have been quite similar, it could be expected that

herbivores (Gompert et al., 2008; Linnen & Farrell, 2007; their speciation patterns would also be similar. However, ΦST Mardulyn et al., 2011; Marshall et al., 2011). Euura bud gallers estimates in these galler groups reveal that differentiation in form at least three clear clusters (S. lapponum, S. lanata + S. nuclear ITS2 is higher in Pontania than in Euura, while the glauca, and S. phylicifolia + S. myrsinifolia) (Nyman, 2002)(II). average levels of differentiation in mitochondrial COI are almost

However, the results also suggest that Euura individuals using the same in both taxa (II). Comparing ΦST estimates across S. hastata as a host form their own species, which would shared willow host species in the Pontania and Euura samples conform to E. hastatae in the sense of Malaise (1920). Pontania revealed no evidence of a correlation in the level of HAD in individuals using S. myrsinifolia and S. borealis are intermixed in either of the genes used. Even though these galler genera live in the trees, which questions the decision to divide leaf gallers on a similar niche environment consisting of Salix species, they these willows into two species (P. varia and P. norvegica, differ in how they experience their host species. Reasons for this respectively) as suggested by Kopelke (1999). Overall, the could include, for example, different chemistry in leaves versus phylogenies and AMOVA results (II; Tables 2 & 3) indicate that buds, and in the phenological availability of these tissues. In host plant has a greater influence on genetic differentiation in general, our results indicate that Euura and Pontania gallers fall gall-inducing sawflies than does geographic region (II). This somewhere between host races and fully independent species result resembles findings of several previous studies (Drès & Mallet, 2002; Peccoud et al., 2009; Powell et al., 2013), as (Drummond et al., 2010; Stireman et al., 2012). species boundaries are not completely developed. Evolution It has been suggested that endophagy promotes HAD, as does not seem to follow the same path even when internal-feeding insects have very close connection with their environmental circumstances are very similar. hosts (Stireman et al., 2005; Dorchin et al., 2009; Dickey & Medina, 2012). Interactions between gall-makers and plants are intensive, as the insects have to manipulate plant tissues to 4.3 TRITROPHIC INTERACTIONS produce the gall that they feed on. These hypotheses seem plausible, because HAD has been found in gall-inducing 4.3.1 The parasitoid community of leaf-mining sawflies sawflies (II), gall midges (Stireman et al., 2005; Dorchin et al., Heterarthrine leafminers are attacked by numerous 2009), gall-making flies (Stireman et al., 2005), and galling pecan hymenopteran parasitoids, but they have lost most of the leaf phylloxera (Dickey & Medina, 2012). However, more hymenopteran, and all dipteran, parasitoids that live on studies are needed from different feeding guilds to confirm this. external-feeding sawflies (Price & Pschorn-Walcher, 1988; Most studies on HAD have been done based on single insect Pschorn-Walcher & Altenhofer, 1989; Richter & Kasparyan, taxa and their hosts (Via et al., 2000; Peccoud et al., 2009; 2013). Some of the parasitoids are polyphagous (e.g., Colastes Dorchin et al., 2009; Hernández-Vera et al., 2010). While there braconius Haliday), and several Pnigalio and Chrysocharis species are a few studies that have broadened their focus to involve even use also other leaf-mining insect taxa as hosts (Godfray, several insect and plant groups (Stireman et al., 2005; Dickey & 1994; Noyes, 2013). Ichneumonids tend to be more specialized Medina, 2010; Egan et al. 2013; II), the repeatability of HAD has than other parasitoids (Pschorn-Walcher & Altenhofer, 1989), not been studied much. Although Pontania and Euura gallers and, in heterarthrine miners, three specialized ichneumonid share many willow host species, they are not sister taxa parasitoids inflicted the highest mortality (III). However, none (Nyman, 2000; 2007). Given that their evolutionary of the parasitoid species on heterarthrines are closely related to

37 Hybridization and introgression is often found in insect opportunities have been quite similar, it could be expected that herbivores (Gompert et al., 2008; Linnen & Farrell, 2007; their speciation patterns would also be similar. However, ΦST Mardulyn et al., 2011; Marshall et al., 2011). Euura bud gallers estimates in these galler groups reveal that differentiation in form at least three clear clusters (S. lapponum, S. lanata + S. nuclear ITS2 is higher in Pontania than in Euura, while the glauca, and S. phylicifolia + S. myrsinifolia) (Nyman, 2002)(II). average levels of differentiation in mitochondrial COI are almost

37 each other (Quicke et al., 2009). Thus, it seems that the miner system, both diversifying forces appear to have played a role. enemy complex has evolved gradually, as the parasitoids As in other tritrophic food webs, associations among plants, represent three families that are widely separated in the heterarthrine miners, and parasitoids have evolved through phylogenies, and also are shared with other herbivore taxa multiple complex mechanisms, which involved host shifts in (Gauthier et al., 2000; Zaldivar-Riverón et al., 2006; Pitz et al., both miners and parasitoids (Lopez-Vaamonde et al., 2005; 2007; Quicke et al., 2009;). In Heterarthrinae, closely related Zaldívar-Riverón et al., 2008). parasitoids tend to attack the same miner species, and the situation therefore differs from the study by Ives & Godfray 4.3.2 The parasitoid community of gall-inducing sawflies (2006) on leaf-mining Phyllonorycter moths and their enemies. Lately, molecular methods have become more popular also in The different results may arise from our wider taxon sampling, studies on multitrophic systems. Identification of parasitoids because the taxonomic scale may influence the phylogenetic based on morphology can be very slow and demanding and, signal arising from the hosts and parasitoids (Cagnolo et al., hence, cryptic species are often found when molecular methods 2011; Desneux et al., 2012). are used (Smith et al., 2006; 2008; 2011; Hayward et al., 2011; There are few studies indicating that diversification can Gebiola et al., 2012). Also rearing can give unreliable results, cascade, meaning that diversification of plants leads to especially if rearing success is low. In gallers, our reared adults diversification of herbivores, which then promotes grouped into 18 distinct clusters based on their DNA barcode diversification of parasitoids (Cronin & Abrahamson, 2001; sequences, and nearly all larval parasitoids could be indentified Althoff, 2008). The focal plants, leaf-mining sawflies, and to species level based on the reference dataset (IV). These results, parasitoids provided good opportunities for studying “bottom– along with a growing body of other studies, indicate that DNA up” and “top–down” diversification forces in multitrophic barcoding is an efficient tool for parasitoid community studies, networks. In miners, the presence of specialized ichneumonids and for indentifying species when morphological identification indicates that some “bottom–up” diversification might have is difficult (Smith et al., 2006; 2008; 2011; Kaartinen et al., 2010; happened in parasitoids following heterarthrine host shifts (III). Wirta et al., 2014). However, the possibility exists that Nevertheless, the existence of many polyphagous eulophid and endoparasitoids were lost during larval collecting, as we only braconid species makes things more complex. Even though the stored visible parasitoid eggs, pupae, and larvae. This may have possibility exists that some of the presumed generalists are a small influence on the obtained results, but it most likely cryptic specialists (cf. Smith et al., 2006; Smith et al., 2007; would not change the overall pattern that we found. The enemy Kaartinen et al., 2010; Hayward et al., 2011), it seems that community of Pontania gallers contained eight common species, “bottom–up” processes cannot completely explain and six species that were encountered only a few times each (IV). diversification in this system. Phylogeny-based analyses of All of the common parasitoids use multiple galler species as miner–parasitoid associations indicate that host plants have a hosts, but our NDMS ordination analyses revealed clear and stronger influence on these associations than does the statistically significant differences in the enemy communities on phylogeny of the miner species (III). If parasitoids are herbivore different galler and willow species, as well as a strong effect of niche specialists, they could influence insect diversification if the galler habitat (boreal–subarctic vs. arctic–alpine). A two-way herbivores can find enemy-free space on novel plant species PERMANOVA also found a relatively weak, but statistically (Bernays & Graham, 1988; Lill et al., 2002; Singer & Stireman, significant, effect for location. 2005). As there are niche- and species-specialist enemies in this

39 each other (Quicke et al., 2009). Thus, it seems that the miner system, both diversifying forces appear to have played a role. enemy complex has evolved gradually, as the parasitoids As in other tritrophic food webs, associations among plants, represent three families that are widely separated in the heterarthrine miners, and parasitoids have evolved through phylogenies, and also are shared with other herbivore taxa multiple complex mechanisms, which involved host shifts in (Gauthier et al., 2000; Zaldivar-Riverón et al., 2006; Pitz et al., both miners and parasitoids (Lopez-Vaamonde et al., 2005; 2007; Quicke et al., 2009;). In Heterarthrinae, closely related Zaldívar-Riverón et al., 2008). parasitoids tend to attack the same miner species, and the situation therefore differs from the study by Ives & Godfray 4.3.2 The parasitoid community of gall-inducing sawflies (2006) on leaf-mining Phyllonorycter moths and their enemies. Lately, molecular methods have become more popular also in The different results may arise from our wider taxon sampling, studies on multitrophic systems. Identification of parasitoids because the taxonomic scale may influence the phylogenetic based on morphology can be very slow and demanding and, signal arising from the hosts and parasitoids (Cagnolo et al., hence, cryptic species are often found when molecular methods 2011; Desneux et al., 2012). are used (Smith et al., 2006; 2008; 2011; Hayward et al., 2011; There are few studies indicating that diversification can Gebiola et al., 2012). Also rearing can give unreliable results, cascade, meaning that diversification of plants leads to especially if rearing success is low. In gallers, our reared adults diversification of herbivores, which then promotes grouped into 18 distinct clusters based on their DNA barcode diversification of parasitoids (Cronin & Abrahamson, 2001; sequences, and nearly all larval parasitoids could be indentified Althoff, 2008). The focal plants, leaf-mining sawflies, and to species level based on the reference dataset (IV). These results, parasitoids provided good opportunities for studying “bottom– along with a growing body of other studies, indicate that DNA up” and “top–down” diversification forces in multitrophic barcoding is an efficient tool for parasitoid community studies, networks. In miners, the presence of specialized ichneumonids and for indentifying species when morphological identification indicates that some “bottom–up” diversification might have is difficult (Smith et al., 2006; 2008; 2011; Kaartinen et al., 2010; happened in parasitoids following heterarthrine host shifts (III). Wirta et al., 2014). However, the possibility exists that Nevertheless, the existence of many polyphagous eulophid and endoparasitoids were lost during larval collecting, as we only braconid species makes things more complex. Even though the stored visible parasitoid eggs, pupae, and larvae. This may have possibility exists that some of the presumed generalists are a small influence on the obtained results, but it most likely cryptic specialists (cf. Smith et al., 2006; Smith et al., 2007; would not change the overall pattern that we found. The enemy Kaartinen et al., 2010; Hayward et al., 2011), it seems that community of Pontania gallers contained eight common species, “bottom–up” processes cannot completely explain and six species that were encountered only a few times each (IV). diversification in this system. Phylogeny-based analyses of All of the common parasitoids use multiple galler species as miner–parasitoid associations indicate that host plants have a hosts, but our NDMS ordination analyses revealed clear and stronger influence on these associations than does the statistically significant differences in the enemy communities on phylogeny of the miner species (III). If parasitoids are herbivore different galler and willow species, as well as a strong effect of niche specialists, they could influence insect diversification if the galler habitat (boreal–subarctic vs. arctic–alpine). A two-way herbivores can find enemy-free space on novel plant species PERMANOVA also found a relatively weak, but statistically (Bernays & Graham, 1988; Lill et al., 2002; Singer & Stireman, significant, effect for location. 2005). As there are niche- and species-specialist enemies in this

39 Parasitoids and parasitic inquilines are the most important speciation, then parasitoids could catch up herbivores on novel mortality factor in many galler species (Kopelke, 1999; Roininen plant species by forming host races, which may lead to further et al., 2005; Craig et al., 2007) and, thus, escape from enemies shifts in the herbivores. Such repeated host-associated shifts could provide a huge selective advantage. If parasitoids use may partly explain the extreme species diversity in these herbivore niches (e.g., plant species or feeding type / tissue interactions. type) as cues for finding the herbivores, then parasitoids could promote herbivore speciation. The observed strong effect of habitat on leaf-galler parasitoid communities indicates that, in addition to host-plant shifts, entering a new habitat may provide shelter for herbivores, and thereby accelerate herbivore speciation (IV). Parasitoids have a close connection to their host and are often host specific (Godfray, 1994; Condon et al., 2014). Thus host phylogeny might influence parasitoid speciation, and parasitoids could speciate in parallel with their hosts. However, phylogenetic studies indicate that the phylogenies of herbivores have relatively little influence on insect–parasitoid associations (Ives & Godfray, 2006; III; IV; but see Deng et al., 2013). Host shifts could also be an important factor influencing parasitoid diversification. A few studies have managed to show that HAD can cascade across trophic levels in plant–herbivore– parasitoid systems (Stireman et al., 2006; Forbes et al., 2009; Feder & Forbes, 2010), but others have not found HAD in insect–parasitoid associations (Cronin & Abrahamson, 2001; Althoff, 2008; Bilodeau et al., 2013). In our study (IV), the largest difference in enemy communities was found between galler species that inhabit the boreal–subarctic vs. arctic–alpine zones in the study areas. Thus, it seems that galler shifts into arctic– alpine habitats have led to speciation in at least two parasitoid groups, suggesting that habitat shifts could accelerate parasitoid diversification. When a herbivore escapes its enemies onto a novel plant species, parasitoids that manage to track the shift have to face a new environment with novel plant volatiles and morphology in order to find their hosts. They also have to overcome novel plant toxins ingested by the herbivores. Overcoming these obstacles may create divergent selection pressures in the parasitoids and promote HAD (Stireman et al., 2006). As parasitoids may be a driving force in herbivore

41 Parasitoids and parasitic inquilines are the most important speciation, then parasitoids could catch up herbivores on novel mortality factor in many galler species (Kopelke, 1999; Roininen plant species by forming host races, which may lead to further et al., 2005; Craig et al., 2007) and, thus, escape from enemies shifts in the herbivores. Such repeated host-associated shifts could provide a huge selective advantage. If parasitoids use may partly explain the extreme species diversity in these herbivore niches (e.g., plant species or feeding type / tissue interactions. type) as cues for finding the herbivores, then parasitoids could promote herbivore speciation. The observed strong effect of habitat on leaf-galler parasitoid communities indicates that, in addition to host-plant shifts, entering a new habitat may provide shelter for herbivores, and thereby accelerate herbivore speciation (IV). Parasitoids have a close connection to their host and are often host specific (Godfray, 1994; Condon et al., 2014). Thus host phylogeny might influence parasitoid speciation, and parasitoids could speciate in parallel with their hosts. However, phylogenetic studies indicate that the phylogenies of herbivores have relatively little influence on insect–parasitoid associations (Ives & Godfray, 2006; III; IV; but see Deng et al., 2013). Host shifts could also be an important factor influencing parasitoid diversification. A few studies have managed to show that HAD can cascade across trophic levels in plant–herbivore– parasitoid systems (Stireman et al., 2006; Forbes et al., 2009; Feder & Forbes, 2010), but others have not found HAD in insect–parasitoid associations (Cronin & Abrahamson, 2001; Althoff, 2008; Bilodeau et al., 2013). In our study (IV), the largest difference in enemy communities was found between galler species that inhabit the boreal–subarctic vs. arctic–alpine zones in the study areas. Thus, it seems that galler shifts into arctic– alpine habitats have led to speciation in at least two parasitoid groups, suggesting that habitat shifts could accelerate parasitoid diversification. When a herbivore escapes its enemies onto a novel plant species, parasitoids that manage to track the shift have to face a new environment with novel plant volatiles and morphology in order to find their hosts. They also have to overcome novel plant toxins ingested by the herbivores. Overcoming these obstacles may create divergent selection pressures in the parasitoids and promote HAD (Stireman et al., 2006). As parasitoids may be a driving force in herbivore

41 5 Conclusions

This thesis presents new information on the role of host plants in speciation of herbivore insects and their parasitoids. Insect– plant interactions are taxonomically unstable, and host shifts to closely related plant species are common. As found in miners, herbivores have usually radiated after their host plants. Thus, current insect–plant interactions have evolved on a background of already existing plant lineages that the miners have colonized, and a few of them have changed host plants at some point during their evolutionary history. Results from gall- inducing sawflies likewise indicate that host plants have a central role in herbivore speciation. Although HAD was found in both galler genera, there are differences in how the two galler taxa use their Salix host species. Evidently, the build-up of HAD is not repeatable, even though evolutionary opportunities for HAD and speciation have been similar in the focal galler groups. DNA barcoding is powerful tool for resolving the structure of parasitoid complexes when species identification is difficult based on morphology. Some parasitoids are species and others niche specialists, indicating that these associations are complex, and that both “bottom–up” and “top–down” diversification forces have played a role in the studied networks. Host-plant shifts could also affect parasitoid speciation, if parasitoids use plants as cues to find their host. At least in the studied subarctic and arctic–alpine habitats, habitat shifts in gallers have led to speciation in two parasitoid groups, and thus the habitat shift seems to have influenced parasitoid diversification. Our finding that both leafminer–parasitoid and galler–parasitoid associations are strongly influenced by the herbivore niches indicates that parasitoids have the potential to spun herbivore insect speciation by driving them into “enemy free space” provided by novel hosts and/or habitats.

43 5 Conclusions

6 References

Abrahamson WG, Blair CP, Eubanks MD & Morehead SA (2003) Sequential radiation of unrelated organisms: the gall fly Eurosta solidaginis and the tumbling flower beetle Mordellistena convicta. Journal of Evolutionary Biology 16:781– 789. Agosta SJ (2006) On ecological fitting, plant-insect associations, herbivore host shifts, and host plant selection. Oikos 114:556– 565. Agosta SJ, Janz N & Brooks DR (2010) How specialists can be generalists: resolving the" parasite paradox" and implications for emerging infectious disease. Zoologia 27:151–162. Agustí N, Bourguet D, Spataro T, Delos M, Eychenne N, Folcher L & Arditi R (2005) Detection, identification and geographical distribution of European corn borer larval parasitoids using molecular markers. Molecular Ecology 14:3267–74. Altenhofer E (2003) Minierende blattwespen (Hym.: Symphyta): ihre minenformen, wirtspflanzen, ökologie und biologie. Gredleriana 5–24. Althoff DM (2008) A test of host-associated differentiation across the 'parasite continuum' in the tri-trophic interaction among yuccas, bogus yucca moths, and parasitoids. Molecular Ecology 17:3917–3927. Althoff DM, Segraves KA & Johnson MT (2014) Testing for coevolutionary diversification: linking pattern with process. Trends in Ecology and Evolution 29:82–89. Argus GW (1997) Infrageneric classification of Salix (Salicaceae) in the new world. Systematic Botany Monographs 52:1–121. Askew RR & Shaw MR (1974) An account of the Chalcidoidea (Hymenoptera) parasitising leaf-mining insects of deciduous trees in Britain. Biological Journal of the Linnean Society 6:289– 335.

6 References

45 Berlocher SH & Feder JL (2002) Sympatric speciation in Craig TP, Itami JK & Horner JD (2007) Geographic variation in phytophagous insects: moving beyond controversy? Annual the evolution and coevolution of a tritrophic interaction. Review of Entomology 47:773–815. Evolution 61:1137–1152. Bernays E & Graham M (1988) On the evolution of host Crespi BJ, Carmean DA & Chapman TW (1997) Ecology and specificity in phytophagous arthropods. Ecology 69:886–892. evolution of galling thrips and their allies. Annual Review of Bilodeau E, Simon J-C, Guay J-F, Turgeon J & Cloutier C (2013) Entomology 42:51–71. Does variation in host plant association and symbiont Cronin JT & Abrahamson WG (2001) Do parasitoids diversify in infection of pea aphid populations induce genetic and response to host-plant shifts by herbivorous insects? behaviour differentiation of its main parasitoid, Aphidius ervi? Ecological Entomology 26:347–355. Evolutionary Ecology 27:165–184. Darriba D, Taboada GL, Doallo R & Posada D (2012) jModelTest Boevé JL, Sonet G, Nagy ZT, Symoens F, Altenhofer E, 2: more models, new heuristics and parallel computing. Häberlein C & Schulz S (2009) Defense by volatiles in leaf- Nature Methods 9:772–772. mining insect larvae. Journal of Chemical Ecology 35:507–517. Davis RB, Baldauf SL & Mayhew PJ (2010) The origins of species Borchsenius F (2009) FastGap 1.2. Available at: richness in the Hymenoptera: insights from a family-level http://www.aubot.dk/FastGap_home.htm. supertree. BMC Evolutionary Biology 10:109. Cagnolo L, Salvo A & Valladares G (2011) Network topology: De Moraes CM, Lewis WJ, Pare PW, Alborn HT & Tumlinson JH patterns and mechanisms in plant-herbivore and host- (1998) Herbivore-infested plants selectively attract parasitoid food webs. Journal of Animal Ecology 80:342–351. parasitoids. Nature 393:570–573. Clarke KR & Gorley RN (2006) PRIMER v6: User De Moraes CM & Mescher MC (2004) Biochemical crypsis in the Manual/Tutorial. PRIMER-E, Plymouth. avoidance of natural enemies by an insect herbivore. Coleman AW (2003) ITS2 is a double-edged tool for eukaryote Proceedings of the National Academy of Sciences of the USA evolutionary comparisons. Trends in Genetics 19:370–375. 101:8993–8997. Colwell RK (2013) Estimate S: Statistical estimation of species Deng J, Yu F, Li H-B, Gebiola M, Desdevises Y, Wu S-A & richness and shared species from samples. Version 9 or Zhang Y-Z (2013) Cophylogenetic relationships between earlier. User's guide and application. Anicetus parasitoids (Hymenoptera: Encyrtidae) and their http://viceroy.eeb.uconn.edu/estimates/index.html. scale insect hosts (Hemiptera: Coccidae). BMC Evolutionary Condon MA, Scheffer SJ, Lewis ML, Wharton R, Adams DC & Biology 13:275. Forbes AA (2014) Lethal interactions between parasites and Desneux N, Blahnik R, Delebecque CJ & Heimpel GE (2012) prey increase niche diversity in a tropical community. Science Host phylogeny and specialisation in parasitoids. Ecology 343:1240–1244. Letters 15:453–460. Connor EF & Taverner MP (1997) The evolution and adaptive Dickey AM & Medina RF (2010) Testing host-associated significance of the leaf-mining habit. Oikos 79:6–25. differentiation in a quasi-endophage and a parthenogen on Cook LG & Gullan PJ (2004) The gall-inducing habit has evolved native trees. Journal of Evolutionary Biology 23:945–56. multiple times among the eriococcid scale insects Dickey AM & Medina RF (2012) Host-associated genetic (Sternorrhyncha: Coccoidea: Eriococcidae). Biological Journal differentiation in pecan leaf phylloxera. Entomologia of the Linnean Society 83:441–452. Experimentalis et Applicata 143:127–137.

47 Berlocher SH & Feder JL (2002) Sympatric speciation in Craig TP, Itami JK & Horner JD (2007) Geographic variation in phytophagous insects: moving beyond controversy? Annual the evolution and coevolution of a tritrophic interaction. Review of Entomology 47:773–815. Evolution 61:1137–1152. Bernays E & Graham M (1988) On the evolution of host Crespi BJ, Carmean DA & Chapman TW (1997) Ecology and specificity in phytophagous arthropods. Ecology 69:886–892. evolution of galling thrips and their allies. Annual Review of Bilodeau E, Simon J-C, Guay J-F, Turgeon J & Cloutier C (2013) Entomology 42:51–71. Does variation in host plant association and symbiont Cronin JT & Abrahamson WG (2001) Do parasitoids diversify in infection of pea aphid populations induce genetic and response to host-plant shifts by herbivorous insects? behaviour differentiation of its main parasitoid, Aphidius ervi? Ecological Entomology 26:347–355. Evolutionary Ecology 27:165–184. Darriba D, Taboada GL, Doallo R & Posada D (2012) jModelTest Boevé JL, Sonet G, Nagy ZT, Symoens F, Altenhofer E, 2: more models, new heuristics and parallel computing. Häberlein C & Schulz S (2009) Defense by volatiles in leaf- Nature Methods 9:772–772. mining insect larvae. Journal of Chemical Ecology 35:507–517. Davis RB, Baldauf SL & Mayhew PJ (2010) The origins of species Borchsenius F (2009) FastGap 1.2. Available at: richness in the Hymenoptera: insights from a family-level http://www.aubot.dk/FastGap_home.htm. supertree. BMC Evolutionary Biology 10:109. Cagnolo L, Salvo A & Valladares G (2011) Network topology: De Moraes CM, Lewis WJ, Pare PW, Alborn HT & Tumlinson JH patterns and mechanisms in plant-herbivore and host- (1998) Herbivore-infested plants selectively attract parasitoid food webs. Journal of Animal Ecology 80:342–351. parasitoids. Nature 393:570–573. Clarke KR & Gorley RN (2006) PRIMER v6: User De Moraes CM & Mescher MC (2004) Biochemical crypsis in the Manual/Tutorial. PRIMER-E, Plymouth. avoidance of natural enemies by an insect herbivore. Coleman AW (2003) ITS2 is a double-edged tool for eukaryote Proceedings of the National Academy of Sciences of the USA evolutionary comparisons. Trends in Genetics 19:370–375. 101:8993–8997. Colwell RK (2013) Estimate S: Statistical estimation of species Deng J, Yu F, Li H-B, Gebiola M, Desdevises Y, Wu S-A & richness and shared species from samples. Version 9 or Zhang Y-Z (2013) Cophylogenetic relationships between earlier. User's guide and application. Anicetus parasitoids (Hymenoptera: Encyrtidae) and their http://viceroy.eeb.uconn.edu/estimates/index.html. scale insect hosts (Hemiptera: Coccidae). BMC Evolutionary Condon MA, Scheffer SJ, Lewis ML, Wharton R, Adams DC & Biology 13:275. Forbes AA (2014) Lethal interactions between parasites and Desneux N, Blahnik R, Delebecque CJ & Heimpel GE (2012) prey increase niche diversity in a tropical community. Science Host phylogeny and specialisation in parasitoids. Ecology 343:1240–1244. Letters 15:453–460. Connor EF & Taverner MP (1997) The evolution and adaptive Dickey AM & Medina RF (2010) Testing host-associated significance of the leaf-mining habit. Oikos 79:6–25. differentiation in a quasi-endophage and a parthenogen on Cook LG & Gullan PJ (2004) The gall-inducing habit has evolved native trees. Journal of Evolutionary Biology 23:945–56. multiple times among the eriococcid scale insects Dickey AM & Medina RF (2012) Host-associated genetic (Sternorrhyncha: Coccoidea: Eriococcidae). Biological Journal differentiation in pecan leaf phylloxera. Entomologia of the Linnean Society 83:441–452. Experimentalis et Applicata 143:127–137.

47 Digweed SC, MacQuarrie CJK, Langor DW, Williams DJM, DNA haplotypes: application to human mitochondrial DNA Spence JR, Nystrom KL & Morneau L (2009) Current Status restriction data. Genetics 131:479–491. of invasive alien birch-leafmining sawflies (Hymenoptera: Farrell BD (1998) "Inordinate fondness" explained: why are there Tenthredinidae) in Canada, with keys to species. Canadian so many beetles? Science 281:555–559. Entomologist 141:201–235. Farrell BD (2001) Evolutionary assembly of the milkweed fauna: Dinca V, Zakharov EV, Hebert PD & Vila R (2010) Complete cytochrome oxidase I and the age of Tetraopes beetles. DNA barcode reference library for a country's butterfly fauna Molecular Phylogenetic and Evolution18:467–78. reveals high performance for temperate Europe. Proceedings Farrell BD, Dussourd DE & Mitter C (1991) Escalation of plant of the Royal Society B: Biological Sciences 278:347–355. defense: do latex and resin canals spur plant diversification? Dorchin N, Scott ER, Clarkin CE, Luongo MP, Jordan S & American Naturalist 138:881–900. Abrahamson WG (2009) Behavioural, ecological and genetic Feder JL & Forbes AA (2010) Sequential speciation and the evidence confirm the occurrence of host-associated diversity of parasitic insects. Ecological Entomology 35:67–76. differentiation in goldenrod gall-midges. Journal of Fitch WM (1971) Toward defining the course of evolution: Evolutionary Biology 22:729–739. minimum change for a specific tree topology. Systematic Drès M & Mallet J (2002) Host races in plant-feeding insects and Biology 20:406–416. their importance in sympatric speciation. Philosophical Forbes AA, Powell TH, Stelinski LL, Smith JJ & Feder JL (2009) Transactions of the Royal Society B: Biological Sciences 357:471– Sequential sympatric speciation across trophic levels. Science 92. 323:776–779. Drummond AJ & Rambaut A (2007) BEAST: Bayesian Fordyce JA (2010) Host shifts and evolutionary radiations of evolutionary analysis by sampling trees. BMC Evolutionary butterflies. Proceedings of the Royal Society B: Biological Sciences Biology 7:214. 277:3735–3743. Drummond CS, Xue HJ, Yoder JB & Pellmyr O (2010) Host- Funk DJ, Filchak KE & Feder JL (2002) Herbivorous insects: associated divergence and incipient speciation in the yucca model systems for the comparative study of speciation moth Prodoxus coloradensis (Lepidoptera: Prodoxidae) on ecology. Genetica 116:251–267. three species of host plants. Heredity 105:183–196. Gauthier N, Lasalle J, Quicke DLJ & Godfray HCJ (2000) Egan SP, Hood GR, DeVela G & Ott JR (2013) Parallel patterns of Phylogeny of Eulophidae (Hymenoptera: Chalcidoidea), with morphological and behavioral variation among host- a reclassification of Eulophinae and the recognition that associated populations of two gall wasp species. PLoS ONE Elasmidae are derived eulophids. Systematic Entomology 8:e54690. 25:521–539. Ehrlich PR & Raven PH (1964) Butterflies and plants: a study in Gebiola M, Gómez-Zurita J, Monti MM, Navone P & Bernardo U coevolution. Evolution 18:586–608. (2012) Integration of molecular, ecological, morphological Excoffier L & Lischer HE (2010) Arlequin suite ver 3.5: a new and endosymbiont data for species delimitation within the series of programs to perform population genetics analyses Pnigalio soemius complex (Hymenoptera: Eulophidae). under Linux and Windows. Molecular Ecology Resources Molecular Ecology 21:1190–1208. 10:564–567. Godfray HC (1994) Parasitoids: Behavioral and Evolutionary Excoffier L, Smouse PE & Quattro JM (1992) Analysis of Ecology. Princeton University Press, Princeton. molecular variance inferred from metric distances among

49 Digweed SC, MacQuarrie CJK, Langor DW, Williams DJM, DNA haplotypes: application to human mitochondrial DNA Spence JR, Nystrom KL & Morneau L (2009) Current Status restriction data. Genetics 131:479–491. of invasive alien birch-leafmining sawflies (Hymenoptera: Farrell BD (1998) "Inordinate fondness" explained: why are there Tenthredinidae) in Canada, with keys to species. Canadian so many beetles? Science 281:555–559. Entomologist 141:201–235. Farrell BD (2001) Evolutionary assembly of the milkweed fauna: Dinca V, Zakharov EV, Hebert PD & Vila R (2010) Complete cytochrome oxidase I and the age of Tetraopes beetles. DNA barcode reference library for a country's butterfly fauna Molecular Phylogenetic and Evolution18:467–78. reveals high performance for temperate Europe. Proceedings Farrell BD, Dussourd DE & Mitter C (1991) Escalation of plant of the Royal Society B: Biological Sciences 278:347–355. defense: do latex and resin canals spur plant diversification? Dorchin N, Scott ER, Clarkin CE, Luongo MP, Jordan S & American Naturalist 138:881–900. Abrahamson WG (2009) Behavioural, ecological and genetic Feder JL & Forbes AA (2010) Sequential speciation and the evidence confirm the occurrence of host-associated diversity of parasitic insects. Ecological Entomology 35:67–76. differentiation in goldenrod gall-midges. Journal of Fitch WM (1971) Toward defining the course of evolution: Evolutionary Biology 22:729–739. minimum change for a specific tree topology. Systematic Drès M & Mallet J (2002) Host races in plant-feeding insects and Biology 20:406–416. their importance in sympatric speciation. Philosophical Forbes AA, Powell TH, Stelinski LL, Smith JJ & Feder JL (2009) Transactions of the Royal Society B: Biological Sciences 357:471– Sequential sympatric speciation across trophic levels. Science 92. 323:776–779. Drummond AJ & Rambaut A (2007) BEAST: Bayesian Fordyce JA (2010) Host shifts and evolutionary radiations of evolutionary analysis by sampling trees. BMC Evolutionary butterflies. Proceedings of the Royal Society B: Biological Sciences Biology 7:214. 277:3735–3743. Drummond CS, Xue HJ, Yoder JB & Pellmyr O (2010) Host- Funk DJ, Filchak KE & Feder JL (2002) Herbivorous insects: associated divergence and incipient speciation in the yucca model systems for the comparative study of speciation moth Prodoxus coloradensis (Lepidoptera: Prodoxidae) on ecology. Genetica 116:251–267. three species of host plants. Heredity 105:183–196. Gauthier N, Lasalle J, Quicke DLJ & Godfray HCJ (2000) Egan SP, Hood GR, DeVela G & Ott JR (2013) Parallel patterns of Phylogeny of Eulophidae (Hymenoptera: Chalcidoidea), with morphological and behavioral variation among host- a reclassification of Eulophinae and the recognition that associated populations of two gall wasp species. PLoS ONE Elasmidae are derived eulophids. Systematic Entomology 8:e54690. 25:521–539. Ehrlich PR & Raven PH (1964) Butterflies and plants: a study in Gebiola M, Gómez-Zurita J, Monti MM, Navone P & Bernardo U coevolution. Evolution 18:586–608. (2012) Integration of molecular, ecological, morphological Excoffier L & Lischer HE (2010) Arlequin suite ver 3.5: a new and endosymbiont data for species delimitation within the series of programs to perform population genetics analyses Pnigalio soemius complex (Hymenoptera: Eulophidae). under Linux and Windows. Molecular Ecology Resources Molecular Ecology 21:1190–1208. 10:564–567. Godfray HC (1994) Parasitoids: Behavioral and Evolutionary Excoffier L, Smouse PE & Quattro JM (1992) Analysis of Ecology. Princeton University Press, Princeton. molecular variance inferred from metric distances among

49 Gompert Z, Forister ML, Fordyce JA & Nice CC (2008) fulgerator. Proceedings of the National Academy of Sciences of the Widespread mito-nuclear discordance with evidence for USA 101:14812–14817. introgressive hybridization and selective sweeps in Lycaeides. Hendry AP, Huber SK, De León LF, Herrel A & Podos J (2009) Molecular Ecology 17:5231–44. Disruptive selection in a bimodal population of Darwin's Goulet H (1992) The genera and subgenera of the sawflies of finches Proceedings of the Royal Society B: Biological Sciences Canada and Alaska (Hymenoptera: Symphyta). Insects and 276:753–759. Arachnids of Canada 20:1–235. Hernández-Vera G, Mitrović M, Jović J, Tosevski I, Caldara R, Goulet H & Huber JT (1993) Hymenoptera of the World: An Gassmann A & Emerson BC (2010) Host-associated genetic Identification Guide to Families. Research Branch Agriculture differentiation in a seed parasitic weevil Rhinusa antirrhini Canada Publication, Ottawa, Ontario. (Coleptera: Curculionidae) revealed by mitochondrial and Gómez-Zurita J, Hunt T, Kopliku F & Vogler AP (2007) nuclear sequence data. Molecular Ecology 19:2286–2300. Recalibrated tree of leaf beetles (Chrysomelidae) indicates Hespenheide HA (1991) Bionomics of leaf-mining insects. independent diversification of angiosperms and their insect Annual Review of Entomology 36:535–560. herbivores. PLoS ONE 2:e360. Hrcek J, Miller SE, Quicke DL & Smith MA (2011) Molecular Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W & detection of trophic links in a complex insect host-parasitoid Gascuel O (2010) New algorithms and methods to estimate food web. Molecular Ecology Resources 11:786–794. maximum-likelihood phylogenies: assessing the performance Hunt T, Bergsten J, Levkanicova Z, Papadopoulou A, John OS, of PhyML 3.0. Systematic Biology 59:307–321. Wild R, Hammond PM, Ahrens D, Balke M, Caterino MS, Hajibabaei M, Janzen DH, Burns JM, Hallwachs W & Hebert PD Gómez-Zurita J, Ribera I, Barraclough TG, Bocakova M, (2006) DNA barcodes distinguish species of tropical Bocak L & Vogler AP (2007) A comprehensive phylogeny of Lepidoptera. Proceedings of the National Academy of Sciences of beetles reveals the evolutionary origins of a superradiation. the USA 103:968–971. Science 318:1913–1916. Hardy NB & Cook LG (2010) Gall-induction in insects: Ives AR & Godfray HC (2006) Phylogenetic analysis of trophic evolutionary dead-end or speciation driver? BMC associations. American Naturalist 168:E1–E14. Evolutionary Biology 10:257. Jackson AP (2004) Cophylogeny of the Ficus microcosm. Hawkins BA (1994) Pattern and Process in Host-parasitoid Biological Reviews 79:751–768. Interactions. Cambridge University Press, Cambridge. Janz N (2011) Ehrlich and Raven revisited: mechanisms Hayward A, McMahon DP & Kathirithamby J (2011) Cryptic underlying codiversification of plants and enemies. Annual diversity and female host specificity in a parasitoid where the Review of Ecology, Evolution, and Systematics 42:71–89. sexes utilize hosts from separate orders. Molecular Ecology Janz N & Nylin S (1998) Butterflies and plants: a phylogenetic 20:1508–1528. study. Evolution 486–502. Hebert PD, Cywinska A, Ball SL & deWaard JR (2003) Biological Kaartinen R, Stone GN, Hearn J, Lohse K & Roslin T (2010) identificaions through DNA barcodes. Proceedings of the Royal Revealing secret liaisons: DNA barcoding changes our Society B: Biological Sciences 270:313–321. understanding of food webs. Ecological Entomology 35:623– Hebert PD, Penton EH, Burns JM, Janzen DH & Hallwachs W 638. (2004) Ten species in one: DNA barcoding reveals cryptic Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, species in the neotropical skipper butterfly Astraptes Ackerly DD, Blomberg SP & Webb CO (2010) Picante: R tools

51 Gompert Z, Forister ML, Fordyce JA & Nice CC (2008) fulgerator. Proceedings of the National Academy of Sciences of the Widespread mito-nuclear discordance with evidence for USA 101:14812–14817. introgressive hybridization and selective sweeps in Lycaeides. Hendry AP, Huber SK, De León LF, Herrel A & Podos J (2009) Molecular Ecology 17:5231–44. Disruptive selection in a bimodal population of Darwin's Goulet H (1992) The genera and subgenera of the sawflies of finches Proceedings of the Royal Society B: Biological Sciences Canada and Alaska (Hymenoptera: Symphyta). Insects and 276:753–759. Arachnids of Canada 20:1–235. Hernández-Vera G, Mitrović M, Jović J, Tosevski I, Caldara R, Goulet H & Huber JT (1993) Hymenoptera of the World: An Gassmann A & Emerson BC (2010) Host-associated genetic Identification Guide to Families. Research Branch Agriculture differentiation in a seed parasitic weevil Rhinusa antirrhini Canada Publication, Ottawa, Ontario. (Coleptera: Curculionidae) revealed by mitochondrial and Gómez-Zurita J, Hunt T, Kopliku F & Vogler AP (2007) nuclear sequence data. Molecular Ecology 19:2286–2300. Recalibrated tree of leaf beetles (Chrysomelidae) indicates Hespenheide HA (1991) Bionomics of leaf-mining insects. independent diversification of angiosperms and their insect Annual Review of Entomology 36:535–560. herbivores. PLoS ONE 2:e360. Hrcek J, Miller SE, Quicke DL & Smith MA (2011) Molecular Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W & detection of trophic links in a complex insect host-parasitoid Gascuel O (2010) New algorithms and methods to estimate food web. Molecular Ecology Resources 11:786–794. maximum-likelihood phylogenies: assessing the performance Hunt T, Bergsten J, Levkanicova Z, Papadopoulou A, John OS, of PhyML 3.0. Systematic Biology 59:307–321. Wild R, Hammond PM, Ahrens D, Balke M, Caterino MS, Hajibabaei M, Janzen DH, Burns JM, Hallwachs W & Hebert PD Gómez-Zurita J, Ribera I, Barraclough TG, Bocakova M, (2006) DNA barcodes distinguish species of tropical Bocak L & Vogler AP (2007) A comprehensive phylogeny of Lepidoptera. Proceedings of the National Academy of Sciences of beetles reveals the evolutionary origins of a superradiation. the USA 103:968–971. Science 318:1913–1916. Hardy NB & Cook LG (2010) Gall-induction in insects: Ives AR & Godfray HC (2006) Phylogenetic analysis of trophic evolutionary dead-end or speciation driver? BMC associations. American Naturalist 168:E1–E14. Evolutionary Biology 10:257. Jackson AP (2004) Cophylogeny of the Ficus microcosm. Hawkins BA (1994) Pattern and Process in Host-parasitoid Biological Reviews 79:751–768. Interactions. Cambridge University Press, Cambridge. Janz N (2011) Ehrlich and Raven revisited: mechanisms Hayward A, McMahon DP & Kathirithamby J (2011) Cryptic underlying codiversification of plants and enemies. Annual diversity and female host specificity in a parasitoid where the Review of Ecology, Evolution, and Systematics 42:71–89. sexes utilize hosts from separate orders. Molecular Ecology Janz N & Nylin S (1998) Butterflies and plants: a phylogenetic 20:1508–1528. study. Evolution 486–502. Hebert PD, Cywinska A, Ball SL & deWaard JR (2003) Biological Kaartinen R, Stone GN, Hearn J, Lohse K & Roslin T (2010) identificaions through DNA barcodes. Proceedings of the Royal Revealing secret liaisons: DNA barcoding changes our Society B: Biological Sciences 270:313–321. understanding of food webs. Ecological Entomology 35:623– Hebert PD, Penton EH, Burns JM, Janzen DH & Hallwachs W 638. (2004) Ten species in one: DNA barcoding reveals cryptic Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, species in the neotropical skipper butterfly Astraptes Ackerly DD, Blomberg SP & Webb CO (2010) Picante: R tools

51 for integrating phylogenies and ecology. Bioinformatics from Bayesian analyses of combined datasets. Molecular 26:1463–1464. Phylogenetics and Evolution 30:686–702. Kergoat GJ, Silvain JF, Delobel A, Tuda M & Anton KW (2007) Linnen CR & Farrell BD (2007) Mitonuclear discordance is Defining the limits of taxonomic conservatism in host-plant caused by rampant mitochondrial introgression in Neodiprion use for phytophagous insects: molecular systematics and (Hymenoptera: Diprionidae) sawflies. Evolution 61:1417–1438. evolution of host-plant associations in the seed-beetle genus Liò P & Goldman N (1998) Models of molecular evolution and Bruchus Linnaeus (Coleoptera: Chrysomelidae: Bruchinae). phylogeny. Genome Research 8:1233–1244. Molecular Phylogenetics and Evolution 43:251–269. Lopez-Vaamonde C, Godfray HC & Cook JM (2003) Kitching RL (2006) Crafting the pieces of the diversity jigsaw Evolutionary dynamics of host-plant use in a genus of leaf- puzzle. Science 313:1055–1057. mining moths. Evolution 57:1804–1821. Kluge AG & Farris JS (1969) Quantitative phyletics and the Lopez-Vaamonde C, Godfray HC, West SA, Hansson C & Cook evolution of anurans. Systematic Biology 18:1–32. JM (2005) The evolution of host use and unusual Kopelke JP (1999) Gallenerzeugende blattwespen Europas: reproductive strategies in Achrysocharoides parasitoid wasps. taxonomische grundlagen, biologie und ökologie Journal of Evolutionary Biology 18:1029–1041. Tenthredinidae: Nematinae: Euura, Phyllocolpa, Pontania). Lopez-Vaamonde C, Wikström N, Labandeira C, Godfray HC, Courier Forschungsinstitut Senckenberg 212:1–183. Goodman SJ & Cook JM (2006) Fossil-calibrated molecular Kopelke JP (2003) Gall-forming Nematinae, their willow hosts phylogenies reveal that leaf-mining moths radiated millions (Salix spp.) and biological strategies (Insecta, Hymenoptera, of years after their host plants. Journal of Evolutionary Biology Symphyta, Tenthredinidae, Nematinae: Euura, Phyllocolpa, 19:1314–1326. Pontania). Senckenbergiana Biologica 82:163–190. Maddison WP & Maddison DR (2010) Mesquite: a modular Labandeira CC & Phillips TL (1996) A Carboniferous insect gall: system for evolutionary analysis. Version 2.74. Available at: insight into early ecologic history of the Holometabola. mesquiteproject. org/mesquite/download/download.html Proceedings of the National Academy of Sciences of the USA Malaise R (1920) Beitrage zur kenntnis schwedischer 93:8470–8474. blattwespen. Entomologisk Tidskrift 40:97–128. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan Malm T & Nyman T (2014) Phylogeny of the symphytan grade PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, of Hymenoptera: new pieces into the old jigsaw(fly) puzzle. Thompson JD, Gibson TJ & Higgins DG (2007) Clustal W and Cladistics, in press. Clustal X version 2.0. Bioinformatics 23:2947–2948. Mardulyn P, Othmezouri N, Mikhailov YE & Pasteels JM (2011) Lemey P, Salemi M & Vandamme A (2009) The Phylogenetic Conflicting mitochondrial and nuclear phylogeographic Handbook. A Practical Approach to Phylogenetic Analysis and signals and evolution of host-plant shifts in the boreo- Hypothesis Testing. Cambridge University Press, Cambridge, montane leaf beetle Chrysomela lapponica. Molecular United Kindom. Phylogenetics and Evolution 61:686–696. Lill JT, Marquis RJ & Ricklefs RE (2002) Host plants influence Marshall DC, Hill KB, Cooley JR & Simon C (2011) parasitism of forest caterpillars. Nature 417:170–173. Hybridization, mitochondrial DNA phylogeography, and Lin C-P & Danforth BN (2004) How do insect nuclear and prediction of the early stages of reproductive isolation: mitochondrial gene substitution patterns differ? Insights lessons from New Zealand cicadas (genus Kikihia). Systematic Biology 60:482–502.

53 for integrating phylogenies and ecology. Bioinformatics from Bayesian analyses of combined datasets. Molecular 26:1463–1464. Phylogenetics and Evolution 30:686–702. Kergoat GJ, Silvain JF, Delobel A, Tuda M & Anton KW (2007) Linnen CR & Farrell BD (2007) Mitonuclear discordance is Defining the limits of taxonomic conservatism in host-plant caused by rampant mitochondrial introgression in Neodiprion use for phytophagous insects: molecular systematics and (Hymenoptera: Diprionidae) sawflies. Evolution 61:1417–1438. evolution of host-plant associations in the seed-beetle genus Liò P & Goldman N (1998) Models of molecular evolution and Bruchus Linnaeus (Coleoptera: Chrysomelidae: Bruchinae). phylogeny. Genome Research 8:1233–1244. Molecular Phylogenetics and Evolution 43:251–269. Lopez-Vaamonde C, Godfray HC & Cook JM (2003) Kitching RL (2006) Crafting the pieces of the diversity jigsaw Evolutionary dynamics of host-plant use in a genus of leaf- puzzle. Science 313:1055–1057. mining moths. Evolution 57:1804–1821. Kluge AG & Farris JS (1969) Quantitative phyletics and the Lopez-Vaamonde C, Godfray HC, West SA, Hansson C & Cook evolution of anurans. Systematic Biology 18:1–32. JM (2005) The evolution of host use and unusual Kopelke JP (1999) Gallenerzeugende blattwespen Europas: reproductive strategies in Achrysocharoides parasitoid wasps. taxonomische grundlagen, biologie und ökologie Journal of Evolutionary Biology 18:1029–1041. Tenthredinidae: Nematinae: Euura, Phyllocolpa, Pontania). Lopez-Vaamonde C, Wikström N, Labandeira C, Godfray HC, Courier Forschungsinstitut Senckenberg 212:1–183. Goodman SJ & Cook JM (2006) Fossil-calibrated molecular Kopelke JP (2003) Gall-forming Nematinae, their willow hosts phylogenies reveal that leaf-mining moths radiated millions (Salix spp.) and biological strategies (Insecta, Hymenoptera, of years after their host plants. Journal of Evolutionary Biology Symphyta, Tenthredinidae, Nematinae: Euura, Phyllocolpa, 19:1314–1326. Pontania). Senckenbergiana Biologica 82:163–190. Maddison WP & Maddison DR (2010) Mesquite: a modular Labandeira CC & Phillips TL (1996) A Carboniferous insect gall: system for evolutionary analysis. Version 2.74. Available at: insight into early ecologic history of the Holometabola. mesquiteproject. org/mesquite/download/download.html Proceedings of the National Academy of Sciences of the USA Malaise R (1920) Beitrage zur kenntnis schwedischer 93:8470–8474. blattwespen. Entomologisk Tidskrift 40:97–128. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan Malm T & Nyman T (2014) Phylogeny of the symphytan grade PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, of Hymenoptera: new pieces into the old jigsaw(fly) puzzle. Thompson JD, Gibson TJ & Higgins DG (2007) Clustal W and Cladistics, in press. Clustal X version 2.0. Bioinformatics 23:2947–2948. Mardulyn P, Othmezouri N, Mikhailov YE & Pasteels JM (2011) Lemey P, Salemi M & Vandamme A (2009) The Phylogenetic Conflicting mitochondrial and nuclear phylogeographic Handbook. A Practical Approach to Phylogenetic Analysis and signals and evolution of host-plant shifts in the boreo- Hypothesis Testing. Cambridge University Press, Cambridge, montane leaf beetle Chrysomela lapponica. Molecular United Kindom. Phylogenetics and Evolution 61:686–696. Lill JT, Marquis RJ & Ricklefs RE (2002) Host plants influence Marshall DC, Hill KB, Cooley JR & Simon C (2011) parasitism of forest caterpillars. Nature 417:170–173. Hybridization, mitochondrial DNA phylogeography, and Lin C-P & Danforth BN (2004) How do insect nuclear and prediction of the early stages of reproductive isolation: mitochondrial gene substitution patterns differ? Insights lessons from New Zealand cicadas (genus Kikihia). Systematic Biology 60:482–502.

53 Mayhew PJ (2007) Why are there so many insect species? Nosil P & Crespi BJ (2006) Experimental evidence that predation Perspectives from fossils and phylogenies. Biological Reviews promotes divergence in adaptive radiation. Proceedings of the 82:425–454. National Academy of Sciences of the USA 103:9090–9095. McCall PJ, Turlings TC, Lewis WJ & Tumlinson JH (1993) Role Novotny V, Basset Y, Miller SE, Weiblen GD, Bremer B, Cizek L of plant volatiles in host location by the specialist parasitoid & Drozd P (2002) Low host specificity of herbivorous insects Microplitis croceipes Cresson (Braconidae: Hymenoptera). in a tropical forest. Nature 416:841–844. Journal of Insect Behavior 6:625–639. Novotny V, Drozd P, Miller SE, Kulfan M, Janda M, Basset Y & McCune B & Mefford MJ (2006) PC-ORD. Multivariate analysis Weiblen GD (2006) Why are there so many species of of ecological data. Version 5. MjM Software Design, herbivorous insects in tropical rainforests? Science 313:1115– Gleneden Beach, Oregon, USA. 1118. McKenna DD, Sequeira AS, Marvaldi AE & Farrell BD (2009) Novotny V, Miller SE, Baje L, Balagawi S, Basset Y, Cizek L, Temporal lags and overlap in the diversification of weevils Craft KJ, Dem F, Drew RA, Hulcr J, Leps J, Lewis OT, Pokon and flowering plants. Proceedings of the National Academy of R, Stewart AJ, Samuelson GA & Weiblen GD (2010) Guild- Sciences of the USA 106:7083–7088. specific patterns of species richness and host specialization in McPeek M & Brown J (2007) Clade age and not diversification plant-herbivore food webs from a tropical forest. Journal of rate explains species richness among animal taxa. American Animal Ecology 79:1193–1203. Naturalist 169:E97–E106. Noyes JS (2013) Universal Chalcidoidea Database. World Wide Meyer J (1987) Plant Galls and Gall Inducers. Gebrüder Web electronic publication, Available at Borntraeger, Berlin. http://www.nhm.ac.uk/research-curation/research/ Miller MA, Pfeiffer W, Schwartz T (2010) Creating the CIPRES projects/chalcidoids/. Science Gateway for inference of large phylogenetic trees. In: Nyman T (2000) Phylogeny and Ecological Evolution of Gall- Proceedings of the Gateway Computing Environments Workshop Inducing Sawflies (Hymenoptera: Tenthredinidae). Ph.D. Thesis, (GCE). 14 Nov 2010, New Orleans, LA. University of Joensuu, Finland. Mitter C, Farrell B & Wiegmann B (1988) The phylogenetic study Nyman T (2002) The willow bud galler Euura mucronata Hartig of adaptive zones: has phytophagy promoted insect (Hymenoptera: Tenthredinidae): one polyphage or many diversification? American Naturalist 132:107–128. monophages? Heredity 88:288–295. Morris RJ, Lewis OT & Godfray HC (2004) Experimental Nyman T (2010) To speciate, or not to speciate? Resource evidence for apparent competition in a tropical forest food heterogeneity, the subjectivity of similarity, and the web. Nature 428:310–313. macroevolutionary consequences of niche-width shifts in Murphy SM (2004) Enemy-free space maintains swallowtail plant-feeding insects. Biological Reviews 85:393–411. butterfly host shift Proceedings of the National Academy of Nyman T, Widmer A & Roininen H (2000) Evolution of gall Sciences of the USA 101:18048–18052. morphology and host-plant relationships in willow-feeding Nixon KC (1999) The parsimony ratchet, a new method for sawflies (Hymenoptera: Tenthredinidae). Evolution 54:526– rapid parsimony analysis. Cladistics 15:407–414. 533. Nosil P & Mooers AO (2005) Testing hypotheses about Nyman T, Farrell BD, Zinovjev AG & Vikberg V (2006a) Larval ecological specialization using phylogenetic trees. Evolution habits, host-plant associations, and speciation in nematine 59:2256–2263.

55 Mayhew PJ (2007) Why are there so many insect species? Nosil P & Crespi BJ (2006) Experimental evidence that predation Perspectives from fossils and phylogenies. Biological Reviews promotes divergence in adaptive radiation. Proceedings of the 82:425–454. National Academy of Sciences of the USA 103:9090–9095. McCall PJ, Turlings TC, Lewis WJ & Tumlinson JH (1993) Role Novotny V, Basset Y, Miller SE, Weiblen GD, Bremer B, Cizek L of plant volatiles in host location by the specialist parasitoid & Drozd P (2002) Low host specificity of herbivorous insects Microplitis croceipes Cresson (Braconidae: Hymenoptera). in a tropical forest. Nature 416:841–844. Journal of Insect Behavior 6:625–639. Novotny V, Drozd P, Miller SE, Kulfan M, Janda M, Basset Y & McCune B & Mefford MJ (2006) PC-ORD. Multivariate analysis Weiblen GD (2006) Why are there so many species of of ecological data. Version 5. MjM Software Design, herbivorous insects in tropical rainforests? Science 313:1115– Gleneden Beach, Oregon, USA. 1118. McKenna DD, Sequeira AS, Marvaldi AE & Farrell BD (2009) Novotny V, Miller SE, Baje L, Balagawi S, Basset Y, Cizek L, Temporal lags and overlap in the diversification of weevils Craft KJ, Dem F, Drew RA, Hulcr J, Leps J, Lewis OT, Pokon and flowering plants. Proceedings of the National Academy of R, Stewart AJ, Samuelson GA & Weiblen GD (2010) Guild- Sciences of the USA 106:7083–7088. specific patterns of species richness and host specialization in McPeek M & Brown J (2007) Clade age and not diversification plant-herbivore food webs from a tropical forest. Journal of rate explains species richness among animal taxa. American Animal Ecology 79:1193–1203. Naturalist 169:E97–E106. Noyes JS (2013) Universal Chalcidoidea Database. World Wide Meyer J (1987) Plant Galls and Gall Inducers. Gebrüder Web electronic publication, Available at Borntraeger, Berlin. http://www.nhm.ac.uk/research-curation/research/ Miller MA, Pfeiffer W, Schwartz T (2010) Creating the CIPRES projects/chalcidoids/. Science Gateway for inference of large phylogenetic trees. In: Nyman T (2000) Phylogeny and Ecological Evolution of Gall- Proceedings of the Gateway Computing Environments Workshop Inducing Sawflies (Hymenoptera: Tenthredinidae). Ph.D. Thesis, (GCE). 14 Nov 2010, New Orleans, LA. University of Joensuu, Finland. Mitter C, Farrell B & Wiegmann B (1988) The phylogenetic study Nyman T (2002) The willow bud galler Euura mucronata Hartig of adaptive zones: has phytophagy promoted insect (Hymenoptera: Tenthredinidae): one polyphage or many diversification? American Naturalist 132:107–128. monophages? Heredity 88:288–295. Morris RJ, Lewis OT & Godfray HC (2004) Experimental Nyman T (2010) To speciate, or not to speciate? Resource evidence for apparent competition in a tropical forest food heterogeneity, the subjectivity of similarity, and the web. Nature 428:310–313. macroevolutionary consequences of niche-width shifts in Murphy SM (2004) Enemy-free space maintains swallowtail plant-feeding insects. Biological Reviews 85:393–411. butterfly host shift Proceedings of the National Academy of Nyman T, Widmer A & Roininen H (2000) Evolution of gall Sciences of the USA 101:18048–18052. morphology and host-plant relationships in willow-feeding Nixon KC (1999) The parsimony ratchet, a new method for sawflies (Hymenoptera: Tenthredinidae). Evolution 54:526– rapid parsimony analysis. Cladistics 15:407–414. 533. Nosil P & Mooers AO (2005) Testing hypotheses about Nyman T, Farrell BD, Zinovjev AG & Vikberg V (2006a) Larval ecological specialization using phylogenetic trees. Evolution habits, host-plant associations, and speciation in nematine 59:2256–2263.

55 sawflies (Hymenoptera: Tenthredinidae). Evolution 60:1622– Aphidius ervi Haliday (Hymenoptera: Braconidae: 1637. Aphidiinae). Biological Control 11:104–112. Nyman T, Zinovjev A, Vikberg V & Farrell B (2006b) Molecular Powell TH, Hood GR, Murphy MO, Heilveil JS, Berlocher SH, phylogeny of the sawfly subfamily Nematinae Nosil P & Feder JL (2013) Genetic divergence along the (Hymenoptera: Tenthredinidae). Systematic Entomology speciation continuum: the transition from host race to species 31:569–583. in Rhagoletis (Diptera: Tephritidae). Evolution 67:2561–76. Nyman T, Bokma F & Kopelke JP (2007) Reciprocal Price PW (2005) Adaptive radiation of gall-inducing insects. diversification in a complex plant-herbivore-parasitoid food Basic and Applied Ecology 6:413–421. web. BMC Biology 5:49. Price PW & Clancy KM (1986) Interactions among three trophic Nyman T, Vikberg V, Smith DR & Boevé JL (2010) How levels: gall size and parasitoid attack. Ecology 1593–1600. common is ecological speciation in plant-feeding insects? A Price PW & Pschorn-Walcher H (1988) Are galling insects better 'Higher' Nematinae perspective. BMC Evolutionary Biology protected against parasitoids than exposed feeders? A test 10:266. using tenthredinid sawflies. Ecological Entomology 13:195–205. Ohshima I, Tanikawa-Dodo Y, Saigusa T, Nishiyama T, Kitani Price PW & Roininen H (1993) Adaptive radiation in gall M, Hasebe M & Mohri H (2010) Phylogeny, biogeography, induction. In Wagner M, and Raffa KF (eds). Sawfly Life and host-plant association in the subfamily Apaturinae History Adaptations to Woody Plants. Academic Press, New (Insecta: Lepidoptera: Nymphalidae) inferred from eight York, pp. 229–257. nuclear and seven mitochondrial genes. Molecular Price PW, Fernandes GW & Waring GL (1987) Adaptive nature Phylogenetics and Evolution 57:1026–1036. of insect galls. Environmental Entomology 16:15–24. Peccoud J, Ollivier A, Plantegenest M & Simon J-C (2009) A Pschorn-Walcher H & Altenhofer E (1989) The parasitoid continuum of genetic divergence from sympatric host races community of leaf-mining sawflies (Fenusini and to species in the pea aphid complex. Proceedings of the National Heterarthrini): a comparative analysis. Zoologisher Anzeiger Academy of Sciences of the USA 106:7495–7500. 222:37–56. Percy DM, Page RDM & Cronk QCB (2004) Plant-insect Quicke DLJ, Laurenne NM, Fitton MG & Broad GR (2009) A interactions: double-dating associated insect and plant thousad and one wasps: a 28S rDNA and morphological lineages reveals asynchronous radiations. Systematic Biology phylogeny of the Ichneumonidae (Insecta: Hymenoptera) 53:120–127. with an investigation into aligment parameter space and Pitz KM, Dowling AP, Sharanowski BJ, Boring CA, Seltmann elision. Journal of Natural History 43:1305–1421. KC & Sharkey MJ (2007) Phylogenetic relationships among Richter A & Kasparyan R (2013) Tachinid (Diptera, Tachinidae) the Braconidae (Hymenoptera: Ichneumonoidea): a parasitoids of sawflies (Hymenoptera, Symphyta). reassessment of Shi et al. (2005). Molecular Phylogenetics and Entomological Review 93:630–633. Evolution 43:338–343. Roderick GK (1997) Herbivorous insects and the Hawaiian Posada D (2008) jModelTest: phylogenetic model averaging. silversword alliance: coevolution or cospeciation? Pacific Molecular Biology and Evolution 25:1253–1256. Science 51:440–449. Powell W, Pennacchio F, Poppy GM & Tremblay E (1998) Roininen H & Danell K (1997) Mortality factors and resource use Strategies involved in the location of hosts by the parasitoid of the bud-galling sawfly, Euura mucronata (Hartig), on willows (Salix spp.) in arctic Eurasia. Polar Biology 18:325–330.

57 sawflies (Hymenoptera: Tenthredinidae). Evolution 60:1622– Aphidius ervi Haliday (Hymenoptera: Braconidae: 1637. Aphidiinae). Biological Control 11:104–112. Nyman T, Zinovjev A, Vikberg V & Farrell B (2006b) Molecular Powell TH, Hood GR, Murphy MO, Heilveil JS, Berlocher SH, phylogeny of the sawfly subfamily Nematinae Nosil P & Feder JL (2013) Genetic divergence along the (Hymenoptera: Tenthredinidae). Systematic Entomology speciation continuum: the transition from host race to species 31:569–583. in Rhagoletis (Diptera: Tephritidae). Evolution 67:2561–76. Nyman T, Bokma F & Kopelke JP (2007) Reciprocal Price PW (2005) Adaptive radiation of gall-inducing insects. diversification in a complex plant-herbivore-parasitoid food Basic and Applied Ecology 6:413–421. web. BMC Biology 5:49. Price PW & Clancy KM (1986) Interactions among three trophic Nyman T, Vikberg V, Smith DR & Boevé JL (2010) How levels: gall size and parasitoid attack. Ecology 1593–1600. common is ecological speciation in plant-feeding insects? A Price PW & Pschorn-Walcher H (1988) Are galling insects better 'Higher' Nematinae perspective. BMC Evolutionary Biology protected against parasitoids than exposed feeders? A test 10:266. using tenthredinid sawflies. Ecological Entomology 13:195–205. Ohshima I, Tanikawa-Dodo Y, Saigusa T, Nishiyama T, Kitani Price PW & Roininen H (1993) Adaptive radiation in gall M, Hasebe M & Mohri H (2010) Phylogeny, biogeography, induction. In Wagner M, and Raffa KF (eds). Sawfly Life and host-plant association in the subfamily Apaturinae History Adaptations to Woody Plants. Academic Press, New (Insecta: Lepidoptera: Nymphalidae) inferred from eight York, pp. 229–257. nuclear and seven mitochondrial genes. Molecular Price PW, Fernandes GW & Waring GL (1987) Adaptive nature Phylogenetics and Evolution 57:1026–1036. of insect galls. Environmental Entomology 16:15–24. Peccoud J, Ollivier A, Plantegenest M & Simon J-C (2009) A Pschorn-Walcher H & Altenhofer E (1989) The parasitoid continuum of genetic divergence from sympatric host races community of leaf-mining sawflies (Fenusini and to species in the pea aphid complex. Proceedings of the National Heterarthrini): a comparative analysis. Zoologisher Anzeiger Academy of Sciences of the USA 106:7495–7500. 222:37–56. Percy DM, Page RDM & Cronk QCB (2004) Plant-insect Quicke DLJ, Laurenne NM, Fitton MG & Broad GR (2009) A interactions: double-dating associated insect and plant thousad and one wasps: a 28S rDNA and morphological lineages reveals asynchronous radiations. Systematic Biology phylogeny of the Ichneumonidae (Insecta: Hymenoptera) 53:120–127. with an investigation into aligment parameter space and Pitz KM, Dowling AP, Sharanowski BJ, Boring CA, Seltmann elision. Journal of Natural History 43:1305–1421. KC & Sharkey MJ (2007) Phylogenetic relationships among Richter A & Kasparyan R (2013) Tachinid (Diptera, Tachinidae) the Braconidae (Hymenoptera: Ichneumonoidea): a parasitoids of sawflies (Hymenoptera, Symphyta). reassessment of Shi et al. (2005). Molecular Phylogenetics and Entomological Review 93:630–633. Evolution 43:338–343. Roderick GK (1997) Herbivorous insects and the Hawaiian Posada D (2008) jModelTest: phylogenetic model averaging. silversword alliance: coevolution or cospeciation? Pacific Molecular Biology and Evolution 25:1253–1256. Science 51:440–449. Powell W, Pennacchio F, Poppy GM & Tremblay E (1998) Roininen H & Danell K (1997) Mortality factors and resource use Strategies involved in the location of hosts by the parasitoid of the bud-galling sawfly, Euura mucronata (Hartig), on willows (Salix spp.) in arctic Eurasia. Polar Biology 18:325–330.

57 Roininen H, Danell K, Zinovjev A, Vikberg V & Virtanen R Schluter D (2001) Ecology and the origin of species. Trends in (2002) Community structure, survival and mortality factors Ecology & Evolution 16:372–380. in arctic populations of Eupontania leaf gallers. Polar Biology Schluter D & Conte GL (2009) Genetics and ecological 25:605–611. speciation. Proceedings of the National Academy of Sciences of the Roininen H, Nyman T & Zinovjev A (2005) Biology, ecology, USA 106:9955–9962. and evolution of gall-inducing sawflies (Hymenoptera: Segar ST, Lopez-Vaamonde C, Rasplus JY & Cook JM (2012) The Tenthredinidae and Xyelidae). In Raman A, Schaefer CW, global phylogeny of the subfamily Sycoryctinae and Withers TM (eds). Biology, Ecology and Evolution of Gall- (Pteromalidae): parasites of an obligate mutualism. Molecular inducing Arthropods, Volumes 1 and 2. Science Publishers, Inc., Phylogenetics and Evolution 65:116–125. pp. 467–494. Sikes DS & Lewis PO (2001) PAUPRat: a tool to implement Ronquist F & Huelsenbeck JP (2003) MrBayes 3: Bayesian parsimony ratchet searches using PAUP*. Available at: phylogenetic inference under mixed models. Bioinformatics http://viceroy. eeb. uconn. edu/paupratweb/pauprat. htm. 19:1572–1574. Simmons, MP & Ochoterena H (2000) Gaps as characters in Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, sequence-based phylogenetic analysis. Systematic Biology Höhna S, Larget B, Liu L, Suchard MA & Huelsenbeck JP 49:369–381. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference Sinclair RJ & Hughes L (2010) Leaf miners: the hidden and model choice across a large model space. Systematic herbivores. Austral Ecology 35:300–313. Biology 61:539–542. Singer MS & Stireman JO (2005) The tri-trophic niche concept Rosenblum EB & Harmon LJ (2011) "Same same but different": and adaptive radiation of phytophagous insects. Ecology Replicated ecological speciation at white sands. Evolution Letters 8:1247–1255. 65:946–960. Skvortsov AK (1999) Willows of Russia and Adjacent Countries. Rott AS & Godfray HCJ (2000) The structure of a leafminer– University of Joensuu, Joensuu. parasitoid community. Journal of Animal Ecology 69:274–289. Slove J & Janz N (2011) The relationship between diet breadth Rundle HD & Nosil P (2005) Ecological speciation. Ecology and geographic range size in the butterfly subfamily Letters 8:336–352. Nymphalinae – a study of global scale. PLoS ONE 6:e16057. Salvo A & Valladares GR (2004) Looks are important: parasitic Smith DR (1971) Nearctic sawflies. III. Heterarthrinae: adults assemblages of agromyzid leafminers (Diptera) in relation to and larvae (Hymenotera: Tenthredinidae). Uniteds States mine shape and contrast. Journal of Animal Ecology 73:494–505. Department of Agriculture Technical Bulletin No. 1420 p. 84. Scheffer SJ & Hawthorne DJ (2007) Molecular evidence of host- Smith DR (1979) Suborder Symphyta. In Krombein KV, Hurd associated genetic divergence in the holly leafminer PD, Smith DR, and Burks BD (eds). Catalog of Hymenoptera in Phytomyza glabricola (Diptera: Agromyzidae): apparent America North of Mexico. Smithsonian Institution Press, discordance among marker systems. Molecular Ecology Washington, DC, pp. 3–137. 16:2627–2637. Smith EL (1970) Biosystematics and morphology of Symphyta. Scheffer SJ, Winkler IS & Wiegmann BM (2007) Phylogenetic II. Biology of gall-making nematine sawflies in the California relationships within the leaf-mining flies (Diptera: region. Annals of the Entomological Society of America 63:36–51. Agromyzidae) inferred from sequence data from multiple Smith MA, Woodley NE, Janzen DH, Hallwachs W & Hebert PD genes. Molecular Phylogenetics and Evolution 42:756–775. (2006) DNA barcodes reveal cryptic host-specificity within

59 Roininen H, Danell K, Zinovjev A, Vikberg V & Virtanen R Schluter D (2001) Ecology and the origin of species. Trends in (2002) Community structure, survival and mortality factors Ecology & Evolution 16:372–380. in arctic populations of Eupontania leaf gallers. Polar Biology Schluter D & Conte GL (2009) Genetics and ecological 25:605–611. speciation. Proceedings of the National Academy of Sciences of the Roininen H, Nyman T & Zinovjev A (2005) Biology, ecology, USA 106:9955–9962. and evolution of gall-inducing sawflies (Hymenoptera: Segar ST, Lopez-Vaamonde C, Rasplus JY & Cook JM (2012) The Tenthredinidae and Xyelidae). In Raman A, Schaefer CW, global phylogeny of the subfamily Sycoryctinae and Withers TM (eds). Biology, Ecology and Evolution of Gall- (Pteromalidae): parasites of an obligate mutualism. Molecular inducing Arthropods, Volumes 1 and 2. Science Publishers, Inc., Phylogenetics and Evolution 65:116–125. pp. 467–494. Sikes DS & Lewis PO (2001) PAUPRat: a tool to implement Ronquist F & Huelsenbeck JP (2003) MrBayes 3: Bayesian parsimony ratchet searches using PAUP*. Available at: phylogenetic inference under mixed models. Bioinformatics http://viceroy. eeb. uconn. edu/paupratweb/pauprat. htm. 19:1572–1574. Simmons, MP & Ochoterena H (2000) Gaps as characters in Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, sequence-based phylogenetic analysis. Systematic Biology Höhna S, Larget B, Liu L, Suchard MA & Huelsenbeck JP 49:369–381. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference Sinclair RJ & Hughes L (2010) Leaf miners: the hidden and model choice across a large model space. Systematic herbivores. Austral Ecology 35:300–313. Biology 61:539–542. Singer MS & Stireman JO (2005) The tri-trophic niche concept Rosenblum EB & Harmon LJ (2011) "Same same but different": and adaptive radiation of phytophagous insects. Ecology Replicated ecological speciation at white sands. Evolution Letters 8:1247–1255. 65:946–960. Skvortsov AK (1999) Willows of Russia and Adjacent Countries. Rott AS & Godfray HCJ (2000) The structure of a leafminer– University of Joensuu, Joensuu. parasitoid community. Journal of Animal Ecology 69:274–289. Slove J & Janz N (2011) The relationship between diet breadth Rundle HD & Nosil P (2005) Ecological speciation. Ecology and geographic range size in the butterfly subfamily Letters 8:336–352. Nymphalinae – a study of global scale. PLoS ONE 6:e16057. Salvo A & Valladares GR (2004) Looks are important: parasitic Smith DR (1971) Nearctic sawflies. III. Heterarthrinae: adults assemblages of agromyzid leafminers (Diptera) in relation to and larvae (Hymenotera: Tenthredinidae). Uniteds States mine shape and contrast. Journal of Animal Ecology 73:494–505. Department of Agriculture Technical Bulletin No. 1420 p. 84. Scheffer SJ & Hawthorne DJ (2007) Molecular evidence of host- Smith DR (1979) Suborder Symphyta. In Krombein KV, Hurd associated genetic divergence in the holly leafminer PD, Smith DR, and Burks BD (eds). Catalog of Hymenoptera in Phytomyza glabricola (Diptera: Agromyzidae): apparent America North of Mexico. Smithsonian Institution Press, discordance among marker systems. Molecular Ecology Washington, DC, pp. 3–137. 16:2627–2637. Smith EL (1970) Biosystematics and morphology of Symphyta. Scheffer SJ, Winkler IS & Wiegmann BM (2007) Phylogenetic II. Biology of gall-making nematine sawflies in the California relationships within the leaf-mining flies (Diptera: region. Annals of the Entomological Society of America 63:36–51. Agromyzidae) inferred from sequence data from multiple Smith MA, Woodley NE, Janzen DH, Hallwachs W & Hebert PD genes. Molecular Phylogenetics and Evolution 42:756–775. (2006) DNA barcodes reveal cryptic host-specificity within

59 the presumed polyphagous members of a genus of parasitoid Stireman JO, Devlin H & Abbot P (2012) Rampant host- and flies (Diptera: Tachinidae). Proceedings of the National Academy defensive phenotype-associated diversification in a of Sciences of the USA 103:3657–3662. goldenrod gall midge. Journal of Evolutionary Biology 25:1991– Smith MA, Wood DM, Janzen DH, Hallwachs W & Hebert PDN 2004. (2007) DNA barcodes affirm that 16 species of apparently Stone GN & Schönrogge K (2003) The adaptive significance of generalist tropical parasitoid flies (Diptera, Tachinidae) are insect gall morphology. Trends in Ecology & Evolution 18:512– not all generalists. Proceedings of the National Academy of 522. Sciences of the USA 104:4967–4972. Strong DR, Lawton JH & Southwood TRE (1984) Insects on Smith MA, Rodriguez JJ, Whitfield JB, Deans AR, Janzen DH, Plants: Community Patterns and Mechanisms. Harvard Univ. Hallwachs W & Hebert PD (2008) Extreme diversity of Press, Cambridge, MA. tropical parasitoid wasps exposed by iterative integration of Swofford DL (2003) PAUP*. Phylogenetic Analysis Using natural history, DNA barcoding, morphology, and Parsimony (*and Other Methods). Version 4. Sinauer collections. Proceedings of the National Academy of Sciences of the Associates, Sunderland, MA. USA 105:12359–12364. Taeger A & Altenhofer E (1998) Kommentare zur Biologie, Smith MA, Eveleigh ES, McCann KS, Merilo MT, McCarthy PC Verbreitung und Gefahrdung der Pflanzenwespen & Van Rooyen KI (2011) Barcoding a quantified food web: Deutschlands (Hymenoptera, Symphyta). In Blank S M & crypsis, concepts, ecology and hypotheses. PLoS ONE Taeger A (eds.) Pflanzenwespen Deutschlands (Hymenoptera. 6:e14424. Symphyta). Kommentierte Bestandsaufnahme. Goecke & Stamatakis A, Hoover P & Rougemont J (2008) A rapid Evers, Keltern. pp. 49–135 bootstrap algorithm for the RAxML Web servers. Systematic Taeger A, Blank SM & Liston AD (2010) World catalog of Biology 57:758–771. Symphyta (Hymenoptera). Zootaxa 2580:1–1064. Stireman JO (2005) The evolution of generalization? Parasitoid Tamura K, Peterson D, Peterson N, Stecher G, Nei M & Kumar S flies and the perils of inferring host range evolution from (2011) MEGA5: molecular evolutionary genetics analysis phylogenies. Journal of Evolutionary Biology 18:325–336. using maximum likelihood, evolutionary distance, and Stireman JO, Nason JD & Heard SB (2005) Host-associated maximum parsimony methods. Molecular Biology and genetic differentiation in phytophagous insects: general Evolution 28:2731–2739. phenomenon or isolated exceptions? Evidence from a Tilmon KJ, Danforth BN, Day WH & Hoffmann MP (2000) goldenrod-insect community. Evolution 59:2573–2587. Determining parasitoid species composition in a host Stireman JO, Nason JD, Heard SB & Seehawer JM (2006) population: a molecular approach. Annals of the Entomological Cascading host-associated genetic differentiation in Society of America 93:640–647. parasitoids of phytophagous insects. Proceedings of the Royal Via S, Bouck AC & Skillman S (2000) Reproductive isolation Society B: Biological Sciences 273:523–530. between divergent races of pea aphids on two hosts. II. Stireman JO, Devlin H, Carr TG & Abbot P (2010) Evolutionary Selection against migrants and hybrids in the parental diversification of the gall midge genus Asteromyia environments. Evolution 54:1626–1637. (Cecidomyiidae) in a multitrophic ecological context. Viitasaari M (2002) A review of the extant families of the Molecular Phylogenetics and Evolution 54:194–210. Hymenoptera. In Viitasaari M (Ed) Sawflies I. Tremex Press, Jyväskylä, pp. 175–182.

61 the presumed polyphagous members of a genus of parasitoid Stireman JO, Devlin H & Abbot P (2012) Rampant host- and flies (Diptera: Tachinidae). Proceedings of the National Academy defensive phenotype-associated diversification in a of Sciences of the USA 103:3657–3662. goldenrod gall midge. Journal of Evolutionary Biology 25:1991– Smith MA, Wood DM, Janzen DH, Hallwachs W & Hebert PDN 2004. (2007) DNA barcodes affirm that 16 species of apparently Stone GN & Schönrogge K (2003) The adaptive significance of generalist tropical parasitoid flies (Diptera, Tachinidae) are insect gall morphology. Trends in Ecology & Evolution 18:512– not all generalists. Proceedings of the National Academy of 522. Sciences of the USA 104:4967–4972. Strong DR, Lawton JH & Southwood TRE (1984) Insects on Smith MA, Rodriguez JJ, Whitfield JB, Deans AR, Janzen DH, Plants: Community Patterns and Mechanisms. Harvard Univ. Hallwachs W & Hebert PD (2008) Extreme diversity of Press, Cambridge, MA. tropical parasitoid wasps exposed by iterative integration of Swofford DL (2003) PAUP*. Phylogenetic Analysis Using natural history, DNA barcoding, morphology, and Parsimony (*and Other Methods). Version 4. Sinauer collections. Proceedings of the National Academy of Sciences of the Associates, Sunderland, MA. USA 105:12359–12364. Taeger A & Altenhofer E (1998) Kommentare zur Biologie, Smith MA, Eveleigh ES, McCann KS, Merilo MT, McCarthy PC Verbreitung und Gefahrdung der Pflanzenwespen & Van Rooyen KI (2011) Barcoding a quantified food web: Deutschlands (Hymenoptera, Symphyta). In Blank S M & crypsis, concepts, ecology and hypotheses. PLoS ONE Taeger A (eds.) Pflanzenwespen Deutschlands (Hymenoptera. 6:e14424. Symphyta). Kommentierte Bestandsaufnahme. Goecke & Stamatakis A, Hoover P & Rougemont J (2008) A rapid Evers, Keltern. pp. 49–135 bootstrap algorithm for the RAxML Web servers. Systematic Taeger A, Blank SM & Liston AD (2010) World catalog of Biology 57:758–771. Symphyta (Hymenoptera). Zootaxa 2580:1–1064. Stireman JO (2005) The evolution of generalization? Parasitoid Tamura K, Peterson D, Peterson N, Stecher G, Nei M & Kumar S flies and the perils of inferring host range evolution from (2011) MEGA5: molecular evolutionary genetics analysis phylogenies. Journal of Evolutionary Biology 18:325–336. using maximum likelihood, evolutionary distance, and Stireman JO, Nason JD & Heard SB (2005) Host-associated maximum parsimony methods. Molecular Biology and genetic differentiation in phytophagous insects: general Evolution 28:2731–2739. phenomenon or isolated exceptions? Evidence from a Tilmon KJ, Danforth BN, Day WH & Hoffmann MP (2000) goldenrod-insect community. Evolution 59:2573–2587. Determining parasitoid species composition in a host Stireman JO, Nason JD, Heard SB & Seehawer JM (2006) population: a molecular approach. Annals of the Entomological Cascading host-associated genetic differentiation in Society of America 93:640–647. parasitoids of phytophagous insects. Proceedings of the Royal Via S, Bouck AC & Skillman S (2000) Reproductive isolation Society B: Biological Sciences 273:523–530. between divergent races of pea aphids on two hosts. II. Stireman JO, Devlin H, Carr TG & Abbot P (2010) Evolutionary Selection against migrants and hybrids in the parental diversification of the gall midge genus Asteromyia environments. Evolution 54:1626–1637. (Cecidomyiidae) in a multitrophic ecological context. Viitasaari M (2002) A review of the extant families of the Molecular Phylogenetics and Evolution 54:194–210. Hymenoptera. In Viitasaari M (Ed) Sawflies I. Tremex Press, Jyväskylä, pp. 175–182.

61 Whitfield JB & Kjer KM (2008) Ancient rapid radiations of I insects: challenges for phylogenetic analysis. Annual Reviews of Entomology 53:449–472. Phylogenetics and evolution of host-plant use in leaf- Wilson JS, Forister ML, Dyer LA, O'Connor JM, Burls K, mining sawflies (Hymenoptera: Tenthredinidae: Feldman CR, Jaramillo MA, Miller JS, Rodríguez-Castañeda Heterarthrinae) G, Tepe EJ, Whitfield JB & Young B (2012) Host conservatism, host shifts and diversification across three trophic levels in two Neotropical forests. Journal of Evolutionary Biology 25:532– Leppänen S A, Altenhofer E, Liston A D, Nyman T (2012) 546. Winkler IS & Mitter C (2008) The phylogenetic dimension of insect-plant interactions: a review of recent evidence. In Molecular Phylogenetics and Evolution 64:331–341 Tilmon K J (Ed) Specialization, Speciation, and Radiation: the

Evolutionary Biology of Herbivorous Insects. University of Reprinted with the kind permission of Elsevier California Press, Berkeley, CA. pp. 240–263. Wirta HK, Hebert PD, Kaartinen R, Prosser SW, Várkonyi G & Roslin T (2014) Complementary molecular information changes our perception of food web structure. Proceedings of the National Academy of Sciences of the USA Early edition, www.pnas.org/cgi/doi/10.1073/pnas.1316990111. Yao H, Song J, Liu C, Luo K, Han J, Li Y, Pang X, Xu H, Zhu Y, Xiao P & Chen S (2010) Use of ITS2 region as the universal DNA barcode for plants and animals. PLoS ONE 5:e13102. Zaldivar-Riverón A, Mori M & Quicke DL (2006) Systematics of the cyclostome subfamilies of braconid parasitic wasps (Hymenoptera: Ichneumonoidea): a simultaneous molecular and morphological Bayesian approach. Molecular Phylogenetics and Evolution 38:130–145. Zaldívar-Riverón A, Shaw MR, Sáez AG, Mori M, Belokoblylskij SA, Shaw SR & Quicke DL (2008) Evolution of the parasitic wasp subfamily Rogadinae (Braconidae): phylogeny and evolution of lepidopteran host ranges and mummy characteristics. BMC Evolutionary Biology 8:329. Ødegaard F, Diserud OH & Østbye K (2005) The importance of plant relatedness for host utilization among phytophagous insects. Ecology Letters 8:612–617.

dissertations

Sanna Leppänen

Effect of host-plant use on | Leppänen | 154 | Sanna speciation and parasitoid community structure in internal-feeding sawflies