Evolution of Cooperation in Ambrosia Beetles

Inauguraldissertation

der Philosophisch-naturwissenschaftlichen Fakultät

der Universität Bern

vorgelegt von

Peter Hans Wilhelm Biedermann

von Trofaiach / Österreich

Leiter der Arbeit:

Prof. Dr. Michael Taborsky

Institut für Ökologie und Evolution

Abteilung Verhaltensökologie

Universität Bern

Evolution of Cooperation in Ambrosia Beetles

Inauguraldissertation

der Philosophisch-naturwissenschaftlichen Fakultät

der Universität Bern

vorgelegt von

Peter Hans Wilhelm Biedermann

von Trofaiach / Österreich

Leiter der Arbeit:

Prof. Dr. Michael Taborsky

Institut für Ökologie und Evolution

Abteilung Verhaltensökologie

Universität Bern

Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen.

Der Dekan: Bern, 20. März 2012 Prof. Dr. Silvio Decurtins

Supervised by: Prof. Dr. Michael Taborsky Department of Behavioural Ecology Institute of Ecology and Evolution University of Bern Wohlenstrasse 50a CH-3032 Hinterkappelen Switzerland

Reviewed by: Prof. Dr. Jacobus J. Boomsma Section for Ecology and Evolution Institute of Biology University of Copenhagen Universitetsparken 15 2100 Copenhagen Denmark

Examined by: Prof. Dr. Heinz Richner, University of Bern (Chair) Prof. Dr. Michael Taborsky, University of Bern Prof. Dr. Jacobus J. Boosma, University of Copenhagen

Copyright Chapter 1 © PNAS 2011 by the National Academy of Sciences of the United States of America, Washington, USA Chapter 2 © Mitt. Dtsch. Ges. allg. angew. Ent. 2011 by the DGaaE, Müncheberg, Gernany Chapter 4 © Zookeys 2010 by Pensoft Publishers, Sofia, Bulgaria Chapter 5 © Behav. Ecol. & Sociobiol. by Springer-Verlag GmbH, Heidelberg, Germany Chapter 9 © J. Bacteriol. by the American Society for Microbiology, Washington, USA General Introduction, Chapter 3, 6, 7, 8, Appendix 1,2, and Summary & Conclusion © Peter H.W. Biedermann

Cover drawing © by Barrett Anthony Klein, Entomoartist, Department of Biology, University of Konstanz, Germany. http://www.pupating.org

Layout by Peter H. W. Biedermann

Printed in Bern, Switzerland by Kopierzentrale der Universität Bern

Für meine Eltern die meine Begeisterung für die Natur erkannt haben

Für Tabea die mich bedingungslos unterstützt

Und Helene Francke-Grosmann, Karl E. Schedl, Dale M. Norris und Richard A. Roeper für ihre Pionierarbeiten auf dem Gebiet der Ambrosiakäfer-Forschung

“There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.”

C. Darwin, Origin of Species 1859

“Ich habe ihm nun geraten, künftig in der Natur nie einen einzelnen Gegenstand alleine herauszuzeichnen, nie einen einzelnen Baum, einen einzelnen Steinhaufen, eine einzelne Hütte, sondern immer zugleich einigen Hintergrund und einige Umgebung mit. Und zwar aus folgenden Ursachen: Wir sehen in der Natur nie etwas als Einzelheit, sondern wir sehen alles in Verbindung mit etwas anderem, das vor ihm, neben ihm, hinter ihm, unter ihm und über ihm sich befindet. Auch fällt uns wohl ein einzelner Gegenstand als besonders schön und malerisch auf; es ist aber nicht der Gegenstand allein, der diese Wirkung hervorbringt, sondern es ist die Verbindung, in der wir ihn sehen.“

Johann Wolfgang von Goethe, zu Eckermann, 5. Juni 1826

Contents

3 General Introduction

11 Chapter 1 Biedermann PHW and M Taborsky (2011): Larval helpers and age polyethism in ambrosia beetles. Proceedings of the National Academy of Science, USA 108(41): 17064-17069. 27 Chapter 2 Biedermann PHW, K Peer and M Taborsky (2011) Female dispersal and reproduction in the ambrosia beetle Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Mitteilungen der deutschen Gesellschaft für allgemeine und angewandte Entomologie 18: in press. 33 Chapter 3 Biedermann PHW and M Taborsky (manuscript in work) Responses to artificial selection on dispersal in a primitively eusocial beetle. 53 Chapter 4 Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus saxesenii Ratzeburg (Scolytinae, Coleoptera). In: Cognato AI, Knížek M (Eds) Sixty years of discovering scolytine and platypodine diversity: A tribute to Stephen L. Wood. Zookeys 56: 253-267. 69 Chapter 5 Biedermann PHW, KD Klepzig and M Taborsky (2011) Costs of delayed dispersal and alloparental care in the fungus-cultivating ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae). Behavioral Ecology and Sociobiology 65:1753–1761. 79 Chapter 6 Biedermann PHW and M Taborsky (manuscript in work) Social fungus farming varies among ambrosia beetles. 101 Chapter 7 Biedermann PHW, KD Klepzig, M Taborsky and DL Six (manuscript in work) Dynamics of filamentous fungi in the complex ambrosia gardens of the primitively eusocial beetle Xyleborinus saxesenii Ratzeburg (Scolytinae; Curculionidae). 123 Chapter 8 De Fine Licht HH and PHW Biedermann (in review) Patterns of functional enzyme activity show that larvae are the key to successful fungus farming by ambrosia beetles. Frontiers in Zoology, submitted 143 Chapter 9 Grubbs KJ, Biedermann PHW, Suen G, Adams SM, Moeller JA, Klassen JL, Goodwinm LA, Woyke T, Munk AC, Bruce D, Detter C, Tapia R, Han CS & CR Currie (2011) The complete genome sequence of Streptomyces cf. griseus (XyelbKG-1 1), an ambrosia beetle-associated actinomycete. Journal of Bacteriology 193(11): 2890-91.

145 Appendix 1 Biedermann PHW, M Taborsky M and DL Six (manuscript in work) Fungal associates and their effect on the behaviours and success of the ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae). 167 Appendix 2 Biedermann PHW, M Taborsky and CR Currie (manuscript in work) Mechanisms of fungus gardening in ambrosia beetles. 175 Summary & Conclusion 183 Acknowledgements 185 Contributions & Funding 189 Curriculum Vitae

Curriculum Vitae

General Introduction

“Again as in the case of corporeal structure, and conformably with my theory, the instinct for each species is good for itself, but has never, as far as we can judge, been produced for the exclusive good of others” C. Darwin, Origin of Species 1859.

Since I got familiar with Darwin’s concept of the survival of the fittest, I am very much fascinated by the question how a behaviour can persist in nature that obviously reduces the direct success (fitness) of an actor, but instead greatly benefits others. We are surrounded by such examples of the success of cooperation: packs of cooperatively hunting carnivores, helpers at bird nests that support the breeding pair in protection and food provisioning, the great success of social insects, where some individuals even refrain from reproduction, and the incredible diverse examples of symbioses, from gut microbes that help humans to digest their food, to pollination of flowers by bees and the fascinating fungus garden that provide nutrients for their insects farmers for getting tended, weeded and provisioned with new substrate. Perhaps less obvious, cooperation can not only be found among selfish biological entities, but is also the foundation of the major evolutionary transitions to stages of higher complexity. The evolution of chromosomes from independent genes, of multicellular organisms from individual cells and of societies from individuals involves a transition such that “entities that were capable of independent replication before the transition can replicate only as part of a larger whole after it” (Maynard-Smith and Szathmáry 1995).

The transition from egoistic individuals to symbioses and to animal societies was already noticed by Darwin as a challenge for his theory of natural selection (Darwin 1859), and has puzzled generations of evolutionary biologists ever since. After the concept of cooperation among animals as “good for the species” has been rejected, W.D. Hamilton’s inclusive fitness theory (also known as kin selection theory; Hamilton 1964) has prevailed as the currently best and most widely accepted theory for understanding the evolution of cooperation. Even though it is perhaps difficult to accept on first sight, there is tremendous evidence that selection acts on the gene level of individuals, meaning that in evolutionary times organisms persist that do not behave for the benefit of themselves, but instead for the highest possible success of the genes they carry. Thus, individuals are expected to maximize the success of their own genes (direct fitness) plus the success of the same genes in other (related) organisms (indirect fitness).

This thesis illustrates the power and validity of Hamilton’s inclusive fitness theory by critically testing its predictions with the social life of ambrosia beetles and their symbioses with fungi. My thesis is a first step for explaining the evolution of sociality and fungus farming in a lineage of beetles, which ancestors are solitary living and feeding on plants. On the next few pages, I will briefly introduce the inclusive fitness concept in more detail and present ecological and genetic factors that promote the evolution of cooperation in nature.

General Introduction

Hamilton’s inclusive fitness theory

„Die Theorie bestimmt was wir beobachten können.“ Albert Einstein.

Modern evolutionary theory on cooperation1 originated with Hamilton’s inclusive fitness theory and his ground-breaking article on the genetic evolution of social behaviour (Hamilton 1963; 1964). This led biologists to realize that altruistic behaviour is based on fundamentally selfish interests, often through helping relatives that possess the same genes, and are never done for the good of the species. (Darwin 1859). This can be also formulated by Hamilton’s rule (Hamilton 1964):

C < r × B

Where C is the fitness cost to the actor, r is the genetic relatedness between the actor and the recipient, and B is the fitness benefit to the recipient.

This formula shows that reproductive altruism can evolve even if it is associated with sterility of the actor, given costs of own reproduction are below the indirect fitness gains via helping a close relative. Hence, if individuals are clones (r = 1), reproductive altruism can evolve simply if it is more efficient than personal reproduction. Hence, theoretically, totipotent clonal individuals are indifferent about who reproduces (Frank 1995). That means, on the other hand, that the higher the relatedness, the lower the threat of cheating (i.e. one partner is not reciprocating the benefit it gets in the relationship) to evolve (e.g., Hamilton 1978; Bourke 2011). The general assumption that relatedness facilitates altruism and mitigates selfishness is only challenged under limited dispersal, when local competition between relatives may fully outweigh the benefits of altruism (Hamilton 1975; West et al. 2002).

Hamilton’s famous theory on kin selection can only be partly applied to the evolution of interspecific mutualism (symbiosis; i.e. when individuals of different species cooperate with each other), however (Foster and Wenseleers 2006). In groups of unrelated individuals, mutually beneficial acts, but not altruism (i.e. a behaviour with net costs for an actor), can evolve under the assumptions of inclusive fitness, because relatedness is needed (r > 0) for altruism to be selected (Hamilton 1964; 1972; Bourke 2011). Therefore, reproductive division of labour (= altruism) can only evolve within societies of related individuals and not in interspecific mutualism, whereas mutually beneficial acts are ubiquitously found in both intra- and interspecific cooperation. Hence, it is important to note that interspecific relationships will only evolve to be mutualistic if the interests of both partners match somehow (e.g. Herre et al. 1999; Foster and Wenseleers 2006). This might be due to a feedback- benefit between the partner’s successes: e.g. if a fungus farmed by insects will produce more fruiting bodies to feed more insects (increasing the fitness of the insects), it will be later dispersed by more insects (increasing its own fitness). Additional to relatedness and aligned fitness interests, there are

1 Here I define a behaviour as cooperative if social partners potentially benefit from its performance, independently of whether this behaviour entails net costs to the actor (Brosnan and de Waal, 2002). This definition includes (i) mutualistic behaviours that are regarded as selfish acts generating benefits to other individuals (common goods) as a by-product , and (ii) altruistic behaviours that bring about net costs to the actor, which are compensated through indirect fitness benefits via kin-selection; they will thus only evolve in groups of relatives (Hamilton, 1964). In the course of social evolution and task specialization shaped by kin selection, mutualistic behaviours may lose their original function and change into truly altruistic behaviors (Lin and Michener, 1972). 2

General Introduction

also ecological and other genetic factors that facilitate inclusive fitness gains for an organism to be higher when joining a group than breeding independently, which will be outlined in the following.

Ecological factors in the evolution of cooperation Ecological factors influence the survival and fecundity of solitary individuals and ones living in differently-sized groups, the costs of dispersal, and the costs and benefits of division of labour (Emlen 1982; Vehrencamp 1983a,b). Certain ecological factors have been repeatedly found to facilitate the evolution of sociality (Korb and Heinze 2008). The first set of factors predisposes for social life because of constraints on solitary breeding (ecological constraint hypothesis; Emlen 1982)): these are (i) limitation of high quality nesting sites, (ii) high mortality during dispersal and difficult nest foundation (e.g. harsh environments), and (iii) demographic factors like minimal group size or population density. The other set of factors facilitates social evolution due to inclusive fitness benefits of philopatry (benefits of philopatry hypothesis (Stacey and Ligon 1991): these are (iv) alloparental brood care, (v) a better common defense under high parasite or predation pressure, (vi) direct fitness through nest inheritance, and (vii) direct and indirect fitness through monopolization of a long-lasting, non-diminishable (bohnanza-like) food source. Ecological factors that have been shown to constrain social evolution (despite other traits that would favour it) are (a) a short breeding season that may allow only one cycle of solitary breeding but no overlapping generations (e.g. facultative eusociality in halictid bees depending on altitude and latitude; Eickwort et al. 1996; Field et al. 2010), and (b) long life spans of individuals relative to duration of the nest, which selects for maintenance of behavioural flexibility (phenotypic plasticity) because inclusive fitness benefits from alternative social strategies should change substantially over time – thus totipotency of helpers should be maintained (e.g. cooperative breeding birds, mammals, lower termites and some ambrosia beetles (Alexander et al. 1991; Choe and Crespi 1997; Korb and Hartfelder 2008, this thesis).

In contrast to the evolution of sociality, there has not been much effort to establish ecological factors that facilitate the evolution of interspecific mutualism (Bourke 2011). It has been noted, however, that (i) mutualisms more commonly evolve in stable than fluctuating environments (e.g. tropical vs. temperate environments), probably because perturbations can hinder the establishment and maintenance of a stable relationship (May 1976), and (b) life-history associated with a sessile habit apparently predisposes species to evolve interspecific mutualisms (Osman and Haugsness 1981; Janzen 1985; Bourke 2011). Three underlying reasons have been proposed: First, an organism with limited dispersal abilities can greatly profit from an association with a partner that exchanges mobility (e.g., evolution of pollen and seed dispersal by animals). Second, sessile organisms often profit from mutualisms that help them acquiring nutrients since they cannot forage actively (e.g., Sachs et al. 2004). Third, being sessile assures partner fidelity and facilitates evolution of mutualism (Nowak and May 1992; Bourke 2011).

Genetic factors in the evolution of cooperation Sub-sociality, monogamy, haplodiploidy and inbreeding have been proposed to predispose for evolution of cooperation – why will be outlined in the following. High relatedness is crucial for the evolution of altruism, but it is not essential for, although facilitates, the evolution of cooperation (see above). Therefore, it must have played a particularly important role in the transition from

General Introduction

cooperative breeding (primitive / facultative eusociality; Michener 1974; Crespi and Yanega 1995)) to obligate eusociality. Cooperative breeders and eusocial societies are both characterized by overlapping parental and offspring generations and alloparental brood care, but they differ in whether the reproductive altruism of some individuals is reversible or permanent (Batra 1966; Wilson 1971). Obligatory eusociality has arisen independently in at least 24, mostly invertebrate, lineages, whereas cooperative breeding evolved many more times in vertebrates (Ligon and Burt 2004; Crozier 2008; Bourke 2011). Current data suggests that permanently sterile eusocial castes only evolved via (i) sub-social colony foundation and (ii) life-time monogamy (Hughes et al. 2008; Boomsma 2009). First, sub-social group formation by a mother and her offspring facilitates the evolution of altruism, because of high within-group relatedness compared to the alternative semi- social group formation of non-siblings, where relatedness is usually low or insecure. Second, if nest foundation is by a monogamous mother, relatedness between daughters and daughter’s own offspring is equal (average r = 0.5; also in haplodiploid systems with unbiased sex-ratios), which implies that any constraint on individual reproduction can favor the evolution of staying, helping and finally sterility (Boomsma 2009). Despite that, multiple queen-mating can be found in many obligatory eusocial clades, but this has been shown to be a derived trait, which evolved after the transition to obligate worker sterility (Hughes et al. 2008; Boomsma 2009).

In haplodiploid species fertilized, diploid eggs develop into females, whereas unfertilized, haploid eggs develop into males. This leads to relatedness asymmetries within subsocial groups of monogamous mothers, which has been proposed to promote social evolution by kin-selection (haplodiploidy hypothesis; (Hamilton 1964): Males pass all their genes to their daughters (r = 1). Full sisters share more genes with each other (r = 0.75), then they would share with their own offspring (r = 0.50). Therefore, if there are opportunities to help relatives producing more offspring, females should prefer to raise sisters compared to selfish reproduction (Hamilton 1964). However, this is only the case under female biased sex-ratios, because under non-biased sex-ratios higher relatedness with sisters (r = 0.75) would be offset by lower relatedness with brothers (r = 0.25) (Trivers and Hare 1976). Female-biased colony sex-ratios compared to population sex-ratio may arise through partial bivoltinism (two overlapping generations per year) and split sex ratios (females contribute different sex-ratios to the same offspring generation) (Grafen 1986; Pamilo 1991; Bourke and Franks 1995). Inbreeding in haplodiploids would not change relatedness asymmetries because relatedness between sisters and to offspring increases to the same degree (Craig 1982). Hence, although inbreeding favours sociality by kin selection as long as local competition between relatives does not outweigh the benefits of altruism (Hamilton 1975; West et al. 2002), inbreeding does not change the proposed role of haplodiploidy (i.e. under female biased colony sex-ratios) for social evolution. Finally, asymmetries in relatedness caused by haplodiploidy that normally facilitate the evolution of female biased helping behaviour (Hamilton 1964; Alexander 1974) are reduced by inbreeding, which increases the likelihood of male helpers to evolve (e.g. thrips; Hamilton 1972; Chapman et al. 2000).

In contrast to intraspecific cooperation the absence of indirect fitness gains (kin selection) in interspecific interactions led to the currently accepted view that mutualisms are reciprocal exploitations that nonetheless provide net benefits to each partner (e.g. Maynard-Smith and Szathmáry 1995; Herre et al. 1999). The interests of interacting partners are hardly ever the same, creating potential for conflicts of interest that shape or destabilize associations. Foster and Wenseleers ( 2006) identified three key factors for the evolution of mutualism in their theoretical model: (1) high benefit to cost ratio, (2) high within-species relatedness (within a species feedbacks benefit those that caused them or close relatives) and (3) high between species fidelity (species stay 4

General Introduction

together long enough for feedbacks to benefit those that caused them). Positive relatedness is essential for all feedback benefits between the partners. Increasing species number on one side (or guild) has the same effect as reducing within species relatedness. An individual will only provide aid to another species, if the individual itself or other positively related individuals receive the feedback aid (Foster and Wenseleers 2006).

A brief introduction to insect sociality and fungus agriculture

“A well-flavoured vegetable is cooked, and the individual is destroyed; but the horticulturist sows seeds of the same stock, and confidently expects to get nearly the same variety…. Thus I believe it has been with social insects: a slight modification of structure, or instinct, correlated with the sterile condition of certain members of the community: consequently the fertile males and females of the same community flourished, and transmitted to their fertile offspring a tendency to produce sterile members having the same modifications.” C. Darwin, Origin of Species 1859.

Over the last decades hymenoptera emerged as the main model system for studying social evolution, because the order consists of species in varying states of sociality, from solitary breeding to obligatory eusociality. It is now clear that genetic (e.g. haplodiploidy, inbreeding), ecological constraints on independent reproduction and benefits of philopatry can predispose individuals to be selected for higher sociality. However, despite a decade of research, there is still discordance of the general importance of genetic factors, like haplodiploidy, relative to ecological factors (e.g. Korb and Heinz 2008). Ecological factors were often underestimated or not studied in insects, while genetic relatedness was sometimes overestimated (e.g. Korb and Lenz 2004).

Insect agriculture is one of the best examples for the success of cooperation in evolution (Farrell et al. 2001; Mueller and Gerardo 2002; Mueller et al. 2005). In total, there are about 3500 species of beetles (the weevil subfamilies Scolytinae and Platypodinae; Farrell et al. 2001), ants (220 described species in the subfamily Myrmicinae; e.g. Weber 1972) and termites (about 330 species; e.g. Aanen et al. 2002) known to grow, tend and harvest fungi, which apparently span the whole range of inter- and intraspecific cooperative relationships (e.g. Wilson 1971; Kent and Simpson 1992; Kirkendall et al. 1997; Hölldobler and Wilson 2010). All three insect farming lineages have evolved the highest levels of intra- and interspecific cooperation, namely eusociality and mutualism with fungi. Advanced fungiculture appears not possible without intraspecific cooperation (Mueller et al. 2005). Obligatory eusociality is the rule in the attine ants and the macrotermitine termites, and has been found in one species of ambrosia beetles up to now (Austroplatypus incompertus; Kent and Simpson 1992). Division of labor sets the stage for the origin of fungiculture in fungus-growing ants and termites (Mueller et al. 2005). Ambrosia beetles originate from solitary or colonial ancestors, and their fungus agriculture may have coevolved with sociality (Kirkendall et al. 1997; Mueller et al. 2005; Peer and Taborsky 2007). They live inside trees, a habitat which extraordinarily favors the evolution of inbreeding, haplodiploidy and sociality of arthropods (Hamilton 1967; 1978). The reason is that a tree often offers a huge and safe food-resource for the tiny exploiters that they can settle and defend for multiple generations. Furthermore, wood is typically low on nutrients (Hunt and Nalepa 1994) and high on poisonous secondary metabolites, including alkaloids, terpenes, and phenols (Highley and

General Introduction

Kirk 1979), which has repeatedly selected for symbioses with decomposing and detoxifying microorganisms (e.g. ambrosia beetles, termites; Hunt and Nalepa 1994; Choe and Crespi 1997; Grunwald et al. 2010). Living within wood has apparently fostered at least eight independent origins of fungiculture in ambrosia beetles (Farrell et al. 2001; Hulcr et al. 2007). Hence, they represent a unique model system to study the evolution of sociality in relation to fungiculture. Interestingly, ambrosia beetles vary in their mating system (inbreeding vs. outbreeding species) and ploidy level (haplodiploid vs. diploid species), which are factors that have been assumed to contribute to social evolution, although their respective roles in social evolution are controversial (see above).

Three types of methods have been suggested for the analysis of ultimate evolutionary questions for the study of adaptation (modified from Crespi and Choe (1997)): (i) functional design analyses encompass descriptive, correlative data collection and manipulative experiments (Williams 1966), (ii) measurement of selection involves quantification of phenotypes and fitness components (Lande and Arnold 1983) and (iii) comparative analysis and phylogenetics. Phylogenetically based tests allow inference of evolutionary trajectories, and thus may help disentangling causes from effect. Among ambrosia beetle species or clades, effects of inbreeding, haplodiploidy or other ecological variables may be analyzed, for example, through comparisons of sister taxa that differ in their social system.

Before such comparative studies can be conducted, however, one needs to establish some basic knowledge on the beetle-fungus interactions, as well as the social system. Four years ago, little was known about the cooperative behaviours within ambrosia beetle nests and whether there is any fungus gardening behaviour of the beetles. Associated fungi had been identified in only a few species of ambrosia beetles. Studies about the role fungal associates play in the nutrition of ambrosia beetles had been missing, and it had not been investigated whether ambrosia beetles can selectively favour or inhibit the growth of specific fungi. However, this knowledge is crucial for the understanding of cooperation inside the beetle galleries. Observations or indications rather than rigorous or scientific analysis had dominated many articles on the social and farming behaviour of ambrosia beetles published in the last 150 years – this is not surprising because it is almost impossible to study these beetles within their natural habitat. Laboratory observation techniques have allowed me to observe the beetles in detail, and study functions of social behaviours, as well as roles of fungi by performing correlational studies and manipulative experiments. Finally, two-year consecutive laboratory rearing enabled artificial selection on dispersal (a correlate of sociality) and its associated traits.

Ambrosia beetles are not only a great model system for evolutionary questions (Crespi and Choe 1997), but invasive fungus-associated beetles pose also a huge threat on naive forests that are already stressed by climate change (Hulcr and Dunn 2011; Six et al. 2011; Choi 2011).

Literature Aanen DK, Eggleton P, Rouland-Lefévre C, Guldberg-Frøslev T, Boomsma JJ & Rosendahl S (2002) The evolution of fungus-growing termites and their mutualistic fungal symbionts. PNAS 99: 14887-14892.

Alexander RD (1974) The Evolution of Social Behavior. Annual Review of Ecology and Systematics 5: 325-383.

Alexander RD, Noonan KM & Crespi BJ (1991) The evolution of eusociality. The Biology of the Naked Mole Rat (Sherman PW, Jarvis JUM & Alexander RD, eds), pp. 1-44. Princeton University Press, Princeton.

Batra SWT (1966) Nests and social behaviour of halictine bees of India. Indian J Enthomology 28: 375-393.

General Introduction

Boomsma JJ (2009) Lifetime monogamy and the evolution of eusociality. Philosophical Transactions of the Royal Society B-Biological Sciences 364: 3191-3207.

Bourke AFG (2011) Principles of Social Evolution. Oxford University Press, Oxford.

Bourke AFG & Franks NR (1995) Social Evolution in Ants. Princeton University Press, Princeton.

Brosnan SF & de Waal FBM (2002) A proximate perspective on reciprocal altruism. Human Nature-An Interdisciplinary Biosocial Perspective 13: 129-152.

Chapman TW, Crespi BJ, Kranz BD & Schwarz MP (2000) High relatedness and inbreeding at the origin of eusociality in gall-inducing thrips. PNAS 97: 1648-1650.

Choe JC & Crespi BJ (1997) The Evolution of Social Behaviour in Insects and Arachnids. Cambridge University Press, Cambridge UK.

Craig R (1982) Evolution of Eusociality by Kin Selection - the Effect of Inbreeding Between Siblings. J. Theor. Biol. 94: 119-128.

Crespi BJ & Choe JC (1997) Explanation and evolution of social systems. The Evolution of Social Behavior in Insects and Arachnids (Choe JC & Crespi BJ, eds), Cambridge University Press, Cambridge.

Crespi BJ & Yanega D (1995) The definition of eusociality. Behav. Ecol. 6: 109-115.

Crozier RH (2008) Advanced eusociality, kin selection and male haploidy. Australian Journal of Entomology 47: 2-8.

Darwin C (1859) The origin of species. Gramercy Books, New York.

Eickwort GC, Eickwort JM, Gordon J & Eickwort MA (1996) Solitary behavior in a high altitude population of the social sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Behav. Ecol. Sociobiol. 38: 227-233.

Emlen ST (1982) The evolution of helping. I. An ecological constraints model. Am. Nat. 119: 29-39.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Field J, Paxton RJ, Soro A & Bridge C (2010) Cryptic Plasticity Underlies a Major Evolutionary Transition. Current Biology 20: 2028-2031.

Foster KR & Wenseleers T (2006) A general model for the evolution of mutualisms. J. Evol. Biol. 19: 1283-1293.

Frank SA (1995) Mutual policing and repression of competition in the evolution of cooperative groups. Nature 377: 520-522.

Grafen A (1986) Split sex ratios and the evolutionary origins of eusociality. J. Theor. Biol. 122: 95-121.

Grunwald S, Pilhofer M & Holl W (2010) Microbial associations in gut systems of wood- and bark-inhabiting longhorned beetles [Coleoptera: Cerambycidae]. Systematic and Applied Microbiology 33: 25-34.

Hamilton WD (1963) Evolution of Altruistic Behavior. Am. Nat. 97: 354-&.

Hamilton WD (1964) The genetical evolution of social behaviour. I+II. J. Theor. Biol. 7: 1-52.

Hamilton WD (1967) Extraordinary sex ratios. Science 156: 477-488.

Hamilton WD (1972) Altruism and related phenomena, mainly in social insects. Ann. Rev. Ecol. Syst. 3: 193-232.

General Introduction

Hamilton WD (1975) Innate social aptitudes of man: an approach from evolutionary genetics. Biosocial Anthropology (Fox R, ed), pp. 133-155. John Wiley and Sons.

Hamilton WD (1978) Evolution and diversity under bark. Diversity of insect faunas (Mound LA & Waloff N, eds), pp. 154-175. Blackwell, Oxford.

Herre EA, Knowlton N, Mueller UG & Rehner SA (1999) The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends Ecol. Evol. 14: 49-53.

Highley TL & Kirk TK (1979) Mechanisms of Wood Decay and the Unique Features of Heart-Rots. Phytopathology 69: 1151-1157.

Hölldobler B & Wilson EO (2010) Der Superorganismus. Springer-Verlag, Berlin, Heidelberg.

Hughes WOH, Oldroyd BP, Beekman M & Ratnieks FLW (2008) Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science 320: 1213-1216.

Hulcr J, Kolarik M & Kirkendall LR (2007) A new record of fungus-beetle symbiosis in Scolytodes bark beetles (Scolytinae, Curculionidae, Coleoptera). Symbiosis 43: 151-159.

Hunt JH & Nalepa CA (1994) Nourishment and Evolution in Insect societies. Westview Press, Boulder.

Janzen DH (1985) The natural history of mutualisms. The biology of mutualism: ecology and evolution. (Boucher DH, ed), pp. 40-99. Oxford University Press, New York.

Kent DS & Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften 79: 86-87.

Kirkendall LR, Kent DS & Raffa KF (1997) Interactions among males, females and offspring in bark and ambrosia beetles: the significance of living in tunnels for the evolution of social behavior. The Evolution of Social Behavior in Insects and Arachnids (Choe JC & Crespi BJ, eds), pp. 181-215. Cambridge University Press.

Korb J & Hartfelder K (2008) Life history and development - a framework for understanding developmental plasticity in lower termites. Biological Reviews 83: 295-313.

Korb J & Heinze J (2008) The ecology of social life: a synthesis. Ecology of social evolution. (Korb J & Heinze J, eds), pp. 245-259. Springer, Berlin Heidelberg.

Korb J & Lenz M (2004) Reproductive decision-making in the termite Cryptotermes secundus (Kalotermitidae) under variable food conditions. Behav. Ecol. 15: 390-395.

Lande R & Arnold SJ (1983) The Measurement of Selection on Correlated Characters. Evolution 37: 1210-1226.

Ligon DJ & Burt DB (2004) Evolutionary origins. Ecology and Evolution of Cooperative Breeding in Birds (Koenig WD, ed), pp. 5-34. Cambridge University Press, Cambridge.

Lin N & Michener CD (1972) Evolution of sociality in insects. Q. Rev. Biol. 47: 131-159.

May RM (1976) Models for two interacting populations. Theoretical ecology: principles and applications. (May RM, ed), pp. 49-70. Blackwell Scientific Publications, Oxford.

Maynard-Smith J & Szathmáry E (1995) The major transitions in evolution. Oxford University Press, Oxford.

Michener CD (1974) The Social Behavior of Bees. Harvard University Press, Cambridge.

Mueller UG & Gerardo N (2002) Fungus-farming insects: Multiple origins and diverse evolutionary histories. Proceedings of the National Academy of Sciences of the United States of America 99: 15247-15249.

General Introduction

Mueller UG, Gerardo NM, Aanen DK, Six DL & Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563-595.

Nowak MA & May RM (1992) Evolutionary Games and Spatial Chaos. Nature 359: 826-829.

Osman RW & Haugsness JA (1981) Mutualism Among Sessile Invertebrates - A Mediator of Competition and Predation. Science 211: 846-848.

Pamilo P (1991) Evolution of the Sterile Caste. J. Theor. Biol. 149: 75-95.

Peer K & Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. Behav. Ecol. Sociobiol. 61: 729-739.

Sachs JL, Mueller UG, Wilcox TP & Bull JJ (2004) The evolution of cooperation. Q. Rev. Biol. 79: 135-160.

Stacey PB & Ligon JD (1991) The benefits-of-philopatry hypothesis for the evolutioni of cooperative breeding: variation in territory quality and group size effects. Am. Nat. 117: 831-846.

Trivers RL & Hare H (1976) Haplodiploidy and the evolution of the social insects. Science 191: 263.

Vehrencamp SL (1983a) A model for the evolution of despotic versus egalitarian societies. Anim. Behav. 31: 667-682.

Vehrencamp SL (1983b) Optimal degree of skew in cooperative societies. Am. Zool. 23: 327-335.

Weber NA (1972) Gardening ants: The attines. American Philosophical Society, Philadelphia.

West SA, Pen I & Griffin AS (2002) Conflict and cooperation - Cooperation and competition between relatives. Science 296: 72-75.

Williams GC (1966) Adaptation and natural selection. Princeton University Press, Princeton, New Jersey.

Wilson EO (1971) The insect societies. Belknap Press of Harvard University Press, Cambridge.

General Introduction

Chapter 1

Larval helpers and age polyethism in ambrosia beetles

Peter H. W. Biedermann1 and Michael Taborsky

Department of Behavioral Ecology, Institute of Ecology and Evolution, University of Bern, CH-3012 Bern, Switzerland

Edited by Bert Hölldobler, Arizona State University, Tempe, AZ, and approved September 1, 2011 (received for review May 14, 2011)

Division of labor among the workers of insect societies is a which is a habitat extraordinarily favoring social evolution (12), conspicuous feature of their biology. Social tasks are commonly apparently having fostered at least seven independent origins of shared among age groups but not between larvae and adults with fungiculture in beetles (13). Hence, they represent a unique model completely different morphologies, as in bees, wasps, ants, and system to study the evolution of sociality in relation to beetles (i.e., Holometabola). A unique yet hardly studied holome- fungiculture. Interestingly, ambrosia beetles vary in their mating tabolous group of insects is the ambrosia beetles. Along with one system (inbreeding vs. outbreeding species) and ploidy level tribe of ants and one subfamily of termites, wood-dwelling ambrosia (haplodiploid vs. diploid species), which are factors that have been beetles are the only insect lineage culturing fungi, a trait predicted to assumed to contribute to social evolution, although their favor cooperation and division of labor. Their sociality has not been respective roles in social evolution are controversial (1). The fully demonstrated, because behavioral observations have been ambrosia beetle subtribe Xyleborini is characterized by regular missing. Here we present behavioral data and experiments from inbreeding, haplodiploidy, and fungiculture (8, 14). High within nests of an ambrosia beetle, Xyleborinus saxesenii. Larval and relatedness and haplodiploidy in combination with an extremely adult offspring of a single foundress cooperate in brood care, gallery female-biased sex ratio are factors predisposing them to advanced maintenance, and fungus gardening, showing a clear division of labor sociality (1). Additionally, cooperation in fungi-culture is likely between larval and adult colony members. Larvae enlarge the gallery because a single individual can hardly maintain a fungus garden on and participate in brood care and gallery hygiene. The cooperative its own (10). Indeed, it was shown that adult female offspring effort of adult females in the colony and the timing of their dispersal delay dispersal from their natal gallery, which results in an overlap depend on the number of sibling recipients (larvae and pupae), on of eggs, larvae, pupae, and at least two generations of adult the presence of the mother, and on the number of adult workers. offspring within a colony (8). A helper effect of philopatric females This suggests that altruistic help is triggered by demands of brood has been indicated by the fact that the number of staying females dependent on care. Thus, ambrosia beetles are not only highly social that do not reproduce correlates positively with gallery but also show a special form of division of labor that is unique among productivity in Xyleborinus saxesenii Ratzeburg (8). Behavioral holometabolous insects. observations of ambrosia beetles within their galleries have been missing so far, however, because it is virtually impossible to altruism | cooperative fungiculture | insect agriculture | larval workers | mutualism observe them in nature inside the wood. The only report on eusociality in ambrosia beetles is not based on behavioral data, ivision of labor enhances work efficiency and is fundamental but reproductive roles have been inferred by destructive sampling Dto the biological evolution of social complexity (1). This of active nests of Austroplatypus incompertus (9). To facilitate conspicuous feature of social insects largely explains their eco- observations of beetle behaviors inside galleries, we developed logical success (2). Task specialization in insects usually occurs artificial observation tubes to contain entire colonies of between different age groups. Individuals either pass through reproducing beetles (15, 16). Here we use this breeding technique consecutive molts during development (Hemimetabola; e.g., of X. saxesenii to ask (i) whether offspring produced in a gallery termites, aphids), with immatures resembling small adults and engage in alloparental brood care and fungus maintenance, (ii) division of labor occurring typically between such larval nymphs whether different types of individuals specialize in divergent tasks, that may show morphological specializations [e.g., first-instar and (iii) whether decisions to help and to disperse relate to the soldier aphids (3)]; or individuals metamorphose by dramatically number of potential beneficiaries and the number of potential reorganizing their morphology during the pupal stage (Hol- workers present in the colony. Furthermore, we evaluate ometabola; e.g., bees, wasps, ants, beetles), which predestines experimentally (iv) whether female dispersal depends on the larvae and adults to specialize in different tasks because of their presence of an egg-laying foundress, because her removal should morphological and physiological differentiation. Indeed, larvae of affect the need for alloparental care. We compare our results with the eusocial Hymenoptera, for example, may produce nest- the patterns of sociality known from other major insect taxa. building silk [weaver ants (4)] and may supply adults with extra enzymes and nutrients [trophallaxis in several wasps and ants (2, Results 5, 6)]. However, all known cooperative actions of holometabolous Age- and Sex-SpecificBehavior. larvae are apparently completely under adult control (2), and despite the potential for the evolution of highly specialized Gallery maintenance and brood care were allocated differently immature helper morphs, larvae of these taxa are largely im- between different age classes within a nest(Fig.1andTable S1). The mobile and helpless and in need of being moved, fed, and cared gallery was extended mainly by larvae (digging), which also reduces the spread of mold (SI Text and Fig. S1), whereas fungus for by adults (7). In ambrosia beetles, however, in which co- care (cropping) and brood protection (blocking) were exclusively operative breeding (8) and eusociality (9) also have been assumed, conducted by adults. Blocking was only larvae can move and forage independently in the nest, which provides great potential for division of labor between larvae and adults. Division of labor sets the stage for the origin of fungiculture in Author contributions: P.H.W.B. and M.T. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B. and M.T. wrote the paper. fungus-growing ants and termites (10). Ambrosia beetles originate from solitary or colonial ancestors, and their fungus agriculture The authors declare no conflict of interest. may have coevolved with sociality (8, 10, 11). However, the role of This article is a PNAS Direct Submission. 1 division of labor is unknown. Ambrosia beetles live inside trees, To whom correspondence should be addressed. E-mail: [email protected].

Chapter 1

Fig. 1. Division of labor and age polyethism between age and sex classes in X. saxesenii. Bars show the mean (±SE) proportions of time larvae, teneral females, mature females, and males performed different cooperative tasks. Statistically significant differences between the classes are denoted by different letters (P < 0.05; GEE, details in Table S1). Note scale differences between A-C and D-F. performed by foundresses and mature females, and in 33 of 35 cases only bodies by allogrooming. In males, allogrooming was frequently followed one individual blocked at a time. In the 25 cases (in 19 of 93 galleries) when by a mating attempt; it may thus serve also to obtain information about we did not observe a blocking female, the following behavioral scan female mating status. revealed that larvae had crawled out of the gallery. At least 71 larvae (x = 3.7 larvae per gallery) died in this manner, which suggests that an Adult Female Behaviors Depending on Gallery Composition. important function of blocking is to prevent larvae from getting accidentally lost. The proportion of time adult females exhibited cropping and allogrooming Larvae and adults took different shares in hygienic behaviors. Only and the occurrence of blocking were higher during gallery stages when larvae compressed dispersed waste into compact balls (balling; Movie S1). brood dependent on care was present in the colony (preadult and larval- Frass-balls, pieces of wood, or parts of dead siblings were moved within adult gallery stages) than during other times (postlarval gallery stage; Table the gallery and pushed out of the entrance (shuffling; Movie S2) mainly by S2). Shuffling was shown equally often before, during, and after larval mature females but to some extent also by larvae and teneral females (Fig. presence within a gallery, whereas adult female digging tended to increase 1 and Table S1). The ultimate waste disposer was always the mature per capita after pupation of the last larva. Cannibalism was not shown female that was blocking the entrance. Cannibalism was directed toward before the hatching of adult daughters, and it occurred more often when adult beetles (1 case), pupae (3 cases), and larvae (37 cases) that were dependent offspring were still present. Dispersal of adult females seemed already dead (in most cases) or that did not respond to being groomed. In to be contingent on brood care demands: it significantly increased in the 10 such cases the groomer (larva or adult) would use its mandibles to open d after a major proportion of larvae had completed their development the body of the groomed sibling within seconds. Allogrooming was shown relative to the 10 d before (Wilcoxon: z = -3, P < 0.001, n = 23; Fig. S3). by all stages and both sexes, and it seemed to be crucial for individual Because brood numbers likely correlate with fungus productivity, dispersal survival: in an experiment with 10 pupae that were singly kept either with could be triggered by alternative factors, however, like the quality of the one or six larvae, the pupae survived in five of five cases with six larvae, but wood/medium or fungus. Nevertheless, in summary there are hints that they survived in only one of five cases with one larva (Fisher exact test: P = some behavioral tasks of adult females functionally relate to the demands 0.048, n = 10; details in SI Text and Fig. S2). In the other four cases with one of care-dependent brood. larva, the pupae were overgrown and killed by fungi. In addition, the body In a second analysis we tested for the relationship between adult female surface of single foundresses that had not yet successfully established a behaviors and the numbers of adult females and brood (pupae and larvae) brood was overgrown by a fungal layer (mainly Paecilomyces variotii and present in the galleries during the larval-adult gallery stage only (Table S2). Fusarium merismoides) within a few weeks, which caused the death of at Analyzed per capita, adult female digging, shuffling, cannibalism, and least 7 of 29 solitary females. Whenever individuals of different stages blocking were all independent of the numbers of adult females and encountered each other, they removed the visible fungal layer on each younger nestmates. However, per capita allogrooming and fungus cropping other's activities significantly increased with increasing numbers of pu-

Chapter 1 pae and larvae, whereas with increasing adult female numbers brood produced by the latter. Dispersal propensity of adult females (i) allogrooming significantly decreased. Adult female dispersal correlated correlates positively with low numbers of dependent young and high negatively with the number of dependent offspring [generalized estimation numbers of adult workers, and (ii) is increased by experimental removal of equation (GEE): P = 0.003; Fig. 2A and Table S2] and positively with adult the foundress. This suggests that philopatry may be related to indirect female numbers (GEE: P < 0.001; Table S2). fitness benefits, because the group members sharing a gallery are all very In 34 of 43 galleries with mature offspring, females delayed their highly related (for another xyleborine beetle see ref. 14). There is no dispersal from the natal nest after maturation. This philopatric period (i.e., morphological differentiation among adult females, and they are all fully the latency from the first appearance of a mature female within a gallery capable of breeding and establishing their own gallery. However, until the first dispersal event) correlated positively with the average dissections of all females in a number of field galleries showed that number of dependent brood (eggs, larvae, pupae) per adult female present daughters cobreed in their natal gallery in only a quarter of colonies (17). In summary, the social and reproductive patterns of X. saxesenii conform to during this period (Spearman rank correlation: RS = 0.379, P = 0.027, n = 34). primitive eusociality, defined by overlapping generations, cooperative brood care, and some reproductive division of labor, despite totipotency in Effect of Foundress Removal. reproduction of all individuals.

Eggs are produced primarily by the foundress, and in 4 of 16 galleries Division of Labor Between Larvae and Adults. dissected in the field eggs were produced also by at least one daughter (on average reproduced 23.9% of all females in these 4 galleries: range, 2-4 Task sharing in X. saxesenii is unequal between the sexes and age classes egg-layers on 4-22 females in total; details in ref. 17). If fewer eggs are pro- for most of the social behaviors described (Fig. 1 and SI Text). Regarding duced, the need for alloparental care declines. Therefore, we predicted gallery hygiene, for example, larvae form balls from dispersed frass, that the number of dispersing daughters will increase if the foundress is whereas the transport and removal of these balls is mainly performed by experimentally removed from the gallery. Our experimental interference adult females. Gallery extension is almost exclusively accomplished by raised the dispersal activity of daughters (relative to the same galleries larvae. A social role of larvae has never been reported in ambrosia beetles, before the manipulation; Fig. 2B) in the treatment (Wilcoxon: z = -2.637, P and gallery maintenance and fungiculture have been attributed solely to = 0.008, n = 10) and in the control groups (Wilcoxon: z = -1.82, P = 0.034, n the foundress (11, 15, 18). Larvae might serve a cooperative function not = 10). Removal of the foundress, however, raised the dispersal of only in Xyleborini but also in other beetles, as anecdotal observations and daughters much more strongly than the control situation (i.e., removal of speculations suggest from the Scolytinae [e.g., Dendroctonus sp.(19); the the medium without the foundress; U test: z = -3.708, P < 0.001, n = 10 + Xyloterini (20)], Platypodinae [e.g., Trachyostus ghanaensis (21); 10). Doliopygus conradti (22)], and Passalidae [Passalus cornatus (23)]. The ultimate cause for the larval specialization in digging may relate to their Discussion repeated molting: as the mandibles gradually wear during excavation (for wood-dwelling termites see ref. 24), adult females showing extensive Here we report on division of labor and age polyethism in Coleoptera, digging behavior would suffer from substantial long-term costs. In contrast, which apparently relate to the fungiculture in ambrosia beetles. It confirms larval mandibles regenerate at each molt. the predicted high degree of sociality of ambrosia beetles (9, 11). In X. Teneral and mature females take over fungus care. Blocking of the saxesenii, all group members contribute to the divergent tasks of gallery gallery entrance is done exclusively by mature females and almost only by maintenance and brood care, and there is correlative evidence that female the foundress (see ref. 14 for similar observations in Xylosandrus offspring delay their dispersal depending on the number of potential germanus). Direct brood care by allogrooming is performed by all age beneficiaries present in their natal gallery. Tasks are typically shared classes. Males groom females at high rates, which may primarily serve differentially between larval and adult colony members, which resembles courtship because mating attempts are always initiated by allogrooming(cf. the division of labor reported from termites and other highly social insect 20). Males do not take part in other social behaviors except for low levels taxa (Table 1). Among holometabolous insects, however, X. saxesenii is the of digging and cropping. Asymmetries in relatedness caused by only species to date known to exhibit an active behavioral task haplodiploidy should favor females to become helpers (25, 26). Inbreeding specialization between larvae and adults. Furthermore, adult daughters in can nevertheless reduce relatedness asymmetries and thus favor also this species specialize in different tasks than the colony foundress, and males to help (27). Male soldiers and brood-caring males have been they seem to adjust their work effort flexibly to the varying size of the documented in haplodiploid and sib-mating thrips and Cardiocondyla ants, respectively (27). In Xyleborini, very few males are produced [approximately 5-12% of offspring (28)], and males do not seem to contribute significantly to gallery function, hygiene, and fungus growth. Instead, they seem to specialize in their reproductive role by continually attempting to locate and fertilize their sisters. Nevertheless, both male larvae and adults are fully capable of performing all of the behaviors exhibited by females, as demonstrated in galleries that contain only males [approximately 2% of X. saxesenii galleries are founded by unfertilized females that produce solely male broods (16, 28)]. The life-history trajectory of X. saxesenii is most similar to that of social aphids and some termite families, where individuals serve as helpers (e.g., workers and soldiers) in their natal colonies before maturation (Table 1). Either sex may help in these taxa, and their flexible developmental period allows individuals to adjust the length of their immature stage to maximize their

Fig. 2. (A) Adult female dispersal correlates negatively with the number of cared brood in the gallery (larvae and pupae; GEE: P = 0.003; statistical details in Table S2). (B) Effects of removal of the blocking foundress. The numbers of dispersing X. saxesenii adult females (medians and quartiles) are shown 4 d before and 4 d after the experimental treatments. Wilcoxon tests: *P < 0.05, **P < 0.01; Mann-Whitney U test: ***P < 0.001; sample sizes were 10 control galleries and 10 galleries from which the foundress was removed.

Chapter 1

Table 1. Forms of division of labor in exemplary species of the most prominent social insect taxa

inclusive fitness. They may either remain in an immature helper phase at In Xyleborini, ideal conditions for the evolution of both mutualism and the nest or develop into adults that can disperse or reproduce in their natal altruism seem to prevail: (i) they are well pre-adapted to brood care colony (3, 29-32). Likewise, the developmental period of second- and third- because they originate from a beetle lineage (Scolytinae) with parental instar larvae of X. saxesenii can vary between 4 and 17 d (16). In addition, care (38); (ii) they breed in isolated galleries that are founded by one there is a striking similarity of these taxa in ecology and mating patterns. reproductive that dominates offspring production [i.e., the foundress (17)]; They all inhabit defensible nests, often within wood, which favors local (iii) they are haplodiploid and mate predominantly among full siblings [e.g., mating and inbreeding – conditions claimed as strongly favoring social inbreeding coefficients of approximately 0.9 in Xylosandrus germanus evolution (12). In many of these cases helping tasks do not seem to curtail Reiter (8)], which increases relatedness within natal colonies; and (iv) they a helper's future reproduction (i.e., because helping is risk free and does disperse solitarily after maturation to reproduce elsewhere. In addition, (v) not reduce a helper's energy stores), which may weaken the tradeoff colony members benefit greatly from cooperation due to their dependence between helping and future reproduction (33, 34). on fungi-culture; and (vi) the wood used as a resource for shelter and substratum for fungi is virtually non-depreciable, which renders resource Role of Individual Selection and Kin Selection. competition negligible. Larval digging, for example, which might be regarded as a by-product of selfish larval feeding, reduces competition We defined a behavior as cooperative if social partners potentially benefit because it generates a common good (i.e., space and substratum for from its performance, independently of whether this behavior entails net fungiculture). Similarly, other group members benefit from gallery hygiene costs to the actor (35). This definition includes (i) mutualistic behaviors that resulting as a by-product from the seemingly selfish adult feeding activities are regarded as selfish acts generating benefits to other individuals cropping and cannibalism. By contrast, balling, shuffling, allogrooming, (common goods) as a by-product (36), and (ii) altruistic behaviors that bring blocking, and adult digging may rather be altruistic. These behaviors ap- about net costs to the actor, which are compensated through indirect parently benefit other group members at the expense of time and energy fitness benefits via kin-selection; they will thus only evolve in groups of costs to the actors, without immediate benefits to the latter. Particularly relatives (25). In the course of social evolution and task specialization dangerous are allogrooming and blocking, because they expose the shaped by kin selection, mutualistic behaviors may lose their original performers to pathogens, parasites, and predators. function and change into truly altruistic behaviors (36). High relatedness Adult X saxesenii females showed a strong incentive to stay and and spatial separation between groups are very favorable to the evolution cooperate in a productive natal nest. Partly this may be of cooperative behaviors (12), as long as local competition between relatives does not oppose this force (37).

Chapter 1 selfish, because up to one quarter of females may get the chance to Adult female behaviors depending on gallery composition. To check for potential reproduce (17), and females might build up reserves during their effects of gallery age on behavior, we discriminated between the following philopatric period for subsequent dispersal and nest foundation. An successive stages of gallery development: (i) preadult gallery stage: founder experimental study in the ambrosia beetle Xyleborus affinis suggests, female and dependent offspring (larvae and pupae) with or without eggs present (n = 2 galleries); (ii) larval-adult gallery stage: all age classes (mature/ teneral however, that females do not build up reserves during their philopatric beetles, larvae, and pupae) with or without eggs present (n = 3 1 galleries); and period but rather suffer direct fitness costs (39). Alternatively, indirect (iii) postlarval gallery stage: only adults without eggs, larvae, or pupae present (n fitness benefits may be involved in helping to raise siblings (8). Three = 16 galleries). Variation in sample size of galleries was caused by the fact that not results suggest that adult female cooperation is triggered by the all age classes were visible in every gallery. We used a first series of GEEs to demands of brood dependent on care: (i) helping effort of adult identify the effect of the particular stage of gallery development on the females rose with a greater number of brood dependent on care, (ii) proportion of time adult females (= teneral and mature females) spent with a adult females dispersed at a higher rate with an increasing number of certain behavior. In a second GEE series performed only with data of the larval- workers in the colony, and (iii) dispersal rate of females increased in adult gallery stage, we analyzed whether and how task performance of adult response to experimental removal of the foundress, which indicates females related to the number of adult females present and to the number of pupae and larvae present in the gallery (Table S2). that delayed dispersal of females is not primarily motivated by the Female dispersal. For each gallery we measured the retention period between potential to breed in the natal gallery (see also ref. 8). the first appearance of a mature daughter (fully sclerotized, ready to disperse) In conclusion, the high degree of sociality in ambrosia beetles seems within the gallery and the first female dispersal event (i.e., the female had left to result from a combination of four major factors: (i) parental care as a through the gallery entrance and sat on top of the medium, where it tried to preadaptation for the evolution of sociality in the ancestors of modern escape through the cap). Using a Spearman rank correlation analysis we tested ambrosia beetles (38); (ii) very high relatedness within families due to whether the dispersal delay interval related to the mean number of offspring haplodiploidy and inbreeding; (iii) a proliferating, monopolizable attended, divided by the mean number of adult females present at that time. resource providing ample food for many individuals, which needs to be Influence of foundress on offspring dispersal. In the entrance tunnel foundresses tended and protected; and (iv) high costs of dispersal (for another either block (sit still and fully close the tunnel) or move back and forth when shuffling material to the dumps. To determine the influence of the foundress' scolytine beetle see ref. 40) due to the difficulties of finding a suitable presence on the behavior and dispersal propensity of mature females, we host tree, of nest foundation, and a successful start of fungiculture (11, experimentally removed the first centimeter of the entrance tunnel when a 16), which render predispersal cooperation particularly worthwhile. X. female was present in there (n = 11 galleries), at a stage when eggs, larvae, and saxesenii larvae are predisposed to assume certain tasks like balling of adult daughters were present together with the foundress. We determined the frass and gallery enlargement (digging) because of their body reproductive status of the removed female by dissection to check whether we morphology and the frequent renewal of mandibles by molting. Thus, had successfully removed the foundress. We excluded one gallery from the behavioral tasks are shared between larval and adult stages. This has treatment group, where we found an immature female blocking instead of the not been shown for beetles, and the described division of labor foundress. In the control group (n = 10 galleries) we removed the first centimeter between immature and adult stages seems to be unique among of medium when no female was present in this part of the entrance tunnel. Experimental galleries were randomly assigned to treatment and control groups. holometabolous social insects at large. Dispersal of the daughters was measured in both groups for 4 d before and 4 d after the intervention by collecting the females on top of the medium that had Materials and Methods left the gallery through the entrance and tried to disperse through the cap of the Study System. X. saxesenii galleries are founded by individual females that tube (Fig. 2B). transmit spores of the species-specific ambrosia fungus Ambrosiella sulfurea Batra in their gut from the natal to the new gallery (41, 42). This fungus forms a yellow Statistics. We used a series of GEEs [lmer in R (43)], which are an extension of layer of fruiting cells on the surface of gallery walls (Fig. S4A). After landing on a generalized linear models with an exchangeable correlation structure of the tree trunk, the foundress excavates a straight tunnel into the xylem with a small response variable within a cluster (= gallery identity), to analyze effects of egg niche at its end. As soon as fungus beds emerge, she feeds on them and starts dependent variables on correlated binary response variables (proportional data egg-laying (16). During offspring development the egg niche is enlarged to a single were transformed to binary data) and to identify factors affecting the relative flat brood chamber of up to a few square centimeters in size and ss1-mm height. behavioral frequencies per class (larvae, adult females, and males) (44). First, we There, most of the fungus garden grows, and individuals of all age classes live in tested whether the larvae, teneral females, mature females, and males show close contact with each other (Fig. S4C). In this study we bred X. saxesenii in glass different tendencies to express the cooperative behaviors (Fig. 1 and Table S1). tubes filled with artificial medium that mainly contained sawdust (for details on Second, we compared these frequencies between foundresses and their mature this method see SI Text and ref. 16). daughters (SI Text and Table S4). In a third series of GEEs we determined the influence of a particular developmental stage of the gallery (preadult, larval-adult, Behavioral Recordings and Analyses. In total, 93 of roughly 500 galleries were and postlarval gallery stages) on the relative behavioral frequency per class (Table founded successfu lly in the laboratory (i.e., eggs were laid), and in 43 of these S2). Finally, we modeled whether larvae and adult female numbers affected the galleries individuals reached adulthood. These galleries were used for behavioral relative frequencies of cooperative behaviors in adult females (Table S2). For the analyses. We distinguished 11 behaviors (Table S3), which comprised seven removal experiment we compared behavioral frequencies and dispersal activity cooperative behaviors that apparently raise the fitness of colony members. Every between the groups using Mann-Whitney U tests, and within groups using second to third day we performed scan observations of all individuals visible Wilcoxon matched-pairs signed-ranks tests (Fig. 2B). All statistical analyses were within a gallery (details in SI Text). performed with SPSS version 15.0 and R version 2.8.1 (43). Age- and sex-specific behavior. We used GEEs (details below) with an exchangeable correlation structure of the response variable within a cluster (= ACKNOWLEDGMENTS. The previous work of K. Peer provided the basis of this gallery identity) to identify effects of the four classes of individuals (larvae, teneral study. We thank C. Arnold and S. Knecht for help in a pilot study, and U.G. females, mature females, males) on the proportion of time spent with a certain Mueller, T. Turrini, R. Bergmueller, K. Peer, and two anonymous reviewers for behavior, using binomial error distributions (Fig. 1 and Table S1). helpful comments on the manuscript. P.H.W.B. was partially supported by the Roche Research Foundation and a DOC grant from the Austrian Academy of Science.

Chapter 1

1. Bourke AFG (2011) Principles of Social Evolution (Oxford Univ Press, Oxford). 28. Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus 2. Holldobler B, Wilson EO (1990) The Ants (Belknap Press of Harvard Univ Press, Cambridge, saxesenii Ratzeburg (Scolytinae, Coleoptera). Zookeys 56:253-267. MA). 29. Thorne BL (1997) Evolution of eusociality in termites. Annu Rev Ecol Syst 28:27-54. 3. Stern DL (1997) Foster WA. Social Behaviour in Insects and Arachnids, eds Choe JC, Crespi BJ 30. Crespi BJ (1997) Mound LA. The Evolution of Social Behaviour in Insects and Arachnids, eds (Cambridge Univ Press, Cambridge, U.K.), pp 150-165. Choe JC, Crespi BJ (Cambridge Univ Press, Cambridge, U.K.), pp 166-180. 4. Wilson EO, Holldobler B (1980) Sex differences in cooperative silk-spinning by weaver ant 31. Chapman TW, Crespi BJ, Perry SP (2008) Ecology of Social Evolution, eds Korb J, Heinze J larvae. Proc Natl Acad Sci USA 77:2343-2347. (Springer, Berlin), pp 57-83. 5. Ishay J, Ikan R (1968) Food exchange between adults and larvae in Vespa orientalis F. Anim 32. Korb J (2008) Ecology of Social Evolution, eds Korb J, Heinze J (Springer, Berlin), pp 151- Behav 16:298-303. 174. 6. Hunt JH, Baker I, Baker HG (1982) Similarity of amino-acids in nectar and larval saliva - the 33. Queller DC, Strassmann JE (1998) Kin selection and social insects. Bioscience 48: 165-175. nutritional basis for trophallaxis in social wasps. Evolution 36:1318-1322. 34. Korb J (2008) Ecology of Social Evolution, eds Korb J, Heinze J (Springer, Berlin), pp 7. Wilson EO (1971) The Insect Societies (Belknap Press of Harvard Univ Press, Cambridge, MA). 245-259. 8. Peer K, Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in 35. Brosnan S, de Waal F (2002) A proximate perspective on reciprocal altruism. Hum Nat ambrosia beetles. Behav Ecol Sociobiol 61:729-739. 13:129-152. 9. Kent DS, Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus 36. Lin N, Michener CD (1972) Evolution of sociality in insects. Q Rev Biol 47:131 -159. (Coleoptera: Platypodidae). Naturwissenschaften 79:86-87. 37. Hamilton WD (1975) Biosocial Anthropology, ed Fox R (John Wiley & Sons, New York), 10. Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR (2005) The evolution of agriculture pp 133-155. in insects. Annu Rev Ecol Evol Syst 36:563-595. 38. Jordal BH, Sequeira AS, Cognato AI (2011) The age and phylogeny of wood boring weevils 11. Kirkendall LR, Kent DS, Raffa KF(1997) The Evolution of Social Behavior in Insects and and the origin of subsociality. Mol Phylogenet Evol 59:708-724. Arachnids, eds Choe JC, Crespi BJ (Cambridge Univ Press, Cambridge, U.K.), pp 39. Biedermann PHW, Klepzig KD, Taborsky M (2011) Costs of delayed dispersal and al- 181 -215. loparental care in the fungus-cultivating ambrosia beetle Xyleborus affinis Eichhoff 12. Hamilton WD (1978) Diversityof Insect Faunas, eds Mound LA, Waloff N (Blackwell, (Scolytinae: Curculionidae). Behav Ecol Sociobiol 65:1753-1761. Oxford), pp 154-175. 40. Garraway E, Freeman BE (1981) Population-dynamics of the juniper bark beetle 13. Farrell BD, et al. (2001) The evolution of agriculture in beetles (Curculionidae: Scolytinae Phloeosinus neotropicus in Jamaica. Oikos 37:363-368. and Platypodinae). Evolution 55:2011 -2027. 41. Batra LR (1966) Ambrosia fungi: Extent of specificity to ambrosia beetles. Science 153: 193- 14. Peer K, Taborsky M (2005) Outbreeding depression, but no inbreeding depression in 195. haplodiploid Ambrosia beetles with regular sibling mating. Evolution 59:317-323. 42. Francke-Grosmann H (1975) Zur epizoischen und endozoischen Übertragung der 15. Saunders JL, Knoke JK (1967) Diets for rearing the ambrosia beetle Xyleborus ferrugineus symbiotischen Pilze des Ambrosiakäfers Xyleborus saxeseni (Coleoptera: Scolitidae). (Fabricius) in vitro. Science 15:463. Entomologica Germanica 1:279-292. 16. Biedermann PHW, Klepzig KD, Taborsky M (2009) Fungus cultivation by ambrosia beetles: 43. R Development Core Team (2008) R: A Language and Environment for Statistical Behavior and laboratory breeding success in three xyleborine species. Environ Entomol Computing (R Foundation for Statistical Computing, Vienna, Austria). 38:1096-1105. 44. Zeger SL, Liang KY, Albert PS (1988) Models for longitudinal data: A generalized estimating 17. Biedermann PHW, Peer K, Taborsky M (2011) Female dispersal and reproduction in the equation approach. Biometrics 44:1049-1060. ambrosia beetle Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Mitt Dtsch Ges 45. Crosland MWJ, Traniello JFA, Scheffrahn RH (2004) Social organization in the dry-wood Allg Angew Entomol. termite, Cryptotermes cavifrons: Is there polyethism among instars? Ethol Ecol Evol 16:117- 18. Francke-Grosmann H (1967) Symbiosis, ed Henry SM (Academic Press, New York), pp 132. 141 -205. 46. Machida M, Miura T, Kitade O, Matsumoto T (2001) Sexual polyethism of founding 19. Deneubourg JL, Gregoire JC, LeFort E (1990) Kinetics of larval gregarious behaviour in the reproductives in incipient colonies of the Japanese damp-wood termite Hodo- bark beele Dendroctonus micans (Coleoptera: Scolytidae). J Insect Behav 3:169-182. termopsisjaponica (Isoptera: Termopsidae). Sociobiology38:501-512. 20. Wichmann HE (1967) Die Wirkungsbreite des Ausstossreflexes bei Borkenkäfern. Anzeiger 47. Crosland MWJ, Ren SX, Traniello JFA (1998) Division of labour among workers in the für Schädlingskunde. J Pest Sci 40:184-187. termite, Reticulitermes fukienensis (Isoptera: Rhinotermitidae). Ethology104:57-67. 21. Roberts H (1968) Notes on biology of ambrosia beetles of genus Trachyostus Schedl 48. Badertscher S, Gerber C, Leuthold RH (1983)Polyethism in food-supply and processing in (Coleoptera - Platypodidae) in West Africa. Bull Entomol Res 58:325-352. termite colonies of Macrotermes subhyalinus (Isoptera). Behav Ecol Sociobiol 12: 115-119. 22. Browne FG (1963) Notes on the habits and distribution of some Ghanaian bark beetles and 49. Benton TG, Foster WA (1992) Altruistic housekeeping in a social aphid. Proc R Soc ambrosia beetles (Coleoptera: Scolytidae and Platypodidae). BullEntomol Lond B Biol Sci 247:199-202. Res 54:229-266. 50. Crespi BJ (1992) Behavioral ecology of Australian gall thrips (Insecta, Thysanoptera). 23. Miller WC (1932) The pupa-case building activities of Passalus cornutus Fab. (La- J Nat Hist 26:769-809. mellicornia). Ann Entomol Soc Am 25:709-712. 51. Choe JC, Crespi BJ (1997) The Evolution of Social Behaviour in Insects and Arachnids 24. Roisin Y (2000) Termites: Evolution, Sociality, Symbioses, Ecology, eds Abe T, Bignell DE, (Cambridge Univ Press, Cambridge, U.K.). Higashi M (Kluwer Academic, Dordrecht, The Netherlands), pp 95-119. 52. Lindauer M (1953) Division of labour in the honey bee colony. Bee World 34:63-73. 25. Hamilton WD (1964) The genetical evolution of social behaviour. I. J Theor Biol 7: 53. Korb J, Heinze J (2008) Ecology of Social Evolution (Springer, Berlin). 1 -16. 26. Alexander RD (1974) The evolution of social behavior. Annu Rev Ecol Syst 5:325-383. 27. Hamilton WD (1972) Altruism and related phenomena, mainly in social insects. Annu Rev Ecol Syst 3:193-232.

Chapter 1

survival of pupae, and that (ii) six digging larvae should be more successful than one digging larva to hinder the spread of mold on the chamber walls. Survival of pupae was significantly affected by the number of coinhabiting larvae (Fisher exact test: P = 0.048, n = 10). All five pupae Supporting Information survived in the chambers with six larvae, whereas only one of the five pupae survived in the chamber with one larva (Fig. S2). Pupae died because they were overgrown by fungi. Biedermann and Taborsky 10.1073/pnas.1107758108 The percentage of chamber wall area covered with mold after 14 d was significantly lower when the chamber was inhabited by six larvae (median 24.1%) than by one larva [median 90.5%; generalized SI Text estimation equation (GEE): coefficient ± SE = -3.072 ± 0.163, z = 18.86, P Activity of Individuals Dependent on Light and Gravity. All behavioral < 0.001, n = 26; Fig. S1]. Mold fungi covered 93% of the chamber area in observations of Xyleborinus saxesenii in this study were done under a two chambers that were left empty for 14 d. microscope with a 6-W artificial light source. For storage, however, In summary, this experiment revealed the importance of (i) larval tubes were wrapped in paper to keep the gallery dark as if in wood. allogrooming for the survival of pupae and of (ii) larval digging to hinder Therefore, we tested whether the changing of light conditions and of the spread of mold fungi within the gallery. the axis of gravity would affect beetle activity. Five galleries during the larval-adult gallery stage were used to obtain Behaviors of Foundresses vs. Mature Daughters. Usually the foundress five activity measures after three different treatments: the numbers of cannot be distinguished from her mature daughters once the latter are active and inactive individuals (larvae and adults combined) in each fully sclerotized. In six galleries, however, the foundress could be gallery were counted (i) right after uncovering the tubes, (ii) after 30 distinguished from her mature daughters because they remained lighter min of exposure to a 6-W light source, and (iii) after 60 min of light than their mothers for an extended period. In these galleries we made exposure, right after changing the axis of gravity by 90°. 10 focal observations of foundresses and 14 of mature daughters (10 Separate pairwise comparisons (Wilcoxon matched-pairs signed-ranks min each). GEE models showed that foundresses and their mature test) of the five observations within each gallery were combined in a daughters differed significantly in the frequency of cooperative 2 metaanalysis according to Sokal and Rohlf (1): χ (2 x N) = -2 x ∑ ln(π). behaviors, except for allogrooming (Table S4) and cannibalism, which 2 This revealed no significant influence of the light [χ (10) < 18.31; P > were never observed by foundresses. The individual blocking the gallery 0.05] and gravity treatments (P > 0.05) on the activity of both larvae and entrance was always the foundress during the focal animal adults. observations. In the removal experiment, however, 1 of 11 dissections Behaviors of the bark beetle Ips pini also do not differ between of the blocking female revealed immature ovaries, indicating that individuals kept in the dark or exposed to light, nor between individuals daughters may also block the gallery entrance occasionally. Per capita, reared in chambers in a vertical position or in a horizontal position (2). foundresses showed more fungus cropping and shuffling behavior but less gallery extension behavior (digging) than their mature daughters. In Tendency of X. Saxesenii Larvae to Aggregate. For this experiment we summary, foundresses spent more time with cooperative behaviors removed all individuals and tunnels from seven galleries during the than there mature daughters. larval-adult gallery stage, leaving only some centimeters of fungus- infested medium within the tubes. In these seven tubes we created Details on Function of Different Cooperative Behaviors. three artificial chambers of «20 mm2 (and «1-mm height) within the medium, extending next to the glass from the top of the medium Gallery extension – digging. An important common good created by alongside the tube wall. Six to twelve second/third-instar larvae were larval digging is the enlargement of the gallery. Nevertheless, it is placed on top of the medium in each of the tubes and allowed to move probably a mutually beneficial behavior with selfish benefits for the freely. After 24 h we recorded the location of all larvae. larvae because it is also part of their feeding on fungal hyphae In six replicates, all larvae had moved inside one and the same of the penetrating the wood. The consumed wood passes the gut without any three chambers. In only one replicate had larvae split up and were sign of digestion (3, 4), however, because the enzymes required to distributed over two chambers. This suggests that X. saxesenii larvae digest wood are likely missing (5, 6). Because hyphens of several fungi show a strong tendency to aggregate. co-occur together with the ambrosia fungus in the medium/wood (especially when the gallery gets older), larval digging apparently also Effect of Larval Numbers on the Survival of Pupae and on Fungus hindered the spread of mold (see above and Fig. S1). Mold fungi did not Growth. For this experiment we removed all individuals from 13 completely overgrow chambers in the presence of larvae. They probably galleries during the larval-adult gallery stage. In each of these galleries serve the larvae as additional food source, despite the fact that these we created two flat chambers with a height of « 1 m m and an area fungi are usually toxic to arthropods. Previous studies have shown that between 22.86 mm2 and 91.89 mm2 next to one of the existing tunnels toxic secondary metabolites produced by mold fungi as a response to and next to the tube glass (to allow observations). Thereafter, one feeding by arthropods can be overcome if the arthropods feed chamber was filled with one larva and the other one with six larvae gregariously on these fungi (7-9). X. saxesenii larvae showed a strong (chambers were randomly assigned). In 5 of the 13 galleries we also tendency to aggregate within the gallery (see above). placed one pupa each in both chambers, which contained one or six The digging of adults is clearly different from feeding and solely larvae, respectively. For the next 14 d we tracked the survival of the serves gallery extension. Gallery enlargement by digging is a mutual pupae (in 5 galleries) and the appearance of mold on the walls of the benefit to colony members, because it increases the surface where the chambers (in all 13 galleries). Several fungi coexist in the medium at this ambrosia gallery stage, and mold fungi are expected to overtake if they are not controlled by larvae. Therefore, we took pictures of the chambers on day 14 after the treatment and analyzed the chamber area covered and uncovered by mold according to larval numbers, by using the program ImageJ (version 1.44p). We predicted that (i) fungal layers growing on the body surface of pupae should be removed more frequently by six allogrooming larvae than by one allogrooming larva, which may affect

Chapter 1 fungi can grow, in this way lowering within-family competition for food because females dispersed at higher rates after the removal of the and space (10). Gallery surface area positively affects fungus and thus blocking foundress (Fig. 2B). colony productivity (11-13). The contribution of adults to gallery Brood care and body hygiene - allogrooming. Allogrooming was fre- extension is negligible compared with larvae. quently observed between individuals of all stages, and its importance Fungus care – cropping. Adult females are constantly walking and was shown by (i) pupae being overgrown by fungi and dying more screening the fungal layers with their large, disk-shaped antennae (Fig. frequently in the presence of one allogrooming larva compared with six S4D). They stop frequently to move their hairy, comb-like mouthparts allogrooming larvae (see above), and (ii)the repeated observation of through the fungus (Fig. S4E), probably brushing off sprouting ambrosia single foundresses (before a brood was successfully produced) dying structures (Fig. S4B). This cropping behavior of the adults apparently due to a lack of getting groomed. Four solitary foundresses died because serves both nutrient intake and fungus care. It induces the characteristic they got adhered to the gallery wall by a fungal layer on their elytra, or ambrosial growth of the fungus, that is, the formation of copious fungi grew underneath the elytra and caused them to swell so that ambrosia cells (fruiting bodies) and sporodochia [clusters of fungus moving was prevented. Additionally, it has been reported from spores (14-18)] (Fig. S4 A and B). The presence of beetles is thus crucial ambrosia beetles that eggs do not hatch (41) and larvae die (21) in the for the growth of consumable ambrosia fungus structures (15, 19). Ad- absence of the grooming foundress. The gregarious life of all age classes ditionally, cropping is likely to prevent invasions of the ambrosia garden that constantly groom each other might also explain the puzzling low by foreign fungi and microbes (14, 20, 21). Oral secretions of X. saxesenii frequency of parasitoids and mites found in nests of ambrosia beetles contain various bacteria (e.g., refs. 22 and 23) that possibly support compared with those of their close relatives, the non-gregarious, fungus growth either by providing nutrients or by controlling the spread phloem-feeding bark beetles (35, 42, 43). Recent studies of other taxa of pathogens, as shown for bacterial mutualists of fungus-culturing ants have shown that grooming can greatly decrease the success of (24-26). parasitoids [e.g., by grooming their eggs (44)], parasites [e.g., phoretic Gallery hygiene - balling, shuffling, and cannibalism. Gallery hygiene is mites (45, 46) and entomopathogenic fungi (47, 48)]. Costs of grooming important for a flourishing fungus garden as well as a healthy colony. include the time and energy spent and the risk of parasite transmission The larval balling behavior has not been described before. Bending their to the groomer (30, 49, 50). bodies ventrally, larvae form cordlike frass into balls that can be shifted Detailed Materials and Methods. within the gallery or dumped out of the entrance more easily (Movie S1). Frass balls are shifted by shuffling (Movie S2) them toward certain Study system and laboratory breeding. Of 350 galleries of X. saxesenii areas, where they are used either to close parts of tunnels, possibly to studied in the field, the majority produced approximately 10-25 regulate humidity (14); to isolate diseased areas [so-called "death cham- dispersing individuals, 3.6% produced more than 100 individuals, and bers" (27, 28)]; or to be recycled by the fungus. We observed that the one gallery exceeded 300 individuals (51). The beetles show a strong wood chippings in the larval frass are often integrated into the fungal sexual dimorphism: the rare males [the average sex ratio is beds (see also ref. 27), which suggests that the mastication of the wood approximately 1:8 to 1:20 male/female (52)] do not fully sclerotize, stay may facilitate resource utilization by the ambrosia fungi. This would small, and are unable to fly. Most of them die in their natal gallery after explain why nitrogen from beetle excretions had been detected in the females have dispersed (52). growing fungi (29). Perhaps most importantly, frass and debris are We bred X. saxesenii in glass tubes filled with artificial medium that disposed, because space is required for fungi to grow and for beetles to mainly contained sawdust ("test medium" described in ref. 53). We used move (3, 12). Occasionally we observed sibling cannibalism, which may one founder female per tube that had either been collected in the field serve both nutrient recycling and the removal of dead and diseased or originated from a brood raised in the laboratory. Females excavated specimens. The latter is probably essential for hindering the spread of galleries and brood chambers largely along the transparent tube wall, diseases and parasites [including mold (30, 31)], which is particularly which enabled observations of activity and development (Fig. S4A). threatening for highly inbred communities (32, 33). When a gallery had been successfully established, the tube was Gallery protection - blocking. Blocking of the entrance by a gallery wrapped in paper to keep it dark as if in wood, but light could enter member is ubiquitous in ambrosia beetles, although it exposes the through the entrance at the top of the tube. The tube was kept in a blocking individual to predation, for instance by birds (34, 35) or constant light/dark cycle (11 h light/13 h dark) at 28 °C/22 °C and 70% predatory beetles (36). Parasitoids also have been reported to lay eggs humidity, and the paper wrap was only removed during observations on blocking ambrosia beetles (37). Blocking provides a variety of under a microscope (x6.4 to x16 magnification) with an artificial light essential services to the gallery members (see ref. 38 for review), like source (maximum 6 W). Behavior of scolytine beetles is affected neither regulating the microclimate, preventing larvae from falling out of the by light (kept in the dark vs. exposed to light) nor when the gallery is turned by 90°, changing the axis of gravity (see above and ref. 2). gallery, and excluding parasites, parasitoids, predators, and foreign fungi Behavioral recordings and analyses. Every second to third day we from the gallery, which are common threats in bark and ambrosia performed scan observations of all individuals visible within a gallery: beetles. Additionally, it may hinder other ambrosia beetles from we noted the gallery identity, the number of visible individuals, and the entering a proliferating gallery and foreign males from mating with respective behaviors they showed at that moment, the number of females in the gallery, which can detrimentally affect their future dispersing individuals found on the surface of the medium (where they reproductive success due to an outbreeding depression (39). Our study tried to escape through the cap), and we counted the number of eggs cannot distinguish between these potential, not mutually exclusive and pupae. In each scan all gallery parts were browsed one time for functions, but the data suggest that blocking by adult females is visible individuals as described. We did not discriminate between the essential for the safety of larvae. Larvae are very mobile, and in addition three larval instars. Pupae and adult beetles were sexed on the basis of to our own results, the only previous removal experiment of a blocking morphology and size. We discerned teneral adult females that had female in scolytine beetles we know of also resulted in the loss of larvae recently hatched and showed weak sclerotization and brownish elytra, [and eggs; species: Coccotrypes dactyliperda Fabricius (40)]. In line with and mature adult females that were fully sclerotized with dark brown to this, blocking is most common during the presence of larvae and tends black elytra. to decrease in frequency after the first female offspring has matured From each scan observation we noted the proportion of individuals (Table S2). In addition, we found that blocking might also retain workers per class (larvae, teneral females, mature females, males) performing a in the nest, respective behavior. In total we conducted 500 scan observations of 43 galleries per age class (x = 11.9 scans per gallery, range 2-40).

Chapter 1

1. Sokal RR, Rohlf FJ (1995) Biometry: The Principles and Practice of Statistics in Biological 29. Kok LT, Norris DM (1972) Symbiotic interrelationships between microbes and ambrosia Research (W. H. Freeman, New York). beetles 6. Aminoacid composition of ectosymbiotic fungi of Xyleborus ferrugineus 2. Schmitz RF (1972) Behavior of Ips pini during mating, oviposition, and larval development (Coleoptera, Scolytidae). Ann Entomol Soc Am 65:598-602. (Coleoptera, Scolytidae). Can Entomol 104:1723-1728. 30. Cremer S, Armitage SAO, Schmid-Hempel P (2007) Social immunity. Curr Biol 17: R693- 3. Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z R702. Morphol Oekol Tiere 45:275-308. 31. Wilson-Rich N, Spivak M, Fefferman NH, Starks PT (2009) Genetic, individual, and group 4. Batra LR (1963) Ecology of ambrosia fungi and their dissemination by beetles. Trans facilitation of disease resistance in insect societies. Annu Rev Entomol 54: 405-423. Kans Acad Sci 66:213-236. 32. Hamilton WD (1972) Altruism and related phenomena, mainly in social insects. Annu 5. Martin MM, Kukor JJ, Martin JS, O'TooleTE, Johnson MW (1981) Digestive enzymes of fungus- Rev Ecol Syst 3:193-232. feeding beetles. Physiol Zool 54:137-145. 33. Hamilton WD, Axelrod R, Tanese R (1990) Sexual reproduction as an adaptation to resist 6. Martin MM (1983) Cellulose digestion in insects. Comp Biochem Physiol A 75:313-324. parasites (a review). Proc Natl Acad Sci USA 87:3566-3573. 7. Rohlfs M, Churchill ACL (2011) Fungal secondary metabolites as modulators of interactions 34. Moore GE (1972) Southern pine beetle mortality in North Carolina caused by parasites and with insects and other arthropods. Fungal Genet Biol 48:23-34. predators. Environ Entomol 1:58-65. 8. Rohlfs M, Obmann B, Petersen R (2005) Competition with filamentous fungi and its 35. Kenis M, Wermelinger B, Gregoire JC (2004) Bark and Wood Boring Insects in Living Trees in implication for a gregarious lifestyle in insects living on ephemeral resources. Ecol Europe, A Synthesis, ed Lieutier F (Kluwer Academic Publishers, Dordrecht, The Entomol 30:556-563. Netherlands), pp 237-290. 9. Rohlfs M, Hoffmeister TS (2004) Spatial aggregation across ephemeral resource patches in 36. Wichmann HE (1967) Die Wirkungsbreite des Ausstossreflexes bei Borkenkäfern. Anzeiger insect communities: An adaptive response to natural enemies? Oecologia 140:654-661. für Schädlingskunde. J Pest Sci 40:184-187. 10. Bright DE (1973) The Bark and Ambrosia Beetles of California, Coleoptera: Scolytidae and 37. Beaver RA(1986) The taxonomy, mycangia and biology of Hypothenemus curtipennis Platypodidae (Univ California Press, Berkeley, CA). (Schedl), the first known cryphaline ambrosia beetle (Coleoptera: Scolytidae). Ent Scand 11. Kajimura H, Hijii N (1994) Reproduction and resource utilization of the ambrosia beetle, 17:131-135. Xylosandrus mutilatus,in-field and experiment populations. Entomol Exp Appl 71:121-132.

12. Kajimura H, Hijii N (1992) Dymamics of the fungal symbionts in the gallery system and the 38. Kirkendall LR, Kent DS, Raffa KF (1997) The Evolution of Social Behavior in Insects and mycangia of the ambrosia beetle, Xylosandrus mutilatus (Blandford) (Coleoptera, Arachnids, eds Choe JC, Crespi BJ (Cambridge Univ Press, Cambridge, U.K), pp 181-215. Scolytidae). Ecol Res 7:107-117. 39. Peer K, Taborsky M (2005) Outbreeding depression, but no inbreeding depression in 13. Tarno H, et al. (2010) Types of frass produced by the ambrosia beetle Platypus quercivorus haplodiploid Ambrosia beetles with regular sibling mating. Evolution 59:317-323. during gallery construction, and host suitability of five tree species for the beetle. J For Res 40. Herfs A (1950) Studien an dem Steinnussborkenkäfer, Coccotrypes tanganus Eggers. 16:68-75. Höfchen Briefe 3 and 1:1-57. 14. Schneider-Orelli O (1913) Untersuchungen über den pilzzüchtenden Obstbaumborkenkäfer 41. Roeper R, Treeful LM, O'Brien KM, Foote RA, Bunce MA (1980) Life history of the ambrosia Xyleborus (Anisandrus) dispar und seinen Nährpilz. Zentralblatt für Bakteriologie, beetle Xyleborus affinis (Coleoptera: Scolytidae) from in vitro culture. Great Lakes Entomol Parasitenkunde. Infektionskrankheiten und Hygiene II 38:25-110. 13:141-144. 15. Batra LR, Michie MD (1963) Pleomorphism in some ambrosia and related fungi. Trans Kans 42. Schedl W (1964) Biologie des gehöckerten Eichenholzbohrers, Xyleborus monographus Acad Sci 66:470-481. (Scolytidae, Coleoptera). Z Angew Entomol 53:411-428. 16. Batra LR (1966) Ambrosia fungi: Extent of specificity to ambrosia beetles. Science 153: 193- 43. Eichhorn O, Graf P (1974) Über einige Nutzholzborkenkäfer und ihre Feinde. Anzeiger für 195. Schädlingskunde-Journal of Pest Science 47:129-135. 17. French JRJ, Roeper RA (1972) Interactions of ambrosia beetle, Xyleborus dispar (Coleoptera, 44. Vincent CM, Bertram SM (2010) Crickets groom to avoid lethal parasitoids. Anim Scolytidae), with its symbiotic fungus Ambrosiella hartigii (Fungi imperfecti). Can Entomol Behav 79:51 -56. 104:1635-1641. 45. Buchler R, Drescher W, Tornier I (1992) Grooming behavior of Apis cerana, Apis mellifera 18. Norris DM (1979) Nutrition, Mutualism, and Commensalism, ed Batra LR (Allanheld, Osmun and Apis dorsata and its effect on the parasitic mites Varroa jacobsoni and Tropilaelaps & Company, Montclair, NJ), pp 53-63. clareae. Exp Appl Acarol 16:313-319. 19. Francke-Grosmann H (1967) Symbiosis, ed Henry SM (Academic Press, New York), pp 141- 46. Delfinado-Baker M, Rath W, Boecking O (1992) Phoretic bee mites and honeybee grooming 205. behavior. Int J Acarol 18:315-322. 20. Lengerken H (1939) Die Brutfürsorge- und Brutpflegeinstinkte der Käfer (Akademische 47. Rosengaus RB, Maxmen AB, Coates LE, Traniello JFA (1998) Disease resistance: A benefit of Verlagsgesellschaft, Leipzig). sociality in the dampwood termite Zootermopsis angusticollis (Isoptera: Termopsidae).

21. Norris DM (1993) Xyleborus ambrosia beetles—a symbiotic ideal extreme biofacies with Behav Ecol Sociobiol 44:125-134. evolved polyphagous privileges at monophagous prices. Symbiosis 14:229-236. 48. Fernandez-Marin H, Zimmerman JK, Rehner SA, Wcislo WT (2006) Active use of the 22. Haanstad JO, Norris DM (1985) Microbial symbiotes of the ambrosia beetle Xyletorinus metapleural glands by ants in controlling fungal infection. Proc Biol Sci 273: 1689-1695. politus. Microb Ecol 11:267-276. 49. Rosengaus RB, Traniello JFA (1997) Pathobiology and disease transmission in dampwood 23. Grubbs KJ, et al. (2011) Genome sequence of Streptomyces griseus strain XylebKG-1, an termites Zootermopsis angusticollis (Isoptera: Termopsidae) infected with the fungus ambrosia beetle-associated actinomycete. J Bacteriol 193:2890-2891. Metarhizium anisopliae (Deuteromycotina: Hypomycetes). Sociobiology 30:185-195. 24. Currie CR, Stuart AE (2001) Weeding and grooming of pathogens in agriculture by 50. Schmid-Hempel P (1998) Parasites in Social Insects (Princeton Univ Press, Princeton). ants. Proc Biol Sci 268:1033-1039. 51. Hosking GB (1972) Xyleborus saxeseni, its life-history and flight behaviour in New 25. Mueller UG, Gerardo N (2002) Fungus-farming insects: Multiple origins and diverse Zealand. N Z J For Sci 3:37-53. evolutionary histories. Proc Natl Acad Sci USA 99:15247-15249. 52. Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus 26. Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR (2005) The evolution of agriculture saxesenii Ratzeburg (Scolytinae, Coleoptera). Zookeys 56:253-267. in insects. Annu Rev Ecol Evol Syst 36:563-595. 27. Hubbard HG (1897) in Some Miscellaneous Results of the Work of the Division of 53. Biedermann PHW, Klepzig KD, Taborsky M (2009) Fungus cultivation by ambrosia beetles: Entomology(US Department of Agriculture Bureau of EntomologyBulletin No. 7), ed Howard Behavior and laboratory breeding success in three xyleborine species. Environ Entomol 38:1096-1105. LO (US Department of Agriculture, Washington, DC), pp 9-13. 28. Peer K, Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. Behav Ecol Sociobiol 61:729-739.

Chapter 1

Fig. S1. Effect of X. saxesenii larvae on the growth of mold fungi. The number of larvae present within a chamber negatively affected the percentage of wall area covered with mold fungi of artificially created chambers in fungus infested medium. Generalized estimation equation: coeff. ± SE = -3.072 ± 0.163, z = 18.86, P < 0.001, n = 26 chambers in 13 galleries (paired design). Box-whisker plots with medians (bold), 10th, 25th, 75th and 90th percentiles are shown.

Fig. S2. Survival of experimentally isolated pupae of X. saxesenii with one (Left) or six (Right) tending larvae present. Single pupae kept with only one larva were significantly more often overgrown by fungi and killed within 14 d compared with pupae kept with six larvae (Fisher exact test: P = 0.048, n = 10 chambers).

Chapter 1

Fig. S3. Relationship between larval numbers present in a gallery and the dispersal activity of adult females. In 23 galleries with a single offspring generation larval numbers dropped below three at 21-77 d (mean 46 d) after gallery foundation. Numbers of dispersing females were counted 10 d before and 10 d after this date. Significantly more females dispersed after than before this date, that is, when all but the last two larvae had pupated (Wilcoxon matched-pairs signed-ranks tests: z = -3, P < 0.001, n = 23 galleries). This suggests that female dispersal might be triggered by demands of larvae dependent on female care. Productivity of the fungus, which is probably decreasing at the same time (because no new eggs are laid), is a confounding factor of this analysis, however, that we could not control for. Box-whisker plots with medians, 10th, 25th, 75th, and 90th percentiles are shown.

Fig. S4. Morphology of X. saxesenii galleries and their inhabitants (beetles and fungi). (A) Morphology of a brood chamber in artificial medium, with different larval instars and a teneral female. The orange layer on the gallery walls is formed by fruiting structures of the ambrosia fungus Ambrosiella sulfurea Batra. (B) Morphology of the fungal layer depicted by scanning electron microscopy with x300 magnification. The round "balls" are fruiting structures of A. sulfurea.(C) Morphology of a brood chamber in the field, with several third-instar larvae, teneral (light brown), and mature (black) females. A thin, orange fungus layer is lining the gallery walls. (D and E) Head and mouthparts of a X. saxesenii female depicted by scanning electron microscopy with x200 and x500 magnification, respectively. ant., antenna; c.e., compound eye; la., labium; la.p., labial palp; ma., mandible; max., maxilla; max.p., maxillary palp.

Chapter 1

Table S1. Separate GEE models to examine differences (P < 0.05) between the proportion of time the different age and sex classes spent with the observed cooperative behaviors (see Fig. 1 in main text)

Chapter 1

Table S2. GEE models for examining differences (P < 0.05) between the proportion of time adult females spent with each cooperative behavior, according to stage (see footnote)

Chapter 1

Table S3. Ethogram of the observed behaviors of larvae (L), females (F), and males (M)

Table S4. Separate GEE models for examining differences (P < 0.05) between the proportion of time mature daughters and foundresses spent with cooperative behaviors

Movie S1. Balling behavior by a larval worker within the brood chamber. This sequence shows an example of balling by a third-instar larva that is surrounded by other larvae and a pupa. With this behavior larvae collect waste material (sawdust and frass) to form balls that can be shifted in the gallery or dumped out of its entrance.

Chapter 1

Movie S2. Shuffling behavior by an adult female in the entrance tunnel of a gallery. This sequence shows the typical shuffling behavior by an adult female, which serves to shift balls and pieces of waste material (sawdust and frass) out of the entrance tunnel for final disposal. Note that during this video sequence a third-instar larva (first scene: 0-12 s) and a teneral female (last scene: 55-59 s) are providing the shuffling female with new material for disposal.

Chapter 1

Chapter 2

Female dispersal and reproduction in the ambrosia beetle Xyleborinus saxesenii RATZEBURG (Coleoptera; Scolytinae)

Peter H.W. Biedermann, Katharina Peer & Michael Taborsky Abteilung Verhaltensökologie, Institut fur Ökologie & Evolution, Universität Bern

Zusammenfassung: Untersuchungen zur kooperativen Brutpflege des Ambrosiakäfers Xyleborinus saxesenii RATZEBURG (Coleoptera; Scolytinae)

Kooperative Brutpflege und überlappende Generationen bei Arthropoden sind vorwiegend in den Gruppen der Hautflügler und Termiten zu finden. Weniger bekannt ist, dass auch Ambrosiakäfer (Coleoptera; Scolytinae und Platypodinae) in sozialen Gruppen leben. Adulter Nachwuchs verbleibt vor dem Ausfliegen für einige Zeit im Nest, um sich der Brutpflege, der Nestverteidigung, sowie der Zucht von Pilzgärten zu widmen, die Ambrosiakäfer zu Nahrungszwecken an den Wänden ihrer Gangsysteme im Holz anlegen. Mit Hilfe einer optimierten Bruttechnik in Glasröhrchen dokumentieren wir überlappende Generationen von Weibchen in Galerien des Ambrosiakäfers Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Nach Erlangung der Geschlechtsreife verzögerten adulte Tochter den Ausflug für durchschnittlich 23 Tage; nach dem Tod des Gründerweibchens übernahmen einige Töchter die Galerie. Da bei X. saxesenii die Weibchen im mütterlichen Nest durch ihre Brüder besamt werden, können sie ihre Eier auch dort ablegen. In unserer Studie haben sich jedoch nur 25% der Töchter im mütterlichen Nest fortgepflanzt, unabhängig von der Brutgrösse, was darauf hinweist, dass der verzögerte Ausflug vorwiegend indirekte Fitnessvorteile bietet. Diese Annahme wird durch frühere Studien unterstützt, die zeigen, dass der Verbleib der Tochter den Bruterfolg der Mutter verbessert. Dies ist vermutlich darauf zurückzuführen, dass Pilzzucht von sozialen Gruppen effektiver bewerkstelligt werden kann. Es ist daher anzunehmen, dass Sozialität und Pilzzucht bei Ambrosiakäfern in engem Zusammenhang entstanden sind. Key-words: eusociality, insect agriculture, delayed dispersal, cooperative fungiculture, alloparental care, helping, haplodiploidy Peter H.W. Biedermann, Institut fur Ökologie & Evolution, Abteilung Verhaltensökologie, Baltzerstrasse 6, CH-3012 Bern, Schweiz, E-Mail: [email protected]

Introduction

Eusociality stands for the highest level of social organization, which is characterized by cooperative brood care, overlap of parental and offspring generations, and the presence of non-reproductive castes (BATRA 1966). Among insects it is found primarily in Hymenoptera and Isoptera (WILSON 1971). In Coleoptera, the most species-rich order of insects, sociality is surprisingly rare and eusociality has only been described for one species of Platypodid ambrosia beetle (KENT & SIMPSON 1992). Ambrosia beetles live in the heartwood of trees and cultivate fungi for food. Due to their cryptic life inside wood tunnels they are difficult to study, but a recently optimized laboratory breeding technique allows to observe their behaviour (SAUNDERS & KNOKE 1967, BIEDERMANN & al. 2009). Delayed dispersal of adult daughters, and thus overlapping generations, are typical for the ambrosia beetles Xyleborinus saxesenii and Xyleborus affinis, although daughters are fully capable of founding their own nests (PEER & TABORSKY 2007, BIEDERMANN & al. 2011). Both species are members of the Scolytinae subtribe Xyleborini that is predisposed for kin-selected sociality because of high relatedness of families within one nest, due to sib- mating and haplodiploidy. Indeed, delayed dispersing daughters in several Xyleborini have been found to engage in a multitude of cooperative behaviours (e.g. maintenance and protection of the gallery, care of fungi and brood) during their philopatric period (BISCHOFF 2004, 27

Chapter 2

BIEDERMANN 2007). Apparently, this imposes costs on their future reproduction. Mature X. affinis-daughters, for instance, produce more eggs when induced to disperse early compared to voluntarily dispersing females that have undergone a philopatric period (BIEDERMANN & al. 2011). Apart from investing in siblings, however, daughters could potentially also invest in own reproduction before dispersing from the natal nest. Here we describe the typical phenology of Xyleborinus saxesenii in laboratory galleries, which illustrates that parental and offspring generations overlap. Furthermore we dissected the ovaries of all daughters present in field galleries to determine if they reproduced in the natal nest.

Material and Methods

Collection and artificial rearing

We collected adult females of Xyleborinus saxesenii from field galleries by dissecting stumps of beech (Fagus sylvatica) with active galleries in the Spilwald forest near Bern, Switzerland. Individuals used for artificial rearing were immediately transferred to the laboratory, surface-sterilized (by submerging them first in ethanol (95%) and then in distilled water for a few seconds), and afterwards individually put in glass tubes filled with an artificial rearing medium based on sawdust and agar (modified medium; for details see BIEDERMANN & al. 2009). Typically, adult females put onto the medium immediately start to excavate a tunnel system (gallery), the walls of which they inseminate with spores of their mutualistic ambrosia fungus Ambrosiella sulfurea BATRA and several other auxiliary fungi (FRANCKE- GROSMANN 1975). After establishment of the fungus layers, which is a delicate process that is successful only in about 20% of cases, successful foundresses start to lay eggs. Offspring development and brood care can be observed when tunnels and brood chambers are constructed next to the glass of the tubes.

Gallery phenology

For our observations we selected six laboratory galleries that provided almost full insight. We observed these galleries every second day for a period of four months and recorded the numbers of eggs, larvae, pupae, immature females, mature females, and males. Females have a brownish coloration as long as they are immature and turn black when fully sclerotized / mature. Sexes are easily discernable by morphology and body size (see figures in FISCHER 1954). The dates of the first visible egg, the first mature female offspring, the first dispersal of female offspring and the death of the foundress were determined. Dispersal was defined as emergence from the gallery, i.e., when individuals were found on the surface of the medium under the cap of the tube (BIEDERMANN 2007). A single dead female appeared in one third of all galleries within the first 100 days after gallery foundation. Most likely this was the foundress, because (i) they were the oldest members of the gallery, (ii) it was always only one individual and (iii) dead adult beetles are only very rarely found inside galleries in general (pers. obs.). Therefore, the moment of 'death of the foundress' was defined as the day when a single dead female was found on the surface (dead individuals, sawdust and faeces are thrown out of the nest by adult group members). To increase the sample size for a reliable estimate of this date we used our whole sample of laboratory galleries, which comprised 66 observations in total.

Reproduction of daughters

For 16 dissected field galleries we recorded the number of eggs, larvae, pupae, immature and mature females and males. Dispersal had been monitored with dispersal traps (Peer & Taborsky 2007) in 13 of these galleries from the time the gallery was founded until it was dissected ( mean = 72 days, range = 19 -309 d). Dispersal had started in nine of these galleries in the meantime and all dispersing females were preserved. After gallery collection, females were stored in 95% ethanol. Dissection was accomplished from the dorsal surface with high precision tweezers under a binocular (6.4 x - 40 x magnification). We discriminated between immature ovaries (no oocytes visible) and egg-carrying ovaries (the four ovarioles contain one or more oocytes; see figures in FISCHER 1954). In order to determine the number of breeders in the galleries, we dissected all females collected from galleries containing eggs. For galleries without eggs, a minimum of 15 females each were dissected. Females that were accidentally damaged when the gallery was opened could not be dissected (proportion

Chapter 2 dissected females: Mean = 93.5%, Range = 19 -100%, N = 16 galleries). Thus, we consider the number of egg-layers found in a gallery as a minimum estimate. However, we found that the galleries that could not be fully dissected all had more than one egg-layer; i.e. the number of galleries with only a single breeder was correctly determined. We also dissected 45 dispersing females from four different galleries to check their reproductive status.

Results

Gallery phenology

After the successful establishment of a fungal layer on the gallery walls, females started to lay eggs between days 11 and 26 (x ± se: 18 ± 2.2; N = 6 galleries) after gallery foundation. First mature daughters appeared between 16 and 34 days (x ± se: 26.2 ± 2.8; N = 6) after the first egg-laying. These daughters delayed their dispersal from the natal nest for 17 to 38 days ( x ± se: 22.8 ± 3.2; N = 6). Foundresses were found dead on the gallery surface between 34 and 97 days (x ± se: 72.2 ± 1.8; N = 66) after gallery foundation. Two to three peaks of egg-laying, followed by two clear peaks of larval numbers (Fig.1), were found by averaging the offspring numbers of the six selected galleries every day, over the whole observation period (day 18 as the mean date of the first visible egg was set as reference for all six galleries). Egg numbers were smaller than larval numbers because (i) developmental periods are different (eggs : 5 days; 1st to 3rd larval instars: 8 - 17 days) and (ii) eggs are less visible especially at the beginning of gallery development. The first offspring peak was obviously produced by the foundress, but the others were probably the result of egg-laying daughters, as the foundress likely had died before. Note that the second phase of egg-laying started at about the same time when the first mature daughters started to disperse (Fig.1). Offspring production strongly decreased about 90 to 100 days after gallery foundation, when the medium dried out and fungal growth ceased.

Fig.1. Average phenology of six laboratory galleries of Xyleborinus saxesenii. We synchronised the phenologies of the six galleries by matching the appearance of the first eggs (see text).

Reproduction of daughters

A direct benefit of delayed dispersal might be reproduction inside the natal gallery. Four of 16 field galleries contained at least two egg-laying females, (with ovarioles containing oocytes). By contrast, all dispersing females (N = 45 females from 4 galleries) did not lay eggs, i.e. their ovarioles contained no oocytes. More than one egg-layer was only found in galleries where female dispersal had already

Chapter 2 started (N = 9). Before the onset of dispersal, there was always a single reproductive female (i.e., the foundress; N = 4). In all galleries, numbers of sub-adult offspring (only larvae and pupae) and eggs present at the time of dissection were positively associated with the number of egg layers found. This was not a by-effect of gallery size, as there was no correlation between the total number of mature females in the gallery and the number of egg-layers (Table 1).

Table 1. Correlations (Kendall’s Tau) between total number of eggs, sub-adults (larvae and pupae), mature females (fully sclerotized) and the number of egg laying females in field galleries. Significant p-values are printed in bold; α = 0.0083 (Bonferroni corrected α = 0.05/6; N = 14 -16 galleries as noted). Sub-adults Mature Egg-layers total females total total Eggs total T 0.536 0.000 0.681 p (2-tailed) 0.007 1.000 0.003 N 16 16 14 Sub-adults total T 0.199 0.638 p (2-tailed) 0.295 0.004 N 16 14 Mature ♀♀ total T 0.131 p (2-tailed) 0.556 N 14

Discussion

The first mature female offspring in X. saxesenii galleries stayed and helped in the care of fungi and brood for at least 17 days, which is longer than reported for Xyleborus affinis (7 days; ROEPER & al. 1980). Even after dispersal had started, females accumulated in the gallery, which supports the hypothesis that females usually delay their dispersal. This was already suggested by an earlier study, where we showed that delayed dispersal does not serve the accumulation of reserves, but, on the contrary, reduces the fertility of X. affinis females (BIEDERMANN & al. 2011). As these females are mature and potentially fully capable of breeding independently, delayed dispersal must entail other fitness gains if it was not maladaptive. Here we showed that, besides indirect fitness benefits to females through cooperative behaviours raising the production of close kin (BIEDERMANN 2007), some females also benefit directly by producing own offspring. The numbers of females and egg layers did not correlate with each other, however, and on average about three quarters of females were not laying eggs. Interestingly, both in the laboratory and field daughters apparently did not reproduce before the onset of dispersal in the respective galleries. Nevertheless, aggressive interactions, which would suggest competition over reproduction, have never been observed. Future studies need to explore the factors affecting dispersal, helping and breeding decisions of female Xyleborini. Sociality in ambrosia beetles probably evolved in close association with fungus agriculture in wood. Indeed, advanced sociality within Coleoptera is only known from the Scolytinae, Platypodinae, and Passalidae, which are all wood-living insects associated with microbial symbionts. Interestingly, it has been hypothesized that sociality of ants and termites also evolved within the same habitat (HAMILTON 1978).

Acknowledgements This manuscript benefitted greatly from comments of Tabea Turrini. PHWB is a recipient of a DOC fellowship of the Austrian Academy of Sciences at the Department of Behavioural Ecology, University of Bern, and was partly funded by a fellowship of the Roche Research Foundation during this research.

Chapter 2

References

BATRA, S.W.T. (1966): Nests and social behaviour of halictine bees of India. - Indian Journal of Entomology 28: 375-393. BIEDERMANN, P.H.W. (2007): Social behaviour in sib mating fungus farmers.- Master thesis, Zoological Institute, University Bern. BIEDERMANN, P.H.W., KLEPZIG, K.D. & TABORSKY, M. (2009): Fungus cultivation by ambrosia beetles: Behavior and laboratory breeding success in three xyleborine species. - Environmental Entomology 38: 1096-1105. BIEDERMANN, P.H.W., KLEPZIG, K.D. & TABORSKY, M. (2011): Costs of delayed dispersal and alloparental care in the fungus-cultivating ambrosia beetle Xyleborus affinis EICHHOFF (Scolytinae: Curculionidae). Behavioral Ecology & Sociobiology, in press. BISCHOFF, L.L. (2004): The social structure of the haplodiploid bark beetle, Xylosandrus germanus. - Diploma thesis, Zoological Institute, University Bern. FISCHER, M. (1954): Untersuchungen über den kleinen Holzbohrer (Xyleborus saxeseni). - Pflanzenschutzberichte 12: 137-180. FRANCKE-GROSMANN, H. (1975): Zur epizoischen und endozoischen Übertragung der symbiotischen Pilze des Ambrosiakäfers Xyleborus saxeseni (Coleoptera: Scolitidae). - Entomologica Germanica 1: 279-292. HAMILTON, W.D. (1978): Evolution and diversity under bark, pp. 154-175 In L.A. MOUND AND N. WALOFF (eds.), Diversity of insect faunas. - Blackwell, Oxford. KENT, D.S. & SIMPSON, J.A. (1992): Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). - Naturwissenschaften 79: 86-87. PEER, K. & TABORSKY, M. (2007): Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. - Behavioral Ecology & Sociobiology 61: 729-739. ROEPER, R., TREEFUL, L.M., O'BRIEN, K.M., FOOTE, R.A. & BUNCE, M.A. (1980): Life history of the ambrosia beetle Xyleborus affinis (Coleoptera: Scolytidae) from in vitro culture. - Great Lakes Entomologist 13: 141-144. SAUNDERS, J.L. & KNOKE, J.K. (1967): Diets for rearing the ambrosia beetle Xyleborus ferrugineus (FABRICIUS ) in vitro. - Science 15: 463. WILSON, E.O. (1971): The insect societies. - Belknap Press of Harvard University Press, Cambridge.

Chapter 2

Chapter 3

Responses to experimental selection on dispersal in a primitively eusocial beetle

Peter H.W. Biedermann1 and M. Taborsky1

1Department of Behavioural Ecology, Institute of Ecology & Evolution, University of Berne, Baltzerstrasse 6, CH-3012 Bern, Switzerland

Correspondance: [email protected]

Abstract

A fundamental goal of evolutionary biology is to explain the existence of non-reproductive helpers or workers, i.e., individuals that invest resources in helping others to reproduce.

Benefits of philopatry and constraints on independent breeding are important forces for delayed dispersal of offspring in many primitively eusocial groups. From an individual viewpoint, delayed dispersal should be favored by natural selection, if inclusive fitness gains by remaining in the natal group excel gains of independent breeding. This is likely under stable food conditions at home, high costs of independent breeding and the possibility of indirect fitness gains through helping kin.

We used the primitively eusocial ambrosia beetle Xyleborinus saxesenii as a model system in an artificial selection experiment on dispersal. In this species adult daughters show a high plasticity in philopatry at the nest, but whether timing of dispersal is heritable and associated with certain other traits is unknown. We therefore selected one line for early dispersal of daughters (dispersers) and the other line for late dispersal and staying with the mother (cooperators). Despite that laboratory breeding significantly delayed the timing of dispersal in both strains within six laboratory generations, artificial selection on dispersal was proven successful because cooperators emerged from the natal nest significantly later than dispersers. Most remarkably, foundresses in the cooperative strain also produced higher numbers of immature and adult brood, and adult females showed a higher overall activity and engagement in cooperative nest protection. In comparison to the disperser strain, these investments at home, however, traded-off with a lower feeding-rate and a reduced success when founding a nest independently. Hence, we hypothesize early and late dispersal reflect two alternative social strategies, optimized either for selfish or cooperative breeding. In ambrosia beetles both strategies are probably maintained because their individual inclusive fitness advantage depends greatly on the stability of the nest. Ambrosia beetles settle dead wood, which may be suitable for the production of multiple broods (favoring cooperators) or not even a single one (favoring dispersers).

Introduction A fundamental question in evolutionary biology is how natural selection has facilitated the evolution of individuals that refrain from dispersal and own reproduction and instead help others to reproduce (Darwin 1859; Wilson 1971; Wilson 1975). Current knowledge suggests that ecological constraints on dispersal (Emlen 1982), benefits of philopatry (Stacey and Ligon 1991) and indirect fitness payoffs (i.e. kin selection; Hamilton 1964) can facilitate the evolution of delayed dispersal, helping, and reproductive altruism (Bourke 2011). The relative importance of these factors depends on the biology and ecology of the species, as suggested by interspecific comparisons. Comparing hymenopteran species at different sociality levels, for example, has shown that the transition to eusociality, i.e. obligatorily sterile worker castes, has occurred solely in groups founded by lifetime monogamous parents and their offspring (Hughes et al. 2008; Boomsma 2009). Ecological factors and relatedness, however, cannot completely account for the variation in sociality among species, i.e., in

Chapter 3 the timing of dispersal, the frequency of helping behaviour and reproductive skew (Komdeur 2006; Charmantier et al. 2007). Among group plasticity in timing of dispersal and frequency to help are indications for underlying genetic variance and/or phenotypic plasticity (e.g. Via and Lande 1985; Price et al. 2003). The latter is expected to predominate under predictable environmental fluctuations, especially, if they occur within a life-time of an organism (Bradshaw 1965; Levins 1968; but not only, see Scheiner 1993). Conditional dispersal and cooperative brood care, for example, should provide a strong advantage in variable environments, as long as habitat deterioration can be accurately predicted from environmental cues (e.g. Ronce 2007). Nevertheless, for evolution of dispersal to occur genetic differentiation is essential and it is expected to be relatively more important (i) in constant environments if maintenance of plasticity is costly (Dewitt et al. 1998) and (ii) if the plastic response for extreme phenotypes is incomplete (Price et al. 2003). In most social species periods of social living are frequently offset by dispersal events followed by solitary phases. The relative duration of these periods depends on the timing of an individuals’ dispersal from the natal group after reaching adulthood. Delayed dispersal and staying with the mother will only be selected, if inclusive fitness benefits for an individual to remain in the natal group outweigh inclusive fitness benefits of independent breeding. This is likely when constraints on independent breeding or dispersal risks are high and individuals can gain direct or indirect benefits by group living (Stacey and Ligon 1991; Heg et al. 2004). In many species where environmental constraints on independent breeding are relatively stable or cannot be estimated prior to dispersal, individuals may rely on available cues of local resource availability or presence of brood dependent on care (Biedermann et al. 2011). Dispersal is not only important for evading adverse environmental conditions, but it is also essential for long-term survival of populations (Bullock et al. 2002). Dispersing individuals avoid costs of kin competition, inbreeding, and habitat instability (Belichon et al. 1996; Gandon 1999; West et al. 2001). Thus, dispersal allows tracking patches of favorable habitat and escape from deteriorating local conditions (Ronce 2007). On the other hand, dispersal is a costly behaviour: First, mortality during dispersal is usually increased due to adverse environmental conditions and predation (e.g. Heg et al. 2004; Ronce 2007). Second, energy allocated to dispersal capacity cannot be allocated to other functions and therefore it commonly trades off with other life-history traits like delayed maturation, reduced fecundity and shorter lifespan (e.g. migratory syndrome, see Dingle 1996; Roff and Fairbairn 2001; 2007; Dingle 2006). Dispersal by flight in insects, for example, is energetically expensive and a substantial amount of resources is allocated to the flight muscles (e.g. Kammer and Heinrich 1978; Dudley 2000). Hence, many insects, histolyse their flight muscles during reproduction (Johnson 1969; Roff 1989; Zera and Denno 1997; for Scolytinae, see Robertson 1998; Lopez-Guillen et al. 2011). A suite of life-history traits co-evolve with dispersal (Dingle 1996) and are thus expected to be genetically correlated, either by pleiotropy or by linkage disequilibrium (e.g. Fuller et al. 2005). Nevertheless, while dispersing and philopatric individuals exhibit different sets of life-history traits, they may still achieve the same lifetime reproductive success (see review in Belichon et al. 1996). In cooperatively breeding animals, dispersal is associated with a solitary settlement of new habitat, whereas non-dispersal (staying) is associated with a social life-style and cooperative brood care. Although not necessarily linked, the decision to stay is associated with the decision to help in many social species (e.g. Choe and Crespi 1997; Cockburn 1998). Several studies have shown that many fitness-related traits, like fecundity, are genetically correlated with dispersal (reviewed in Roff and Fairbairn 2001). Therefore, apart from morphological changes that go together with dispersal and non-dispersal (and a social versus a solitary life-style), other life history, behavioural and physiological traits should be correlated (Roff and Fairbairn 2001; Dingle 2006). (i) Dispersers are expected to have better abilities to colonize new habitat than staying individuals (when latter are forced to disperse). (ii) Fecundity is expected to be reduced in dispersers relative to stayers and the difference will depend on the energetic costs of adaptations for dispersal. Despite a short-term fitness disadvantage, dispersal is still preserved in a population because of its long-term fitness benefit by the colonization of new habitats. (iii) In cooperatively breeding groups, individuals with a higher dispersal propensity than average should be less inclined to engage in cooperative brood care and sustainable resource use. Among group variability for the propensity to leave and help is found in many social species. Underlying genetic differences are expected because such crucial life history 34

Chapter 3 traits should be subject to selection. Parental effort and helping behaviour, for instance, have been found to be heritable in cooperatively breeding birds (Maccoll and Hatchwell 2003; Charmantier et al. 2007). In social insects, genetic differences for helping behaviour are indicated by recent studies that have identified genes that influence social behaviour (reviewed in Robinson et al. 2008; Keller 2009). However, hitherto we lack information about the evolvability of dispersal and its correlated traits in social taxa, although this is essential for understanding group formation as the first step in social evolution. Scolytine beetles are an ideal model system for the study of the evolution of dispersal in combination with sociality. Many species may switch between a social and a solitary life within a life-time, which seems to depend largely on ecological conditions (Kirkendall et al. 1997). Sociality is particularly facilitated in the fungus farming tribe Xyleborini. They live in haplodiploid, inbred kin-groups in isolated galleries, which increases relatedness within natal colonies. Furthermore, colony members benefit greatly from cooperation due to their dependence on fungiculture in dead wood, used as a resource for shelter and substrate for fungi that is virtually non-depreciable in the short-term (Biedermann and Taborsky 2011). In the long-run, however, individuals need to maintain their dispersal abilities if degradation of the dwelling forces them to leave and start a solitary life. In the best studied species Xyleborinus saxesenii, adult offspring of a foundress usually delay dispersal from the natal nest to benefit from helping close kin and partly from own reproduction (Biedermann and Taborsky 2011; Biedermann et al. 2012). Constraints on independent breeding in this primitively eusocial species appear to be very high, as a result of high costs of dispersal (for another scolytine beetle see Garraway and Freeman 1981) due to the difficulties of (i) finding a suitable host tree, (ii) nest foundation, and (iii) a successful start of fungiculture (Kirkendall et al. 1997; Biedermann et al. 2009). Nevertheless, duration of philopatric periods of adult daughters within a nest appear highly variable (also under optimal conditions), ranging from at least twelve days after maturation to a whole life (> 60 days). During these times daughters seem to invariably engage in helping behaviour at the natal nest (Biedermann and Taborsky 2011), and up to a quarter of them also lay eggs (Biedermann et al. 2012). Despite apparently high levels of inbreeding and genetic homogeneity within galleries, this might suggest genetic variance within these family groups underlying the timing of dispersal and the tendency to engage in cooperative behaviours. Here we test the evolvability of dispersal and its associated traits and trade-offs in an artificial selection experiment with a laboratory population of X. saxesenii. If dispersal is heritable, selecting one line of daughters for early dispersal (dispersers) and one line for late dispersal (cooperators), which is a correlate of cooperation, is expected to reveal significant differences in the mean timing of offspring dispersal between both groups after some generations. Furthermore, early dispersal, relative to late dispersal, may trade-off with fecundity and be associated with traits like reduced engagement in helping behaviours at the natal nest, for example. On the other hand, dispersers may be better adapted to transfer their fungi to new galleries and found nests independently.

Material & Methods Study system and general laboratory breeding The fruit-tree pinhole borer (X. saxesenii; Scolytinae) is a temperate species with a world-wide distribution. It breeds in tunnel systems (galleries), excavated in recently dead wood of a wide variety of host trees. New galleries are always founded by individual females transmitting spores of a complex of fungi, dominated by the species-specific ambrosia fungus Raffaelea sulphurea (formerly Ambrosiella sulfurea; Ascomycota), in specialised organs (mycetangia or gut) from the natal to the new gallery (Biedermann et al. 2012). Broods are produced after a thin fungal layer of fruiting cells has established on the surface of gallery walls. Successful planting of fungus often fails, however, either because no spores of the ambrosia fungus have been transmitted or the substrate is not suitable for the spores to germinate (Biedermann et al. 2009). If planting fails the female sometimes disperses again, but mostly dies. Successful females may, however, produce large broods and lay eggs for several weeks in common chambers excavated by the larval feeding activity (on fungus infested wood; (Biedermann and Taborsky 2011). In the laboratory (25° C) first daughters will reach adulthood about 30 to 40 days after gallery foundation, will be inseminated by a brother and will remain with their sibs in the following for a period of at least two, sometimes many more weeks. If 35

Chapter 3 the woody substrate is durable, a few adult daughters seem to overtake the mother’s gallery and proceed breeding after her death (in the laboratory on average 70 to 80 days after gallery foundation; (Biedermann et al. 2012). If successful, field and laboratory galleries are comparable in size producing approximately 10-50 dispersing female offspring (Biedermann et al. 2009); galleries with several hundred dispersers are possible in the field, but exceptional (Hosking 1972). The small, wingless males are in the minority (the average sex ratio is approximately 1:8 to 1:20; (Biedermann et al. 2009). They disperse if no more offspring is produced, but can only wander the natal log and may introduce in neighbouring galleries there (Biedermann 2010). We bred consecutive laboratory generations of X. saxesenii in transparent plastic tubes filled up to two-thirds with artificial medium that mainly contained agar and sawdust (“test medium” described in Biedermann et al. 2009). One founder female was used per tube and is sat onto the medium after 3 sec rinses in bleach, 90% ethanol and distilled water, to sterilize their body surface from weed fungi. Afterwards tubes are closed by plastic caps and wrapped in paper to keep them dark as if in wood, but light could enter through the tops of the tubes. Tubes are kept in a constant light/dark cycle (13 h light/11 h dark) at 28 °C/22 °C and the paper wrap is only removed during observations under a microscope (×6.4 to ×16 magnification) with an artificial light source (maximum 6 W). Beetles excavate their tunnels largely along the transparent tube wall, which enables observations of their behaviours. Experiments showed that behaviour of scolytine beetles within their galleries is affected neither by light nor by changing the axis of gravity while doing observation (Schmitz 1972; Biedermann and Taborsky 2011). Dispersal is through the entrance tunnel and can be recorded by collection of these individuals from the top of the medium. Pilot studies suggested that periods of diapause are essential for successful long-term breeding of X. saxesenii within the laboratory (see also data on an obligate diapause in Xylosandrus germanus; Weber and McPherson 1983). Therefore and because of logistical reasons (i.e. periods when we were not able to record dispersal every day) we reduced breeding temperature during the F1 to F5 generation once for 8 weeks to 8°C and several times at irregular intervals for 2-3 weeks to 16°C. Egg- laying probably ceased during these periods, because overwintering field galleries of X. saxesenii contain only larvae, pupae and adults (Fischer 1954; unpubl. data). Beetles do not disperse below 18°C (Faccoli and Rukalski 2004); unpubl. data).

Recording of dispersal and artificial selection Our laboratory population stemmed from females collected in November 2009 from overwintering galleries in beech stumps (Fagus sylvatica) in the Spil-forest (560 m asl, 46°95’, 7°31’) close to Berne, Switzerland. We started our selection experiment with its second laboratory generation, making up 30 successfully established galleries (control group, starting generation F0). Every 1-2 days we collected all dispersing adult females that had emerged onto the medium with insect handling tweezers, transferred them to sterilized micro tubes with a moistened piece of filter paper and afterwards stored them in the refrigerator until they were either disposed or used for further breeding. In each generation we aimed to select for (i) early dispersal (= disperser strain) and (ii) remaining within the nest (= cooperator strain). In the disperser strain we used about the first third of all emerging daughters per gallery (starting generation: x = 10.6 females, range 0-35) as founder females for the next generations. The cooperator strain was made up of the last daughters remaining in the natal nest (starting generation: = 5.0 females per gallery, range 0-15). Remaining daughters were collected by gallery dissections close to the end of gallery life. In cases when galleries were already empty upon dissection, we continued breeding with the last dispersing females from the respective gallery that we had stored. After we had recorded the dispersal of the F6 generation, we used a Cox regression model to detect potential effects of long-term laboratory breeding (comparing F0 with F6) and artificial selection (cooperators vs. dispersers in F6) on the dispersal rate of daughters, by controlling for gallery identity and weather parameters for Berne (Source: MeteoSchweiz). A way to increase the incentive of adult daughters to stay within a natal nest and to continue brood care and breeding in the cooperator strain was to extend gallery durability with fresh substrate for the fungus: First, tubes were cut apart 2-3mm below the chamber. Subsequently, a second tube filled

Chapter 3 with freshly prepared, sterile and not yet solidified medium was attached to the first tube and tightly fixed (Fig. 1). We extended tubes about 70 days after gallery foundation and in a few cases a second time about 50 days later. Only about one third of tubes had visible chambers and could be extended this way, however.

Fig. 1. Illustration of the method to increase the longevity of galleries in the cooperators strain.

Behavioural recordings Behaviours were recorded by scan observations during which all gallery parts were browsed one time for inhabitants: we noted the gallery identity and counted the number of eggs and pupae, the number of visible larvae and adults and the respective behaviours they showed at that moment. In our analysis we did not discriminate between the three larval instars. Pupae and adult beetles were sexed on the basis of morphology and size. We discerned teneral adult females that had recently hatched and showed weak sclerotization and brownish elytra, and mature adult females that were fully sclerotized with dark brown to black elytra. From each scan observation we noted the proportion of individuals per class (larvae, teneral females, mature females, males) performing a respective behaviour. We decided between (i) the larval behaviours: digging (= feeding), shuffling, balling, allogrooming, cannibalism, inactivity and locomotion, (ii) the adult female behaviours: blocking, shuffling, allogrooming, cannibalism, digging, cropping fungus (= feeding), inactivity and locomotion, and (ii) the adult male behaviours: allogrooming, cannibalism, cropping fungus, mating attempt, copula, inactivity and locomotion (for detail descriptions of these behaviours see (Biedermann and Taborsky 2011)). In the starting generation (F0) we conducted nine scan observations of all 30 galleries on random days 35-80 days after gallery foundation. The selection lines in the F6 generation we scanned on 35 random days 12-80 days after gallery foundation. We intermixed strains and allocated random numbers to the 33 galleries of the cooperators strain and to the 67 galleries of the disperser strain to be blind on their identity before conducting behavioural observation. We used a series of GEE models to detect potential effects of long-term laboratory breeding (comparing F0 with F6) and artificial selection (cooperators vs. dispersers in F6), by controlling for gallery identity, gallery age and numbers of other immatures and adults present at the respective scan.

Influence of artificial selection on other parameters Apart from dispersal and behaviours we also collected data on gallery composition, enzyme profiles of the fungus gardens, sex ratios and gallery founding success. We dissected galleries in the F6 generation on day 45, 62 and 87 after gallery foundation. All eggs, larvae, pupae, males, teneral and adult females were counted and then stored in 70% ethanol. Numbers of dispersed daughters until this date were also recorded. One piece of fungus garden material from the brood chamber was removed and used for enzyme profiling (see also (De Fine Licht and Biedermann 2012)). We used a series of GLM models to detect potential effects of artificial selection (cooperators vs. dispersers) on gallery composition and the ratio between dispersed and remaining daughters. Final sex ratios of all galleries in the F0 and F6 generation were recorded and analysed whether they are affected by laboratory breeding and artificial selection using a GEE model that controlled for gallery of origin in the F0 generation. Furthermore, we recorded the founding success, i.e. the ability of a founder female to successfully produce at least one adult offspring, of the F1 and the F6 generation in our selection lines. Afterwards we used a GEE model to test the influence of long-term

Chapter 3 laboratory breeding (comparing F1 with F6) and artificial selection (cooperators vs. dispersers in F6) on the binomial success rates, by controlling for gallery identity.

Statistics We used a Cox regression model to detect potential effects of long-term laboratory breeding (comparing F0 with F6) and artificial selection (cooperators vs. dispersers in F6) on the dispersal rate of daughters, by controlling for gallery identity and weather parameters. Furthermore, we analysed several other life-history variables to detect correlative effects and trade-offs in response to our artificial selection. We used a series of GEEs (lmer in R; (R Development Core Team 2008), which are an extension of generalized linear models with an exchangeable correlation structure of the response variable within a cluster (= gallery identity or origin), to analyse effects of dependent variables on correlated binary response variables (proportional data were transformed to binary data) and to identify effects of laboratory breeding and artificial selection (1) on each of the relative behavioural frequencies, (2) on gallery founding success, and (3) on sex ratios. Furthermore, we used a series of GLM models with quasi-binomial error distribution to detect potential effects of artificial selection (cooperators vs. dispersers) on gallery composition and the ratio between dispersed and remaining daughters. All statistical analyses were performed with SPSS version 15.0 and R version 2.8.1 (R Development Core Team 2008).

Results Dispersal of adult daughters from the natal nest Our selective treatment for delayed dispersal of adult daughters (= cooperator strain) increased the likelihood of female dispersal within six generations by a factor of 1.96 in relation to the females selected for early dispersal (= disperser strain; Cox model: p = 0.005; Table 1 and Fig. 1). Half of the adult daughters had been dispersed at day 64 ( x , CI 63 – 64) after gallery foundation in the disperser strain, which is five days before it occurred in the cooperator strain ( x = 69, CI 68 – 70). Compared to the dispersal of the starting generation (= control; = 59, CI 59 – 61), however, laboratory breeding overall resulted in a later dispersal of adult daughters (control vs. disperser strain: odds-ratio = 0.48, p = 0.001; control vs. cooperator strain: odds-ratio = 0.24, p < 0.001). In general, dispersal tended to occur on certain days when barometric pressure had risen during the last 24 hours (p = 0.1).

Table 1. Results of the Cox regression model for the dispersal of adult daughters.

Cox Model1 b2 S.E.3 z p Odds4 Control (N = 226) vs. Cooperators (N = 336) -1.41 0.1 -5.29 <0.001 0.24 Control (N = 226) vs. Dispersers (N = 804) -0.74 0.08 -3.28 0.001 0.48 Cooperators (N = 336) vs. Dispersers (N = 804) 0.67 0.09 2.8 0.005 1.96 Fall vs. Rise of barometric pressure5 0.32 0.07 1.63 0.103 1.38 Wald test: χ2 = 28.3, df = 2, p = <0.001 1 Formula for the final Cox Model in R: coxph(Surv(galleryage, dispersal) ~ Treatment + cluster(Gallerynumber), data). 2 b = regression coefficient of dispersal function for variable. 3 Standard error of regression coefficient. 4 Odds ratio (= exp(b)). 5 Variable excluded from the final model.

Chapter 3

Figure 1. Philopatry of adult females within the natal nest under two selection regimes. The black dotted line refers to the dispersal of the Control galleries before the start of the selection. Cooperators refer to the 6th generation of the selected galleries that were bred for cooperation and late dispersal, respectively. Dispersers refer to the 6th generation of galleries that were selected for early dispersal. Cox regression analysis showed that in both Cooperators (odds ratio = 0.24; z = -5.29, p = <0.001) and Dispersers (odds ratio = 0.48; z = -3.28, p = 0.001) adult females delayed dispersal after six generations of laboratory breeding relative to the females in the Control galleries in the first generation. Our selective treatment increased the likelihood of philopatry in the Cooperators by a factor of 1.96 relative to the Dispersers (z = 2.8, p = 0.005).

Behavioural changes in response to the selection regime

Selecting for early and late dispersal of daughters had strong effects on their behaviours within the natal nests. As expected adult females of the cooperator strain were more active (GEE: p = 0.001 – 0.052; Table 2, Table S1) and were more often engaged in cooperative behaviours than the adult females in the disperser strain (p = 0.005) and the control group(p < 0.001), which was mainly an effect of an increased blocking frequency (p = 0.003 – 0.004; see Table S1). Cropping of the fungus was less commonly observed in the cooperator strain in relation to the dispersers strain (p < 0.001) and to the control (p < 0.001). Consecutive laboratory breeding led to an overall increase of adult female activity (= decrease of inactivity; p = 0.001 – 0.18), of cannibalism (p = 0.024 – 0.043) and locomotion (p < 0.001), but reduced their engagement in cropping of the fungus (p < 0.001).

Behaviours of larval stages were not differently affected by the two selection regimes. In relation to the control group, however, larvae in both lines decreased their activity (GEE: p = 0.002 – 0.011; Table S2) , the frequency of locomotion (p < 0.001), and tended to engage less often in allogrooming (p = 0.028 – 0.18). Male behaviours were unaffected by laboratory breeding and the selection regimes, except for allogrooming that was shown less often in the dispersers than the controls (p = 0.041; Table S3).

Chapter 3 Gallery composition in dependence of the selection regime The first offspring normally emerges from the pupal stage at a gallery-age between 30 and 40 days. In our study, significant dispersal activity started about 14 days later (see Fig. 1). Eleven cooperator and disperser galleries were dissected on day 45. Offspring composition (i.e. number of eggs, larvae, pupae, males, teneral and adult females) did not differ between the selection regimes at this gallery stage (GLMs: p < 0.05, Table S4). Adult female dispersal had only started in one exceptional gallery of the cooperator strain. Around the time when half of the daughters had dispersed from their natal gallery, at a gallery age of 62 days, we found offspring composition to differ markedly between the selected strains: Cooperator galleries (N = 11 galleries) contained on average a higher number of immature offspring (p < 0.001) and adult daughters (p < 0.001) than disperser galleries (N = 40). Latter is partly caused by the smaller number of dispersed daughters until this date in the cooperator strain (p = 0.039). Close to the end of gallery life when egg-laying had ceased, at a gallery-age of 87 days, however, the small numbers of immatures and adult females left within the galleries did not differ between cooperator (N = 13) and disperser strain ( N = 5). The small average productivity of the disperser galleries ( x = 8.8 ind.) may explain the higher absolute and relative numbers of daughters that had dispersed from the cooperator galleries until this date.

Table 2. Separate GEE models to examine differences (p < 0.05) between the proportion of time adult females spent with the observed behaviours.

Behaviour Parameter Coeff.  SE z p Direction Adult ♀♀ activity Intercept (Control) 2.91 0.44 6.63 <0.001 Contrast Control vs. Cooperators 0.92 0.29 3.13 0.001 +++ Contrast Control vs. Dispersers 0.33 0.25 1.33 0.18 Contrast Cooperators vs. Dispersers -0.58 0.3 -1.94 0.052 (-) Age of gallery -0.02 0.01 -3.08 0.002 - -

Adult ♀♀ cooperative Intercept (Control) 1.7 0.49 3.45 <0.001 behaviours (blocking, shuffling, Contrast Control vs. Cooperators 1.35 0.34 4 <0.001 +++ allogrooming and cannibalism) Contrast Control vs. Dispersers 0.5 0.31 1.6 0.11 Contrast Cooperators vs. Dispersers -0.85 0.3 -2.83 0.005 - - Age of gallery -0.08 0.01 -9.41 <0.001 - - - Number of larvae 0.06 0.01 4.78 <0.001 +++

Adult ♀♀ cropping fungus Intercept (Control) -0.66 0.27 -2.45 0.014 Contrast Control vs. Cooperators -1.28 0.16 -8.06 <0.001 - - - Contrast Control vs. Dispersers -0.67 0.14 -4.64 <0.001 - - - Contrast Cooperators vs. Dispersers 0.61 0.16 3.92 <0.001 +++ Age of gallery 0.02 0.01 3.98 <0.001 +++ Number of larvae -0.02 0.01 -2.15 0.031 -

We used GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery) to identify effects of the selection regime on the total amount of time each behaviour was observed, by controlling for influences of gallery-age and number of larvae (for the full model see Table S3). The influences of the regimes on the behavioural frequencies are displayed as contrasts between classes.

Chapter 3

Figure 2. Composition of galleries in the two selected strains at an age of 62 days after gallery foundation. Mean (± standard errors) numbers, for the cooperator and disperser galleries are given. GLM: *** - p < 0.001, * - p < 0.05 (for exact values see Table S4).

Gallery founding success

At the beginning of our selection experiment we measured the founding success (i.e. the ability to produce at least one adult daughter) of the first 269 dispersing daughters (1st generation of disperser strain) and the 115 last dispersing / remaining daughters (1st generation of cooperator strain) from the same galleries. Founding success did not differ between the strains in the first generation and also compared to the daughters in the cooperator strain after six generations (GEE: p > 0.05; Fig. 3, Table S5). Early dispersing daughters in the disperser strain, however, markedly improved their founding success until the sixth generation (p < 0.001).

Chapter 3

Figure 3. Rate of successful brood establishment by daughters, when reared independently after the first and the sixth laboratory generation and depending on the selection regime. Different letters denote significant differences in the success rates (GEEs: p < 0.05; for details see Table S4).

Discussion

The evolution and maintenance of sociality can only be understood by deciphering the interaction between genetics and environment in shaping a complex phenotype (Komdeur 2006; Charmantier et al. 2007). Our study is the first to show the response of dispersal to selection in a social taxon. It helps to understand how social phenotypes are selected over non-social phenotypes when sociality is favoured by specific environmental conditions. Adult female X. saxesenii selected for early dispersal (and hence reduced sociality) exhibited smaller offspring numbers and a shorter lifespan compared to those selected for late dispersal. In the latter, however, only 32% of breeding attempts were successful, compared to 69% in early dispersers. Under natural conditions, however, the higher success in nest foundation in early dispersers is likely offset by predation and adverse environmental conditions (for a scolytine beetle see Garraway and Freeman 1981), potentially resulting in similar reproductive outputs.

After six generations of artificial selection on early and late dispersal, we found daughters of the early disperser strain to leave on average five days before those of the philopatric strain. Such a shift was not detected, however, when we compared dispersal between each parental gallery at the beginning of the experiment with dispersal in its daughter-galleries at the end. Hence, differences between the selected strains after six generations are not a result of among-gallery selection on early and late dispersers, but these arise because of between-gallery differences in timing of dispersal. Although we did not particularly design our experiment for selection on between-gallery differences, this result is not surprising. Ambrosia beetles in the tribe Xyleborini are highly inbred and thus within- family genetic variability is very low and offering no possibility for selection. The only genetic study on a xyleborine beetle reported an inbreeding coefficient of F = 0.88 in Xylosandrus germanus, which suggests that only 3% of all matings are with non-sibs, and individuals within a gallery are almost clones (Keller et al. 2011).

The selectability of the timing of dispersal in X. saxesenii indicates that this behaviour can respond to environmental factors. A crucial environmental factor for ambrosia beetle sociality is the unpredictability of the habitat they settle. Constraints on independent breeding are always very high in these beetles because dying trees are patchily distributed (and thus difficult to locate) and the plantation of ambrosia gardens often fails. This has apparently led to females in temperate xyleborine ambrosia beetles to ubiquitously remain and help in the natal nest for a few days after reaching adulthood. Dead wood offers a habitat that is ephemeral, but for awhile highly productive. Depending on the part of the tree, which is settled, galleries may be productive for one or several beetle generations, which selects for simpler or more complex sociality, respectively. Although X.

Chapter 3 saxesenii prefers a relatively long-lived and stable woody substrate (hardwood of large diameter trees in an early stage of the degradation process; Pfeffer 1995), host quality can deteriorate quickly, e.g. due to desiccation or competition with other xylophagous organisms, forcing beetles to disperse. Such an environment with unpredictable, but often extensive durability selects for the maintenance of both early and late dispersers, since the constraints on independent breeding are high. Ambrosia beetle females dispersing early have a better chance of locating new hosts and may, therefore, be favoured by selection late dispersers. In a large-diameter tree stump that is able to nourish multiple generations of beetles, on the other hand, early dispersal would probably have higher fitness costs, as the likelihood for not finding a better resource is high. Indeed, current knowledge suggests that primitive eusociality (=cooperative breeding) in ambrosia beetles is only found under exactly this condition (e.g. X. saxesenii and X. affinis; Biedermann et al. 2011; 2012; Biedermann and Taborsky 2011). Sub-sociality, on the other hand, appears to be present in species predominantly attacking short-lived resources (e.g. Xylosandrus germanus; Bischoff 2004), like small diameter branches, whereas true eusociality has evolved only within living trees, which are a long-lasting and stable habitat (in Austroplatypus incompertus; Kent and Simpson 1992).

The cooperators strain produced females that lived longer and produced more adult offspring compared to the dispersers strain. This resembles the syndrome of mobile and sedentary phenotypes in other insects living on ephemeral resources (e.g. fruit-feeding moths; Gu and Danthanarayana 1992; Gu et al. 2006; Torriani et al. 2010). A major difference to these solitary species is, however, that X. saxesenii females may additional gain indirect fitness benefits by staying and helping in the natal nest. Our recent studies on cooperatively breeding xyleborine ambrosia beetles (X. affinis and X. saxesenii) showed that adult offspring staying in the natal group invariably engage in the following cooperative behaviours: (i) gallery protection by blocking, (ii) brood care by allogrooming, (iii) gallery hygiene by shuffling faeces and sawdust (Biedermann and Taborsky 2011), and (iv) cropping the fungal layers, which may serve the removal of fungal weeds (Norris 1979). Furthermore, (v) the mere presence of adult females (and also immatures) profits fungus growth, most likely through their excretions (Anisandrus dispar; Batra and Michie 1963; French and Roeper 1972). In summary, these behaviours all help to raise closely related kin and are expressed depending on current needs in X. saxesenii; dispersal and help are partly triggered by the amount of brood depending on care and the number of other helpers (Biedermann and Taborsky 2011).

Female activity tended to be higher (p = 0.052) in the cooperator than in the disperser strain. Additionally, we found cooperators to engage slightly more frequently in the combined cooperative behaviours blocking, shuffling, allogrooming, and cannibalism. This resulted solely from a significant increase in blocking frequency, however. The magnitude of received help is much higher in cooperators if the total help is examined: Many more females stayed and helped, for several days more, within the natal nest than in dispersers. This may be one reason why the females in the cooperator strain produced more offspring than in the disperser strain in the sixth generation. Only one of the behaviours that are regarded mutually beneficial – fungus cropping – was more commonly expressed by the females in the disperser strain. There are two possible explanations: Females following the non-social strategy (i) might have a higher food requirement, because they need to build up an energy-expensive flight apparatus, and (ii) they should have less of an interest in long- term management of their fungus cultures compared to cooperators, and thus possibly “overgrazed” their gardens. It will be interesting to compare the growth and quality (i.e. fungal species diversity) of the fungus gardens between the strains in a future study. Genetic differences have been repeatedly shown to be responsible for such distinct behavioural strategies (e.g. in birds; Sih et al. 2004) and may also underlie our findings. Also adult males exhibited significantly more allogrooming in the cooperative strain, which suggests that their behavioural changes were also triggered by genetic or non-genetic effects correlating with our selected traits.

Here we show a heritable component of dispersal (and sociality), which means that each individual can adjust its behaviour only within a certain range. There are only two other studies in cooperatively breeding birds that indicate genetic differences to partly underlie individual variation in social behaviours (i.e. parental care; Maccoll and Hatchwell 2003; Charmantier et al. 2007). Besides genetic inheritance, however, there are other modes of non-genetic inheritance. (a) The phenotype of an organism can be determined not only by its genotype and the environment the organism

Chapter 3 experiences, but also by the environment and phenotype of its mother. This may be due to mothers supplying eggs with mRNA or proteins. Insect’s phenotypes have been frequently reported to be affected by such maternal effects (e.g. Mousseau and Dingle 1991; Li and Margolies 1994) and they may also play a role in X. saxesenii. (b) Alternatively, in insects living in symbioses with microorganisms like ambrosia beetles, the symbionts might directly shape behaviours and life- histories of their hosts. If xyleborine ambrosia beetles found new galleries, they always transfer spores of the fungal symbionts from the natal nest in their mycetangia (vertical transmission; Francke-Grosmann 1956). Additionally, a few spores of commensal and parasitic fungi get usually vectored on the body surface of beetles and this fungi increase in abundance during aging of the gallery (e.g. Francke-Grosmann 1967; Beaver 1989; Kajimura and Hijii 1992). Hence, a community of microbes is continuously associated and vertically transferred from one generation to the next, probably without much horizontal (between galleries) exchange. “Inheritance” of a microbial community could give rise to repeated patterns of behavioural responses (e.g. dispersal, cooperative behaviours) and life-history traits (e.g. fecundity, life-span) within families (cf. comparable to territory inheritance in cooperatively breeding birds; Komdeur 1992; Charmantier et al. 2007). Cross- fostering experiments (exchanging offspring or fungal communities) should be used in the future to reveal the potential role of non-genetic modes of inheritance.

Despite a heritable component, there is also a huge within gallery variation in the timing of dispersal. The most likely mechanism underlying this variation is variability in individual response thresholds to environmental conditions. In this study, for example, we found evidence for increased dispersal after a rise in barometric pressure. Barometric pressure is closely associated with weather conditions and serves as a dispersal clue for many small insects (e.g. Zettel 1984; Li and Margolies 1994). Also, there are indications that the quality of the fungal resource, i.e., fungal species composition, affects dispersal (Kajimura and Hijii 1992). Furthermore, in another study we found that dispersal correlated with number of gallery inhabitants. Larvae, which depend on care, lead to delayed dispersal, whereas other helpers promoted dispersal (Biedermann and Taborsky 2011). An alternative explanation for variable dispersal is enforced philopatry through maternal manipulation. Maternal blocking of the entrance tunnel of the gallery has been shown to mainly serve (a) the protection of the gallery from intruders (e.g. predators, parasites, foreign ambrosia beetles), (b) the regulation of the microclimate, (c) preventing larvae from falling out of the gallery, and (d) possibly preventing helping daughters from dispersal (for details see Kirkendall et al. 1997; Biedermann and Taborsky 2011). Recently we found adult daughters to disperse at higher rates after the removal of the blocking female (Biedermann and Taborsky 2011). Blocking females are usually egg-layers and they might profit from the presence of the helpers (Biedermann and Taborsky 2011). In this study we found adult female blocking behaviour to significantly respond to our selective treatment. After six generations it was exhibited more frequently in the cooperators than the dispersers strain. Hence, we might have indirectly selected for high and low blocking frequencies that consequently have led to females dispersing late or early, respectively. The possible role of blocking in enforcing females to help needs to be explored in future studies.

Foundresses produced more offspring in the cooperators strain than in the dispersers strain. First, this may be a result of higher adult female activity and frequency of helping, which may have pushed fungus productivity (see above). Adult female helping behaviours are allocated to brood numbers (Biedermann and Taborsky 2011). Differences in the amount of help between strains, however, are not an artefact of the higher offspring numbers in the cooperators, because we controlled for brood numbers in our model when we compared the frequency of behaviours between the two strains. Second, foundresses may have adopted different energy allocation strategies: They may have invested either in helping and reproduction, or in dispersal. Dissections of X. saxesenii females revealed that dispersing females always have non-mature ovaries (Biedermann et al. 2012). For females in the disperser strain, on the other hand, a more efficient flight-machinery can be assumed. Consequently, they face a dispersal-reproduction trade-off, which is typically found in insects (Gu and Danthanarayana 1992; Zera and Denno 1997; Gu et al. 2006). Although scolytine beetles can apparently fully histolyse their wing muscles (e.g. Robertson 1998), this may consume energy that is later missing for reproduction. Furthermore, preparation for dispersal may be associated with a higher metabolic turnover (e.g. Hanski et al. 2006), which is generally known to constrain energy reserves and survival (e.g. Van Voorhies 2001). Still it cannot be excluded that females preparing for 44

Chapter 3 dispersal may adopt a risk-spreading strategy (Kisdi 2002) and lay some eggs before departure from the natal habitat (e.g. Hughes and Dorn 2002; Biedermann et al. 2011), and then produce less offspring after founding a new nest.

Independent of the artificial selection on timing of dispersal we found consecutive laboratory rearing to generally lead to later dispersal and behavioural changes. This suggests adaptations to laboratory conditions. Artificial medium is a nutrient-rich and relatively long-lived breeding substrate, and this apparently shifted dispersal in both strains over time. Furthermore, it led to larvae being less active, walking less and engaging less in allogrooming. Adult female activity, locomotion and cannibalism increased, whereas cropping fungus strongly decreased. The latter, together with the reduced larval allogrooming, indicates that infestation with fungal parasites decreased during consecutive laboratory rearing. In summary, laboratory conditions seem to favour the cooperative and late- dispersing strategy, as both strains were selected for later dispersal, independent of the treatment. This is probably a result of increased gallery productivity and longevity. High quality food conditions have been repeatedly shown to strongly select against a dispersers phenotype in arthropods that live on ephemeral resources (e.g. Li and Margolies 1994; Gu et al. 2006).

Many ambrosia beetles have recently developed into invasive pests of forests and fruit-tree plantations (Hulcr and Dunn 2011). Invasiveness is often correlated with a high dispersal ability (e.g. Sakai et al. 2001).Our study suggests that populations might differ in their dispersal capability and probably several other correlated life-history traits. It would be interesting to compare invasive and natal populations with regard to their dispersal and life-history traits. Deciphering the underlying heritable differences may greatly help for understanding the evolution of high dispersal abilities and invasiveness and the other extreme philopatry and social evolution. Clearly, ecological conditions play a major role. In unpredictable environments, populations are expected to exhibit polymorphisms for dispersal traits (Southwood 1977). A genetic basis for this polymorphism is well documented in many arthropod populations (Harrison 1980; Roff and Fairbairn 2001; 2007). This suggests that variation between individuals in the propensity to stay, help and breed in the natal nest does not arise solely from individuals maximizing inclusive fitness, but are partly heritable. Heritability found in this study is a prerequisite for evolution of these traits.

Literature

Batra LR & Michie MD (1963) Pleomorphism in some ambrosia and related fungi. Transactions of the Kansas Academy of Science 66: 470-481.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Belichon S, Clobert J & Massot M (1996) Are there differences in fitness components between philopatric and dispersing individuals? Acta Oecologica-International Journal of Ecology 17: 503-517.

Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus saxesenii Ratzeburg (Scolytinae, Coleoptera). Zookeys 56: 253-267.

Biedermann PHW, Klepzig KD, Ott E, Taborsky M & Six DL (2012) Dynamics of filamentous fungi in the ambrosia gardens of the primitively eusocial beetle Xyleborinus saxesenii Ratzeburg (Scolytinae: Curculionidae). in prep.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW, Klepzig KD & Taborsky M (2011) Costs of delayed dispersal and alloparental care in the fungus-cultivating ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae). Behav. Ecol. Sociobiol. 65: 1753-1761.

Biedermann PHW, Peer K & Taborsky M (2012) Female dispersal and reproduction in the ambrosia beetle Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Mitteilungen der Deutschen Gesellschaft für allgemeine und angewandte Entomologie in review.

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

Chapter 3 Bischoff, L L (2004) The social structure of the haplodiploid bark beetle, Xylosandrus germanus. Diploma thesis, Zoological Institute, Univ. Bern.

Boomsma JJ (2009) Lifetime monogamy and the evolution of eusociality. Philosophical Transactions of the Royal Society B- Biological Sciences 364: 3191-3207.

Bourke AFG (2011) Principles of Social Evolution. Oxford University Press, Oxford.

Bradshaw AD (1965) Evolutionary significance of phenotypic plasticity in plants. Advances in Genetics (Caspari EW, ed), pp. 115-157. Academic Press, New York.

Bullock JM, Kenward RE & Hails RS (2002) Dispersal ecology. Blackwell, Oxford.

Charmantier A, Keyser AJ & Promislow DE (2007) First evidence for heritable variation in cooperative breeding behaviour. Proceedings of the Royal Society B-Biological Sciences 274: 1757-1761.

Choe JC & Crespi BJ (1997) The Evolution of Social Behaviour in Insects and Arachnids. Cambridge University Press, Cambridge UK.

Cockburn A (1998) Evolution of Helping Behavior in Cooperatively Breeding Birds. Annual Review of Ecology and Systematics 29: 141-177.

Darwin C (1859) The origin of species. Gramercy Books, New York.

De Fine Licht HH & Biedermann PHW (2012) Patterns of functional enzyme activity suggest that larvae are the key to successful fungus farming by ambrosia beetles. Frontiers in Zoology in review.

Dewitt TJ, Sih A & Wilson DS (1998) Costs and limits of phenotypic plasticity. Trends Ecol. Evol. 13: 77-81.

Dingle H (1996) Migration: The biology of life on the move. Oxford University Press, New York.

Dingle H (2006) Animal migration: is there a common migratory syndrome? Journal of Ornithology 147: 212-220.

Dudley R (2000) The biomechanics of insect flight. Princeton University Press, Princeton NJ.

Emlen ST (1982) The evolution of helping. I. An ecological constraints model. Am. Nat. 119: 29-39.

Faccoli M & Rukalski JP (2004) Attractiveness of artificially killed red oaks (Quercus rubra) to ambrosia beetles (Coleoptera, Scolytidae). Invertibrati di una foresta della Pianura Padana, Bosco della Fontana. (Ceretti P, Hardersen S, Mason F, Nardi G, Tisato M & Zapparoli M, eds), pp. 171-179. 3. Cierre Grafica Editore, Verona.

Fischer M (1954) Untersuchungen über den kleinen Holzbohrer (Xyleborus saxeseni). Pflanzenschutzberichte 12: 137-180.

Francke-Grosmann H (1956) Zur Übertragung der Nährpilze bei Ambrosiakäfern. Naturwissenschaften 43: 286-287.

Francke-Grosmann H (1967) Ectosymbiosis in wood-inhabiting beetles. Symbiosis (Henry SM, ed), pp. 141-205. Academic Press, New York.

French JRJ & Roeper RA (1972) Interactions of ambrosia beetle, Xyleborus dispar (Coleoptera, Scolytidae), with its symbiotic fungus Ambrosiella hartigii (Fungi imperfecti). Canadian Entomologist 104: 1635-1641.

Fuller RC, Baer CF & Travis J (2005) How and When Selection Experiments Might Actually be Useful. Integrative and Comparative Biology 45: 391-404.

Gandon S (1999) Kin competition, the cost of inbreeding and the evolution of dispersal. J. Theor. Biol. 200: 345-364.

Garraway E & Freeman BE (1981) Population-dynamics of the juniper bark beetle Phloeosinus neotropicus in Jamaica. Oikos 37: 363-368.

Gu HN & Danthanarayana W (1992) Quantitative Genetic-Analysis of Dispersal in Epiphyas-Postvittana .2. Genetic Covariations Between Flight Capacity and Life-History Traits. Heredity 68: 61-69.

Gu HN, Hughes J & Dorn S (2006) Trade-off between mobility and fitness in Cydia pomonella L. (Lepidoptera : Tortricidae). Ecological Entomology 31: 68-74.

Hamilton WD (1964) The genetical evolution of social behaviour. I+II. J. Theor. Biol. 7: 1-52.

Chapter 3 Hanski I, Saastamoinen M & Ovaskainen O (2006) Dispersal-related life-history trade-offs in a butterfly metapopulation. Journal of Animal Ecology 75: 91-100.

Harrison RG (1980) Dispersal polymorphisms in insects. Ann. Rev. Ecol. Syst. 11: 95-118.

Heg D, Bachar Z, Brouwer L & Taborsky M (2004) Predation risk is an ecological constraint for helper dispersal in a cooperatively breeding cichlid. Proceedings of the Royal Society of London Series B-Biological Sciences 271: 2367-2374.

Hosking GB (1972) Xyleborus saxeseni, its life-history and flight behaviour in New Zealand. N. Z. J. Forest Science 3: 37-53.

Hughes J & Dorn S (2002) Sexual differences in the flight performance of the oriental fruit moth, Cydia molesta. Entomologia Experimentalis et Applicata 103: 171-182.

Hughes WOH, Oldroyd BP, Beekman M & Ratnieks FLW (2008) Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science 320: 1213-1216.

Hulcr J & Dunn RR (2011) The sudden emergence of pathogenicity in insect-fungus symbioses threatens naive forest ecosystems. Proceedings of the Royal Society B: Biological Sciences.

Johnson CG (1969) Migration and dispersal of insects by flight. Methuen, London.

Kajimura H & Hijii N (1992) Dymamics of the fungal symbionts in the gallery system and the mycangia of the ambrosia beetle, Xylosandrus mutilatus (Blandford) (Coleoptera, Scolytidae). Ecological Research 7: 107-117.

Kammer AE & Heinrich B (1978) Insect flight metabolism. Advances in insect physiology (Treherne JE, Berridge MJ & Wigglesworth VB, eds), pp. 133-228. Academic Press, London.

Keller L (2009) Adaptation and the genetics of social behaviour. Philosophical Transactions of the Royal Society B-Biological Sciences 364: 3209-3216.

Keller L, Peer K, Bernasconi C, Taborsky M & Shuker DM (2011) Inbreeding and selection on sex ratio in the bark beetle Xylosandrus germanus. submitted.

Kent DS & Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften 79: 86-87.

Kisdi E (2002) Dispersal: Risk spreading versus local adaptation. Am. Nat. 159: 579-596.

Komdeur J (1992) Importance of Habitat Saturation and Territory Quality for Evolution of Cooperative Breeding in the Seychelles Warbler. Nature 358: 493-495.

Komdeur J (2006) Variation in individual investment strategies among social animals. Ethology 112: 729-747.

Levins R (1968) Evolution in changing environments. Princeton University Press, Princeton NJ.

Li JB & Margolies DC (1994) Barometric-Pressure Influences Initiation of Aerial Dispersal in the 2-Spotted Spider-Mite. Journal of the Kansas Entomological Society 67: 386-393.

Li JB & Margolies DC (1994) Responses to Direct and Indirect Selection on Aerial Dispersal Behavior in Tetranychus-Urticae. Heredity 72: 10-22.

Lopez-Guillen G, Carrasco JV, Cruz-Lopez L, Barrera J, Malo E & Rojas J (2011) Morphology and Structural Changes in Flight Muscles of Hypothenemus hampei (Coleoptera: Curculionidae) Females. Environmental Entomology 40: 441-448.

Maccoll ADC & Hatchwell BJ (2003) Heritability of parental effort in a passerine bird. Evolution 57: 2191-2195.

Mousseau TA & Dingle H (1991) Maternal Effects in Insect Life Histories. Annual Review of Entomology 36: 511-534.

Norris DM (1979) The mutualistic fungi of Xyleborini beetles. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 53-63. Allanheld, Osmun & Company, Montclair.

Pfeffer A (1995) Zentral- und westpaläarktische Borken- und Kernkäfer (Coleoptera: Scolytidae, Platypodidae). Pro Entomologia, c/o Naturhistorisches Museum Basel, Basel. 47

Chapter 3 Price TD, Qvarnstrom A & Irwin DE (2003) The role of phenotypic plasticity in driving genetic evolution. Proceedings of the Royal Society of London Series B-Biological Sciences 270: 1433-1440.

R Development Core Team (2008) R: A language and environment for statistical computing. Vienna, Austria, R Foundation for Statistical Computing.

Robertson IC (1998) Flight muscle changes in male pine engraver beetles during reproduction: the effects of body size, mating status and breeding failure. Physiological Entomology 23: 75-80.

Robinson GE, Fernald RD & Clayton DF (2008) Genes and Social Behavior. Science 322: 896-900.

Roff DA & Fairbairn DJ (2001) The genetic basis of dispersal and migration and its consequences for the evolution of correlated traits. Dispersal (Clobert J, Danchin E, Dhondt AA & Nichols JD, eds), pp. 191-202. Oxford University Press, New York.

Roff DA & Fairbairn DJ (2007) The evolution and genetics of migration in insects. Bioscience 57: 155-164.

Roff DA (1989) Exaptation and the evolution of dealation in insects. J. Evol. Biol. 2: 109-123.

Ronce O (2007) How does it feel to be like a rolling stone? Ten questions about dispersal evolution.

Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, McCauley DE, O'Neil P, Parker IM, THOMPSON JN & Weller SG (2001) The population biology of invasive species. Annual Review of Ecology and Systematics 32: 305-332.

Scheiner SM (1993) Genetics and Evolution of Phenotypic Plasticity. Annual Review of Ecology and Systematics 24: 35-68.

Schmitz RF (1972) Behavior of Ips pini during mating, oviposition, and larval development (Coleoptera, Scolytidae). Canadian Entomologist 104: 1723-1728.

Sih A, Bell AM, Johnson JC & Ziemba RE (2004) Behavioral syndromes: An integrative overview. Q. Rev. Biol. 79: 241-277.

Southwood TRE (1977) Habitat, Templet for Ecological Strategies - Presidential-Address to British-Ecological-Society, 5 January 1977. Journal of Animal Ecology 46: 337-365.

Stacey PB & Ligon JD (1991) The benefits-of-philopatry hypothesis for the evolutioni of cooperative breeding: variation in territory quality and group size effects. Am. Nat. 117: 831-846.

Torriani MVG, Mazzi D, Hein S & Dorn S (2010) Direct and correlated responses to artificial selection on flight activity in the oriental fruit moth (Lepidoptera: Tortricidae). Biol. J. Linn. Soc. 100: 879-889.

Van Voorhies WA (2001) Metabolism and lifespan. Experimental Gerontology 36: 55-64.

Via S & Lande R (1985) Genotype-Environment Interaction and the Evolution of Phenotypic Plasticity. Evolution 39: 505-522.

Weber BC & McPherson JE (1983) Life history of the ambrosia beetle Xylosandrus germanus (Coleoptera: Scolytidae). Annals of the Entomological Society of America 76: 455-462.

West SA, Murray MG, Machado CA, Griffin AS & Herre EA (2001) Testing Hamilton's rule with competition between relatives. Nature 409: 510-512.

Wilson EO (1971) The insect societies. Belknap Press of Harvard University Press, Cambridge.

Wilson EO (1975) Sociobiology. The new synthesis. Belknap Press of Harvard University Press, Cambridge.

Zera AJ & Denno RF (1997) Physiology and ecology of dispersal polymorphism in insects. Annual Review of Entomology 42: 207-230.

Zettel J (1984) The Significance of Temperature and Barometric-Pressure Changes for the Snow Surface-Activity of Isotoma- Hiemalis (Collembola). Experientia 40: 1369-1372.

Chapter 3 Supplementary information

Table S1. Separate GEE models to examine differences (p < 0.05) between the proportion of time larval stages spent with the observed behaviours.

Behaviour Parameter Coeff.  SE z p digging / feeding Mean frequency (Control) 0.89 0.15 5.92 <0.001 Contrast Control vs. Cooperators -0.01 0.1 -0.06 0.95 Contrast Control vs. Dispersers 0.04 0.1 0.42 0.67 Contrast Cooperators vs. Dispersers 0.05 0.07 0.69 0.49 Age of gallery -0.01 0.0 -3.62 <0.001 Number of larvae -0.02 0.0 -6.96 <0.001 Number of staying adult females 0.02 0.01 1.69 0.09 gallery hygiene Mean frequency (Control) -3.1 0.16 -19.11 <0.001 (shuffling and balling) Contrast Control vs. Cooperators -0.22 0.18 -1.25 0.21 Contrast Control vs. Dispersers -0.11 0.16 -0.68 0.49 Contrast Cooperators vs. Dispersers 0.11 0.13 0.8 0.42 Age of gallery* 0.01 0.01 1.3 0.19 Number of larvae 0.02 0.01 3.53 <0.001 Number of staying adult females* -0.01 0.02 -0.67 0.51 allogrooming Mean frequency (Control) -2.28 0.12 -18.3 <0.001 Contrast Control vs. Cooperators -0.19 0.14 -1.35 0.18 Contrast Control vs. Dispersers -0.29 0.13 -2.2 0.028 Contrast Cooperators vs. Dispersers -0.1 0.1 -1 0.34 Age of gallery* -0.0 0.0 -0.64 0.53 Number of larvae 0.03 0.0 7.26 <0.001 Number of staying adult females* -0.02 0.01 -1.55 0.12 inactivity Mean frequency (Control) -1.91 0.1 -19.2 <0.001 Contrast Control vs. Cooperators 0.31 0.12 2.54 0.011 Contrast Control vs. Dispersers 0.36 0.12 3.16 0.002 Contrast Cooperators vs. Dispersers 0.06 0.08 0.66 0.51 Age of gallery* 0.0 0.0 1.01 0.31 Number of larvae* -0.0 0.0 -0.09 0.93 Number of staying adult females* -0.02 0.01 -1.46 0.15 locomotion Mean frequency (Control) -2.17 0.14 -15.5 <0.001 Contrast Control vs. Cooperators -0.49 0.15 -3.3 <0.001 Contrast Control vs. Dispersers -0.6 0.14 -4.16 <0.001 Contrast Cooperators vs. Dispersers -0.11 0.11 -1.05 0.3 Age of gallery* -0.0 0.0 -0.58 0.56 Number of larvae 0.03 0.0 5.36 <0.001 Number of staying adult females -0.03 0.01 -2.18 0.029

We used GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery) to identify effects of the selection regime on the total amount of time each behaviour was observed, by controlling for influences of gallery-age, number of larvae and number of staying adult females. The influences of the regimes on the behavioural frequencies are displayed as contrasts between classes. *variable not in the final model after step-wise model reduction using ANOVA analysis of log-likelihood scores.

Chapter 3 Table S2. Separate GEE models to examine differences (p < 0.05) between the proportion of time adult males spent with the observed behaviours.

Behaviour Parameter Coeff.  SE z p cropping fungus Mean frequency (Control) -0.81 0.32 -2.55 0.011 Contrast Control vs. Cooperators -0.42 0.39 -1.07 0.28 Contrast Control vs. Dispersers -0.06 0.35 -0.17 0.86 Contrast Cooperators vs. Dispersers 0.36 0.36 1 0.32 Age of gallery* -0.01 0.02 -0.48 0.63 Number of larvae* 0.02 0.02 1.15 0.25 Number of males* 0.06 0.08 0.72 0.47 Number of teneral females* -0.05 0.05 -0.89 0.38 Number of staying adult females -0.05 0.03 -1.71 0.088 allogrooming Mean frequency (Control) 0.25 0.9 0.28 0.78 Contrast Control vs. Cooperators 0.38 0.42 0.92 0.36 Contrast Control vs. Dispersers -0.49 0.35 -1.39 0.16 Contrast Cooperators vs. Dispersers -0.87 0.43 -2.04 0.041 Age of gallery -0.03 0.02 -1.53 0.13 Number of larvae* -0.02 0.02 -0.95 0.34 Number of males 0.13 0.08 1.57 0.12 Number of teneral females* 0.06 0.04 1.39 0.16 Number of staying adult females* -0.02 0.03 -0.6 0.55 mating attempt and Mean frequency (Control) -1.98 0.39 -5.05 <0.001 copula Contrast Control vs. Cooperators -0.12 0.48 -0.25 0.8 Contrast Control vs. Dispersers -0.04 0.43 -0.08 0.93 Contrast Cooperators vs. Dispersers 0.09 0.48 0.18 0.86 Age of gallery* 0.02 0.02 0.89 0.37 Number of larvae 0.03 0.02 1.56 0.12 Number of males* -0.12 0.14 -0.9 0.37 Number of teneral females 0.08 0.05 1.57 0.12 Number of staying adult females* 0.02 0.04 0.42 0.67 locomotion Mean frequency (Control) -0.98 0.31 -3.13 0.002 Contrast Control vs. Cooperators -0.57 0.41 -1.38 0.17 Contrast Control vs. Dispersers -0.38 0.36 -1.04 0.3 Contrast Cooperators vs. Dispersers 0.19 0.41 0.47 0.64 Age of gallery* 0.0 0.02 0.14 0.89 Number of larvae -0.04 0.02 -1.52 0.13 Number of males* -0.11 0.11 -1.02 0.31 Number of teneral females* -0.05 0.06 -0.84 0.4 Number of staying adult females* 0.03 0.03 0.95 0.34

We used GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery) to identify effects of the selection regime on the total amount of time each behaviour was observed, by controlling for influences of gallery-age, number of larvae, number of males, number of teneral females and number of staying adult females. The influences of the regimes on the behavioural frequencies are displayed as contrasts between classes. *variable not in the final model after step-wise model reduction using ANOVA analysis of log-likelihood scores.

Chapter 3 Table S3. Separate GEE models to examine differences (p < 0.05) between the proportion of time adult females spent with the observed behaviours. Behaviour Parameter Coeff.  SE z p Cooperation (blocking, Mean frequency (Control) 1.7 0.49 3.45 <0.001 shuffling, Contrast Control vs. Cooperators 1.35 0.34 4 <0.001 allogrooming, Contrast Control vs. Dispersers 0.5 0.31 1.6 0.11 cannibalism) Contrast Cooperators vs. Dispersers -0.85 0.3 -2.83 0.005 Age of gallery -0.08 0.01 -9.41 <0.001 Number of larvae 0.06 0.01 4.78 <0.001 Number of staying adult females* 0.00 0.02 0.24 0.81 blocking Mean frequency (Control) 0.27 0.54 0.49 0.62 Contrast Control vs. Cooperators 1.1 0.39 2.85 0.004 Contrast Control vs. Dispersers 0.1 0.38 0.26 0.8 Contrast Cooperators vs. Dispersers -1.01 0.33 -3.02 0.003 Age of gallery -0.05 0.01 -4.73 <0.001 Number of larvae 0.06 0.01 4.45 <0.001 Number of staying adult females -0.25 0.05 -5.32 <0.001 shuffling Mean frequency (Control) 0.33 0.64 0.51 0.61 Contrast Control vs. Cooperators 0.03 0.45 0.08 0.94 Contrast Control vs. Dispersers -0.58 0.45 -1.28 0.2 Contrast Cooperators vs. Dispersers -0.61 0.44 -1.38 0.17 Age of gallery -0.07 0.01 -4.82 <0.001 Number of larvae 0.05 0.02 2.67 0.008 Number of staying adult females -0.16 0.05 -2.95 0.003 allogrooming Mean frequency (Control) -3.78 0.34 -11.13 <0.001 Contrast Control vs. Cooperators -0.29 0.4 -0.73 0.47 Contrast Control vs. Dispersers -0.03 0.39 -0.07 0.95 Contrast Cooperators vs. Dispersers 0.27 0.41 0.65 0.52 Age of gallery* -0.01 0.01 -0.84 0.4 Number of larvae 0.03 0.02 2.13 0.033 Number of staying adult females 0.05 0.02 2.52 0.012 cannibalism Mean frequency (Control) -7.01 1.13 -6.22 <0.001 Contrast Control vs. Cooperators 2.43 1.2 2.03 0.043 Contrast Control vs. Dispersers 2.69 1.19 2.26 0.024 Contrast Cooperators vs. Dispersers 0.25 0.56 0.46 0.65 Age of gallery* -0.02 0.02 -0.84 0.4 Number of larvae* -0.03 0.05 -0.68 0.49 Number of staying adult females* 0.06 0.05 1.05 0.3 cropping fungus Mean frequency (Control) -0.66 0.27 -2.45 0.014 Contrast Control vs. Cooperators -1.28 0.16 -8.06 <0.001 Contrast Control vs. Dispersers -0.67 0.14 -4.64 <0.001 Contrast Cooperators vs. Dispersers 0.61 0.16 3.92 <0.001 Age of gallery 0.02 0.01 3.98 <0.001 Number of larvae -0.02 0.01 -2.15 0.031 Number of staying adult females 0.01 0.01 1.35 0.18 inactivity Mean frequency (Control) -2.91 0.44 -6.63 <0.001 Contrast Control vs. Cooperators -0.92 0.29 -3.13 0.001 Contrast Control vs. Dispersers -0.33 0.25 -1.33 0.18 Contrast Cooperators vs. Dispersers 0.58 0.3 1.94 0.052 Age of gallery 0.02 0.01 3.08 0.002 Number of larvae* -0.02 0.01 -1.14 0.25 Number of staying adult females* 0.01 0.01 0.81 0.42 locomotion Mean frequency (Control) -2.73 0.15 -18.4 <0.001 Contrast Control vs. Cooperators 0.77 0.2 3.94 <0.001 Contrast Control vs. Dispersers 0.87 0.2 4.47 <0.001 Contrast Cooperators vs. Dispersers 0.11 0.18 0.6 0.55 Age of gallery* 0.0 0.01 0.56 0.58 Number of larvae* -0.01 0.01 -0.96 0.34 Number of staying adult females* -0.01 0.02 -0.52 0.6

We used GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery) to identify effects of the selection regime on the total amount of time each behaviour was observed, by controlling for influences of gallery-age, number of larvae and number of staying adult females. The influences of the regimes on the behavioural frequencies are displayed as contrasts between classes. *variable not in the final model after step-wise model reduction using ANOVA analysis of log-likelihood scores. 51

Chapter 3

Table S4. Separate GLM models to examine differences (p < 0.05) between the composition of the galleries in the two selection regimes on day 45, day 62 and day 87 after gallery foundation. Variable Contrasts Coeff. ± SE z p Direction

Gallery founding success Cooperators 1st vs. Dispersers 1st -0.08 ± 0.26 -0.32 0.75 ns rate Cooperators 6th vs. Dispersers 6th 1.82 ± 0.42 4.3 <0.001 +++ Cooperators 1st vs. Cooperators -0.46 ± 0.33 -1.4 0.16 ns 6th Dispersers 1st vs. Dispersers 6th 1.45 ± 0.34 4.31 <0.001 +++ Model coefficients are reported as Coeff. ± SE (standard error of the estimate). The influence of the selection regime (cooperators vs. dispersers) on the offspring numbers and proportion of staying females was modelled for each day separately with quasipoisson or quasibinomial error distributions. A positive coefficient denotes that the mean of the cooperators is higher than the mean of the dispersers; a negative contrast denotes the reverse. Significant differences (p < 0.05) are displayed in bold; trends for differences (p < 0.05) are indicated by brackets. C = cooperators, D = dispersers.

Table S5. GEE model to examine differences (p < 0.05) in the success rates of daughters to produce a brood when reared independently between the selection regimes and at the start (1st generation) and at the end (6th generation) of laboratory breeding.

Offspring stage Age of gallery Coeff. ± SE t p Direction

Number of immature stages (eggs, 45 days 0.61 ± 0.39 1.56 0.13 C = D 62 days -2.69 ± 0.67 -4.03 <0.001 C > D larvae, pupae) 87 days 0.47 ± 0.87 0.54 0.6 C = D

Number of staying adult ♀♀ 45 days -0.72 ± 0.43 -1.7 0.11 C = D 62 days -1.77 ± 0.09 -19.2 <0.001 C > D 87 days 0.39 ± 0.72 0.54 0.6 C = D

Number of dispersed adult ♀♀ 45 days dispersal only in one cooperators gallery 62 days 1.1 ± 0.49 2.23 0.03 C < D 87 days -1.15 ± 0.51 -2.24 0.039 C > D

Proportion of staying adult ♀♀ 45 days dispersal only in one cooperators gallery 62 days -2.87 ± 0.57 -5.04 <0.001 C > D 87 days 1.54 ± 0.8 1.93 0.072 (C < D)

Total adult individuals (♂,♀) 45 days -0.68 ± 0.36 -1.93 0.068 (C > D) 62 days -0.51 ± 0.19 -2.67 0.01 C > D 87 days -0.52 ± 0.36 -1.42 0.18 C = D We used GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery) and binomial error distribution (successful brood production yes or no). The influences of the regimes on the success rates are displayed as contrasts between classes.

Chapter 4

Abstract Strongly female-biased sex ratios are typical for the fungal feeding haplodiploid Xyleborini (Scolytinae, Coleoptera), and are a result of inbreeding and local mate competition (LMC). These ambrosia beetles are hardly ever found outside of trees, and thus male frequency and behavior have not been addressed in any empirical studies to date. In fact, for most species the males remain undescribed. Data on sex ratios and male behavior could, however, provide important insights into the Xyleborini's mating system and the evolution of inbreeding and LMC in general. In this study, I used in vitro rearing methods to obtain the first observational data on sex ratio, male production, male and female dispersal, and mating behavior in a xyleborine ambrosia beetle. Females of Xyleborinus saxesenii Ratzeburg produced between 0 and 3 sons per brood, and the absence of males was relatively independent of the number of daughters to be fertilized and the maternal brood sex ratio. Both conformed to a strict LMC strategy with a relatively precise and constant number of males. If males were present they eclosed just before the first females dispersed, and stayed in the gallery until all female offspring had matured. They constantly wandered through the gallery system, presumably in search of unfertilized females, and attempted to mate with larvae, other males, and females of all ages. Copulations, however, only occurred with immature females. From galleries with males, nearly all females dispersed fertilized. Only a few left the natal gallery without being fertilized, and subsequently went on to produce large and solely male broods. If broods were male-less, dispersing females always failed to found new galleries.

Keywords inbreeding, haplodiploidy, ambrosia beetles, all male broods, LMC, sub-sociality

Chapter 4

Introduction In panmictic (randomly mating) species, natural selection usually favors balanced sex ratios (Fisher 1930). In settings where male and female offspring are of unequal value to the mother, however, optimal sex ratios may be unbalanced. For example, in a non-randomly mating population or species with strictly local offspring reproduction, a mother will gain highest fitness by producing as many daughters as possible and just enough sons to ensure fertilization of all sisters in the brood (Hamilton 1967). This extreme economy in the production of males is common in small arthropods with regular brother-sister mating and local mate competition (LMC) between brothers (Hamilton 1967; Norris 1993). Usually it is associated with arrhenotokous haplodiploidy (Hamilton 1978 referred to arthropods exhibiting these characters as "the biofacies of extreme sex ratios and arrhenotoky"). Arrhenotoky potentially gives a mother precise control over the sex of each offspring, as she may choose to produce diploid daughters by fertilizing an egg, or haploid sons by leaving it unfertilized (Hamilton 1967; Charnov 1982). Hamilton (1978) suggested that the ancestral habitat for the aforementioned biofacies is situated under the bark of dead trees, even if the support for this quotation is weak and claims on the original habitat of ancient lineages are highly speculative (Normark et al. 1999). A prominent group of beetles living under the bark is the weevil sub-family Scolytinae, that includes species with varying mating systems (Kirkendall 1983; 1993) and ploidy (diploid, functional haploid, haplodiploid; Kirkendall 1993; Brun et al. 1995; Normark et al. 1999). Both characters are selected to the extremes in all species belonging to the subtribe Xyleborini, which are a typical example for Hamilton's biofacies since they are characterized by strong inbreeding, LMC and haplodiploidy. These so-called ambrosia beetles solely feed on a microbial complex (fungi, yeasts and bacteria; Haanstad and Norris 1985), which they grow on the walls of self-constructed tunnel systems ("galleries") in the heart- or sapwood of trees. Until recently, this habitat has virtually been impossible to access and observe without destruction, which is why basic data on life histories, mating systems and behavior of most Xyleborini is still missing. The adaptation of an in vitro rearing technique (Norris and Baker 1967; Biedermann et al. 2009) finally may lead to the establishment of an excellent new model system for studying the independent evolution of inbreeding and haplodiploidy in a weevil lineage. The few observational studies on Xyleborini that exist suggest that these beetles behave sub-socially (Kirkendall et al. 1997; Mueller et al. 2005). Sub-sociality in Xyleborinus saxesenii Ratzeburg and other Xyleborini is indicated by the fact that mature female offspring maintain the natal gallery throughout a few days or even weeks at a time when they could already found their own nests (Kirkendall et al. 1997; Mueller et al. 2005; Biedermann 2007). Costs of independent breeding have been found to be high in this species due to risky dispersal and low founding success, which in former studies did not exceed 20% neither in the field nor when in vitro culturing was used (Biedermann et al. 2009). Therefore, it might pay for daughters to stay in productive

Chapter 4 galleries, where they either reproduce themselves, or gain indirect fitness benefits by helping to produce more sisters (e.g. Peer and Taborsky 2007). The help of males (if there is) is assumed to be of minor importance to the family, given their tiny size (figures in Fischer 1954) and their underrepresentation in relation to females. If males do not play an important role for gallery maintenance and protection, male numbers should be minimized in order to lower the cost of LMC. The optimal number should be affected by the maximum number of females one male is able to fertilize, the timing of male production, male survivorship and longevity, as well as male dispersal (see Kirkendall 1993 for a detailed review; Borsa and Kjellberg 1996). In Xyleborini males are outnumbered by females by a ratio of 1:5 to 1:30, depending on the species (e.g. Schedl 1962). Males supposedly hatch before females (observed in Xyleborus affinis Eichhoff; Roeper et al. 1980), but their life expectancy is unknown. As they lack functional flight wings, they are assumed to never leave the nest (Roeper et al. 1980; Kirkendall 1983; 1993), however, recent data suggests that males of Xylosandrus germanus Reiter sometimes leave their natal gallery to crawl on the host tree in search for outbreeding opportunities, i.e., disperse (Peer and Taborsky 2004). The objective of this study was to observe males of X. saxesenii inside their galleries for the first time to report their number and hatching time relative to females, as well as to determine whether or not males disperse and if they share in gallery maintenance and protection. Furthermore, I aimed at clarifying the existence of all male broods, which has been remarked upon in several studies on Xyleborini (e.g. Fischer 1954; Norris and Baker 1967; Kirkendall 1993). Such broods are assumed to be produced by unfertilized females that later mate with one of their sons to subsequently produce "normal" mixed broods (Herfs 1959). Their existence was tested by following the fate of unfertilized females. All these studies were conducted by consecutively rearing several generations of X. saxesenii families in an artificial medium in glass tubes that allow for behavioral observations (Biedermann et al. 2009).

Methods

Study species

In the study species Xyleborinus saxesenii Ratzeburg, the ambrosial complex is typically dominated by the anamorphic fungus Ambrosiella sulfurea Batra (Batra 1967), and serves as the sole food for adults as well as the developing larvae. The latter feed xylomycetophagously (on fungus and wood; Schedl 1958), in this way digging out a single large brood chamber where individuals of all age classes live in close vicinity to each other and also to their fungal food source. Galleries are always founded by individual females that usually have been fertilized by a brother prior to their emergence from the natal nest. The first offspring mostly stay with their mother after reaching adult-hood while the subsequent offspring generations develop (e.g. Kalshoven 1962; Schedl

Chapter 4

1966; Bischoff 2004; Biedermann et al. 2009). Brood care and fungus tending in X. saxesenii are hitherto unknown, but expected to occur, because in case the foundress dies before the first brood has eclosed, the brood dies as well and the fungal garden de-grades (Batra and Michie 1963; Norris 1979; 1993). Furthermore, gallery protection by blocking the entrance with the abdomen or a plug of frass and brood care behaviors have been observed in adult daughters of this as well as other xyleborine species (Kirkendall et al. 1997; Bischoff 2004; Biedermann 2007; personal observations).

Laboratory breeding and data collection

X. saxesenii beetles were bred in artificial agar-sawdust based "standard medium" in glass tubes (Biedermann et al. 2009). Single females were surface sterilized by sub-merging them first in ethanol (95%) and then in distilled water for a few seconds and subsequently put directly onto the medium. They usually started to excavate a tunnel system within two days and would not lay eggs until their mutualistic ambrosia fungus had started to grow. About 20% of the females started to lay eggs, which resembles the breeding success found in the field (Biedermann et al. 2009). The developing brood would subsequently enlarge the gallery system which was often constructed next to the tube wall, which facilitated the observations necessary for this study. Parts of the gallery inside the medium could not be accessed with this method, but as individuals move a lot, this should not have influenced my results.

Observations

I scanned the beetles in 70 observable galleries every second to third day and recorded male and female numbers, allocating the individuals to different developmental stages (eggs, larvae, male and female pupae, immature females, mature females, and males). Male and female pupae and adults were easy to differentiate because of their strong sexual dimorphism in size (Fig. 1, see figures in Fischer 1954). Immature females were identified by their light brown coloration that would turn black after maturation. As a result of sex specific mortality rates and siblicide, the so gained "secondary sex ratio" of the immature and mature offspring may differ from the mother's optimal "primary sex ratio" of the eggs. Where I could witness male-female interactions, I recorded the age of the partners and whether it was a mating attempt or successful mating. Whenever possible, I determined the dates when the first egg was present (N = 70 galleries), when the first male and female offspring hatched (N = 29 galleries), when the first and last female maturated (N = 26), and when the first male dispersed (N = 13). Dispersal was defined as emergence from the gallery, i.e., when individuals were found on the surface of the medium under the cap of the tube (Biedermann 2007). I stopped monitoring and dissected the galleries either when eclosion of new beetles ceased within about 3

Chapter 4

months (N = 41 galleries), or when all adult females had dispersed (N = 29 galleries). In the latter case, galleries were used to monitor male dispersal. I measured the breeding success (defined as the successful start of egg-laying within 40 days) and brood size of 311 daughters out of 33 families (median = 5 daughters/ family, range = 1—41) to determine whether they differ between daughters of galleries with males and male-less ones. Subsequently, I correlated the secondary sex ratio of 62 daughter- families that had successfully produced a brood with the secondary sex ratio of their mothers' families to test whether secondary sex ratios might be heritable. Additionally, I observed the behavior of eight males from different galleries for 10 min and calculated the proportion of time spent on different behaviors. I differentiated between shuffling (moving frass and feces with the legs under the body and with the abdomen), digging (excavating new tunnels), cannibalism (feeding on a larva, pupa, or adult beetle), courtship behavior (grooming an egg, larva, pupa, or adult beetle with maxillae and labium), walking, cropping (feeding on the fungal layer covering gallery walls), and mating attempt (mounting or copulating with a female) (Biedermann 2007).

Statistical analyses

The association between numbers of male and female offspring was explored using linear regression. The number of males within a brood of a given size should have

Chapter 4 a binomial distribution under random sex determination. To test if this was the case, I compared the variance of the male numbers observed with that expected if the numbers of males were binomially distributed using a Chi2-test (Green et al. 1982). The same test was also used to analyze whether the number of males per family resembled a Poisson distribution. Male mating preference with either immature or adult females was analyzed using Fisher's exact test. Usually the data were not normally distributed (which was determined with Kolmogorov-Smirnov tests), consequently non-parametric statistics were conducted. The Mann-Whitney U-test for comparisons between two independent groups was applied to determine productivity differences between male and male-less broods, and to check for differences in the timing of male and female hatching and dispersal. The inheritance of sex ratios between mother and daughter galleries was analyzed with a Spearman's rank correlation. Analyses were performed with SPSS (Release 14.0; SPSS, Chicago, IL) and R (R Development Core Team 2008).

Results

Sex ratio and breeding success

Males were extremely scarce in X. saxesenii galleries (mean abundance per gallery = 0.96, SE = 0.09, N = 70; Fig. 2). Male production by mothers was not random, as their overall number did not resemble a Poisson distribution. They produced more one- and two-male broods, fewer male-less and three-male broods than expected, and no four-male brood, which suggests the optimal number is one or two males per brood. Nearly one third of the mothers produced no males (N = 21). Families with males were significantly larger than families without males (Fig. 3). In accordance with this finding, males were absent from 50% of the galleries with only ten daughters or fewer, whereas only 8.3% of the galleries with more than ten daughters were male-less. However, overall the number of males per gallery was only weakly affected by the number of adult females (Linear regression: F = 3.824, p = 0.055, DF= 1, N = 70; Fig. 4). The residual mean squared error of the number of males (MSE = 0.597) was much smaller than the mean variance expected under binomial distribution (0.9398; Chi2-test: χ2= 43.2, DF = 68, p > 0.05), which indicates that the production of males was not random, but relatively constant and precise (around one or two males per brood). Breeding success (i.e., the likelihood to lay at least one egg within 40 days) of daughters per family ranged between 0% and 78% (mean = 15.2%, median = 0, N = 311 daughters from 33 families), and was 0% for all male-less broods. A statistical difference was however not detectable because of the high variation of breeding success across all families tested (U- test: T = 77,p = 0.067, N = 26 vs. 7). The secondary sex ratios of the 62 successfully founded daughter-galleries did not correlate with the secondary sex ratios of their mothers galleries (Spearman's rank correlation: r = 0.038, p = 0.771, N = 62).

Chapter 4

Male presence and behavior

The first male per gallery usually eclosed a few days after his first sister did so (Fig. 5), but before her maturation (N = 29 galleries). Following eclosion, males spent most of their time with cropping fungus and walking through the gallery system. When encountering a possible partner, the male started to courtship her, which resembles grooming the body of the female, and afterwards attempted to mate with her. Females seemed relatively reluctant, and thus mating attempts usually lasted several minutes and were rarely successful. Thirty mating attempts were observed, during which males mounted immature (N = 7 of 30 times) and mature females (N = 11 of 30 times) as well as pieces of wood, larvae, and other males (N = 12 of 30 times). Successful copulations, however, were only observed with three immature females, which hints towards a preference to mate with them (Fisher's exact test: p = 0.09, N = 21; Fig. 6). Only two females were found to emerge unfertilized from their galleries (indicated by the production of an exclusively male brood later on); they emerged from galleries with brood sex ratios of 2:16 and 1:86, respectively, which indicates that one male is capable of fertilizing up to 60 sisters on its own (no unfertilized females were detected in a gallery with a sex ratio of 1:60). Other male behaviors which were rarely exhibited included digging, cannibalism, and the shuffling of saw-dust and feces with the legs (all with median = 0%). Although males are wingless, some of them (at least one male in 13 out of 29 galleries) dispersed, i.e., were found on the surface of the medium, when all offspring had matured and no new eggs were laid. Males did not disperse randomly together with

Chapter 4

their sisters throughout gallery life, but in most cases the first male dispersal occurred after the last female (in a gallery) had matured (Fig. 5).

Discussion

Under inbreeding conditions, natural selection should favor the production of that number of males that maximizes the mean number of inseminated females dispersing from a brood. Thus, in case males are able to inseminate only a limited number of females, the numbers of males and females should be correlated (e.g. Green et al. 1982; Borsa and Kjellberg 1996). In contrast to these predictions, male numbers in X. saxesenii families were relatively independent of the number of sisters to be fertilized, ranging from 0 to 3 males per 1 to 86 females, and also independent of maternal brood sex ratios. These facts conform to a strict LMC strategy (e.g. Hamilton 1967; Bernal et al. 2001) with a relatively precise and constant number of males per brood. Males were apparently extremely successful to locate all their unfertilized sister. Only ~3% (2 of 70 galleries) of the successfully breeding females turned out to be unfertilized (they produced solely male broods), which is approximately the same ratio found for Xyleborus compactus Eichhoff in the field (Brader 1964). Addition-ally, in a family with a sex ratio of 1:60 no unfertilized females were found during consecutive lab rearing. The success of males is probably explainable by the gallery morphology. Xyleborini in the genus Xyleborinus usually construct brood chambers ("caves"; Wood 1982; Roeper 1995) instead of branching tunnel systems as they are excavated by species of the genus Xyleborus (J. Hulcr personal communication), where males and females probably meet more regularly and thus fewer males may be sufficient to fertilize all sisters. In line with this idea, male and female numbers are correlated in the branching tunnel systems of Xyleborus affinis Eichhoff (Xyleborina, Scolytinae), as founder females lay eggs in clusters that always contain a male egg (Roeper et al. 1980). At least one male per family should, however, be expected under the assumption of strict inbreeding. Contrary to that, 30% of all families were male- less; a proportion which is not untypical for Xyleborini (19% of all families in Xylosandrus germanus Reiter in the field; Peer and Taborsky 2004). As these measures were taken on adult offspring, the easiest explanation for that finding is that the secondary sex ratios reported, differ from the primary sex ratios allocated by the mother to the eggs, i.e., males had somehow been eliminated. Such secondary male killing could have been accomplished either by selectively cannibalizing offspring or by selecting mechanisms of the fungal cultivars2 (for attine ants; Mueller 2002) or of bacteria in the ovaries (Peleg and Norris 1972; Kawasaki et al. 2009). Given that both microbes are transmitted only from mothers to daughters, they would presumably profit from female biased broods (Dyson and Hurst 2004; Normark 2004). Studies on the primary sex ratio

2 The fungal manipulation hypothesis did not hold up in the only explicit tests so far (see Dijkstra & Boomsma (2008) in Oikos 117: 1892-1906. [This note was added after the publication of this article.] 62

Chapter 4 allocated by the mother to the eggs are therefore necessary to clarify the role of secondary sex ratio distorters as a cause for the high frequency of male-

less families in Xyleborini. If, however, it is assumed that there is no secondary male killing and that the absence of males is intended by the mother beetle from the very beginning, then either the chances for mating with foreign partners (and thus, outbreeding) must be higher than previously expected, or selection against male-less broods must be weak. Meeting and outbreeding with foreign mates is obviously possible, as males dispersed from nearly half of the galleries when brood production ceased and females sometimes produced all male broods, whose males also all emerged. It fits these data that X. saxesenii males can sometimes be seen outside their nests, occasionally walking on their natal host trees (personal field data). The idea that males disperse from their natal nest in search for outbreeding opportunities implies, however, that they are able to enter foreign galleries, what seems so be complicated by the fact that females or more rarely males (in the genera Coptodryas and Cyclorhipidion; J. Hulcr personal communication) block the gallery entrances (e.g. Kirkendall et al. 1997; personal observations). On the other hand, male- less families were always very small (equal to or less than 10 females; see similar results in Hypothenemus hampei Ferrari by Borsa and Kjellberg 1996), and in a previous study daughters of small families (less than 20 females) were found to contribute much less offspring to the next generation than the daughters from larger families, presumably because the latter are in better body condition and have a more productive complex of microbes to nourish them (Biedermann et al. 2009). Therefore, it is possible that these small male- less galleries would anyway not contribute much fitness to the next generation and hence selection against the absence of males within small broods is probably weak. Relatively independent from the fact if secondary manipulations of the sex ratio occur, male numbers should be higher under conditions that allow outbreeding, as has been shown in X. germanus (Peer and Taborsky 2004). In case X. saxesenii mothers also allocate offspring sex according to outbreeding opportunities, then less males should have been found in my lab study compared to the field, where outbreeding opportunities can be expected to arise regularly. Contrary to these expectations, I found a sex ratio of 1:8 (m:f) on average, which is well below the average 1:20 sex ratio previously found in the field both in Australia (Hosking 1972) and Switzerland (personal observations). Although the methods for measuring sex ratios differed between lab and field studies, these strongly contradicting results are not explainable by methodological differences alone and may point to other (environmental) factors influencing sex allocation in X. saxesenii. Males hatched significantly later than their first sisters did, which contrasts with findings in X. affinis, whose males hatch before females (Roeper et al. 1980). I hypothesize that in X. saxesenii fertilization of all females by their brothers was ensured because females need about ten days to fully develop (Biedermann et al. 2009), and males hatched on average 4 days after females. Furthermore, fertilization of all females was made possible by the high sexual activity of males that did not engage in social behaviors. Instead, they constantly wandered the gallery in search for virgin females, which they started to courtship (groom) upon encounter, followed

Chapter 4 by mounting and copulation. Although rarely seen, these copulations occurred exclusively with immature females. Concordantly, males did not disperse as long as new immatures eclosed and could be fertilized, but tended to leave the natal nest as soon as eclosion ceased, presumably in search for outbreeding opportunities. This indicates that they try to fertilize as many sisters as possible, but search for outbreeding opportunities as soon as direct fitness maximization within the natal nest is no longer possible. Unfertilized females produced solely male broods, but did not mate with their sons to subsequently produce "normal" mixed broods. This expectation was based on observations of Coccotrypes dactyliperda Fabricius (Scolytinae) females that do so (Herfs 1959). The behavioral difference between these two species is, however, plausible when one takes into account the extreme longevity and fertility of C. dactyliperda females, which raise up to five broods, in this way producing up to 144 individuals in total. In contrast, X. saxesenii females usually die soon after their first offspring matures (personal observations), which in case they mated with one of their sons would mean shortly after doing so. Even in case she was nevertheless able to lay a few fertilized eggs before her death, the brood would still be doomed as the mother's presence and her brood care appears crucial for successful brood development (e.g. Norris 1993; Kirkendall et al. 1997; Biedermann 2007). Another fascinating explanation could be that male broods are intended by the females and represent an alternative reproductive tactic under certain conditions when fitness gains through outbreeding sons are higher than those that can be achieved with the production and maintenance of a mixed brood. Such a condition may, for example, be an environment with a lot of small and male-less broods (e.g., under high density), a situation in which selection would favor a few females to just produce sons that visit their male-less neighboring families. My data shed light on the cryptic life and mating system of xyleborine ambrosia beetles within their galleries. In the future, molecular studies will be crucial for determining whether the observed inter- and intraspecific variance in sex ratios is an adaption to outbreeding in some Xyleborini or a hint for the existence of sex ratio distorters, for example, their symbionts. My data provide a starting point for future studies dealing with factors that influence xyleborine sex ratios and the frequency of outbreeding events. Ambrosia beetles are one of the best model systems for studying the evolution of inbreeding, haplodiploidy, sociality and symbioses, but at the same time one of the least known. Before comparative studies on different species can be done to address the issues mentioned, further basic data on the behaviors of these beetles inside their galleries have to be collected.

Conclusion

My data suggest that males in Xyleborinus saxesenii Ratzeburg are extremely successful in locating and fertilizing all their sisters in the natal gallery. On average there is only

Chapter 4 about one male per family relatively independent of the number of sisters and the maternal brood sex ratio. Apparently, this is sufficient to fertilize all females, probably because the morphology of the gallery system with a central "brood chamber" makes it easy for males to locate them. Only about 3% of the females from broods with males disperse unfertilized from the natal nest and subsequently produce all male broods, but do not mate with one of their sons to produce a mixed brood afterwards. Despite the fact that males are not the first to hatch within a brood, they do so before the first females mature, and only disperse from the natal gallery once the last female has finished her development. They do not engage in gallery maintenance, gallery protection and brood care, but constantly wander the gallery, presumably in search of unfertilized females. Although males attempt to mate with all individuals independent of age and sex, they were observed to copulate only with immature females.

Acknowledgements I want to thank J. Hulcr and A. Cognato for giving me the chance to be part of this festschrift. I hold S. L. Wood's accomplishments in great esteem, and I regret that I have never had the chance to meet him in person. This manuscript emerged at the Institute of Ecology and Evolution, University of Bern, Switzerland, under the supervision of M. Taborsky. Additionally to discussions with him, it profited a lot from suggestions made by T. Turrini, R. Roeper, K. Peer, E. Ott, and two anonymous reviewers. During parts of the project, I was financially supported by a grant of the Roche Foundation and a DOC grant of the Austrian Academy of Science.

References Batra LR (1967) Ambrosia fungi - A taxonomic revision and nutritional studies of some species. Mycologia 59: 976-1017. Batra LR, Michie MD (1963) Pleomorphism in Some Ambrosia and Related Fungi. Transactions of the Kansas Academy of Science 66: 470-481. Bernal JS, Gillogly PO, Griset J (2001) Family planning in a stemborer parasitoid: sex ratio, brood size and size-fitness relationships in Parallorhogaspyralophagus (Hymenoptera : Braconidae), and implications for biological control. Bulletin of Entomological Research 91: 255-264. Biedermann PHW (2007) Social behaviour in sib mating fungus farmers. Master thesis, Berne, Switzerland: University of Berne. Biedermann PHW, Klepzig KD, Taborsky M (2009) Fungus cultivation by ambrosia beetles: Behavior and laboratory breeding success in three Xyleborine species. Environmental Entomology 38 (4): 1096-1105. Bischoff LL (2004) The social structure of the haplodiploid bark beetle, Xylosandrus germanus. Diploma thesis, Berne, Switzerland: University of Berne.

Chapter 4

Borsa P, Kjellberg F (1996) Secondary sex ratio adjustment in a pseudo-arrhenotokous insect, Hypothenemus hampei (Coleoptera: Scolytidae). Comptes Rendus del Academie des Sciences Serie Iii-Sciences de la Vie-Life Sciences 319: 1159- 1166. Brader L (1964) Etude de la relation entre le scolyte des rameaux du cafeir, Xyleborus compactus Eichh. (X. morstatti Hag.), et sa plantehote. Mededelingen van de andbouwhogeschool te Wageningen 64: 1-109. Brun LO, Borsa P, Gaudichon V, Stuart JJ, Aronstein K, Coustau C, Ffrench-Constant RH (1995) Functional haplodiploidy. Nature 374: 506. Charnov EL (1982) The Theory of Sex Allocation. Princeton University Press, Princeton, 355 pp. Dyson EA, Hurst GDD (2004) Persistence of an extreme sex-ratio bias in a natural population. Proceedings of the National Academy of Sciences of the United States of America 101: 6520-6523. Fischer M (1954) Untersuchungen über den kleinen Holzbohrer (Xyleborus saxeseni). Pflanzenschutzberichte 12: 137-180. Fisher RA (1930) The Genetical Theory ofNatural Selection. Oxford University Press, Oxford, 370 pp. Green RF, Gordh G, Hawkins BA (1982) Precise Sex-Ratios in Highly Inbred Parasitic Wasps. American Naturalist 120: 653-665. Haanstad JO, Norris DM (1985) Microbial symbiotes of the ambrosia beetle Xyletorinus politus. Microbial Ecology 11: 267-276. Hamilton WD (1967) Extraordinary sex ratios. Science 156: 477-488. Hamilton WD (1978) Evolution and diversity under bark, In: Mound LA, Waloff N (Eds) Diversity of insect faunas. Blackwell, Oxford, 154-175. Herfs A (1959) Über den Steinnussborkenkäfer Coccotrypes dactyliperda F. Anzeiger für Schädlingskunde - Journal of Pest Science 32: 1-4. Hosking GB (1972) Xyleborus saxeseni, its life-history and flight behaviour in New Zealand. New Zealand Journal for Forest Science 3: 37-53. Kalshoven LGE (1962) Note on the habits of Xyleborus destruens Bldf., the near-primary borer of teak trees on Java. Entomologische Berichten 22: 7-18. Kawasaki YIM, Miura K, Kajimura H (2009) Superinfection of five Wolbachia in the alnus ambrosia beetle, Xylosandrus germanus (Blandford) (Coleoptera: Curuculionidae). Bulletin of Entomological Research 100: 231-239. Kirkendall LR (1983) The evolution of mating systems in bark and ambrosia beetles (Coleoptera: Scolytidae and Platypodidae). Zoological Journal of the Linnean Society 77: 293-352. Kirkendall LR (1993) Ecology and evolution of biased sex ratios in bark and ambrosia beetles. In: Wrensch DL, Ebbert MA (Eds) Evolution and diversity of sex ratio in insects and mites. Chapman and Hall, New York, 235-345. Kirkendall LR, Kent DS, Raffa KF (1997) Interactions among males, females and offspring in bark and ambrosia beetles: the significance of living in tunnels for the evolution of social behavior. In: Choe JC, Crespi BJ (Eds) The Evolution of Social Behavior in Insects and Arachnids. Cambridge University Press, Cambridge, 181-215. Mueller UG (2002) Ant versus fungus versus mutualism: Ant-cultivar conflict and the deconstruction of the attine ant-fungus symbiosis. American Naturalist 160: S67-S98.

Chapter 4

Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563—595. Normark BB (2004) Haplodiploidy as an outcome of coevolution between male-killing cyto- plasmatic elements and their hosts. Evolution 58: 790—798. Normark BB, Jordal BH, Farrell BD (1999) Origin of a haplodiploid beetle lineage. Proceedings of the Royal Society London B 266: 2253—2259. Norris DM (1979) The mutualistic fungi of Xyleborini beetles. In: Batra LR (Ed) Nutrition, Mutualism, and Commensalism. Allanheld, Osmun and Company, Montclair, 55—63. Norris DM (1993) Xyleborus ambrosia beetles - a symbiotic ideal extreme biofacies with evolved polyphagous privileges at monophagous prices. Symbiosis 14: 229—236. Norris DM, Baker JK (1967) Symbiosis: effects of a mutualistic fungus upon the growth and reproduction of Xyleborus ferrugineus. Science 156: 1120—1122. Peer K, Taborsky M (2004) Female ambrosia beetles adjust their offspring sex ratio according to outbreeding opportunities for their sons. Journal of Evolutionary Biology 17: 257—264. Peer K, Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. Behavioral Ecology and Sociobiology 61: 729—739. Peleg B, Norris DM (1972) Bacterial symbiote activation of insect parthenogenetic reproduction. Nature New Biology 236: 111—112. R Development Core Team (2008) R: A language and environment for statistical computing. Vienna, Austria, R Foundation for Statistical Computing. Roeper RA, Treeful LM, O'Brien KM, Foote RA, Bunce MA (1980) Life history of the ambrosia beetle Xyleborus affinis (Coleoptera: Scolytidae) from in vitro culture. Great Lakes Entomologist 13: 141—144. Roeper RA (1995) Patterns of mycetophagy in Michigan ambrosia beetles. Michigan Academian 27: 153—161. Schedl KE (1958) Breeding habits of arboricole insects in Central Africa. In: Becker EC (Ed) Proceedings of the 10th International Congress of Entomology, Montreal (Canada), August 1956, 183—197. Schedl KE (1962) Scolytidae und Platypodidae Afrikas, II. Revista de Entomologia de Mocamique 5: 1—594. Schedl W (1966) Zur Verbreitung und Autokologie von Xyleborus eurygraphus Ratzeburg. Bericht des Naturwissenschaftlichen - Medizinischen Vereins Innsbruck 54: 61—74. Wood SL (1982) The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Naturalist Memoirs 6: 1—1359.

Chapter 5

Behav Ecol Sociobiol (2011) 65:1753-1761 DOI 10.1007/s00265-011-1183-5

Received: 30 August 2010/Revised: 14 March 2011 /Accepted: 31 March 2011 /Published online: 15 April 2011 © Springer-Verlag 2011

Abstract Body reserves may determine the reproductive output However, induced dispersers produced more offspring than of animals, depending on their resource allocation strategy. In delayed dispersers within a test period of 40 days. This suggests insects, an accumulation of reserves for reproduction is often that delayed dispersal comes at a cost to females, which may obtained before dispersal by pre-emergence (or maturation) result primarily from alloparental care and leads to a reduced feeding. This has been assumed to be an important cause of reproductive output. Alternatively, females might have delayed dispersal from the natal nest in scolytine beetles. In the reproduced prior to dispersal. This is unlikely, however, for the cooperatively breeding ambrosia beetles, this is of special majority of dispersing females because of the small numbers of interest because in this group delayed dispersal could serve two offspring present in the gallery when females dispersed, alternative purposes: "selfish" maturation feeding or "altruistic" suggesting that mainly the foundress had reproduced. In alloparental care. To distinguish between these two possibilities, addition, "gallery of origin" was a strong predictor of the we have experimentally studied the effect of delayed dispersal reproductive success of females, which may reflect variation in on future reproductive output in the xyleborine ambrosia beetle the microbial complex transmitted vertically from the natal nest Xyleborus affinis. Females experimentally induced to disperse to the daughter colony, or variation of genetic quality. These and delayed dispersing females did not differ in their body results have important implications for the understanding of condition at dispersal and in their founding success afterwards, proximate mechanisms selecting for philopatry and alloparental which indicates that females disperse independently of care in highly social ambrosia beetles and other cooperatively condition, and staying adult females are fully mature and would breeding arthropods. be able to breed. Keywords Resource allocation . Capital breeding . Bark beetles. Sociality. Fungus gardening. Cooperative breeding Communicated by N. Wedell P. H. W. Biedermann (*) • M. Taborsky Behavioural Ecology, Institute of Ecology and Evolution, Introduction University of Berne, Baltzerstrasse 6, 3012 Bern, Switzerland Living in groups involves three key decisions of totipotent *e-mail: [email protected] individuals (Helms Cahan et al. 2002): whether or not to disperse (Stacey and Ligon 1991; Kokko and Ekman 2002), whether or not P. H. W. Biedermann : K. D. Klepzig USDA Forest Service, Southern Research Station, to breed (Keller and Reeve 1994; Hager and Jones 2009) and 2500 Shreveport Hwy, whether or not to help other group members to raise their Pineville, LA 71360, USA offspring (Eden 1987;Stacey and Koenig 1990). Dispersal, as the K. D. Klepzig first and most basal strategic decision, should depend on relative US Forest Service, Southern Research Station, fitness effects of staying and leaving, which are a function of 200 WT Weaver Blvd, Asheville, NC 28804, USA ecological conditions like resource availability, population density and predation risk (Koenig et al. 1992; Heg et al. 2004;

Chapter 5

Bruintjes et al. 2010). Prior to dispersal, however, it is often (Kirkendall et al. 1997; Mueller et al. 2005). Evidence for difficult to estimate potential costs imposed by external factors. such cooperative brood care exists in X. saxesenii, where the Therefore, only internal cues may be available such as body number of larvae and pupae is proportional to the number condition (i.e. reserves), or local resource availability. of adult female helpers in a gallery (Peer and Taborsky Studies of the fitness consequences of sociality typically 2007), and behavioural observations revealed cooperative involve highly cooperative vertebrate and insect societies, in care (Biedermann 2007; Biedermann et al. 2009). If females which reproduction is restricted to only a few, specialized dispose of the potential to help raising broods but suffer individuals per group. An ideal model system to unravel the from limited fertility, they may benefit by caring for the importance of intrinsic factors and external causes selecting for brood of relatives instead of taking the risk to disperse and advanced sociality should, however, preferably comprise breed independently (West-Eberhard 1975; Craig 1983; totipotent individuals that flexibly engage in dispersal, Roisin 1994). 3. Direct fitness benefits through reproduction reproduction and helping depending on conditions (e.g. Costa in the natal gallery 2006). These requirements are met by the polyphyletic ambrosia Staying adult daughters might also reproduce in the natal beetles, which cultivate ambrosia fungi as food inside galleries gallery. This was observed in X. saxesenii, where one quarter excavated in the wood of freshly dead trees (e.g. Schedl 1956; of the females were found to lay eggs in their natal nest Beaver 1989; Farrell et al. 2001). Dispersal from the natal group (Biedermann 2007). in order to find an own gallery is associated with high fitness costs, as suitable wood is patchily distributed and the success In our study species, Xyleborus affinis Eichhoff (Xyleborini, rate of establishing a new fungus garden is low (about 20% in Scolytinae), daughters flexibly disperse over a period of about 50 Xyleborinus saxesenii Ratzeburg; Biedermann et al. 2009). days. Prior to dispersal, all females cooperatively care for the Therefore, if fungus productivity in the natal gallery is good and brood and fungal cultures in their natal, commonly defended thus optimal feeding conditions prevail, philopatry should be gallery for at least 1 week (Roeper et al. 1980). Here we aim to favoured. In fact, such age-dependent, delayed dispersal of unravel whether prolonged philopatry of females causes costs or scolytine beetles has been witnessed since a long time (e.g. benefits regarding body condition and future reproductive Eichhoff 1881; Whitney 1971; Botterweg 1982; Krausseopatz et success, i.e. to distinguish between hypothesis (1) and the two al. 1995; McNee et al. 2000). Three non-exclusive hypotheses for alternatives (2 and 3) above. To this end, we experimentally this trait have been proposed: induced dispersal of adult females and measured their body weight, size and reproductive success in comparison to a control 1. Direct fitness benefits through maturation feeding group that was allowed to disperse deliberately without Delayed dispersal has been associated with pre- experimental interference. If maturation feeding occurred, emergence or maturation feeding (Eichhoff 1881; induced dispersers should be less successful in gallery foundation Botterweg 1982; McNee et al. 2000). Evidence for and offspring production than voluntarily delayed dispersers maturation feeding exists from phloem-feeding mountain (hypothesis 1) because the former had stayed shorter in the pine beetles, where females were experimentally pre- natal gallery than the latter before experimental collection. If in vented from feeding after eclosion. They matured normally contrast induced dispersers are more successful, this would but were less likely to breed successfully and laid smaller reveal that delayed dispersal is associated with a reduction in eggs (Elkin and Reid 2005). Scolytine ambrosia beetles, body condition, indicating fitness costs by cooperative brood however, have a different feeding habit, using ambrosia care (hypothesis 2) or reproduction prior to dispersal (hypothesis fungi as their sole nutritional source. Female ambrosia 3). The latter possibility we checked by counting offspring beetles were found to lay eggs only after growing their own numbers at the time of collection and by determining the ovarian fungus garden on which they fed (French and Roeper 1975; status of all collected females. Kingsolver and Norris 1977; Roeper et al. 1980; Beaver 1986). Hence, it is presently unclear whether reserves accumulated before emergence will raise the productivity of Material and methods those beetles sufficiently to outweigh the fitness costs of delayed dispersal. Study species 2. Indirect fitness benefits through cooperative brood care Females delaying dispersal could help to raise siblings, X. affinis is a tropical and subtropical member of the scolytine especially since the fungus cultivation of ambrosia beetles subtribe Xyleborini, which are characterized by may benefit from the cooperation of several individuals

Chapter 5 haplodiploidy, strongly female-biased sex ratios (X affinis: 1:8.5 they inoculate the tunnel walls with ambrosia fungus spores males/females - Roeper et al. 1980; 1:15.2 males/ females - from their mycetangia, which results in a fungus layer lining the Biedermann, unpublished data) and matings between full tunnel walls within a week after insertion (Roeper et al. 1980). siblings in their natal gallery (Xylosandrus germanus: Peer and At that time, females start to produce eggs, while feeding on Taborsky 2004, 2005). Only female beetles disperse from their fungus providing the essential nutrients (Kingsolver and Norris host trees by flight. They transmit spores of species-specific 1977). The progeny passes through three larval instars and a ambrosia fungi to the new gallery in a spore carrying organ pupal stage. After eclosion, it takes a few days until the beetles ("mycetangium"; Francke-Grosmann 1956), or in some cases in fully sclerotize. The first adult offspring appear in the tunnels the hindgut (Francke-Grosmann 1975). X. affinis galleries deeply around 38 days after gallery foundation. The first daughters start penetrate the wood of deciduous trees with single tunnels to disperse around day 50 after gallery foundation, but most of extending over 6 m that may be inhabited by several generations them delay dispersal from their natal nest for much longer. for up to 4 years (Schneider 1987). Under laboratory conditions, Daughters generally delay their dispersal: (1) The first developing two to three generations may develop within a gallery, which female offspring stay on average for 12 days after full will host up to 100 individuals of all developmental stages sclerotization before they start to disperse (Fig. 1; see also concurrently (Biedermann et al. 2009; Biedermann, personal Roeper et al. 1980); (2) also after dispersal has started, females observation). accumulate in the gallery as the rate of females eclosing from the pupal stage is higher than the dispersal rate. About 80-90 Preparation of artificial medium days from gallery foundation, the medium deteriorates, which is when production of new progeny has ceased and all individuals We filled sterile glass tubes (18 mm diameter x 150 mm length; leave the gallery (Fig. 1). Bellco Glass, Vineland, NJ, USA) with standard medium which consists of 0.35 g streptomycin, 1 g Wesson's salt mixture, 5 g Experimental manipulations yeast, 5 g casein, 5 g starch, 10 g sucrose, 20 g agar and 75 g oak tree sawdust (Biedermann et al. 2009). All ingredients were We observed seven galleries of the second laboratory generation mixed and supplemented by 2.5 ml of wheat germ oil, 5 ml 95% during days 57-63 after gallery foundation and immediately ethanol and 500 ml of deionised water. The mixture was collected the first females emerging on the surface of the autoclaved for 20 min at 124°C and covered immediately with medium to disperse (henceforth called delayed dispersers; Fig. sterile plastic caps (Bellco Glass kap-uts, Vineland, NJ, USA). After 1). Onthesameday,we dissected each of these the medium had cooled down, we scratched its surface with a galleries and collected the same number of mature females (fully sterile scalpel to facilitate the onset of tunnel excavation by a sclerotized) from the founder female. Then we closed the tubes again with the plastic caps and left them to set for 4 to 5 days.

Laboratory rearing of the beetles

All females used in this study were descendants of the first laboratory generation of females collected from oak logs in Pineville, LA, USA (123 ft asl; 31°20', 92°24')in summer 2007. Dispersing females emerged from their natal gallery on the surface of the artificial medium and were collected for rearing of consecutive laboratory generations (Biedermann et al. 2009). Before starting a new gallery, females were surface-sterilized by washing them first for a few seconds with 95% ethanol and then with deionised water. Then each female was placed singly on the prepared medium in a separate glass tube. These tubes were Fig. 1 Typical phenology of a laboratory gallery in X. affinis and the timing of our manipulations. E1 start of egg laying of the founder female, reclosed with the caps and stored at room temperature (~23°C) F first full sclerotization of a daughter, E2 start of daughter egg laying, in darkness (wrapped in paper, but light could shine on the while the foundress usually continues to lay eggs, D first daughters start entrance). to disperse from gallery, Ee end of egg laying because of deterioration of Following insertion on the medium, females immediately the medium. We collected dispersing and staying females between days 57 and 63, either to take measurements (dry weight, body length, ovary started to bore a tunnel. As they penetrate the medium, development) or to let them breed independently for 40 days until we dissected the daughter galleries to count their offspring

Chapter 5 same gallery as had emerged on the surface (henceforth called females) on (1) the ability to found a gallery successfully by using induced dispersers). Thus, we sampled two to 24 females per binomial error distributions and (2) the offspring numbers gallery, resulting in a total of 38 induced and 37 delayed (numbers of eggs, larvae, pupae and adults) by using normal dispersing females (one delayed disperser died during handling). error distributions. In a GEE model, the correlation structure of All delayed and induced dispersers were then inserted singly into the non-independent measurements (i.e. the influence of gallery new tubes to find their own galleries. Forty days later, we identity) is modelled separately and is not of primary interest. opened all galleries, checked whether a brood had been However, as breeding success and productivity appeared to be produced and counted offspring numbers. In the analyses of highly variable between progeny of different galleries, we treatment effects on offspring numbers, we excluded females additionally analysed the effects of gallery of origin. We used that did not produce any offspring (N=11) and one family where gallery of origin as the sole explanatory variable in a logistic only one induced dispersing female reproduced at all (so no regression (LR) to determine its influence on breeding success comparison with a sibling delayed dispersing female was and in a general linear model (GLM) to determine its influence possible). on productivity. Another GEE model was performed to analyse dry weight differences between delayed and induced dispersing Measurements of delayed and induced dispersing females females. Statistical analyses were performed with SPSS (Version 15.0, ©SPSSInc., Chicago, IL,USA,1989-2005) and R (R At the time we started our experimental manipulations, we also Development Core Team 2008). Model coefficients are reported dissected eight galleries and stored all dispersing and staying as B±standard error of the estimate (SE) throughout, with the females (equivalent to the delayed and induced dispersing induced dispersers as the reference category (coefficient set to females) in 95% ethanol. At a later date, we dry-weighed all zero). Significance level α=0.05. females from four of these galleries with a high-precision scale (precision 0.01 mg; Sartorius ME215S-OCE, Göttingen, Germany) after a 24-h drying process (oven, 80°C). We measured their Results body length to the nearest 0.01 mm using a microscope (x6.4- x40 magnification) with an ocular micrometer. Using the same Experimental manipulations microscope, we also dissected the ovaries of all females from the remaining four galleries from the dorsal side with high-precision Of the 75 experimental foundresses, 64 produced a brood tweezers. We classified ovaries as either immature (ovaries successfully. Treatment did not affect whether a brood was rudimentarily developed), mature (fully developed ovaries but successfully produced or not (seven of 37 delayed dispersing no oocytes) or egg carrying (four ovarioles containing one or females vs. four of 38 induced dispersing females failed to more oocytes; see figures in Fischer 1954). produce a brood; GEE: B±SE 0.653± 2 0.546, χ = 1.429, df =1, P=0.232), and there was no effect of Statistical analyses gallery identity on founding success (LR: χ2=0.001, df=1, P=0.97, N=75). Our nested design with variable numbers of repeated measures Of the 64 successful foundresses, induced dispersers from the experimental galleries generated matched, non- produced more offspring than delayed dispersers (GEE: B± SE independent measurements. Linear mixed models may be used 9.137±4.426, χ2=4.263, df=1, P=0.039), which rejects the when the distribution of the repeated responses for a subject maturation feeding hypothesis. This relationship was observed in has a multivariate normal distribu-tion. This is unlikely when the four out of six galleries (Fig. 2a). If the total offspring numbers dependent variable is binary or count data (Norusis 2007). were split up between different developmental stages, it became Therefore, generalized estimating equations (GEE), which are an clear that this result was solely caused by the variance in the extension of generalized linear models, were used to analyse numbers of laid eggs (GEE: B± SE 5.175±1.6, χ2=10.466, df=1, effects of dependent variables on correlated binary or count P=0.001), but not by the numbers of larvae (GEE: B±SE response variables (Liang and Zeger 1986; Zeger and Liang 1986). 4.263±3.797, χ2=1.26, df=1, P=0.262), pupae (GEE: B±SE We used GEEs with an exchangeable correlation structure of the 2.756±2.589, χ2=1.133, df=1, P=0.287) and adults (GEE: B±SE response variable within a cluster (= gallery identity) to identify 2.703±3.492, χ2=0.599, df=1, P=0.439) present 40 days after the effects of our treatment (induced vs. delayed dispersing treatment (Fig. 2b). This difference in egg numbers is not

Chapter 5

Fig. 2 Comparison of total numbers of offspring produced by the two in the four developmental offspring stages. The differences between the experimental groups of females 40 days after the treatment. a Offspring numbers of eggs, larvae and pupae reflect the average duration of these numbers in dependence of the gallery of origin and the overall total. stages (for X. saxesenii: egg stage - 5 days (range=5), three larval instars - Numbers of daughter galleries included (induced dispersers/delayed 11 days (range=8-21), pupal stage - 7 days (range=6-7); Biedermann et al. dispersers): gallery1 (2/2), g2 (5/6), g3 (5/5), g4 (5/2), g5 (4/4) and g6 2009). Arithmetic means of daughter gallery offspring numbers are (12/11); g7 was omitted because only one female of this gallery shown with their standard errors. GEE: *P< 0.05; **P<0.001 reproduced successfully. b The overall total is split up

an artefact from a few galleries, since we found the mean egg dispersing females never had eggs in their ovaries (14 adult numbers to be higher in the induced dispersers than in the females from four galleries). delayed dispersers in all six galleries analysed (Fig 3 in the "Appendix"). The total number of offspring produced by successful foundresses was strongly affected by the gallery of Discussion origin (GLM: F5,,,63 = 5.6, P<0.001). The first developing female offspring in all X. affinis galleries Measurements of delayed and induced dispersing females stayed and helped in brood care for at least

Body length and weight did not differ significantly between induced (x = 2.85mm, SE < 0.01;x = 0.41mg, SE = 0.01, N = 31) Gallery Total Ovary A B C D (%) and delayed dispersing females (x = 2.85mm, SE < 0.01;x = development 0.4mg, SE = 0.02, N = 13; GEE body length: no variability, 2 P>0.05; GEE body weight: B±SE -0.03±0.52, χ ="0.06, df=1, Dispersing Immature 0 1 8 0 64.3 P=0.95; see Fig. 4 in the "Appendix"). females Mature 2 0 0 3 35.7 Egg-carrying 0 0 0 0 0 About 24% of the induced dispersing females had ovaries Total 2 1 8 3 100 containing eggs (14 of 59 adult females from four galleries; Table 1), which suggests that staying females often produce Staying females Immature 2 2 3 13 33.9 eggs in their natal gallery. The number of delayed dispersers Mature 6 2 1 16 42.4 Egg-carrying 1 5 3 5 23.7 was independent of the number of staying females (= induced Total 9 9 7 34 100 dispersers; Spearman: R=-0.32, P=0.68) and of the number of females among them laying eggs (Spearman: R=-0.21, P= 0.79, Table 1 Developmental status of the ovary of delayed and induced N=4 galleries). The likelihood to breed differed between dispersing females in four dissected galleries galleries (Fisher's exact test: P=0.03, N=59). Delayed

Chapter 5

1week (cf. Roeper et al. 1980) and on average for 12 days after own egg laying in the natal gallery. However, it is unlikely that full sclerotization before they started to disperse (Fig. 1). Also this would fully explain the productivity differences found after dispersal has started, the rate of female eclosion from the between the two groups of females as the total numbers of eggs, pupal stage is higher than the dispersal rate, which causes larvae, pupae and teneral beetles present in the galleries in females to accumulate in the gallery. This suggests that females relation to the total number of adult females were very small at usually delay dispersal, probably in order to help in care of the the time of our treatment (day 60: mean= 2.57, SE=1.15, brood produced by the foundress (i.e. their mother) or later on range=0-8, N=7 galleries), suggesting that there was only very potentially also by sisters. sparse egg laying, if any, by females other than the foundress. (2) Given that females delay their dispersal, our data reject the The delayed dispersing females may have developed flight predictions of the maturation feeding hypothesis. In contrast, muscles for dispersal which the induced dispersers might have the predictions of the other two hypotheses were confirmed: lacked; the reuse of energy when transferring nutrients from Delayed dispersal seems to entail long-term costs, most likely by flight muscles to ovaries may bring about extra synthesis costs the investment in cooperative care, and some females reproduce (Zera and Denno 1997), which could affect productivity (e.g. Elkin in their natal gallery. The data reveal also that philopatric and Reid 2005). This possibility needs to be scrutinised in future females are fully capable to found an own gallery and to start studies focusing on the energetics of dispersal, reproduction and reproducing at any point of time, disproving the assumption that reorganisation of tissue in these beetles. they are infertile before dispersal (Graham 1961). Hitherto, Independently of whether females might benefit from staying prolonged philopatry, which is a common feature of female life by raising their inclusive or direct fitness, dispersal, by contrast, is histories in many phloem (e.g. Kirkendall et al. 1997) and costly due to high mortality risk (Milne and Giese 1970; Dahlsten ambrosia feeding Scolytinae (e.g. Kalshoven 1962; Peer and 1982). In addition, establishing a fungus culture after dispersal Taborsky 2007;Biedermannetal.2009), has been attributed fails very often (only 20% of founded fungus gardens are usually to maturation feeding (Eichhoff 1881; Botterweg 1982; successful in X. saxesenii; Peer and Taborsky 2007; Biedermann McNee et al. 2000). In contrast, our findings indicate that in X. et al. 2009). Therefore, it is conceivable that females might affinis other benefits select for philopatry. Firstly, philopatry can compete for staying in a productive nest (Kokko and Ekman raise the inclusive fitness of females by enhancing the 2002). If this was the case, dispersing females could be the less production of close relatives (Bischoff 2004; Biedermann 2007; competitive individuals, which might explain why they were less Peer and Taborsky 2007). successful in their breeding attempts after dispersal when Galleries with more adult females produced more offspring in compared to philopatric females. However, two results indicate the closely related species X. saxesenii (Peer and Taborsky 2007), that competition for staying does not occur or has little effect on also over multiple generations (Biedermann et al. 2009). Direct dispersal in X. affinis. First, we found no difference in size or body observations of X. affinis and other xyleborine species revealed condition between dispersing and staying females. Second, the that staying adult females share in brood care and fungus numbers of females present in the gallery (including egg layers maintenance (Roeper et al. and non-reproductives) did not relate to the numbers of 1980; Bischoff 2004; Biedermann 2007; Biedermann et al. 2009). dispersers in our experiment. A relationship between female The members of a gallery are almost clones due to mating density and dispersal would be expected, however, if occurring virtually exclusively among full siblings (for X. competition triggers dispersal, regard-less whether dispersal germanus: sib-mating estimate = 97% of matings, Keller et al., submitted for publication; see also Peer and Taborsky 2005). occurs voluntarily or is forced by other colony members. Secondly, direct fitness benefits can apply for females that Delayed dispersing females differed in their ovary devel- reproduce at home, which was true for nearly a quarter of opment from those collected in the gallery, as egg-carrying staying females in our sample. This is a similar proportion as ovaries were only found among the latter. This might suggest observed in X. saxesenii (Biedermann 2007). two different reproductive strategies of females: (1) A dispersal There might be alternative causes to costly helping for the phenotype showing delayed ovary maturation and perhaps also a lowered productivity of delayed dispersers: (1) The reduced strongly developed flight apparatus and (2) a philopatric post-dispersal productivity of females may have resulted from phenotype that stays, helps in brood care and eventually breeds their in the natal gallery. Such strategies are common in insects because there is often a trade-off between

Chapter 5 construction of the flight apparatus and ovary development Therefore, this group can provide insights into intrinsic and (Zera and Denno 1997). Nevertheless, the existence of two ecological factors inducing individuals to stay at home rather than distinct phenotypes is yet unknown from scolytine beetles, and to disperse and to help rather than to reproduce independently, we also do not find evidence for this possibility in our data. First, which are basic components of social evolution. we did not find any differences in body size and weight, which contrasts with theoretical predictions that dispersers should be the stronger competitors under kin competition (Gyllenberg et Acknowledgements We are grateful to Stacy Blomqvist and Eric Ott for collecting X. affinis in the field and starting the first lab galleries. This al. 2008; Bonte and De La Pena 2009); this has been confirmed manuscript benefitted greatly from comments of Tabea Turrini, Dik Heg also in several vertebrate (e.g. Kawata 1987; Cote et al. 2007) and two anonymous reviewers. The study was supported by a cooperative and insect taxa (e.g. Sundstrom 1995; Moore et al. 2006). Body agreement between the Department of Behavioral Ecology, University of condition is an important factor determining successful host Bern and the Southern Research Station, USDA Forest Service. PHWB is a finding and gallery foundation in scolytine beetles (Dendroctonus recipient of a DOC fellowship of the Austrian Academy of Sciences at the Department of Behavioural Ecology, University of Bern, and was partly ponderosae; Latty and Reid 2010). Second, selection for fixed funded by a fellowship of the Roche Research Foundation. behavioural strategies, which would be associated with a dispersal polymorphism, is probably weak for Scolytinae because of their unpredictable environmental conditions. Indeed, there is a multitude of studies showing their enormous flexibility to react adaptively to changing environ-mental conditions, for example, Appendix by transferring nutrients between flight muscles and ovaries back and forth within a few days (e.g. Reid 1958; McNee et al. 2000). This might explain why both female groups showed the same founding success and about the same timing of first egg laying (reflected by the equal number of adult offspring in both treatments), which would be unlikely assuming two distinct phenotypes. In summary, although there is variation in reproductive success and probably flight performance, dis-persal and philopatry are likely rather plastic, condition-dependent strategies which are predicted to be superior to fixed strategies in many cases (e.g. Ims and Hjermann 2001; Bowler and Benton 2005). "Gallery of origin" significantly affected both the number of offspring produced after dispersal and the number of egg layers among staying females. It has been suggested that the mutualistic microbial complex (certain fungi and bacteria) maintained in the gallery is the major factor influencing gallery productivity via the quality of the transmitted mutualistic microbes (Baker and Norris 1968; Kok et al. 1970; Kingsolver and Norris 1977; Batra 1979; Kajimura and Hijii 1994). The beetles' genetic quality might be another factor that has not yet been explored. Unfortunately, our approach does not allow distinguishing between these two potential causes of the observed gallery effects. Xyleborine ambrosia beetles live under the very conditions where higher sociality has probably evolved multiple times Fig. 3 Total numbers of eggs produced by the two experimental groups of females 40 days after the treatment. Induced dispersers laid more eggs (Hamilton 1978), which includes high levels of inbreeding, than delayed dispersers overall (GEE: P<0.001). Numbers of daughter protection of a common nest and a virtually unlimited food galleries included (induced dispersers/delayed dispersers): galley (2/2), g2 source (e.g. some hymenoptera, gall-thrips, aphids and lower (5/6), g3 (5/5), g4 (5/2), g5 (4/4) and g6 (12/11). Arithmetic means are termites: Choe and Crespi 1997; Korb and Heinze 2008). shown with their standard errors

Chapter 5

Bowler DE, Benton TG (2005) Causes and consequences of animal dispersal strategies: relating individual behaviour to spatial dynamics. Biol Rev 80:205-225 Bruintjes R, Hekman R, Taborsky M (2010) Experimental global food reduction raises resource acquisition costs of brood care helpers and reduces their helping effort. Funct Ecol 24:1054- 1063. Helms Cahan S, Blumstein DT, Sundstrom L, Liebig J, Griffin A (2002) Social trajectories and the evolution of social behaviour. Oikos 96:206-216 Choe JC, Crespi BJ (1997) The evolution of social behaviour in insects and arachnids. Cambridge University Press, Cambridge. Costa JT (2006) The other insect societies. Belknap Press of Harvard University Press, Cambridge. Cote J, Clobert J, Fitze PS (2007) Mother-offspring competition promotes colonization success. PNAS 104:9703-9708 Craig R (1983) Subfertility and the evolution of eusociality by kin selection. J Theor Biol 100:379-397 Dahlsten DL (1982) Relationship between bark beetles and their natural enemies. In: Mitton JB, Sturgeon KB (eds) Bark beetles in North American conifers. University of Texas Press, Austin, pp 140-182 Eden SF (1987) When do helpers help - food availability and helping in the moorhen, Gallinula-chloropus. Behav Ecol Sociobiol 21:191-195 Eichhoff W (1881) Die Europäischen Borkenkäfer. Springer, Berlin Fig. 4 Comparison of the body weight (mg) of females from the two Elkin CM, Reid ML (2005) Low energy reserves and energy allocation experimental groups collected in four galleries. There was no significant decisions affect reproduction by mountain pine beetles, difference between induced and delayed dispersers (GEE: P=0.95). Dendroctonus ponderosae. Funct Ecol 19:102-109 Numbers of galleries included (induced dispersers/delayed dispersers): Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH, Jordal BH (2001) The evolution of agriculture in beetles gallerya (9/2), gb (12/7), gc (4/1) and gd (6/3). Arithmetic means of dry weight are shown with their standard errors (Curculionidae: Scolytinae and Platypodinae). Evolution 55:2011-2027 Fischer M (1954) Untersuchungen über den kleinen Holzbohrer (Xyleborus saxeseni). Pflanzenschutzberichte 12:137-180 Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei References Ambrosiakäfern. Z Morph Ökol Tiere 45:275-308 Francke-Grosmann H (1975) Zur epizoischen und endozoischen Baker JM, Norris DM (1968) A complex of fungi mutualistically involved in Übertragung der symbiotischen Pilze des Ambrosiakäfers nutrition of ambrosia beetle Xyleborus ferrugineus .JInv Path Xyleborus saxeseni (Coleoptera: Scolitidae). Entomologica 11:246-250 Germanica 1:279-292 French JRJ, Roeper RA (1975) Studies on the biology of the ambrosia Batra LR (1979) Insect-fungus symbiosis. Nutrition, mutualism, and beetle Xyleborus dispar. J Appl Entomol 78:241-247 commensalism. Wiley, New York Graham K (1961) Air-swallowing—mechanism in photic reversal of the Beaver RA (1986) The taxonomy, mycangia and biology of Hypothenemus beetle Trypodendron. Nature 191:519-520 curtipennis (Schedl), the first known cryphaline ambrosia beetle Gyllenberg M, Kisdi E, Utz M (2008) Evolution of condition-dependent (Coleoptera: Scolytidae). Ent Scand 17:131-135 dispersal under kin competition. J Math Biol 57:285-307 Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia Hager R, Jones CB (2009) Reproductive skew in vertebrates: proximate beetles. In: Wilding N, Collins NM, Hammond PM, Webber JF and ultimate causes. Cambridge University Press, Cambridge (eds) Insect-fungus interactions. Academic, London, pp 121 -143 Hamilton WD (1978) Evolution and diversity under bark. In: Mound LA, Biedermann PHW (2007) Social behaviour in sib mating fungus farmers. Waloff N (eds) Diversity of insect faunas. Blackwell, Oxford, Master thesis, University of Bern pp 154-175 Biedermann PHW, Klepzig KD, Taborsky M (2009) Fungus cultiva-tion by Heg D, Bachar Z, Brouwer L, Taborsky M (2004) Predation risk is an ambrosia beetles: behavior and laboratory breeding success in ecological constraint for helper dispersal in a cooperatively three xyleborine species. Environ Entomol 38:1096-1105 breeding cichlid. Proc R Soc Lond B 271:2367-2374 Bischoff LL (2004) The social structure of the haplodiploid bark beetle, Xylosandrus germanus. Diploma thesis, University of Bern Ims RA, Hjermann DO (2001) Condition-dependent dispersal. In: Clobert Bonte D, De La Pena E (2009) Evolution of body condition-dependent J, Danchin E, Dhondt AA, Nichols JD (eds) Dispersal. Oxford dispersal in metapopulations. J Evol Biol 22:1242-1251 University Press, Oxford, pp 203-216 Botterweg PF (1982) Dispersal and flight behavior of the spruce bark beetle Kajimura H, Hijii N (1994) Reproduction and resource utilization of the Ips typographus in relation to sex, size and fat-content. Z Angew ambrosia beetle, Xylosandrus mutilatus, in-field and Entomol 94:466-489. experiment populations. Entomol Exp Appl 71:121-132 Kalshoven LGE (1962) Note on the habits of Xyleborus destruens Bldf., the near-primary borer of teak trees on Java. Entomol Berichten 22:7-18

Chapter 5 Kawata M (1987) The effect of kinship on spacing among female Norusis M (2007) Generalized estimating equations. In: Norusis red-backed voles, Clethrionomys rufocanus M (ed) SPSS 15.0 advanced statistical procedures bedfordiae. Oecologia 72:115-122 companion. Prentice Hall, Upper Saddle River, pp Keller L, Reeve HK (1994) Partitioning of reproduction in animal 279-296 societies. Trends Ecol Evol 9:98-102 Peer K, Taborsky M (2004) Female ambrosia beetles adjust their Kingsolver JG, Norris DM (1977) The interaction of Xyleborus offspring sex ratio according to outbreeding ferrugineus Fabr. (Coleoptera: Scolytidae) behavior opportunities for their sons. J Evol Biol 17:257-264 and initial reproduction in relation to its symbiotic Peer K, Taborsky M (2005) Outbreeding depression, but no fungi. Ann Entomol Soc Am 70:1-4 inbreeding depression in haplodiploid ambrosia Kirkendall LR, Kent DS, Raffa KF (1997) Interactions among beetles with regular sibling mating. Evolution males, females and offspring in bark and ambrosia 59:317-323 beetles: the significance of living in tunnels for the Peer K, Taborsky M (2007) Delayed dispersal as a potential evolution of social behavior. In: Choe JC, Crespi BJ route to cooperative breeding in ambrosia beetles. (eds) The evolution of social behavior in insects and Behav Ecol Sociobiol 61:729-739 arachnids. Cambridge University Press, Cambridge, R Development Core Team (2008) R: a language and pp 181 -215 environment for statistical computing. R Foundation Koenig WD, Pitelka FA, Carmen WJ, Mumme RL (1992) The for Statistical Computing, Vienna evolution of delayed dispersal in cooperative Reid RW (1958) The behaviour of the mountain pine beetle, breeders. Q Rev Biol 67:111-150 Dendroctonus monticolae Hopk., during mating, egg Kok LT, Norris DM, Chu HM (1970) Sterol metabolism as a basis laying, and gallery construction. Can Entomol for a mutualistic symbiosis. Nature 225:661 -662 90:505-509 Kokko H, Ekman J (2002) Delayed dispersal as a route to Roeper R, Treeful LM, O'Brien KM, Foote RA, Bunce MA (1980) breeding: territorial inheritance, safe havens, and Life history of the ambrosia beetle Xyleborus affinis ecological constraints. Am Nat 160:468-484 (Coleoptera: Scolytidae) from in vitro culture. Great Korb J, Heinze J (2008) The ecology of social life: a synthesis. In: Lakes Entomol 13:141-144 Korb J, Heinze J (eds) Ecology of social evolution. Roisin Y (1994) Intragroup conflicts and the evolution of sterile Springer, Berlin, pp 245-259 castes in termites. Am Nat 143:751-765 Krausseopatz B, Kohler U, Schopf R (1995) The energetic state Schedl KE (1956) Breeding habits of arboricole insects in Central of lps typographus L. (Coleoptera, Scolytidae) during Africa. Verh 10 Int Kongr Entomol (Vienna) 1:183- the life-cycle. Z Angew Entomol 119:185-194 197 Liang KY, Zeger SL (1986) Longitudinal data-analysis using Schneider I (1987) Distribution, fungus-transfer and gallery generalized linear-models. Biometrika 73:13-22 construction of the ambrosia beetle Xyleborus affinis Latty TM, Reid ML (2010) Who goes first? Condition and danger in comparison with X. mascarensis (Coleoptera, dependent pioneering in a group-living bark beetle Scolytidae). Entomologia Generalis 12:267-275 (Dendroctonus ponderosae). Behav Ecol Sociobiol Stacey PB, Koenig WD (1990) Cooperative breeding in birds: 64:639-646 long-term studies of ecology and behavior. McNee WR, Wood DL, Storer AJ (2000) Pre-emergence feeding Cambridge University Press, Cambridge in bark beetles (Coleoptera: Scolytidae). Environ Stacey PB, Ligon JD (1991) The benefits-of-philopatry Entomol 29:495-501 hypothesis for the evolution of cooperative Milne DH, Giese RL (1970) Biology of the Columbian timber breeding: variation in territory quality and group size beetle, Corthylus columbianus (Coleoptera: effects. Am Nat 117:831-846 Scolytidae). 10. Comparison of yearly mortality and Sundstrom L (1995) Dispersal polymorphism and physiological dispersal losses with population densities. Entomol condition of males and females in the ant, Formica News 81:12-24 truncorum. Behav Ecol 6:132-139 Moore JC, Loggenberg A, Greeff JM (2006) Kin competition West-Eberhard MJ (1975) The evolution of social behaviour by promotes dispersal in a male pollinating fig wasp. kin selection. Q Rev Biol 50:1-33 Biol Lett 2:17-19 Whitney HS (1971) Association of Dendroctonus ponderosae Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR (2005) with blue stain fungi and yeasts during brood The evolution of agriculture in insects. Ann Rev Ecol development in lodgepole pine. Can Entomol Evol Syst 36:563-595 103:1495-1503 Zeger SL, Liang KY (1986) Longitudinal data-analysis for discrete and continuous outcomes. Biometrics 42:121 -130 Zera AJ, Denno RF (1997) Physiology and ecology of dispersal polymorphism in insects. Ann Rev Entomol 42:207- 230

Chapter 6

Social Fungus Farming Varies among Ambrosia Beetles

Peter H.W. Biedermann1,2 and M. Taborsky1

1Department of Behavioural Ecology, Institute of Ecology & Evolution, University of Berne, Baltzerstrasse 6, CH-3012 Bern, Switzerland

2 USDA Forest Service, Southern Research Station, 2500 Shreveport Hwy, Pineville/LA, 71360, USA

Correspondance: [email protected]

Running title: Social fungus farming by ambrosia beetles

Abstract Advanced fungus agriculture independently evolved once in ants, once in termites, and seven times in weevils. In the latter, the so-called ambrosia beetles, this evolution was accompanied by an increase of social complexity leading to division of labour. Ambrosia beetles live in subsocial or eusocial family groups within self-built galleries in wood, where they farm ambrosia fungi for food. Gallery morphologies vary considerably between species, which is assumed to affect the frequency of interactions between group members, i.e. adults and brood, as well as their mutualistic ambrosia fungi. Here we illustrate the intra- and interspecific variation in ambrosia beetle gallery morphology and describe how this affects within-family interactions in two exemplary species. Xyleborinus saxesenii inhabits cave-like galleries, whereas Xyleborus affinis (like most other ambrosia beetles) digs out branching tunnel systems. Life histories of both species are very similar, but mature X. affinis engage less in brood care and fungus cropping relative to mature X. saxesenii. Remarkably, however, reduced brood care by adults is compensated with increased grooming by larvae in X. affinis, which is likely an effect of the spatial separation between adults and brood as well as fungus in the tunnel systems. By contrast, in the caves of X. saxesenii individuals of all age classes are mixed and closely interact with the fungal layers lining the gallery walls. The cave-like gallery pattern results from larvae feeding on wood and fungus instead of just grazing off the fungus layer from tunnel walls. Two fundamentally different ways of fungus agriculture are thereby indicated to exist in the subtribe Xyleborini (Scolytinae) and much more variation may still be uncovered in other ambrosia beetle lineages.

Introduction

The cultivation of crops for nourishment is not restricted to humans. Three lineages of insects – fungus-growing ants (Attini), fungus-growing termites (Macrotermitinae), and ambrosia beetles (Curculionidae: Scolytinae, Platypodinae) – evolved the skills for fungus agriculture. This includes habitual planting or insemination, cultivation, protection and harvesting of the cultivar, on which they inevitably depend for nutrition (Mueller et al. 2005). These complex tasks are shared between different individuals in a gallery. Sociality and division of labour has therefore been proposed to have played a crucial role in the evolution of fungus agriculture (Mueller et al. 2005; Biedermann and Taborsky 2011).

Eusociality and division of labour have apparently facilitated the evolution of agriculture in ants and termites, because agricultural tasks may be partitioned between group members (Hölldobler and Wilson 1990; Hart et al. 2002). Different worker castes in ants and termites are specialized in either foraging, chopping and cleaning the substrate before incorporation into the

Chapter 6 garden, planting of mycelium onto new substrate, monitoring and weeding of the garden, and disposal of diseased or senescent parts of the garden (Traniello and Leuthold 2000; Bot et al. 2001; Hart et al. 2002; Mueller et al. 2005). Ambrosia beetle agriculture must have originated differently, because they originate from a non-social lineage with only parental care (Farrell et al. 2001). Interestingly, the transition to a fungal diet evolved convergently at least eight times in weevils (Farrell et al. 2001). In all three clades that have hitherto been examined more closely – Xyleborini, Xyloterini, and Platypodinae – this transition was accompanied by an increase in social complexity (Kirkendall et al. 1997). Female offspring often delay dispersal from the natal nest, resulting in overlapping generations (e.g. Merkl and Tusnadi 1992; Peer and Taborsky 2007; Biedermann et al. 2009; 2012), and a recent direct behavioural study revealed cooperative breeding and a complex system of task sharing between adult and larval stages in Xyleborinus saxesenii (Scolytinae: Xyleborini; (Biedermann and Taborsky 2011). Eusociality, i.e., permanently sterile and reproductive castes, was not observed in this species, but was suggested in the ambrosia beetle Austroplatypus incompertus (Platypodinae) after examination of morphological differences between adult females (Kent and Simpson 1992).

Social evolution is generally triggered by constraints on independent breeding and / or benefits of philopatry (e.g. Selander 1964; Stacey and Ligon 1987; Kokko and Ekman 2002; Korb 2008). In ambrosia beetles constraints on independent breeding are probably immense. Freshly dead wood, where ambrosia beetles culture their fungi, is a patchily distributed resource, and mortality due to predation or adverse environmental conditions during dispersal is assumed to be very high (Milne and Giese 1970; Dahlsten 1982). Once a suitable host tree has been detected, a founder beetle has to excavate a gallery, overcome tree defences like resins, and successfully establish the ambrosia fungus, which fails in 70-80% of all females (e.g. Fischer 1954; Biedermann 2007; Peer and Taborsky 2007). By contrast, benefits of philopatry in a productive natal gallery may include fitness gains by (1) co-breeding and (2) helping close relatives to produce more offspring (Biedermann et al. 2011). All Xyleborini are predisposed for kin-selected benefits as a result of haplodiploidy and obligate brother-sister matings in the natal gallery (Peer and Taborsky 2005), which lead to high genetic relatedness (Peer and Taborsky 2007). Indeed, delayed dispersal of daughters appears to be the rule in Xyleborini, which help their mothers with gallery protection and hygiene, brood and fungus care (Biedermann et al. 2012). The participation of larvae in these cooperative tasks is unique among Holometabola (Biedermann and Taborsky 2011).

The beetles’ agriculture differs from that of other insects in one important feature: the longevity/stability of the substrate on which that fungi grow. Ants and termites collect the substrate (woody branches, leaves and grass) for their crops while the beetles live within the substrate that os consumed by the fungus (wood). In the first case, substrate is renewable and fungus gardens are potentially long-lasting, while in the second case resources are locally exhaustible as nutrients wear out (Kirkendall et al. 1997). Thus, ambrosia beetles that culture their fungi in tunnel systems (galleries) in the sap- and heartwood of trees need to constantly enlarge the system to continuously provide new substrate for their fungi. In addition, most ambrosia beetles settle in freshly dead trees, where their fungi are faced with an increased competition from other microorganisms over time. Recycling of beetle excretions by fungi (French and Roeper 1973; Norris 1975), or hygienic behaviours (Cardoza et al. 2006) and bacteria producing antibiotics (Scott et al. 2008) may postpone the spread of competing microorganisms, but their conquest cannot be hindered completely. Only the colonization of gradually dying trees, which sometimes occurs in X. saxesenii and X. affinis, may enable continuous breeding and several beetle generations to stay and overlap within one gallery; single X. saxesenii galleries with more than 300 adult inhabitants have been reported (Hosking 1972) and tunnels of a single X. affinis gallery may be inhabited for more than four years and reach a total length of 6m (Schneider 1987). Nevertheless, such cases are probably rare and not the rule, and breeding in dead wood certainly constrains the longevity of fungus gardens and the evolution of social complexity also in these two species. This may explain why eusociality has yet been suggested

Chapter 6 only for A. incompertus, a species breeding in long-lasting galleries (> 36 years) in living trees (Kent and Simpson 1992).

Another important factor potentially affecting ambrosia beetle sociality is the morphology of their galleries. A gallery made of a branching tunnel system is typical for the majority of Xyleborini and Platypodinae. Whilst adult stages are apparently randomly spaced, their brood is only found in the tips of newly expanded tunnels, where the fungus growth is highest (Fig. 2). Like adults, larvae are mycetophagous and feed on the palisade-like fungal layer on the tunnel walls. By contrast, species in the xyleborine genera Xyleborinus and Xylosandrus construct cave-like brood chambers, where adults and offspring live in close contact to each other and their wall-lining fungus gardens (Fig.1). These chambers result from the effort of larvae that feed xylo-mycetophagously, not only on the wall-covering fungal layers but also on fungus-infested wood (Schedl 1956; Roeper 1995). There is little digestion of the swallowed wood (Francke-Grosmann 1967; Roeper 1995), but it is likely that decomposition of wood is enhanced due to the spreading of larval faeces on the tunnel walls that may get reutilized by the growing fungi (Hubbard 1897, Francke-Grosmann 1967). In the other tunnel-constructing ambrosia beetle lineages, social interactions may be constrained if the tunnels inhibit the free passing of individuals (Kirkendall et al. 1997).

In this study we compare two cooperatively breeding species of fungus-growing ambrosia beetles that differ in the type of galleries produced. We describe the behavioural caste that is responsible for digging the galleries and estimate the extent of social behaviours. As model species we use X. saxesenii that build cave-like brood chambers and X. affinis that build the more common branching-tunnel galleries. This comparison will hint on whether the complex system of division of labour found in X. saxesenii (Biedermann and Taborsky 2011) can be assumed to exist only in species living in brood chambers, or also in the majority of ambrosia beetles, which live in tunnels. Finally, we describe the interspecific and intraspecific variation of the structure of their gallery systems.

Material & Methods

Study Species

Xyleborinus saxesenii Ratzeburg and Xyleborus affinis Eichhoff are probably the most common members of the ambrosia beetle tribe Xyleborini (Curculionidae: Scolytinae) in the Southern states of the US, and among the most widely distributed ambrosia beetles around the world today (Wood 1982). X. saxesenii originates from temperate Eurasia, whereas X. affinis is native to the American tropics and subtropics, and both are highly successful invaders of other continents. Like all ambrosia beetles, they are solely living on a complex of self-cultured fungi, grown within a tunnel system bored into the xylem of trees (Beaver 1989). Hosts are recently dead or highly stressed trees that emit ethanol, which attracts founder females with fungus-filled mycetangia (spore carrying organs; Ranger et al. 2010). As both species are attacking a wide variety of deciduous and coniferous tree species, it appears that their main mutualistic fungi are very polyphagous (Wood 1982). These fungi dominate the microbial community within healthy beetle galleries (Kajimura and Hijii 1992). An unidentified Raffaelea sp. (Roeper and French, 1981) or Cephalosporium pallidum Verrall (Verrall 1943) serve as primary food for X. affinis; Raffaelea sulphurea (L.R. Batra) T.C. Harr. (syn. Ambrosiella sulfurea; Harrington et al. 2010) serves as primary food for X. saxesenii (Batra 1966; Francke- Grosmann 1975). Both beetles show a strongly female-biased sex-ratio due to local mate competition as a result of brother-sister mating (1:8-20; Roeper et al. 1980; Biedermann 2010; Biedermann et al. 2011; cf. Peer and Taborsky 2004).

Illustrations of Gallery Morphology

Chapter 6

For studying the morphology of X. saxesenii galleries we dissected a disc (80 × 15cm - diameter × height) cut off from a trunk of beech (Fagus sylvatica) near Bern/CH (560 m asl, 46°95’, 7°31’) on 9th November 2008. The trunk was still rooted in the ground and part of a tree felled about one year before. Boring holes of beetles were solely found on the moist, north-facing side of the trunk. The disc was brought to the lab, and beetle tunnels were carefully dissected. Apart from at least 40 galleries of X. saxesenii, the disc contained also galleries of the ambrosia beetles Xylosandrus germanus (N = 5 galleries) and Anisandrus dispar (N = 2), as well as the ship-timber beetle Hylecoetus dermestoides (Lymexylidae; N > 50).

Inhabitants of all X. saxesenii galleries were counted and the shape of the galleries was traced in original size on paper whenever possible. These drawings were scanned and colour modified with Adobe Photoshop CS2 (Version 9.0, © Adobe Systems Inc. San Jose/CA, USA, 1999- 2005) to suit graphical illustration. Gallery size was analysed by using GSA Image Analyser (Demoversion 2011-02-10; © Software development & Analytics Bansemer & Scheel GbR, Rostock, Germany).

To display the morphology of X. affinis galleries we scanned eight original-sized gallery drawings from (Schedl 1962) (pp. 362-365) and modified them with Adobe Photoshop CS2 in the same way as described above.

Laboratory Rearing and Observations.

X. affinis females used in this study were collected from oak logs, whereas X. saxesenii females were caught with Lindgren funnel traps baited with ethanol, both near Pineville/LA, USA (123ft asl; 31°20’, 92°24’) in Summer 2007. After bringing all females to the lab we prepared them for laboratory rearing by surface sterilization, rinsing them for some seconds with 95% ethanol and afterwards with deionised water. This treatment reduces fungal contamination by elimination of fungal spores sticking to the body surface of the beetles, but does not harm the fungal spores of the cultivar within the mycetangium, where they are kept for transmission by the beetles (Francke- Grosmann 1956). Afterwards we placed each female singly on a standard rearing medium of agar and sawdust that was prepared in glass tubes (for more details on this technique see Biedermann and others 2009). About 50 tubes per species were established in this way. Tubes were closed with plastic caps, stored at room temperature (~23°C) and wrapped in paper, allowing light only to shine onto the surface of the media. This way beetles bore tunnels frequently next to the tube glass, which allows behavioural observations when the paper is removed (Fig. 2C).

A female immediately starts boring an entrance tunnel when placed onto the medium. While boring it inseminates the tunnels with ambrosia fungus spores from its mycetangium or the gut (Francke-Grosmann 1956; Francke-Grosmann 1975). If she is successful this results in fungus growth on the walls within a few days (Roeper et al. 1980). After feeding on the fungus she starts to lay eggs (Kingsolver and Norris 1977) and usually proceeds to do so for the next 40 days. During that time her offspring pass three larval instars and a pupal stage, and finally they pass through a short maturation period until they are fully sclerotized. First adult progeny appears around day 27 after gallery initiation (Roeper et al. 1980; Biedermann et al. 2011).

Every third to fourth day we observed the behaviour of all individuals in 19 successfully established galleries per species, between days 22 and 42 after initiation. As not all tubes had tunnels next to the glass that could be observed from the beginning, observations ranged from 3 to 6 per gallery. During each observation we noted the number of eggs and pupae and scanned the behaviour of all 1st instar larvae, 2nd/3rd instar larvae, teneral females (not fully sclerotized), mature females and males. Age and sex classes can be easily discerned by their appearance (for details see Fischer 1954 and Biedermann et al. 2009).

Chapter 6

Behavioural and Statistical Analyses

All ambrosia beetle behaviours have been described by Biedermann and Taborsky (2011). Here we focus only on the behaviours of larvae and adults that are of interest in a social context (see Table 1). Digging behaviour of developing offspring increases the size of the egg-niche to a brood chamber and thus creates wall surface for the fungal gardens to flourish (Bright 1973). Digging is part of larval feeding on fungal hyphae that penetrate the wood (wood is not digested; (Batra 1966; Haack and Slansky 1987) and thus cannot be separated from it. In contrast, the digging of adults is clearly different from feeding, and solely serves gallery extension. By cropping, adults graze off ambrosia (Fig. 1), which apparently enhances fungus growth (Biedermann and Currie, unpubl. data). Consequently, weeds do not invade (see references in Lengerken 1939; Norris 1993) and asexual fruiting (= ambrosial growth) is induced (Schneider-Orelli 1913; Batra and Michie 1963; French and Roeper 1972). Gallery hygiene is important for a flourishing fungus garden as well as a healthy colony and can be attributed to three behaviours: balling, shuffling and cannibalising. By balling, larvae compress faeces and small pieces of wood into compact balls, which facilitates further transport. Those frass-balls are moved towards fungus beds by shuffling, where some of their nutrients appear to get recycled (Xyleborus ferrugineus; Kok and Norris 1972). Frass, pieces of wood, and sclerotized parts of dead siblings get also shuffled towards unused tunnel-parts or to the entrance, where they are thrown out of the gallery. The function of cannibalising is not completely clear as on the one hand it often involves dead or hardly moving siblings that probably need to be removed to maintain gallery hygiene, but occasionally also individuals that seem to move normally are consumed. The health status of colony members might be checked during allogrooming, which appears crucial for individual survival, as microorganisms quickly cover uncleaned beetle bodies and brood (Biedermann and Taborsky 2011). Finally, blocking of the gallery entrance has been suggested to serve a variety of functions like protection of the brood and regulation of the microclimate inside the gallery (see Kirkendall 1993 and Kirkendall et al. 1997 for review).

Generalized estimating equations (GEEs) were used to analyse species differences in the relative behavioural frequency (binary response variable) per class (1st instar larvae, 2nd/3rd instar larvae, teneral females, mature females, and males) by controlling for gallery age. The latter had no significant effect on the behavioural frequencies, so it was excluded from the models. GEEs are an extension of generalized linear models with an exchangeable correlation structure of the response variable within a cluster, which allows controlling for the variation between observations from a single gallery. GEEs were performed using R (lmer in R; Version 2.12.1; R Development Core Team 2008).

Results

Gallery Morphology

Eighteen of >40 dissected field galleries of X. saxesenii were successfully traced and analysed. Interestingly, although distances between tunnels of different wood boring beetles in the disc were very close, they never merged. X. saxesenii galleries are illustrated in Fig. 1. All galleries were flat with a clearance of about 1mm and usually expanded within one plane (except Fig. 1F,G), always along the direction of the wood fibres, often in a right angle to the annual rings. Entrance tunnels were directed vertically to the tree surface, had a length of 5 to 50mm and were followed by one or several distinct brood chambers of different shapes. Galleries with a single, large brood chamber seemed to harbour the most individuals (Fig. 1O-R; Table S1). About half of the galleries (N = 8) were inhabited by less than 10 individuals, and about a quarter by more than 50 individuals (N = 4). The number of individuals correlated strongly with gallery area (R2 = 0.62; linear regression: F = 26.2, p < 0.001, DF= 1, N = 18). The galleries encompassed on average 126.1 mm2 (± 17 se) and inhabited 23.2 ± 6.5 individuals, with an adult sex-ratio of 1.2 : 8.4 (m : f). Hence, there was about one individual per 83

Chapter 6

5 mm2, which makes about 10 mm2 gallery wall space per individual (both sides of the gallery). Brood and adults were randomly distributed within a chamber (Fig. 1R). In contrast, brood in X. affinis galleries was only located in the tips of the tunnels (Fig. 2B,C). X. affinis galleries apparently merged sometimes or alternatively females bored a second exit (Fig. 2I; Schedl 1962). Schedl (1962) described tunnels to usually expand within one plane, but also above and below if attack densities get very high. Tunnels of X. affinis can be frequently found also in the phloem, but still with their normal mycetophagous habit. The eight traced galleries (Fig. 2B-I) that were not fully traced, had a mean total length of 172.8 mm (Range: 71-294 mm), had between 3 and 13 side-tunnels and split the first time on average 11.8 mm (Range: 4-26 mm) behind the entrance. Schedl (1962) reported a mean of 61.1 (Range: 4-223; N = 18 galleries) adult females and 1.5 (Range: 0-4) adult males per gallery; this gave a strongly female-biased mean sex-ratio of 0.024.

Behavioural differences between X. affinis and X. saxesenii

Larval instars

The way of feeding differed between the species. X. affinis larvae are mycetophagous, solely cropping the fungal layers from their gallery walls. By contrast, larvae of X. saxesenii expand the gallery by their xylo-mycetophagous feeding behaviour (= digging). When feeding frequency was compared between species, it turned out that digging was more common than cropping (GEE: p = 0.007; Fig. 3, Table 3) in 2nd/3rd instars.

There were no behavioural differences among the 1st instar larvae of the two species, except that allogrooming was significantly more common in X. affinis (GEE: p = 0.028). This difference was also highly significant in 2nd/3rd instars (GEE: p < 0.001; Fig. 3, Table S2). Balling was only found in 2nd/3rd instars of X. saxesenii and served to compile faeces and sawdust that accumulated in large amounts from the larval digging habit.

Teneral and mature females, and males

Behavioural analyses revealed no statistically significant difference between the behaviour of teneral females. Mature females of both species did not differ in their time spent with blocking and digging (GEE: p > 0.05; Fig. 4, Table S2). Cropping of the fungus (GEE: p = 0.02) and allogrooming (GEE: p < 0.001) was more common in mature females of X. saxesenii, whereas X. affinis mature females showed more shuffling (GEE: p < 0.001). Males of both species did not differ in any of the behaviours observed.

Discussion

The existence of long-lived kin groups is the most likely precondition for the evolution of agriculture (Brock et al. 2011). Examples include fungus-growing ants and termites (Mueller et al. 2005), bacteria-growing social amoeba (Brock et al. 2011) and ambrosia beetles of the Xyleborini tribe (Scolytinae; (Schneider-Orelli 1913; Francke-Grosmann 1967; Norris 1993; Biedermann et al. 2009). The latter are characterized by fungus agriculture, inbreeding, haplodiploidy, and varying social organizations (Peer and Taborsky 2007). Cooperative breeding and larval workers have been recently described for X. saxesenii (Biedermann and Taborsky 2011) and may exist also in other Xyleborini.

Here we have aimed to illustrate the complex cave-tunnel and branching-tunnel gallery systems of Xyleborini, which are two of six types of gallery morphologies known from ambrosia beetles (for details see Fig.5, Table 2). Apart from the mating system (inbreeding within the nest vs. 84

Chapter 6 outbreeding), the main factor affecting gallery structure is the feeding habit of the larvae, which is either xylomycetophagous on wood and fungus (digging in X. saxesenii) or mycetophagous on fungus only (cropping in X. affinis). This has implications interactions among gallery members and between beetles and their fungi. In the cave-tunnel systems of X. saxesenii, where brood and adults are intermingled and surrounded by their fungi, adults showed more brood care (allogrooming) and interactions with the fungus (cropping) than X. affinis adults in their galleries. Wood-containing, larval faeces were either smeared on the tunnel walls (Hubbard 1897) probably for reutilization by the fungus, or prepared by the larvae for removal (balling). By contrast, brood was constrained to the tips of the branching-tunnels in X. affinis. This spacial separation between adults and larvae was probably the reason for larvae partly undertaking the allogrooming of the mature females. The latter interacted also less frequently with the fungi (cropping), but instead spent half of their time in queues with others, shuffling the tunnels free of waste. Previously, individuals of both species have been observed to adjust their behaviours to the demands in their galleries: (i) X. saxesenii mature females adjust their per capita allogrooming frequency to the number of brood present (Biedermann and Taborsky 2011), and (ii) X. affinis larvae and adults reduce cropping- and allogrooming-activities in the entrance part of the gallery relative to the tips of the tunnels, where most of the brood and fungus is present (Biedermann, Taborsky and Six, in preparation). These observations suggest that social task sharing between larvae and adults is not restricted to cave-building species (Biedermann and Taborsky 2011), but might also exist in the majority of ambrosia beetles inhabiting branching- tunnel systems.

Gallery architecture All X. saxesenii galleries we dissected had an entrance tunnel and a cave-like brood chamber expanding along the wood grain. The shape of this chamber was highly variable and in some cases several chambers were lined up in sequence. This suggests that the mycelium of Raffaelea sulfurea, the mutualistic fungus, develops along the grain, but penetrates only the wood cells close to the gallery surface (e.g. Leach et al. 1940; Francke-Grosmann 1963); this growth is probably followed by the larval gnawing activity and thus the chamber gets a drawn-out shape. The existence of multiple chambers may indicate that the fungus does not grow equally well in all parts of the gallery.

The high numbers of brood and adults found in field galleries in November suggest that X. saxesenii galleries persist longer than for one season. It is very likely that the galleries had been founded about 6 months earlier, because the trees had been felled one year before and the main founding period of galleries for X. saxesenii is spring (Fischer 1954; Peer and Taborsky 2007). Brood was abundant and the wood around galleries showed no signs of microbial decay when it was dissected (although it was densely settled by other wood-boring beetles already). Therefore, it seems likely that in most galleries the adults, which are descendants of the founder female, would have continued breeding in the following year. In the laboratory, founder females die about two months after gallery foundation (Biedermann et al. 2012), and mature daughters may then inherit the gallery (Biedermann et al. 2009; 2012).

Dissections of all mature females in X. affinis laboratory galleries showed that around the time when the founder female dies more than half of the daughters reproduce (unpublished data; see Appendix 1 of the thesis). Field galleries of the sub-tropical X. affinis were not studied here, but previous studies suggest that they may get remarkably large (up to 6 m in length) and old (up to 4 years; Schedl 1962, Scheider 1987). This information is important because it shows that with a generation time of about a month (in the laboratory; Biedermann et al. 2009), multiple overlapping generations within one gallery are possible, which increases the probability that complex social systems evolve (Kirkendall et al. 1997; Peer and Taborsky 2007).

Chapter 6

Behavioural differences Gallery differences morphology between Xyleborus affinis and Xyleborinus saxesenii were confirmed also when breeding the beetles in artificial medium in the laboratory, suggesting that divergent gallery structure is not a result of breeding in a different substrate. We never observed gallery enlargement by larvae of X. affinis during the trials, although they were reared in an agar- sawdust-medium that is much easier to gnaw than solid wood. By contrast, in X. saxesenii almost all gallery excavation is due to the larval digging activity (Biedermann and Taborsky 2011). Digging by adults was seen in very low and similar frequencies in both species. These observations are in line with (Roeper 1995) suggestion that different feeding modes of larval instars and not the digging activity of adults are responsible for the different gallery structures of ambrosia beetles.

Interestingly, X. saxesenii larval digging was more common than X. affinis larval cropping. Considering the similar size and developmental period length of both species X. saxesenii larvae may need to spend more cooperative effort than X. affinis larvae. Digging in wood probably demands more energy expenditure than cropping off fungus, and in addition there are probably less nutrients in fungus-infested wood material (wood cannot be digested by the larval digestive system; (Francke- Grosmann 1967) than in pure fungal material. So why do larvae of X. saxesenii engage in digging instead of cropping, if the cost-benefit ratio is presumably worse? First, it appears that Raffaelea sulfurea (the primary mutualist of X. saxesenii) does not produce a fungal layer comparable in thickness to the unidentified primary mutualist of X. affinis (which might be an unidentified Raffaelea sp. Roeper and French (1981), or Cephalosporium pallidum Verrall (1943)), so the lower productivity of the fungus on the gallery surface could force larvae to dig, which may raise its productivity by providing more space for sporulation. In this case, selection of the larval behaviour would be driven by the fungus. Second, in X. saxesenii galleries digestion of the wood may be achieved by the joint action of larval offspring and Raffaelea sulfurea. X. saxesenii adults and larvae show different enzymatic profiles, with xylanases only present in larvae (De Fine Licht and Biedermann, submitted). Additionally, chewed-up wood swallowed by larvae provides more surface for the fungal enzymes to interact with (similar to wood-feeding termites; Watanabe and Tokuda 2009), both within intestines of larvae and in their faeces that are partly smeared on the gallery walls (Hubbard 1897; Biedermann and Taborsky 2011), probably for further utilization by the fungus. Third, larval digging may be selected because it creates more space for fungal growth and additional brood.

In the field, X. saxesenii gallery size correlated with the number of colony members. inhabited more individuals. In contrast to X. affinis , in X. saxesenii other group members benefit from the larval feeding activity because this increases space and thus reduces within-group competition for food (Biedermann and Taborsky 2011). Larval digging, however, also produces large amounts of faecal waste, which is removed from the gallery in two steps: (i) by the balling behaviour of larvae that apparently serves to pile up the faeces in balls, which are then displaced and shuffled out of the gallery by adult females. Balling was not observed in X. affinis larvae; much less faeces were produced in their galleries and the faeces did not contain woody parts. Adult X. affinis seem to be much more important for gallery hygiene given their shuffling behaviour relative to X. saxesenii adults, because waste can quickly block a tunnel if it is not immediately removed. In X. affinis, multiple shuffling adults were often queuing within tunnels, passing along faeces and sawdust for final disposal outside the entrance.

Brood in X. affinis galleries was confined to the ends of the tunnels were fungi seemed to grow best, while adults usually stayed in other parts of the tunnels. By contrast, in X. saxesenii brood and adults were fully mixed in the common brood chamber together with the fungus, where adults engaged in (a) allogrooming and (b) cropping of the fungus. Allogrooming is necessary, because individuals that are not cleaned by others are quickly overgrown by mold (Paecilomyces sp. and Fusicolla sp.; Biedermann et al., submitted), which can quickly kill them (Biedermann and Taborsky 2011). Interestingly, in X. affinis larvae show more allogrooming than adults and also more than X. saxesenii larvae, probably because X. affinis larvae more frequently meet each other due to the 86

Chapter 6 spatial conditions in the galleries. Cropping of the fungus apparently serves both nutrient intake and fungus care. It has been shown to induce the ambrosial growth of the fungus, which is necessary for consumption (French and Roeper 1972) and to prevent the invasion of fungal pathogens (Schneider- Orelli 1913; Norris 1993). Oral secretions might be involved, which in other scolytine beetles were shown to contain bacteria with antibiotic effects against antagonistic fungi (Cardoza et al. 2006; Scott et al. 2008). X. saxesenii adults showed more cropping of the fungus than X. affinis adults, but this difference cannot be easily interpreted as these species use different fungi for agriculture.

In conclusion, it seems that living in tunnels does not restrict the evolution of social complexity in ambrosia beetles (Fig. 5; Table 2). In contrast to previous suggestions (Kirkendall et al. 1997), individuals in X. affinis tunnel systems can easily pass each other, allowing for similarly complex social interactions as previously found in X. saxesenii (Biedermann and Taborsky 2011), a species with a cave-like gallery structure. Cooperative task specialisation and age polyethism is also shown by X. affinis. Most remarkably, even larvae of X. affinis behave cooperatively, suggesting that larval workers are probably the rule rather than the exception in gregariously living weevils, the only holometabolous insect group in which active social behaviours by larvae have been reported (Biedermann and Taborsky 2011). Hence, we propose that there might be modes of fungus agriculture differing mainly in the contribution of larvae and adult offspring (Fig. 5; Table 2): (1) Cooperative xylomycetophagous fungus agriculture (in Xyleborinus sp., Xylosandrus sp. and Platypodinae), where larvae excavate a brood chamber or tunnels by feeding on fungus infested wood. Large amounts of wood are swallowed and probably partly digested with the help of enzymes (De Fine Licht and Biedermann 2012), and faeces may be smeared against the walls for reutilization by the growing fungus. A common brood chamber facilitates interactions between individuals and with the fungus. Larvae may engage in balling of faeces and sawdust (yet reported only for Xyleborinus saxesenii). (2) Cooperative mycetophagous fungus agriculture (in Xyleborus sp.), where larvae feed on fungal layers covering the gallery walls, and adults enlarge the gallery system by digging. Faeces do not contain woody parts but are still partly reutilized by the growing fungi (Kok and Norris 1972; Norris 1975). Larvae harvest the fungus and interact with each other in the tips of the tunnels, and thus are largely separated from the adults that keep the afferent tunnels free of waste. (3) Subsocial xylomycetophagous fungus agriculture (in Trypodendron sp.), where larvae excavate a cradle (pupal niche) for themselves and do not interact with other brood, but there are observations that they interact with the adults (Hubbard 1897; Borden 1988). Parents bore the tunnels and keep them free of larval frass and protect the nest (Borden 1988; Kirkendall et al. 1997).

The first two modes of fungus agriculture may coincide with a (A) simple or (B) complex social type, which is reflected in the complexity of the gallery pattern. Species exhibiting a simple social type (e.g. Xylosandrus sp., Anisandrus sp., Trypodendron sp.) usually use galleries only for the development of a single offspring generation (Fig. 5A,C,E), whereas species exhibiting a more complex social type (e.g. Xyleborinus sp., some Xyleborus sp., Platypus sp.) typically grow several offspring generations within one gallery (Fig. 5B,D,F). Division of labour between larvae and adults and delayed dispersal of adult offspring may occur in species of both social types, but these traits are much more pronounced in species of the complex social type (see Table 2). The development of several offspring generations in succession requires rather long-lived substrate conditions. Thus, whereas species of the simple social type can settle in dead branches or tree trunks of small diameters, species of the complex social type appear to settle typically in large-diameter trunks of recently dead trees or even in living trees (Schedl 1962; Wood 1982). Nevertheless, in dead wood it is probably unpredictable whether successive offspring generations will be able to use the same nest, because of the exploitation of the substrate by a plethora of competing microorganisms. Selection will therefore favour totipotent individuals that may either disperse and breed independently, or stay and help, or stay and co-breed with relatives (e.g. X. saxesenii: Peer and Taborsky (2007) and Biedermann and Taborsky (2011); X. affinis: Biedermann et al. 2011). Eusociality, classified by a permanently sterile and a reproductive caste, will only evolve under stable resource conditions in living trees, like found for instance in A. incompertus (Kent and Simpson 1992). Hence, the durability 87

Chapter 6 of the substrate may be the major selective force deciding about the evolution of ambrosia beetle social complexity.

The ambrosia beetle tribe Xyleborini (Scolytinae) radiated about 20 million years ago and stands for one of eight independent origins of fungus agriculture in the curculionid beetles (Farrell et al. 2001; Hulcr et al. 2007), which is probably closely associated with the evolution of division of labour and advanced sociality (Kirkendall et al. 1997; Biedermann and Taborsky 2011). This contrasts with the other fungus culturing insect groups, attine ants and macrotermitine termites, that lived in eusocial groups already when fungiculture evolved (Wilson 1971), which makes ambrosia beetles a unique model system to study the evolution of sociality in relation to fungiculture (Mueller et al. 2005). The evolutionary history of sociality in ambrosia beetles becomes even more interesting when considering that they vary in their mating patterns (inbreeding vs. outbreeding species) and ploidy level (haplodiploid vs. diploid species; see Table 2), which are factors proposed to contribute to social evolution (Choe and Crespi 1997; Bourke 2011), but their respective roles are controversial (West et al. 2007; Bourke 2011).

Acknowledgements

We thank Stacy Blomquist and Eric Ott for their help in collecting X. affinis in the field and starting the first laboratory galleries. This study was performed at the Institute of Ecology and Evolution, University of Berne, Switzerland, and at the Southern Research Station in Pineville. This work benefited greatly from help and discussions with Kier Klepzig and commentaries on the manuscript by Tabea Turrini, Henrik de Fine Licht, Florian Menzel and Markus Zöttl. Studies in Pineville, LA, were supported by a cooperative agreement with the Southern Research Station, USDA Forest Service. PHWB received a DOC-fellowship from the Austrian Academy of Sciences and a Roche Research Fellowship at the Department of Behavioural Ecology, University of Berne, Switzerland.

Literature

Batra LR (1966) Ambrosia fungi: extent of specifity to ambrosia beetles. Science 153: 193-195.

Batra LR & Michie MD (1963) Pleomorphism in some ambrosia and related fungi. Transactions of the Kansas Academy of Science 66: 470-481.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Biedermann PHW (2007) Towards eusociality in ambrosia beetles. Masters thesis, University of Berne, Switzerland.

Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus saxesenii Ratzeburg (Scolytinae, Coleoptera). Zookeys 56: 253-267.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: Behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Chapter 6

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

Borden JH (1988) The striped ambrosia beetle. Dynamics of Forest Insect Populations. (Berryman AA, ed), pp. 579-596. Plenum, New York.

Bot ANM, Currie CR & Boomsma JJ (2001) Waste management in leaf-cutting ants. Ethology Ecology & Evolution 13: 225-237.

Bourke AFG (2011) Principles of Social Evolution. Oxford University Press, Oxford.

Bright DE (1973) The bark and ambrosia beetles of California, Coleoptera: Scolytidae and Platypodidae. University of California Press, Berkeley, CA.

Brock DA, Douglas TE, Queller DC & Strassmann JE (2011) Primitive agriculture in a social amoeba. Nature 469: 393-396.

Cardoza YJ, Klepzig KD & Raffa KF (2006) Bacteria in oral secretions of an endophytic insect inhibit antagonistic fungi. Ecological Entomology 31: 636-645.

Choe JC & Crespi BJ (1997) The Evolution of Social Behaviour in Insects and Arachnids. Cambridge University Press, Cambridge UK.

Dahlsten DL (1982) Relationship between bark beetles and their natural enemies. Bark Beetles in North American Conifers (Mitton & Sturgeon, eds), pp. 140-182. University of Texas Press, Austin.

De Fine Licht HH & Biedermann PHW (2012) Patterns of functional enzyme activity suggest that larvae are the key to successful fungus farming by ambrosia beetles. Frontiers in Zoology in review.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Fischer M (1954) Untersuchungen über den kleinen Holzbohrer (Xyleborus saxeseni). Pflanzenschutzberichte 12: 137-180.

Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z. Morph. u. Ökol. Tiere 45: 275-308.

Francke-Grosmann H (1963) Some new aspects in forest entomology. Annual Review of Entomology 8: 415-438.

Francke-Grosmann H (1967) Ectosymbiosis in wood-inhabiting beetles. Symbiosis (Henry SM, ed), pp. 141-205. Academic Press, New York.

Francke-Grosmann H (1975) Zur epizoischen und endozoischen Übertragung der symbiotischen Pilze des Ambrosiakäfers Xyleborus saxeseni (Coleoptera: Scolitidae). Entomologica Germanica 1: 279-292.

French JRJ & Roeper RA (1973) Patterns of nitrogen utilization between the ambrosia beetle Xyleborus dispar and its symbiotic fungus. Journal of Insect Physiology 19: 593-605.

Haack RA & Slansky F (1987) Nutritional ecology of wood feeding Coleoptera,Lepidoptera and Hymenoptera. Nutritional Ecology of Insects, Mites and Spiders (Slansky F & Rodriguez JG, eds), pp. 449-486. Wiley, New York. 89

Chapter 6

Harrington TC, Aghayeva DN & Fraedrich SW (2010) New combinations in Raffaelea, Ambrosiella, and Hyalorhinocladiella, and four new species from the redbay ambrosia beetle, Xyleborus glabratus. Mycotaxon 111: 337-361.

Hart A, Anderson C & Ratnieks F (2002) Task partitioning in leafcutting ants. Acta ethologica 5: 1-11.

Hölldobler B & Wilson EO (1990) The ants. The Belknap Press of Harvard University Press, Cambridge, MA.

Hosking GB (1972) Xyleborus saxeseni, its life-history and flight behaviour in New Zealand. N. Z. J. Forest Science 3: 37-53.

Hubbard HG (1897) The ambrosia beetles of the United States. Some Miscellaneous Results of the Work of the Division of Entomology (Howard LO, ed), pp. 9-13. U.S. Dept. of Agriculture Bureau of Entomology Bull. No. 7.

Hulcr J, Kolarik M & Kirkendall LR (2007) A new record of fungus-beetle symbiosis in Scolytodes bark beetles (Scolytinae, Curculionidae, Coleoptera). Symbiosis 43: 151-159.

Kajimura H & Hijii N (1992) Dymamics of the fungal symbionts in the gallery system and the mycangia of the ambrosia beetle, Xylosandrus mutilatus (BLANDFORD) (Coleoptera, Scolytidae). Ecological Research 7: 107-117.

Kent DS & Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften 79: 86-87.

Kingsolver JG & Norris DM (1977) The interaction of Xyleborus ferrugineus (Fabr.) (Coleoptera: Scolytidae) behavior and initial reproduction in relation to its symbiotic fungi. Annals of the Entomological Society of America 70: 1-4.

Kok LT & Norris DM (1972) Symbiotic interrelationships between microbes and ambrosia beetles 6. Aminoacid composition of ectosymbiotic fungi of Xyleborus ferrugineus (Coleoptera, Scolytidae). Annals of the Entomological Society of America 65: 598-602.

Kokko H & Ekman J (2002) Delayed dispersal as a route to breeding: territorial inheritance, safe havens, and ecological constraints. Am. Nat. 160: 468-484.

Korb J (2008) The ecology of social evolution in termites. Ecology of Social Evolution (Korb J & Heinze J, eds), pp. 151-174. Springer, Berlin, Heidelberg.

Leach JG, Hodson AC, Chilton SJP & Christensen CM (1940) Observations on two ambrosia beetles and their associated fungi. Phytopathology 30: 227-236.

Lengerken H (1939) Die Brutfürsorge- und Brutpflegeinstinkte der Käfer. Akademische Verlagsgesellschaft m.b.H., Leipzig.

Merkl O & Tusnadi CsK (1992) First introduction of Xyleborus affinis (Coleoptera: Scolytidae), a pest of Dracaena fragrans 'Massangeana', to Hungary. Folia Entom. Hung. 52: 67-72.

Milne DH & Giese RL (1970) Biology of the Columbian Timber Beetle, Corthylus columbianus (Coleoptera: Scolytidae). 10. Comparison f yearly mortality and dispersal losses with population densities. Entomological News 81: 12-24.

Mueller UG, Gerardo NM, Aanen DK, Six DL & Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563-595.

Chapter 6

Norris DM (1975) Chemical interdependence among Xyleborus spp. ambrosia beetles and their symbiotic microbes. Material und Organismen 3: 479-788.

Norris DM (1993) Xyleborus ambrosia beetles - a symbiotic ideal extreme biofacies with evolved polyphagous privileges at monophagous prices. Symbiosis 14: 229-236.

Peer K & Taborsky M (2004) Female ambrosia beetles adjust their offspring sex ratio according to outbreeding opportunities for their sons. J. Evol. Biol. 17: 257-264.

Peer K & Taborsky M (2005) Outbreeding depression, but no inbreeding depression in haplodiploid ambrosia beetles with regular sibling mating. Evolution 59: 317-323.

Peer K & Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. Behav. Ecol. Sociobiol. 61: 729-739.

R Development Core Team (2008). R: A language and environment for statistical computing. Vienna, Austria, R Foundation for Statistical Computing.

Ranger CM, Reding ME, Persad AB & Herms DA (2010) Ability of stress-related volatiles to attract and induce attacks by Xylosandrus germanus and other ambrosia beetles. Agricultural and Forest Entomology 12: 177-185.

Roeper R, Treeful LM, O'Brien KM, Foote RA & Bunce MA (1980) Life history of the ambrosia beetle Xyleborus affinis (Coleoptera: Scolytidae) from in vitro culture. Great Lakes Entomologist 13: 141-144.

Roeper RA (1995) Patterns of mycetophagy in Michigan ambrosia beetles. Michigan Academian 27: 153-161.

Roeper RA & French JRJ (1981) Ambrosia fungi of the Western United States and Canada - beetle assocaitions (Coleoptera: Scolytidae), Tree Hosts, and Distribution. Northwest Science 55: 305-309.

Schedl KE (1956) Breeding habits of arboricole insects in Central Africa. Proceedings 10th International Congress of Entomology, Montreal, pp. 183-197.

Schedl KE (1962) Scolytidae und Platypodidae Afrikas, II. Revista de Entomologia de Mocamique 5: 1-594.

Schneider I (1987) Distribution, fungus-fransfer and gallery construction of the ambrosia beetle Xyleborus affinis in comparison with X. mascarensis (Coleoptera, Scolytidae). Entomologia Generalis 12: 267-275.

Schneider-Orelli O (1913) Untersuchungen über den pilzzüchtenden Obstbaumborkenkäfer Xyleborus (Anisandrus) dispar und seinen Nährpilz. Zentralblatt für Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene II 38: 25-110.

Scott JJ, Oh DC, Yuceer MC, Klepzig KD, Clardy J & Currie CR (2008) Bacterial protection of beetle-fungus mutualism. Science 322: 63.

Selander RK (1964) Speciation in Wrens of the Genus Campylorhynchus. University of California Publications in Zoology 74: 1-305.

Stacey PB & Ligon JD (1987) Territory Quality and Dispersal Options in the Acorn Woodpecker, and A Challenge to the Habitat-Saturation Model of Cooperative Breeding. Am. Nat. 130: 654-676.

Traniello JFA & Leuthold RH (2000) The behavioral ecology of foraging in termites. Termites: Evolution, Sociality, Symbiosis, Ecology. (Abe T, Bignell DE & Higashi M, eds), pp. 141-168. Kluwer, Dordrecht.

Verrall AF (1943) Fungi associated with certain ambrosia beetles. Journal of Agricultural Research 66: 135-144.

Watanabe H & Tokuda G (2009) Cellulolytic Systems in Insects. Annu. Rev. Entomol. 55: 609-632.

Chapter 6

West SA, Griffin AS & Gardner A (2007) Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. J. Evol. Biol. 20: 415-432.

Wilson EO (1971) The Insect Societies. Belknap Press, Cambridge.

Wood SL (1982) The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Naturalist Memoirs 6: 1-1359.

Graphs & Tables

Table 1. Behaviours of larval and adult xyleborine ambrosia beetles likely resulting in mutual benefits for other group members in the natal gallery (modified and extended from Biedermann & Taborsky, 2011).

Class Behaviour Definition Mutual X. X. benefit saxesenii affinis

Larvae enlarging the brood chamber by feeding on fungus- gallery Digging x infested substrate extension

Cropping grazing on the fungal layer on the gallery walls fungus care (?) x forming balls of frass and sawdust by repeated ventral Balling hygiene x body contractions grooming an egg, larva, pupa, or adult beetle with the brood care, Allogrooming x x mouthparts hygiene

Adults staying inactive in the entrance tunnel and plugging it gallery Blocking x x with the body (abdomen directed to the outside) protection

gallery Digging excavating new tunnels x x extension grazing on the fungal layer on the gallery walls with the Cropping fungus care x x maxillae and/or mandibles Shuffling moving frass and sawdust with the legs and elytra hygiene x x grooming an egg, larva, pupa, or adult beetle with the brood care, Allogrooming x x mouthparts (i.e., maxillae, labium) hygiene

Chapter 6 Table 2. Overview of the gallery architecture and modes of fungiculture and social life of representative ambrosia beetle genera.

Family Curculionidae Lymexylidae

Tribe Xyleborini Xyloterini Cryphalini Platypodini

Genus Xylosandrus Xyleborinus Anisandrus Xyleborus Trypodendron Gnathotrichus Platypus Hylocoetus Gallery architectu re Branching Branching Branching Type of Branching Branching Cave-tunnel Cave-tunnel tunnels with tunnels with tunnels with Single tunnel gallery tunnels tunnels larval cradles larval cradles larval cradles Structure Simple1 Complex2 Simple1 Complex2 Simple1 Complex2 Complex2 na Reference A B C D E F F Not shown in Fig. 5 Haplodiploi Haplodiploi Ploidy Haplodiploid Haplodiploid Diploid Diploid Diploid Diploid d d Breeding Inbreeding Inbreeding Inbreeding Inbreeding Outbreeding Outbreeding Outbreeding Outbreeding structure Mode of feeding Xylomycetopha Xylomycetopha Mycetopha Mycetopha Xylomycetopha Xylomycetopha Xylomycetopha Xylomycetopha Larvae gy3 (?) gy3 gy4 gy4 gy3 gy3 gy3 gy3 Mycetopha Mycetopha Xylomycetopha Adults Mycetophagy4 Mycetophagy4 Mycetophagy4 Mycetophagy4 No feeding gy4 gy4 gy3 (?) Social structure Communa Yes Yes Yes Yes Yes Yes Yes No l nesting Location Tips of Tips of Larval cradles, Brood chamber Brood chamber Larval cradles Larval cradles na of brood tunnels (?) tunnels in tunnels Separatio n of larvae No No Partly Partly Yes Yes Partly na and adults Number of generatio 1 Several 1 Several 1 Several (?) Several na ns per nest Number of 1 Several 1 Several 1 Several (?) Several (?) na breeding females Social Sub-social, Primitive Sub-social, organizati Sub-social Sub-social primitive Sub-social Sub-social Solitary eusocial eusocial on eusocial Division of labour GE(?), BC, Larvae GE, BC, GH BC(?), GH(?) BC, GH - ? GE, BC(?), GH na GH(?) Adult GP, BC, GH, GE, GP, BC, GE, GP, BC, GE, GP, GH, GE, GP, GH, GE, GP, BC, GH, GP, BC, GH, FC na females FC(?) GH, FC GH, FC FC(?) FC(?) FC Adult I I I I GP, GH GP, GH GP, GH na males

1 Simple gallery structure: tunnels branching usually not branching more than two times; 2 Complex gallery: tunnel branching several times; 3 Xylomycetophagy: eating xylem and fungal biomass; 4 Mycetophagy: eating fungal fruiting structures. Abbreviations: BC – brood care, FC – fungus care, GE – gallery extension, GP – gallery protection, GH – gallery hygiene, I – insemination.

Chapter 6

Fig.1. Illustrations of the morphologies of Xyleborinus saxesenii galleries of three size classes in beech (Fagus sylvatica). Drawings made from original, overwintering galleries (N = 18) dissected from a single trunk in late fall (9th November 2008, in Bern/CH, 560 m asl, 46°95’, 7°31’). Numbers of inhabitants are given in the lower left hand corner of each figure (immat – immatures (larvae, pupae), ♀ - adult females, ♂ - adult males; for more details see Table 2 in the Appendix section). The entrance tunnel is always directed to the side of the trunk, except in figures (O) and (P) where beetles bored into the trunk from the top. Brood chambers of X. saxesenii are flat with a height of ~1mm and mostly span in one or two planes in the direction of the wood grains; chambers diverging out of the pictured plane are indicated by white dashed lines in figures (F) and (G). Gallery parts that could not be fully traced are indicated by black dashed lines. Figure (R) is modified from a photo (Biedermann et al. 2009) to illustrate larval and adult inhabitants (the right side could not be traced). The size of an adult female is given on top of the scale. Original size.

Chapter 6

Fig.2. Illustrations of the morphology of laboratory and field Xyleborus affinis galleries. (A) Photo of a laboratory gallery in artificial medium. (B-I) Field galleries adapted from drawings of Schedl (1962; pp. 362-365;N=8). Tunnels have a diameter of ~1mm and extend in different directions, independent of the wood grain, either in xylem or phloem. There is no communal brood chamber, but brood is found in tips of freshly bored tunnels as indicated in figures (A) and (C). It is unclear whether galleries with two entrances (arrows in I) result from two merging galleries or the construction of a second exit tunnel. The size of an adult female is given on top of the scale. Original size.

Chapter 6

Fig.3. Species comparison for cooperative (probably mutually beneficial) behaviours of the 2nd and 3rd larval instars. § - Digging and Cropping denote the different feeding modes of Xyleborus affinis and Xyleborinus saxesenii larvae (see text). The bars show the mean (± se) proportion of time. Statistically significant differences between the species are given (*** - p < 0.001, ** - p < 0.01; GEE, for details see Table 3 in the Appendix).

Chapter 6

Fig.4. Species comparison for cooperative (probably mutually beneficial) behaviours of adult females. The bars show the mean (± se) proportion of time of Xyleborus affinis and Xyleborinus saxesenii larvae. Statistically significant differences between the species are given (*** - p < 0.001, * - p < 0.05; GEE, for details see Table 3 in the Appendix).

Chapter 6

Fig.5. Types of ambrosia beetle galleries and important representative genera. Black tunnels are excavated by adult beetles, whereas pink areas are excavated by larvae (only tunnels in F are excavated by both adults and larvae). Cooperative stands for division of labour between adults and larvae. Subsocial stands for parental care without division of labour between adults and larvae. For details on the biology of the groups see Table 2.

Chapter 6 Supplementary material

Table S1. Xyleborinus saxesenii gallery sizes and compositions in winter. For schemes of gallery morphology see Fig.1. Sizes in brackets denote galleries that could not be fully measured.

nd rd st 2 & 3 Teneral Adult 2 Gallery 1 instars Pupae ♂♂ Σ Size [mm ] instars ♀♀ ♀♀ A 0 (86.2) B 2 1 2 5 (75.2) C 1 3 4 8 93 D 1 2 4 7 126 E 3 1 1 5 26.6 F 6 2 8 (232) G 2 2 (77.2) H 1 1 50.8 I 1 6 7 47.7 J 1 11 12 107.2 K 4 10 14 (89.7) L 3 4 3 11 1 22 137.3 M 3 2 14 2 21 (114) N 21 1 22 107.4 O 8 32 1 8 10 7 66 232.7 P 2 24 7 16 6 55 172 Q 5 32 1 7 29 1 75 (286.4) R 12 69 1 5 87 (209) Mean  se 2.1  0.8 9.9 4.3 0.3 0.3 1.3 0.7 8.4 1.8 1.2 0.5 23.2 6.5 126.13 17

Chapter 6 Table S2. Behavioural differences (p < 0.05) between the proportions of time the two study species spent with the observed cooperative behaviours (for graphical illustration see Fig.3. and Fig.4.). GEEs with an exchangeable correlation structure of the response variable within a cluster (= gallery- identity) were used to identify effects of the species on the total amount of time each behaviour was observed. Model coefficients are reported as coeff.  se (standard error of the estimate), with Xyleborus affinis in the first row of the model as the reference category (coefficient set to zero). A positive contrast denotes that the mean of Xyleborinus saxesenii is higher than the mean of X. affinis; a negative contrast denotes the reverse.

Class Behaviour Model coeff.  se z p

2nd &3rd digging Only in X. saxesenii instar larvae cropping Only in X. affinis

cropping vs. Intercept (X. affinis) -0.39 0.2 -1.9 0.06 digging Contrast X. affinis vs. X. saxesenii 0.84 0.31 2.7 0.007

balling Only in X. saxesenii

allogrooming Intercept (X. affinis) -2.12 0.21 -10.3 <0.001 Contrast X. affinis vs. X. saxesenii -1.67 0.45 -3.74 <0.001 adult ♀♀ blocking Intercept (X. affinis) -2.91 0.21 -13.9 <0.001 Contrast X. affinis vs. X. saxesenii -0.17 0.63 -0.27 0.79

digging Intercept (X. affinis) -5.28 0.68 -7.82 <0.001 Contrast X. affinis vs. X. saxesenii 0.68 1.45 0.47 0.64

cropping Intercept (X. affinis) -2.36 0.23 -10.5 <0.001 Contrast X. affinis vs. X. saxesenii 1.04 0.44 2.37 0.018

shuffling Intercept (X. affinis) -0.07 0.09 -0.79 0.43 Contrast X. affinis vs. X. saxesenii -2.26 0.44 -5.17 <0.001

allogrooming Intercept (X. affinis) -4.04 0.36 -11.3 <0.001 Contrast X. affinis vs. X. saxesenii 2.03 0.52 3.91 <0.001

100

Chapter 7

1 Dynamics of filamentous fungi in the complex ambrosia gardens of the 2 primitively eusocial beetle Xyleborinus saxesenii Ratzeburg 3 (Scolytinae: Curculionidae)

Peter H.W. Biedermann1,2, Kier D. Klepzig2,3, Michael Taborsky1 and Diana L. Six4

1Department of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland 2USDA Forest Service, Southern Research Station, 2500 Shreveport Hwy, Pineville, LA 71360, USA 3US Forest Service, Southern Research Station, 200 WT Weaver Blvd, Asheville, NC 28804, USA 4Department of Ecosystem and Conservation Sciences, College of Forestry and Conservation, The University of Montana, Missoula, MT 59812, USA

Corresponding author: Peter H.W. Biedermann Baltzerstrasse 6, CH-3012 Bern, Switzerland phone: 0041 31 631 3015 e-mail: [email protected]

Species identity for Raffaelea-like sp. B and Artrinium sp. are not yet confirmed by sequencing! – These species have been marked in grey throughout the text.

Abstract Insect fungus gardens consist usually of a community of interacting microbes with both beneficial and detrimental effects to the farmers. In contrast to fungus-farming ants and termites the fungal communities of ambrosia beetles and the effects of particular fungal species on the farmers are largely unknown. Here we used a laboratory breeding technique for studying the filamentous fungal garden community of the xyleborine ambrosia beetle, Xyleborinus saxesenii Ratzeburg (Scolytinae: Curculionidae), which cultures fungi in tunnel systems excavated within recently dead trees. Raffaelea sulphurea and a Fusicolla acetilerea (Tubaki, C. Booth & T. Harada) (formerly Fusarium merismoides var. acetilereum) -like fungus were transmitted in spore-carrying organs (mycetangia) by gallery founding females and established first in new gardens. Abundance of R. sulphurea had positive effects on the egg-laying of the female and on larval numbers. Over time, four other fungal species emerged in the gardens, of which two, Paecilomyces variotii and Penicillium decaturense, became increasingly abundant. P. variotii had a negative impact on larval numbers and led to the death of parental females by forming biofilms on their bodies. It also comprised the main portion of garden material removed from galleries by adults. Our data suggests that two mutualistic, several commensalistic and one to two pathogenic filamentous fungi are associated with X. saxesenii. Fungal diversity in gardens of ambrosia beetles appears at least ten times lower than in gardens of fungus culturing ants.

Keywords: insect-fungus agriculture, cooperative breeding, mycetangium, mycangium, ambrosia fungus gardens, mutualism, social behaviour, commensalism, mycophagy, eusociality, subsocial

Introduction Mycophagy by insects has evolved in several lineages including springtails, flies, moths, wood wasps, termites, ants and beetles (Wheeler and Blackwell 1984; Martin 1987; Wilding et al. 1989). Among these, only attine ants, macrotermitine termites and curculionid ambrosia beetles evolved advanced fungus agriculture (Mueller et al. 2005). This involves (1) obligate nutritional dependence 101

Chapter 7 on fungal food for adults and their brood, (2) translocation of their fungal crops by spore- or propagule-carrying organs within nests and when founding new nests, and (3) cultivation and management of the fungal crops (i.e., continuous monitoring, sustainable management and protection, weeding and biological control of alien microbes). As the latter can be easier managed by a group of individuals partitioning the labour, advanced fungus agriculture is often associated either with a subsocial (most ambrosia beetles) or eusocial life strategy (all farming ants and termites, one ambrosia beetle; Mueller et al., 2005). Fungus agriculture has been well studied in the fungus-farming ants and termites, but is yet little understood in ambrosia beetles. These beetles dwell in the wood of (usually) recently dead or weakened trees, where they construct tunnel systems (galleries) upon the walls of which they nurture ambrosia gardens. Ambrosia gardens consist of fungi the beetles carry into trees in their intestines or, more commonly, in specialized structures called mycetangia or mycangia (Francke- Grosmann 1956; 1975). This vertical transmission of ambrosia fungi from the beetle`s natal galleries to newly founded nests is associated with the co-evolution between fungi and beetles and often a species-specific association between partners (Six 2003). For many species, healthy fungus gardens are dominated by mutualistic ambrosia fungi of the genera Raffaelea and Ambrosiella (Ascomycota; Harrington et al. 2010). These species usually show an ambrosial- or kohlrabi-like growth form within the gardens, forming nutrient-rich fruiting structures (e.g. conidiospores) that are grazed by adult beetles and their offspring. Gardens also contain a complex of other filamentous fungi, yeasts and bacteria (e.g.Haanstad and Norris 1985), which are often transmitted by spores sticking to the body of founding females (Francke-Grosmann 1967). The effects of these microbes are largely unknown. The associated fungi of only three of about 3500 ambrosia beetles worldwide have yet been investigated (Kajimura and Hijii 1992; Harrington and Fraedrich 2010; Endoh et al. 2011), and the fungal dynamics in mycangia and galleries have been studied only in one ambrosia beetle, Xylosandrus mutilatus (Kajimura and Hijii 1992). The interaction between ambrosia beetles and their associates has not yet been studied, although both adults and larvae can probably alter the fungal community of their ambrosia gardens, as has been observed in fungus farming ants (Mueller et al. 2001). Adult ambrosia beetles are particularly attracted to their primary ambrosia fungus and repelled by fungal pathogens (Hulcr et al. 2011). An observational study of ambrosia beetle behaviours within their nests revealed that adults and their larvae closely interact with their gardens (Biedermann and Taborsky 2011). Larvae were observed to cooperate in various duties, which is exceptional for holometabolous insects,; they enlarge the gallery by digging and thereby create space for the fungi to spread, they fertilize fungi with their excretions, clean colony members and gallery walls which prevents the spreading of mold, and participate in the removal of waste from the tunnel system (Biedermann and Taborsky 2011). Adults were observed to block tunnels, thus potentially regulating the microclimate of the gardens (Kirkendall et al. 1997), and to graze their gardens, which affects fungal growth (French and Roeper 1972) and potentially also species composition (Biedermann and Taborsky 2011). Ambrosia fungi have been shown to dominate gardens in recently founded galleries and newly built tunnel systems, and decrease in abundance relative to invading weed fungi at the end of gallery life, when beetles leave the nest, and in old parts of the tunnel system (e.g. Kajimura and Hijii 1992). The abundance of certain fungi may positively or negatively affect the brood. For example, fungal associates of Xyleborus ferrugineus (Fabricius) vary considerably in sterol, lipid and amino acid content, and thus in their nutritional quality for the developing brood (Kok and Norris 1972a; Kok and Norris 1972b; Kok and Norris 1973). Here we report a comprehensive survey of the filamentous fungi closely associated with the ambrosia beetle Xyleborinus saxesenii Ratzeburg, and their dynamics in relation to the life history of this species. In addition we report which fungal associates of X. saxesenii are carried in the mycetangia and guts of females during their dispersal flight. Based on previous studies we expected to find Raffaelea sulphurea (L.R. Batra) T.C. Harr. (syn. Ambrosiella sulfurea; (Harrington et al. 2010) as the primary symbiont (Francke-Grosmann 1956; 1975; Batra 1967; Roeper et al. 1980; Roeper and French 1981), but hitherto the identity of other fungal associates has been unknown. To follow the dynamics of all fungi within the beetle gardens and their effect on the brood, we sampled garden

102

Chapter 7 material over the entire developmental period of a brood and recorded brood numbers and offspring development. This was possible with help of a laboratory rearing technique allowing to observe the beetles within their galleries (Biedermann et al. 2009). Furthermore, we identified (i) fungi removed from the galleries by adults (to the dumps; i.e. the material disposed out of the entrance tunnel), and (ii) detrimental fungi growing on the bodies of beetles. Finally, we discuss our results in comparison to fungal communities found within gardens of fungus-growing ants.

Materials & Methods

Study species Xyleborinus saxesenii is one of the most common ambrosia beetles in temperate zones worldwide. Originally native to Eurasia, over the last 200 years it has been introduced into parts of Africa, Oceania, as well as South and North America (for the actual distribution see http://xyleborini.tamu.edu/public/site/scolytinae/home). The species is still spreading, facilitated by the shipment of woody products around the world, and by characteristics of its own biology. X. saxesenii shows little host tree preference and probably lacks on inbreeding depression because of a mating system in which sib-mating is the rule between haploid brothers and diploid sisters in their natal nest (as in Xylosandrus germanus; Peer and Taborsky 2005). Therefore, the translocation of a single female may be sufficient for the successful establishment of a new population without negative effects of a genetic bottle-neck. Galleries of X. saxesenii are founded by single females which dig a vertical entrance tunnel of a few centimetres into a tree trunk. They inoculate gallery walls with fungi, lay eggs when fungal gardens have established, and later care for the developing brood. The larvae feed on fungus-infested wood and in this way gradually enlarge the tunnel to a flat brood chamber (Roeper 1995). This xylomycetophagous feeding is typical for larvae in the genus Xyleborinus and likely serves to reduce kin competition, as it increases the space for ambrosia gardens to grow and improves the breakdown of wood by enzymes (De Fine Licht and Biedermann 2012). Wood passes the guts of larvae without being digested, but finely grounded and readily utilized by the fungi. Such woody frass is partly spread on the ambrosia garden microbes, which probably recycle and fully breakdown this material (Biedermann and Taborsky 2011). Overlapping generations are typical in X. saxesenii nests, because adult females delay dispersal after maturation and fertilization by a brother (Peer and Taborsky 2007; Biedermann et al. 2011). During this time they engage in brood and fungus care, thereby increasing gallery productivity (Biedermann and Taborsky 2011). Additionally, fungal gardens benefit from the recycling of their excretions (Abrahamson and Norris 1970). About 20 % of daughters also reproduce in their natal nest (Biedermann 2007; Biedermann et al. 2011).

Beetle collection and laboratory breeding About 100 X. saxesenii females were caught alive with Lindgren funnel traps baited with ethanol in Pineville, LA, USA (38m asl; 31°20’, 92°24’) in Summer 2007. Collection cups were filled with damp sterile filter paper and emptied twice daily to avoid microbial contaminations of the beetles (Benjamin et al. 2004). In the lab we surface-sterilized the beetles by rinsing them twice for a few seconds, first with 70% ethanol and afterwards with deionised water. This treatment does not harm the fungal spores of the cultivar within the mycetangium, but should reduce fungal contamination by eliminating some of the spore-load sticking to the body surface of the beetles (e.g. molds). It is necessary for laboratory breeding of ambrosia beetles, because these “contaminants” establish relatively easily in our homogenic artificial media than under natural conditions (Biedermann et al. 2009), where beetles largely surface-sterilize themselves boring through bark rich in fungitoxins and other antibiotics substances (Berryman 1989). Surface-sterilization – both in the laboratory and in the field – is likely incomplete, however, because specific contaminants take-over old galleries (Kajimura and Hijii 1992), which must have been transmitted initially as sticky spores in pits of the exoskeleton (Bridges et al. 1984). Thirteen females were used for fungal isolations (see 103

Chapter 7 below). The remaining females were placed singly on an agar-sawdust-based rearing medium in glass tubes (for details on this technique and ingredients of the modified medium we used see (Biedermann et al. 2009). Tubes were closed with plastic caps, stored vertically, and wrapped in paper in a way that allowed light to penetrate the tube only from the top. This way beetles frequently bored tunnels next to walls of the glass tube, allowing observations of brood development when the paper was removed (Biedermann 2007). Tubes were kept at 23°C.

Fungus isolations from adult females captured during dispersal flight After surface-sterilization, we aseptically dissected mycetangia from 13 adult females using fine tweezers under a microscope (6.4× – 40× magnification). The mycetangium in X. saxesenii is a paired cavity at the basis of the females’ elytra; for isolating the spores present in the mycetangia, we removed the two elytra and placed their bases on malt agar (MA: 25 g malt extract, 20 g agar, 1 l deionized H2O) plates. Elytral mycetangia were too small to be dissected successfully, so we cannot exclude that our isolations also contain fungi sticking to the upper and the bottom side of the elytra. Guts of the beetles were dissected and squashed in a sterile Petri dish. Gut material was then spread across the surface of MA plates using a sterile metal loop. All cultures were then incubated at 25°C in darkness for about two weeks, and purified by subculturing.

Fungal isolations from laboratory galleries After introduction into tubes, females usually bored an entrance tunnel and inoculated the medium with fungi. In previous lab studies, it was observed that ambrosia beetle females first feed on the fungus and then lay eggs (e.g. Kingsolver and Norris 1977) for about 40 days. During that time, four periods of gallery development can be discerned: (1) three to five days with only the foundress and eggs present, (2) at least ten days with foundress, eggs and immatures (larvae and pupae) present, (3) about 40 days with foundress, eggs, immatures and adult offspring present, and (4) the nest-leaving phase when the foundress has died, offspring have matured and gradually disperse (Biedermann et al. 2011). We timed our sampling of brood and fungi in accordance with this gallery development pattern: we dissected eight galleries in period 1, ten galleries in period 2, and nine galleries in period 4. We did no dissections during period 3, because we expected only gradual changes in between fungal frequencies found in period 2 and period 4, and thus decided to not reduce the sample sizes in the other groups any further. Additionally, we dissected eight galleries that did not produce brood within one month post-introduction of the female, and nine galleries where all larvae died. From each of the galleries, we took eight samples from the gallery wall of the entrance tunnel (= oldest part of gallery) and eight samples from the gallery wall of the brood chamber (where most inhabitants were present) using a sterile needle. Four of these samples we placed on MA and four on cycloheximide-streptomycin malt agar (CSMA: 10 g malt extract, 15 g agar, 20 ml filter sterilized CSMA stock solution containing 2 mg of Cycloheximide and 1 mg Streptomycin,

1 L deionized H2O) plates. Samples were placed in the middle of the plates.

Additional isolations Single, live and healthy larvae from five different galleries were squashed in sterile Petri dishes and then plated on MA and CSMA agar using sterile metal loops. Four live adult females were found to be overgrown by a biofilm of fungi, which we isolated by scrapping parts off with a sterile needle and plating four samples from each insect on MA. The fungal composition of eight gallery dumps (= frass and sawdust shuffled out of the nest onto the surface of the medium) was analysed by plating four samples of each on MA and CSMA.

Identification of filamentous fungus isolates Isolated fungi were initially placed into groups based on cultural colony characteristics (i.e. morphology and color of mycelium and fruiting structures). Representative samples were used for DNA sequencing. To extract DNA, a small amount of mycelium and conidia was scraped from the surface of young, relatively unmelanized colonies growing on MA, or hyphae were taken from cultures grown in 2% malt extract broth (MEB). The mycelium was macerated in 200 µl PrepMan 104

Chapter 7

Ultra (Applied Biosystems, USA), incubated at 95°C for 10 min, and then centrifuged. The supernatant containing DNA was then used for PCR amplification of a portion of the ribosomal RNA encoding region and partial b-tubulin gene with the primer pairs ITS3 (White et al. 1990) and LR3 (Vilgalys and Hester 1990), and Bt2b (Glass and Donaldson 1995) and T10 (O'Donnell and Cigelnik 1997). PCR conditions used have been described by (Six et al. 2009). Amplicons were purified using a High Pure PCR Product Purification Kit (Roche, Germany) and sequencing was performed on an ABI 3130 automated sequencer (Perkin–Elmer Inc, USA) at the Murdock Sequencing Facility (University of Montana, Missoula, MT USA). DNA sequences of representative isolates were deposited in GenBank (Table 1). Contigs of forward and reverse sequences obtained with each primer pair were aligned in MEGA4 (Tamura et al. 2007). BLAST searches were done with sequences of each isolate in the NCBI GenBank database (http://www.ncbi.nlm.nih.gov). Cultures of isolates from this study were deposited in the culture collection of Diana Six (DLS) at the University of Montana, Missoula, MT, USA, and the Centraalbureau voor Schimmelcultures (CBS), Utrecht, The Netherlands (Table 1).

Statistical analyses Using all 16 samples taken from each laboratory gallery we estimated the frequency of each fungal species. Additionally, we determined whether a fungus was present or not in the eight samples taken from the main tunnel and the brood chamber of each gallery. For each fungal species we analysed how its frequency and presence (dependent variables) were affected by the culture medium (MA, CSMA), the location within the gallery (brood chamber, main tunnel) and the period of offspring development (foundress with eggs, foundress with larvae, only adult progeny; fixed factors) by controlling for gallery of origin (random factor). Using data from the first two periods only (foundress with eggs, foundress with larvae) we also tested if numbers of eggs and larvae correlated with fungal frequencies. All analyses were done using generalized estimating equations (GEEs) in R (lmer; Version 2.12.1; (R Development Core Team 2008). GEEs are an extension of generalized linear models with an exchangeable correlation structure of the response variable within a cluster, which allows for controlling the variation between observations from a single gallery. This was necessary because of variation in sample sizes between galleries, as plates had to be excluded from the analyses if they did not yield any microbial growth.

Results

Our morphological data in combination with DNA sequencing revealed that six species of filamentous fungi were associated with X. saxesenii (Table 1). As expected, Raffaelea sulphurea (formerly Ambrosiella sulfurea) was regularly isolated from X. saxesenii bodies and galleries. Both -tubulin) generated for this fungus matched those in GenBank (accessions of subject sequences) for this species at 100%. The isolates also exhibited the distinct morphology of this species which is….. ITS sequences for the other fungus isolated regularly most closely matched sequences for Fusicolla acetilerea (formerly Fusarium merismoides var. acetilereum) in GenBank (accessions) -tubulin sequence generated for this fungus (closest match, 90% to Fusarium ciliatum). The dark morphospecies isolated from gallery dumps and biofilms found on dead insects matched morphological descriptions for Paecilomyces variotii. Sequences for this fungus also matched those for this species in GenBank (accessions ITS, -tubulin, 99% 1 base pair difference). Less commonly isolated fungi included Penicillium decaturense (accessions ITS 99% match in GenBank, morphological and cultural characters also well matched to species description), a Raffaelea-like sp. B fungus (no close match (<90%) in GenBank to either ITS or b-tubulin sequences; id not finished yet), and an Artrinium sp. (sequencing not finished).

Overall, R. sulphurea (GEE: p < 0.001), the Raffaelea-like sp. B (p = 0.003) and Artrinium sp. (only on CSMA) were more commonly detected on CSMA than on MA, whereas the opposite was

105

Chapter 7 true for F. acetilerea-like sp. (p < 0.001; Table 2). The other species were isolated equally often from CSMA and MA.

Mycetangially-transmitted fungi F. acetilerea-like sp. dominated in mycetangia of all thirteen dissected females and was found in six of their guts. R. sulphurea was present in mycetangia of only one of these females, but was isolated from 9 of 13 female guts (Fig. 1).

Fungus dynamics in relation to offspring development R. sulphurea and F. acetilerea-like dominated the gardens of freshly founded galleries after the foundresses had started to lay eggs (period 1, Fig. 1). Egg numbers tended to increase with increasing abundance of R. sulphurea (GEE: pfrequency = 0.09; Table S2). Fungus composition (frequency and presence of single species) of unsuccessful galleries without any eggs did not differ from galleries with eggs (p < 0.05). All six species of fungi were isolated from samples taken during the period after eggs had hatched (period 2). F. acetilerea-like sp. increased in abundance (period 1 vs. 2: ppresence = 0.02; Table S1). This had no effect on larval numbers, however (p = 0.92; Table S2). Larval numbers were positively correlated with the frequency of R. sulphurea (p = 0.035), and tended to correlate negatively with the frequency of P. variotii (p = 0.063; Fig.2, Table S2). Fungus composition (frequency and presence of single species) of galleries in which all larvae died during development did not differ from galleries with successfully developing larvae (p > 0.05).

Abundance of R. sulphurea (period 1 vs. 4: pfrequency = 0.028) and P. variotii (period 2 vs. 4; pfrequency = 0.001) were significantly lower after maturation of all offspring (Table S1). Only Pe. decaturense increased towards this period (period 1+2 vs. 4: ppresence = 0.02). Data for F. acetilerea- like sp. is contradictory; the number of galleries where it was present decreased (period 2 vs. 4: ppresence = 0.03), whereas the relative frequency within galleries increased (period 2 vs. 4: pfrequency = 0.002).

Fungal composition in relation to location R. sulphurea (pfrequency < 0.001, ppresence = 0.12) and F. acetilerea-like sp. (pfrequency = 0.002, ppresence = 0.08) were more common in the brood chamber than in the main tunnel of the galleries, while an opposite trend P. variotii (pfrequency < 0.001, ppresence = 0.01; Table S1). Data on frequency and presence for Raffaelea-like sp. was contradictory (increase pfrequency = 0.09, decrease ppresence = 0.07). P. variotii was the dominant species growing in the dumps of the beetles (present in 7 of 8 galleries), followed by F. acetilerea-like sp. (in 2 of 8 galleries), Raffaelea-like sp. (in 1 of 8 galleries) and Pe. decaturense (in 1 of 8 galleries; Fig.1).

Fungi on the body of adults We frequently saw foundresses that were overgrown with a thin layer of fungi. If they were not able to successfully produce offspring (who would have groomed off this layer) this likely led to the death of these females, because at some point this layer became so thick that it constricted movements of beetles through the tunnels (Biedermann and Taborsky 2011). P. variotii (present on 4 of 4 beetles) was the main component forming this layer, but F. acetilerea-like sp. (in 2 of 4 beetles; Fig. 1) was also found.

Discussion

Fungus isolations from laboratory-reared galleries revealed that two filamentous fungi, R. sulphurea and F. acetilerea-like sp. are regularly associated with X. saxesenii. Nest-founding females transferred both species from the natal nest in their spore-carrying organs (mycetangia and guts), which led to the initial prevalence of these species in the fungal gardens and during the period of 106

Chapter 7 larval development. The primary food fungus, R. sulphurea, formed thin ambrosia layers that were fed upon by the adults (Fig. 3; Biedermann and Taborsky 2011). F. acetilerea-like sp. probably served as a secondary and/or larval food source. However, over time five other filamentous fungi, Arthrinium sp., Raffaelea-like sp. B, Pe. decaturense and P. variotii, appeared within the fungal gardens, of which the first two seem to be commensal. Pe. decaturense is probably weakly parasitic. P. variotii appears strongly pathogenic to the beetles, because it invaded the fungal gardens over time and negatively affected larvae and adults.

The mutualistic associates of X. saxesenii R. sulphurea and F. acetilerea-like were the only fungi isolated from the spore-carrying organs of dispersing females ready to found new fungus gardens. Their elytral pouches (mycetangia) contained mostly spores of F. acetilerea-like sp., whereas R. sulphurea dominated in gut samples. While fungus farming ants and termites actively pick symbiont propagules to found new fungus gardens (Mueller et al. 2005), ambrosia beetles passively take up spores into external mycetangia from the surrounding environment when moving within their natal nest before dispersal (Beaver 1989). At that time (which is the period when only adult progeny is present) galleries were already heavily contaminated by four other non-mutualistic fungi. Selective substances produced in the gut lumen and by numerous glands lining the beetles` mycetangia (Schneider and Rudinsky 1969; Schneider 1991) likely assure the exclusive transmission of R. sulphurea and F. acetilerea-like sp.; both symbionts were isolated from both organs, although at different frequencies. In summary, this confirmed the existence of two spore carrying modes in X. saxesenii (Francke-Grosmann 1975) and suggests important roles for both fungi in the life-cycle of this beetle. Several observations suggest a strong mutualistic role of R. sulphurea in this system. First, it prevailed in galleries shortly after their foundation and throughout brood development, where it formed characteristic ambrosial layers of densely packed conidia with large nutritional conidiospores on the gallery walls (Fig. 3). Ambrosia layers are predominately formed within the brood chambers, where adult daughters and larval offspring aggregate and crop off the nutritional conidia (Biedermann and Taborsky 2011). Second, the number of eggs produced by the foundress and larval numbers tended to correlate positively with the frequency of R. sulphurea. Third, the abundance of R. sulphurea was lowest in galleries with only adult progeny, suggesting that egg production ceased when productivity of this fungus dropped below a certain threshold. Fourth, R. sulphurea has been isolated from mycetangia and galleries of X. saxesenii originating researchers from all over the US and Europe (Francke-Grosmann 1956; 1975; Batra 1967; Roeper et al. 1980; Roeper and French 1981). Raffaelea (Ascomycota: Ophiostomatales) and Ambrosiella (Microascales) are also the primary mutualists of many other temperate ambrosia beetles (Roeper and French 1981; Farrell et al. 2001; Harrington et al. 2010). The role of F. acetilerea-like sp. for the beetles is less clear. It was part of the fungal layer that sometimes formed on adults. Its ability to grow on body surfaces might explain why it was so frequently isolated from the elytral mycetangia of beetles. On the other hand, observations suggest that Fusarium sensu lato may also grow in a yeast phase within galleries, similar to the ambrosial growth of Raffaelea species (Abrahamson 1969; R.A. Roeper, personal communication), which might have been present also in our galleries. It also was common within brood chambers during development of larvae indicating that it may serve as an additional food source for them. Larvae feed differently than adults that feed on ambrosial layers, however, with most feeding occurring on fungus-infested wood and faeces of other larvae (Biedermann and Taborsky 2011). Enzymatic activity in guts of larvae and adults also appear to differ (De Fine Licht and Biedermann 2012), which indicates that they may feed on different fungi. Furthermore, Fusarium merismoides has recently been found to be of some nutritional value for larvae of the ambrosia beetle Platypus quercivorus (Platypodinae; Qi et al. 2011). While Raffaelea and Ambrosiella are often cited as the main food fungi associated with ambrosia beetles, Fusarium sensu lato appear to be extremely common, but its role for the beetles is obscure and might depend on the host. Fusarium solani, for example, is weakly pathogenic for the bark beetle Dendroctonus frontalis (Moore 1973). In galleries of the ambrosia

107

Chapter 7 beetles Anisandrus dispar and Xyleborus ferrugineus it is very common, however, and in the latter it can even serve as the primary food fungus (Zimmermann 1973; Norris 1979). Species of the genus Fusarium sensu lato are frequently isolated from soil and plants, and also are commonly associated with ambrosia beetles (Norris 1979). Unpublished studies by R.A. Roeper suggest that Fusarium associated with ambrosia beetles may have higher temperature growth optima than Raffaelea species, which might explain why Fusarium species are more commonly found as associates of ambrosia beetles in the tropics and in the southern US than in the northern US (our study was done in Louisiana; R.A. Roeper, personal communication). Fusarium has long been acknowledged as a complex, unresolved polyphyletic group of fungi. Recently, the taxonomy of Fusarium has been revisited with the result that some species have been moved to different genera, including Fusicolla (Grafenhan et al. 2011). It is likely that many ambrosia beetle associates formerly identified as Fusarium will be reclassified in new genera. In summary, these observations suggest a secondary mutualistic role of F. acetilerea-like sp. for X. saxesenii, but experimental studies are needed to determine whether this is really the case. The cultivation of two or more mutualists is apparently common in ambrosia beetles (Norris 1979; Haanstad and Norris 1985; Harrington and Fraedrich 2010; Endoh et al. 2011), against predictions of most hypotheses regarding the formation and maintenance of mutualism. Symbiont competition can generate selection for symbiont traits that enhance their competitive ability but harm the host (Frank 1996; Mueller 2002). Additionally, there should be strong selection for a ‘best symbiont’, over time leading to its fixation with a host. However, while some symbioses, including fungus-farming ants and termites, involve only one mutualistic partner (e.g. Mueller et al. 2005), many others involve multiple symbionts indicating mechanisms that allow coexistence. In the case of symbionts in ambrosia beetle gardens, niche differences of the various fungi may reduce competition. If the fungi exploit different resources in the tree this may alleviate selection against any one partner and help to maintain a community of symbionts rather than a single fungal partner. In X. saxesenii, one fungus species might serve as food for adults and the other one as food for larvae (De Fine Licht and Biedermann 2012). Cooperation between symbionts is also possible. Laboratory studies showed that R. sulfurea has a need for B-vitamins to grow, which might be provided by other microbes (filamentous fungi, yeasts or bacteria; R.A. Roeper, personal communication). In Dendroctonus bark beetles (Scolytinae), the possession of several apparently redundant fungal symbionts with differing environmental tolerances may reduce the risk of the host being left aposymbiotic when environmental conditions shift over a season and from year to year (Six and Bentz 2007). Experimental studies are needed to clarify the roles and interactions of the various symbionts associated with bark and ambrosia beetles.

Other fungal associates Species of the anamorphic genera Arthrinium, Paecilomyces, and Penicillium were also associated with X. saxesenii, without being transmitted by founder females in their spore-carrying organs. Instead, spores of such fungi have been found to be vectored in small quantities on females` body surfaces (Francke-Grosmann 1967). Arthrinium sp. was isolated at low frequencies from all gallery-classes and also from larval bodies. The presence of Arthrinium sp. did not affect adult beetles or larvae in this study Pe. decaturense and P. variotii predominated in old galleries, at the entrance tunnel and in gallery dumps. We regard both as parasites, although we found no negative effects of Pe. decaturense in this study. However, its known to produce antiinsectan compounds (Zhang et al. 2003) and previous to this study Pe. decaturense has been only isolated living (and probably feeding) on a wood decay fungus (Peterson et al. 2004). Penicillium sp. often compete with insects for ephemeral resources and thus regularly produce compounds against insect feeding (Peterson et al. 2004; Rohlfs and Churchill 2011).Unidentified Penicillium sp. have been frequently reported from old galleries of X. saxesenii and have been regarded weak antagonists (Fischer 1953; Francke-Grosmann 1975). For P. variotii there is clear evidence for its antagonistic effect on X. saxesenii: Its abundance negatively affected larval numbers and it formed a deadly fungal layer on the surface of adult

108

Chapter 7 beetles. In a previous study we found this layer to have caused the death of at least 7 out of 29 isolated females (Biedermann and Taborsky 2011). The genus Paecilomyces includes many entomopathogenic species and also plant saprobes belonging to the earliest colonizers of recently dead plants (e.g. Kim et al. 2001; Tang et al. 2005).

Fungus dynamics in a laboratory setting vs. field galleries Ambrosia beetles live in the wood of trees, where they can only be studied by destructive gallery dissection. Thus a laboratory setting was required to study the fungus dynamics in relation to the dynamics of the beetles’ life history within galleries. It is intrinsic to all laboratory studies, however, that results might be influenced by the difference between laboratory and field conditions. Our artificial breeding medium, for example, is richer in nutrients and moisture than natural wood (Saunders and Knoke 1967). Therefore it is important to determine whether these differences could have influenced the conclusions of our study: First, total offspring numbers of X. saxesenii in field and laboratory galleries are almost identical (Biedermann et al. 2009). Second, different substrate conditions might influence the abundance of particular fungi relative to others, but should not affect within species dynamics (e.g. time course of abundance in dependence of gallery stage and composition). Thus, we believe that our conclusions are generally valid but the results on relative fungal abundance should not be overinterpreted.

Can beetles influence the community of their gardens? Larvae and adult X. saxesenii constantly remove the growth of F. acetilerea-like sp. and P. variotii from their body surface by grooming each other. They also constantly crop their gardens and hinder the spread of pathogens by dumping old sawdust, faeces, fungal material and dead individuals out of the gallery entrance (Biedermann and Taborsky 2011). Symbionts may profit from direct fertilization by the faeces of larvae and adults; nitrogen is known to be recycled by the primary fungi of X. ferrugineus (Norris 1975). If larvae and adults are removed from a gallery, its fungus gardens are overrun by saprobic fungi (normally coexisting at low levels) within 1-2 days (Leach et al. 1940; Batra 1979; Norris 1979; Biedermann and Taborsky 2011), which demonstrates that the beetles play an active role in maintaining the composition of their gardens. As mechanical removal is likely to be only partially effective against weed-fungi, beetle-associated antibiotics producing bacteria may play a role in controlling weeds and pathogens as in other fungus-culturing insects (Mueller et al. 2005). In Dendroctonus bark beetles several bacterial groups are known to reduce the growth of antagonistic fungi (Scott et al. 2008) and some of them are actively applied within oral secretions during specialised cleaning behaviours by the adults (Cardoza et al. 2006). Streptomyces griseus, which is known to produce antibiotics, has been recently isolated from X. saxesenii galleries (Grubbs et al. 2011), but whether bacteria influence the composition of ambrosia gardens should be investigated in future studies. This is the first study reporting correlative evidence for fitness effects of particular fungi on an ambrosia beetle. A relatively high number of filamentous fungi are commonly associated with the ambrosia beetle X. saxesenii, which suggests a predictable rather than an accidental presence in the fungus gardens. Interestingly, most of the secondary fungal flora that we reported has been also isolated from nests of fungus-growing ants (Rodrigues et al. 2008; 2011): Fusarium sp., Paecilomyces sp. and Penicillium sp. are regularly isolated from different attine ant species. The reason might be that these genera are often associated with plants, either as endophytes or early saprobes. Thus, these fungi are likely present in the plant material the ants collect to provision their gardens (Rodrigues et al. 2011) as well as in the wood surrounding ambrosia beetle galleries; although in the present study fungi must have been vectored by the dispersing females. The absence of a strong association between these secondary associates and the farming insects in both systems suggests that most of these filamentous fungi are transient components of the gardens. This does not mean, however, that secondary microbes do not influence insect-fungus symbioses. Whereas research on ambrosia beetle microbial symbioses is still at the beginning, many important mutualists and

109

Chapter 7 antagonists have been reported from the fungus –farming ants (Silva et al. 2006; Pinto-Tomas et al. 2009). Our study revealed six fungal species within ambrosia beetle gardens, which is at least ten times less then recently isolated from fungus-growing ant gardens in the field (between 66 and 106 fungal species depending on the ant species; (Rodrigues et al. 2011). Comparing species accumulation curves of both studies (Fig. S1 vs. Fig. 2 in Rodrigues et al. 2011) suggests that in this study we have sampled all fungi present, whereas many more might be detected by increasing sampling effort in the ants. Of course these differences in absolute species numbers between beetles and ants is in part an effect of laboratory breeding, but this must not hide the fact that the beetle galleries are a more closed system than the ant nests (U.G. Mueller, personal communication). First, the ants build their nests in soil, which is heavily contaminated by microbes. Second, they have to leave the nest to forage and thus are exposed to all kinds of contaminants. Finally, they have to use substrate that contains all kinds of endophytic and epiphytic microbes. The beetles, in contrast, bury into dying or recently dead wood, a substrate, which is much less contaminated by microbes than soil. Thus, it seems that ambrosia beetles have an easier time to sequester their gardens from foreign fungi than the fungus-farming ants, despite the fact that beetles cannot protect their gardens from fungi that grow in from the surrounding wood (the ants sequester their gardens from the surrounding soil; Hölldobler and Wilson 1990) that serves as the food-substrate for their fungi. In summary, sophisticated techniques for weeding and disinfection like in the ant gardens (e.g. Currie et al. 1999; Currie and Stuart 2001; Boomsma and Aanen 2009) might not be needed in beetle gardens. The biggest threat to ambrosia gardens are probably diminishing nutrients, so recycling of excretions may be important. Indeed, some evidence of nutrient cycling between beetles and fungi exist (Kok and Norris 1972a; 1972c; Biedermann and Taborsky 2011).

Literature

Abrahamson LP (1969). Physiological interrelationships between ambrosia beetles and their symbiotic fungi. Ph.D.-thesis, University of Wisconsin, Madison.

Abrahamson LP & Norris DM (1970) Symbiotic interrelationships between microbes and ambrosia beetles. 5.Amino acids as a source of nitrogen to fungi in beetle. Annals of the Entomological Society of America 63: 177-180.

Batra LR (1967) Ambrosia fungi - A taxonomic revision and nutritional studies of some species. Mycologia 59: 976-1017.

Batra LR (1979) Symbiosis, commensalism and aposymbiosis - conclusions. Insect-Fungus Symbiosis. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 259-265. John Wiley & Sons, New York.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Benjamin RK, Blackwell M, Chapela IH, Humber RA, Kevin G, Klepzig KD, Lichtwardt RW, Malloch D, Noda H, Roeper RA, Spatafora JW & Weir A (2004) Insect-and Other Arthropod-Associated Fungi. Biodiversity of Fungipp. 395-433. Academic Press, Burlington.

Berryman AA (1989) Adaptive pathways in scolytid-fungus associations. Insect-Fungus Interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 145-159. Academic Press, London.

Biedermann, PHW (2007) Social behaviour in sib mating fungus farmers. Master thesis, University of Berne, Switzerland.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: Behavior and laboratory 110

Chapter 7 breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW, Peer K & Taborsky M (2011) Female dispersal and reproduction in the ambrosia beetle Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Mitteilungen der Deutschen Gesellschaft für allgemeine und angewandte Entomologie in review.

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

Boomsma JJ & Aanen DK (2009) Rethinking crop-disease management in fungus-growing ants. Proceedings of the National Academy of Sciences of the United States of America 106: 17611-17612.

Bridges JR, Marler JE & McSparrin BH (1984) A quantitative study of the yeasts and bacteria associated with laboratory-reared Dendroctonus frontalis Zimm. (Coleopt., Scolytidae)1. Zeitschrift für Angewandte Entomologie 97: 261-267.

Cardoza YJ, Klepzig KD & Raffa KF (2006) Bacteria in oral secretions of an endophytic insect inhibit antagonistic fungi. Ecological Entomology 31: 636-645.

Currie CR, Scott JA, Summerbell RC & Malloch D (1999) Fungus-growing ants use antibiotic-producing bacteria to control garden parasites. Nature 398: 701-704.

Currie CR & Stuart AE (2001) Weeding and grooming of pathogens in agriculture by ants. Proceedings of the Royal Society of London Series B-Biological Sciences 268: 1033-1039.

De Fine Licht HH & Biedermann PHW (2012) Patterns of functional enzyme activity suggest that larvae are the key to successful fungus farming by ambrosia beetles. Front.Zool., in review.

Endoh R, Suzuki M, Okada G, Takeuchi Y & Futai K (2011) Fungus Symbionts Colonizing the Galleries of the Ambrosia Beetle Platypus quercivorus. Microbial Ecology 62: 106-120.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z. Morph. u. Ökol. Tiere 45: 275-308.

Francke-Grosmann H (1967) Ectosymbiosis in wood-inhabiting beetles. Symbiosis (Henry SM, ed), pp. 141-205. Academic Press, New York.

Francke-Grosmann H (1975) Zur epizoischen und endozoischen Übertragung der symbiotischen Pilze des Ambrosiakäfers Xyleborus saxeseni (Coleoptera: Scolitidae). Entomologica Germanica 1: 279-292.

Frank SA (1996) Host-symbiont conflict over the mixing of symbiotic lineages. Proceedings of the Royal Society of London Series B-Biological Sciences 263: 339-344.

Glass NL & Donaldson GC (1995) Development of primer sets designed for use with the PCR to amplify conserved genes from filamentous ascomycetes. Appl. Environ. Microbiol. 61: 1323-1330.

Grafenhan T, Schroers HJ, Nirenberg HI & Seifert KA (2011) An overview of the taxonomy, phylogeny, and typification of nectriaceous fungi in Cosmospora, Acremonium, Fusarium, Stilbella, and Volutella. Studies in Mycology 68: 79-113. 111

Chapter 7

Grubbs KJ, Biedermann PHW, Suen G, Adams SM, Moeller JA, Klassen JL, Goodwin LA, Woyke T, Munk AC, Bruce D, Detter C, Tapia R, Han CS & Currie CR (2011) The Complete Genome Sequence of Streptomyces cf. griseus (XyelbKG-1), an Ambrosia Beetle-Associated Actinomycete. J. Bacteriol. 193: 2890-2891.

Haanstad JO & Norris DM (1985) Microbial symbiotes of the ambrosia beetle Xyletorinus politus. Microbial Ecology 11: 267-276.

Harrington TC & Fraedrich SW (2010) Quantification of Propagules of the Laurel Wilt Fungus and Other Mycangial Fungi from the Redbay Ambrosia Beetle, Xyleborus glabratus. Phytopathology 100: 1118-1123.

Hölldobler B & Wilson EO (1990) The ants. The Belknap Press of Harvard University Press, Cambridge, MA.

Hulcr J, Mann R & Stelinski LL (2011) The Scent of a Partner: Ambrosia Beetles Are Attracted to Volatiles from Their Fungal Symbionts. Journal of Chemical Ecology 37: 1374-1377.

Kim JJ, Ra JB, Son DS & Kim GH (2001) Fungi Colonizing Sapwood of Japanese Red Pine Logs in Storage. Mycobiology 29: 205-209.

Kok LT & Norris DM (1972a) Comparative phospholipid compositions of adult female Xyleborus ferrugineus and its mutualistic fungal ectosymbionts. Comparative Biochemistry and Physiology 42: 245-254.

Kok LT & Norris DM (1972b) Phospholipid composition of fungi mutualistic with Xyleborus ferrugineus. Phytochemistry 11: 1449-1453.

Kok LT & Norris DM (1972c) Symbiotic interrelationships between microbes and ambrosia beetles 6. Aminoacid composition of ectosymbiotic fungi of Xyleborus ferrugineus (Coleoptera, Scolytidae). Annals of the Entomological Society of America 65: 598-602.

Kok LT & Norris DM (1973) Comparative sterol compositions of adult female Xyleborus ferrugineus and its mutualistic fungal ectosymbionts. Comparative Biochemistry and Physiology 44: 499-505.

Leach JG, Hodson AC, Chilton SJP & Christensen CM (1940) Observations on two ambrosia beetles and their associated fungi. Phytopathology 30: 227-236.

Martin MM (1987) Invertebrate-Microbial Interactions . Cornell University Press, Ithaca.

Moore GE (1973) Pathogenicity of three entomogenous fungi to the Southern Pine Beetle at various temperatures and humidities. Environmental Entomology 2: 54-57.

Mueller UG (2002) Ant versus fungus versus mutualism: Ant-cultivar conflict and the deconstruction of the attine ant-fungus symbiosis. Am. Nat. 160: S67-S98.

112

Chapter 7

Mueller UG, Gerardo NM, Aanen DK, Six DL & Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563-595.

Mueller UG, Schultz TR, Currie CR, Adams RMM & Malloch D (2001) The origin of the attine ant-fungus mutualism. Q. Rev. Biol. 76: 169-197.

Norris DM (1975) Chemical interdependence among Xyleborus spp. ambrosia beetles and their symbiotic microbes. Material und Organismen 3: 479-788.

Norris DM (1979) The mutualistic fungi of Xyleborini beetles. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 53-63. Allanheld, Osmun & Company, Montclair.

O'Donnell K & Cigelnik E (1997) Two Divergent Intragenomic rDNA ITS2 Types within a monophyletic lineage of the fungus Fusarium are nonorthologous. Molecular Phylogenetics and Evolution 7: 103-116.

Peer K & Taborsky M (2005) Outbreeding depression, but no inbreeding depression in haplodiploid ambrosia beetles with regular sibling mating. Evolution 59: 317-323.

Peer K & Taborsky M (2007) Delayed dispersal as a potential route to cooperative breeding in ambrosia beetles. Behav. Ecol. Sociobiol. 61: 729-739.

Peterson SW, Bayer EM & Wicklow DT (2004) Penicillium thiersii, Penicillium angulare and Penicillium decaturense, new species isolated from wood-decay fungi in North America and their phylogenetic placement from multilocus DNA sequence analysis. Mycologia 96: 1280-1293.

Pinto-Tomas AA, Anderson MA, Suen G, Stevenson DM, Chu FST, Cleland WW, Weimer PJ & Currie CR (2009) Symbiotic Nitrogen Fixation in the Fungus Gardens of Leaf-Cutter Ants. Science 326: 1120-1123.

Qi HY, Wang JG, Endoh R, Takeuchi Y, Tarno H & Futai K (2011) Pathogenicity of microorganisms isolated from the oak platypodid, Platypus quercivorus (Murayama) (Coleoptera: Platypodidae). Applied Entomology and Zoology 46: 201-210.

R Development Core Team (2008) R: A language and environment for statistical computing. Vienna, Austria, R Foundation for Statistical Computing.

Rodrigues A, Bacci M, Mueller UG, Ortiz A & Pagnocca FC (2008) Microfungal "weeds" in the leafcutter ant symbiosis. Microbial Ecology 56: 604-614.

Rodrigues A, Mueller UG, Ishak HD, Bacci Jr Mc & Pagnocca FC (2011) Ecology of microfungal communities in gardens of fungus-growing ants (Hymenoptera: Formicidae): a year-long survey of three species of attine ants in Central Texas. FEMS Microbiology Ecology 78: 244-255.

Roeper RA (1995) Patterns of mycetophagy in Michigan ambrosia beetles. Michigan Academian 27: 153-161.

Roeper RA & French JRJ (1981) Ambrosia fungi of the Western United States and Canada - beetle assocaitions (Coleoptera: Scolytidae), Tree Hosts, and Distribution. Northwest Science 55: 305-309.

Roeper RA, Hazen CR, Helsel DK & Bunce MA (1980) Studies on Michigan ambrosia fungi. The Michigan Botanist 19: 69-73.

Rohlfs M & Churchill ACL (2011) Fungal secondary metabolites as modulators of interactions with insects and other arthropods. Fungal Genetics and Biology 48: 23-34.

Saunders JL & Knoke JK (1967) Diets for rearing the ambrosia beetle Xyleborus ferrugineus (Fabricius) in vitro. Science 15: 463.

Schneider I (1991) Some ecological aspects of the ambrosia-symbiosis. Anzeiger für Schadlingskunde Pflanzenschutz Umweltschutz 64: 41-45. 113

Chapter 7

Schneider IA & Rudinsky JA (1969) Mycetangial glands and their seasonal changes in Gnathotrichus retusus and G. sulcatus. Annals of the Entomological Society of America 62: 39-43.

Scott JJ, Oh DC, Yuceer MC, Klepzig KD, Clardy J & Currie CR (2008) Bacterial protection of beetle-fungus mutualism. Science 322: 63.

Silva A, Rodrigues A, Bacci M, Pagnocca FC & Bueno OC (2006) Susceptibility of the ant-cultivated fungus Leucoagaricus gongylophorus (Agaricales : Basidiomycota) towards microfungi. Mycopathologia 162: 115-119.

Six DL (2003) Bark beetle-fungus symbioses. Insect Symbiosis (Bourtzis K & Miller TA, eds), pp. 97-114. CRC Press, Boca Raton.

Six DL & Bentz BJ (2007) Temperature determines symbiont abundance in a multipartite bark beetle-fungus ectosymbiosis. Microbial Ecology 54: 112-118.

Six D, Doug Stone W, de Beer Z & Woolfolk S (2009) Ambrosiella beaveri sp. nov., Associated with an exotic ambrosia beetle, Xylosandrus mutilatus (Coleoptera: Curculionidae, Scolytinae), in Mississippi, USA. Antonie van Leeuwenhoek 96: 17-29.

Tamura K, Dudley J, Nei M & Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24: 1596-1599.

Tang AMC, Jeewon R & Hyde KD (2005) Succession of microfungal communities on decaying leaves of Castanopsis fissa. Canadian Journal of Microbiology 51: 967-974.

Vilgalys R & Hester M (1990) Rapid genetic identification and mapping of enzymatically amplified ribosomal DNA from several Cryptococcus species. J. Bacteriol. 172: 4238-4246.

Wheeler Q & Blackwell M (1984) Fungus-Insect Relationships . Columbia University Press , New York.

Wilding N, Collins NM, Hammond PM & Webber JF (1989) Insect-Fungus Interactions. Academic Press, London.

Zhang YC, Li C, Swenson DC, Gloer JB, Wicklow DT & DOWD PF (2003) Novel antiinsectan oxalicine alkaloids from two undescribed fungicolous Penicillium spp. Organic Letters 5: 773-776.

Zimmermann G (1973) Die Pilzflora einiger im Holz lebender Borkenkäfer. Material und Organismen 8: 121-131.

114

Chapter 7

Graphs

proportion of samples 0.00 0.20 0.40 0.60 0.80 1.00

Raffaelea sulphurea

mycangium (N = 13) Arthrinium spp.??

Raffaelea-like sp. B

Fusicolla acetilerea-like sp. gut (N = 13)

transmitted fungi transmitted Penicillium decaturense

Paecilomyces variotii

foundress & eggs (N = 8)

foundress & larvae (N = 10)

only adults (N = 9) fungi within gallery-classes within fungi

larval bodies (N = 9)

fungus growing on adults (N = 4) other samples other

gallery dump (N = 8)

Fig. 1. Percentage of fungal species isolated relative to total number of samples taken. Isolations from the spore-carrying organs (mycetangium, gut), of different gallery-classes, of whole larval bodies, of the fungal layer growing on adult beetles, and of the gallery dump are displayed. N = sample size of dissected adult beetles, galleries, or larvae.

115

Chapter 7

Fig.2. Relationship of different fungi with numbers of Xyleborinus saxesenii larvae during the immature stage of colonies. During the stage with foundress, eggs, larvae and pupae the frequency of Raffaelea sulphurea correlated positively with the numbers of larvae present (GEE: p = 0.035; LR: r2 = 0.11). In contrast, the frequency of Paecilomyces variotii tended to correlate negatively with the numbers of larvae present (GEE: p = 0.063; LR: r2 = 0.22). All other fungal species had no significant relationship with offspring numbers. N = 10 galleries; for statistical details see Table 3.

Fig.3. Morphology of the ambrosial layer produced by Raffaelea sulfurea within laboratory galleries of Xyleborinus saxesenii. Conidiophores with conidia are visible. Note the area on the left side where several conidia are missing on top of the conidiophores – adult beetles have probably cropped them off. The picture was taken using Scanning Electron Microscopy (SEM) with 400× magnification.

116

Chapter 7

Table 1. Characteristics of the six fungal species isolated from laboratory galleries of Xyleborinus saxesenii.

Order Species Medium Location Fungal species dynamic Influence on progeny Dominance ( > 50% of Proposed numbers samples) association Ophiostomatales Raffaelea in gut, CSMA > MA BC > MT 1 foundress/eggs > adult progeny 1 positive (eggs*, larvae) mutualist sulphurea during foundress/eggs Raffaelea- CSMA > MA BC < MT*1 not present during foundress/eggs ns - ? like sp. B Sordariales Arthrinium only on CSMA ns ns. ns - ? sp.? Hypocreales Fusicolla foundress/eggs < foundress/larvae 2 acetilerea- CSMA < MA BC > MT 1 foundress/larvae < adult progeny 1 ns in mycetangium mutualist like sp. foundress/larvae > adult progeny 2 Penicillium foundress/eggs < adult progeny 2 ns ns ns - parasite (?) decaturense foundress/larvae < adult progeny 2 Eurotiales Paecilomyces not present during foundress/eggs on adult bodies, ns BC < MT 2 negative (larvae*) parasite variotii foundress/larvae > adult progeny 1,2 in gallery dump Significant differences (p < 0.05) and trends (p < 0.01) indicated with * are shown, except were marked with ns (not significant). Abbreviations: CSMA – cycloheximide-streptomycin-malt agar, MA – malt agar; BC – brood chamber, MT – main tunnel; “…>…” /”… <…” denote the direction of the significant difference in the isolation frequency between the two media/locations/gallery classes. 1 Significant difference in the frequency of the fungus (p < 0.05) 2 Significant difference in the presence of the fungus (p < 0.05)

117

Chapter 7 Supplementary Information

Table S1. Factors influencing the abundance of fungi isolated from X. saxesenii. Separate (1) GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences between the frequency, and (2) GLM models for examining differences between the presence (yes/no) of the isolated fungal species, influenced by culture medium (CSMA – cycloheximide-streptomycin- malt agar, MA – malt extract agar), location within the gallery (brood chamber or main tunnel), and stage of gallery development. Model coefficients are reported as coeff.  se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of main factors on the fungal frequencies are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse. Significant relationships are highlighted in bold face. coeff.  se t p

Raffaelea sulphurea Intercept of frequency (CSMA, brood chamber, 1.65 0.85 1.95 0.05 foundress with eggs) Contrast CSMA vs. MA -2.6 0.37 -7.03 <0.001 Contrast brood chamber vs. main tunnel -2.02 0.36 -5.62 <0.001 Contrast foundress with eggs vs. foundress with -1.28 1.05 -1.22 0.22 larvae Contrast foundress with eggs vs. only adult progeny -2.41 1.09 -2.2 0.028 Contrast foundress with larvae vs. only adult -1.13 0.82 -1.37 0.17 progeny

Intercept of presence (brood chamber, foundress 1.34 0.69 1.94 0.058 with eggs) Contrast brood chamber vs. main tunnel -1 0.64 -1.57 0.12 Contrast foundress with eggs vs. foundress with -0.16 0.78 -0.2 0.84 larvae Contrast foundress with eggs vs. only adult progeny -0.63 0.79 -0.8 0.43 Contrast foundress with larvae vs. only adult -0.48 0.76 -0.63 0.53 progeny

Artrinium sp. Intercept of frequency (CSMA, brood chamber, -11.7 23.3 -0.5 0.61 foundress with eggs) Contrast CSMA vs. MA Only present on CSMA Contrast brood chamber vs. main tunnel -5.66 4.88 -1.16 0.25 Contrast foundress with eggs vs. foundress with -3.8 32.9 -0.12 0.91 larvae Contrast foundress with eggs vs. only adult progeny -1.3 32.8 -0.04 0.97 Contrast foundress with larvae vs. only adult -2.45 1.69 -1.45 0.15 progeny

Intercept of presence (brood chamber, foundress -2.45 1.18 -2.08 0.04 with eggs) Contrast brood chamber vs. main tunnel -0.57 1.31 -0.44 0.67 Contrast foundress with eggs vs. foundress with -0.16 1.52 -0.1 0.92 larvae Contrast foundress with eggs vs. only adult progeny -0.03 1.52 -0.02 0.98 Contrast foundress with larvae vs. only adult 0.12 1.51 0.08 0.94 progeny

Raffaelea-like sp. B Intercept of frequency (CSMA, brood chamber, -31.9 999 -0.00 1 foundress with eggs) Contrast CSMA vs. MA -4.83 1.6 -3.02 0.003 Contrast brood chamber vs. main tunnel 8.35 4.91 1.7 0.09 Contrast foundress with eggs vs. foundress with Only present during foundress with larvae larvae Contrast foundress with eggs vs. only adult progeny Only present during only adult progeny Contrast foundress with larvae vs. only adult -2.26 1.74 -1.3 0.19 progeny

Intercept of presence (brood chamber, foundress 0.45 0.69 0.65 0.52 with eggs) Contrast brood chamber vs. main tunnel -1.47 0.79 -1.86 0.07 Contrast foundress with eggs vs. foundress with Only present during foundress with larvae larvae 118

Chapter 7 Contrast foundress with eggs vs. only adult progeny Only present during only adult progeny Contrast foundress with larvae vs. only adult 1.37  1.21 1.13 0.26 progeny

Fusicolla acetilerea- Intercept of frequency (CSMA, brood chamber, -8.04 2.63 -3.05 <0.002 like sp. foundress with eggs) Contrast CSMA vs. MA 9.32 1.88 4.96 <0.001 Contrast brood chamber vs. main tunnel -1.67 0.51 -3.24 0.001 Contrast foundress with eggs vs. foundress with -0.19 2.43 -0.08 0.94 larvae Contrast foundress with eggs vs. only adult progeny 4.05 2.52 1.61 0.11 Contrast foundress with larvae vs. only adult 4.17 1.33 3.13 0.002 progeny

Intercept of presence (brood chamber, foundress 0.39 0.67 0.58 0.56 with eggs) Contrast brood chamber vs. main tunnel -1.34 0.75 -1.8 0.08 Contrast foundress with eggs vs. foundress with 2.46 1.03 2.38 0.02 larvae Contrast foundress with eggs vs. only adult progeny 0.19 0.82 0.23 0.82 Contrast foundress with larvae vs. only adult -2.27 1.03 -2.2 0.03 progeny

Penicillium Intercept of frequency (CSMA, brood chamber, -10.6 12.1 0.88 0.38 decaturense foundress with eggs) Contrast CSMA vs. MA -11.9 8.66 -1.38 0.17 Contrast brood chamber vs. main tunnel -0.06 0.66 -0.09 0.93 Contrast foundress with eggs vs. foundress with 0.71 13.8 0.05 0.96 larvae Contrast foundress with eggs vs. only adult progeny 2.76 13.2 0.21 0.83 Contrast foundress with larvae vs. only adult 2.05 7.05 0.29 0.77 progeny

Intercept of presence (brood chamber, foundress -2.37 1.17 -2.02 0.05 with eggs) Contrast brood chamber vs. main tunnel -0.83 0.89 -0.93 0.36 Contrast foundress with eggs vs. foundress with 0.59 1.4 0.42 0.68 larvae Contrast foundress with eggs vs. only adult progeny 2.99 1.27 2.35 0.02 Contrast foundress with larvae vs. only adult 2.4 1.01 2.37 0.02 progeny

Paecilomyces Intercept of frequency (CSMA, brood chamber, -21.8 999 -0.01 1 variotii foundress with eggs) Contrast CSMA vs. MA -0.08 0.37 -0.23 0.82 Contrast brood chamber vs. main tunnel 2.39 0.39 6.22 <0.001 Contrast foundress with eggs vs. foundress with Only present during foundress with larvae larvae Contrast foundress with eggs vs. only adult progeny Only present during only adult progeny Contrast foundress with larvae vs. only adult -2.02 0.62 -3.25 0.001 progeny

Intercept of presence (brood chamber, foundress -20.87 999 -0.01 1 with eggs) Contrast brood chamber vs. main tunnel 2.01 0.78 2.56 0.01 Contrast foundress with eggs vs. foundress with Only present during foundress with larvae larvae Contrast foundress with eggs vs. only adult progeny Only present during only adult progeny Contrast foundress with larvae vs. only adult -2.53 0.77 -3.3 0.002 progeny

119

Chapter 7 Table S2. The relationship between the most common fungal species and the number of X. saxesenii offspring. Separate GEE models were performed with an exchangeable correlation structure of the response variable within a cluster (gallery identity) to examine the potential influence of the frequency of fungal species on offspring numbers, controlling for the influence of medium (CSMA – cycloheximide-streptomycin-malt agar; MA – malt extract agar). The potential effects on egg and larval numbers during particular periods of gallery development are shown (for graphical illustration see fig.2). Model coefficients are reported as coeff.  se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). A positive coefficient denotes a positive relationship; a negative coefficient denotes a negative relationship.

coeff.  se t p

Foundress with eggs, brood chamber

Raffaelea sulphurea Intercept of frequency (CSMA) -0.14 1.45 -0.01 0.92 Contrast CSMA vs. MA -3.49 1.04 -3.37 <0.001 Number of eggs 0.9 0.53 1.72 0.09

Fusicolla acetilerea-like sp. Intercept of frequency (CSMA) -22.9 999 0 1 Contrast CSMA vs. MA Only present on MA Number of eggs -2.91 4.24 -0.69 0.49

Paecilomyces variotii Not present

Foundress with larvae, brood chamber

Raffaelea sulphurea Intercept of frequency (CSMA) -0.15 0.51 -0.3 0.77 Contrast CSMA vs. MA -2.49 0.61 -4.08 <0.001 Number of larvae 0.05 0.02 2.11 0.035

Fusicolla acetilerea-like sp. Intercept of frequency (CSMA) -7.88 4.82 -1.63 0.1 Contrast CSMA vs. MA 13.6 5.04 2.69 0.007 Number of larvae 0.02 0.21 0.1 0.92

Paecilomyces variotii Intercept of frequency (CSMA) -0.86 0.49 -1.76 0.08 Contrast CSMA vs. MA -0.46 0.65 -0.71 0.48 Number of larvae -0.06 0.03 -1.86 0.063

120

Chapter 7

pecies richness

0 20 40 60 80 100

Number of samples

Fig. S1. Individual-based rarefaction curves of fungal communities associated with galleries of X. saxesenii. Bars indicate 95% confidence intervals.

121

Chapter 7

122

Chapter 8

Patterns of functional enzyme activity show that larvae are the key to successful fungus farming by ambrosia beetles

Henrik H. de Fine Licht1* and Peter H. W. Biedermann2

1Section of Microbial Ecology, Department of Biology, Lund University, Solvegatan 37, SE-223 62 Lund, Sweden 2Department of Behavioural Ecology, Institute of Ecology & Evolution, University of Berne, Baltzerstrasse 6, CH- 3012 Bern, Switzerland, Email: [email protected].

* Corresponding Author: Henrik H. de Fine Licht, Section of Microbial Ecology, Department of Biology, Lund University, Solvegatan 37, SE-223 62 Lund, Sweden, Email: [email protected], Tel: (+46) 46 222 3763.

Running title: Ambrosia symbiosis enzyme activity

Submitted to Frontiers in Zoology, in review

Abstract

Introduction: Fungus-farming by insects depends on the collaboration and division of labour of multiple individuals. In wood-dwelling fungus-farming weevils, the so-called ambrosia beetles, larval and adult stages have been recently shown to specialize in different social tasks. Wood is converted into fungal food by the combined action of digging larvae (which create more space for the fungi to grow), fungus tending adults and enzymatic hydrolysis by the cultivars. However, whether adults, larvae and the mutualistic fungi are producing different enzymes in this process is unknown. Here we characterize the enzyme profiles of all partners in the ambrosia beetle microbial symbiosis for the first time. We measure 13 distinct plant cell-wall degrading enzymes in Xyleborinus saxesenii adults, larvae and the fungus garden microbial consortium at different ages and locations within the nest. Finally, we compare the enzyme profiles with other fungus-farming insects. Results: We discovered that wood and fungus feeding larvae are important for successful wood degradation and fungiculture: Cellulase activity was identified in whole-body extracts of both larval and adult X. saxesenii, whereas xylanase activity was exclusively detected in larvae. These enzymes apparently (pre-) digest plant cell-wall structures within larval guts, because larvae feed on fungus- infested wood, unlike adults that only feed on fungus. Furthermore, larval feces have been found to be applied on the fungus gardens after gut passage, likely for further utilization by the mutualistic fungi. Conclusion: The enzymatic profile of the ambrosia symbiosis resembles other fungus-growing insect systems. However, remarkably in ambrosia beetle societies adults and larvae do not compete for the same food within their nests - in contrast, larvae increase colony fitness by facilitating wood degradation and fungus cultivation. The differential enzyme profiles add an additional layer of complexity to the division of behavioural tasks between life-stages already reported within the primitively eusocial X. saxesenii. In general, enzymatic division of labour may be an unrecognized more common feature of holometabolous insects, which dramatically reorganize their body during metamorphosis. Larval physiology and enzymatic profile may provide novel evolutionary opportunities to social insect societies capable of utilizing this resource.

Keywords Symbiosis, Insoluble chromogenic enzyme substrates, xylomycetophagy, Xyloborinus saxesenii, insect fungus farming, social evolution, enzymatic division of labor

123

Chapter 8 Introduction

Insects are the most abundant and diverse animal class on earth [1]. A key factor for their enormous success are adaptations to novel environments and food sources by the help of symbiotic microorganisms [2]. Insect hosts maintain prokaryotic, fungal, and bacterial associates in a variety of ways, which help them in nutrient acquisition and recycling, environmental detoxification, and defense against antagonists. By means of microbial symbionts insects are able to produce food out of plant material low on insect-accessible nutrients, but rich in structural polysaccharides (cross-linking glycans, cellulose, and lignin) and toxic metabolites [3]. In most instances this occurs internally inside the insect host with the aid of an abundant microbial gut flora [4-6], however there are several notable exceptions of insect lineages that cultivate microbes externally on plant material [7, 8]. These ectosymbionts are either kept in “gardens” and consumed directly by their insect hosts (e.g. certain lineages of fungus-growing ants, termites, and ambrosia beetles; [9]), or contribute indirectly by increasing the nutrient content of decomposing material (e.g. bark beetles; [10]), by degradation of toxic plant compounds (e.g. terpenes by bark beetle associated fungi [11, 12]), or by provisioning of extracellular enzymes which may facilitate insect ingestion or wood burrowing (e.g. wood wasps; [5, 13]). External symbionts of insects are typically filamentous fungi, yeasts and bacteria that may be transported in mycetangia (also termed mycangia [14]), which are specialized organs for fungal spore transmission that ensure successful re-establishment of the nutritional symbiosis after dispersal. Mycetangia have evolved independently several times and are known from many different fungus associated insects such as wood and phloem feeding beetles, gall midges and wood wasps [15-17]. The active care and maintenance of the fungal crops by the insect hosts after dispersal is, however, rare. Only three insect lineages, notably the fungus-growing ants, termites and ambrosia beetles, are truly farmers of their fungal crops. Within task sharing societies they not only propagate, but also actively cultivate and sustainably harvest their crops in microbial gardens (i.e., advanced fungiculture [9]). Ambrosia beetle is an ecological term used for all weevils that farm fungi within tunnel systems (galleries) in the wood of trees. Ambrosia farming is only found in Scolytinae and Platypodinae and evolved repeatedly at least nine times from the phloem feeding habit without any known reversal to non-farming [18, 19]. Ambrosia beetles usually dwell in recently dead wood that seems almost free of microbes, where they initiate nest building by inoculating tunnel walls with mutualistic fungi. The beetle-fungus relationship is often species (or genus) specific, with highly selective transmission of the primary symbionts in mycetangia of dispersing beetles [20, 21]. These so-called ambrosia fungi (usually species of the ascomycete genera Ambrosiella and Raffaelea) are forming layers of conidiophores on the tunnel walls that produce nutrient rich conidiospores for larval and adult beetle nutrition. Secondary symbionts, such as other filamentous fungi (e.g. Fusarium sp., Graphium sp., Ophiostoma sp., Paecilomyces sp., Penicillium sp. [22]), yeasts (e.g. Candida sp. [23]) and bacteria [24, 25], are also present within galleries and are often passively vectored in small amounts attached to the surface of dispersing females [15, 20]. However, the primary mutualistic ambrosia fungus is known for only a minority of the 3000 species worldwide [26- 28], and there has only been a single attempt to characterize the entire microbiome of an ambrosia gallery [29]. Studies on the dynamics of filamentous fungi in xyleborine ambrosia beetle galleries suggest that propagates of mutualistic ambrosia fungi (Ambrosiella and Raffaelea) are passively spread on tunnel walls from the mycetangia or via beetle feces during the excavation by the gallery founding female. This ensures that the mutualistic fungi dominate the gallery microbial flora initially while eggs are laid and larvae develop [22, 30]. Later, when the first offspring mature, other saprobic fungi (secondary symbionts like Penicillium sp. and Paecilomyces sp.) start to appear and increase in frequency over time. These opportunistic fungi dominate the microbial gallery flora at the time when the gallery is abandoned and all individuals disperse to found new galleries [22, 30]. In the ambrosia beetle Xyleborinus saxesenii (Fruit-tree pinhole borer) their abundance negatively affects the number of larvae and they are more commonly encountered within the entrance tunnel than within the brood chambers and also dominate in the gallery dumps [30]. Larvae of X. saxesenii do not only feed on ambrosia fungi, like the adults and larvae of many other ambrosia beetles, but feed xylomycetophagously on fungus infested wood [31]. In this way they (a) create more space for the developing fungus to form fruiting bodies on the gallery walls, (b) lower competition between group members by enlargement of the nest space, (c), reduce the 124

Chapter 8 growth of not identified molds, possibly by gregariously feeding on them [32] and (d) probably most importantly benefit the growing ambrosia fungus, because the finely fragmented woody sawdust within the larval feces is smeared on the gallery walls after defecation [32, 33]. However, nothing is known about the mechanism or the enzymatic machinery whereby the ambrosia beetles in collaboration with the consortium of symbiotic fungi degrade the surrounding wood. In larvae of a related phloem feeding bark beetle (Phloeosinus bicolor) α-amylase-, invertase-, maltase-, lactase-, and protease-activities were detected together with some hydrolytic activity on a substrate of cross- linking glycans (hemi-cellulose) but not on cellulose [34]. Similarly, in adults of the phloem feeding Ips cembrae consistent activity against cross-linking glycans together with pectinase-, α-glucosidase-, β-glucosidase-, α-galactosidase-, β-galactosidase-, trehalase-, tryptase-, peptitase-, and lipase- activities were detected in the intestinal lumen [35]. Many of these enzymes are probably of microbial origin, because bark beetles (i) carry yeasts and bacteria in their intestines [36-38] and (ii) feed on phloem that is often infested by Ophiostomatoid fungi [16]. The bark beetle associated fungi (e.g. the genera Ceratosytiopsis, Entomocorticium and Ophiostoma [39, 40]), in addition to associated yeasts [12, 36, 39], and bacteria [41, 42] are capable of producing a variety of enzymes catalyzing (a) protein/peptide degradation (endo-, exoproteases and peptidases), (b) polysaccharide/starch/sugar degradation (glycoside-hydrolytic enzymes) and (c) fat/fatty acid degradation (lipases) [39, 40, 43- 49]. Here the activities of the major groups of plant cell-wall degrading enzymes: cellulases, endo- xylanases, pectinases, and also endo-proteases and α-amylases in the ambrosia beetle system are tested for the first time. We measured the enzyme activity in larval and adult female X. saxesenii and of the whole microbiome from three distinct locations within laboratory galleries (gallery dump, entrance tunnel, and brood chamber) and find enzymatic division of labor between adult beetles and larvae. The latter possess a distinct cell-wall degrading endo-β-1,4-xylanase activity not found in adults. This was done by a high-throughput screening method of 13 different enzymes in combination with an artificial rearing technique of beetle galleries. Laboratory rearing enabled us to precisely track the development of broods, and take enzyme measurements at particular time points: (a) with abundant immature brood, (b) with both immature brood and mature offspring and (c) at the end of gallery life with almost exclusively adult offspring.

Results

Of the 13 enzyme substrates screened in this study, six specific enzyme activities (endo-β-1,4- glucanase, endo-β-1,3(4)-glucanase, endo-β-1,4-xylanase (xylan and arabinoxylan), endo-β-1,4- mannanase, and endo-protease (casein)) were consistently detected in all samples (Fig. 1A, [see Additional file 1]) and varied significantly between the three different gallery compartments: entrance, brood chamber and gallery dump (log-likelihood ANOVA comparison of final mixed models with reduced null models: likelihood-ratio3,5 = 14.1 – 50.4, p = < 0.0001 – 0.0009). For these six enzyme activities the predictor variables: ‘gallery class’ and ‘total number of animals’ where not significant (log-likelihood ANOVA comparison of final mixed models with reduced null models: likelihood-ratio5,11 = 2.0 – 8.7, p = 0.1884 – 0.9169) when included as interaction terms with gallery compartment (entrance, brood chamber or gallery dump). However, endo-β-1,4-xylanase activity against the cross-linking glycans (xylan and arabinoxylan) and endo-β-1,4-mannanase against pectin varied not only with gallery compartment but also with the age of the gallery (log-likelihood ANOVA comparison of final mixed models with reduced null models: likelihood-ratio5,11 = 12.7 – 16.9, p = 0.0095 – 0.0472, Fig. 1A, [see Additional file 1]). The final reduced and most parsimonious models for endo-β-1,4-xylanase and endo-β-1,4-mannanase enzyme activity therefore included both gallery compartment and the interaction term of gallery compartment and gallery age as fixed factors, whereas the final model of endo-β-1,4-glucanase, endo-β-1,3(4)-glucanase and endo-protease (casein) only consisted of gallery compartment as the fixed factor. Enzyme activity against the substrates xyloglucan, galactan, rhamnogalacturonan, debranched arabinan and amylose where sporadic across the gallery compartments with many zero activities detected (Fig. 1A, [see Additional file 1]), and these enzyme activities were therefore more appropriately analyzed separately for each age cohort with a non-parametric Kruskal-Wallis test and the significance evaluated using a chi- square distribution [see Additional file 1]. In all samples no enzyme activity against the substrates dextran and collagen were detected (Fig. 1A, [see Additional file 1]). 125

Chapter 8 The plant cell-wall degrading cellulases, endo-xylanases and pectinases had a consistently higher activity in the gallery dump material compared to the entrance tunnel and the brood chamber (Fig. 1A), whereas endo-protease activity against casein showed the opposite trend with the highest enzyme activity in the entrance tunnel (Fig. 1A, [see Additional file 1]). Cellulolytic activity was similar between the entrance tunnel and brood chamber across gallery ages, whereas endo-β-1,4-xylanase (xylan and arabinoxylan) and endo-β-1,4-mannanase activity changed across age cohorts most notably with an increase in enzyme activity in the gallery dump with age (Fig. 1A, [see Additional file 1]). For these three enzymes we also noted a consistent but non-significant trend of high activity in the entrance tunnel compared to the brood chamber at age 45, similar activity at age 62 and the opposite pattern at age 87 with the highest activity in the brood chamber compared to the entrance tunnel (Fig. 1A, [see Additional file 1]). The increased enzyme activity of plant cell-wall degrading enzymes in the gallery dump was also evident from the partial least square regression analysis because these specific enzymes correlated (i.e. clustered) more closely to the gallery dump than either the entrance tunnel or the brood chamber [see Additional file 1]. Screening for specific endo-β-1,4-glucanase and endo-protease (casein) activity did not reveal any enzyme activity of adult and larval fungus-growing beetles (Fig. 2). In contrast, endo-β-1,3(4)- glucanase (beta-glucan) activity were consistently present between 1st, 2nd/3rd instar larvae and adults (Fig. 2). Endo-β-1,4-xylanase activity was detected in both 1st and even more so in 2nd/3rd instar larvae, but not in adult beetles, in which not a single sample of the 14 x 3 beetles measured showed any enzyme activity against xylan (Fig. 2).

Discussion

Glycoside hydrolases are very important enzymes in the fungus-growing ambrosia beetle symbiosis. These enzymes aid in the degradation of the carbohydrate rich cell walls and thereby sustain the conversion of wood that surrounds beetle galleries into fungal biomass. The usage of the comprehensive set of plant cell-wall degrading enzymes generally produced by saprobic fungi [50-52] in combination with mechanical preprocessing of the wood by beetle activities have allowed ambrosia beetles to utilize their present niche deep inside recently dead wood inaccessible to other non-symbiotic fungi and insects that only penetrate the outer bark layers [21]. By using a controlled laboratory rearing method of X. saxesenii fungus-growing ambrosia beetles [53], the present study provides a first glimpse of the intricate biochemical mechanisms whereby beetles with the help of symbionts are able to exclusively live from wood. As expected, plant cell-wall degrading cellulases, endo-xylanases and the pectinolytic endo-β- 1,4-mannanase dominate the enzymatic profile but also consistent endo-protease activity against casein were detected at all measured time-points in all three gallery compartments (Fig. 1A). Taken together the enzymatic profile of the microbial consortium of X. saxesenii ambrosia galleries resembles common wood degrading ascomycete and basidiomycete fungi [50, 51, 54], highlighting the universal similarity of enzymes required in the initial degradation of recently dead wood material. However, the production of extracellular enzymes by filamentous fungi is highly dependent on the growth medium and external conditions like temperature and moisture etc. [50], and it is therefore extremely difficult if not impossible to obtain natural enzyme activity profiles under in-vitro laboratory conditions because the actual micro-habitat experienced by microbes in nature cannot be satisfactorily reproduced in the laboratory. Despite these caveats, the detailed enzymatic measurements of laboratory reared and age-controlled beetle galleries containing all the naturally vectored symbiotic microbes still provide an informative substitute for natural measurements of the usually inaccessible ambrosia beetle galleries deep inside wood. Ambrosia beetle galleries are not static environments because the composition of the associated microbial consortium changes with gallery age [22, 30]. However, despite this gradual turnover in the microbial flora, cellulase and endo-protease activity were remarkably similar at all three stages of gallery development measured (Fig. 1A). In contrast, endo-xylanase activity significantly increased in the brood chamber but decreased in the entrance tunnel with gallery age (Fig. 1A). These distinct changes in endo-xylanase activity most likely reflect that the wood surrounding the gallery walls is in different stages of degradation at the different gallery ages, which thereby influence the set of hydrolytic enzymes required for efficient hydrolysis. However, we are unable to distinguish whether these shifts in enzyme activity is due to changes in endo-xylanase 126

Chapter 8 production by the resident microbes, the succession of microbes in the galleries or beetle activities. In addition, when comparing the endo-β-1,4-glucanase and endo-β-1,4-mannanase activity between compartments within galleries, the entrance tunnel and brood chamber showed remarkably similar enzymatic profiles whereas the gallery dump has much higher activity (Fig. 1A). The expelled sawdust material in the gallery dump material most likely represents a more accessible carbohydrate resource for microbial hydrolysis than the galleries themselves and the observed enzyme activity is likely due to opportunistic bacteria and fungi not necessarily associated with the beetle galleries. Endo-symbionts play a crucial role in nutrient acquisition in many wood-feeding arthropods, like termites or wood-boring beetles [2, 4, 49]. In bark and ambrosia beetles they seem of minor importance because these beetles either feed on nutrient rich phloem, xylomycetophagously on nutrient enriched fungus infested wood or mycetophagously only on fungal tissues. The gut flora of ambrosia beetles has not been studied, but for bark beetles the species richness in larval and adult guts is relatively low [37, 42]. The endo-symbiotic yeasts and bacteria in bark beetles have been shown to detoxify poisonous wood compounds (e.g. tannins [11] and fix nitrogen [55]). However, their role for wood-degradation appears rather small compared to the primary fungal symbionts that are growing within the galleries of Ips and Dendroctonus beetles [37, 42]. For ambrosia beetles, which only feed on fungi and not on woody material, an endogenous production of wood degrading enzymes either by the beetles or associated endo-symbionts is therefore not expected. Hence, the endo-β-1,3(4)-glucanase activity observed in whole-body extracts of X. saxesenii larvae and adults is likely used to degrade the Ophiostomatoid fungal cultivars (Raffaelea sulfurea), which unusually for fungi contain cellulose in their cell walls [27, 56]. However endogenous endo-β-1,3(4)-glucanase production in arthropods is rare and has in certain cases been found to be horizontally acquired from bacteria [57, 58], whereas the alternative explanation of acquisition and utilization of ingested fungal enzymes is well known from several fungal-insect mutualisms (c.f. the acquired enzyme hypothesis [13, 59], table 2). The enzymatic capabilities of ambrosia fungi are unknown, but other bark-beetle associated Ophiostomatoid fungi are known to produce a variety of cellulolytic enzymes [44, 45, 48, 52, 60], although “blue-stain” fungi of which Ophiostoma sp. belong in general leave the cellulose and cross-linking glycans mostly intact and instead utilize storage products in the living ray parenchyma [27]. Unlike X. saxesenii adults, larvae of all species in the ambrosia beetle genus Xyleborinus are xylomycetophagous and ingest both fungal tissue and particles of wood while feeding [31, 32]. Interestingly this difference in nutrition was also reflected by endo-β-1,4-xylanase activity observed in whole-body extracts of larvae, but not of adults (Fig. 2). Again, like endo-β-1,3(4)-glucanase, this endo-β-1,4-xylanase might be fungus derived, but that would imply that the fungus exclusively produces endo-β-1,4-xylanase in the structures eaten by the larvae (e.g. xylanase activity in the mycelial form of Ophiostoma ulmi is higher than in the yeast-like stage [45]) and not by the adults or that the larvae but not the adults avoid internal proteolysis of this enzyme. On the other hand, a handful of beetles has been shown to be capable of endogenous xylanase production, for example larvae and adults of the wood-boring beetle Phaedon cochleariae [61] and the scolytid beetle Hypothenemus hampei [Uniprot: E2J6M9]. Alternatively, endo-β-1,4-xylanase enzymes might be produced by gut endo-symbionts that are either specific to the larvae (larval specific bacteria are known from Ips and Dendroctonus bark beetles [37, 38]), or the endo-symbionts facultatively produce and secrete endo-β-1,4-xylanases depending on context. Irrespective of enzymatic origin, the breakdown of cross-linking glycans within the larval intestinal tract likely (i) has a positive influence on larval nutrition and (ii) is enhanced by active mixing of small woody particles with endo- β-1,4-xylanase and endo-β-1,3(4)-glucanase.

Conclusions

Despite differences in the type of substrate used to cultivate symbiotic fungi a striking, but perhaps not surprising, commonality between the major insect fungus-growing systems is the direct or indirect use of similar fungal carbohydrate active enzymes to utilize recalcitrant plant material as a stable source of food (Table 2). Plant cell-wall degrading xylanases, pectinases and perhaps to a lesser degree cellulases in all cases dominate the enzymatic profiles, although inherent variation between fungus-growing systems are certainly present at the level of specific enzymes. Endogenously produced cellulase enzymes for example are not common among arthropods [62], 127

Chapter 8 which indicates that the provision of essential carbohydrate active fungal enzymes heavily shift the cost/benefit ratio of fungus culturing in favour of symbiotic coevolution. Feeding activity of X. saxesenii larvae not only benefits other group members by creating more space for the ambrosia fungus to form ambrosial layers on the gallery walls, but here we show that it also enhances wood degradation and nutrient cycling. Predigested larval feces, which contains small woody particles and probably also enzymes, is smeared on gallery walls after defecation [33]. The wood particles in this fecal inoculum apparently get further degraded and nitrogenous excretions are recycled by the ambrosia fungi growing there [63]. This may in turn explain the positive effect of larval numbers on group productivity in X. saxesenii [32], and demonstrates a synergism between age groups that prevents competition for fungal food, because adults and larvae feed differently and apparently use a complementary set of enzymes. The differences in the enzymatic capabilities between X. saxesenii larvae and the adults are of high interest for understanding the social system of this species. X. saxesenii is the only primitively eusocial ambrosia beetle described and similarly to the highly eusocial ants, bees and termites exhibit division of labor not only between the sexes, but most importantly also between larval and adult offspring. Differential enzyme activity therefore adds an additional layer of complexity to the behavioural division of labour between adults and larvae. Production of extra enzymes and nutrients by larvae (and their trophallaxis to adults) has also been reported from other social insects, like ants and wasps [64-66], and larvae of the leaf-cutting ant Acromyrmex subterraneous have even been denoted the “digestive caste” of the colony based on the extensive enzymatic machinery detected in their gut lumen [67]. We suggest that larvae in holometabolous insect societies may play a much bigger role in resource utilization than is currently recognized.

Materials and methods

Biological samples X. saxesenii adult females were collected in the Spilwald forest (560 m asl; 46°95’, 7°31’) close to Berne, Switzerland in January 2010, by dissection of galleries from stumps of beech trees (Fagus sylvatica) that had been cut about a year earlier. Individual adult females were brought to the laboratory and sat up in rearing tubes as previously described [53], using a sterile nutrient enriched beech saw-dust media solidified with agar in ~15 mL plastic tubes. Due to their obligate sib-mating and vertical transmission of symbionts, emerging female offspring can be collected from the surface of the media and immediately used for consecutive breeding by placing them onto new media. Galleries used in this study are from the 5th laboratory generation. X. saxesenii galleries typically consist of a straight entrance tunnel dug perpendicular into the media for about 2-5 cm where it reaches a flat brood chamber of 2-3 cm2 and a height of 1 mm (Fig. 1B, S3). As long as there are larvae present, this chamber is continuously expanded by their feeding activity. Thus larvae constantly create fresh substrate and more space for the ambrosia fungus to form ambrosia layers on the chamber walls, which are mainly grazed off by the adult beetles. Larval feces containing woody pieces are smeared on gallery walls, apparently for further degradation and recycling of nutrients by the fungus [33]. Major parts of the finely fragmented sawdust material and feces are, however, disposed through the entrance in a pile of dump on the surface of the media (Fig. 1B). In this study we collected samples from the gallery and the beetles and larvae themselves. The gallery samples were obtained from three different locations and were collected at three different time points and the specific enzyme activity immediately measured for 13 different plant cell wall degrading enzymes (Table 1). Samples contained the entire microbiome (i.e. fungi, yeasts and bacteria) (a) from the gallery dump, and (b) from wall material of the entrance tunnel and (c) the brood chamber. These samples were taken at different time points by dissecting galleries either 45 days, 62 days and 87 days after gallery foundation, when also the number of adults, pupae, 1st and 2nd/3rd instar larvae were counted. These sampling points roughly correspond to different stages in the life-cycle of a gallery: at day 45 there are on average few adults and many 1st and 2nd/3rd instar larvae present in the gallery and the primary fungus in the microbiome at this stage is the Raffaelea sulfurea symbiont [22, 30]; at day 62 there are more adults present on average but fewer larvae, dispersal of adults is just starting and the microbiome has changed and is no longer completely dominated by R. sulfurea but a mixture of several saprobes [22, 30]; and at day 87 almost all beetles present are adults, many adults have left the gallery and the rest is close to leaving. At this stage 128

Chapter 8 gallery walls are heavily melanized and many different species of fungi and yeasts are present [22, 30].

Protein extraction Total proteins were extracted by grinding 30 mg material (wet weight) of entrance tunnel, brood chamber or gallery dump respectively in an Eppendorf tube with 260 µl ddH20 containing 0.1 % Tween20 with a small plastic pestle. Samples were vortexed and centrifuged at 15.000g for 15 min at 4 °C. after which the supernatant was immediately measured for enzyme activity to minimize internal proteolysis. Tween20 was added to the extraction water to keep enzymes in suspension. In total 11, 15 and 8 galleries from day 45, 62 and 87, respectively, were used giving a total sample size of 91 extracts from 34 galleries. 20 gallery dump samples had to be discarded because of insufficient material. In addition, total enzymes were extracted from adult beetles, 1st and 2nd/3rd instar larvae from the gallery cohort dissected at day 45. All individuals were surface sterilized once in bleach and once in 96% alcohol. Thereafter, either three adults, four 2nd/3rd instar larvae or twelve 1st larvae were combined to normalize the amount of biological material to approximately 30 mg biomass and grinded in 60 µl ddH20 containing 0.1 % Tween20.

Enzyme activity measurements Enzyme activity was assayed with Azurine-Crosslinked (AZCL) polysaccharides that are purified polysaccharides cross-linked with a blue dye to form a water insoluble substrate commercially available from Megazyme© (Bray, Ireland) in the form of a powder (table 1). Assay plates were prepared as previously described [68, 69] with a medium consisting of 1% agarose, 23 mM phosphoric acid, 23 mM acetic acid and 23 mM boric acid, mixed and adjusted to pH = 6. The medium was heated using a microwave to melt the agarose. When the medium had cooled to 65°C, 0.1 % weight/volume AZCL substrate wetted in 96% ethanol was added. The medium was then poured into Petri dishes and allowed to solidify before ~4 mm2 wells were made with a cut off pipette tip. 15 µl supernatant of each enzyme extract were applied per well and after 24 hours of incubation at room temperature (ca. 21°C) in the dark the plates were photographed and later analysed by measuring the area of the blue halo surrounding each well with image analysis software (ImageJ ver. 1.37v, W. Rasband, http://rsb.info.nih.gov/ij/). The area surrounding the blue halo is a quantitative measure of enzyme activity that can be compared between samples but does not provide absolute values of enzyme activity [68, 70, 71]. A pilot study showed no activity of either gallery, beetle or larval extracts against the substrates AZCL-pullulan, AZCL-chitosan, AZCL-curdlan, and AZCL-pachyman (data not shown) and these substrates where therefore omitted from this experiment (Table 1).

Data analysis Enzyme activity of the gallery data were ’log + 1’ transformed to improve normalization of the data. Enzyme activity were analyzed separately for each substrate with a mixed linear model with the factorial variables sample (three levels: ‘entrance’, ‘brood chamber’ and ‘gallery dump’) and the interaction between sample and gallery age (three levels: ‘age45’, ‘age62’ and ‘age87’), sample and gallery class (three levels: ‘adult beetles and immatures present’, ‘only immatures present’ and ‘no beetles or larvae present’) and the continuous numerical variable ‘total number of animals in gallery present’ as fixed effects. Because entrance, brood chamber and gallery dump samples from the same gallery are not independent measures each gallery were assigned a code that was included as a random factor in the model. Model estimation was performed with Maximum Likelihood using the lme function implemented in R [72] and each variable evaluated by ANOVA analysis of log-likelihood scores using a step-wise model reduction scheme. Specific means were compared with Tukey’s multiple comparisons of the final model. The correlation of a particular enzyme activity and sample location in the gallery were further analyzed using partial least square regression of a matrix consisting of three x-variables (sample locations: entrance, brood chamber and gallery dump) and 13 y-variables (enzyme activity for each substrate screened) using the R package pls [73]. No statistical analysis were performed on enzyme activities extracted from larvae or beetles because although samples were approximately standardized to the same total biomass, much individual variation and the inherent physiological difference between larval and adult morphology would render the result very ambiguous. 129

Chapter 8

Authors’ contributions HHDFL and PHWB designed the study. PHWB collected and maintained laboratory beetle galleries. HHDFL and PHWB conducted enzyme measurements and HHDFL analyzed the data. HHDFL and PHWB wrote the manuscript in collaboration.

Authors’ information HHDFL is a postdoctoral fellow at the section for microbial ecology at Lund University, Sweden, and studies the ecology and evolution of mutualistic systems. PHWB is Ph.D. student at the Department of Behavioural Ecology at University of Berne, Switzerland. PHWB studies the evolutionary mechanisms of social and mutualistic interactions.

Acknowledgements The authors thank Marko Rohlfs, Anders Tunlid and Morten Schiøtt for helpful comments on a previous version of this manuscript and Jacobus J. Boomsma for providing the AZCL reagents. HHdFL gratefully acknowledges the M. P. Christiansen and wife foundation administered by the Danish Mycological Society for a research travel grant. HHdFL is supported by a fellowship from the Danish Research Council | Natural Sciences and PHWB partially by a DOC fellowship of the Austrian Academy of Sciences and by a fellowship of the Roche Research Foundation.

References 1. Mora C, Tittensor DP, Adl S, Simpson AGB, Worm B: How many species are there on earth and in the ocean. PLoS Biol 2011, 9:e1001127. 2. Bourtzis K, Miller TA: Insect symbiosis. Boca Raton: CRC Press; 2003. 3. Hillis WE: Heartwood and tree exudates. New York: Springer; 1987. 4. Warnecke F, Luginbuhl P, Ivanova N, Ghassemian M, Richardson TH, Stege JT, Cayouette M, McHardy AC, Djordjevic G, Aboushadi N, Sorek R, Tringe SG, Podar M, Martin HG, Kunin V, Dalevi D, Madejska J, Kirton E, Platt D, Szeto E, Salamov A, Barry K, Mikhailova N, Kyrpides NC, Matson EG, Ottesen EA, Zhang XN, Hernandez M, Murillo C, Acosta LG, et al: Metagenomic and functional analysis of hindgut microbiota of a wood-feeding higher termite. Nature 2007, 450:560-U517. 5. Martin MM: Invertebrate-microbial interactions. Ithaca: Cornell University Press; 1987. 6. Grunwald S, Pilhofer M, Holl W: Microbial associations in gut systems of wood- and bark- inhabiting longhorned beetles [Coleoptera: Cerambycidae]. Syst Appl Microbiol 2010, 33:25- 34. 7. Anagnostou C, Dorsch M, Rohlfs M: Influence of dietary yeasts on Drosophila melanogaster life-history traits. Entomol Exp Appl 2010, 136:1-11. 8. Hammond PM, Lawrence JF: Mycophagy in insects: A summary. In Insect-fungal associations. New York: Oxford University Press; 2005: 275-283 9. Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR: The evolution of agriculture in insects. Annu Rev Ecol Syst 2005, 36:563-595. 10. Ayres MP, Wilkens RT, Ruel JJ, Lombardero MJ, Vallery E: Nitrogen budgets of phloem- feeding bark beetles with and without symbiotic fungi. Ecology 2000, 81:2198-2210. 11. Hunt DWA, Borden JH: Conversion of Verbenols to verbenone by yeasts isolated from Dendroctonus ponderosae (Coleoptera, Scolytidae). J Chem Ecol 1990, 16:1385-1397. 12. Dowd PF: Insect fungal symbionts - A promising source of detoxifying enzymes. J Ind Microbiol 1992, 9:149-161. 13. Kukor JJ, Martin MM: Acquisition of digestive enzymes by Siricid woodwasps from their fungal symbiont. Science 1983, 220:1161-1163. 14. Francke-Grosmann H: Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z Morphol Tiere 1956, 45:275-308. 15. Francke-Grosmann H: Ectosymbiosis in wood-inhabiting beetles. In Symbiosis. Edited by Henry SM. New York: Academic Press; 1967: 141-205 16. Six DL: Bark beetle-fungus symbioses. In Insect Symbiosis. Edited by Bourtzis K, Miller TA. Boca Raton: CRC Press; 2003: 97-114

130

Chapter 8 17. Grebennikov VV, Leschen RAB: External exoskeletal cavities in Coleoptera and their possible mycangial functions. Entomol Sci 2010, 13:81-98. 18. Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH, Jordal BH: The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 2001, 55:2011-2027. 19. Hulcr J, Kolarik M, Kirkendall LR: A new record of fungus-beetle symbiosis in Scolytodes bark beetles (Scolytinae, Curculionidae, Coleoptera). Symbiosis 2007, 43:151-159. 20. Beaver RA: Insect-fungus relationships in the bark and ambrosia beetles. In Insect-fungus interactions. Edited by Wilding N, Collins NM, Hammond PM, Webber JF. London: Academic Press; 1989: 121-143 21. Harrington TC: Ecology and evolution of mycophagous bark beetles and their fungal partners. In Insect-fungal associations. New York: Oxford University Press; 2005: 257-295 22. Kajimura H, Hijii N: Dymamics of the fungal symbionts in the gallery system and the mycangia of the ambrosia beetle, Xylosandrus mutilatus (Blandford) (Coleoptera, Scolytidae). Ecol Res 1992, 7:107-117. 23. Kurtzman CP, Robnett CJ: Three new insect-associated species of the yeast genus Candida. Can J Microbiol 1998, 44:965-973. 24. Haanstad JO, Norris DM: Microbial symbiotes of the ambrosia beetle Xyletorinus politus. Microb Ecol 1985, 11:267-276. 25. Grubbs KJ, Biedermann PHW, Suen G, Adams SM, Moeller JA, Klassen JL, Goodwin LA, Woyke T, Munk AC, Bruce D, et al: The complete genome sequence of Streptomyces cf. griseus (XyelbKG-1), an Ambrosia beetle-associated Actinomycete. J Bacteriol 2011, 193:2890-2891. 26. Roeper RA, French JRJ: Ambrosia fungi of the Western United States and Canada - beetle assocaitions (Coleoptera: Scolytidae), tree hosts, and distribution. Northwest Science 1981, 55:305-309. 27. Kirisits T: Fungal associates of European bark beetles with special emphasis on the ophiostomatoid fungi. In Bark and Wood Boring Insects in Living Trees in Europe, a Synthesis. Edited by Lieutier F, Keith RD, Battisti A, Gregoire JC, Evans HF. Dordrecht: Springer; 2004: 181-237 28. Alamouti SM, Tsui CKM, Breuil C: Multigene phylogeny of filamentous ambrosia fungi associated with ambrosia and bark beetles. Mycol Res 2009, 113:822-835. 29. Endoh R, Suzuki M, Okada G, Takeuchi Y, Futai K: Fungus symbionts colonizing the galleries of the Ambrosia beetle Platypus quercivorus. Microb Ecol 2011, 62:106-120. 30. Biedermann PHW, Klepzig KD, Ott E, Taborsky M, Six DL: Dynamics of filamentous fungi in the ambrosia gardens of the primitively eusocial beetle Xyleborinus saxesenii Ratzeburg (Scolytinae: Curculionidae). submitted. 31. Roeper RA: Patterns of mycetophagy in Michigan ambrosia beetles. Michigan Academian 1995, 27:153-161. 32. Biedermann PHW, Taborsky M: Larval helpers and age polyethism in ambrosia beetles. Proc Natl Acad Sci USA 2011:1-6. 33. Hubbard HG: Some miscellaneous results of the work of the division of entomology. In US department of agriculture bureau of entomology bulletin No 7. Edited by Howard LO. Washington, DC.: US department of agriculture; 1897: 9-13. 34. Parkin EA: The digestive enzymes of some woodboring beetle larvae. J Exp Biol 1940, 17:364-377. 35. Balogun RA: Digestive enzymes of alimentary canal of larch bark beetle Ips cembrae Heer. Comp Biochem Physiol 1969, 29:1267-1270. 36. Rivera FN, Gonzalez E, Gomez Z, Lopez N, Hernandez-Rodriguez C, Berkov A, Zuniga G: Gut- associated yeast in bark beetles of the genus Dendroctonus erichson (Coleoptera: Curculionidae: Scolytinae). Biol J Linn Soc 2009, 98:325-342. 37. Delalibera I, Vasanthakumar A, Burwitz BJ, Schloss PD, Klepzig KD, Handelsman J, Raffa KF: Composition of the bacterial community in the gut of the pine engraver, Ips pini (Say) (Coleoptera) colonizing red pine. Symbiosis 2007, 43:97-104. 38. Vasanthakumar A, Delalibera I, Handelsman J, Klepzig KD, Schloss PD, Raffa KF: Characterization of gut-associated bacteria in larvae and adults of the southern pine beetle, Dendroctonus frontalis Zimmermann. Environ Entomol 2006, 35:1710-1717.

131

Chapter 8 39. Valiev A, Ogel ZB, Klepzig KD: Analysis of cellulase and polyphenol oxidase production by southern pine beetle associated fungi. Symbiosis 2009, 49:37-42. 40. Geib SM, Filley TR, Hatcher PG, Hoover K, Carlson JE, Jimenez-Gasco MD, Nakagawa-Izumi A, Sleighter RL, Tien M: Lignin degradation in wood-feeding insects. Proc Natl Acad Sci USA 2008, 105:12932-12937. 41. Schmidt O, Dietrichs HH: Zur Aktivität von Bakterien gegenüber Holzkomponenten. In Organismen und Holz. Edited by Becker G, Liese W. Berlin: Duncker und Humblot; 1976: 91- 102 42. Delalibera I, Handelsman J, Raffa KF: Contrasts in cellulolytic activities of gut microorganisms between the wood borer, Saperda vestita (Coleoptera: Cerambycidae), and the bark beetles, Ips pini and Dendroctonus frontalis (Coleoptera: Curculionidae). Environ Entomol 2005, 34:541-547. 43. Rosch R, Liese W, Berndt H: Studies on enzymes of blue-stain fungi .I. Cellulase-, Polygalacturonase-, Pectinesterase- and Laccase-Activity. Arch Mikrobiol 1969, 67:28-50. 44. Przybyl K, Dahm H, Ciesielska A, Molinski K: Cellulolytic activity and virulence of Ophiostoma ulmi and O. novo-ulmi isolates. Forest Pathol 2006, 36:58-67. 45. Binz T, Canevascini G: Xylanases from the Dutch elm disease pathogens Ophiostoma ulmi and Ophiostoma novo-ulmi. Physiol Mol Plant P 1996, 49:159-175. 46. Binz T, Gremaud C, Canevascini G: Production and purification of an extracellular beta- galactosidase from the Dutch elm disease fungus Ophiostoma novo-ulmi. Can J Microbiol 1997, 43:1011-1016. 47. Beckman CH: Production of Pectinase, Cellulases, and growth-promoting substance by Ceratostomella Ulmi. Phytopathology 1956, 46:605-609. 48. Svaldi R, Elgersma DM: Further studies on the activity of cell wall degrading enzymes of aggressive and non-aggressive isolates of Ophiostoma ulmi. Eur J Forest Pathol 1982, 12:29- 36. 49. Vega FE, Dowd PF: The role of yeasts as insect endosymbionts. In Insect-Fungal Assocations: Ecology and Evolution. Edited by Vega FE, Blackwell M. New York: Oxford University Press; 2005: 211-243 50. Cooke RC, Rayner ADM: Ecology of saprotrophic fungi. Harlow: Longmann Group limited; 1984. 51. Dighton J: Nutrient cycling by saprotrophic fungi in terrestrial habitats. In The Mycota IV: Environmental and microbial relationships. Edited by Esser K, Lemke PA. Berlin: Springer Verlag; 1997: 287-300. 52. Nilsson T: Soft-rot fungi - decay patterns and enzyme production. In Organismen und Holz. Edited by Becker G, Liese W. Berlin: Duncker und Humblot; 1976: 103-112. 53. Biedermann PHW, Klepzig KD, Taborsky M: Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environ Entomol 2009, 38:1096-1105. 54. Dix NJ, Webster J: Fungal Ecology. London: Chapman & Hall; 1995. 55. Bridges JR: Nitrogen-fixing bacteria associated with bark beetles. Microb Ecol 1981, 7:131- 137. 56. Hoog GSD, Scheffer RJ: Ceratocystis versus Ophiostoma: A reappraisal. Mycologia 1984, 76:292-299. 57. Kikuchi T, Shibuya H, Jones JJ: Molecular and biochemical characterization of an endo-β-1,3- glucanase from the pinewood nematode Bursaphelenchus xylophilus acquired by horizontal gene transfer from bacteria. Biochem J 2005, 389:117-125. 58. Song JM, Nam K, Sun YU, Kang MH, Kim CG, Kwon ST, Lee J, Lee YH: Molecular and biochemical characterizations of a novel arthropod endo-beta-1,3-glucanase from the Antarctic springtail, Cryptopygus antarcticus, horizontally acquired from bacteria. Comp Biochem Phys B 2010, 155:403-412. 59. Martin MM, Martin JS: Cellulose digestion in the midgut of the fungus-growing termite Macrotermes natalensis: The role of acquired digestive enzymes. Science 1977, 199:1453- 1455. 60. Tamerler C, Keshavarz T: Lipolytic enzyme production in batch and fed-batch cultures of Ophiostoma piceae and Fusarium oxysporum. J Chem Tech Biotech 2000, 75:785-790.

132

Chapter 8 61. Girard C, Jouanin L: Molecular cloning of cDNAs encoding a range of digestive enzymes from a phytophagous beetle, Phaedon cochleariae. Insect Biochem Molec 1999, 29:1129- 1142. 62. Watanabe H, Tokuda G: Animal cellulases. Cell Mol Life Sci 2001, 58:1167-1178. 63. Norris DM: Chemical interdependence among Xyleborus spp. ambrosia beetles and their symbiotic microbes. Mater Organismen 1975, 3:479-788. 64. Ishay J, Ikan R: Food exchange between adults and larvae in Vespa orientalis F. Anim Behav 1968, 16:298-303. 65. Hunt JH, Baker I, Baker HG: Similarity of amino-acids in nectar and larval saliva - the nutritional basis for trophallaxis in social wasps. Evolution 1982, 36:1318-1322. 66. Hölldobler B, Wilson EO: The Ants. Cambridge: Harvard University Press; 1990. 67. Erthal MJ, Silva CP, Samuels RI: Digestive enzymes in larvae of the leaf cutting ant, Acromyrmex subterraneus (Hymenoptera: Formicidae: Attini). J Insect Physiol 2007, 53:1101-1111. 68. De Fine Licht HH, Schiøtt M, Mueller UG, Boomsma JJ: Evolutionary transitions in enzyme activity of ant fungus gardens. Evolution 2010, 64:2055-2069. 69. Schiøtt M, De Fine Licht HH, Lange L, Boomsma JJ: Towards a molecular understanding of symbiont function: Identification of a fungal gene for the degradation of xylan in the fungus gardens of leaf-cutting ants. BMC Microbiol 2008, 8. 70. Pedersen M, Hollensted M, Lange L, Andersen B: Screening for cellulose and hemicellulose degrading enzymes from the fungal genus Ulocladium. Int Biodeter Biodegr 2009, 63:484- 489. 71. Kooij PW, Schiott M, Boomsma JJ, Licht HHD: Rapid shifts in Atta cephalotes fungus-garden enzyme activity after a change in fungal substrate (Attini, Formicidae). Insec Soc 2011, 58:145-151. 72. R development core team: R: A language and environment for statistical computing. Vienna, Austria; 2011. 73. Mevik BH, Wehrens R: The pls package: principal component and partial least squares regression in R. J stat softw 2007, 18:1-24.

133

Chapter 8 Figures

Figure 1 Glycoside hydrolytic enzyme activity of X. saxesenii ambrosia beetle galleries. A. Enzyme activity of 13 specific carbohydrate active enzymes presented as a heatmap with darker coloration showing higher enzyme activity. Enzyme activity was measured when only immature brood was present, when both immature and adult brood were present, and finally when only adult brood were present (45, 62 and 87 days respectively, after gallery foundation by a single mated female). Enzymes are divided into functional groups according to the plant cell structure functioning as substrate for enzymatic hydrolysis. B. Picture of a X. saxesenii gallery in artificial media around day 45 after gallery foundation. Note the three distinct compartments, where enzyme activities were measured: entrance tunnel, brood chamber, and gallery dump. Many white larvae and a few light brown teneral females are visible in the brood chamber and the lower part of the entrance tunnel.

134

Chapter 8

Figure 2 Endo-β-1,4-glucanase, endo-β-1,3(4)-glucanase, endo-β-1,4-xylanse and endo-protease activity (mean±SE) of adult (n = 14×3 adults), large (2nd/3rd instar, n = 14×4 larvae) and small (1st instar, n = 8×12 larvae) X. saxesenii larvae, respectively. Endo-β-1,3(4)-glucanase activity is present in all three life stages, whereas endo-β-1,4-xylanase activity is not present in adult beetles and only detected in large and small larvae.

Table 1. Insoluble chromogenic substrates used to test for enzyme activity and the specific type of enzymes measured. Substrate Enzyme Starch AZCL-Amylose α-amylase Protein AZCL-Casein endo-protease AZCL-Collagen endo-protease Pectin AZCL-Debr. Arabinan endo-α-1,5-arabinase AZCL-Rhamnogalacturonan rhamnogalacturonanase AZCL-Galactomannan endo-β-1,4-mannanase AZCL-Galactan endo-β-1,4-galactanase Cellulose AZCL-HE-Cellulose cellulase (endo-β-1,4-glucanase) AZCL-Barley β-Glucan cellulase (endo-β-1,3-1,4-glucanase) AZCL-Xyloglucan endo-β-1,4-xyloglucanase Cross-linking Glycans AZCL-Xylan endo-β-1,4-xylanase AZCL-Arabinoxylan endo-β-1,4-xylanase AZCL-Dextran endo-α-1,6-dextranase AZCL = Azurine cross-linked polysaccharides (Megazyme©, Bray, Ireland).

135

Chapter 8 Table 2. Overview of highly derived, obligate nutritional symbioses between insects and fungi. Coleoptera Diptera Hymenoptera Isoptera Ambrosia Bark beetles1 Ship-timber Gall midges Wood wasps Fungus-growing Fungus-growing beetles beetles ants termites Insect family Curculionidae Curculionidae Lymexylidae Cecidomyiidae Xiphydriidae, Formicidae Termitidae Orussidae, Anaxyelidae, Siricidae

Mutualistic Ascomycota Ascomycota Ascomycota Ascomycota Basidiomycota Basidiomycota Basidiomycota fungi (Ambrosiella, (Ophiostoma, (Endomyces) (Lasioptera, (Cerrena, (Leucocoprinus, (Termitomyces) Raffaelea, Ceratocysti Ramichlori Stereum, Leucoagaricus Fusarium) opsis, dium) Amylostereum); and the family Grosmannia, Ascomycota Pterulaceae) Ceratocystis) (Daldinia Basidiomycota decipiens, (Entomocortici Entonaema um) cinnabarina)

Age (Mya) 21–60 ? ? ? ? 45–65 24–34

Agriculture Mode of Multi-species Multi-species Monoculture Monoculture Monoculture Monoculture Monoculture cultivation community community Mode of Wood galleries Phloem Wood tunnels Plant galls Wood tunnels Subterranean Subterranean nesting & chambers galleries & nests (occ. nests and chambers mounds) mounds

Substrate for Surrounding Surrounding Surrounding Surrounding Surrounding Collected plant Collected plant fungi wood phloem (and wood plant tissue wood material material wood) (twigs, (dry leaf litter, caterpillar feces, twigs, wood) leaf litter, flowers, fruits, fresh leaves) Mode of Advanced Primitive Primitive ? Primitive Advanced Advanced agriculture2 (possibly advanced in Dendroctonus )

Enzymatic profile Fungus ? ? Saprotrophism garden (incl. Wood Wood Wood Saprotrophism (plant cell-wall microbial degrading degrading degrading (saprobic and degrading) community) saprotrophism saprotrophism saprotrophism biotrophic in (this study) leaf-cutting ants) Fungus Possible Possible ? ? Present Present Present acquired (this study) (e.g. in enzymes3 Dendroc tonus)

Mode of feeding4 Adults Mycetophagy Phloeo- No food Plant sap No food Mycetophagy, Mycetophagy, mycetophagy (plant material) (plant material) Larvae Mycetophagy Phloeo- Xylo- Mycetophagy Xylo- Mycetophagy Mycetophagy (Xylo- mycetophagy mycetophagy mycetophagy mycetophagy5) 1 Here we only refer to bark beetles in nutritional symbioses with fungi. 2 Primitive fungiculture is defined only by dispersal and seeding of fungi; advanced fungiculture additionally involves active care of the fungal crops (cf. Mueller et al. 2005). 3 Evidence for fungus acquired enzymes that are active in the insect gut or fecal exudates (Martin and Martin 1977; Kukor and Martin 1983). 4 Distinctions originating from the fungus-growing beetle literature: Mycetophagy = eating fungal mycelium, fruiting bodies or specific fungal structures, Phloeomycethophagy = eating phloem and fungal biomass, Xylomycetophagy = eating xylem and fungal biomass. 5 Only in larvae of the genus Xyleborinus and probably Xylosandrus.

136

Chapter 8 SUPPORTING ONLINE MATERIAL

Patterns of functional enzyme activity in the fungus-growing ambrosia beetle mutualism reveal that larvae may be key to successful wood degradation

Henrik H. de Fine Licht1 and Peter H. W. Biedermann2

1Section of Microbial Ecology, Department of Biology, Lund University, Solvegatan 37, SE-223 62 Lund, Sweden, Email: [email protected].

2Department of Behavioural Ecology, Institute of Ecology & Evolution, University of Berne, Baltzerstrasse 6, CH- 3012 Bern, Switzerland, Email: [email protected].

Figure S1. X. saxesenii laboratory gallery enzyme activity (mean±SE) against 13 specific carbohydrate substrates measured at 45, 62 and 87 days after gallery foundation by a single mated female. The same data is presented as a heatmap in figure 1 in the manuscript. For ease of interpretation the polysaccharide substrates are divided into five major groups based on enzyme activity: Cellulases – cellulose, beta-glucan, xyloglucan; Cross-linking glycans – xylan, arabinoxylan, dextran; pectinases – galactomannan, galactan, rhamnogalacturonan, debranched arabinan; Proteases – casein, collagen; and Amylase – amylose. Different letters above horizontal columns indicate significant different means in post-hoc Tukey’s tests. Ns = non-significant, k-w = Kruskal-Wallis test with a * denoting significance at p < 0.05. Sample sizes for entrance tunnel, brood chamber, and gallery dump respectively are: n = 16, n = 16, n = 11 at age 45, n = 11, n = 13, n = 5 at age 62, and n = 8, n = 8, n = 2 at age 87.

Figure S2.

Principal component analyses of gallery enzyme activity at age 45 (A), age 62 (B), and age 87 (C). X loadings are gallery sampling points (brood chamber, entrance tunnel, and gallery dump); Y loadings are specific enzyme activity: cellulos = HE-cellulose (cellulase), betaglu = β-glucan (cellulase / ß- glucanase), xyloglu = xyloglucan (cellulase), xylan = endo-xylanase, axyl = arabinoxylan (endo- xylanase), galman = galactomannan (endo-1,4-ß-mannanase), galac = galactan (endo-1,4-ß-galactanase), debarab = debranched arabinan (endo-arabinase), casein = endo-protease, amyl = amylose (α- amylase). The closer an enzyme substrate is to one of the three samples, the higher the particular enzyme activity is in that sample.

Figure S3.

Picture of a X. saxesenii brood chamber (top) with 1st and 2nd/3rd instar larvae, pupae and teneral females surrounded by lush mycelial growth of the symbiont Raffaelea sulfurea and close-up of two adult females (bottom) inside a gallery with mites. Both pictures are form laboratory galleries and the top picture is taken through the plastic tube containing the gallery. © P. H. W. Biedermann.

137

Chapter 8 Figure S1

138

Chapter 8 Figure S2A

Gallery age = 45

139

Chapter 8 Figure S2B

Gallery age = 62

140

Chapter 8 Figure S2C

Gallery age = 87

141

Chapter 8

Figure S3

142

Chapter 9

Genome Sequence of Streptomyces griseus Strain XylebKG-1, an Ambrosia Beetle-Associated Actinomycete Kirk J. Grubbs,1,2 Peter H. W. Biedermann,1,3 Garret Suen,1,4 Sandra M. Adams,1,4 Joseph A. Moeller,1 Jonathan L. Klassen,1 Lynne A. Goodwin,5,6 Tanja Woyke,5 A. Christine Munk,5,6 David Bruce,5,6 Chris Detter,5,6 Roxanne Tapia,5,6 Cliff S. Han,5,6 and Cameron R. Currie1,4* Department of Bacteriology, University of Wisconsin—Madison, Madison, Wisconsin1; Cellular and Molecular Pathology Graduate Program, University of Wisconsin—Madison, Madison, Wisconsin2; Institute of Ecology and Evolution, Division Behavioral Ecology, University of Berne, Bern, Switzerland3; DOE Great Lakes Bioenergy Research Center, University of Wisconsin - Madison, Madison, Wisconsin4; DOE Joint Genome Institute, Walnut Creek, California5; and Los Alamos National Laboratory, Bioscience Division, Los Alamos, New Mexico6

Received 3 March 2011/Accepted 25 March 2011

Streptomyces griseus strain XylebKG-1 is an insect-associated strain of the well-studied actinobacterial species S. griseus. Here, we present the genome of XylebKG-1 and discuss its similarity to the genome of S. griseus subsp. griseus NBRC13350. XylebKG-1 was isolated from the fungus-cultivating Xyleborinus sax-esenii system. Given its similarity to free-living S. griseus subsp. griseus NBRC13350, comparative genom-ics will elucidate critical components of bacterial interactions with insects.

Streptomyces griseus is a soil bacterium known for its produc- Cheng, unpublished) primer walks. The total size of the ge- tion of secondary metabolites, including streptomycin, the first nome is 8,727,768 bp, and the final assembly is based on 352.4 effective antibiotic for tuberculosis (19). Here, we present the Mbp of 454 sequence (38 X coverage) and 5.3 Gbp of Illumina genome sequence of Streptomyces griseus strain XylebKG-1, sequence (257 X coverage). which, to our knowledge, is the first strain of S. griseus associ- The overall genome G+C content is 72.1%. Generation (http: ated with an insect. XylebKG-1 was isolated from the ambrosia //compbio.ornl.gov/generation/), Glimmer (5), and Critica (ver- beetle Xyleborinus saxeseni, which cultivates a fungus for food sion 1.05) (1) were used to predict a total of 7,265 candidate (2). Actinobacteria-specific isolations from both beetles and their protein-encoding gene models. RNAmmer (13) annotated six 16S fungal galleries resulted in isolation of XylebKG-1. rRNAs and six 23S rRNAs. A tRNAscan-SE (14) search revealed 66 DNA from pure isolates was extracted using a bead-beating tRNAs corresponding to all 20 standard amino acids. Additional protocol (10), and the genome was sequenced at the DOE Joint PRIAM (4), KEGG (12), and COG (18) analyses were completed, Genome Institute (JGI). A noncontiguous finished genome of and the results can be accessed at http://genome.ornl XylebKG-1 was generated using a shotgun approach employing a .gov/microbial/streACT1. combination of Illumina (3) and 454 sequencing technologies Evidence for XylebKG-1 as a strain of S. griseus is based on (15). An Illumina shotgun library (69,927,062 reads totaling ~5.3 similarity between the genomes of XylebKG-1 and the type strain Gbp) and two 454 GS (FLX Titanium) shotgun libraries (688,595 S. griseus subsp. griseus NBRC13350 (16). Both have similar standard reads and 172,566 20-kbp paired-end reads totaling genome sizes (8.7 Mbp to 8.5 Mbp) and G+ C contents (72.1% to ~352 Mbp) were sequenced and assembled. The 454 data were 72.2%), six rRNA operons, and 66 tRNAs. An average nucleotide assembled using Newbler, version 2.3 (Roche), and Illumina identity analysis conducted between both genomes using sequencing data were assembled with Velvet, version 0.7.63 (20). Jspecies (version 1.2.1) (17) revealed a 98.98 ANIb value (91.68% All assemblies were integrated using parallel Phrap, version SPS D genome alignment) and a 98.95 ANIm value (94.13% genome 4.24 (High Performance Software, LLC). Illumina data were used alignment), indi-cating a species level degree of similarity (9). to correct potential base errors and increase consensus quality Given this similarity, XylebKG-1 represents a unique opportunity using the software program Polisher, developed at the JGI (Alla to study genetic elements involved in Actinobacteria-insect as- Lapidus, unpublished). Possible misassembled regions were sociations. corrected using gapResolution (Cliff Han, unpublished) or Nucleotide sequence accession number. The genome se- Dupfinisher (11) or by sequencing cloned bridging PCR fragments quence of Streptomyces griseus strain XylebKG-1 has been de- with subcloning. Gaps between contigs were closed by editing in posited in GenBank under accession no. ADFC00000000. Consed (6-8), by PCR, and by Bubble PCR (J.-F.

This work was funded by the DOE Great Lakes Bioenergy Research * Corresponding author. Mailing address: DOE Great Lakes Bioenergy Center (DOE BER Office of Science DE-FC02-07ER64494), supporting G.S., Research Center and Department of Bacteriology, 6155 MSB, 1550 Linden S.M.A., J.A.M., and C.R.C. This work was also funded by the National Drive, University of Wisconsin - Madison, Madison, WI 53706. Phone: Institute of Food and Agriculture, U.S. Department of Agriculture, under (608) 265-8034. Fax: (608) 262-9865. E-mail: currie @bact.wisc.edu. identification number WISO1321, sup-

143

Chapter 9 porting K.J.G. P.H.W.B. was partly supported by a DOC fellowship 9. Goris, J., et al. 2007. DNA-DNA hybridization values and their relationship to whole- of the Austrian Academy of Sciences at the Department of genome sequence similarities. Int. J. Syst. Evol. Microbiol. 57:81-91. Behavioral Ecology, University of Bern. The work conducted by the 10. Graff, A., and R. Conrad. 2005. Impact of flooding on soil bacterial communities U.S. Department of Energy Joint Genome Institute is supported by associated with poplar (Populus sp.) trees. FEMS Microbiol. Ecol. 53:401-415. the Office of Science of the U.S. Department of Energy under 11. Han, C., and P. Chain. 2006. Finishing repetitive regions automatically with contract no. DE-AC02-05CH11231. Additional support was provided Dupfinisher, p. 142-147. In H. Valafar (ed.), Proceedings of the 2006 International by the National Science Foundation (MCB-0702025). Support for Conference on Bioinformatics Computational Biology, BIOCOMP'06. CSREA Press, Las J.L.K. comes from an NSERC postdoctoral fellowship. Vegas, NV. 12. Kanehisa, M., S. Goto, M. Furumichi, M. Tanabe, and M. Hirakawa. 2010. KEGG for representation and analysis of molecular networks involving diseases and REFERENCES drugs. Nucleic Acids Res. 38:D355-D360. 1. Badger, J. H., and G. J. Olsen. 1999. CRITICA: coding region identification tool invoking 13. Lagesen, K., et al. 2007. RNAmmer: consistent and rapid annotation of comparative analysis. Mol. Biol. Evol. 16:512-524. ribosomal RNA genes. Nucleic Acids Res. 35:3100-3108. 2. Batra, L. R. 1967. Ambrosia fungi: a taxonomic revision, and nutritional studies of 14. Lowe, T. M., and S. R. Eddy. 1997. tRNAscan-SE: a program for improved some species. Mycologia 59:976-1017. detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 3. Bennett, S. 2004. Solexa Ltd. Pharmacogenomics 5:433-438. 25:0955-0964. 4. Claudel-Renard, C., C. Chevalet, T. Faraut, and D. Kahn. 2003. Enzyme-specific profiles 15. Margulies, M., et al. 2005. Genome sequencing in microfabricated high-density for genome annotation: PRIAM. Nucleic Acids Res. 31: picolitre reactors. Nature 437:376-380. 6633-6639. 16. Ohnishi, Y., et al. 2008. Genome sequence of the streptomycin-producing 5. Delcher, A. L., K. A. Bratke, E. C. Powers, and S. L. Salzberg. 2007. Identifying bacterial microorganism streptomyces griseus IFO 13350. J. Bacteriol. 190:4050-4060. genes and endosymbiont DNA with Glimmer. Bioinformat-ics 23:673-679. 17. Richter, M., and R. Rossello-Mora 2009. Shifting the genomic gold standard for the 6. Ewing, B., and P. Green. 1998. Base-calling of automated sequencer traces using prokaryotic species definition. Proc. Natl. Acad. Sci. U. S. A. 106: Phred. II. Error probabilities. Genome Res. 8:186-194. 19126-19131. 7. Ewing, B., L. Hillier, M. C. Wendl, and P. Green. 1998. Base-calling of 18. Tatusov, R. L., M. Y. Galperin, D. A. Natale, and E. V. Koonin. 2000. The COG automated sequencer traces using Phred. I. Accuracy assessment. Genome database: a tool for genome-scale analysis of protein functions and evolution. Res. 8:175-185. Nucleic Acids Res. 28:33-36. 8. Gordon, D., C. Abajian, and P. Green. 1998. Consed: A graphical tool for sequence 19. Waksman, S. A. 1953. Streptomycin: background, isolation, properties, and finishing. Genome Res. 8:195-202. utilization. Science 118:259-266. 20. Zerbino, D. R., and E. Birney. 2008. Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res. 18:821-829.

144

Appendix 1

Fungal associates and their effects on the behaviours and success of the ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae)

Peter H.W. Biedermann 1,2, Michael Taborsky1 and Diana L. Six3

1Department of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland 2USDA Forest Service, Southern Research Station, 2500 Shreveport Hwy, Pineville, LA 71360, USA 3Department of Ecosystem and Conservation Sciences, College of Forestry and Conservation, The University of Montana, Missoula, MT 59812, USA

Corresponding author: Peter H.W. Biedermann, Baltzerstrasse 6, CH-3012 Bern, Switzerland; phone: 0041 31 631 3015; e-mail: [email protected]

Running title: Effects of fungi on ambrosia beetles

Manuscript in work

Note: All fungal species mentioned in this manuscript are not confirmed by molecular sequencing yet!

Abstract

True fungus gardens are found in fungus-farming ants, fungus-farming termites and ambrosia beetles (Scolytinae, Platypodinae). Ant and termite fungus gardens consist of a complex community of fungi, which are under control by the ants that aim to bias the biomass in favor of one cultivar. Fungus garden communities of ambrosia beetles are largely unknown and studies on beetle behaviours in response to different fungi are missing. The role of the different fungal associates for the beetles is unclear, but observations of larval and adult behaviours in response to fungal garden communities can help to reveal fungal roles and behavioural functions.

Using a laboratory breeding and observation technique we revealed a cooperatively breeding system in the xyleborine ambrosia beetle Xyleborus affinis, which is characterized by delayed dispersal of adult daughters, cooperative brood care by larvae and adults, and about half of the totipotent adult daughters laying eggs within the natal nest. Most interesting, egg-laying females are more likely to engage in cooperative behaviours. Furthermore, fungus gardens covering gallery walls composed of five different filamentous fungi, of which a Raffaelea sp. A was predominant and likely served as the main food for adults and larvae. Two species, Mucor sp. and Phaeoacremonium rubrigenum were most abundant in the oldest gallery part close to the entrance, suggesting their parasitic role. Indeed, presence of P. rubrigenum was also negatively correlated with the numbers of females nearby. Presence of fungi affected larval and adult behaviours differently. Adult fungus cropping behaviour, which is feeding but may also serve weeding, was more commonly observed in the presence of Unknown sp. In conclusion, species-diversity is much lower in ambrosia gardens compared to gardens of fungus-farming ants. Behavioural control of fungal weeds may be of minor importance in the ambrosia beetle mutualism.

145

Appendix 1 Introduction

Advanced fungus agriculture by insects has evolved once in ants, once in termites and eight times independently in weevils (Farrell et al. 2001; Hulcr et al. 2007). Latter, combined in the ecological group ambrosia beetles, are apparently especially prone for the evolution of fungus farming. They are regarded to descend from phloem-feeding weevils, which shared their habitat with plant-pathogenic, endophytic and saprobic fungal species (Klepzig and Six 2004). At the beginning, some of these fungi probably used the beetles for getting dispersed to new plant material by producing sticky spores, as the insects have target-oriented dispersal abilities. A nutritional dependence and fungus agriculture may have evolved in the following (cf. the transmission-first model; Mueller et al. 2005). Advanced fungus agriculture by insects today involves by definition (i) the insects’ obligate nutritional dependence on the fungi, (ii) adaptions by the insect for transmission of fungal propagules, and (iii) the active management (farming) of the fungal crops (Mueller et al. 2005).

The active management of fungal crops is dependent on sociality. Ants and termites had been living in eusocial societies (characterized by reproductive division of labor, cooperative brood care, and overlapping generations (Wilson 1971)) already at the origin of fungus agriculture. Ancestors of ambrosia beetles, however, lived solitarily or gregariously, suggesting that active farming must have evolved in association with social evolution (Mueller et al. 2005). One ambrosia beetle is known to be eusocial, living within small family-groups that are composed by a reproducing queen and a few sterile workers (Austroplatypus incompertus; Kent and Simpson 1992). At least one other ambrosia beetle, probably many more, are primitively eusocial, living within family-groups with multiple reproductives, but no permanent sterile worker caste (e.g. Xyleborinus saxesenii; 2012). The remainder of ambrosia beetles are either subsocial (a single female cares for the brood) or communal (several reproductive females cooperate in brood care and farming) (Kirkendall et al. 1997; Biedermann and Taborsky 2012). The level of sociality typically relates to the longevity of the woody resource: The eusocial beetle breeds in living trees – an infinite resource; the habitat of the others range from dying or dead large-diameter trees to small diameter branches. Fungus agriculture is favoured by sociality because of the advantage of division of labour, which is long known to strongly increase work efficiency of social groups (Smith 1776). Division of labour is ubiquitous in the ant and termite farmers and has been recently also found in an ambrosia beetle (Biedermann and Taborsky 2011). Xyleborinus saxesenii adult females engage in protection and maintenance of the nest while larvae overtake gallery enlargement and some hygienic tasks. Larvae are especially important because (i) they increase the wall surface for the fungus gardens to grow, (ii) they suppress the spread of pathogenic fungi, (iii) their feces appears to be recycled by the fungus gardens and (iv) unlike adults they produce xylanases (hemicellulases) for wood and fungus degradation (Biedermann and Taborsky 2011; De Fine Licht and Biedermann 2012).

Ambrosia beetle galleries are of one of three types: (i) the brood-chamber type, (ii) the branching tunnels type and (iii) the branching tunnels type with larval cradles (Biedermann and Taborsky 2012). If these are settled for multiple generations (i.e., overtaken by the daughters) complex gallery patterns result, characterized by tunnels (or brood-chambers) that branch several times and expand dozens of centimeters (up to meters - drilled by beetles that are often less than 3 mm) into the wood (Schedl 1962; Schneider 1987). The pattern is regarded important for social and beetle-fungus interactions: In branching tunnel systems adults, brood and fungus gardens are spatially more separated than in communal brood chambers (Biedermann and Taborsky 2011). Primitive eusociality, division of labour and fungus farming has been recently reported from the brood-chamber type breeding species X. saxesenii (Biedermann and Taborsky 2011), but the social system and fungus farming behaviours of species that construct branching tunnels remain enigmatic. Promising model species are other species in the tribe Xyleborini, because this group of beetles is ubiquitously predisposed for kin-selected sociality because of haplodiploid and inbreeding within the natal nest (Peer and Taborsky 2004; 2005).

In general, healthy fungus gardens of ants, termites and beetles can be viewed as a complex community of microbes that is dominated by the main cultivar fungus, which form the main part of the garden biomass. In ambrosia beetles all developmental stages depend on the main ambrosia fungus as their primary food source (Francke-Grosmann 1956; Batra 1963; Beaver 1989). Progeny complete their life cycle in extensive gallery systems the beetles typically bore within the sap-wood of recently dead host trees. Fungi are fed from a thin layer (= ambrosia gardens) covering the tunnel walls. Apart from bacteria and yeasts (Haanstad and Norris 1985), this layer is composed of different filamentous fungal species 146

Appendix 1 simultaneously: (i) Auxiliary ambrosia fungi, which appear in associations with different host beetles and (ii) the primary ambrosia fungus (usually a single species), which is highly beetle-specific (Batra 1966; Beaver 1989) and cannot survive outside the symbiosis. Most described primary cultivars are from the asexual, ophiostomatoid genera Ambrosiella and Raffaelea, which are not related, but convergently evolved the typical ambrosial growth morphology (i.e. large and nutritional conidiospores, which are grazed by the beetles; (Cassar and Blackwell 1996; Farrell et al. 2001; Rollins et al. 2001; Gebhardt et al. 2004). Characteristic for these species is furthermore the high sensitivity to even short periods of desiccation (Zimmermann and Butin 1973). Therefore during beetle dispersal, when the adult is searching for a new breeding substrate, spores of the primary ambrosia fungus get safely stored either within the gut (Francke-Grosmann 1975) or more commonly within a specialized cuticular pouch, the mycetangium (Francke-Grosmann 1956).

The fungal community of ambrosia gardens has been described for less than a handful of the 3500 ambrosia beetle species worldwide (Kajimura and Hijii 1992; Harrington and Fraedrich 2010; Endoh et al. 2011; Biedermann et al. 2012). In X. saxesenii only two fungal species are intentionally transmitted to new galleries by the beetles, but after egg-laying and offspring development their relative abundance in the fungus gardens decreases (Biedermann et al. 2012). Particular auxiliary species, like Penicillium sp. and Paecilomyces sp., are associated with (i) old galleries, (ii) waste that is removed from the gallery and (iii) dead individuals (Biedermann et al. 2012). An experiment showed that X. saxesenii larvae are able to control the spread of these weeds to some extent (Biedermann and Taborsky 2011). It is unknown however, if specialized behaviours play a role in this process, for example like fungus tending and weeding in fungus-farming ants (Mueller et al. 2001). It is likely because several studies have reported an immediate destruction of fungus gardens by auxiliary molds in response to the removal of beetles (Schneider-Orelli 1913; Francke-Grosmann 1956; Norris 1993). Some auxiliary fungi may also have an nutritional value for the beetles. Fusarium-like sp., for example, are common auxiliary species in nests of fungus-growing ants and ambrosia beetles (Rodrigues et al. 2011), and several studies suggest for particular species a slightly mutualistic function (Norris 1979; Qi et al. 2011; Biedermann et al. 2012). A first step to determine whether a specific fungus has a mutualistic, commensal or pathogenic for the beetles in the symbiosis is to document how fungus abundance relates to beetle numbers and their behaviours. Up to now, however, no study has investigated if adult beetles and their larvae alter their behaviours depending on the composition of the fungus gardens they experience.

Here we used artificial observation tubes that contained entire colonies of reproducing X. affinis beetles and their fungus gardens (i) to determine the social system of an ambrosia beetle breeding in branching- tunnel galleries, which may constrain social interactions, (ii) to describe the fungal community of gardens and relate it to beetle’s breeding success and behaviours. In particular we ask (a) whether offspring produced in a gallery engage in alloparental brood care and fungus maintenance, (b) whether decisions of adult females to help, to breed and to disperse relate to the number of potential beneficiaries, the number of potential competitors and depend on the location within the nest (old vs. freshly excavated gallery parts), and (c) whether reproduction affects the propensity to engage in cooperative behaviours. Ambrosia beetle behaviours have been studied in detail for only a single species yet (X. saxesenii; (Biedermann and Taborsky 2011) and they have not been related to certain gallery compartments and the fungal flora, which is crucial for understanding behavioural functions in more details and the role of different symbionts for the beetles. We should expect a primary mutualistic fungus in the genus Ambrosiella or Raffaelea, as known from other ambrosia beetles (Harrington 2005), to dominate freshly excavated gallery parts, where probably most developing brood is located. Feeding and brood care behaviours might dominate there, whereas hygienic gallery maintenance, gallery protection and courtship might be more frequent in the older entrance part of the nest because of higher expected abundances of auxiliary weeds. Artificial observation tubes (Biedermann et al., 2009) allowed us targeted fungal analyses of gallery compartments of different age, as well as the exact tracking of brood, adult numbers and their behaviours.

147

Appendix 1 Materials & Methods

Study species

The sugar-cane borer Xyleborus affinis (Xyleborini, Scolytinae, Curcilionidae) is one of the most common ambrosia beetles in tropical and sub-tropical zones worldwide. Best known for its damages in sugar-cane plantations it also has been reported to attack more than 248 other woody plant species (Schedl 1962). Depending on the host species, the available woody material and its resistance to degradation single plants may be settled by the beetles for one or several generations. A gallery is always founded by a single female that is attracted by the volatiles (mostly ethanol) of fairly stressed or recently dead plants. While boring an entrance tunnel perpendicular into the wood she will intentionally inseminate its tunnel walls with spores of an unknown ambrosia fungus (possibly Cephalosporium pallidum (Verrall 1943)) she carries in her oral mycetangia (Francke-Grosmann and Schedl 1960; Schneider 1987). If successful, layers (= gardens) of this fungus will appear on the walls and the foundress starts to lay eggs. By feeding solely on the fungi larvae pass through three instars followed by pupation (Biedermann and Taborsky 2012). When reaching adulthood many daughters will not disperse immediately but remain in the natal nest where they probably help in gallery expansion, brood and fungus care, and may overtake breeding after the foundress has died (Biedermann et al. 2011). Thus gallery tunnels with an initial length of about 20-30 cm after one generation may expand over 6 m (within 4 years) by the work of several consecutive overlapping generations (Schneider 1987). Staying and consecutive breeding within one gallery is possible because X. affinis (i) produces its own fungal food and (ii) is an inbreeding species with regular brother-sister mating and a haplodiploid sex determination system. Haplodiploidy enables founder females to assign optimal brood sex-ratios (by laying unfertilized eggs that develop into males and fertilized eggs that develop into females), which are strongly female biased. Ninety-five laboratory galleries have produced on average 19.4 ♀ (range 0 – 86) and 0.96 ♂ (range 0 - 6) offspring in the first generation (Biedermann et al. 2009), and a field study has reported a mean sex ratio of ♀ : ♂ = 41.6 : 1 from 15 galleries (Schedl 1962). Eighty- five percent of laboratory galleries contain only a single male, which is hatching first and obviously capable to fertilize a huge number of sisters (Roeper et al. 1980). Males are flightless and may disperse only by food, what they do after offspring matured (Biedermann, unpublished data).

Beetle collection, laboratory breeding and phenology

Females used for breeding in this study were directly collected from oak logs in Pineville, LA, USA (123 ft asl; 31°20’, 92°24’) in June 2007. Carried in sterile glass vials, they were immediately brought to the laboratory and used for artificial rearing: Females were surface-sterilized (by washing them first for a few seconds with 95% ethanol and then with distilled water) and afterwards singly placed on prepared sterile artificial medium in separate glass tubes (18 mm diameter × 150 mm length; Bellco Glass, Vineland, NJ, USA) that were covered by sterile plastic caps (Bellco Glass kap-uts, Vineland, NJ, USA). Artificial medium had been rested for 5 days and consisted of an autoclaved mixture of 0.35 g streptomycin, 1 g Wesson’s salt mixture, 5 g yeast, 5 g casein, 5 g starch, 10 g sucrose, 20 g agar, 75 g oak tree sawdust (freshly grounded and oven dried), 2.5 ml wheat germ oil, 5 ml 95% ethanol and 500 ml deionized water (for details on the preparation see Biedermann et al. (2009)). After introduction of the beetles, tubes were stored at room temperature (∼23°C) in darkness (wrapped in paper, but light could shine on the entrance). This way, females will start digging tunnels as if in wood, which they often build adjacent to the glass of the tube. Behaviour of adults and brood can be observed in those tunnels when the paper is removed (Biedermann et al. 2009). At 23°C a fungal layer and first eggs appear in successfully founded galleries around 10 days after gallery foundation. The first adults hatch about a month later and females remain in the natal nest for at least a week before they disperse by moving onto the surface of the media and try to fly away (where they can be collected for consecutive breeding). The highest productivity is reached around 60 days after gallery foundation when the first females start to disperse and offspring of all stages are present together. However, in the laboratory more offspring develop typically for further 20- 30 days until the medium dries out and all individuals leave the gallery (Roeper et al. 1980; Biedermann et al. 2011).

148

Appendix 1 Behavioural observations, gallery dissections and fungal isolations

For our study we used 23 successfully founded galleries. Eight of them were excluded after the behavioural observations because they were used in another study. Tunnels were only partly visible in another seven galleries, which were not used for the behavioural observations and just for the dissections and fungal isolations. In summary, we observed the behaviours of individuals in 16 galleries, dissected and isolated fungi of 15 galleries, and could relate behaviours with fungal isolations in eight of them. We started our experiment at about the peak of gallery productivity, when our study galleries ranged between 51 and 60 days of age (after gallery foundation), before females had started to disperse. In the three days before the treatment we conducted daily behavioural observations of all visible individuals. For each individual we noted the developmental stage and sex (larva, adult female, adult male), its location within the gallery (main-tunnel, side-tunnel, brood-tunnel; see Fig. 1) and its behaviour at the moment of observation. We decided between the larval behaviours allogrooming, cropping, cannibalism, locomotion, being pushed by an adult female and inactivity, and the adult behaviours shuffling, blocking, digging, allogrooming, self-grooming, cropping fungus, cannibalism, locomotion, pushing larva, inactivity and the male mating (attempt) (for details see Table S1). On the fourth day we brought galleries individually to a sterile bench, carefully broke the glass of the tubes, and took 12 fungus samples each from the main-tunnel, the side-tunnel and the brood-tunnel with a forceps that was flame-sterilized after each sample. Four samples were placed on malt agar plates (MA:

25 g malt extract, 20 g agar, 1 l deionized H2O), four samples on cycloheximide-streptomycin malt agar plates (CSMA: 10 g malt extract, 15 g agar, 20 ml filter sterilized CSMA stock solution containing 2 mg of cycloheximidin and 1 mg streptomycin, 1 L deionized H2O), and the last four on benomyl malt agar plates (BMA: 10 g malt extract, 15 g agar, 1 ml filter sterilized BMA stock solution containing 1 mg benomyl dissolved in dimethylsulfoxid, 1 L dionized H2O). Samples were spotted in the centre of the plates. Afterwards we fully dissected the whole gallery system and counted all eggs, larvae, pupae, adult females and males.

Identification of filamentous fungus isolates

Isolated fungi were initially placed into groups based on cultural colony characteristics (i.e. morphology and color of mycelium and fruiting structures). Representative samples were used for DNA sequencing (for details follow the protocol in Biedermann et al. (2012)).

Statistical analyses

Using all 12 samples taken from each location (main-, side-, brood-tunnel) of the laboratory gallery (independent variable) we estimated the frequency and presence (yes/no) of each fungal species (dependent variables) by controlling for medium (MA, CSMA, BMA; fixed factors) and gallery of origin (random factor). We also analysed how number of eggs, larvae and adult females (dependent variable) were affected by the location within the gallery (fixed factor) and the presence of the different fungi (fixed factor) by controlling for gallery of origin (random factor). Differences in the frequency of male observations between locations within the gallery were tested using Fisher’s exact test because of the small number of males observed. In a third series of models we analysed whether frequencies of larval and adult behaviours (dependent variable) were influenced by the location within the gallery (fixed factor). Additionally, we first determined whether the frequency of major larval and adult behaviours (dependent variable) were affected by the presence of the different fungi (fixed factor) by controlling for location within the gallery (fixed factor) and gallery of origin (random factor) and second whether each fungal frequency (dependent variable) was affected by the frequency of larval and adult fungus cropping and adult shuffling behaviour (fixed factors) and gallery of origin (random factor). All these analyses were done using generalized estimating equations (GEEs) in R (lmer; Version 2.12.1; R Development Core Team, 2008). GEEs are an extension of generalized linear models with an exchangeable correlation structure of the response variable within a cluster, which allows for controlling the variation between observations from a single gallery. This was necessary because of variation in sample sizes between galleries as a few plates had to be excluded from the analyses (they did not yield any microbial growth). Finally, we used GLMs to model how the proportion of egg-laying and dispersing females and females with developed 149

Appendix 1 ovaries (dependent variables) are affected by the number of brood and of other females (fixed factors) within the nest.

Results

Overlapping generations and factors influencing reproductive division of labour

Eight experimental galleries at an age of about 60 days, shortly after the first generation of offspring had reached adulthood and started to disperse, contained on average 33.6 (± 10 se) eggs, 7.1 (± 2.2) larvae, 5.4 (± 3) pupae, 0.3 (± 0.3) immature females, 6.8 (± 0.9) adult females and 0.9 (± 0.1) adult males. Of the adult females a mean of 26.8% (± 9.7) had non-developed ovaries, 19.2% (± 7.9) had developed ovaries, and 54.1% (± 13.3) were laying eggs (Table S9). All 14 dispersing adult females that were dissected had either non-developed (N = 9) or developed ovaries (N = 5).

Egg-laying was unequally distributed among females and there was no fixed proportion of females laying eggs per gallery: First, egg numbers were independent of the number of potential egg-layers (= females with developed ovaries and egg-laying females) (GLM: p > 0.05) and second the proportion of egg-laying females correlated negatively with number of females with non-developed ovaries (GLM: p = 0.007; Table S7) and developed ovaries (p = 0.009). Proportion of females with developed ovaries tended to be negatively correlated with the number of egg-layers (p=0.054), which might suggest that development of ovaries is triggered by the opportunity to breed. The propensity of females to disperse was reduced with increasing numbers of larvae dependent on brood care (p = 0.019) and females with non-developed ovaries (p = 0.028), but was enhanced by increasing numbers of females with developed ovaries (p = 0.011). Interestingly, we found a trend for adult females with non-developed ovaries (N = 11) to engage less likely in mutually beneficial behaviours (only allogrooming, cropping and shuffling were observed) immediately before galleries were dissected, than females potentially laying eggs (Fisher’s exact test: p = 0.085, N = 21). Additionally, potential egg-layers tended to be found more frequently close to the brood (i.e. in the side-tunnel) than in the main-tunnel compared to females with non-developed ovaries (Chi2- test: χ2 = 2.94, df = 1, p = 0.087, N = 93)

Offspring and adult numbers and their behaviours in relation to the three gallery compartments

Eggs and pupae were only found in the brood-tunnels of the galleries. Larvae were most commonly found in the brood-tunnels, followed by the side-tunnels (GEE: p = 0.03; Table S3, Fig 3), and the main tunnel (p < 0.001). When present in the main-tunnel they only showed locomotion, which was also often observed in the side-tunnel, but only rarely in the brood-tunnel (p < 0.001; Table S4). Larval allogrooming and cropping were most commonly detected in the brood-tunnels (p < 0.05). Larval cannibalism was only present at low rates in the side tunnel.

Adult females and males were least commonly observed in the brood-tunnels (p < 0.05; Table S3, Fig. 4). Males tended to stay in the side tunnels (p = 0.064). The most frequent adult female behaviours were shuffling and fungus cropping, with the last most common in the brood-tunnel (p = 0.001; Table S5, Fig. 4). Inactivity tended to be was most commonly shown in the main-tunnel (p < 0.066). Blocking was, by definition, only found in the main tunnel, whereas digging, cannibalism and pushing others was never found there. All other female behaviours were equally common in all three gallery compartments (p > 0.05). Adult males spent their time mainly with cropping (27% of their time), locomotion (25%), inactivity (16%), followed by attempting to mate (11%), and allogrooming (11%). Self-grooming, mating, being pushed and shuffling were only rarely shown by males (< 4%).

Fungal isolations

Our isolations revealed five species of filamentous fungi associated with Xyleborus affinis. Two of these fungi, Mucor sp. and Raffaelea sp., were only morphologically identified, because unfortunately they could not be revived after cool storage before sequencing. Sequencing revealed the three fungi Fusarium 150

Appendix 1 sp., Phaeoacremonium rubrigenum and Unknown. As expected (Roeper and French 1981), an unknown Raffaelea sp. was most commonly isolated from X. affinis. Detailed description of sequencing results by Diana.

Overall, Raffaelea sp., P. rubrigenum and Unknown were more commonly detected on CSMA than MA (GEEs: p = 0.017-0.001; Table S2), which shows that these three species are resistant to the antibiotic Cycloheximide. Interestingly they grew also better on BMA than MA (p = 0.007-0.001). The opposite was found in Mucor sp. and Fusarium sp., which were most frequently isolated from MA followed by BMA and CSMA (p < 0.01).

Fungal diversity in relation to gallery compartment and age of gallery

The Raffaelea sp. was isolated from all compartments of all galleries, and was most commonly detected in the brood-tunnel (GEEs: p = 0.003-0.001; Table S2, Fig 2 and S1), which strongly suggests its nutritional role for X. affinis. Fusarium sp., Mucor sp. and Unknown were present in about 60% of all galleries. Unknown and Fusarium sp. were isolated equally often from all compartments (p > 0.05), whereas Mucor sp. and P. rubrigenum occurred more frequently near the entrance than the brood (p = 0.003-0.006). Fusarium sp. and Mucor sp. were strongly associated with old galleries as well as dead females (i.e., isolation rate > 60%; Fig. S1). The other three fungi were detected in < 20% of samples from these specimens (except Raffaelea sp., which was also present in > 80% of old galleries).

Influence of fungal species on beetle productivity and their farming behaviours

The regular presence of Raffaelea sp. in samples from all compartments of all galleries supports its essential function for the beetles, but inhibited to test the influence of its presence on beetle productivity and behaviours. It is likely to be the only fungus with a nutritional values for the brood, because egg and larval numbers were both not affected by any of the other species (GEEs: p > 0.05; Table S3). In contrary, adult female numbers were negatively affected by the presence of P. rubrigenum (p = 0.005). Adult females increased cropping frequency in the presence of Unknown (p = 0.038; Table S6, Fig 5), increased shuffling by decreasing inactivity levels in the presence of Fusarium sp. (p = 0.048-0.02) and increased inactivity levels in the presence of P. rubrigenum (p = 0.002). None of the fungal species influenced the activity of larvae, their cropping and shuffling behaviours (p > 0.05; data not shown).

Influence of larval and adult farming behaviours on the isolation frequency of the fungi

Raffaelea sp. were more commonly detected in galleries, which larvae had higher inactivity levels in the three days before the isolations (p < 0.05; Table S6). Unknown was positively affected by the larval cropping frequency (p = 0.004). Adult fungus cropping, shuffling or inactivity levels had no significant effect on abundance of the five fungi. There was, however, a trend for a higher isolation frequency of Mucor sp. and P. rubrigenum sp. from galleries, where adults showed higher inactivity rates in the days before (p < 0.12).

Discussion

Here we report correlative evidence for ambrosia beetles to actively manage their fungus gardens, i.e. behavioural changes in response to the fungal species composition of their ambrosia gardens. For each of the three X. affinis gallery compartments – brood-tunnel, side-tunnel, and main-tunnel – we give an overview of the number of larval and adult inhabitants, their respective behaviours and the diversity of filamentous fungi. Ambrosia beetle behaviour normally cannot be observed within solid wood, so here we rear beetles in artificial medium in glass tubes (Biedermann et al. 2009). That gave us the chance for the first targeted fungal samplings in combination with behavioural observations.

151

Appendix 1 Observations and female dissections revealed primitive eusociality in X. affinis, which was characterized by adult daughters remaining and helping, with only half of them reproducing in the natal gallery. The number of egg-laying daughters correlated positively with the total number of adults present in the natal nest. Interestingly, females that carried eggs in their ovaries tended to engage more frequently in social behaviours immediately before the dissection. They were cropping their fungal gardens, for example, which we found to be predominated by one ambrosia fungus, a Raffaelea sp.; it was invariably associated with galleries of X. affinis. This species formed thick ambrosia layers on the walls of the side- and brood- tunnels, which were fed upon by larvae and adults. Behavioural observations suggested Unknown to function probably as a secondary food source and three other isolates – Fusarium sp., Phaeoacremonium rubrigenum and Mucor sp. – to have negative effects for the beetles in the symbiosis. Raffaelea sp. was typically associated with freshly excavated brood-tunnels were individuals preferably fed, whereas the three proposed pathogens were predominantly found in the main-tunnels, old galleries and associated with dead individuals. P. rubrigenum had negative effects on the number of adult females present, and the other three either increased their shuffling (Fusarium sp.) or cropping intensity (Mucor sp., Unknown).

The social structure within branching-tunnel galleries of X. affinis

X. affinis galleries contained high numbers of individuals of all age groups at around day 60, which is shortly after adult daughters start to disperse from galleries (Biedermann et al. 2011). Multiple egg-layers were already present and on average about half of the adult daughters that had remained in the natal nest had started to lay eggs. Although all adult females participate in social tasks in the natal nest, egg- laying females were more inclined to do so. This higher investment of egg-laying individuals into brood care than of non-reproducing helpers is typical for social groups that have not taken the transition to eusociality. Reproducing females in the primitive eusocial Xyleborinus saxesenii, for example, typically protect the gallery entrance against the introduction of predators, which is the most risky task in an ambrosia beetle society (Biedermann and Taborsky 2011). By contrast, after the evolutionary transition to eusociality, i.e. the evolution of a reproductive and a sterile worker caste, non-egg-layers (=workers) are not only overtaking the most dangerous task but usually all the work that needs to be done (Wilson 1971; Bourke 2011).

Larvae are predominantly found in the brood-tunnels, whereas adult females and males prefer to stay in the side-tunnels. Males are waiting in the side tunnels for freshly emerging females to mate with.

Literature

Batra LR (1963) Ecology of ambrosia fungi and their dissemination by beetles. Trans. Kansas Acad. Sci. 66: 213-236.

Batra LR (1966) Ambrosia fungi: extent of specifity to ambrosia beetles. Science 153: 193-195.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW, Klepzig KD & Taborsky M (2011) Costs of delayed dispersal and alloparental care in the fungus- cultivating ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae). Behav. Ecol. Sociobiol. 65: 1753- 1761.

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

152

Appendix 1 Biedermann PHW & Taborsky M (2012) Social fungus farming varies among ambrosia beetles. in prep.

Bourke AFG (2011) Principles of Social Evolution. Oxford University Press, Oxford.

Cassar S & Blackwell M (1996) Convergent origins of ambrosia fungi. Mycologia 88: 596-601.

De Fine Licht HH & Biedermann PHW (2012) Patterns of functional enzyme activity suggest that larvae are the key to successful fungus farming by ambrosia beetles. Frontiers in Zoology in review.

Endoh R, Suzuki M, Okada G, Takeuchi Y & Futai K (2011) Fungus Symbionts Colonizing the Galleries of the Ambrosia Beetle Platypus quercivorus. Microbial Ecology 62: 106-120.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z. Morph. u. Ökol. Tiere 45: 275-308.

Francke-Grosmann H (1975) Zur epizoischen und endozoischen Übertragung der symbiotischen Pilze des Ambrosiakäfers Xyleborus saxeseni (Coleoptera: Scolitidae). Entomologica Germanica 1: 279-292.

Francke-Grosmann H & Schedl W (1960) Ein orales Übertragungsorgan der Nährpilze bei Xyleborus mascarensis Eichh (Scolytidae). Naturwissenschaften 47: 405.

Gebhardt H, Begerow D & Oberwinkler F (2004) Identification of the ambrosia fungus of Xyleborus monographus and X. dryographus. Mycol. Progr. 3: 95-102.

Haanstad JO & Norris DM (1985) Microbial symbiotes of the ambrosia beetle Xyletorinus politus. Microbial Ecology 11: 267-276.

Harrington TC (2005) Ecology and evolution of mycophagous bark beetles and their fungal partners. Insect-fungal associationspp. 257-295. Oxford University Press, New York.

Hulcr J, Kolarik M & Kirkendall LR (2007) A new record of fungus-beetle symbiosis in Scolytodes bark beetles (Scolytinae, Curculionidae, Coleoptera). Symbiosis 43: 151-159.

Kent DS & Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften 79: 86-87.

Klepzig KD & Six DL (2004) Bark beetle-fungal symbiosis: Context dependency in complex associations. Symbiosis 37: 189-205.

Mueller UG, Gerardo NM, Aanen DK, Six DL & Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563-595.

Mueller UG, Schultz TR, Currie CR, Adams RMM & Malloch D (2001) The origin of the attine ant-fungus mutualism. Q. Rev. Biol. 76: 169-197.

Norris DM (1979) The mutualistic fungi of Xyleborini beetles. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 53-63. Allanheld, Osmun & Company, Montclair.

Norris DM (1993) Xyleborus ambrosia beetles - a symbiotic ideal extreme biofacies with evolved polyphagous privileges at monophagous prices. Symbiosis 14: 229-236. 153

Appendix 1 Peer K & Taborsky M (2004) Female ambrosia beetles adjust their offspring sex ratio according to outbreeding opportunities for their sons. J. Evol. Biol. 17: 257-264.

Peer K & Taborsky M (2005) Outbreeding depression, but no inbreeding depression in haplodiploid ambrosia beetles with regular sibling mating. Evolution 59: 317-323.

Roeper RA & French JRJ (1981) Ambrosia fungi of the Western United States and Canada - beetle assocaitions (Coleoptera: Scolytidae), Tree Hosts, and Distribution. Northwest Science 55: 305-309.

Rollins F, Jones KG, Krokene P, Solheim H & Blackwell M (2001) Phylogeny of asexual fungi associated with bark and ambrosia beetles. Mycologia 93: 991-996.

Schedl KE (1962) Scolytidae und Platypodidae Afrikas, II+III. Revista de Entomologia de Mocamique 5: 1-1352.

Smith A (1776) An Inquiry into the Nature and Causes of the Wealth of Nations. W.Strahan and T.Cadell, London.

Verrall AF (1943) Fungi associated with certain ambrosia beetles. Journal of Agricultural Research 66: 135-144.

Wilson EO (1971) The insect societies. Belknap Press of Harvard University Press, Cambridge.

Zimmermann G & Butin H (1973) Untersuchungen über die Hitze- und Trockenresistenz holzbewohnender Pilze. Flora 162: 393-419.

154

Appendix 1 Graphs

Fig.1. Morphology of a X. affinis laboratory gallery in artificial medium at about 60 days after its foundation. Note the species characteristic tunnel system with the four compartments: S – surface, MT – main tunnel, ST – side tunnel, BT – brood tunnel. A white fungal layer and white larvae are visible within the ST and BTs. Several adult beetles are sitting in the MT.

Fig. 2. Percentage of fungal species isolated relative to total number of samples taken per gallery compartment. Mean ± SE; N = 12 samples per compartment; main-tunnel – 8 galleries, side – tunnel – 6 galleries, brood-tunnel – 7 galleries.

155

Appendix 1

Fig. 3. Behaviours of X. affinis larvae in the three gallery compartments. Mean ± se; different letters denote significant differences (p < 0.05; for exact values see Table 3) between compartments. The mean (± se) numbers of larvae observed in the three compartments are given in the box on the right.

Fig. 4. Behaviours of X. affinis adult females in the three gallery compartments. Mean ± se; different letters denote significant differences (p < 0.05; for exact values see Table 4) between compartments. The mean (± se) numbers of adult females observed in the three compartments are given in the box on the right.

156

Appendix 1

Fig. 5. Effect of the presence of the five different fungi on the frequency of adult female cropping. Mean (± se) frequencies, for fungus present or not per gallery, are given. Raffaelea sp. A was always present. * - p < 0.05 (for exact values see Table 5).

157

Appendix 1

Supplementary Information

Table S1. Ethogram of the observed behaviours of X. affinis larvae (L), females (F), and males (M) (modified from Biedermann and Taborsky, 2011).

Behaviour Shown by Definition Mutual benefit Digging F excavating new tunnels gallery extension Cropping L, F, M grazing on the fungal layer covering gallery walls with fungus care (?) the maxillae and/or mandibles Shuffling F, M moving frass and sawdust with the legs and elytra hygiene Cannibalism L, F, M feeding on a larva, pupa or adult beetle that is usually hygiene (?) dead Allogrooming L, F, M grooming an egg, larva, pupa or adult beetle with the brood care, hygiene mouthparts (i.e., maxillae, labium) Blocking F staying inactive in the entrance tunnel and plugging it protection with the body (abdomen directed to the outside) Pushing others F shifting a larva or male within a tunnel protection Self-grooming F, M grooming oneself with the legs - Inactive L, F, M being inactive without moving - Locomotion L, F, M creeping (L), or walking on the tibia with back-folded - tarsi (F, M) Being pushed L, M getting shifted by an adult female - Mating (attempt) M mounting a female or copulating with her -

158

Appendix 1 Table S2. Influence of culture media and origin of sample (gallery compartment) upon the abundance of fungi associated with Xyleborus affinis.

Fungal species Parameter coeff. ± se z p

Mucor sp. Intercept (BMA) -2.16 ± 0.86 -2.2 0.012 Contrast BMA vs. MA -0.23 ± 0.51 -0.44 0.66 Contrast BMA vs. CSMA -1.56 ± 0.61 -2.59 0.01 Contrast MA vs. CSMA -1.34 ± 0.42 -3.22 0.001

Intercept (main tunnel) -1.88 ± 0.74 -2.55 0.011 Contrast main vs. side tunnel -0.41 ± 0.38 -1.08 0.28 Contrast main vs. brood-tunnel -1.78 ± 0.46 -3.83 <0.001 Contrast side vs. brood-tunnel -1.37 ± 0.46 -3.1 0.003

Raffaelea sp. A Intercept (BMA) 1.19 ± 0.38 3.1 0.002 Contrast BMA vs. MA -1.49 ± 0.35 -4.27 <0.001 Contrast BMA vs. CSMA 0.08 ± 0.39 0.21 0.83 Contrast MA vs. CSMA 1.57 ± 0.3 5.26 <0.001

Intercept (main tunnel) 0.36 ± 0.28 1.26 0.21 Contrast main vs. side tunnel -0.23 ± 0.28 -0.82 0.41 Contrast main vs. brood-tunnel 0.44 ± 0.29 1.55 0.12 Contrast side vs. brood-tunnel 0.67 ± 0.29 2.35 0.019

Fusarium sp. Intercept (BMA) -2.03 ± 0.73 -2.76 0.006 Contrast BMA vs. MA 1.25 ± 0.41 3.04 0.002 Contrast BMA vs. CSMA -1.38 ± 0.53 -2.6 0.009 Contrast MA vs. CSMA -2.63 ± 0.5 -5.24 <0.001

Intercept (main tunnel) -1.3 ± 0.63 -2.07 0.039 Contrast main vs. side tunnel -0.36 ± 0.36 -0.98 0.33 Contrast main vs. brood-tunnel -0.25 ± 0.35 -0.7 0.49 Contrast side vs. brood-tunnel 0.11 ± 0.36 0.3 0.76

Phaeoacremonium rubrigenum Intercept (BMA) -2.33 ± 0.52 -4.52 <0.001 Contrast BMA vs. MA -1.85 ± 0.68 -2.71 0.007 Contrast BMA vs. CSMA -0.27 ± 0.63 -0.42 0.67 Contrast MA vs. CSMA 1.59 ± 0.66 2.39 0.017

Intercept (main tunnel) -2.16 ± 0.41 -5.29 <0.001 Contrast main vs. side tunnel -1.58 ± 0.61 -2.59 0.01 Contrast main vs. brood-tunnel -1.68 ± 0.61 -2.75 0.006 Contrast side vs. brood-tunnel -0.1 ± 0.76 -0.13 0.9

Unknown fungus Intercept (BMA) -4.67 ± 0.87 -5.37 <0.001 Contrast BMA vs. MA Only present on BMA Contrast BMA vs. CSMA 3.11 ± 0.84 3.73 <0.001 Contrast MA vs. CSMA Only present on CSMA

Intercept (main tunnel) -3.74 ± 0.63 -5.94 <0.001 Contrast main vs. side tunnel 0.07 ± 0.58 0.12 0.91 Contrast main vs. brood-tunnel 0.82 ± 0.5 1.63 0.1 Contrast side vs. brood-tunnel 0.75 ± 0.52 1.44 0.15 Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in the isolation frequency between the culture media (BMA – benomyl-malt agar vs. MA – malt-extract agar vs. CSMA – cycloheximide- streptomycin-malt agar) and locations within the gallery (main tunnel vs. side tunnel vs. brood tunnel). Model coefficients are reported as coeff. ± se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of independent variables on the fungal frequencies are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse.

159

Appendix 1

Fig. S1. Presence of fungal species in the three gallery compartments, old galleries and associated with dead adult females. Twelve samples were taken from each compartment and from old galleries; four samples from dead females. Presence was recorded as yes or no.

160

Appendix 1 Table S3. Influence of gallery compartment and presence of the different fungi upon the number of inhabitants.

Offspring numbers Parameter coeff. ± se z p

Eggs Intercept -15.2 ± 999 0.0 1 Presence of Mucor sp. -1.07 ± 1.03 -1.04 0.36 Presence of Raffaelea sp. A always present Presence of Fusarium sp. 19.03 ± 999 0 1 Presence of P. rubrigenum -1.18 ± 1.15 -1.02 0.36 Presence of Unknown fungus 0.17 ± 0.69 0.25 0.82

Larvae Intercept (main tunnel) -1.26 ± 0.86 -1.46 0.14 Contrast main vs. side-tunnel 2.57 ± 0.79 3.24 0.001 Contrast main vs. brood-tunnel 3.55 ± 0.86 4.15 <0.001 Contrast side vs. brood-tunnel 0.98 ± 0.45 2.18 0.03 Presence of Mucor sp. -0.64 ± 0.45 -1.42 0.16 Presence of Raffaelea sp. A always present Presence of Fusarium sp. -0.15 ± 0.35 -0.43 0.67 Presence of P. rubrigenum 0.07 ± 0.56 0.12 0.91 Presence of Unknown fungus -0.04 ± 0.54 -0.08 0.94

Adult ♀♀ Intercept (main tunnel) 2.8 ± 0.32 8.77 <0.001 Contrast main vs. side-tunnel 0.16 ± 0.18 0.9 0.37 Contrast main vs. brood-tunnel -0.41 ± 0.21 -1.93 0.054 Contrast side vs. brood-tunnel -0.57 ± 0.21 -2.71 0.007 Presence of Mucor sp. -0.13 ± 0.23 -0.54 0.59 Presence of Raffaelea sp. A always present Presence of Fusarium sp. 0.08 ± 0.22 0.39 0.69 Presence of P. rubrigenum -0.58 ± 0.21 -2.8 0.005 Presence of Unknown fungus -0.25 ± 0.24 -1.02 0.31

Adult ♂♂ Fisher’s exact test (N = 17 observations) Contrast main (N = 3) vs. side-tunnel (N = 12) 0.064 Contrast main (N = 3) vs. brood-tunnel (N = 2) 1 Contrast side (N = 12) vs. brood-tunnel (N = 2) 0.026 Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in the number of eggs, larvae, adult females and males between the three gallery compartments and depending on the presence of the five fungal species. Eggs were only found in the brood-tunnel. Model coefficients are reported as coeff. ± se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of independent variables on the fungal frequencies are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse.

161

Appendix 1 Table S4. Influence of gallery compartment upon the behaviour of larvae. Proportion of larvae Parameter coeff. ± se z p

Allogrooming1 Intercept (side-tunnel) -2.51 ± 0.53 -4.76 <0.001 Contrast side vs. brood-tunnel 1.23 ± 0.56 2.2 0.028

Cropping1 Intercept (side-tunnel) -1.57 ± 0.39 -4.06 <0.001 Contrast side vs. brood-tunnel 1.02 ± 0.41 2.47 0.013

Cannibalism1 Intercept (side-tunnel) -3.93 ± 1.01 -3.89 <0.001 Contrast side vs. brood-tunnel Only present in side-tunnel

Inactivity1 Intercept (side-tunnel) -2.25 ± 0.68 -3.29 0.001 Contrast side vs. brood-tunnel 0.79 ± 0.59 1.34 0.18

Locomotion Intercept (brood-tunnel) -2.07 ± 0.37 -5.53 <0.001 Contrast main vs. side tunnel Only behaviour in main-tunnel Contrast main vs. brood-tunnel Only behaviour in main-tunnel Contrast side vs. brood-tunnel -2.17 ± 0.41 -5.26 <0.001

Being pushed by adult1 Intercept (brood-tunnel) -4.61 ± 0.83 -5.56 <0.001 Contrast side vs. brood-tunnel Only present in brood-tunnel Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in behavioural frequencies between the three gallery compartments. Model coefficients are reported as coeff. ± se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of independent variables are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse. 1 Not present in main-tunnel.

162

Appendix 1 Table S5. Influence of gallery compartment upon the behaviour of adult females. Proportion of adult ♀♀ Parameter coeff. ± se z p

Shuffling Intercept (side-tunnel) -0.6 ± 0.14 -4.14 <0.001 Contrast main vs. side tunnel -0.42 ± 0.22 -1.86 0.062 Contrast main vs. brood-tunnel -0.04 ± 0.31 -0.12 0.9 Contrast side vs. brood-tunnel 0.46 ± 0.3 1.52 0.13

Blocking Intercept (main-tunnel) -2.42 ± 1.32 -1.84 0.066 Contrast main vs. side tunnel Only present in main-tunnel Contrast main vs. brood-tunnel Only present in main-tunnel

Digging Intercept (side-tunnel) 0.0 ± 1.41 0.0 1 Contrast main vs. side tunnel Only present in side-tunnel Contrast main vs. brood-tunnel Only present in brood-tunnel Contrast side vs. brood-tunnel 24.6 ± 999 0.0 1

Allogrooming Intercept (side-tunnel) -1.5 ± 0.23 -6.5 <0.001 Contrast main vs. side tunnel -0.11 ± 0.42 -0.27 0.79 Contrast main vs. brood-tunnel 0.56 ± 0.48 1.17 0.24 Contrast side vs. brood-tunnel 0.67 ± 0.39 1.7 0.089

Self-grooming Intercept (side-tunnel) -1.83 ± 0.36 -5.09 <0.001 Contrast main vs. side tunnel -0.51 ± 0.48 -1.05 0.3 Contrast main vs. brood-tunnel 0.22 ± 1.2 0.19 0.85 Contrast side vs. brood-tunnel 0.73 ± 1.21 0.6 0.55

Cropping fungus Intercept (main-tunnel) -1.18 ± 0.29 -4.12 <0.001 Contrast main vs. side tunnel 0.19 ± 0.33 0.57 0.57 Contrast main vs. brood-tunnel 1.18 ± 0.36 3.23 0.001 Contrast side vs. brood-tunnel 0.99 ± 0.27 3.57 <0.001

Cannibalism Intercept (brood-tunnel) -1.39 ± 0.79 -1.75 0.08 Contrast main vs. side tunnel Only present in side-tunnel Contrast main vs. brood-tunnel Only present in brood-tunnel Contrast side vs. brood-tunnel -0.29 ± 1.14 -0.25 0.8

Inactivity Intercept (side-tunnel) -1.13 ± 0.23 -4.88 <0.001 Contrast main vs. side tunnel -0.89 ± 0.3 -3.0 0.003 Contrast main vs. brood-tunnel -0.77 ± 0.42 -1.84 0.066 Contrast side vs. brood-tunnel 0.12 ± 0.43 0.27 0.78

Locomotion Intercept (brood-tunnel) 0.0 ± 1.41 0 1 Contrast main vs. side tunnel -0.23 ± 0.32 -0.7 0.48 Contrast main vs. brood-tunnel 0.77 ± 1.44 0.53 0.59 Contrast side vs. brood-tunnel 1.0 ± 1.42 0.7 0.48

Pushing others Intercept (brood-tunnel) -1.61 ± 1.1 -1.47 0.14 Contrast main vs. side tunnel Only present in side-tunnel Contrast main vs. brood-tunnel Only present in brood-tunnel Contrast side vs. brood-tunnel -0.54 ± 1.19 -0.46 0.65 Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in behavioural frequencies between the three gallery compartments. Model coefficients are reported as coeff. ± se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of independent variables are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse.

163

Appendix 1 Table S6. Influence of gallery compartment and presence of the different fungi upon the frequency of particular adult female behaviours. Proportion of adult ♀♀ Parameter coeff. ± se z p

Cropping Intercept (main tunnel) -2.77 ± 0.52 -5.33 <0.001 Contrast main vs. side-tunnel 1.05 ± 0.45 2.34 0.019 Contrast main vs. brood-tunnel 1.49 ± 0.46 3.21 0.001 Contrast side vs. brood-tunnel 0.45 ± 0.42 1.06 0.29 Presence of Mucor sp. 0.68 ± 0.39 1.73 0.084 Presence of Raffaelea sp. A always present Presence of Fusarium sp. 0.22 ± 0.38 0.59 0.56 Presence of P. rubrigenum -0.41 ± 0.43 -0.95 0.34 Presence of Unknown fungus 0.94 ± 0.46 2.07 0.038

Shuffling Intercept (main tunnel) -0.9 ± 0.42 -2.13 0.033 Contrast main vs. side-tunnel -0.63 ± 0.34 -1.87 0.061 Contrast main vs. brood-tunnel -0.68 ± 0.39 -1.74 0.082 Contrast side vs. brood-tunnel -0.04 ± 0.42 -0.11 0.92 Presence of Mucor sp. 0.44 ± 0.37 1.2 0.23 Presence of Raffaelea sp. A always present Presence of Fusarium sp. 0.71 ± 0.36 1.98 0.048 Presence of P. rubrigenum -0.11 ± 0.31 -0.35 0.72 Presence of Unknown fungus 0.27 ± 0.37 0.72 0.47

Inactive Intercept (main tunnel) -1.13 ± 0.79 -1.43 0.15 Contrast main vs. side-tunnel 0.07 ± 0.51 0.14 0.89 Contrast main vs. brood-tunnel -0.52 ± 0.61 -0.85 0.39 Contrast side vs. brood-tunnel -0.59 ± 0.67 -0.88 0.38 Presence of Mucor sp. -0.47 ± 0.64 -0.73 0.47 Presence of Raffaelea sp. A always present Presence of Fusarium sp. -1.38 ± 0.6 -2.32 0.02 Presence of P. rubrigenum 1.79 ± 0.57 3.14 0.002 Presence of Unknown fungus -0.51 ± 0.65 -0.79 0.43

Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in behavioural frequencies between the three gallery compartments and the five fungal species. Model coefficients are reported as coeff. ± se (standard error of the estimate), with the group in brackets in the first row of the model as the reference category (coefficient set to zero). The influences of independent variables on the behavioural frequencies are displayed as contrasts between classes to give a better understanding. A positive contrast denotes that the mean of the second class is higher than the mean of the first class; a negative contrast denotes the reverse.

164

Appendix 1 Table S7. Influence of the frequency of specific larval and adult female behaviours upon the frequency of the different fungal species. Fungal abundances Parameter coeff. ± se z p

Mucor sp. Intercept -1.9 ± 0.56 -3.4 <0.001 Cropping larvae 0.49 ± 0.96 0.51 0.61 Inactive larvae -1.1 ± 1.51 -0.73 0.47

Intercept -2.64 ± 0.98 -2.7 0.007 Cropping adult females 1.48 ± 1.29 1.15 0.25 Inactive adult females 2.56 ± 1.4 1.83 0.067 Shuffling adult females -1.1 ± 1.23 -0.9 0.37

Raffaelea sp. Intercept 0.24 ± 0.29 0.84 0.4 Cropping larvae -0.7 ± 0.65 -1.08 0.28 Inactive larvae 2.3 ± 0.95 2.42 0.016

Intercept -0.3 ± 0.61 -0.5 0.62 Cropping adult females 0.42 ± 0.94 0.45 0.66 Inactive adult females 0.97 ± 1.09 0.89 0.37 Shuffling adult females 0.91 ± 0.89 1.03 0.3

Fusarium sp. Intercept -1.08 ± 0.81 -1.34 0.18 Cropping larvae -0.45 ± 0.82 -0.55 0.58 Inactive larvae -2.74 ± 2.3 -1.19 0.23

Intercept -2.3 ± 1.23 -1.88 0.06 Cropping adult females 0.68 ± 1.28 0.53 0.6 Inactive adult females 0.85 ± 2.02 0.42 0.67 Shuffling adult females 0.91 ± 1.42 0.64 0.52

Phaeoacremonium rubrigenum Intercept -3.51 ± 0.9 -3.91 <0.001 Cropping larvae 1.71 ± 1.48 1.16 0.25 Inactive larvae -3.09 ± 4.78 -0.65 0.52

Intercept -4.04 ± 1.32 -3.06 0.002 Cropping adult females 2.01 ± 1.97 1.02 0.31 Inactive adult females 3.18 ± 2.03 1.57 0.12 Shuffling adult females -0.08 ± 1.88 -0.04 0.97

Unknown fungus Intercept -2.36 ± 0.48 -4.91 <0.001 Cropping larvae 2.36 ± 0.82 2.86 0.004 Inactive larvae -4.36 ± 3.49 -1.25 0.21

Intercept -2.66 ± 1.06 -2.52 0.012 Cropping adult females 0.77 ± 1.54 0.5 0.62 Inactive adult females -2.2 ± 2.46 -0.9 0.37 Shuffling adult females 1.67 ± 1.46 1.15 0.25 Separate GEE models with an exchangeable correlation structure of the response variable within a cluster (gallery-identity), for examining differences in fungal frequencies per gallery depending on the frequency of inactivity and of potential fungus cleaning behaviours (cropping and shuffling). Model coefficients are reported as coeff. ± se (standard error of the estimate).

165

Appendix 1 Table S8. Separate GLM models to examine factors influencing adult females to lay eggs, develop ovaries or disperse.

Proportion of adult ♀♀ Parameter coeff. ± se z p

Egg-laying Intercept 2.9 ± 1.0 2.89 0.004 Number of immature offspring* -0.02 ± 0.03 -0.79 0.43 Number of ♀♀, non-developed ovaries -0.82 ± 0.3 -2.72 0.007 Number of ♀♀, developed ovaries -0.57 ± 0.22 -2.62 0.009 Number of ♀♀* 0.29 ± 0.47 0.61 0.54

Developed ovaries Intercept -3.74 ± 2.47 -1.52 0.13 Number of immature offspring* 0.02 ± 0.02 0.75 0.46 Number of ♀♀, egg-laying -0.5 ± 0.26 -1.93 0.054 Number of ♀♀, non-developed ovaries -0.59 ± 0.24 -2.44 0.015 Number of ♀♀ 0.69 ± 0.29 2.38 0.018

Dispersing Intercept 0.49 ± 0.48 1.03 0.31 Number of eggs* -0.01 ± 0.01 -0.86 0.39 Number of larvae -0.24 ± 0.1 -2.34 0.019 Number of ♀♀, non-developed ovaries -0.26 ± 0.12 -2.19 0.028 Number of ♀♀, developed ovaries 0.48 ± 0.19 2.53 0.011 Number of ♀♀* -0.1 ± 0.41 -0.24 0.81 Model coefficients are reported as coeff. ± se (standard error of the estimate). * Variable not in the final model after step-wise model reduction using ANOVA analysis of loglikelihood scores.

Table S9. Composition of X. affinis laboratory galleries about 60 days after gallery foundation. Immature stages Adult ♀♀ 1st 2nd/3rd Teneral Adult non-developed developed egg- Gallery Eggs instar instar Pupae ♀♀ ♂♂ ovaries ovaries laying dispersing A 18 1 4 24 - 1 7 3 - 3 B 42 1 - - - 1 - - 4 2 C 20 3 - 1 - - - - 3 5 D 22 8 3 1 2 - 2 6 1 14 E 53 8 5 - - 1 2 2 5 1 F 19 - 1 5 - 1 3 1 3 8 G 92 5 - - - 1 - 1 4 3 H 3 - - 12 - 1 4 - 3 -

166

Appendix 2

Mechanisms of fungus gardening in ambrosia beetles

Peter H.W. Biedermann1,2, Michael Taborsky1 and Cameron R. Currie2

1Department of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland 2Department of Bacteriology, University of Wisconsin-Madison, Madison, USA

Corresponding author: Peter H.W. Biedermann Baltzerstrasse 6, CH-3012 Bern, Switzerland phone: 0041 31 631 3015 e-mail: [email protected]

Manuscript in work

Introduction

The most elaborate of insect–fungus interactions is advanced fungus farming, in beetles called ambrosia symbiosis. This relationship evolved repeatedly in beetles, at least eight times in weevils, and ambrosia beetle is an ecological classification, not a phylogenetic group (Hulcr and Dunn 2011). In comparison to bark beetles they differ in how they colonize trees and rely on fungi for food. Unlike bark beetles, which live under the bark where they feed on fungus infested tree tissues, ambrosia beetles build tunnel systems (galleries) deep within the sapwood, and they actively cultivate mutualistic fungi in “ambrosia gardens” on the walls of these galleries as their sole food source (Francke-Grosmann 1967; Norris 1979; Beaver 1989). Primary symbionts, so called ambrosia fungi in the two anamorph genera Ambrosiella and Raffaelea (Ascomycota) are dominating these gardens (Beaver 1989; Farrell et al. 2001). Both genera produce asexual fruiting structures (conidiophores and – spores) in the presence of the beetles. This so called “ambrosial growth” forms large mats on the tunnel walls and is essential for beetle nutrition. Convergent fruiting is also known from the cultivars of farming ants (= gongolydia; Bass and Cherrett 1996) and termites (= nodules ; Aanen et al. 2002), but in none of these systems it is fully understood how this specialized growth is triggered by the insects.

Studies on Xyleborinus saxesenii and Xyleborus affinis (Chapter 1-6, Appendix 1) showed that in ambrosia beetles larval and adult offspring of a single foundress cooperate in brood care, gallery maintenance, and fungus gardening. The latter is termed cropping and is one of the most common behaviours of adult females within galleries. Its role for the health and productivity of the fungus gardens is unknown. In a descriptive study it had been mentioned, however, that pure cultures of Ambrosiella hartigii, the mutualistic fungus of the ambrosia beetle Anisandrus dispar, produce fruiting structures, after they have been in contact with adult females (see Fig. 1 from Batra 1967). This so-called ambrosial growth is ubiquitous in ambrosia beetle galleries, but seems to occur only in the presence of the beetles.

In this study we aimed at (i) quantifying Batra’s (1967) results experimentally and (ii) testing their general validity for ambrosia beetles. In particular we tested (a) if ambrosial growth can be induced by adult females of four different ambrosia beetle species (Cnestus mutilatus, Anisandrus sayi, Xylosandrus germanus, Xyleborinus saxesenii) in pure cultures of their mutualists, which we had isolated from their mycetangia before. Additionally we also tested (b) if it can also be induced by larvae and pupae, and (c) what are the potential underlying triggers for this phenomenon (e.g., mechanical disturbance, larval frass and chemical substances).

167

Appendix 2 Material & Methods

Study species Xyleborinus saxesenii and Xylosandrus germanus are found in the temperate zones worldwide and the most common Xyleborini in Central Europe (Wood 1982). They differ in their social system and mode of fungiculture. X. saxesenii is a primitively eusocial species with multiple developing generations within one nest and larvae that feed on fungus infested wood (Biedermann and Taborsky 2011). X. germanus on the other hand has only a single generation per nest and larvae that feed predominantly on fungus (Bischoff 2004). This may relate to the different natures of their ambrosia fungi, which are from two distantly related clades of ascomycete fungi in which ambrosial growth convergently evolved as a result of beetle feeding: Ambrosiella hartigii, the symbiont of X. germanus, belongs to the Ceratocystis clade, produces larger fruiting structures and does not survive the winter in galleries (Zimmermann 1973). It is also associated with Anisandrus dispar, which had been found by Batra to trigger the fruiting of A. hartigii on plates in culture (Batra 1967). Raffaelea sulfurea, the symbiont of X. saxesenii, belongs to the Ophiostoma clade, is cold tolerant and like all Ophiostoma fungi insensitive to the antibiotic cycloheximide, which is produced by Streptomyces bacteria (Cassar and Blackwell 1996). This might be an adaption to grow in the presence of such bacteria.

Experimental procedure We caught adult females of the four different beetle species in ethanol baited traps in the surrounding of Madison, USA during May 2009. Specimens were brought to the laboratory and surface sterilized. The mycetangium from five individuals per species was dissected and containing fungal spores were cultured on plates. Most of these isolations gave pure cultures and were stored in the refrigerator. The rest of the beetles were put in rearing tubes on artificial medium to start laboratory colonies (Biedermann et al. 2009). This was only successful in X. saxesenii and X. germanus. Larvae, pupae, larval frass and medium (as a control for larval frass) for the experiment were taken from these lab colonies. Five days before the start of the experiment we set up pure cultures of the four different fungi and malt agar plates. Thus, at the start of the experiment fungal mycelium had covered about 2/3 of the plate’s diameter. Adult females (either from lab cultures or the field), pupae, larvae (when available) and two potential excretions (larval frass and proline) were applied on pure cultures on agar plates of the four different ambrosia fungi. Solitary individuals were directly placed onto the fungal mycelium and subsequently covered with a small plastic cup. In all groups we had a control area, where nothing was applied. In the X. saxesenii treatment we also controlled for the cup, and placed an empty cup onto the medium. Sterile artificial medium was used as negative control for the larval frass. Distilled, sterile water served as a negative control for proline, which was diluted in this water. Proline was tested because a study on the fungal growth on artificial media suggested that the ambrosial growth form of the fungus in the mycetangia is attributed to an extreme abundance of the free amino acid proline in the insect’s haemolymph and body secretions (French and Roeper 1973). Over a period of 13 days we compared the change of fungus growth to a control area where nothing was applied. We took daily pictures of the plates and later analysed the amount of ambrosial growth or aerial mycelium using image analysis software.

Results

Cnestus mutilatus. Ambrosiella beaver, the ambrosia fungus of Cnestus mutilatus produced ambrosial growth that was randomly distributed on the plate and none of the treatments differed from the control area where nothing was applied (Fig. 3). Anisandrus sayi. The unknown ambrosia fungus of Anisandrus sayi did not produce any ambrosial growth in culture on plates, although fruiting structures were visible within the gallery systems of this beetle in artificial medium. We probably did not isolate the mutualistic ambrosia fungus of A. sayi. Xylosandrus germanus. Ambrosiella hartigii, the ambrosia fungus of Xylosandrus germanus, produced fruiting structures (ambrosial growth) in the presence of adult X. germanus females, pupae, larvae and larval frass (p < 0.001; Fig. 2). In did not produce any ambrosial growth neither in the control treatment, when proline or sterile artificial medium was applied nor when mycelium was mechanically disturbed with a scalpel (p < 0.05). Interestingly, fruiting was also induced by the presence of adults of Anisandrus sayi – a different ambrosia beetle, but only if the beetles were in direct contact with the fungus.

168

Appendix 2 Xyleborinus saxesenii. Raffaelea sulfurea, the ambrosia fungus of Xyleborinus saxesenii did not produce any ambrosial growth on plates, but instead produced plenty of aerial mycelium. Analyses revealed larvae (LME model: p < 0.001), larval frass (p < 0.05) and adult females (p < 0.05) to significantly induce the growth of aerial mycelium in comparison to a control area, where nothing was applied (Fig. 4). Placing the empty cup without larvae and adults on the fungus also tended to induce this growth (p = 0.06), however, which may suggest that the changed gas content or air moisture triggered the fungal pleomorphism. Taking the empty cup as new control group, only larvae further increased mycelium induction (p < 0.001).

Discussion

In this study we show that some ambrosia beetles are able to induce growth or fruiting in their mutualistic ambrosia fungus. Only one fungus, Ambrosiella beaveri fruited independent of the presence and activity of the beetles. The induction of aerial mycelium by larvae in X. saxesenii and ambrosial growth (i.e., nutritional fruiting bodies) by larvae, pupae and adult females in X. germanus is an extraordinarily important result for my thesis, because it shows for the first time that immatures (larvae, pupae) and adults are able to increase the productivity of their fungus. Hence, the presence of immatures and delayed dispersing adults is likely to increase the productivity of the fungus, and thus the amount of food available for the whole colony. It remains unknown what substance(s) and/or behaviour(s) induce the fungal pleomorphism in X. saxesenii and X. germanus. The different gas content or air humidity within our cups, may resemble aerial conditions within the tunnels. Unknown X. saxesenii larval secretions or excretions seem to play an additional role. The induction by non-moving pupae in X. germanus strongly suggests that chemical substances produced either by the insects or other microbes are involved. Ambrosial growth in galleries is covered with unstudied bacteria and many microbes can be isolated from the body surface and mouth parts of adult beetles and larvae (own unpublished data). Most importantly, (i) the application of unidentified yeasts associated with the ambrosia beetle Xyleborus dispar triggered the ambrosial growth in its ambrosia fungus in culture (Peklo and Satava 1950), and (ii) the bark beetle associate Leptographium procerum, a relative of the ambrosia fungi, was strongly induced to produce asexual fruiting structures in the presence of several bacteria (e.g. Pantoea sp., Pseudomonas sp.; Adams et al. 2009). A positive role of yeasts in the gardens of ambrosia beetles has been frequently claimed (Webb 1945; Francke-Grosmann 1963; Francke-Grosmann 1967; Beaver 1989), but never tested. In summary, there are strong indications that symbiotic microbes might be involved in the fruiting of the ambrosia fungi, which would be a unique new function of mutualistic microbes in fungus cultivating insects.

Literature

Aanen DK, Eggleton P, Rouland-Lefévre C, Guldberg-Frøslev T, Boomsma JJ & Rosendahl S (2002) The evolution of fungus-growing termites and their mutualistic fungal symbionts. PNAS 99: 14887-14892.

Adams AS, Currie CR, Cardoza Y, Klepzig KD & Raffa KF (2009) Effects of symbiotic bacteria and tree chemistry on the growth and reproduction of bark beetle fungal symbionts. Canadian Journal of Forest Research-Revue Canadienne de Recherche Forestiere 39: 1133-1147.

Bass M & Cherrett JM (1996) Leaf-cutting ants (Formicidae, Attini) prune their fungus to increase and direct its productivity. Functional Ecology 10: 55-61.

Batra LR (1967) Ambrosia fungi - A taxonomic revision and nutritional studies of some species. Mycologia 59: 976- 1017.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Biedermann PHW, Klepzig KD & Taborsky M (2009) Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environmental Entomology 38: 1096-1105.

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

169

Appendix 2 Bischoff LL (2004) The social structure of the haplodiploid bark beetle, Xylosandrus germanus. Diploma Thesis, Zoological Institute, University of Berne, Switzerland.

Cassar S & Blackwell M (1996) Convergent origins of ambrosia fungi. Mycologia 88: 596-601.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Francke-Grosmann H (1963) Some new aspects in forest entomology. Annual Review of Entomology 8: 415-438.

Francke-Grosmann H (1967) Ectosymbiosis in wood-inhabiting beetles. Symbiosis (Henry SM, ed), pp. 141-205. Academic Press, New York.

French JRJ & Roeper RA (1973) Patterns of nitrogen utilization between the ambrosia beetle Xyleborus dispar and its symbiotic fungus. Journal of Insect Physiology 19: 593-605.

Hulcr J & Dunn RR (2011) The sudden emergence of pathogenicity in insect-fungus symbioses threatens naive forest ecosystems. Proceedings of the Royal Society B: Biological Sciences.

Norris DM (1979) The mutualistic fungi of Xyleborini beetles. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 53-63. Allanheld, Osmun & Company, Montclair.

Peklo J & Satava J (1950) Fixation of Free Nitrogen by Insects. Experientia 6: 190-192.

Webb S (1945) Australian ambrosia fungi. Proceedings of the Royal Society of Victoria 57: 57-79.

Wood SL (1982) The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Naturalist Memoirs 6: 1-1359.

Zimmermann G (1973) Vergleichende ökologisch-physiologische Untersuchungen an Ambrosiapilzen, Assoziierten Bläuepilzen und Luftbläuepilzen. Doctoral thesis, Georg-August University Göttingen, Germany.

Graphs

Fig. 1. An agar plate with a pure culture of Ambrosiella hartigii, the mutualistic fungus of Anisandrus dispar (the mycelial growth form is black). White fruiting structures were induced at locations where adult females had been crawling before (Figure from Batra 1967).

170

Appendix 2

Fig. 2. Induction of ambrosial growth in Ambrosiella hartigii, the ambrosia fungus of Xylosandrus germanus. A Graph showing the induction of spores and spore-like structures by different treatments involving different stages of X. germanus and adult females of Anisandrus sayi. Significant differences to the control are marked by ***- p < 0.001 (LME models); number of tested samples are given in brackets. B Example of the experimental procedure. Individuals and certain substances were confined to / applied on pure cultures of the ambrosia fungus A. hartigii. Reaction of the fungus is displayed six days after the treatment. Note that mycelial growth is black, while ambrosial growth is white.

171

Appendix 2

Fig. 3. Ambrosial growth in Ambrosiella beaveri, the ambrosia fungus of Cnestus mutilatus. A Graph showing the number of spores and spore-like structures depending on adult females and different concentrations of proline (and associated controls) applied on the fungus. There are no significant differences to the control: p > 0.05 (LME models); number of tested samples are given in brackets. B Picture of A. beaveri and the results of the treatment after six days. Note the dark mycelial growth and the white ambrosial growth.

172

Appendix 2

Fig. 4. Growth of aerial mycelium in Raffaelea sulphurea, the ambrosia fungus of Xyleborinus saxesenii. A Quantity of aerial mycelium induced by larvae, pupae and adult female X. saxesenii and other treatments. Significant differences to the control are marked by ***- p < 0.001, * - p < 0.05, (*) – p < 0.1 (LME models); number of tested samples are given in brackets. B Picture of Raffaelea sulphurea six days after the treatment. Note the darker mycelial growth and the whitish ambrosial growth.

173

Appendix 2

174

Summary & Conclusion

The most famous farming insects are, without doubt, the attine ants. Rightly so, but they are not the only ones...

The ancient and highly evolved mutualisms between insects and fungi are a textbook example of symbiosis (Mueller et al. 2005). Ants, termites and ambrosia beetles independently evolved the ability to grow fungi for food about 40–60 million years before the origin of human agriculture (Mueller and Gerardo 2002). Fungus farming evolved once in ants and termites and at least nine times in beetles (Farrell et al. 2001; Hulcr et al. 2007), with no known reversal to non-farming. The evolution of fungiculture seems to be always accompanied by a large radiation and a huge ecological success (Farrell et al. 2001; Mueller et al. 2005). Like in humans, farming by insects can only be managed in cooperative societies, in which planting, protecting, cultivating and harvesting of the crops are shared by several individuals. Thus, it is not surprising that ambrosia beetles evolved the highest sociality among beetles, as one ambrosia beetle is regarded eusocial (Kent and Simpson 1992) and others at least sub-social (Kirkendall et al. 1997).

In this Ph.D. project I aimed to contribute a puzzle piece to the understanding of social evolution and evolution of symbioses in nature. In particular, I investigated the social system and the mode of fungiculture focusing on two exemplary species of the ambrosia beetle tribe Xyleborini, the native fruit-tree pinhole borer Xyleborinus saxesenii and the American sugarcane shot-hole borer Xyleborus affinis. I assume that they stand for two fundamentally different social systems and modes of fungus agriculture, because X. saxesenii inhabits cave-like galleries, whereas X. affinis, like most other ambrosia beetles, digs out branching tunnel systems.

New insights into the social system of xyleborine ambrosia beetles Until recently, behaviours of ambrosia beetles within their galleries had not been studied, because they live a hidden life inside the wood, where they are virtually impossible to observe. The only report on eusociality in ambrosia beetles was not based on behavioural data, but reproductive roles were inferred by destructive sampling of active nests and the finding that only one Austroplatypus incompertus female per colony had developed ovaries (Kent and Simpson 1992). An artificial rearing and observation technique (Saunders and Knoke 1967) was rediscovered by our group (Peer & Taborsky, 2004; 2005), however, which I was able to adapt for laboratory rearing of several xyleborine ambrosia-beetle species (Biedermann et al. 2009). Among those, Xyleborinus saxesenii and Xyleborus affinis were supposed to have a high potential for advanced sociality and fungus gardening, because (i) field studies had shown that dozens of larvae and adults gregariously live in galleries that are sometimes settled by several offspring generations (Hosking 1972; Schneider 1987) and (ii) correlative data suggested that X. saxesenii adult daughters delay dispersal from the natal nest depending on the number of brood – possibly depending on care (Peer and Taborsky 2007). The first observation with our laboratory colonies in tubes was unique for holometabolous insects: larval and adult offspring of a single foundress cooperate in brood care, gallery maintenance, and fungus gardening, showing a clear division of labour between larval and adult colony members (Chapter 1 and 6, Appendix 1). Larvae of both species participate in brood care, and in X. saxesenii larvae also enlarge the gallery and participate in gallery hygiene. The cooperative effort of adult females in the colony and the timing of their dispersal depend on the number of sibling depending on brood care (larvae and pupae) and on the number of adult workers in both species. This suggests that

175

Summary & Conclusion

staying and helping in the nest – for gaining indirect fitness – is triggered by demands of brood dependent on care. Insects with a metamorphosis between the larval and adult stage (Holometabola; e.g., bees, wasps, ants, beetles) dramatically reorganize their morphology during ontogenesis, because the two stages represent distinct developmental and evolvable modules (Yang 2001). In some ambrosia beetles, like X. saxesenii and some platypodine species, this has apparently led to larvae that are predisposed by their body morphology and the frequent renewal of mandibles by moulting to assume certain tasks, like balling of frass and gallery enlargement (Chapter 1). Additionally, X. saxesenii larvae seem to be capable of producing wood-degrading enzymes (Hemicellulases), which are not found in the bodies of adults (Chapter 8). By contrast, larvae of solely mycetophagous ambrosia beetles, like X. affinis, play a minor role for gallery productivity, because the only cooperative behaviour they engage in is allogrooming (Chapter 6, Appendix 1). The role of solely mycetophagous larvae in the induction of ambrosial growth has not been tested yet, however (Appendix 2). In summary, larvae in most ambrosia-beetle species are free to move within the natal nest, and are not confined to small areas or brood cells like those of most hymenopteran social societies (Wilson 1971; Hölldobler and Wilson 1990). This, in combination with different capabilities, predisposed ambrosia beetles for division of labour between larval and adult stages. Importance and specific roles of larvae in the galleries appear to vary between species (Chapter 6). Males also take part in cooperative behaviours (X. saxesenii; Chapter 1 and 4), which suggests that relatedness asymmetries caused by haplodiploidy, which would favour female-biased help, are probably offset by inbreeding in this species (comparable to thrips; Hamilton 1972; Chapman et al. 2000). Nevertheless, because of strong local male competition, there are only one or two males on ten to eighty females per gallery, and thus their help is of minor importance. The only exception are male-only galleries (founded by unfertilized females) that develop normally and can produce as much offspring as normal galleries (Chapter 4). Adult female offspring exhibit philopatric periods of at least 7 days (X. affinis) / 17 days (X. saxesenii; Chapter 2 and 5), and some even stay all their life in the natal nest (X. saxesenii: Chapter 3, Peer and Taborsky 2007). Removal and independent breeding of staying X. affinis females showed that they are totipotent and fully capable of founding their own nest (Chapter 5). Ovary dissections revealed that one (X. saxesenii field galleries: Chapter 2) to two quarters (X. affinis laboratory galleries; Appendix 1) of staying females also lay eggs in the natal nest during their philopatric period (Chapter 2 and 5). These egg-laying females (i) engaged more often in social behaviours than non-egg-laying females (X. affinis: Appendix 1) and (ii) they appeared to be the only ones overtaking the most dangerous blocking behaviour (X. saxesenii: Chapter 1), which exposes females to predation and parasites. Numbers of egg-layers did neither correlate with the number of staying adult females nor with the number of eggs, which suggests that egg numbers are adjusted to fungus productivity. Interestingly, delayed dispersal imposes fitness costs on independent reproduction, because X. affinis females that disperse after their philopatric period produced fewer eggs in their brood when bred independently than females removed from the gallery before their philopatric period, suggesting that delayed dispersal comes at a cost to X. affinis females (Chapter 5). This cost probably results from reproduction and alloparental care in the natal nest during philopatry. Hence, we could reject the hypothesis that females engage in maturation feeding (i.e., store up body reserves) during this period. Artificial selection on early and late dispersal in X. saxesenii resulted within six generations in a significantly different timing of dispersal between the two lines (Chapter 3). Late dispersal was correlated with a higher number of offspring produced by the mother and more adults staying within the nest, which suggests that delayed dispersal is linked with a cooperative strategy. Indeed, adult female activity and the frequency of cooperative nest protection are higher in the late-dispersing strain than the early-dispersing strain. The cooperative strategy of late dispersers is, however, associated with a lower success of independent nest foundation (Chapter 3). In summary, X. saxesenii and X. affinis are typical cooperative breeders, characterized by overlapping generations, cooperative brood care, and totipotent workers. Their galleries are relatively long-lived in comparison to other subsocial ambrosia beetles (e.g., Anisandrus dispar, Xylosandrus germanus;

176

Summary & Conclusion

Schneider-Orelli 1913; Bischoff 2004), and thus several females in consecutive offspring generations can sometimes continue breeding in the same protected nest. In the temperate X. saxesenii, this is possible because larval and pupal stages produced in fall stay together with adults in the natal nest during the winter and adults may restart egg-laying in the following spring. Females in the temperate X. germanus also delay dispersal, but have only one breeding cycle per gallery; they overwinter exclusively as adults (Heidenreich 1960; 1964). It is possible that cold-tolerance of immature X. saxesenii has facilitated the evolution of cooperative breeding / facultative eusociality in this species (cf. facultative eusociality in halictid bees; Eickwort et al. 1996; Field et al. 2010). In conclusion, the high degree of sociality in X. saxesenii and X. affinis (and likely many other xyleborine ambrosia beetles) seems to result from a combination of four major factors (for details see Chapter 1 and 6): (i) parental care as a preadaptation for the evolution of sociality in the ancestors of modern ambrosia beetles; (ii) very high relatedness within families due to haplodiploidy and inbreeding; (iii) a proliferating, monopolisable nutritional resource providing ample food for many individuals that needs to be tended and protected; and (iv) high costs of dispersal due to the difficulties of finding a suitable host tree, of nest foundation, and a successful start of fungiculture, which render predispersal cooperation particularly worthwhile. Current knowledge suggests that all ambrosia beetles are at least sub-social, which seems necessary for successful fungus gardening, protection of the nest and brood care. Cooperative breeding may evolve in species, which galleries may persist for several consecutive offspring generations. This depends (a) on the generation time, which is not more than 30 days under optimal conditions in the majority of xyleborine ambrosia beetles, and (b) on the longevity of the breeding substrate (i.e., competition with other ambrosia beetles, the timing of beetle attack in the dying process of a tree, diameter of branches/stems, wood type). Totipotency of all females is, however, expected to be maintained in all species settling dying or dead trees that are able to disperse if the productivity of the nest cannot be guaranteed any longer. In X. saxesenii, for example, many galleries need to be left after a single generation, despite the fact that some galleries are productive for several offspring cycles. The obligatorily eusocial ambrosia beetle Austroplatypus incompertus settles living tress – a habitat with almost endless productivity. Indeed, single galleries can be productive for at least 36 years, although productivity is apparently very low because beetles need six years to reach adulthood (Kent and Simpson 1992).

New insights into the feedback between sociality and fungus-culturing in xyleborine ambrosia beetles Division of labour and sociality in ambrosia beetles evolved multiple times in association with fungiculture, and every step towards higher sociality possibly facilitated the beetle-fungus mutualism, or vice versa. This contrasts with fungus-farming ants and termites, which already lived in eusocial groups at the origin of fungiculture (Mueller and Gerardo 2002; Mueller et al. 2005), and makes ambrosia beetles a unique model system to study the evolution of sociality in relation to fungiculture (Mueller et al. 2005; Biedermann and Taborsky 2011). But why should fungiculture profit from sociality and division of labour? First, farmed fungi can profit from a group of individuals that jointly excavate a gallery, protect the gardens, and maintain optimal growing conditions by regulating humidity, for example by blocking (Chapter1), and weeding the gardens. The latter has not been fully demonstrated in ambrosia beetles, but in an experiment we showed that X. saxesenii larvae can reduce the growth of mould (Chapter 1), and correlative behavioural studies showed that certain hygienic behaviours of adult female X. affinis (e.g. shuffling, fungus cropping) are more commonly shown in the presence of fungal pathogens (Appendix 1). The strongest evidence for a positive feedback between sociality and fungus-farming, however, is our finding that in some species fruiting of ambrosia fungi can be induced, apparently simply by the physical contact of the fungus with either larvae, pupae, or adult females (see below, Appendix 2). The resulting fruiting structures apparently serve as the dominant food for the whole colony. Hence, the more immatures and adults are present within a gallery, the stronger they induce ambrosial

177

Summary & Conclusion

growth and the better is the food supply for the developing brood. This suggests that daughters that delay dispersal enhance gallery productivity simply by their presence.

New insights into the xyleborine ambrosia beetle-fungus mutualism Ambrosia beetles are close relatives of bark beetles, but they differ in their way to colonize trees and their dependence on fungi for food. Unlike bark beetles, which live under the bark where they feed on fungus- infested tree tissues, ambrosia beetles build tunnel systems (galleries) deep within the sapwood, and they actively cultivate mutualistic fungi in “ambrosia gardens” on the walls of these galleries as their sole food source (Francke-Grosmann 1967; Norris 1979; Beaver 1989). Primary symbionts, in the two anamorph genera Ambrosiella and Raffaelea (Ascomycota), dominate these gardens (Beaver 1989; Farrell et al. 2001). Interestingly, both genera originate from two different clades of fungal plant pathogens (Ceratocystis and Ophiostoma; Cassar and Blackwell 1996; Harrington et al. 2010), but they show convergent morphological adaptations for the symbiosis with beetles. They are strongly polymorphic and change their mycelial growth either in a yeast-like form within their host beetle mycetangia, or produce asexual fruiting structures (conidiophores and conidiospores) in the presence of the beetles. This ambrosial growth forms large mats on the tunnel walls (Fig. 1) and is essential for beetle nutrition. Convergent fruiting is also known from the cultivars of farming ants (= gongolydia; Bass and Cherrett 1996) and termites (= nodules; Aanen et al. 2002), but in none of these systems it is fully understood how the insect trigger this specialized growth. Adult female beetles induce ambrosial growth in pure cultures of the fungus on plates via an unknown mechanism (in Xyleborus dispar; Batra and Michie 1963; French and Roeper 1972), and one of my studies showed that also larvae and pupae can trigger ambrosial growth (in Xylosandrus germanus; Appendix 2). The induction by non-moving pupae strongly suggests that chemical substances produced either by the insects or other microbes are involved. Ambrosia layers within galleries are covered with unstudied bacteria (see Fig. 1) and yeasts, bacteria and fungi can be isolated from the body surface and mouth-parts of adult beetles and larvae (own unpubl. data). A positive role of yeasts in the gardens of ambrosia beetles has been frequently claimed (Webb 1945; Francke-Grosmann 1963; 1967; Beaver 1989), but never tested. In summary, there are strong indications that symbiotic microbes might be involved in the fruiting of the ambrosia fungi, which would be a unique new function of mutualistic microbes in fungus-cultivating insects. Originally thought to just involve the beetles and ambrosia fungi, fungus gardens of ambrosia beetles are now considered to inhabit multiple other microbial associates (Haanstad and Norris 1985; Kajimura and Hijii 1992), like other filamentous fungi (e.g. Fusarium sp., Penicillium sp., Paecilomyces sp.; Francke-Grosmann 1967; Beaver 1989; Six et al. 2011), yeasts and bacteria (Haanstad and Norris 1985; Canganella et al. 1994; Endoh et al. 2008; Ninomiya et al. 2010; Chapter 9). A community of microbes with different enzymatic capabilities may have advantages to detoxify and degrade wood compounds synergistically and may profit from nutrient exchange (Shifrine and Phaff 1956; Canganella et al. 1994; Adams et al. 2008; Chapter 8). Additionally, antibiotics-producing symbionts (e.g. Streptomycetes species) are known from several other insects and defend their hosts against pathogens (Currie et al. 1999; Kaltenpoth et al. 2005; Martin 2009). Pathogens and parasites are probably also a threat to ambrosia gardens. However, although fresh wood is very attractive for many wood-degrading microbes, beetle fungus gardens are well protected against environmental microbes within tunnels of almost sterile wood. In contrast, fungus-farming ants and termites need

178

Summary & Conclusion

to (i) sequester their gardens from the surrounding soil that is heavily contaminated with microbes and developed, and (ii) have advanced microbial defences against microbes that they constantly bring into the nest with the substrate they collect to feed their fungi (Mueller and Gerardo 2002; Mueller et al. 2005). Nevertheless, my studies showed that several pathogenic moulds (Paecilomyces sp., Mucor sp., etc.) also co-exist in ambrosia gardens (Chapter 7, Appendix 1), and that both larvae and adults are able to hinder their spread within their galleries in unknown ways (Chapter 2), which may also involve other microbes. Antibiotics-producing bacteria inhibit the spread of an antagonistic fungus in farming ants (Currie et al. 1999; Currie and Stuart 2001) and other bark beetles (Cardoza et al. 2006; Scott et al. 2008). Additionally, insect-associated yeasts have been shown to curtail growth of antagonistic fungi (Liu et al. 2007; Adams et al. 2008; Rodrigues et al. 2009). Such a defence by microbes has the advantage that, unlike the insects, microbes can potentially evolve at the same rate as pathogens, enabling mutualistic insect-microbe systems to respond more rapidly to the emergence of novel disease genotypes (Currie et al. 1999; Denison et al. 2003; Adams et al. 2009).

Literature

Aanen DK, Eggleton P, Rouland-Lefévre C, Guldberg-Frøslev T, Boomsma JJ & Rosendahl S (2002) The evolution of fungus-growing termites and their mutualistic fungal symbionts. PNAS 99: 14887-14892.

Adams AS, Six DL, Adams SM & Holben WE (2008) In vitro interactions between yeasts and bacteria and the fungal symbionts of the mountain pine beetle (Dendroctonus ponderosae). Microbial Ecology 56: 460-466.

Bass M & Cherrett JM (1996) Leaf-cutting ants (Formicidae, Attini) prune their fungus to increase and direct its productivity. Functional Ecology 10: 55-61.

Batra LR & Michie MD (1963) Pleomorphism in some ambrosia and related fungi. Transactions of the Kansas Academy of Science 66: 470-481.

Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. Insect-fungus interactions (Wilding N, Collins NM, Hammond PM & Webber JF, eds), pp. 121-143. Academic Press, London.

Biedermann PHW & Taborsky M (2011) Larval helpers and age polyethism in ambrosia beetles. Proc. Natl Acad. Sci. USA 108: 17064-17069.

Bischoff LL (2004) The social structure of the haplodiploid bark beetle, Xylosandrus germanus. Diploma Thesis, Zoological Institute, University of Berne, Switzerland.

Canganella F, Paparatti B & Natali V (1994) Microbial Species Isolated from the Bark Beetle Anisandrus-Dispar F. Microbiological Research 149: 123-128.

Cardoza YJ, Klepzig KD & Raffa KF (2006) Bacteria in oral secretions of an endophytic insect inhibit antagonistic fungi. Ecological Entomology 31: 636-645.

Cassar S & Blackwell M (1996) Convergent origins of ambrosia fungi. Mycologia 88: 596-601.

Chapman TW, Crespi BJ, Kranz BD & Schwarz MP (2000) High relatedness and inbreeding at the origin of eusociality in gall-inducing thrips. PNAS 97: 1648-1650.

Currie CR, Scott JA, Summerbell RC & Malloch D (1999) Fungus-growing ants use antibiotic-producing bacteria to control garden parasites. Nature 398: 701-704.

179

Summary & Conclusion

Currie CR & Stuart AE (2001) Weeding and grooming of pathogens in agriculture by ants. Proceedings of the Royal Society of London Series B-Biological Sciences 268: 1033-1039.

Denison RF, Kiers ET & West SA (2003) Darwinian agriculture: When can humans find solutions beyond the reach of natural selection? Q. Rev. Biol. 78: 145-168.

Endoh R, Suzuki M & Benno Y (2008) Ambrosiozyma akamigamensis sp. nov. and A. neoplatypodis sp. nov., two new ascomycetous yeasts from ambrosia beetle galleries. Antonie van Leeuwenhoek 94: 365-376.

Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH & Jordal BH (2001) The evolution of agriculture in beetles (Curculionidae : Scolytinae and Platypodinae). Evolution 55: 2011-2027.

Field J, Paxton RJ, Soro A & Bridge C (2010) Cryptic Plasticity Underlies a Major Evolutionary Transition. Current Biology 20: 2028-2031.

Francke-Grosmann H (1956) Hautdrüsen als Träger der Pilzsymbiose bei Ambrosiakäfern. Z. Morph. u. Ökol. Tiere 45: 275-308.

Francke-Grosmann H (1963) Some new aspects in forest entomology. Annual Review of Entomology 8: 415-438.

Francke-Grosmann H (1967) Ectosymbiosis in wood-inhabiting beetles. Symbiosis (Henry SM, ed), pp. 141-205. Academic Press, New York.

Haanstad JO & Norris DM (1985) Microbial symbiotes of the ambrosia beetle Xyletorinus politus. Microbial Ecology 11: 267-276.

Hamilton WD (1972) Altruism and related phenomena, mainly in social insects. Ann. Rev. Ecol. Syst. 3: 193-232.

Heidenreich E (1960) Primärbefall durch Xylosandrus germanus an Jungeichen. Anzeiger fur Schadlingskunde Pflanzenschutz Umweltschutz 23: 5-10.

Heidenreich E (1964) Ökologische Bedingungen für Primärbefall durch Xylosandrus germanus. J. Appl. Entom. 54: 131-140.

Hölldobler B & Wilson EO (1990) The ants. The Belknap Press of Harvard University Press, Cambridge, MA.

Hosking GB (1972) Xyleborus saxeseni, its life-history and flight behaviour in New Zealand. N. Z. J. Forest Science 3: 37-53.

Hulcr J, Kolarik M & Kirkendall LR (2007) A new record of fungus-beetle symbiosis in Scolytodes bark beetles (Scolytinae, Curculionidae, Coleoptera). Symbiosis 43: 151-159.

Kaltenpoth M, Gottler W, Herzner G & Strohm E (2005) Symbiotic Bacteria Protect Wasp Larvae from Fungal Infestation. Current Biology 15: 475-479.

180

Summary & Conclusion

Kent DS & Simpson JA (1992) Eusociality in the beetle Austroplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften 79: 86-87.

Liu X, Wang J, Gou P, Mao C, Zhu ZR & Li H (2007) In vitro inhibition of postharvest pathogens of fruit and control of gray mold of strawberry and green mold of citrus by aureobasidin A. International Journal of Food Microbiology 119: 223-229.

Martin K (2009) Actinobacteria as mutualists: general healthcare for insects? Trends in Microbiology 17: 529- 535.

Mueller UG, Gerardo NM, Aanen DK, Six DL & Schultz TR (2005) The evolution of agriculture in insects. Annual Review of Ecology Evolution and Systematics 36: 563-595.

Ninomiya S, Mikata K, Nakagiri A, Nakase T & Kawasaki H (2010) Pichia porticicola sp. nov., a novel ascomycetous yeast related to Pichia acaciae isolated from galleries of ambrosia beetles in Japan. The Journal of General and Applied Microbiology 56: 281-286.

Norris DM (1979) The mutualistic fungi of Xyleborini beetles. Nutrition, Mutualism, and Commensalism (Batra LR, ed), pp. 53-63. Allanheld, Osmun & Company, Montclair.

Rodrigues A, Cable RN, Mueller UG, Bacci M & Pagnocca FC (2009) Antagonistic interactions between garden yeasts and microfungal garden pathogens of leaf-cutting ants. Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 96: 331-342.

Scott JJ, Oh DC, Yuceer MC, Klepzig KD, Clardy J & Currie CR (2008) Bacterial protection of beetle-fungus mutualism. Science 322: 63.

Shifrine M & Phaff HJ (1956) The Association of Yeasts with Certain Bark Beetles. Mycologia 48: 41-55.

Six DL (2003) Bark beetle-fungus symbioses. Insect Symbiosis (Bourtzis K & Miller TA, eds), pp. 97-114. CRC Press, Boca Raton.

Six D, Poulsen M, Hansen A, Wingfield M, Roux J, Eggleton P, Slippers B & Paine T (2011) Anthropogenic effects on interaction outcomes: examples from insect-microbial symbioses in forest and savanna ecosystems. Symbiosis1-21.

Webb S (1945) Australian ambrosia fungi. Proceedings of the Royal Society of Victoria 57: 57-79.

Wilson EO (1971) The insect societies. Belknap Press of Harvard University Press, Cambridge.

181

Summary & Conclusion

182

Acknowledgements

Dankeschön. Merci. Thank you.

Eine Lebensabschnitt – 7 Jahre am Hasli – gehen für mich nun zu Ende!

Michael, vielen Dank für deine Ermunterung, nach Bern zu kommen, und das Sozialverhalten der Ambrosiakäfer zu erforschen. Ich habe durch dein „Riegersburg-Seminar“ nicht nur eine Passion für Ambroiakäfer und eine Nische für meine weitere Karriere als Wissenschaftler entdeckt, sondern habe auch – 2 Jahre später – meine baldige Frau Tabea kennengelernt. Danke, dass du mich immer fachlich, aber auch finanziell, voll und ganz unterstützt hast. Als Betreuer hast du mir eine klare Richtung für meine Forschung gewiesen, mir aber gelichzeitig viel Freiheit bei der Durchfürung gelassen. Das war genau richtig für mich. Mit deinem breiten Fachwissen hast du immer die treffenden Fragen gestellt und neue Lösungswege aufgezeigt. Auch wenn du mir keine finanzierte Doktorandenstelle anbieten konntest, hast du mich immer ermuntert, selbst Forschungsanträge zu schreiben. Ich war gezwungen, selbst aktiv zu werden, Projekte selbst zu planen, und mir Netzwerke mit anderen Wissenschaftlern aufzubauen. Obwohl das Schreiben eigener Anträge mich sehr gefordert hat, und ich mir damals oft eine fixe Stelle gewünscht hätte, bin ich im Nachhinein sehr froh über die lehrreichen Erfahrungen, die ich in dieser Zeit machen konnte. Mit dem über die Jahre aufgebauten Rüstzeug fühle ich mich jetzt sehr gut auf eine unabhängige wissenschaftliche Karriere vorbereitet. Ich danke dir ausserdem, dass du mich dabei unterstützt, auch nach dem Abschied aus Bern mit dem Modellsystem „Ambrosiakäfer“ weiterzuarbeiten. Danke für die tolle und unglaublich lehrreiche Zeit am Hasli!

Während dieser Arbeit habe ich grosses Entgegenkommen von mehreren Personen am Institut für Forstentomologie der BOKU Wien erfahren. Christian, vielen Dank für deine sofortige Zusage, meine Doktorarbeit und den Antrag bei der ÖAW (Österreichischen Akademie der Wissenschaften) zu unterstützen. Du hast mir Anfang 2008 mit dem gemeinsamen Durcharbeiten und „Zuspitzen“ den Antrag für mein DOC- Stipendium gerettet! Ich bin mir sicher, dass ich die 2 jährige Finanzierung ohne deine Hilfe nicht bekommen hätte. Dein Wohlwollen und dein Glaube an meine wissenschaftlichen Fähigkeiten haben mir immer sehr viel Selbstvertrauen gegeben. Thomas, dir möchte ich vor allem für die Vermittlung der Kontakte in die USA danken. Wo würde ich heute ohne die Erfahrungen und das Wissen, das ich mir in den USA angeeignet habe, stehen?! Rudi und Axel, euch danke ich für die spannenden Einblicke in die angewandte Borkenkäferforschung. Am gesamten Institut für Forstentomologie habe ich immer sehr viel Menschlichkeit und ehrliche Freundschaft erlebt – danke den Führungspersonen Axel, Rudi und Christian für dieses tolle Klima.

In the first year of my Ph.D.-studies I got the chance to visit the Southern Research Station (Louisiana, USA) for a few months where I learned how to isolate and work with fungi. My deepest thanks go to Kier for enabling this collaboration. I am so grateful that you invited me, even though you did not know me personally. I profited so much from your experience in working with fungi, and your financial support for visiting the labs of Diana, Cameron, Ken and Ulrich. The scientific input from these people was absolutely essential for the success of my Ph.D. I also want to thank Stacy and Chi for giving me social support during this visit in Pineville. I want to thank you two, Dan, Bill and Michael for the hospitality. Discussions with Dr. Moser, Brian and Will were a big joy. I got to meet Diana and Eric in Missoula, which resulted in an exciting collaboration on fungus identifications. Thank you, Diana, for sharing your knowledge on fungi with me. Eric, thanks for the adventures during free time. Cameron, thank you so much for your invitation to Madison and the months I could spend in your lab. The insights you gave me into ant-fungus interactions were a great inspiration. I enjoyed meeting Frank, Kirk, Garret and Ken, and became friends with Sandye, Aaron, Michael and Jiri. Thank you all for helpful discussions.

183

Acknowledgements

Ulrich, your passion on insect-fungus symbioses was immediately contagious. It was a great experience to go out in the field and dig up ant Atta-nest together with you. I profited a lot from the many discussions we had in Austin and elsewhere and inputs you gave me were crucial for several of my experiments, and your comments facilitated our PNAS publication. In Austin I also became friends with Barrett – thank you so much for the wonderful beetle drawings (on my wall and on the cover of this thesis) and the passion and joy you radiate.

Thanks also to Prof. Heinz Richner for chairing my Ph.D.-defense and to Prof. Koos Boomsma who kindly agreed to be my external referee.

Alex, Markus und Arne danke für eure Freundschaft und für den Spass, den wir gemeinsam hatten – sei es bei Partys, bei Wanderungen in der Natur, unseren „Badminton-Männerabenden“ oder den vielen lustigen Spieleabenden. Besonders geschätzt habe ich eure Hingabe bei wissenschaftlichen Diskussionen. Auch Babsi, Thomas und Stefan danke ich für die gemeinsamen Erlebnisse. Cori, Michael und Andrea, euch danke ich für die nette Gemeinschaft im Büro und den netten Gesprächen. Marcel, Florian und Claudia, danke für eure Freundschaft. Ihr wart wichtige Gesprächspartner für mich, nicht nur bei beruflichen, sondern auch bei privaten Themen und die gemeinsamen Unternehmungen bleiben mir in schöner Erinnerung. Schliesslich möchte ich all den vielen herzlichen Menschen am Hasli und in der Baltzerstrasse danken, in deren Gemeinschaft ich mich aufgehoben gefühlt habe.

Ich danke meinen Eltern, ohne die ich nicht der wäre, der ich bin. Ihr habt mich in meiner Leidenschaft für die Natur immer unterstützt und mir auch bei Schwierigkeiten geholfen, dranzubleiben. „Bua“, dir danke ich für die Besuche und die gemeinsamen Unternehmungen in der Schweiz und, dass ich mich immer auf dich verlassen kann. Jörg und Michael, für eure langjährige Verbundenheit. Auch wenn wir uns alle vier einmal lange nicht gesehen haben, ist es doch dann schnell wieder so nah wie bei den gemeinsamen Schulferien am Mitterberg. Meinen langjährigen Freunden in Österreich – Gernot, Franz und Martin, danke ich für eure Verbundenheit auch auf die Distanz, die Gespräche über Skype und gemeinsamen Unternehmungen. Tabea, dir danke ich, für deine seelische wie auch fachliche Unterstützung. Viele der Texte in dieser Doktorarbeit hast du mehrmals korrigiert und mit deinen Sprachkenntnissen weitreichend verbessert. Du bist die wichtige Stütze und starke Frau an meiner Seite!

184

Contributions & Funding

Chapter 1 P.H.W.B. and M.T. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B. and M.T. wrote the paper.

Chapter 2 P.H.W.B., K.P. and M.T. designed research; P.H.W.B. and K.T. performed research; P.H.W.B. analyzed data; and P.H.W.B. and M.T. wrote the paper.

Chapter 3 P.H.W.B. and M.T. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B. and M.T. wrote the paper.

Chapter 4 P.H.W.B. designed research, performed research, analyzed data and wrote the paper.

Chapter 5 P.H.W.B. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B., K.D.K. and M.T. wrote the paper.

Chapter 6 P.H.W.B. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B. and M.T. wrote the paper.

Chapter 7 P.H.W.B. and K.D.K. designed research; P.H.W.B. and D.L.S. performed research; P.H.W.B. analyzed data; and P.H.W.B., M.T. and D.L.S. wrote the paper.

Chapter 8 P.H.W.B. and H.H.d.F.L. designed research; P.H.W.B. and H.H.d.F.L. performed research; H.H.d.F.L. analyzed data; and P.H.W.B. and H.H.d.F.L. wrote the paper.

Chapter 9 P.H.W.B. provided research material.

Appendix 1 P.H.W.B. designed research; P.H.W.B. and D.L.S. performed research; P.H.W.B. analyzed data; and P.H.W.B., M.T. and D.L.S. wrote the paper.

Appendix 2 P.H.W.B. and C.R.C. designed research; P.H.W.B. performed research; P.H.W.B. analyzed data; and P.H.W.B. wrote the paper.

This project was funded by the following organizations:

• Austrian Academy of Science (ÖAW)

• Roche Research Foundation

• US-Department of Agriculture (USDA)

• Department of Bacteriology, University of Wisconsin-Madison

• Phil. nat. Faculty, University of Bern

• Swiss Zoological Society (SZS)

• Conférence universitaire de Suisse occidentale (CUSO)

• Ethological Society

• International Union for the Study of Social Insects (IUSSI)

185

186

187

188

Curriculum Vitae

Name: Peter Hans Wilhelm Biedermann

Date of birth: 21st June 1981 in Leoben, Austria

Nationality: Austrian

Address: Lorystrasse 12, CH-3008 Bern, Switzerland tel.: +41 31 534 9019 e-mail: [email protected]

E D U C A T I O N

1/2009- Ph.D. candidate Thesis on “The Evolution of Cooperation in Ambrosia beetles” Institute of Ecology & Evolution, University of Bern, Switzerland; Supervisor: Prof. Dr. Michael Taborsky

7/2007 Master degree Master of Science in Ecology and Evolution, awarded with distinction (“summa cum laude”)

4/2005-7/2007 Study of Ecology and Evolution Thesis on “Social behaviour of Xyleborinus saxesenii” Institute of Zoology, University of Bern, Switzerland; Supervisor: Prof. Dr. Michael Taborsky

1/2005 Bachelor degree Bachelor in Biology; Thesis on “Hidden leks in the yellow browed warbler Phylloscopus inornatus”; Institute of Zoology, University of Graz, Austria; Supervisor: Prof. Dr. Heiner Römer

2002-2005 Study of Biology/Zoology Major in Ethology and Neurobiology; University of Graz, University of Vienna, Austria

2000 -2001 Study of Physics University of Graz, Austria

1999 School leaving examination (Matura) awarded with distinction, BRG Leoben, Austria

F U R T H E R E X P E R I E N C E S

4-8/2009 Visiting researcher “Microbial communities of Ambrosia beetles” Department of Bacteriology, UW Madison/WI, USA; Supervisor: Ass.-Prof. Dr. Cameron Currie

8-11/2004 Internship on “Sexual selection in blue tits” Max Planck Institute for Ornithology, Seewiesen, Germany; Supervisors: Prof. Dr. Bart Kempenaers

4-8/2004 Project on “Drone (Apis mellifera) – swallow interactions” Institute of Zoology, University of Graz, Austria. supervisor: Prof. Dr. Karl Crailsheim

2-3/2004 Field assistant “Neurophysiology of katydids in relation to predation by bats” Smithsonian Tropical Research Institute, BCI, Panama; Supervisor: Prof. Dr. Heiner Römer

4-8/2003 Project on “Breeding-ecology of Yellow-browed Warblers” in Mongolia University of Göttingen, Germany, funded by the DAAD; supervisor: Prof. Dr. M. Mühlenberg

8-10/2002 Field assistant Thunder Cape Bird Observatory ThunderBay/ON, Canada; banding of songbirds and raptors, bird observations

6-8/2001 & 2002 Field assistant “Sex differences in Sandhill-cranes” Intern. Crane Foundation, Baraboo/WI, USA; banding, radio-tracking cranes, prairie restoration

189

Curriculum Vitae

5/2001 Field assistant for whale and dolphin research Tethys organization, Mediterranean sea

2000-2005 Bird monitoring for “Atlas of breeding-birds of Graz” Museum of Natural History (Joanneum), Graz, Austria

P O S I T I O N S H E L D

1/2009- Ph.D. student with Prof. Michael Taborsky Institute of Ecology & Evolution, University of Bern, Switzerland

8-12/2007 Visiting researcher with Dr. Kier Klepzig Southern Research Station, USDA Forest Service, Pineville/LA, USA

I N V I T E D T A L K S A N D S E M I N A R S

2011 - Max-Planck Institute for Chemical Ecology, Jena, Germany

2010 - Department of Environmental Sciences, ETH Zurich, Switzerland - Institute for Zoology, University of Regensburg, Germany - Seminar in Ecology & Evolution, University of Neuchatel, Switzerland - Zoological Colloquium, University of Graz, Austria

2007 - Section of Integrative Biology, University of Texas, Austin, USA - Department of Bacteriology, University of Wisconsin, Madison, USA

T A L K S A T M E E T I N G S A N D C O N F E R E N C E S

2011 - Genetics of Bark Beetles and Associated Microorganisms (IUFRO), Sopron, Hungary - European Society of Evolutionary Biology (ESEB), Tübingen, Germany (Poster) - Deutsche Gesellschaft für allgemeine und angewandte Entomologie (DGaaE), Berlin, D - Austrian Entomological Society (öEG), Graz, Austria - Biology2011, Zurich, Switzerland

2010 - International Union for the Study of Social Insects (IUSSI), Kopenhagen, Denmark - COST meeting, Diversity of symbioses in arthropods, Bad Bevensen, Germany (Poster) - Biology2010, Neuchatel, Switzerland

2009 - International Union for the Study of Social Insects (IUSSI), Frauenchiemsee, Germany* * Awarded with the Kutter-prize - European Society of Evolutionary Biology (ESEB), Torino, Italy (Poster) - 57th Annual Meeting Entomological Society of America, Indianapolis, USA (Virtual Poster)

2008 - Behavioural Biology, Dijon, France - Ethological Society, Regensburg, Germany

2007 - East Texas Forest Entomology Seminar (ETFES07), Nacogdotches/TX, USA - International Union of Forest Research (IUFRO07), Vienna, Austria - D-Day, Lausanne, Switzerland (Poster) - Meeting of PhD Students in Evolutionary Biology (EMPSEB13), Lund, Sweden - Xylobionten-Meeting, Bern, Switzerland - DZG and Ethological Society, Göttingen, Germany - Biology2007, Zurich, Switzerland (Poster)

2006 - Ethological Society, Bielefeld, Germany (Poster)

2005 - International Union for the Study of Social Insects (IUSSI), Halle, Germany - Zoological Colloquium, Graz, Austria

190

Curriculum Vitae

List of Publications

P U B L I C A T I O N S (peer reviewed)

(1) Biedermann PHW (2006) Hidden leks in the Yellow-browed Warbler (Phylloscopus inornatus)? - Investigations from the Khan Khentey Reserve (Mongolia). Acrocephalus 27: 233-247.

(2) Biedermann PHW & Kärcher MH (2008) Weather-dependent activity and flying height of Barn Swallows (Hirundo rustica) and House Martins (Delichon urbica) in southwestern Styria. Egretta 50: 76-81.

(3) Kärcher MH, Biedermann PHW, Hrassnigg N & Crailsheim K (2008) Predator-prey interaction between drones A. m. carnica and swallows Hirundo rustica or Delichon urbica. Apidology 39(3): 302-309.

(4) Delhey K, Peters A, Biedermann PHW & Kempenaers B (2008) Optical properties of the uropygial gland secretion: no evidence for UV cosmetics in birds. Naturwissenschaften 95(10): 939-946.

(5) Biedermann PHW, Klepzig KR & Taborsky M (2009) Fungus cultivation by ambrosia beetles: Behavior and laboratory breeding success in three Xyleborine species. Environmental Entomology 38(4): 1096- 1105.

(6) Biedermann PHW (2010) Observations on sex ratio and behavior of males in Xyleborinus saxesenii Ratzeburg (Scolytinae, Coleoptera). In: Cognato AI, Knížek M (Eds) Sixty years of discovering scolytine and platypodine diversity: A tribute to Stephen L. Wood. Zookeys 56: 253-267.

(7) Grubbs K.J., Biedermann P.H.W., Suen G., Adams S.M., Moeller J.A., Klassen J.L., Goodwinm L.A., Woyke T., Munk A.C., Bruce D., Detter C., Tapia R., Han C.S. & Currie C.R. (2011): Complete Genome Sequence of Streptomyces cf. griseus (XyelbKG-1 1), an Ambrosia Beetle-Associated Actinomycete. Journal of Bacteriology 193(11): 2890–2891.

(8) Biedermann PHW, Klepzig KD & Taborsky M (2011): Costs of delayed dispersal and alloparental care in the fungus-cultivating ambrosia beetle Xyleborus affinis Eichhoff (Scolytinae: Curculionidae). Behavioral Ecology & Sociobiology 65:1753–1761.

(9) Biedermann PHW & Taborsky M (2011): Larval helpers and age polyethism in ambrosia beetles. Proceedings of the National Academy of Science USA 108(41): 17064-17069.

(10) Biedermann PHW, Peer K & Taborsky M (2011): Female dispersal and reproduction in the ambrosia beetle Xyleborinus saxesenii Ratzeburg (Coleoptera; Scolytinae). Mitteilungen der Deutschen Gesellschaft für allgemeine und angewandte Entomologie 18: in press.

Submitted

(11) De Fine Licht HH & Biedermann PHW (submitted): Patterns of functional enzyme activity in Xyleborinus saxesenii fungus-growing ambrosia beetles. Frontiers in Zoology, in review.

O T H E R S

(1) Biedermann PHW (2003) Die Kraniche der Welt. Zool. Newsletter 2; Landesmuseum Joanneum Graz.

191