Interactions Between Liometopum Microcephalum (Formicidae) and Other Dominant Ant Species of Sympatric Occurrence

MASARYK UNIVERSITY FACULTY OF SCIENCE DEPARTMENT OF BOTANY AND ZOOLOGY

Populations of the ant Liometopum microcephalum (Panzer, 1798) at different spatial scales

Ph.D. Dissertation

Lenka Petráková

Supervisor: Doc. Jiří Schlaghamerský, Ph.D. Brno 2016

Bibliographic Entry

Author Mgr. Lenka Petráková Faculty of Science, Masaryk University Department of Botany and Zoology

Title of Thesis: Populations of Liometopum microcephalum (Panzer, 1798) at different spatial scales

Degree programme: Biology

Field of Study: Zoology

Supervisor: Doc. Jiří Schlaghamerský, Ph.D.

Academic Year: 2015/2016

Number of Pages: 166

Keywords: Liometopum microcephalum, ant colonies, competitors, territories, worker size polymorphism, trophobiosis, dispersal, genetic diversity, South Moravian population, distribution range, lineage divergence, Pleistocene refugia

Bibliografický záznam

Autor: Mgr. Lenka Petráková Přírodovědecká fakulta, Masarykova univerzita Ústav botaniky a zoologie

Název práce: Populace mravence Liometopum microcephalum (Panzer, 1798) na různých prostorových škálách

Studijní program: Biologie

Studijní obor: Zoologie

Vedoucí práce: Doc. Jiří Schlaghamerský, Ph.D.

Akademický rok: 2015/2016

Počet stran: 166

Klíčová slova: Liometopum microcephalum, mravenčí kolonie, konkurenti, teritoria, velikostní polymorfismus dělnic, trofobioza, šíření, genetická diverzita, jihomoravská populace, areál druhu, rozdílnost linií, pleistocenní refugia

Abstract

The ant Liometopum microcephalum is scattered from Italy in the west to the Lower Volga River in Russia and western Iran in the east, from South Moravia in the Czech Republic in the north to northern Israel in the south. The studied species is the only representative of the genus Liometopum occurring in Europe, albeit 14 fossil species of this genus have been found here. Despite the fact that this arboricolous species is locally ecologically important and its colonies consist of hundreds of thousands of workers, several questions regarding its biology remain unknown. The present thesis includes studies on L. microcephalum conducted at three spatial scales: local situations of one or several neighbouring colonies, the South Moravian population and the entire distribution area. Individual colonies within the South Moravian population were studied in regard to territory size and shape, interactions with competitors, trophobiosis with aphids, and morphological and genetic variability. The morphological and genetic variability was compared among colonies from different sites in South Moravia. The relatedness among colonies and populations was further assessed across the species’ distribution range to explore the phylogeography of the species. The thesis includes three scientific papers, which were published in ISI-indexed journals, and one submitted manuscript, which has already undergone a first review. The first paper deals with the position of the species in the hierarchy of ant communities and with its interactions with two other territorial ants, Lasius fuliginosus and Formica rufa, occurring in the same habitats. Liometopum microcephalum was able to withstand the competitive pressure of these competitors. Aggression was observed particularly in the beginning of the vegetation season, when the ants fought about foraging trees. Later in the season, conflicts occurred only when workers of different species were experimentally transferred into a competitor’s territory. Otherwise the competitors avoided each other. When defending its territories, L. microcephalum profited from the cooperation of its workers, which were biting the body of their rivals with their strong mandibles. The aim of the second study was to clarify the extent of morphological variability within and among L. microcephalum colonies. In all fifteen studied colonies a broad range of worker sizes was recorded. Variability was continuous and there were no distinct morphs. Several degrees of polymorphism were found in the studied colonies. These degrees did not correlate with the measured variables, but often varied in the course of time. Workers sampled in spring were significantly bigger than the workers collected in summer. There was a positive relationship between worker sizes and the colonies’ territory areas. In the third study the importance of honeydew intake in the studied species was evaluated by measuring the amount of total and reducing sugars in ant gasters of workers collected in different situations. The highest amount of sugars occurred in workers descending from foraging trees. This confirmed that the ants visit those trees to tend aphids. The amount of sugars in the gasters increased from April to June, but decreased rapidly in July. That is apparently related to the aphids’ life cycles and to the increased need of protein-rich food. In the fourth study the relatedness of colonies sampled at 40 sites across the distribution range was determined. In the studied colonies, 36 mitochondrial haplotypes were found. The most recent common ancestor of all studied colonies lived almost four million years ago. The Levantine population differed distinctly from the rest of the samples, having evolved separately since the Pliocene. The species probably dispersed from Anatolia in north-western direction. The origin of the European clades is related to the Pleistocene climate oscillations. The species survived glaciations in traditional Mediterranean refugia (the Balkan and Apennine Peninsulas) but also within the Carpathian Arc, at the Black Sea coast and at the south-eastern margin of the Alps. Abstrakt

Mravenec lužní (Liometopum microcephalum) je ostrůvkovitě rozšířený od Itálie na západě po dolní tok Volhy v Rusku a západní Írán na východě, od Jižní Moravy na severu po severní Izrael na jihu. Studovaný druh je jediným zástupcem svého rodu v celé Evropě, ačkoli zde bylo nalezeno 14 fosilních druhů tohoto rodu. Přestože je tento arborikolní druh lokálně ekologicky významný a jeho kolonie čítají až stovky tisíc dělnic, různé aspekty jeho biologie jsou stále neznámé. Práce zahrnuje studie druhu L. microcephalum provedené na třech prostorových měřítkách: lokální situace u jedné nebo několika sousedících kolonií, v jihomoravské populaci a v rámci celého areálu druhu. U jednotlivých kolonií uvnitř jihomoravské populace byla studována velikost a tvar teritorií, vztahy s konkurenty, trofobioza se mšicemi, morfologická a genetická variabilita. Morfologická a genetická variabilita byla porovnávána mezi koloniemi z různých stanovišť na jižní Moravě. Příbuznost kolonií a populací pak byla sledována napříč areálem druhu s cílem objasnit fylogeografii druhu. Součástí práce jsou tři vědecké články publikované v časopisech s impakt faktorem (WoS) a jeden rukopis zaslaný do redakce, který již prošel první recenzí. První studie se zabývá pozicí druhu v hierarchii mravenčích společenstev a jeho vztahy ke dvěma dalším teritoriálním druhům, Lasius fuliginosus a Formica rufa, které obývají stejný habitat. Liometopum microcephalum dokázal obstát v konkurenci těchto soupeřů. Agresivní chování bylo pozorováno zejména na počátku vegetační sezóny, kdy docházelo k bojům o potravní stromy. Později ke střetům docházelo pouze tehdy, pokud byly jedinci jiného druhu experimentálně přeneseni do konkurentova teritoria. Jinak se všechny studované druhy svým konkurentům vyhýbaly. Při obraně svých teritorií L. microcephalum těžil ze spolupráce mezi dělnicemi, které vetřelce napadaly svými silných kusadly. Druhá studie měla za cíl objasnit míru morfologické variability uvnitř a mezi koloniemi mravence lužního. U všech patnácti studovaných kolonií byla zjištěna široká škála velikostí dělnic. Proměnlivost byla plynulá, bez jasně vymezených morf. U jednotlivých kolonií byly zjištěny různé stupně polymorfismu. Tyto stupně nesouvisely s žádnou z měřených proměnných, měnily se však v čase. Dělnice vzorkované na jaře byly významně větší než dělnice sbírané v létě. Byla potvrzena pozitivní korelace mezi velikostí dělnic a rozlohou teritoria kolonie. Ve třetí studii byl hodnocen význam příjmu medovice u mravence lužního pomocí množství celkových a redukujících cukrů v gasterech dělnic sbíraných v různých situacích. Nejvyšší obsah cukrů byl naměřen u dělnic sestupujících z potravních stromů. Tím se potvrdilo, že mravenci tyto stromy navštěvují za účelem dojení mšic. Obsah zjištěných cukrů narůstal od dubna do června, v červenci však rapidně poklesl. To patrně souvisí s generačními cykly mšic a zvýšenou potřebou proteinů v potravě. Ve čtvrté studii byla zjišťována příbuznost kolonií ze 40 lokalit napříč areálem druhu. U studovaných kolonií bylo nalezeno celkem 36 různých mitochondriálních haplotypů. Nejbližší společný předek všech studovaných kolonií žil před téměř čtyřmi miliony lety. Populace z Levanty se zřetelně odlišovala od ostatních vzorků, vyvíjela se samostatně od Pliocénu. Druh se pravděpodobně šířil z Anatolie severozápadním směrem. Původ evropských kládů souvisí s pleistocenními oscilacemi klimatu. Druh přežíval glaciály v tradičních mediteránních refugiích (Balkánský a Apeninský poloostrov), ale i v Karpatském oblouku, na pobřeží Černého moře a pod jihovýchodním okrajem Alp.

Acknowledgements

I would like to thank all those people, who have trusted me and believed that my work had sense. My special thanks go to my supervisor Jiří Schlaghamerský for his ideas, advice, comments, and especially for his patience with corrections of my texts written in English. I´d like to thank him also for the opportunity to study this remarkable ant species and to begin with DNA-based methods. A big thank you goes to Andrea Tóthová, who supervised me in the molecular lab in the beginning of my Ph.D. studies and showed me how to analyse DNA data. Thanks to Kristina Civáňová and Kristýna Koukalová for their help with sequencing, and to Tatiana Aghová, who advised me how to run BEAST analyses. Thanks to Stano Pekár and Víťa Syrovátka for their help with statistical modelling and to Ondra Hájek for creating maps for the manuscripts. Big thanks go to my love Petr Silvestr Dušátko, who has brought balance and happiness into my life. Thank you, Petr, for your support and patience, especially during the last months, during which I spent almost every evening and weekend working. Thanks to my family that they exist and that they gave me a chance to live better than they do. Many friends of mine kept their fingers crossed, I thank also to them and I apologize that I was often too busy and had to give priority to my work instead of joining them in other activities.

The thesis was financially supported by the Ministry of Education, Youth and Sports of the Czech Republic, Research Plan No. MSM0021622416 and by the Czech Science Foundation, Grant No. GD526/09/H025.

Table of Contents

1. Introduction ...... 1 2. Distribution of Liometopum microcephalum ...... 5 3. Methods ...... 11 3.1 The studies in South Moravia ...... 11 3.2 The study across the species’ range ...... 13 4. Study A: Interactions with other dominant ants ...... 15 5. Study B: Worker size polymorphism ...... 35 6. Study C: Trophobiosis...... 83 7. Study D: Genetic diversity of the South Moravian population ...... 95 8. Study E: Phylogeography of the species ...... 101 9. Conclusions ...... 129 10. References ...... 131 11. Author’s contributions to the papers ...... 147 12. List of publications ...... 149 1. Introduction

Ants (Hymenoptera, Formicidae) are one of the most important insect groups, having strong impact on their environment and producing a huge amount of biomass. They affect nutrient fluxes in ecosystems (for example, they may facilitate decomposition of organic matter) and the composition of plant and animal communities (e.g. by dispersing seeds of flowering plants and by interacting with other animals, for example defending their own nests or tending aphids). However, each species has its unique requirements and its real effect depends on many factors, such as colony size, preferred habitat, food type, or physical conditions of the site. The oldest ant fossils (from the subfamily Sphecomyrminae), are known from the Early Cretaceous, approximately 90-100 million years ago (LAPOLLA et al. 2013). Since that time, the ants have undergone a very long journey forming their ways of life, diverging in appearance and filling free ecological niches. MOREAU et al. (2006) proposed that extant subfamilies began to diversify during the Late Cretaceous or Early Eocene, at the same time when flowering plants arose. So far, more than 13,000 extant ant species classified into 16 subfamilies have been described (BOLTON 2014), but the actual species number has been assessed as much higher, perhaps exceeding 20,000 species (HÖLLDOBLER & WILSON 1990). Liometopum ants (also called velvety tree ants in English) are considered ecologically important albeit rare and understudied. All species have a similar way of life, forming huge colonies that aggressively defend their nests, which are situated either in tree trunks or in soil. In the last decade, in particular North American species of the genus Liometopum have received increasing attention as it is shown by the number of studies concerning their biology (e.g. WANG et al. 2010, HOEY-CHAMBERLAIN et al. 2013, CRUZ-LABANA et al. 2014, HOEY-CHAMBERLAIN & RUST 2014) and their importance as structural pests in wood (e.g. KLOTZ et al. 2010). Moreover, L. apiculatum is exploited economically in Mexico – its larvae and pupae called “escamoles” are collected and prepared as a traditional meal. Their nests are disturbed even twice a year and high demand for “escamoles” threatens their populations (LARA-JUÁREZ et al. 2015). The genus Liometopum belongs to the subfamily Dolichoderinae, which includes 707 extant species in 28 genera, 134 fossil species and 20 fossil genera (BOLTON 2014). Dolichoderinae dominate in the fossil records from the Tertiary; since then their numbers have been decreasing (WARD et al. 2010). The genus Liometopum is generally considered to be an ancient one. Regarding its distribution, age and number of extant and fossil species, a relict character of the genus had been suggested (e.g. RADCHENKO 2005). At present, seven extant species of the genus Liometopum are considered valid, and twenty fossil species are recognized (DEL TORO et al. 2009, BOLTON 2014). All extant species of the genus occur exclusively in the northern hemisphere, inhabiting rather islands of suitable habitats (e.g. oak or pine forests) than vast continuous areas. Three species (L. apiculatum Mayr 1870, L. luctuosum Wheeler 1905 and L. occidentale Emery 1895) inhabit the western part of North America, partially south to the Central American Yucatan Peninsula, three species (L. lindgreeni Forel 1902, L. orientale Karavaiev 1927 and L. sinense Wheeler 1921) are known from the (south-)eastern part of Asia. Only one species (L. microcephalum Panzer 1798) occurs in Europe although fourteen fossil species are known from there. Liometopum oligocenicum (Wheeler, 1915) was found in Baltic amber from the Eocene. Four fossil species. i.e. Liometopum brunascens (Heer, 1867), L. croaticum (Heer, 1849), L. imhoffii (Heer, 1849), and L. pallidum (Heer, 1867), dating to the Early Miocene, were found in

Croatia (HEER 1849, DLUSSKY & PUTYATINA 2014). Seven fossil species, i.e. L. crassinervis Heer, 1849; L. escheri (Heer, 1867), L. globosum (Heer, 1849), L. longaevum (Heer, 1849), L. stygium (Heer, 1867), L. venerarium (Heer, 1864), L. ventrosum (Heer, 1849) from the Miocene were discovered in Switzerland (HEER 1849, HEER 1865, HEER 1867). In Germany, MEUNIER (1917) found a species, Liometopum rhenana (Meunier, 1917), from the Tertiary. Another fossil species, L. bogdassarovi (Nazaraw, Bagdasaraw & Uriew 1994), was found in Belarus (HETERICK & SHATTUCK 2011). Most of those species were originally described within the genus Poneropsis, which has been recently declared a junior synonym of Liometopum (DLUSSKY & PUTYATINA 2014). Liometopum microcephalum was originally described as Formica microcephala PANZER 1798 based on a male collected in Austria. That explains the origin of the name “microcephala” as ant males generally have rather small heads. Later, Formica austriaca MAYR 1853 was described based on a worker found in Austria. In 1861, MAYR found a queen and classified all those specimens as a single species, Liometopum microcephalum. As mentioned above, the species is the only representative of the genus occurring in Europe (and in the adjacent part of western Asia). Its closest relatives inhabit the Far East in Asia: Liometopum orientale KARAVAIEV 1927 occurs in the Primorski region of Russia and in North Korea (KUPYANSKAYA 1988, RADCHENKO 2005), Liometopum sinense WHEELER 1921 is endemic to China (DEL TORO et al. 2009, GUÉNARD & DUNN 2012) and Liometopum lindgreeni FOREL 1902 lives in the north- eastern part of India (Assam, Meghalaya and Arunachal Pradesh states; MATHEW & TIWARI 2000; BHARTI et al. 2016) and in Myanmar (MATHEW & TIWARI 2000). Liometopum microcephalum is patchily distributed across the Ponto-Mediterranean region. Its present distribution area reaches from the Apennine Peninsula in the west up to Lower Volga region in Russia and western Iran in the east. The southern border is delimited by the Mediterranean Sea continuing to northern Israel, and the northern border passes through Austria, the Czech Republic, Slovakia and the Ukraine up to the Lower Volga River in Russia. The populations are scattered in fragmented islands of suitable habitats, often isolated by several hundreds of kilometres. Information about its occurrence in most countries is scarce or outdated, in some countries it has been red-listed. Records on L. microcephalum distribution were summarized by M. OMELKOVÁ in her master thesis (OMELKOVÁ 2005). She presented a detailed list of sites based on records that had been published mainly in the first half of the 20th century. Recently, several ant checklists of eastern European countries were published (e.g. LAPEVA-GJONOVA et al. 2010, KARAMAN 2011, BOROWIEC & SALATA 2012, BRAČKO et al. 2014 a, b). Unfortunately, erroneous and outdated data about the species’ occurrence are cited in some of the checklists (e.g. BOROWIEC & SALATA 2012, BOROWIEC 2014), and that presents a serious problem. Moreover, geographic names are sometimes given with mistakes (e.g. many cases in DEL TORO et al. 2009). Therefore I have tried to complete a detailed survey of the species’ present occurrence in individual countries (see the Chapter 2, Distribution of Liometopum microcephalum). The information is based on published records, checklists and personal explorations. Liometopum microcephalum is a noteworthy species with specific habitat requirements. It is a thermophilous ant, and particularly in the north of its range it is restricted to warm floodplains of large rivers or other humid and warm sites. It is strictly associated with large trees, occurring both in forests as well as in open areas. Its carton nests are situated in cavities of tree trunks or thick limbs, most often high above ground (MAKAREVICH 2003). This could present some kind of 2 adaptation to floods. It strongly prefers oaks but nests have been found also inside chestnuts, poplars, ashes and maples (SCHLAGHAMERSKÝ & OMELKOVÁ 2007), generally in trees with rough bark. The species forms very populous colonies. In the hierarchy of ant communities, L. microcephalum is a dominant species (PETRÁKOVÁ & SCHLAGHAMERSKÝ 2011). Its territories are defended by aggressive workers against other ants, affecting the distribution and structure of invertebrate communities (GRABENWEGER et al. 2005, TARTALLY 2006, PETRÁKOVÁ & SCHLAGHAMERSKÝ 2011). Its territories consist of nest trees, foraging trees, food resources and trails connecting these places. The ants forage mainly on trees, termed foraging trees when not hosting their nest. They are predators with very good visual orientation (preying on Hexapoda, Myriapoda, Araneae, Lumbricidae; WIEST 1966, 1967), scavengers, and were also observed feeding on nectar (e.g. maples, pers. obs.) and tending aphids (trophobiosis; MAKAREVICH 2003, SCHLAGHAMERSKÝ et al. 2013). Its colonies are apparently polydomous but information about colony founding is scarce and ambiguous. Although WIEST (1967) observed females laying eggs, the females did not survive more than five months. WIEST suggested that the females have to overwinter in their natal nest. Mating occurs from May to July (SEIFERT 2007). Males and winged females were observed in June and July outside the nest both at the bottom of tree trunks (own unpublished observations, J. SCHLAGHAMERSKÝ and A. TARTALLY, pers. comm.) as well as in tree crowns (D. HORAL, pers. comm.). Due to the inaccessibility of the nest it is unknown what happens inside, and also the number of queens in a colony remains unexplored. Very few studies on the biology of L. microcephalum have been published. WIEST studied its gland structures and behaviour in her PhD thesis (WIEST 1966), one year later she published her observations on sexual individuals and the species’ foraging trees and prey (WIEST 1967). Forty years later, MAKAREVICH (2003), published a study about several colonies from the Lower Dnepr River in the Ukraine, dealing mainly with space partitioning of the colonies, i.e. the exploitation of their nest and foraging trees. Although these studies present important findings about the species’ biology, the latter two are based on direct observations and lack a detailed description of methodology. Only a few colonies at single sites were studied and thus these studies lack any comparison of different habitats. SCHLAGHAMERSKÝ & OMELKOVÁ (2007) studied the species distribution in South Moravia (which represents the northwestern-most part of the species range), trying to explain the species’ preference for nest sites. I decided to study some of the remaining unknown (or unclear) aspects of L. microcephalum biology. In my master theses I focused on interactions of the species with other sympatrically occurring and behaviourally dominant ants, both in the field and under laboratory conditions (PETRÁKOVÁ & SCHLAGHAMERSKÝ 2011). Although the data were collected for my master thesis, and first presented herein, a more sophisticated approach of data analysis was required before the results could be published. In the first phase of my PhD study I continued working on this topic, conducting these new statistical analyses and writing the manuscript. The study itself has brought important findings about the species’ ecology and I often refer to the corresponding paper in my subsequent studies. For the above reasons I am including the published paper as a part of my present thesis. Due to the nests’ position inside tree trunks and due to workers’ aggressive protection of their nest trees against any intruders, the colony structure has never been studied in detail. In particular, the use of molecular markers, as a non-invasive technique, could help us to resolve some

of the questions regarding the relatedness of workers, colonies and populations (having unknown sizes). I have taken into account mainly three spatial scales:

(1) Individual colonies from different sites within the South Moravian subpopulation were studied with the aim to describe intracolonial genetic and morphologic variation. a) Does the ant have any distinct worker castes? b) How many reproducing queens live in a single colony?

(2) Comparing the colonies among sites within the South Moravian subpopulation a) Is there morphological variability among the colonies and if so, is it affected by ecological factors? b) What is the genetic variation within the South Moravian subpopulation of L. microcephalum?

(3) Samples from different sites across the species distribution area were compared to assess a) What is the genetic diversity of populations across the species range? b) What is the origin of the individual subpopulations? c) Which colonisation routes could the species have used when dispersing to its present range?

All but one study resulted in scientific papers. Each paper is preceded by a brief introduction explaining why I studied the topic, why I used particular methods and what were the most important results. The details are presented in the papers, thus it would be redundant to repeat exactly the same information in detail in these introductory sections. In addition, I have included findings or details (such as graphs and figures of territories) which were made as important parts of my research but did not become part of the published papers, mostly because of space limitations. The study on the genetic diversity of the South Moravian population has not been published because I did not obtain positive results due to inappropriate genetic markers used (which unfortunately were the only available). The corresponding chapter is therefore structured differently from the other parts of this thesis, with several subsections corresponding to introduction, methods, results and discussion. The phylogeographic study is presented as a manuscript re-submitted to the Journal of Biogeography. This is the second submitted version, modified based on the comments of the editor and three reviewers. After another revision it has been designated as “requiring minor revision” (Journal of Biogeography, 6th May 2016). I am also including a paper presenting results of a study on trophobiosis in L. microcephalum. I participated in that study, being responsible for the final statistical analyses that were presented in the paper. This was not one of the original aims of my PhD thesis, but presents a useful complement by looking at another important aspect of the species’ ecological function and its interaction with other organisms.

2. Distribution of Liometopum microcephalum Italy: The species was reported from many regions at the end of the 19th century and during the first half of the 20th century. In the north, it was recorded north of Genova (Serravalle Scrivia; EMERY 1878, MANTERO 1889) and in the Venezia Giulia region (MOCSÁRY 1918, MÜLLER 1921, MÜLLER 1923; at present the western part of that region belongs to Italy, the northeast is a part of Slovenia and the south-east belongs to Croatia). It was common in the province of Modena and in Reggio Emilia (MAYR 1855, EMERY 1878, CECCONI 1903, MENOZZI 1918, MENOZZI 1924, GRANDI 1935). It occurred only rarely along the western and eastern coasts in the middle part of the peninsula, and was missing at higher altitudes of the Apennine Mts. (EMERY 1869, FOREL 1909, CASTELLANI 1937, MENOZZI 1921, MENOZZI 1942, CONSANI 1949, BARONI URBANI 1964). It occurred also in Calabria (EMERY & CAVANNA 1880, MENOZZI 1921, BARONI URBANI 1964). DE STEFANI (1888) and EMERY (1915, 1916) reported L. microcephalum from the Madonie Mountains in Sicily. Concerning the present distribution, no details have been published. In the last checklist of the Italian ant fauna (POLDI et al. 1995), the species was reported from „the north, the south and from Sicily“. Recently, it undoubtedly occurs in eastern Lombardy in Mantova (own observation; JIŘÍ SCHLAGHAMERSKÝ, pers. comm.), south of Firenze (VERONIKA JÍLKOVÁ, pers. comm.), Perugia in central Italy (MEI 2016), near Rome (Bracciano National Park - FABRIZIO RIGATO, pers. comm.; Sughereta di San Vito - NICKLAS JANSSON, pers. comm.), in the Gargano National Park (JIŘÍ PROCHÁZKA, pers. comm.), and in Calabria (Pollino National Park - DAVID HAUCK, pers. comm.). It seems that it has disappeared from the nearby Bosco della Fontana site (near Mantova) where it was observed in 2003 (JIŘÍ SCHLAGHAMERSKÝ, pers. comm.). The species has not been found recently at several sites from which it had been reported in the past (ZANGHERI 1969): neither in Serravalle Scrivia, nor in Pineta di Classe near Ravenna and in Bosco di Scardavilla near Forli (JIŘÍ SCHLAGHAMERSKÝ, pers. comm.) I have looked for the ant species in the Friuli-Venezia Giulia region, in Sambiase (Calabria) and in Pompei, but without success. Potentially suitable nest sites were mostly occupied by Linepithema humile and Crematogaster scutellaris. Its present occurrence in Sicily remains questionable. It was not included in the checklist by DONISTHORPE (1926). My own search in the Madonie Mountains was not successful (though I was not able to explore the whole Madonie area).Furthermore, in the Natural History Museum in Castelbuono in the Madonie Regional Natural Park I have not found any specimen of that species. Austria: The species inhabits the Morava River floodplain in the north-eastern part of the country, but it is very rare there (SCHLICK-STEINER et al. 2003). The most known site is at Marchegg (MELANIE TISTA, pers. comm.). In the past, the species was abundant also in Wien and near Eisenstadt but has decreased substantially (WIEST 1967). DEL TORO et al. (2009) has examined a specimen collected by B.C. SCHLICK-STEINER in Laxenburg, just south of Vienna, in 2002. Recently, it has been found also along the Danube River in Donau- Auen National Park (JIŘÍ SCHLAGHAMERSKÝ, pers. comm.).

Czech Republic: The species is restricted to the region of South Moravia (south- eastern part of the country). Its occurrence has been well documented (e.g. ZDOBNITZKY 1910, SOUDEK 1922, KRATOCHVÍL 1936, KRATOCHVÍL 1938, ZÁLESKÝ 1939, BEZDĚČKA 1995, BEZDĚČKA 1996a). SCHLAGHAMERSKÝ & OMELKOVÁ (2007) published a detailed survey of its recent sites, and the information has been updated in SCHLAGHAMERSKÝ & PETRÁKOVÁ (2014). The species occurs along the lower Dyje and Morava rivers up to the village of Mikulčice in the east, to Vrkoč pond in the west (KRÁSA 2014) and to a forest east of Židlochovice in the north (DAVID HAUCK, pers. comm.). The latter site represents the northernmost point of its actual distribution in South Moravia (perhaps even in the whole distribution area). Despite of its position at the margin of the species’ distribution range, the South Moravian population is the largest one which has been documented so far, consisting of almost 900 colonies (SCHLAGHAMERSKÝ & OMELKOVÁ 2007). Slovakia: The distribution was well documented in the first half of 20th century (MOCSÁRY 1918, ZÁLESKÝ 1939). The species was known from several sites in the southern part of Slovakia: Levice, Rimavská Sobota, near Košice and on the Silica Plateau. BEZDĚČKA (1996b) published records from southwestern Slovakia, from sites along the Danube River. Recent information about its occurrence at the sites in the middle and south-eastern part of Slovakia is missing. The species was found in the Danube floodplains near Lučenec (pers. obs.) and near Kamenice nad Hronom (PAVEL PRŮDEK, pers. comm.), but my search on the Silica Plateau was not successful in summer 2012. In the past, it was recorded from Somotor village (MOCSÁRY 1918) which is situated close to the Latorica Protected Landscape Area. However, the species does not occur there anymore (S. MIŇOVÁ, pers. comm.) although that area with its warm floodplain forests resembles typical habitats of this species. Hungary: The species has been reported mainly from the north-eastern part of Hungary (Tiszadob, Túristvándi, Debrecen, Hortobágyi National Park; GALLÉ 1981, GALLÉ et al. 2005, ANDRÁSZ TARTALLY, pers. comm.) and from the area northeast of Szeged (Makó, Gyula, Körös; GALLÉ et al. 2005). It is abundant even in Budapest (MARKÓ & CSÖSZ 2002; DAVID HORAL, pers. comm.). It occurs also in Kiskunság National Park, north and northeast of Lake Balaton and along the Drava River (LÁSZLO GALLÉ, pers. comm.). It seems that the species still colonizes new sites on the Hungarian Plain (LÁSZLO GALLÉ, pers. comm.). Slovenia: The species occurs within the Sub-Pannonian region in the north-eastern part of the country near Lendava (close to the borders to Hungary and Croatia; BRAČKO 2007, GREGOR BRAČKO, pers. comm.). Croatia: The species was rather abundant in the past as is evident from the number of records published between the mids of the 19th and 20th centuries (MAYR 1855, KORLEVIĆ 1890, MÜLLER 1923, ZIMMERMANN 1934, VOGRIN 1955 in BRAČKO 2006). It was found at the coast of the Adriatic Sea (Učka, Rijeka, near Skradin, Split, near Dubrovnik) and near Zagreb. Detailed information about its present occurrence is missing. Recently it was collected in Lonjsko Polje, at the border with Bosnia (LUKÁŠ ČÍŽEK, pers. comm.) and in Zagreb (MATE ZEC, pers. comm.). According to ADI VESNIC (pers. comm.) it occurs also in the south, in the Imotski region.

Bosnia: The species occurs in the north of the country near Banja Luka, and in the south near Cernica (PETROV & COLLINGWOOD 1992), Drinovci (ADI VESNIC, pers. comm.) and south of Mostar (DEL TORO et al. 2009). Montenegro: The species occurs in the south of the country, especially close to the Skadar Lake (KARAMAN 2011, MARKO KARAMAN, pers. comm.), along the Cijevna river (BRAČKO et al. 2014a) and also in the north in Boka Kotorska near Herceg Novi (ADI VESNIC, pers. comm.). Serbia: The species lives in the north that is in the province of Vojvodina (PETROV & COLLINGWOOD 1992, PETROV 2002, KARAMAN & KARAMAN 2003, MARKO KARAMAN, pers. comm.). DEL TORO et al. (2009) examined material collected in Kosovo. Albania: No checklist of Albanian ants has been published. According to AGOSTI & COLINGWOOD (1987), there were no records of L. microcephalum from Albania. DEL TORO et al. (2009), however, examined a specimen from Çorovodë at Osum River. In 2010, it was found in the Drin i Zi River bank westward of Peshkopi (in the northeast of the country; D. HORAL, pers. comm.). Macedonia: According to KARAMAN (2000, 2009), the species was recorded in the east of the country in the beginning of the 20th century (Kalučkovo, Štip and Bajlik village in Belasica Mts.; DOFLEIN 1920). PETROV & COLLINGWOOD (1992) reported it from Dojran Lake in the south. Recently, BRAČKO et al. (2014b) found the species near the town Demir Kapija in the south-eastern part of the country. Greece: According to BOROWIEC & SALATA (2012), the species occurs in the Aegean and Ionian Islands, in Epirus, Macedonia, Central Greece including Thessaly, Euboia and Sporades Islands, and in the Peloponnese - that means in almost all provinces within the country, but they did not specify the sites in more detail (except two new records from the Olympus and Ossa Mts. in eastern Greece). Older records provided evidence of its occurrence in Lefkada (DEL TORO et al. 2009) and in the Greveniti and Taygetos Mts. (LEGAKIS 1984). We obtained samples from the western coast (from Amvrakikos Bay; JIŘÍ PROCHÁZKA, pers. comm.) and from the north-eastern part, i.e. from the environs of Xanthi and Arriana (GREGOR BRAČKO, pers. comm.). Bulgaria: The presence of L. microcephalum in Bulgaria has been rather well documented (FOREL 1892, ATANASSOV 1934, ATANASSOV 1936, ATANASSOV 1957, ATANASSOV 1964, ATANASSOV & DLUSSKIJ 1992, LAPEVA-GJONOVA 2004, LAPEVA- GJONOVA 2011). At present, the species is scattered within the whole country, particularly in the Danubian Plain, Stara Planina Mts., Sredna Gora Mts., Rila and Pirin Mts., Rhodopi Mts. and Strandzha Mts. (LAPEVA-GJONOVA et al. 2010). The species is abundant almost everywhere up to 1000 m a.s.l. (TOSHKO LJUBOMIROV, pers. comm.). We obtained samples from the northwest (near Vratsa), southwest (close to the border to Macedonia, TOSHKO LJUBOMIROV, pers. comm.) and the southeast (south-east of Kardzhali; DAVID HORAL, pers. comm.). Romania: No detailed information about the occurrence of L. microcephalum in Romania has been published. The species occurs east of Arad close to the border with Hungary (GALLÉ et al. 2005), near Timisoara (JIŘÍ PROCHÁZKA, pers. comm.), in Bihor

County and near Tăşnad (BÁLLINT MARKÓ pers. comm.). It has also been found at Târgu Jiu, in the Danube floodplain south of Bucharest, at the Black Sea coast and nearby Galati at the border with Moldova and the Ukraine (BÁLLINT MARKÓ, LUKÁŠ ČÍŽEK, pers. comm.). Moldova: To my knowledge, no comprehensive information on its historic or present distribution has been published. According to POPESCU & DAVIDEANU (2009) the species occurs along Prut River at the border with Romania (IRINEL EUGEN POPESCU, pers. comm.). In 2013 and 2014, it was observed and collected near Cahul (DAVID HORAL, pers. comm.). Ukraine: Liometopum microcephalum is very rare in the Ukraine (RADCHENKO 2011). Only four sites are known from the country: the Ukrainian Carpathians (without further details), forests along the Samara River in the Depropetrovsk region (APOSTOLOV & LIKHOVIDOV 1973), an island within the Danubian Delta in the Odessa region (AKIMOVA 2009) and Kozatsky island in the Lower Dnepr River near Nova Kakhovka in the Kherson Region (MAKAREVICH 2003). Russia: In the past, the species had been found near Tuapse on the north-eastern coast of the Black Sea (RUSZKY 1905). At present it was reported from several sites in the Lower Volga region near Volgograd and from Tsagan Aman (SAVRANSKAJA 1998, GREBENNIKOV 2007, DMITRY A. DUBOVIKOV, pers. comm), which represents the north-easternmost point of the species’ distribution area. Georgia: Up-to-date data about the occurrence of the species are lacking. The data in the latest checklist (GRATIASHVILI & BARJADZE 2008) are based on old records. The species was recorded near Tbilisi (NASONOV 1889, RUZSKY 1905, JIJILASHVILI 1964 in GRATIASHVILI & BARJADZE 2008) and from Baghdati and Zekari Pass in the western part of the country (RUZSKY 1905 in GRATIASHVILI & BARJADZE 2008). Turkey: The species is most abundant in European Turkey (Thrace), particularly in Edirne, Kırklareli, Bursa and Balikesir regions (AKTAÇ & ÇAMLITEPE 1987, ARAS & AKTAÇ 1987, ÇAMLITEPE & AKTAÇ 1987, ARAS & AKTAÇ 1990, AKTAÇ et al. 1994, KIRAN & AKTAÇ 2006, LAPEVA-GJONOVA & KIRAN 2012 in KIRAN & KARAMAN 2012). It is scattered in western Anatolia, in Hatay province in the south (NICKLAS JANSSON and KADRI KIRAN, pers. comm.) and in Siirt province in the south-east of Anatolia (AKTAÇ 1976, NICKLAS JANSSON, pers. comm.). Lebanon: DEL TORO et al. (2009) analysed a specimen collected in 1953 at the Hasbani River, situated at the border with Israel. TOHMÉ & TOHMÉ (2014) found the species near Quartaba (about 50 km southward of Tripoli). Israel: The species occurs in the north of the country: along the Tal stream (EYGEN AVITAL, pers. comm.) close to the Lebanese border, and close by the Sea of Galilee (KUGLER 1988, VONSHAK & IONESCU-HIRSCH 2009). Iran: The species was recently recorded in Saqqez in the western part of the country (PAKNIA & KAMI 2007, PAKNIA et al. 2008).

Disputable records: According to the new checklist of Greek ants (BOROWIEC & SALATA 2012) and the Catalogue of European Ants (BOROWIEC 2014), L. microcephalum occurs also in Portugal, the Balearic Islands, Corsica, Sardinia, mainland France and even in Switzerland. Unfortunately, the authors did not refer to any sources of this information. No specimens are known from Switzerland or Germany. MARTÍNEZ & TINAUT (2001) have definitely disclaimed the occurrence of L. microcephalum in Spain. The “record for Spain” was based on the European Catalogue of Hymenoptera (KIRCHNER 1867) and there was neither a voucher specimen nor information about the locality. The species has not been found in the Iberian Peninsula since the time of Kirchner´s record. After COLINGWOOD (1978) had studied the ants of the Iberian Peninsula, he doubted the species´ occurrence there. No records from Portugal have been ever published. BERNARD (1968) refused the possibility that the species’ could occur in mainland France, Corsica and Sardinia, and no further records have been published later. However, the crabronid wasp Tracheliodes varus, which is known to be a specialized hunter of L. microcephalum ants, was reported from Corsica (BITSCH & LECLERCQ 1993). According to ZETTEL et al. (2004) the wasps apparently must be hunting some other ant species there. The only record from the Algerian port of Annaba, formerly Bône (BERNARD 1968), was probably based on a temporary introduction (if we take into account the importance of this harbour since the Middle Ages). No other record is known from northern Afrika. Although Armenia is situated between Turkey and Georgia, no records of the species’ occurrence have been published. ARAKELIAN (1994) did not include the species in the list of Armenian ants although he included it in the associated identification key.

3. Methods

3.1 The studies in South Moravia

In South Moravia, L. microcephalum was studied at the colony level and at the South Moravian population level – the morphological and genetic variability was compared among colonies and among sites. I use the term South Moravian population for the colonies present in South Moravia, where they live on more or less spatially isolated islands of suitable habitat in less than 15 km (mostly less than 5 km) from each other. The closest Austrian and Slovak sites are situated about 50 km from the southernmost South Moravian sites. However, gene flow between these populations cannot be excluded. The colonies were sampled at 12 study sites. In this area the species inhabits at least three different habitat types, which differ mainly by the age of nest trees, by the distance between the trees and by humidity levels. The habitat types were as follows:

1) Floodplain forests (or at least moist forests; Figure 1a), where colonies of L. microcephalum were abundant. Oaks (Quercus robur) and ashes dominated in those forests, accompanied by elms, elders, maples and limes. The forests had dense undergrowth, including seedlings, and coarse woody debris was abundant. The study sites Křivé Jezero, forests near Lednice, forests near Pohansko, Ranšpurk forest reserve, and Sekulská Morava are situated in the Dyje River floodplain, whereas the forest at the Rendezvous National Nature Monument site (here mostly Quercus cerris) is moist due to the marshy banks of a small pond and high level of ground water. The forest east of Tvrdonice is situated between the Morava River and Kopanice Brook.

2) Moist meadows with solitary trees or trees in small groups are typical of the Lednice Castle Park (Figure 1b). The park is situated in the Dyje River floodplain, the Stará Dyje and Zámecká Dyje river branches flow through the park. The species lives here in mighty oaks more than 200 years old, which often have tree trunks exceeding one metre in diameter. There are also large limes, horse chestnuts, plane trees and conifers.

3) The xerothermous forest near Bulhary (Bulhary game enclosure) is relatively young, formed mainly by oaks (Quercus petraea), maples, and limes. Only three colonies were found, all in the lightest and oldest part of the forest in 70-110 year-old oaks.

In total, I sampled 41 colonies: five colonies at the Křivé Jezero site, three colonies at the Bulhary game-enclosure, eight colonies in the Lednice Castle Park, two colonies in the Lednice floodplain forest, five in Kančí Obora forest north of Břeclav, seven colonies in the Rendezvous forest, five in forests near Pohansko, five in the Ranšpurk forest reserve, and one at the Sekulská Morava site. Samples from two colonies inhabiting the northernmost sites (from a forest east of Židlochovice and a forest near Uherčice) were provided by David Hauck, Jiří Schlaghamerský sampled three colonies at Tvrdonice site and Jan Kašpar collected samples for sugar content analyses from several colonies at the Rendezvous site.

Figure 1: A moist forest habitat at Rendezvous site (at the top), park landscape habitat at Lednice Castle Park (in the middle), and the xerothernous forest near Bulhary (at the bottom); photographs by L. Petráková.

Sometimes it was difficult to distinguish between nest and foraging trees. Particularly in spring, high numbers of ant workers ascend almost all trees in the vicinity of their nest. Usually, the tree with the highest diameter is the nest tree, but there are exceptions. The exploitation of foraging trees changes during the ants’ annual activity period, whereas the nest trees are occupied throughout the whole activity period. The most reliable way how to detect

12 the nest tree is to check the colonies in October, when their activity rapidly decreases and the workers move only on their nest trees and in their closest surroundings. As most of the studied colonies were observed for several years, I could be sure where the nest was situated. Some of the colonies occupy several interconnected nests. I observed such cases mainly at the Rendezvous site. When several nests occurred close to each other, I performed aggressiveness tests to see if they belonged to distinct colonies or if they represented a polydomous colony. To do that, I collected several workers from a nest tree, transferred those onto the second nest tree and observed the behaviour of both groups. If they did not begin to fight I considered them to belong to a single colony. Sampling design and methodology are described in detail in the particular studies, as each study required a specific approach. Pitfall trapping, food bait experiments and direct observations were employed in the study A. In the study B, 150-200 ant workers per colony were sampled from 15 colonies in total. These were stored in 60% ethanol and their body sizes were measured. In the latter two studies, ants’ territories were observed. The territories consisted of nest trees, foraging trees, food resources, trails connecting all these places and areas on the soil surface where the ants searched for temporary food resources. The trails on the ground were sometimes even 50 cm wide (as I observed for instance in Lednice Castle Park). In the study C, the workers were frozen immediately after collecting and later the amount of sugars in their gasters was assessed by chemical and photometric methods. In the studies D and E, the samples were stored in pure ethanol to prevent DNA degradation. Mitochondrial DNA and microsatellites were PCR amplified and sequenced, and the relatedness between the colonies was assessed. Unfortunately, the same colonies received different names in the individual papers. The first letter always corresponded to a site, and the following number indicated the colony, but depended on the number of colonies included in the particular study. That led to several inconsistencies in the colonies’ names. Thus colony L1 in the study A corresponds to colony Z1 in study B, The colonies P1 and P2 in study A correspond to L2 and L1, respectively, in the study B, whereas colony R1 is the same in both studies, and colony R3 corresponds to colony R2 in the study B.

3.2 The study across the species’ range

The aim of the phylogeographic study was to obtain samples from as many sites across the entire distribution area of L. microcephalum as possible. My supervisor and I asked myrmecologists from most of the countries where L. microcephalum occurs to provide samples of workers from single colonies in pure ethanol or at least to provide information about the present distribution of the species (for details see Chapter 2: Distribution of Liometopum microcephalum). Some samples were also provided by my supervisor and colleagues visiting the relevant countries. Additional to sampling in South Moravia, I searched for the ant at five sites in Italy in 2011 (north-western Italy, Mantova, Pompei, Lamezia Terme in Calabria, and Madonie Mountains in Sicily), and in Slovakia in 2012 (floodplain of the Danube River in the south-west, Lučenec surroundings, Silica plateau, Latorica floodplain). These sites were selected based mainly on old records published mostly in the first half of the 20th century, or unpublished

observations, taking into account the present site conditions, in particular the occurrence of deciduous forests or at least suitable individual trees. The specimens were stored in pure ethanol, DNA was isolated from three workers per colony. In most cases, samples from a single colony per site were processed. Three mitochondrial and one nuclear DNA fragment were PCR-amplified and sequenced. For details of the genetic analyses see Chapter 8: Phylogeography of the species.

4. Study A: Interactions with other dominant ants

Availability of nest trees and the amount of food resources are the main factors affecting the presence or absence of an arboricolous ant species at a given site. As the high quality resources available at each site are limited, species with similar requirements usually compete, trying to monopolize those resources. VEPSÄLÄINEN & PISARSKI (1982) proposed a three-level hierarchy of ant communities, which clearly reflects relationships among coexisting ant species, with submissive species at the bottom and territorial species at the top of the community. Although L. microcephalum forms huge colonies, which often occupy several nest and foraging trees, nobody had assigned the species to a certain position in that hierarchy and it remained unclear how the species interacts with co-occurring behaviourally dominant ants. Is it able to withstand the pressure from its competitors or even to outcompete them? Six colonies were studied at three sites in South Moravia. At the study sites, L. microcephalum co-occurred with two behaviourally dominant ant species, that is Lasius fuliginosus and Formica rufa. All three species have similar trophic requirements. L. microcephalum and L. fuliginosus compete also for nest trees. I studied the spatial distribution of workers between nests of the competing species. Lines of pitfall traps were placed between nest trees of L. microcephalum and L. fuliginosus. One situation (in a forest near Lednice) was observed from May to July 2005, two different situations (at Rendezvous site) were studied from May to September 2007 (Appendix A1). In all cases the traps were emptied in two-week intervals. Liometopum microcephalum workers were most active in June and in the first half of July. Later the numbers of trapped workers decreased. In the preliminary study, conducted in 2005, only three workers and two mated females of L. fuliginosus were trapped in May. Later the area was fully controlled by L. microcephalum. In the subsequent study, conducted at another site, both species moved mostly in the vicinity of their own nest and the number of trapped individuals decreased with proximity to the competitor’s nest. Both species moved mainly on trails and were rarely recorded nearby the competitor’s nest. Lasius fuliginosus stayed active for a longer time period within the year. Differences between space partitioning were evident among individual colonies, apparently affected by the distribution of food resources. Space partitioning was observed in the field, suggesting that control over the foraging trees changes between years but remains mostly stable during one annual activity period. Three maps of the studied species’ territories are shown in Appendix A2. Food baits were placed between nests and between trails of the competitors. The trials were conducted in all studied situations several times, but these experiments did not yield usable results as the workers of the studied colonies did not take notice of the baits. Permanent food resources situated in foraging trees were apparently more attractive for the ant workers. I observed the behaviour of the ants in one-on-one combats in the lab, setting up 50 runs for each species pair. Liometopum microcephalum lost the combat in the majority of those trials. All three species use defensive chemicals, which proved most effective in L. fuliginosus, whose workers were able to paralyse their enemies completely in the enclosed

space of the arena (Petri dish). Workers of L. microcephalum attacked their rivals by their strong mandibles. In the field, L. microcephalum strongly profited from cooperation among workers. Fighting for foraging trees took place, in particular, in early spring. Apart from the trees, the competitors tended to avoid each other. Aggressive behaviour was observed in the vicinity of their nest trees, on trails and at food resources, indicating that L. microcephalum is a territorial species standing at the top position in the hierarchy of those ant communities of which it is a part. Six years after those observations were conducted, I observed a total displacement of competitors by L. microcephalum in two situations. In P1, the nest tree that had been occupied by L. fuliginosus was later found fully controlled by L. microcephalum. In R3, the F. rufa nest situated between nest trees of L. microcephalum and L. fuliginosus was empty, and I did not find any F. rufa workers in the vicinity of the nests of the remaining two ant species. Thus L. microcephalum colonies are not only able to withstand competive pressure from their neighbours, but they can even win the whole “war” when competing for space.

COMMUNITY ECOLOGY 12(1): 9-17, 2011 1585-8553/$20.00 © Akadémiai Kiadó, Budapest DOI: 10.1556/ComEc.12.2011.1.2

Interactions between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence

L. Petráková1, 2 and J. Schlaghamerský1

Department of Botany and Zoology, Faculty of Science, Masaryk University, Kotláøská 2, 611 37 Brno, Czech Republic

Corresponding author. Fax: +420 532 146 213; e-mail: [email protected]

Keywods: Encounters, Space partitioning, Territorial ants. Abstract. Interactions of Liometopum microcephalum with other two territorial ants also nesting in or foraging on trees, Lasius fuliginosus and Formica rufa, were studied in South Moravia (Czech Republic), at the northwestern border of its range, in 2005-2009. L. microcephalum defends its nest and foraging trees. Its distribution area is fragmented and in the north restricted to river floodplains. We investigated whether competition by other behaviourally dominant ant species could limit its distribution. We found six sites where nests of potential competitors were situated close to a L. microcephalum nest tree. We studied the partitioning of space (occupancy of foraging trees) between the species by observation and by lines of pitfall traps placed between nest trees of L. microcephalum and L. fuliginosus. These species avoided each other; the territory border changed over time. Worker interactions, including combats between L. microcephalum and F. rufa colonies, were observed in the field. Aggressive behaviour occurred close to the nest, on trails, foraging trees, and occasionally at food baits. In such situations, L. microcephalum took advantage of worker cooperation. Encounters of single workers were observed in laboratory experiments. L. microcephalum, attacking primarily by biting, lost almost all combats with L. fuliginosus, being paralysed by its secretions. One-on-one encounters with F. rufa led less frequently to combat and chances were more even, but F. rufa prevailed more often. We confirmed the territorial behaviour of L. microcephalum. In the rare situations, in which its colonies occurred together with other territorial species, we observed conflicts but no total displacement of one species by another.

Abbreviation: GEEGeneralized Estimating Equations.

Introduction the trails must be renewed after the winter pause (Quinet et al. 1997). Combats comprise sharp physical aggression (Ma- Conflicts are common in ants as a consequence of niche belis 2003) and spraying chemicals, although they can also overlap and resource competition (e.g., Czechowski 1985, be ritualized as menacing postures without attacks (Höll- 1999, Vepsäläinen and Savolainen 1990). Success depends dobler and Wilson 1990). Competitive interactions can lead on the social organisation of the colony, the workers’ mor- to replacement (Savolainen and Vepsäläinen 1989), perma- phology, the scouts’ ability to localise new resources, the ef- nent conflict (Pereira et al. 2003) or tolerance (Temeles ficiency of recruitment, and the aggressiveness level (Pisar- 1994). Coexistence is possible only if the colonies divide the ski and Vepsäläinen 1989, Holway 1999) as well as physical resources (food, space) complementarily, e.g., food speciali- conditions of a site (Mabelis 1977, Vepsäläinen and Savo- zation, shifts in the use of individual vegetation layers and lainen 1988). Usually, either the strongest competitor or the foraging periods (Bernstein 1975, Hunt 1974, Vepsäläinen one that arrives first takes control over the resource (Höll- and Savolainen 1988). Polydomous systems of colonies fa- dobler and Wilson 1990), whereas the weaker competitors cilitate the exploitation of a foraging area and thus support are pushed back to the less optimal parts of the habitat and colony development. Polydomous ants tend to be also have to use food of lower quality (Sanders and Gordon ecologically dominant (Debout et al. 2007). 2000). This leads to an asymmetric use of space (Pereira et al. 2003). Many myrmecologists suggested a hierarchal order In the present study, we look at the interaction of the within the ant community with varying number of levels dolichoderine ant Liometopum microcephalum (Panzer, (e.g., Vepsäläinen and Pisarski 1982, Schoener 1983). Gen- 1798) with two behaviourally dominant, territorial ant spe- erally, “submissive species” are adaptable and often shift cies with similar requirements. L. microcephalum is a ther- their resource preferences in the presence of a stronger ant mophilous, submediteranean and pontocaspic species. In the species (Savolainen et al. 1989) whereas “territorial species” north of its range the species is restricted to the floodplains at the top are able to defend an area around their nest, trails of large rivers and its distribution area is substantially frag- and monopolized food resources (Levings and Traniello mented; thus the species is considered threatened in many 1981, Hölldobler and Wilson 1990). These species form countries (Makarevich 2003, Bezdìèka 2005, Seifert 2007). huge colonies and often determine the structure of ant com- It reaches its northwestern border of distribution in the Dyje munities (Savolainen et al. 1989, Czechowski and Markó and Morava floodplains in South Moravia (Czech Republic) 2005). Territory borders change dynamically in the course of where its population is large and vital despite its marginal time. Ants fight mostly in the spring when food is scarce and position (Schlaghamerský and Omelková 2007). The ants 10 Petráková and Schlaghamerský build nests several metres above ground in old but living often extending below ground. According to Seifert (2007) it trees, especially oaks. Most foraging takes place on other has a particularly close association with the aphid Stomaphis trees in the vicinity of the nest tree. Several authors (e.g., quercus, which is also tended by L. microcephalum (Wiest Forel 1892, Makarevich 2003) have reported at least some 1967). Workers of similar size (4-6 mm) as those of L. mi- colonies of this species to be polydomous. L. microcephalum crocephalum use effective chemicals (dendrolasin, undecan) is a very efficient hunter (e.g., Zettel et al. 2004, Graben- instead of physical aggression in combat. The species is poly- weger et al. 2005), although aphid tending seems to be also domous. Formica rufa L. 1761 colonies are mostly monodo- an important part of its foraging behaviour (Wiest 1967, Seif- mous (Seifert 2007), building typical ant heaps, most often ert 2007). Workers (3-7 mm) have been reported to be very in coniferous forests (Czechowski et al. 2002). Its workers aggressive towards other ant species (e.g., Forel 1892, Wiest are of bigger size (8–10 mm) than those of the two above- 1967, Grabenweger et al. 2005, Seifert 2007). They attack by mentioned species and very aggressive, using strong mandi- biting and by spraying a secretion which repels enemies, bles as well as chemical weapons (formic acid, decan, unde- marks prey and initiates alarm behaviour (Wilson and Pavan can). 1959, Wiest 1966, Wiest 1967). Andersen (1997) classified We took into account only those sites where the nests of the North American Liometopum species as behaviourally potentially competing species were less than 25 m apart, as- dominant. Most myrmecologists who have observed L. mi- suming that the probability of conflict increased with de- crocephalum would probably consider it a territorial, behav- creasing colony distance. The 25 m threshold was a compro- iourally dominant species (e.g., Wiest 1967, Tartally 2006). mise between the need to find a sufficient number of sites to We posed the question of whether L. microcephalum was study and our assumption that interactions and their effect on indeed a territorial species and if its colonies were able to distribution would be difficult to detect over a longer dis- withstand the competition pressure of sympatric, behaviour- tance. We checked the trees at these sites in autumn, when ally dominant ants. As only a small proportion of L. micro- the activity of workers was markedly lower and the ants cephalum colonies in South Moravia exist outside of the his- moved only close to their nest trees, to tell apart nest trees of toric floodplain (no longer exposed to regular flooding in small colonies and favoured foraging trees (with many work- most parts, due to massive river regulation in the 1970s- ers running up and down the trunk from spring to summer). 1980s), and as we had anecdotal evidence about replacement Six study sites were identified and used within the pre- of its colonies by Formica spp. in dry localities (Josef Chytil, sent study. These sites were situated within three study areas pers. comm.), we assumed that the observed distribution pat- o that covered an oak-dominated floodplain forest (L1; 48 48´ tern might be (amongst other things) the result of competi- o N, 16 49´ E), a moist oak old-growth outside of the flood- tion. The devastating effect of floods on ants nesting at o o plain (R1, R2, R3; 48 44´ N, 16 47´ E), and a large land- ground level had been observed in the very same floodplain scape park with ancient oak trees in moist meadows (P1, P2; (Schlaghamerský 2000). Nests in tree trunks well above the o o 48 48´ N, 16 48´ E). No suitable site was found in the area’s ground level should give L. microcephalum a competitive ad- xerothermic forests. We observed several workers of For- vantage over ants with ground nests in active floodplains, mica polyctena Förster 1850 (a species closely related to F. whereas such species might outcompete L. microcephalum rufa) not far from a L. microcephalum colony at such a site, when and where floods do not occur. This might have con- but we were not able to find a Formica nest within a radius tributed to the fragmented distribution of the species and of 100 m. At all study sites Lasius fuliginosus was the poten- could also lead to a further decline of its population in South tial competitor, but only at one site (R3) was Formica rufa Moravia and elsewhere, as dominant ant species with a less present as well. Table 1 shows the distances between nests of arboricolous way of life might get established in the historic potentially competing colonies. We were not able to find a floodplains. To get a better understanding of the interactions site with neighbouring L. microcephalum and F. rufa colo- of L. microcephalum and its potential competitors, we studied nies without L. fuliginosus in their vicinity. The distribution the space partitioning between their colonies and its changes and abundance of ant species was studied by means of lines over time as well as behaviour during direct encounters. of pitfall traps. Interactions between potentially competing species were assessed using food baits, field observations Materials and methods and observations of one-on-one combats of workers in the laboratory. The study was conducted in South Moravia (south-eastern Czech Republic) in 2005-2009. We searched for sites Spatial distribution of species where L. microcephalum nests were located close to colonies of other dominant species. In South Moravia, L. micro- Only three sites were suitable for pitfall trapping: L1 was cephalum meets two territorial ant species, both having simi- studied from May to July 2005 and R1 and R2 from May to lar trophic requirements as L. microcephalum (predation and September 2007 (at the latter two the same colony of L. aphid tending) and, in contrast to L. microcephalum, both fuliginosus but two different colonies of L. microcephalum abundant throughout most of the Palearctic (Seifert 2007). were involved). The other sites were located near frequented Lasius fuliginosus (Latreille 1798) also prefers trunks of old pathways or in a castle park, where the risk of the traps being trees, including oaks, but its nests are close to the ground, damaged or lost was high. In all cases the traps were arranged Interactions between ants 11

Table 1. Distances (m) between nests of potentially competing the relationship was different for both species. Actual dis- colonies of L. microcephalum and other behaviourally dominant tances of individual traps were used for the GEE modelling. species at all study sites; underlined distances correspond with Further, we tested for each line of pitfall traps if workers of transects studied by pitfall trapping, first number in brackets the two species occured randomly in the traps or if individual equals the number of trap lines perpendicular to the transect, second number in brackets equals the number of traps in each traps had captured preferentially either one or the other spe- line; two numbers for distances separated by a semicolon indi- cies, which would point at mutual avoidance (Wilcoxon cate that there were two nests of the given species. Paired Test). Additionally, we assessed the differences in numbers of other ant species among the lines of traps (Kruskal-Wallis Test).

Field observations

We tried to evoke conflicts between colonies by placing food baits (tuna with honey or canned meat for cats) between nests or trails of potentially competing species. These baiting in lines perpendicular to a transect between nest trees of po- experiments (n = 25) were conducted at all three study areas tentially competing species (L. microcephalum and L. fuligi- during spring and summer 2007 and 2008 and in spring 2009. nosus) to form a grid, which would cover the space with the The baits were placed in the field between 10.30 and 14.00 highest probability of encounters between the members of and exposed for two hours. We placed always five or ten both colonies. The number of lines was adapted to the dis- baits in equal intervals along a transect between nests (distance between nest trees; standard intervals between lines tances between baits were chosen depending on the transect and between traps were preserved within each site (Fig.1). length). The individual baits were randomly shifted up to one For easier orientation the lines were assigned letters in alpha- metre to the right or left from the transects. Control baits were betical order (A, B, C, etc.), line A being closest to the L. placed in a distance of one metre from the nest or directly on microcephalum nest tree. Small plastic cups (film canisters; the nest tree (ant heap in case of F. rufa). When we found diameter 3.2 cm, depth 4.5 cm) with ethylene glycol as kill- trails of both species leading close to each other, we placed ing and preserving agent served as pitfall traps. The catch one bait between them. was retrieved every two weeks and the cups were exchanged for new ones with new preservative on this occasion. In addition to the baiting experiments, we mapped the Trapped ants were identified to species (according to Seifert position of trunk trails connecting nests with stable food re- 2007 and Czechowski et al. 2002) and counted. For the pur- sources and monitored changes in space and resource parti- pose of data analysis we pooled data of three subsequent trap- tioning in the course of an annual activity period and between ping periods, obtaining three six-week periods for which sta- those of subsequent years. Each site was visited at least two tistical tests were performed separately. The distribution of times per year (spring and summer; 2007-2009). We ob- the datasets did not fulfill the conditions of normality. To served the behaviour of the workers in the area between nests model the relationship between the abundance of both ant for at least 30 minutes during each visit at a given site, i.e., species and the distance from their competitor’s nest tree we how far from their nest towards the rival’s nest the workers used Generalized Estimating Equations (GEE). This regres- moved and if the species encountered (L1 - 14 visits, R - 24 sion method allows implementation of an association struc- visits,P-11visits).Wealsoexperimentally transferred sin- ture that models spatial correlation between captures of both gle workers of the rival species onto the trunks of nest trees species (Hardin and Hilbe 2003). We used unstructured as- of L. microcephalum and L. fuliginosus (surface of ant hill in sociation structure and GEE with Poisson error structure as case of F. rufa) to observe the reaction of the domestic ant the abundances were counts. For each site and date an AN- colony (R1, R2, R3; summer 2007 and 2008, ten replicates COVA type of model was fitted in order to reveal whether for each species in total).

Figure 1. The arrangement of pitfall traps (1-9= traps, A-E = lines of traps; intervals between traps in each line = 1 m, intervals between lines = 2-2.5 m). 12 Petráková and Schlaghamerský

Aggressiveness tests Table 2. Percentage representation of ant species captured during the whole trapping season along the three transects studied One-on-one encounters were observed in closed Petri by pitfall trapping (total number of individuals captured at L1: dishes. The ants were collected at the sites L1 and R1-3 and n = 1695, R1: n = 6522, R2: n = 2118). the experiments were carried out on the same day or the day after the ant collection. For each of the two species pairs (L. microcephalum versus L. fuliginosus, L. microcephalum versus F. rufa) 50 runs were observed. The observation ended when one of the rivals was killed or after one hour in the case of no aggression. We never used the same worker for more than one experiment. After each run the dish was wiped out with 70% ethanol to remove smells. We took notes on the behaviour of both ants, in particular, if they started to fight, which species initiated the fight, what tactics they used, and what was the outcome. The encounters were observed by the first author and partially also recorded with a video camera. We used McNemar’s Test to evaluate equal probability of wins (i.e., if forces of both species were balanced). both species were active all over the area for most of the season, but finally L. fuliginosus occupied a larger portion (Fig. Results 3d). The occurrence of both species differed across traps (Wilcoxon Paired Test: P < 0.05) for each line except the Spatial distribution of species middle line C at R1 (all periods) and line A at R2 (first period, In total, 13 ant species were found at the study sites (Ta- when a trail of L. fuliginosus led close to the L. micro- ble 2). Regarding non-dominant species, numbers of indi- cephalum nest tree ). Numbers of ants in traps of a given line viduals did not differ significantly among the trap lines. At varied a lot and highest numbers were found in traps that L1, the catch of the first trapping period in May 2005 showed were placed close to trees or shrubs (visited by foraging that L. fuliginosus workers were penetrating into the area ants). L. microcephalum workers were most active in June close to the L. microcephalum nest tree, reaching line B (Fig. and July. After that their activity decreased gradually. The abundance of ants of both species increased exponentially 2a). After that, the whole area was fully controlled by L. mi- 2 crocephalum (Fig. 2b) till the end of the trapping season. For with the distance from the tree of the competitor (GEE, X 1 that reason we did not test the significance of distribution > 7.81, P < 0.01), but for most study periods the increase was data for this two species obtained at L1. Subsequent research significantly different for the two species (Fig. 4). At R1 the increase of L. microcephalum was significantly steeper than at other sites (R1, R2) in 2007 showed quite different results. 2 At R1 both species were abundant in traps all over the area that of L. fuliginosus in the first two periods (GEE, X 1 >7.3, mainly in spring (Fig. 3a). Later each species dominated the P < 0.001). In the last period the increase of L. fuliginosus was significantly steeper than that of L. microcephalum lines close to its nest tree (Fig. 3b-c). At the second site (R2) 2 (GEE, X 1 = 16.4, P < 0.0001). In this time period, L. micro-

Figure 2. Spatial distribution of workers between nest trees in site L1 in 2005. (black columns = Liometopum microcephalum, grey columns = Lasius fuliginosus, white square = empty trap, missing square = lost trap; line A = closest to L. microcephalum nest, line D = closest to L. fuliginosus nest). Interactions between ants 13

Figure 3. Spatial distribution of workers between nest trees at the sites R1 (a-c) and R2 (d) in 2007 (black columns = Liometopum microcephalum, grey columns = Lasius fuliginosus, white square = empty trap, missing square = lost trap; line A = closest to L. microcephalum nest, line E or C = closest to L. fuliginosus nest).

Figure 4. Relationships between abundance of L. microcephalum and L. fuliginosus and the distance from the competitors nest tree in the three time periods at two study sites (in the first period, curves of both species at R2 overlap). Fitted models were obtained from GEE. 14 Petráková and Schlaghamerský cephalum was already reducing its activity outside of nest microcephalum worker with its mandibles and ran back with trees (see also numbers of individuals). At R2 the increase it. Only single L. microcephalum workers on the ground were 2 was similar for both species in the first period (GEE, X 1 = attacked. This went on for almost two hours, after which ob- 0.1, P = 0.81); in the second period the increase of L. fuligi- servation had to be interrupted. Checking again one week nosus was significantly steeper than that of L. micro- later, the situation was similar, but there were far fewer F. 2 cephalum (GEE, X 1 = 5.3, P = 0.02), and in the last period rufa workers present. During the same time period the same the increase of L. microcephalum was significantly steeper F. rufa colony also lost one foraging tree (an old oak) to L. 2 than that of L. fuliginosus (GEE, X 1 = 4.7, P = 0.03). Taking fuliginosus. F. rufa workers were not able to climb up either into consideration the numbers of captured ants, a sharp de- of these two oaks for the rest of the season. Space partitioning crease of L. microcephalum towards the competitor’s nest (of foraging trees with resources such as prey and honeydew) tree was observed during periods of high activity, in particu- thus occurred in early spring and stayed rather stable during lar at R1 (Fig. 4). the rest of the ants’ annual activity period. We were not successful in using food baits to evoke a Field observations conflict between colonies, although we tried several modifi- L. microcephalum and L. fuliginosus moved mainly on cations of setting up the baits. The baits were most often oc- trails of dendritic shape, connecting nest trees with foraging cupied by Myrmica rubra or M. ruginodis. L. microcephalum trees or shrubs, whereas F. rufa workers were often present and L. fuliginosus noticed only those baits placed closest to off their trails. Workers of competing species going out of their nest or trail, whereas F. rufa fed even on baits farther their nests to forage tended to avoid the direction of the com- than the middle of the transect. In three of nine cases, where petitor’s nest. Single foraging trees were monopolized either trails of L. microcephalum and L. fuliginosus came close (1-2 by one or the other species. In several cases we observed that m), both species detected the bait. The species that had ar- the “possession” of foraging trees changed over time, in par- rived first always gained control over the bait (in two cases ticular from one year to another. At L1 we noted no conflict this was Liometopum microcephalum, in one case Lasius and both ant species kept the same foraging trees during the fuliginosus); the late-comer retreated after a very short fight whole period of our study. At the sites R1-3 almost all the or exhibition of menacing behaviour. F. rufa and Liome- single trees, as well as saplings and shrubs in the dense un- topum workers never exploited the same bait. derstorey, were occupied by one or the other of the studied Observing ant behaviour at all study sites (independent ant species. The “possession” of foraging trees often changed of the presence of baits), we observed only a few cases of in a subsequent year but remained stable throughout the ant’s direct aggression. Mostly when a Liometopum worker acci- activity season. In the Lednice Castle Park (P1), where dentally met its potential enemy, both of them jumped aside mainly ancient oaks were scattered in grassland, we observed and ran away. We observed single workers of competing spe- the take-over of a foraging tree of L. fuliginosus by L. micro- cies moving in 2-5 cm distance from each other without cephalum. At another site in the same area (P2) we observed showing aggression. However, aggressive defence behav- the opposite outcome (L. fuliginosus displaced L. micro- iour was apparent on Liometopum trails and close to their cephalum). The only open aggression between colonies that nest trees. Here, and particularly on the nest tree trunks, the we were able to observe occurred in April 2009 in R3 and Liometopum workers attacked the intruder (experimentally involved L. microcephalum and F. rufa. The base of an oak transferred to the trunk) immediately, showing a high level trunk, which had been frequented by F. rufa in 2007 and of cooperation. Several workers held the intruder’s legs 2008, was occupied by hundreds of L. microcephalum work- stretched apart, while others bit into its body. Typical alarm ers. Hundreds of F. rufa workers had assembled on the behaviour was observed – other workers ran around the en- ground around the trunk base, with their mandibles open. emy with wide opened mandibles and lifted gaster. Single L. FromtimetotimeaF. rufa worker ran forward, grabbed a L. fuliginosus workers were more easily defeated than F. rufa

Figure 5. Aggressiveness tests outcomes of one-on-one combats observed in the lab; a) L. microcephalum vs. L. fuliginosus: McNemars X = 15.04, df = 1, p-value = 0.0001; b) L. microcephalum vs. F. rufa: McNemars X = 4.65, df = 1, p-value = 0.03. Interactions between ants 15

Table 3. Aggressivenes tests - number of cases in which ant pices) and too low humidity levels (Schlaghamerský and workers applied particular tactics against the competitor during Omelková 2007). Better knowledge of the drier habitats oc- one-on-one combats observed in the lab (species abbreviations: cupied by L. microcephalum in some parts of its range would LIM = Liometopum microcephalum, LAF = Lasius fuliginosus, FOR = Formica rufa). help to assess the importance of these factors. The location of nest trees restricted the use of pitfall trapping. Pitfall trap data reflect the changes in worker activity during the season (Czechowski et al. 1995, Vepsäläinen et al. 2000) and allow to detect in which places workers of a given species occur on the ground. The absence of L. fuliginosus workers in traps at L1 originally gave us the impression that L. microcephalum was by far the stronger competitor workers, which were even able to walk with immobile L. microcephalum workers clinging to their legs. L. micro- (Petráková and Schlaghamerský 2007). Subsequent observa- cephalum workers were also attacked immediately after the tions imply that it was caused by the distribution of food re- transfer to the nest of the other competitor species. L. fuligi- sources at L1 – the only permanent trail of L. fuliginosus led nosus workers always paralysed L. microcephalum worker, from the nest to a lime tree situated outside the area covered whereas F. rufa aggressively bit into the rival’s body. by the traps. Data from R1 and R2 confirmed that L. microcephalum and L. fuliginosus avoided each other. The two Aggressiveness tests species met most often in the middle between their nest trees, where worker numbers of both species were relatively low in Workers usually did not fight immediately when they comparison to the trap lines close to their nest trees. The were put into the arena but mostly after they ran into each modelling of the relationship between worker abundance and other or when one of them fell on the other accidentally. In distance to the competitor’s nest tree also showed avoidance, one-on-one combats, L. fuliginosus as well as F. rufa seemed in particular L. microcephalum did not approach nest of L. to be stronger than L. microcephalum, although a great deal fuliginosus at R1 for most of season. Results from R2 were of experiments resulted in both species affected (Fig. 5). The ambiguous; worker activity between the two nest trees was most frequent tactics of L. microcephalum was biting into lower than in R1 and both species were also active close to legs or antennae (Tab. 3). Usually the Liometopum worker the competitor’s colonies, which were rather close. The ter- held on to its rival’s limb until paralysed by the rival’s chemi- ritory borders changes over time as reported for many species cal weapons. A strong dose of these substances induced body by other authors (e.g., Hölldobler and Wilson 1990, Akino trembling, repeated opening and closing of mandibles, and and Yamaoka 1999). The territory of L. microcephalum con- loss of coordination down to rigidity. L. microcephalum was sists of nest trees, trails and foraging trees (food resources) paralysed more often and quicker by L. fuliginosus than by that are defended together with their closest surroundings. F. rufa. Liometopum workers were able to paralyse F. rufa This strategy is more efficient than the defence of a continu- workers but were not successful in biting off their legs de- ous area covering also land between the trails (Wilson 1975). spite attempts to do so. On the other hand, Liometopum work- This allows submissive ant species to live in the non-de- ers did not paralyse L. fuliginosus workers but bit off their fended patches. Our expectation to find most individuals of legs easily. other (submissive) species in the traps of the transects’ middle lines (“no-man’s land”) was not met. Submissive species Discussion are very adaptable and tend to quickly localise small food items or forage at different times than dominant ants (Fellers Intensive searching in the field yielded only a few suit- 1987, Czechowski and Markó 2005), which enables them to able sites with nests of other behaviourally dominant species coexist with dominant species. close to L. microcephalum colonies, and in xerothermic oak woods we did not succeed at all. Therefore, we could not make We assume that the mutual avoidance of L. micro- direct observations that would confirm our assumption that cephalum and L. fuliginosus is probably caused by species- competitors largely limit the occurrence of L. microcephalum specific odours. Wiest (1967) observed that L. micro- to floodplains (and other water-logged sites). On the other cephalum avoided L. fuliginosus workers, including dead hand, the lack of strong colonies of potential competitors in ones. Mabelis (2003) pointed out that Formica species avoid the vicinity of the few and small L. microcephalum colonies L. fuliginosus as well. The chemical weapons used by L. present in South Moravia’s xerothermic forests might ex- fuliginosus have been reported to be very powerful plain why L. microcephalum was able to become established (Czechowski 1999), which was also confirmed in our experi- at these sites. In the latest of several reported cases of replace- mental one-on-one combats with L. microcephalum. How- ment by Formica spp., an isolated L. microcephalum colony ever, the chemical substances used should have a stronger ef- at a dry site had been replaced by a Formica rufa colony fect in a closed Petri dish than in a real-life situation. The within a few years in the 1990s (Josef Chytil, pers. comm.). most effective weapon of single L. microcephalum workers Other factors potentially responsible for the observed distribu- was their dentate mandibles (see also Grabenweger et al. tion of L. microcephalum in South Moravia are the lack of suit- 2005). However, a crucial component of the species’ strength able nest trees in the area’s xerothermic oak forests (former cop- in combat is cooperation. Ant species of bigger body size 16 Petráková and Schlaghamerský more often use one-on-one tactics, and profit from their body with L. fuliginosus and F. rufa, the two major sympatric ter- size, whereas small-bodied ants prefer collective combat ritorial ants in South Moravia. However, the coordinated ac- (McGlynn 2000). Wiest (1967) reported that the small myr- tion of workers seems to be its strongest weapon. At sites micine Tetramorium caespitum workers were able to kill where its colonies occurred together with the other territorial several L. microcephalum workers in an experiment and its species, we observed conflicts, but no total displacement of colonies were equally abundant inside and outside of L. mi- one species by another over a study period of several years. crocephalum territories. She also published that only night- Such sites were, however, rare and limited to floodplains and active arboricolous ants lived in its territory and that species similar moist habitats. Here L. microcephalum seems able to nesting on the ground formed smaller nests within L. micro- withstand competition by the top dominant ant species. cephalum territories than outside. We have no experimental Acknowledgements. The study was supported by the Min- data on encounters of worker groups, but we observed coor- istry of Education, Youth and Sports of the Czech Republic, dinated action of fighting workers in the field. Our attempts Research Plan No. MSM0021622416. The Czech Ministry to initiate and observe conflicts of colonies at food baits, of the Environment granted access to the Rendezvous Na- however, were not successful, although this method has been tional Nature Monument and allowed us to collect L. micro- widely used in similar experimental studies (e.g., cephalum specimens by its Exception No. 8375/04- Czechowski 1985, Human and Gordon 1996, Gibb and Ho- 620/1377/04. The Master’s thesis of M. Omelková provided chuli 2004). Workers probably paid little attention to our valuable information on the biology and distribution of L. food baits because they were attracted by bigger and more at- microcephalum as a starting point for the present study. We tractive resources (Sanders and Gordon 2000, Czechowski and are particularly grateful to our colleague S. Pekár, who pro- Markó 2005) - most foraging seems to take place in the trees and vided statistical advise and conducted the GEE modelling. M. Barclay (Natural History Museum, London) kindly also includes aphid tending (Wiest 1967, Makarevich 2003, own checked the English. unpublished results). Other explanations might be “route fidel- ity”, which means that the workers hardly leave their trails (Sa- volainen and Vepsäläinen 1989), or that L. microcephalum References workers did not notice immobile prey. These ants react strongly Akino, T. and R. Yamaoka. 1999. Trunk trail network of Lasius to movement because of their excellent visual orientation (Wiest fuliginosus Latreille (Hymenoptera: Formicidae): Distribution 1966, Dietrich and Busch 2004). between conspecific neighboring colonies. Entomol. Sci. 2: Space and resource partitioning between ant colonies 341-346. takes place in early spring when the trails leading to food re- Andersen, A. N. 1997. Functional groups and patterns of organization in North American ant communities: a comparison with sources have to be renewed after the winter pause (e.g., Ak- Australia. J. Biogeogr. 24: 433-460. ino and Yamaoka 1999). This was also the case for our spe- Bernstein, R. A. 1975. Foraging strategies in ants in response to cies. In the years of our study, spring started repeatedly very variable food density. Ecology 56: 213-219. early with a fast and unexpected rise in temperature. There- Bezdìèka, P. 2005. Formicoidea (mravenci). In: Farkaè, J., D. Král fore, we were not always able to observe the initial phase of and M. korpík (eds.), Red List of Threatened Species in the worker activity and we observed fights between colonies Czech Republic: Invertebrates. Agency for Nature Conserva- only in spring 2009 at R3. We did not observe any mass kill- tion and Landscape Protection of the Czech Republic (AOPK), ing on that occasion; from the workers’ behaviour we con- Praha. pp. 384-386. clude that their aggression was ritualized (Temeles 1994, Case, T. J. and M. E. Gilpin. 1794. Interference competition and Mercier et al. 1997). Once a species had taken possession of niche theory. Proc. Nat. Acad. Sci. 71: 3073-3077. a foraging tree and successfully defended it for some time, it Czechowski, W. 1985. Competition between Myrmica laevinodis kept this tree for the rest of the season. The studied species Nyl. and Lasius niger (L.) (Hymenoptera, Formicoidea). Ann. often changed their foraging trees (particularly Zool. 39: 153173. smaller/younger ones) in the course of subsequent years at Czechowski, W. 1999. Lasius fuliginosus (Latr.) on a sandy dune sites where these resources were abundant (where trees were its living conditions and interference during raids of Formica sanguinea Latr. (Hymenoptera, Formicidae). Ann. Zool. 49: interspersed in grassland, the large nest trees probably also 117-123. provided most food sources). Ants are highly adaptive and Czechowski, W. and B. Markó. 2005. Competition between For- tend to prefer food that is most available at the given moment mica cinerea Mayr (Hymenoptera: Formicidae) and co-occur- (Czechowski et al. 1995). Large colonies, however, prefer ring ant species, with special reference to Formica rufa L.: stable food resources for colony maintenance (Hölldobler direct and indirect interferences. Polish J. Ecol. 53: 467-487. and Wilson 1990). In the case of the studied species these Czechowski, W., B. Pisarski and K. Yamauchi. 1995. Succession were the foraging trees with aphid colonies and other sources of ant communities (Hymenoptera, Formicidae) in moist pine of food. The sufficient availability of food sources is surely forest. Fragmenta Faunistica 38 (24): 447-488. a crucial factor for their coexistence (Case and Gilpin 1974). Czechowski, W., A. Radchenko and W. Czechowska. 2002. The ants (Hymenoptera: Formicidae) of Poland. Museum and Institute of To conclude, L. microcephalum indeed shows territorial Zoology PAS, Warszawa. behaviour but as a truly arboricolous species it is much more Debout, G., B. Schatz, M. Elias and D. McKay. 2007. Polydomy in active in defending nest and foraging trees than temporary ants: what we know, what we think we know and what re- food sources on the ground. Its single workers lose combats mains to be done. Biol. J. Linnean Soc. 90: 319-348. Interactions between ants 17

Dietrich, C. O. and T. Busch. 2004. Arboricaria sociabilis (Kul- fuliginosus (Hymenoptera: Formicidae). Behav. Processes 40: czynski, 1897) (Araneae: Gnaphosidae) neu für Österreich: Ein 75-83. spezialisierter myrmekoider Räuber von Liometopum micro- Sanders, N. and D. M. Gordon. 2000. The effects on interspecific cephalum (Panzer, 1798) (Hymenoptera: Formicidae). Wissen- interactions on resource use and behavior in a desert ant. Oe- schaftliche Mitteilungen aus dem Niederösterreichischen cologia 125: 436-443. Landesmuseum 16: 33-46. Savolainen, R. and K. Vepsälainen. 1989. Niche differentiation of Fellers, J. H. 1987. Interference and exploitation in a guild of ant species within territories of the wood ant Formica woodland ants. Ecology 68: 1466-1478. polyctena. Oikos 56: 3-16. Forel, A. 1892. Die Ameisenfauna Bulgariens. Verhandlungen der Savolainen, R., K., Vepsälainen and H. Wuorenrinne. 1989. Ant k. k. zoologisch-botanischen Gesellschaft in Wien 42: 305-318. assemblages in the taiga biome: testing the role of territorial Gibb, H. and D. F. Hochuli. 2004. Removal experiment reveals wood ants. Oecologia 81: 481-486. limited effects of a behaviorally dominant species on ant as- Schlaghamerský, J. 2000. The saproxylic beetles (Coleoptera) and semblages. Ecology 85: 648-657. ants (Formicidae) of Central European hardwood floodplain Grabenweger, G., P. Kehrli, B. Schlick-Steiner, F. Steiner, M. forests. Folia Facultatis scientiarium naturalium Universitatis Stolz, and S. Bacher. 2005. Predator complex of the horse Masarykianae Brunensis, Biologia 103: 1-205. chestnut leafminer Cameraria ohridella: identification and im- Schlaghamerský, J. and M. Omelková. 2007. The present distribu- pact assessment. J. App. Entomol. 129: 353-362. tion and nest tree characteristics of Liometopum micro- Hardin, J. W. and J. M. Hilbe. 2003. Generalized Estimating Equa- cephalum (Panzer, 1798) (Hymenoptera: Formicidae) in South tions. Chapman and Hall/CRC, Boca Raton. Moravia. Myrmecological News 10: 85-90. Hölldobler, B. and E. O. Wilson. 1990. The Ants. Harvard Univ. Schoener, T. W. 1983. Field experiments on interspecific competi- Press, Cambridge, MA. tion. American Naturalist 122: 240-285. Holway, D. A. 1999. Competitive mechanisms underlying the dis- Seifert, B. 2007. Die Ameisen Mittel- und Nordeuropas.Lu- placement of native ants by the invasive Argentine ant. Ecol- traVerlags- und Vertriebsgesellschaft, Görlitz/Tauer. ogy 80: 238-251. Tartally, A. 2006. Long term expansion of a supercolony of the in- Human, K. G. and D. M. Gordon. 1996. Exploitation and interfer- vasive garden ant, Lasius neglectus (Hymenoptera: Formici- ence competition between the invasive Argentine ant, Linepi- dae). Myrmekologische Nachrichten 9: 21-25. thema humile, and native ant species. Oecologia 105: 405-412. Temeles, E. J. 1994. The role of neighbors in territorial systems: when are they dear enemies? Animal Behavior 47: 339-350. Hunt, J. H. 1974. Temporal activity patterns in two competing ant species (Hymenoptera: Formicidae). Psyche 81: 237-242. Vepsälainen, K. and B. Pisarski. 1982. Assembly of island ant communities. Ann. Zool. Fenn. 19: 327-335. Levings, S. C. and J. F. A. Traniello. 1981. Territoriality, nest dis- persion and community structure in ants. Psyche 88: 265-319. Vepsäläinen, K. and R. Savolainen. 1988. A competition hierarchy Mabelis, A. A. 1977. Artenreichtum von Ameisen in einigen among boreal ants: impact on resource partitioning and com- Waldtypen. Berichte der Internationalen Symposien der Inter- munity structure. Oikos 51: 135-155. nationalen Vereinigung für Vegetationskunde (Tüxen, R. ed.), Vepsäläinen, K. and R. Savolainen. 1990. The effect of interfer- Vaduz, pp. 187-208. ence by Formicinae ants on the foraging of Myrmica. J. Anim. Mabelis, A. A. 2003. Do Formica species (Hymenoptera: Formici- Ecol. 59: 643-654. dae) have a different attack mode? Ann. Zool. 53: 667-668. Vepsäläinen, K., R. Savolainen, J. Tiainen and J. Vilén. 2000. Suc- Makarevich, O. N. 2003. Liometopum microcephalum (Hymenop- cessional changes of ant assemblages: from virgin and ditched tera: Formicidae) in the Lower Dnepr. Vestnik zoologii 37: 51- bogs to forests. Ann. Zool. Fenn. 37: 135-149. 56. Wiest, L. 1966. Über die Funktion des Gasters bei der Ver- McGlynn, T. P. 2000. Do Lanchester´s laws of combat describe ständigung von Ameisen. Doctoral thesis (unpublished), Uni- competition in ants? Behavioral Ecology 11: 686-690. versität Wien. Mercier, J. L., A. Lenoir and A. Dejean. 1997. Ritualised versus Wiest, L. 1967. Zur Biologie der Ameise Liometopum micro- aggressive behaviours displayed by Polyrhachis laboriosa (F. cephalum Panz. Wissenschaftliche Arbeiten aus dem Burgen- Smith) during intraspecific competition. Behav. Processes 41: land 38: 136-144. 39-50. Wilson, E. O. 1975. Enemy specification in the alarm-recruitment Pereira, H. M., A. Bergman and J. Roughgarden. 2003. Socially system of an ant. Science 190: 798-800. stable territories: The negotiation of space by interacting fora- Wilson, E. O. and M. Pavan. 1959. Glandular sources and specific- gers. Amer. Nat. 161: 143-152. ity of some chemical releasers of social behavior in Petráková, L. and J. Schlaghamerský. 2007. Preliminary results on dolichoderine ants. Psyche 66: 70-76. the interaction of Liometopum microcephalum (Panzer, 1798) Zettel, H., T. Ljubomirov, M. F. Steiner, C. B. Schlick-Steiner, G. with other ants (Hymenoptera: Formicidae). Myrmecological Grabenweger, and H. Wiesbauer. 2004. The European ant News 10: 118. hunters Tracheliodes curvitarsis and T. varus (Hymenoptera: Pisarski, B. and K. Vepsälainen. 1989. Competition hierarchies in Crabronidae): Taxonomy, species discrimination, distribution ant communities (Hymenoptera, Formicidae). Ann. Zool. 42: and biology. Myrmecologische Nachrichten 6: 39-47. 321-329. Received March 31, 2010 Quinet, Y., J. C. De-Biseau and J. M. Pasteels. 1997. Food recruit- Revised September 8, November 19, 2010 ment as a component of the trunk-trail behaviour of Lasius Accepted December 6, 2010

Appendix A1:

Spatial distribution of workers captured in pitfall traps between nest trees of Liometopum microcephalum and Lasius fuliginosus at Rendezvous site in 2007.

15.5.-29.5. 29.5.-12.6. 12.6.-27.6.

27.6.-10.7. 10.7.-27.7. 27.7.-9.8.

9.8.-25.8. 25.8.-10.9. 10.9.-26.9.

Figure A1-1: Spatial distribution of workers captured in pitfall traps between nest trees of Liometopum microcephalum (red columns) and Lasius fuliginosus (blue columns) in the situation R1 at Rendezvous site in 2007. (A = line of traps placed closest to L. microcephalum nest, E = line of traps placed closest to L. fuliginosus nest; white square = no behaviourally dominant ant species captured; missing square = lost trap).

15.5.-29.5. 29.5.-12.6. 12.6.-27.6.

27.6.-10.7. 10.7.-27.7. 27.7.-9.8.

9.8.-25.8. 25.8.-10.9. 10.9.-26.9.

Figure A1-2: Spatial distribution of workers captured in pitfall traps between nest trees of Liometopum microcephalum (red columns) and Lasius fuliginosus (blue columns) in the situation R2 at Rendezvous site in 2007. (F = line of traps placed closest to L. fuliginosus nest, H = line of traps placed closest to L. microcephalum nest; white square = no behaviourally dominant ant species captured; missing square = lost trap).

Appendix A2:

Space partitioning between Liometopum microcephalum and its competitors (maps of territories) in three different situations showing exploitation of foraging trees and trails leading to food resources. The legend included in Figure A2-1 applies to all three figures.

Figure A2-1: Space partitioning between colonies of Liometopum microcephalum, Lasius fuliginosus and Formica rufa in situation R3; Rendezvous site, spring 2009.

Figure A2-2: Space partitioning in situations R1 (left) and R2 (right); Rendezvous site, summer 2008. Trails and trees occupied by Liometopum microcephalum are shown in black, grey colour shows trails and trees used by Lasius fuliginosus; for details see the legend in Figure A2-1.

Figure A2-3: Space partitioning in situation P1; Lednice Castle Park, 2007 and 2009 compared. Trails and trees occupied by Liometopum microcephalum are shown in black, grey colour shows trails and trees used by Lasius fuliginosus; for details see the legend in Figure A2-1.

5. Study B: Worker size polymorphism

Our preliminary observations suggested that Liometopum microcephalum workers were rather variable, within and among colonies. Their body length ranges from 3 to 7 mm (EMERY 1916, STITZ 1939, SEIFERT 2007). Contradictory information about the species’ degree of polymorphism had been published and it remained unclear, if the species was monomorphic (EMERY 1912), polymorphic (SHATTUCK 1992) or dimorphic (SEIFERT 2007). The degree of polymorphism is determined by iso- or allometric growth of body parts. WILSON (1953) distinguished five degrees, which corresponded to the evolution of ant worker castes: monomorphism (the lowest level), simple allometry, diphasic allometry, triphasic allometry, and complete dimorphism (the highest level). Schemes of relations between body parts that are characteristic for the above mentioned degrees, are shown in Figure 2. The existence of polymorphism is closely connected to labour division within a colony, but it is also affected by several other factors, such as the physical conditions at a site, presence of competitors in the vicinity of a nest (PASSERA et al. 1996, MCGLYNN & OWEN 2002, MCGLYNN et al. 2012) and the nutrition of larvae during their development (WHEELER 1991). Therefore, I considered in my study several factors that could potentially affect the worker size and degree of polymorphism. The type of the site (habitat) is linked with the amount of available food resources. I compared three habitat types. The number of foraging trees was highest in floodplain forests, accompanied by many potential food sources in bushes and numerous seedlings. The park landscape was characterised by large trees and bushes in meadows, and in xerothermous forests only trees of smaller diameter and sparse understorey (compared to the other two habitats) were present. Another factor potentially affecting worker size was the presence of a competitor in the vicinity of a L. microcephalum nest tree. The main competitors were Lasius fuliginosus, Formica rufa (present only at one site) and other colonies of L. microcephalum. Colonies without any competitor in their vicinity were included as controls. As the degree of polymorphism can change in the course of time, I sampled the colonies twice a year, in the beginning (April-May) and in the second half (July-August) of the species’ annual activity period. I tracked their trails and recorded the areas occupied by the ants, drawing their position and borders with help of a squared grid. The territories of all studied colonies are shown in Appendix B1. The territory size was calculated as the sum of distinct areas occupied by the ants on the ground surface, including nest and foraging trees trunk cross- section (I did not include the area of tree crowns because I was not able to determine the portion of a given tree crown truly occupied by the ants). Territory size should correspond to the size, vitality and success of each colony. Workers were collected from the trunks of their nest trees, from trails on the ground 5-10 meters far from the nest tree (depending on the territory size) and from the trunks of their nest trees after I had transferred their competitors onto these (presuming that such situation would lead to the appearance of potentially specialized defenders - majors or even soldiers if such existed).

Figure 2: Allometric relations of measured body parts (x, y) corresponding to the five degrees of polymorphism proposed by WILSON (1953). The degrees are ordered from the lowest on the left to the highest on the right; k = slope of the regression line fitted in the bivariate plot; the histograms in the right corner of each plot show uni- or bimodal distribution of size categories of the measured body parts.

Before I developed the final study design, I had measured about 300 workers from three different colonies (one colony from Pohansko site, two colonies from Rendezvous site). I had measured six characters: head width, head length, alitrunk length, pronotum height, hind femur length and hind tibia length. After preliminary analysis, I decided to measure only four characters. In a polymorphic species, the head shape and size are usually the most variable (WILSON 1953) whereas legs are proportional to the total body size (TSCHINKEL et al. 2003). Measurements of alitrunk and pronotum can be biased by the position of the ant under the microscope and thus they may not be absolutely accurate. Moreover, the latter two characters did not reflect the allometric relations better than the head size, therefore I did not include them in the subsequent analyses. Workers collected from each colony were divided into two or three groups using the k- means clustering method. I constructed bivariate plots with worker sizes, fitted a regression line in each cluster of sizes, and tested the differences between slopes of the lines using analysis of covariance (see the paper for details). Differences between groups of workers collected at different places (within a single colony) were tested using Linear Mixed Models. The significance of factors potentially affecting the worker size (territory size, season, competitor, and type of habitat) was calculated using marginal models with General Least Squares function (PINHEIRO & BATES 2000). Minors, media and majors, with continuous variability (a gradual increase) of their sizes, and thus not representing clearly distinguishable castes, were found in all studied colonies (Figure 3). In all colonies at least one measured body part showed an allometric relationship with femur length. That confirms that L. microcephalum is indeed a polymorphic species. Bivariate plots of all measured characters are shown in Appendix B2. From one colony, two individuals conspicuously different from the other workers were collected (Figure 4). My first impression was that these might represent a distinct worker caste. However, my search for additional big-headed individuals was not successful. They were neither found in any other colony in the same year nor in the same or other colonies in the subsequent year. They probably represented workers infested by parasites, which had caused a head deformation (WHEELER 1928, TRABALON et al. 2000, CSÖSZ 2012). Workers collected on their nest trees did not differ from those collected far from the nest. An apparent difference found between groups of workers collected before and after

Figure 3: Selected heads of measured workers from a single colony (R1) showing the continuous variability in their sizes.

Figure 4: One of the two distinct workers (left) found in colony L3 having a much bigger head of different shape than had the other, normal, workers in that colony (right).

simulated enemy attacks was non-significant. Workers collected in spring were significantly larger than those collected in summer. Overwintering workers prevail in ant colonies in early spring while in summer the colonies produce large numbers of workers, probably giving priority to numbers instead of body size. Worker size correlated positively with territory area. The territory areas were larger in spring and were affected by the presence of competitors (colonies neighbouring L. fuliginosus nests defended the largest territories whereas those having conspecific neighbours had the smallest territories). The effect of habitat was not significant, although it seemed that in the xerothermous forest the workers and territories were smaller than

in the other habitats. However, this observation was based on merely two colonies found in this habitat. The sampling was more or less selective due to my effort to capture workers of all possible sizes present within a colony. This surely resulted in a bias in regard to the numbers of workers within specific size categories. For example, in some colonies a lower number of medium-sized workers was collected. That resulted in a “gap” in the bivariate plots that could be misinterpreted as a tendency to dimorphism. However, the slopes of the fitted lines did not differ significantly. In some colonies, outliers were present in the bivariate plots. Although 200 workers sampled per colony should represent a sufficiently big sample, accuracy would further increase with the number of workers measured. Therefore I cannot exclude that the plots (and corresponding slopes of the regression lines) would be slightly different if I would have measured, for instance, two thousand workers per colony.

Myrmecological News 20 101-111 Vienna, September 2014

Worker polymorphism in the arboricolous ant Liometopum microcephalum (Hymeno- ptera: Formicidae: Dolichoderinae): Is it related to territory size?

Lenka PETRÁKOVÁ & Jiří SCHLAGHAMERSKÝ

Abstract

Liometopum microcephalum (PANZER, 1798) is a rare arboricolous ant, which forms large colonies of high ecological importance and ranks at the top position in the hierarchy of ant assemblages. Many aspects of the species' biology remain unknown due to its scattered occurrence and bad nest accessibility. Published information is sometimes incon- sistent, such as in the case of worker polymorphism. Our objectives were (1) to determine the level of polymorphism, (2) to ascertain if workers occupied by different tasks differ in size, and (3) to assess the effect of competitors, habitat type and territory size (as a proxy for colony size) on worker size. Fifteen colonies, with or without a competing ant species in their vicinity, were sampled in spring and summer 2011 in the south-eastern part of the Czech Republic. Head width, head length, hind femur and tibia lengths were measured as indices of worker size. Territory areas, assessed during each sampling, served as an indicator of the size and vitality of individual colonies. Worker size variability was continuous, with a broad size range for all measured characters within each studied colony. We found different levels of polymorphism for measured body parts in individual colonies: isometry and simple, diphasic and triphasic allometry; in most colonies, the level changed in the course of time. Generally, workers collected in spring were larger than those collected in summer (p < 0.0001). We did not find any differences between workers performing different tasks outside the nest. We found a positive correlation between territory size and body size, represented by mean femur length (p = 0.0068). Territory size was affected by the presence of behaviourally dominant ant species (p = 0.0036), in particular Lasius fuliginosus (LATREILLE, 1798). We conclude that in contrast to information in literature the species is not truly dimorphic and even colonies seemingly made of workers of one size class contain a wide range of worker sizes. Key words: Liometopum, worker size, polymorphism, allometry, territory, competitors. Myrmecol. News 20: 101-111 (online 3 July 2014) ISSN 1994-4136 (print), ISSN 1997-3500 (online) Received 13 August 2013; revision received 19 February 2014; accepted 24 February 2014 Subject Editor: Alexander S. Mikheyev Lenka Petráková (contact author) & Jiří Schlaghamerský, Department of Botany and Zoology, Faculty of Science, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic. E-mail: [email protected]

Introduction The existence of polymorphism is closely connected to whereas elder workers perform tasks in the open. This is labour division within a colony of social insects and helps a case of polyethism (= behavioural change during an in- to increase colony fitness through more efficient exploita- dividual's life). Polymorphism and polyethism do not ex- tion of available resources (OSTER & WILSON 1978, WIL- clude each other and can both occur in the same species SON 1980). Polymorphism arose many times in the phylo- (MADI & JAFFE 2006). Some studies indicate that at least genetic history of ants (WILSON 1953). At present, 15 - 20% in some ant species the ratio between minor workers and of all described ant species are said to have polymorphic major workers or soldiers can shift based on environmental workers (HÖLLDOBLER & WILSON 1990). Striking differ- conditions, in particular competition with other ant colo- ences between worker castes present within one colony are nies of the same or other species (PASSERA & al. 1996, seen mainly in tropical representatives such as in army ants MCGLYNN & OWEN 2002, MCGLYNN & al. 2012). (Eciton spp.) or leaf-cutting ants (Atta spp.). In the tem- Worker polymorphism is characterized by allometric perate zone, Camponotus and Messor are examples of ant growth of body parts. Particular morphs can differ not genera occurring in Europe that have distinct worker castes. only in size and shape of some body parts (mostly head In ants with worker polymorphism, morphologically dis- and mandibles; WILSON 1953) but also in physiological tinct worker individuals, which often perform specialized features (production of different gland secretions or hind functions, are produced within a single colony. Typical- gut enzymes; ROBINSON 2009). They can even have dif- ly, the majors (or soldiers) defend the nest and territory, ferent qualities, for instance in terms of longevity, running whereas minors ensure brood care (HIGASHI & PEETERS speed or tolerance to high or low temperatures (CERDÁ & 1990, SCHWANDER & al. 2005, MERTL & TRANIELLO 2009). RETANA 1997). Size variation is not always determined by In some species, morphologically distinct castes are re- genetic factors (SCHWANDER & al. 2005). It seems that the placed by age groups: Young workers stay inside the nest, nutrition of larvae is crucial for their future appearance as adults (WHEELER 1991). During metamorphosis, tissues of minors present in some colonies. Our observations on the a developing individual interact and compete for limited large population in South Moravia (SCHLAGHAMERSKÝ & nutrients; thereby some body parts grow faster than others OMELKOVÁ 2007, PETRÁKOVÁ & SCHLAGHAMERSKÝ 2011, (NIJHOUT & WHEELER 1996). This disproportionate growth SCHLAGHAMERSKÝ & al. 2013) confirmed the occurrence results in allometry (TSCHINKEL & al. 2003). Generally, of a wide range of worker sizes even within single colo- well-fed larvae develop into majors, whereas starving lar- nies; moreover some colonies seemed to consist of subs- vae become minors. The number of individuals belonging tantially smaller workers than other colonies. The latter ob- to a given caste can be modified at colony level through servation might be the consequence of the age and vitali- hormonal regulation (WHEELER & NIJHOUT 1984). The de- ty of a colony: Presumably young (newly established) or gree of polymorphism can change during a colony's life very old (declining) colonies might produce mostly minors, (TSCHINKEL 1988). At least one reason is the high ener- the same could apply for mid-aged colonies barely survi- getic cost of the production of large workers, hardly af- ving under adverse conditions. fordable for a newly established colony. Owing to the above-mentioned biology of the species According to WILSON (1953) several levels of poly- it would be very difficult to measure the size, age and vi- morphism, which respond to the evolution of castes, are dis- tality of colonies. Thus we decided to study worker poly- tinguished. At the bottom of that hierarchy is monomor- morphism in relation to territory area. Territory size should phism, which is characterized by isometric growth of body be correlated with colony size but may also reflect colony parts, a narrow range of size variability and a unimodal success in contests with competitors and food density in a frequency distribution of measured sizes. The following, given habitat. The shape and size of ant territories change polymorphic, levels are characterized by allometry. Allo- in the course of time (e.g., AKINO & YAMAOKA 1999, TAR- k metric growth can be expressed as y = b * x , where x and TALLY 2006) as a consequence of resource preference and y represent two body parts, b is the initial growth index and competition with neighbouring colonies (of the same as k is a growth constant obtained as the slope of the regres- well as other species of similar requirements and com- sion line fitted to log-transformed data in a bivariate plot petitive strength). We observed the same in Liometopum of the two measured body parts. In the case of allometry microcephalum during a preceding study (PETRÁKOVÁ & the slope of the line always differs from one. The lowest SCHLAGHAMERSKÝ 2011). Defending the territory can be level is called monophasic (or simple) allometry: The fre- costly and thus large and vital colonies are favoured. We quency distribution of measured values is unimodal, the hypothesized that vital and successful colonies, i.e., colo- size variability is small, but its lower and upper extremes nies with large territories, would produce workers of big- can represent functional castes. In diphasic allometry, the ger size than colonies defending smaller territories. regression line breaks into two segments with different In the present study we want to quantify worker size slopes, corresponding to minor and major workers; the fre- variability and to clarify the level of polymorphism in Lio- quency distribution is bimodal. Similarly, in the case of metopum microcephalum. Is it the same in all colonies triphasic allometry, three line segments with different slopes within our study area? Does it change in the course of the are present, often forming a curve of sigmoid shape; the annual activity period? Further we ask if worker size is outer segments represent relatively stable minor and major connected to the different tasks performed by these indi- worker castes, whereas the middle segment, representing viduals (defence, foraging) within a single colony. Finally, intermediate workers, is often unstable. Stabilizing selec- we investigate if worker size depends on the colony's ter- tion leads to complete dimorphism, which is considered ritory size, the presence of major competitors or the type to be the highest degree of polymorphism: There are two of habitat. segments of the regression line, corresponding to minor Methods and major workers, that are separated by a gap. Liometopum microcephalum (PANZER, 1798) is a rare The study was carried out on 15 colonies of Liometopum arboricolous ant of Pontomediterranean distribution in Eur- microcephalum in South Moravia (south-eastern part of the ope and the Middle East. It forms large, sometimes poly- Czech Republic). Five of the studied colonies neighboured domous, colonies of high ecological importance and be- with Lasius fuliginosus (LATREILLE, 1798) or Formica havioural dominance within ant communities (PETRÁKOVÁ rufa LINNAEUS, 1761 colonies, another five were found & SCHLAGHAMERSKÝ 2011). Workers are aggressive and close (less than 30 m) to another L. microcephalum colony actively defend their territories. They are partly zoophag- (we first transferred several workers from one nest tree to ous, partly feed on honeydew and nectar (WIEST 1967, the other and then observed the workers' reaction as to SCHLAGHAMERSKÝ & al. 2013). The species builds its nests see if the workers belonged to the same or to two distinct within trunks or limbs of old and mighty trees, usually colonies). The remaining five colonies had no nest of a several metres above ground. Due to its limited distribu- strong competitor (in terms of behaviourally dominant ant tion and bad nest accessibility, many aspects of the spe- species) in their vicinity. The study sites included three cies' biology remain unknown. Published reports on the ant types of habitat, representing different conditions in terms give a range of worker body length of 3 - 7 mm (EMERY of moisture and resource availability (nest sites, food re- 1916, STITZ 1939) and rather contradictory information on sources): A landscape park with ancient oaks in meadows the level of polymorphism. According to EMERY (1912), had the largest oaks of all compared habitats, with large L. microcephalum is monomorphic, whereas according to interspaces partially filled by shrubs (site L – seven colo- SHATTUCK (1992) the species is polymorphic, with minor nies), floodplain forests with oak as one of the dominant workers lacking ocelli. However, according to AGOSTI & tree species had the highest density of trees, saplings and COLLINGWOOD (1987) all workers possess ocelli. Accord- shrubs (sites: P – one colony, R – four colonies, Z – one ing to SEIFERT (2007) the species is dimorphic, with only colony) and a xerothermous forest (site M – two colo-

102 nies) that provided the least suitable conditions (lack of fitted regression lines through each group of points in all old trees of large diameter for nest building, lower vege- bivariate plots and assessed their slopes (outliers were ex- tation density in comparison to the floodplain forests). cluded from the analysis only when their Cook's distances We sampled the colonies twice in 2011: In spring (from were higher than 0.5). Differences between slopes of re- mid April to mid May) and in summer (from end of June gression lines found in a given colony per sampling date to early August), except three colonies that were sampled were tested using analysis of covariance (e.g., HW = α + only once due to their poor accessibility. At both times we βFL + FL : group). If an interaction between continuous collected the ants from the trunks of the nest trees. Fur- (FL) and categorical (group) variables was significant on a thermore, in spring we collected workers that were mov- p-level lower than 0.1, the slopes of the fitted regression ing on trails in 5 - 10 metres distance from their nest trees. lines were considered different (p-levels under 0.05 indi- In the case of colonies neighbouring with competing col- cated diphasic allometry, or dimorphism – when there was onies, we also sampled workers from their nest trees after a gap between two segments, or triphasic allometry – when induced enemy attacks (we transferred 20 - 30 workers of there were three segments; p-levels between 0.1 and 0.05 the competitor to the nest tree trunk and after 10 - 15 mi- indicated slightly diphasic allometry). The approximate nutes we collected workers of the defending colony from position of breakpoints was then assessed as the range be- this trunk segment) during the summer sampling. At each tween the maximal value measured in the group of "minor" sampling date we drew maps of territories consisting of workers and the minimal value in the group of "majors". trails connecting nest trees with foraging trees and with If there were no differences between the slopes represent- patches on the ground where workers foraged, either indi- ing "minors" and "majors", a single slope was calculated vidually or in foraging columns. Based on these maps we for all data points representing given body parts measured assessed the territory area for each colony and date sepa- for a single colony at a given date. Such colonies were con- rately. sidered to be simple allometric or isometric (if the slopes The ants were sampled selectively, with the objective to of the regression lines did not differ from 1). Allometric capture workers of all sizes present and thus to record the relations were tested also with ANCOVA as differences of entire size variability. From each group of workers (col- the given slopes from a slope equal to 1. lected from nest trees, from trails and from nest trees af- To compare groups of workers collected from nest ter induced enemy attacks) we measured 50 individuals. trees and from trails as well as from nest trees before and Maximal head width (HW), head length (HL; from lower after induced competitor attack we used a Linear Mixed margin of clypeus to the end of vertex), hind femur length Model (lme function in nlme package; PINHEIRO & BATES (FL), and hind tibia length (TL) were measured. Head 2000) with "colony" as random effect, which takes into width is the most variable character in polymorphic ant account possible differences between individual colonies. species (WILSON 1953, ARAUJO & TSCHINKEL 2010) where- Additionally, we tested the differences in head widths and as femur length rather correlates with total body size. As femur lengths with the t-test, for each colony separately. shown, for instance, in Solenopsis invicta BUREN, 1972, the The same approach was used in comparing sizes of wor- relation between leg length and total body size is isometric kers collected in spring and summer. (TSCHINKEL & al. 2003). Therefore femur length served To assess the dependence of worker size on territory as an index of worker body size in our study. We did not size as well as the effect of season, competitor and type measure total worker size to avoid bias caused by variable of habitat, we used marginal models with General Least gaster size – workers returning from foraging trees back Squares function (nlme package, PINHEIRO & BATES 2000). to the nest may have a markedly larger gaster than others As almost all colonies were sampled two times, measure- due to honeydew collection (SCHLAGHAMERSKÝ & al. 2013). ments were not independent, so we inserted correlation Additionally, we analysed workers conserved in ethanol structure (compound symmetry correlation) into the mod- from the inner part of a nest. This sample had been col- els, taking colony identity into account. Two datasets were lected by our colleague M. Omelková from an oak imme- tested: The first included all measured workers, the sec- diately after its uprooting in summer 2005 (also in South ond the average values calculated from the largest and Moravia). Larvae and pupae prevailed in the sample. We the smallest worker sampled per colony and sampling date measured all 119 workers in the sample to assess the body (i.e., one data point per colony and sampling date). The size of workers taking care of the brood. latter values were used to avoid bias caused by sampling Statistical analysis: All statistical analyses were done – in some colonies a higher percentage of majors than in in R 2.10.1. software (R DEVELOPMENT CORE TEAM 2011). others could have been sampled accidentally. The analyses script is available in the Appendix S1 (as di- Results gital supplementary material to this article, at the journal's web pages). For each colony and date we constructed bi- Worker size variability, degree of polymorphism: We variate plots of the measured body parts to determine the found a broad range of measured sizes for all characters degree of polymorphism, with the femur length always de- in every studied colony. Head width (HW) varied from picted on the x-axis. We also plotted frequency distribu- 0.825 mm (± 0.044 mm SD; average of minimal values tions of the measured sizes to see if the variability of sizes measured in each colony) to 1.634 mm (± 0.068 mm SD; was continuous or not. The workers were then divided into average of maximal values), head length (HL) varied from groups corresponding to the number of segments (clouds 0.806 mm (± 0.06 mm SD) to 1.542 mm (± 0.109 mm of points) using k-means cluster analysis (HARTIGAN & SD). Femur length (FL) varied from 0.794 mm (± 0.076 mm WONG 1979). That method allows partitioning the data SD) to 1.532 mm (± 0.052 mm SD) and tibia length (TL) points into k groups by finding the minimal sum of the varied from 0.731 mm (± 0.071 mm SD) to 1.460 mm (± within-groups sum of squares (BORCARD & al. 2011). We 0.053 mm SD) on average. All measurements are avail-

103

Fig. 1: Examples of different allometric relations between head width and femur length. Points show individual ant workers, lines show the fitted regressions with different slopes. a) Colony L3 (park landscape, close to another Liometopum microcephalum nest): isometry in summer; b) colony Z1 (floodplain forest, without competitor): simple allometry in spring; c) colony R3 (floodplain forest, close to another L. microcephalum nest): diphasic allometry in summer – minors = black, majors = grey; d) colony R2 (floodplain forest, close to Lasius fuliginosus and Formica rufa nests): triphasic allometry in summer – minors = black, majors = grey, intermediates = white triangles. See Table 1 for slope values.

able in the Appendix S2. Frequency distribution plots as to 1.168 for TL / FL. In those cases in which R2 values were well as bivariate plots showed that the worker size variabi- low and SE values high in the regression models (Tab. 1) lity was continuous without any conspicuous gap between this was the consequence of low numbers of individuals minors and majors. within the corresponding groups or of a small range of wor- Different body parts had often different degrees of al- ker sizes within the groups. In colonies with diphasic allo- lometry, even within the same colony. We detected both metries, the breakpoints of the regression lines were situated positive (k > 1) and negative (k < 1) allometries and also approximately between 1.10 mm and 1.30 mm of femur isometry (k = 1). The TL / FL relationship was in most length (variation between colonies). Workers with femur cases isometric whereas the HW / FL relationship was lengths above these values had ocelli, whereas minor wor- mostly simply allometric or diphasic (see Tab. 1). Colo- kers lacked them. In the case of triphasic allometry two nies varied in allometries also in the course of the season. breakpoints were present (at 0.99 - 1.04 mm and 1.16 - One colony (R2) that had been diphasic in spring was tri- 1.24 mm of FL). Intermediate workers did not have true phasic in summer (Fig. 1). ocelli but instead only three small dark and raised planes In spring, slopes of regression lines varied from 0.808 could be recognized in their place. to 1.233 for the HW / FL relationship, from 0.859 to 1.113 In spring 2011, we found in one colony (L3, neigh- for HL / FL and from 0.695 to 1.103 for TL / FL. In sum- bouring with two other Liometopum microcephalum colo- mer, they varied from 0.532 to 1.300 for the HW / FL rela- nies) two individuals distinctly different from the rest of the tionship, from 0.594 to 0.994 for HL / FL and from 0.777 measured workers. These had conspicuously large, rounded

104 Tab. 1: Slopes of the regression lines fitted into bivariate plots. The p-values show the significance of differences between the slopes of the regression lines. In cases of simple allometry the values resulting from testing the difference between groups obtained by k-means cluster analysis are presented in brackets. Where two values are given in one row, the top values correspond to the group of minor workers and the bottom ones to that of majors. R2 (coefficient of determination) indicates the quality of the fitted regression model. Abbreviations: BP = breakpoint, HW = head width, HL = head length, TL = tibia length, FL = femur length.

spring summer Colony Ratio BP BP Allometry P Slope SE R2 Df Allometry P Slope SE R2 Df (μm) (μm) HW/FL simple - (0.6739) 1.094 0.03 0.95 98 simple - (0.11) 1.073 0.03 0.96 47 Z1 HL/FL isometry - (0.8883) 0.969 0.03 0.90 98 simple - (0.919) 0.920 0.03 0.94 47

TL/FL isometry - (0.5612) 0.998 0.03 0.92 98 isometry - (0.8778) 1.007 0.04 0.94 48 slightly 1141- 1.030 0.05 0.91 34 HW/FL 0.0612 simple - (0.4438) 1.085 0.03 0.92 97 diphasic 1265 0.809 0.10 0.50 60 L1 HL/FL simple - (0.2063) 0.913 0.04 0.82 98 simple - (0.7553) 0.932 0.03 0.89 97 TL/FL simple - (0.2451) 0.920 0.03 0.89 98 isometry - (0.4000) 1.021 0.03 0.94 96 HW/FL isometry - (0.8621) 1.045 0.04 0.88 98 simple - (0.1492) 1.115 0.03 0.95 97 L2 HL/FL isometry - (0.5364) 0.953 0.04 0.85 98 isometry - (0.3029) 0.953 0.03 0.93 97 TL/FL isometry - (0.9577) 0.973 0.03 0.91 98 isometry - (0.4671) 0.960 0.03 0.93 97 1248- 0.808 0.08 0.77 30 HW/FL diphasic 0.0409 isometry - (0.4407) 1.017 0.04 0.88 98 L3 1257 1.076 0.10 0.64 66 HL/FL isometry - (0.1530) 0.962 0.04 0.84 98 simple - (0.1468) 0.917 0.03 0.88 98 TL/FL isometry - 0.1033 0.998 0.03 0.93 98 simple - (0.8135) 0.958 0.02 0.94 98 HW/FL simple - (0.1716) 1.097 0.03 0.93 96 simple - (0.9639) 1.242 0.05 0.94 46 HL/FL isometry - (0.4632) 0.962 0.04 0.87 97 isometry - (0.4782) 0.985 0.04 0.91 46 L4 slightly 1210- 1.103 0.08 0.83 39 1237- 0.777 0.18 0.49 31 TL/FL 0.0728 diphasic 0.0237 * diphasic 1236 0.900 0.08 0.71 57 1279 1.168 0.18 0.64 10 HW/FL simple - (0.3864) 1.069 0.04 0.88 98 simple - (0.6448) 1.098 0.02 0.98 95 1150- 0.629 0.07 0.75 28 L5 HL/FL isometry - (0.8589) 0.978 0.04 0.89 98 diphasic 0.0026 * 1173 0.886 0.05 0.80 68 TL/FL isometry - (0.5868) 1.008 0.03 0.91 98 isometry - (0.8161) 0.994 0.02 0.94 97

HW/FL simple - 1.070 0.03 0.92 98 simple - (0.5274) 1.146 0.05 0.93 47 (0.9011) L6 1221- 0.982 0.05 0.92 27 HL/FL diphasic 0.0193 * isometry - 0.9122 0.982 0.05 0.90 47 1261 0.860 0.08 0.64 68 TL/FL isometry - (0.5897) 0.965 0.03 0.94 98 isometry - (0.4457) 1.052 0.03 0.95 48 1243- 1.020 0.06 0.90 30 slightly 1163- 0.950 0.08 0.92 12 HW/FL diphasic 0.049 * 0.0689 1268 1.233 0.09 0.77 62 diphasic 1184 1.028 0.08 0.83 33 L7 1243- 0.859 0.08 0.80 30 HL/FL diphasic 0.0561 simple - (0.5102) 0.949 0.03 0.97 47 1268 1.113 0.11 0.64 62 TL/FL isometry - (0.2929) 0.998 0.02 0.95 98 isometry - (0.6819) 1.013 0.03 0.95 47 1237- 1.001 0.05 0.87 66 HW/FL ------diphasic 0.0515 1265 1.161 0.06 0.84 79 M1 HL/FL ------isometry - (0.5546) 0.989 0.02 0.93 147 TL/FL ------isometry - (0.4382) 0.985 0.03 0.87 123 1129- 0.734 0.08 0.67 43 HW/FL ------diphasic 0.0028 * 1163 1.087 0.06 0.79 102 M2 1129- 0.594 0.07 0.63 43 HL/FL ------diphasic 0.0072 * 1163 0.894 0.05 0.72 102 TL/FL ------isometry - 0.6647 1.026 0.02 0.92 148 HW/FL simple - (0.3909) 1.131 0.05 0.86 97 ------P1 HL/FL isometry - (0.8693) 0.983 0.05 0.78 98 ------TL/FL isometry - (0.2206) 1.033 0.04 0.87 98 ------slightly 1211- 0.956 0.09 0.84 23 HW/FL simple - (0.4719) 1.171 0.03 0.95 98 0.08 diphasic 1226 1.152 0.07 0.77 73 R1 slightly 1211- 0.804 0.08 0.81 23 HL/FL simple - (0.2507) 0.946 0.03 0.93 98 0.0853 diphasic 1226 0.994 0.07 0.71 73 TL/FL diphasic 1283- 0.0282 0.990 0.09 0.77 33 isometry - (0.2494) 0.984 0.02 0.96 98

105

1292 0.695 0.09 0.46 63 992- 0.0082 0.946 0.12 0.74 23 slightly 1190- 0.866 0.06 0.88 30 1037, HW/FL 0.0667 triphasic 0.532 0.09 0.65 18 R2 diphasic 1273 1.044 0.08 0.74 66 1160- 0.0001 1.227 0.11 0.74 48 1235

HL/FL simple - (0.1823) 0.926 0.03 0.91 98 simple - (0.5362) 0.966 0.03 0.94 93 TL/FL isometry - (0.9881) 0.967 0.03 0.92 98 isometry - (0.4441) 1.008 0.02 0.97 97 1202- 0.0320 0.878 0.09 0.77 27 HW/FL simple - (0.5744) 1.194 0.04 0.91 96 diphasic 1207 * 1.300 0.13 0.83 21 R3 HL/FL simple - (0.3308) 1.061 0.04 0.88 96 isometry - (0.3615) 0.972 0.04 0.91 48 TL/FL simple - (0.5148) 1.057 0.03 0.91 98 isometry - (0.7772) 1.012 0.03 0.96 47 1048- 0.0006 0.785 0.07 0.76 38 HW/FL simple - (0.449) 1.095 0.03 0.94 96 diphasic 1089 * 1.035 0.06 0.86 52 R4 1048- 0.0055 0.656 0.07 0.70 40 HL/FL isometry - (0.7962) 1.030 0.03 0.90 96 diphasic 1089 * 0.920 0.07 0.78 54 TL/FL isometry - (0.3128) 0.968 0.04 0.91 98 isometry - (0.4162) 0.957 0.04 0.93 98

heads (HW: 1.71 and 1.72 mm) disproportionate to their previously studied colonies. However, neither rather small bodies (FL: 1.42 and 1.35 mm) and, in differences in head width nor in femur length were particular, to their small gasters. significant. We did not find any additional workers with We found differences in worker size between the two conspicuously big heads. sampling dates: In most colonies workers were bigger in We compared worker sizes of individuals collected spring than in summer (Fig. 2). These differences were from the inside of a nest (including also numerous brood; significant in five colonies (Welch Two Sample t-test; collecting site: Lednické rybníky, summer 2005 - exact R2 – FL: t = 2.57, df = 190, p = 0.0109; R3 – HW: t = sampling date unknown, leg. M. Omelková) with that of 3.99, df = 89, p = 0.0001, FL: t = 4.23, df = 88, p < every colony sampled by us in 2011. The workers from 0.0001; R4 – HW: t = 5.55, df = 197, p < 0.0001, HW: t most of our colonies (n = 11) were significantly smaller = 4.73, df = 198, p < 0.0001; L3 – HW: t = 2.82, df = than the workers found inside the nest (Welch Two 198, p = 0.0053, FL: t = 2.34, df = 197, p = 0.0202; L6 – Sample t-test; FL: M1 - t = -4.067, df = 260, p < 0.0001, FL: t = 2.55, df = 101, p = 0.0122). Only in one case M2 - t = -7.141, df = 242, p < 0.0001; Z1: t = -6.455, df were workers significantly smaller in spring than in = 256, p < 0.0001; L1: t = -4.518, df = 228, p < 0.0001; summer: that colony (L4) was situated in a landscape L3: t = -5.676, df = 269, p < 0.0001; L4: t = -3.449, df = park and neighboured two independent L. 238, p = 0.0007; L5: t = -3.848, df = 257, p = 0.0002; L7: microcephalum colonies (Welch Two Sample t-test, HW: t = -2.666, df = 265, p = 0.0081; R2: t = -4.402, df = t = -2.68, df = 104, p = 0.0086; FL: t = -2.51, df = 113, p 286, p < 0.0001; R3: t = -2.003, df = 251, p = 0.0463; = 0.0133). When all colonies were tested together, the R4: t = -10.68, df = 246, p < 0.0001). Only one colony difference between spring and summer was also (P1, situated in a floodplain forest and neighbouring with significant (LME; FL: F = 39.45, df = 2533, p < 0.0001; a Lasius fuliginosus colony) consisted of workers that HW: F = 32.81, df = 2533, p < 0.0001). were larger than those sampled inside the nest (Welch Workers size versus tasks partitioning: Workers Two Sample t-test; FL: t = 3.263, df = 210, p = 0.0013). collected on nest trees had a similar size as the workers We found diphasic allometry in the workers collected collected on trails (LME; HW: F = 0.705, df = 1484, p = from the nest (slopes of regression lines in HW: k1 = 0.4013; FL: F = 0.027, df = 1484, p = 0.8695). Testing 0.884, k2 = 1.112, slopes significantly different: p = individual colonies, none showed significant differences 0.0096). in this respect. Factors potentially affecting worker size: When we Workers collected after simulated attacks were used all measured femur lengths for modelling of the somewhat larger than workers collected before these relationship between worker size and all explanatory attacks, however, only in one of nine colonies this variables, we found that the sampling date had a crucial difference was statistically significant (R1, situated in a effect on the worker size (femur length). Workers wet forest and neighbouring with a Lasius fuliginosus collected in spring were bigger than workers collected in colony; Welch Two Sample t-test; HW: t = -2.88, df = summer (GLS: F = 39.2, df = 2549, p < 0.0001). The 98, p = 0.0049; FL: t = -2.9, df = 98, p = 0.0046). When second most important, albeit only marginally workers from all colonies were tested together the significant, explanatory variable was the territory size difference between the size of workers collected before (GLS: F = 3.51, df = 2549, p = 0.0612). Colonies and after an attack was not statistically significant (LME; neighbouring with Lasius fuliginosus nests consisted of HW: F = 2.121, df = 889, p = 0.1456; FL: F = 2.877, df = the biggest workers (Fig 4). However, taking into 889, p = 0.0902; Fig. 3). We repeated the experiment account colony identity, the differences were not with simulated attacks in spring 2012 using five of the statistically significant (GLS: F = 1.95, df = 2549, p =

106

0.1426). Differences between colonies neighbouring with titor (GLS: F = 3.81, df = 27, p = 0.0366). Colonies L. fuliginosus and those neighbouring with other neighbouring with L. fuliginosus had the largest

Fig. 2: Measured sizes (head width and femur length range) of workers collected in spring (shown in dark grey) and in Fig. 4: Differences in femur lengths of L. microcephalum summer (in light grey). Measured values: medians (bars) workers depending on the presence of a competitor (either with 95% confidence limits (notches), 25–75% quantiles Lasius fuliginosus or another colony of L. microcephalum). (boxes), 1.5 interquartile ranges (whiskers) and outliers Values predicted using marginal models with General Least (circles); sample sizes: 1300 workers were measured in Squares function and correlation struc-ture: femur length spring, 1249 workers in summer. mean values with 95% confidence intervals; sample size: 2550 workers (900 workers collected from colonies neighbouring L. fuliginosus colonies, 900 workers from colonies adjacent to other, hostile L. microcephalum colonies and 750 workers from colonies not adjacent to colonies of any strong competitor).

territories of all studied colonies (Fig. 6), differing in size from colonies neighbouring with other L. microcephalum colonies (t = -2.58, df = 27, p = 0.0163). We did not confirm any relationship between territory size and type of habitat. Neither territory area nor average worker size affected the level of polymorphism. Discussion Based on our results we consider Liometopum microcephalum a polymorphic species - in all studied colonies always at least one measured body part showed an allo- Fig. 3: Size of workers (femur lengths) collected before and metric relationship with femur length. Worker size varia- after simulated attacks in spring 2011. Values predicted bility was in all colonies continuous on a broad size range using Linear Mixed Models (LME): femur length mean for all measured characters, i.e. worker size changed values with 95% confidence intervals; sample size 900 gradually and no distinct morphs were detected. The range workers (450 workers measured before and 450 workers of femur lengths was a bit smaller than the head width after the attacks). range. Bivariate plots are often used to visualize physical L. microcephalum colonies were marginally significant (t variability of ant workers (e.g. WILSON 1953, ESPA-LADER = -1.82, df = 2549, p = 0.068). In comparison with the & al. 1990, WHEELER 1991, FERNÁNDEZ & al. 1994, sampling date, territory size, presence of competitors and FRASER & al. 2000, TSCHINKEL & al. 2003). Our results type of habitat had a negligible effect on worker size. showed that different body parts had sometimes different However, we found a positive correlation between terri- degrees of allometry, even within the same colony. We tory size and mean femur length (i.e. mean of the smallest found variability in slopes among colonies, however no and the largest worker collected in a given colony; GLS: F correlation (not even a trend) with the presence of = 8.78, df = 27, p = 0.0068; regression coefficient = competitors, territory size or type of habitat was observed. 0.5433; Fig. 5). Also the sampling date had a significant We consider the colonies to be simply allometric only effect in this model (GLS: F = 4.42, df = 27, p = 0.0462). when the difference of their slopes was above p = 0.1, The territory size depended on the presence of a compe- because even some colonies differing, e.g., at p = 0.08

107 seemed markedly diphasic in the bivariate plots. As the The bivariate plots showed that workers with femur worker size changed in the course of time, so did the lengths exceeding 1.2 mm prevailed in most colonies. This could have been caused by unintentional selective sampling in the field. However, the breakpoints in the bivariate plots were situated near this value. Workers with a femur longer than 1.2 mm usually possessed ocelli. This was in agreement with SHATTUCK (1992), who pointed out the presence of ocelli in majors only. Intermediate workers had only rudiments of ocelli and we did not find any traces of ocelli in minor workers. However, this was also true in colonies with simple allometry and even isometry. We did not find any case of a distinctly dimorphic colony. Di- morphism is characterized by the absence of intermediates, whereas in partial dimorphism at least one measured character forms a continuous chain of points in the bivariate plot and the frequency distribution is bimodal (WILSON 1953). The low number of intermediate workers can be interpreted as a consequence of stabilizing selection, which leads to the existence of distinct morphs.

We observed a similar pattern in two diphasic colonies Fig. 5: Differences in femur lengths of L. microcephalum (L7, R1), which can be interpreted as a tendency to partial workers depending on territory area and sampling date. dimorphism in summer. In another three simply allometric Measured values: white circles = mean femur lengths in colonies (L3, L4 and L5) we also noted lower numbers of spring, black circles = mean femur lengths in summer; individuals with femur lengths approaching 1.2 mm. values predicted using marginal models with General Least Workers collected from the trails were approximately Squares function and correlation structure: dashed line = correlation between territory area and mean femur length in the same size as workers collected from nest trees. The spring, solid line = correlation between territory area and nest of this species is in most cases situated high up in the mean femur length in summer. trunk or in large limbs and thus trails are also present on the tree trunk. The nest trees were occupied exclusively by L. microcephalum workers, and we had assumed that the proportion of larger workers would increase with the distance from the nest trees and thus with an increasing probability of enemy attacks. Majors should be better adapted to move far from the nest – e.g. they have longer legs than minors and their bodies lose water more slowly than minors (LIGHTON & al. 1994). However, also the opposite distribution of majors and minors in space can be found, for instance in the driver ant Dorylus molestus, in which soldiers were found to be more abundant in the vicinity of nests containing brood than in outer parts of its territory (BRAENDLE & al. 2003). We did neither find larger L. microcephalum workers to be more frequent in greater distance from their nest trees than on its trunk, nor the opposite. Also minor workers moved far away from their nest trees or even close to enemies during attacks. PFEIFFER & LINSENMAIR (2001) had observed the same in Camponotus gigas. A broad range of worker sizes participating in foraging allows better exploitation of Fig. 6: Differences in territory area depending on the pres- various food items and prey of a broader size range ence of competitors. Measured values: medians (bars), 25 - (DAVIDSON 1978). 75% quantiles (boxes), 1.5 interquartile ranges (whis-kers) Of the almost 3000 workers measured, only two and outliers (circles); sample size: 27 territories of 15 Lio- individuals were conspicuously different from the others. metopum microcephalum colonies were measured, 13 in Originally, we assumed that they were representing a spring, 14 in summer.). distinct soldier caste, because soldiers are often big- headed. Soldiers of Pheidologeton obtunospinosa, for slopes of regression lines in the bivariate plots. Caste example, plug the nest entrance with their heads and thus structure is not a fixed characteristic – it can change with prevent enemies to intrude into the nest (HUANG & colony age and size (TSCHINKEL 1988). However, we did WHEELER 2011); the same adaptation occurs in Campo- not observe any pattern in these changes. In summer, the notus truncatus (see SEIFERT 2007) and in the genus regression lines were “broken” and in a few cases the Cephalotes – according to POWELL (2009) this is a mor- number of workers decreased near these breakpoints.

108 phological adaptation having direct consequences for territory size in Pogonomyrmex barbatus (Myrmicinae) colony reproduction. That was the reason why we tried to changes rapidly until five years after colony foundation, induce defensive reaction through simulated enemy and then colonies become fully mature. Territory size attacks on the nest trees and thus to provoke the changes over time as a consequence of resource prefe- appearance of these soldiers. We did not find any rences and interactions with competitors. Costs of territory additional individuals of that appearance during these defence are closely connected to costs of the production of trials. We undertook another attempt to get hold of majors or soldiers. Immediately after colony founding, all additional specimens of this assumed soldier caste in April workers are very small as production of major workers is 2012, when we sampled five colonies at the same site quite costly and the colony invests energy into the where the two aberrant specimens had been collected increasing of the worker number instead. In the course of before (we supposed that more soldiers could be produced time increasingly larger workers are produced and once a in spring, when colonies establish their territory borders). colony reaches sufficient size a new worker caste can arise Despite this effort we did not succeed in finding any (HÖLLDOBLER & WILSON 1990). Worker size polymor- similar individuals. We cannot confirm the existence of a phism should be positively correlated with colony age distinct soldiers caste just based on the two individuals (WOOD & TSCHINKEL 1981) and negatively correlated found. A more probable explanation could be that they with the intensity of competition (DAVIDSON 1978). Our presented an aberration. For instance, we cannot exclude a results do not support the latter hypothesis. possible effect of parasites, which can induce morpho- The biggest workers were noted in colonies neigh- logical changes in ants, such as differences in shape and bouring with Lasius fuliginosus nests. These colonies also size of the head, mesonotum and petiolus or missing ocelli occupied the largest territories. L. fuliginosus is an impor- (WHEELER 1928, TRABALON & al. 2000, CSÖSZ 2012). tant competitor of L. microcephalum. Both exploit to a We expected a higher percentage of minors inside the large extent the same resources in terms of food and nest – smaller workers should be more vulnerable and habitat (L. fuliginosus is also preferably arboricolous). L. slower than larger workers. Based on the comparison of fuliginosus is able to combat L. microcephalum very workers found inside of a nest (the only sample from successfully, using effective chemical weapons, whereas inside a nest available to us, collected in 2005) with all the strength of L. microcephalum is based on the workers collected in 2011, the opposite seems to be true. quantitative preponderance of workers biting with their However, the nest sample had been taken at a different site sharp mandibles (PETRÁKOVÁ & SCHLAGHAMERSKÝ (although in the same area) from a downed tree (the 2011). If a colony of L. microcephalum can survive close colony therefore later ceased to exist) and no correspond- to a L. fuliginosus nest, it must be strong enough to ding sample of ants collected outside the nest (and under withstand its competition as well as direct confrontation. undisturbed conditions) had been taken. On the other We propose that selective pressure in that case could lead hand, we can hardly imagine that workers of this colony to increased production of workers with big body size. moving outside the nest could have been larger than those According to MCGLYNN & OWEN (2002) the presentation found within the nest, which were of extraordinary size. of clumped food baits, which had been shown to attract Colonies were sampled two times per year: in spring, more competitors, leads to an increased production of when workers were searching for new food resources and soldiers in the dimorphic, tropical Pheidole flavens. They often moved in foraging columns on the ground, and in assumed that both inter- and intraspecific competition summer, when workers ran mainly along trunk trails were probable factors. On the other hand, we hypothesize leading to foraging trees or other permanent food that L. microcephalum colonies neighbouring with con- resources. That was the reason why the assessed territory specific colonies (thus having the very same niche) limit areas were larger in spring than in summer. Workers each other in terms of colony growth and therefore small- collected in spring were obviously individuals that had bodied workers prevail. However, we have to say that overwintered. Large workers are more resistant to PASSERA & al. (1996) found a higher percentage of suboptimal conditions and live longer than minors soldiers in colonies of the tropical Pheidole pallidula (PORTER & TSCHINKEL 1985b). When a colony is starving experimentally exposed to intraspecific competition com- or the queen dies or ceases oviposition for long periods of pared to colonies without contact with conspecific time, the majors : minors ratio rises. For instance, in the competitors. genus Solenopsis, majors are much more numerous than Positive correlation between worker size and colony minors in early spring because of a seasonally high rate of size was noted in many ant genera (GRAY 1971, PORTER brood production (MARKIN & DILLIER 1971, MARKIN & & TSCHINKEL 1985a), the same holds for colony size and al. 1974). Thus one explanation of the difference in territory area (TSCHINKEL & al. 1995). Younger colonies, worker body size between spring and summer may be that which have usually lower competitive strength, are gene- the workers were representatives of different, albeit rally only able to defend small territories. We considered overlapping, generations. the territory size as an indicator of the colony’s vitality and All studied colonies were at least five years old (i.e. success in competition. In accordance with this as- had been observed in other studies before, e.g. sumption, we observed a positive correlation of worker SCHLAGHAMERSKÝ & OMELKOVÁ 2007, PETRÁKOVÁ & body size with territory area. If a colony reaches a stable SCHLAGHAMERSKÝ 2011), thus potential effects of colony size (in terms of worker number) and has optimal age were likely negligible. According to GORDON (1995), conditions for further growth, it begins to produce sexuals

109 and major workers. If the colony is successful, it can BATCHELOR, T.P. & BRIFFA, M. 2010: Influences of resource- consist of larger workers than a colony living under holding potential during dangerous group contests between suboptimal conditions. Moreover, the level of wood ants. – Animal Behaviour 80: 443-449. polymorphism could be affected by the availability of food BORCARD, D., GILLET, F. & LEGENDRE, P. 2011: Numerical eco- resources because the nutrition of larvae is important for logy with R. – Springer, New York, 306 pp. worker development. Colonies consisting of workers with BRAENDLE, C., HOCKLEY, N., BREVIG, T., SHINGLETON, A.V. & smaller body size have lower resource-holding potential in KELLER, L. 2003: Size-correlated division of labour and spatial comparison with colonies of the same species that have distribution of workers in the driver ant, Dorylus molestus. – larger workers (BATCHELOR & BRIFFA 2010). Another Naturwissenschaften 90: 277–281. aspect that we did not consider in our study is the genetic structure of colonies. As shown for the ant genus CERDÁ, X. & RETANA, J. 1997: Links between worker polymorphism and thermal biology in a thermophilic ant Pheidole, the number of matings of the queen with dif- species. – Oikos 78: 467-474. ferent males correlates positively with the diversity in worker body size within a colony (HUANG & al. 2013) CSÖSZ, S. 2012: Nematode infection as significant source of We summarize that worker size variability in L. unjustified taxonomic descriptions in ants (Hymenoptera: microcephalum was high both within colonies and among Formicidae). – Myrmecological News 17: 27-31. colonies. Colonies having four different levels of poly- DAVIDSON, D.W. 1978: Size variability in the worker caste of a morphism were found within a small part of the species’ social insect (Veromessor pergandei MAYR) as a function of distribution area. Small workers did not possess ocelli, in the competitive environment. – The American Naturalist 112: somewhat larger workers only rudimentary ocelli were 523-532. present, whereas the ocelli of larger workers were well EMERY, C. 1912: Hymenoptera, fam. Formicidae, subfam. developed. Worker size within a colony was positively Dolichoderinae. In: Wytsman, P. (Ed.): Genera insectorum correlated with territory size and marginally with the 137. – V. Verteneuil & L. Desmet, Bruxelles, pp. 1-50. presence of a strong competitor. Workers collected in EMERY, C., 1916: Fauna entomologica italiana I.: Hymenoptera - summer were smaller than those collected in spring. We Formicidae. – Bullettino della Societa Entomologica Italiana found no effect of the type of habitat on worker size (but 47: 79-275. we had only a small sample from dry oak-hornbeam forests due to the rare occurrence of the species in this ESPALADER, X., RETANA, J. & CERDÁ, X. 1990: The caste system of Camponotus foreli EMERY (Hymenoptera: Formicidae). – habitat in our study area). Not a single distinctly dimor- Sociobiology 17 (2): 299-312. phic colony was found within the South Moravian population of L. microcephalum. FERNÁNDEZ, I., BALLESTA, M. & TINAUT, A. 1994: Worker polymorphism in Proformica longiseta (Hymenoptera: Formi- Acknowledgements cidae). – Sociobiology 24 (1): 39-46.

We received funding from the Ministry of Education, FRASER, V.S., KAUFMANN, B., OLDROYD, B.P. & CROZIER, R.H. Youth and Sports of the Czech Republic (Research Plan 2000: Genetic influence on caste in the ant Camponotus con- No. MSM0021622416) and the Czech Science Foundation sobrinus. – Behavioral Ecology and Sociobiology 47: 188-194. (GAČR; grant No. 526/09/H025). Markéta Omelková GORDON, D.M. 1995: The development of ant colony’s foraging sampled workers and brood from the nest analyzed in our range. – Animal Behaviour 49: 649-659. study, Šárka Mašová kindly helped with high power GRAY, B. 1971: A morphometric study of the ant species, Myr- microscopy (investigating the presence of ocelli across the mecia dispar (CLARK) (Hymenoptera: Formicidae). – Insectes worker size range) and Vít Syrovátka gave us advice on Sociaux 18: 95-110. data processing. Max Barclay (Natural History Museum, London) kindly checked the English. Alexander Mik- HARTIGAN, J.A. & WONG, M.A. 1979: A K-means clustering heyev, Terry McGlynn and an anonymous reviewer were algorithm. – Applied Statistics 28: 100–108. instrumental in improving the quality of the article. HIGASHI, S. & PEETERS, C.P. 1990: Worker polymorphism and nest structure in Myrmecia brevinoda Forel (Hymenoptera: References: Formicidae). – Journal of the Australian Entomological Society 29: 327-331. AGOSTI, D.  COLLINGWOOD, C.A. 1987: A provisional list of Balkan ants (Hymenoptera, Formicidae) and a key to the HÖLLDOBLER, B. & WILSON, E.O. 1990: The Ants. – Harvard worker caste, II. Key to the worker caste, including the University Press, Cambridge, MA, 732 pp. European species without the Iberian. – Bulletin de la Société HUANG, M.H. & WHEELER, D.E. 2011: Colony demographics of Entomologique Suisse 60: 261-293. rare soldier-polymorphic worker caste system in Pheidole ants (Hymenoptera, Formicidae). – Insectes Sociaux 58: 539-549. AKINO, T. & YAMAOKA, R. 1999: Trunk trail network of Lasius fuliginosus LATREILLE (Hymenoptera: Formicidae): Distribu- HUANG, M.H., WHEELER, D.E. & FJERDINGSTAD, E.J. 2013: tion between conspecific neighbouring colonies. – Entomolo- Mating system evolution and worker caste diversity in gical Science 2: 341-346. Pheidole ants. – Molecular Ecology 22: 1998-2010.

ARAUJO, M.B. & TSCHINKEL, W. 2010: Worker allometry in LIGHTON, J.R.B., QUINLAN, M.C. & FEENER, D.H. 1994: Is bigger relation of colony size and social form in the fire ant Solenopsis better – water-balance in the polymorphic desert harvester ant invicta. – Journal of Insect Science 10 (94): 1-12. Messor pergandei. – Physiological entomology 19 (4): 325- 334. 110

MADI, Y. & JAFFE, K. 2006: On foraging behavior of the SCHLAGHAMERSKÝ, J. & OMELKOVÁ, M. 2007: The present polymorphic tree dwelling ant Daceton armigerum (Hymeno- distribution and nest tree characteristics of Liometopum ptera: Formicidae). – Entomotropica 21(2): 117-123. microcephalum (PANZER, 1798) (Hymenoptera: Formicidae) in South Moravia. – Myrmecological News 10: 85-90. MARKIN, G.P. & DILLIER, J.H. 1971: The seasonal life cycle of the imported fire ant, Solenopsis saevissima richteri, on the SCHLAGHAMERSKÝ, J., KAŠPAR, J., PETRÁKOVÁ, L. & ŠUSTR, V. gulf coast of Mississippi. – Annals of the Entomological So- 2013: Trophobiosis in the arboricolous ant Liometopum mic- ciety of America 64: 562-565. rocephalum (Hymenoptera: Formicidae: Dolichoderinae). – European Journal of Entomology 110: 231-239. MARKIN, G.P., O’NEAL, J., DILLIER, J.H. & COLLINS, H.L. 1974: Regional variation in the seasonal activity of the imported fire SCHWANDER, T., ROSSET, H. & CHAPUISAT, M. 2005: Division of ant, Solenopsis saevissima richteri. – Environmental Entomo- labour and worker size polymorphism in ant colonies: the logy 3: 446-462. impact of social and genetic factors. – Behavioral Ecology and Sociobiology 59: 215-221. MCGLYNN, T.P. & OWEN, J.P. 2002: Food supplementation alters caste allocation in a natural population of Pheidole flavens, a SEIFERT, B. 2007: Die Ameisen Mittel- und Nordeuropas. – lutra dimorphic leaf-litter dwelling ant. – Insectes Sociaux 49: 8 -14. Verlags- und Vertriebsgesellschaft, Görlitz / Tauer, 368 pp.

MCGLYNN, T.P., DIAMOND, S.E. & DUNN, R.R. 2012: Tradeoffs SHATTUCK, S.O. 1992: Generic revision of the Ant Subfamily in the Evolution of Caste and Body Size in the Hyperdiverse Dolichoderinae (Hymenoptera: Formicidae). – Sociobiology 21 Ant Genus Pheidole. – PLoS ONE 7(10): e48202. (1): 1-176.

MERTL, A.L. & TRANIELLO, J.F.A. 2009: Behavioral evolution in STITZ, H. 1939: Hautflügler oder Hymenoptera I: Ameisen oder the major worker subcaste of twig-nesting Pheidole Formicidae. In: DAHL, F., DAHL, M. & BISCHOF, H. (Eds.): Die (Hymenoptera: Formicidae): does morphological specialization Tierwelt Deutschlands und der angrenzenden Meeresteile nach influence task plasticity? – Behavioral Ecology and ihren Merkmalen und nach ihrer Lebensweise 37. – Gustav Sociobiology 63: 1411-1426. Fischer, Jena, 428 pp.

NIJHOUT, H.F. & WHEELER, D.E. 1996: Growth models of com- TRABALON, M., PLATEAUX, L., PÉRU, L., BAGNÈRES, A.G. & plex allometries in holometabolous insects. – The American HARTMANN, N. 2000: Modification of morphological chara- Naturalist 148: 40-56. cters and cuticular compounds in worker ants Leptothorax nylanderi induced by endoparasites Anomotaenia brevis. – OSTER, G.F. & WILSON, E.O. 1978: Caste and ecology in the Journal of Insect Physiology 46 (2): 169-178. social insects. – Princeton University Press, Princeton, N. J., 355 pp. TARTALLY, A. 2006: Long term expansion of a supercolony of the invasive garden ant, Lasius neglectus (Hymenoptera: For- PASSERA, L., RONCIN, E., KAUFMANN, B. & KELLER, L. 1996: micidae). – Myrmecologische Nachrichten 9: 21-25. Increased soldier production in ant colonies exposed to intraspecific competition. – Nature 379: 630-631. TSCHINKEL, W.R. 1988: Colony growth and the ontogeny of worker polymorphism in the fire ant, Solenopsis invicta. – PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. 2011: Interactions Behavioral Ecology and Sociobiology 22(2): 103-115. between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence. – Community TSCHINKEL, W.R., ADAMS, E.S. & MACOM, T. 1995: Territory Ecology 12: 9-17. area and colony size in the fire ant Solenopsis invicta. – Journal of Animal Ecology 64: 473-480. PFEIFFER, M. & LINSENMAIR, K.E. 2001: Territoriality in the Malaysian giant ant Camponotus gigas (Hymenoptera/ TSCHINKEL, W.R., MIKHEYEV, A.S. & STORZ, S. 2003: Allometry Formicidae). – Journal of Ethology 19: 75-85. of workers of the fire ant, Solenopsis invicta. – Journal of Insect Science 3:2, 11 pp. PINHEIRO, J. C. & BATES, D. M. 2000: Mixed-Effects Models in S and S-PLUS. - Statistics and Computing Series, Springer- WHEELER, D.E. & NIJHOUT, H.F. 1984: Solder determination in Verlag, New York, NY, 530 pp. the ant Pheidole bicarinata: inhibition by adult soldiers. – Journal of Insect Physiology 30: 127-135. PORTER, S.D. & TSCHINKEL, W.R. 1985a: Fire ant polymorphism (Hymenoptera: Formicidae): Factors affecting worker size. – WHEELER, D.E. 1991: The developmental basis of worker caste Annals of the Entomological Society of America 78(3): 381- polymorphism in ants. – The American Naturalist 138: 1218- 386. 1238.

PORTER, S.D. & TSCHINKEL, W.R. 1985b: Fire ant polymorphism: WHEELER, W.M. 1928: Mermis parasitism and intercastes among the ergonomics of brood production. – Behavioral Ecology and ants. – Journal of Experimental Zoology 50: 165–237. Sociobiology 16 (4): 323-336. WILSON, E.O. 1953: The origin and evolution of polymorphism POWELL, S. 2009: How ecology shapes caste evolution: linking in ants. – Quarterly Review of Biology 28: 136-156. resource use, morphology, performance and fitness in a superorganism. – Journal of Evolutionary Biology 22: 1004- WILSON, E.O. 1980: Caste and division of labour in leaf-cutter 1013. ants (Hymenoptera: Formicidae: Atta), I. The overall pattern in A. sexdens. – Behavioral Ecology and Sociobiology 7: 143- R DEVELOPMENT CORE TEAM 2011: A Language and 156. Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. – , WOOD, L.A. & TSCHINKEL, W.R. 1981: Quantification and modi- retrieved on 9 August 2011 fication of worker size variation in the fire ant Solenopsis invicta. – Insectes Sociaux 28: 117-128. ROBINSON, E.J.H. 2009: Physiology as a caste-defining feature. – Insectes Sociaux 56: 1-6.

111

112

Appendix B1:

Territories of the studied colonies. In all figures, circles represent individual trees, either serving as nest or foraging trees. Lines represent ant trails; in the light red areas ant workers dispersed, moving away from the trails. One square on the grid corresponds to 1m2.

Figure B1-1: The territory of the colony L1 (Lednice Park site); a) spring (21 April 2011), b) summer (12 July 2011).

Figure B1-2: The territory of the colony L2 (Lednice Park site); a) spring (21 April 2011), b) summer (12 July 2011).

Figure B1-3: The territory of the colony L3 (Lednice Park site); a) spring (6 May 2011), b) summer (1 August 2011).

Figure B1-4: The territory of the colony L4 (Lednice Park site); a) spring (6 May 2011), b) summer (1 August 2011).

Figure B1-5: The territory of the colony L5 (Lednice Park site); a) spring (21 April 2011), b) summer (1 August 2011).

Figure B1-6: The territory of the colony L6 (Lednice Park site); a) spring (6 May 2011), b) summer (12 July 2011).

Figure B1-7: The territory of the colony L7 (Lednice Park site); a) spring (28 May 2011), b) summer (12 July 2011).

Figure B1-8: The territory of the colony R1 (Rendezvous forest site); a) spring (17 May 2011), b) summer (29 July 2011); red = Liometopum microcephalum, yellow = Lasius fuliginosus.

Figure B1-9: The territory of the colony R2 (Rendezvous forest site); a) spring (17 May 2011), b) summer (29 July 2011); red = Liometopum microcephalum, yellow = Lasius fuliginosus, black/grey = Formica rufa.

Figure B1-10: The territory of the colony R3 (Rendezvous forest site); a) spring (17 May 2011), b) summer (29 July 2011).

Figure B1-11: The territory of the colony R4 (Rendezvous forest site); a) spring (17 May 2011), b) summer (29 July 2011).

Figure B1-12: The territory of the colony Z1 (Lednice forest site); a) spring (28 May 2011), b) summer (1 August 2011).

Figure B1-13: The territory of the colony M1 (Bulhary game enclosure forest); summer (11 July 2011).

Figure B1-14: The territory of the colony M2 (Bulhary game enclosure forest); summer (11 July 2011).

Figure B1-15: The territory of the colony P (forest site near Pohansko); spring (21 May 2012).

Appendix B2:

Allometric relations between measured characters (HW = head width, HL = head length, TL = tibia length) and femur length (FL). See the paper for details.

Figure B2-1: Allometric relations between measured characters in the colony L1 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-2: Allometric relations between measured characters in the colony L2 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line.

Figure B2-3: Allometric relations between measured characters in the colony L3 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-4: Allometric relations between measured characters in the colony L4 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-5: Allometric relations between measured characters in the colony L5 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-6: Allometric relations between measured characters in the colony L6 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-7: Allometric relations between measured characters in the colony L7 sampled at the Lednice Park site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-8: Allometric relations between measured characters in the colonies M1 and M2 sampled at the Bulhary game enclosure forest site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-9: Allometric relations between measured characters in the colony P1 sampled at the forest site near Pohansko (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-10: Allometric relations between measured characters in the colony R1 sampled at the Rendezvous forest site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-11: Allometric relations between measured characters in the colony R2 sampled at Rendezvous forest site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-12: Allometric relations between measured characters in the colony R3 sampled at Rendezvous forest site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-13: Allometric relations between measured characters in the colony R4 sampled at Rendezvous forest site (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

Figure B2-14: Allometric relations between measured characters in the colony Z1 sampled at the forest site near Lednice (HW = head width, HL = head length, TL = tibia length, FL = femur length); k represents the slope of the regression line (k1 and k2 correspond to minor and major workers, respectively).

6. Study C: Trophobiosis

Ants generally need both types of nutrients, that is proteins and carbohydrates (VEPSÄLAINEN & SAVOLAINEN 1990, HÖLLBODLER & WILSON 1990). Aphid tending by L. microcephalum workers had been proposed by MAYR in 1855, whereas other contemporary myrmecologists, EMERY (1891) and FOREL (1892), considered the species a strict predator. Much later, WIEST (1967) and MAKAREVICH (2003) observed both predatory behaviour and trophobiosis with aphids of the genus Stomaphis in L. microcephalum, HALMÁGYI (1968) published that L. microcephalum workers tend Lachnus longipes. In South Moravia, Lachnus roboris and Stomaphis quercus was observed in nest and foraging trees of L. microcephalum (OMELKOVÁ 2003). Our aim was to confirm that the species does visit its foraging trees for the purpose of aphid tending. If so, we wanted to see if there were any seasonal differences in honeydew collecting. Honeydew intake is detectable by the presence of reducing sugars in ant gasters. Honeydew consists mainly of non-reducing sugars (trehalose, melezitose, saccharose) whereas reducing sugars (glucose, maltose) form a smaller part of honeydew. Trehalose, however, is an important component of insect haemolymph and thus it is not suitable for honeydew detection. In contrast, the reducing sugars are rare in the haemolymph and thus allow reliable detection of honeydew intake. Workers ascending and descending nest and foraging trees were sampled at Rendezvous site once per month (from April to July 2009) and stored immediately in carbon-dioxide ice. In total, 320 workers were analysed (80 workers per sampling date: 20 workers from each category - descending and ascending a nest tree, ascending and descending a foraging tree). Each ant gaster was weighted and the amount of total and reducing sugars were measured by photometric methods (see the paper for details). Relationships among the mean amount of sugars in an ant gaster and the type of tree, direction of movement and date were modelled using Generalized Linear Models (GLM). Gaster masses of “satiated” workers, i.e. those going down from foraging trees and up their nest trees, were higher (p < 0.05) than in workers moving in the opposite direction (descending nest trees and ascending foraging trees). Workers descending foraging trees had the highest amount of total and reducing sugars. However, those workers carried higher amounts of sugars than workers ascending the nest trees. Perhaps some of the sampled workers ascending nest trees were not returning from foraging trees, possibly they were guarding their nest tree. Another explanation would be the occurrence of trophallaxis: The workers returning from foraging trees back to their nest might have passed a portion of the collected honeydew to other workers moving apart from their nest tree. Although total sugar contents in ant gasters were high in April, reducing sugar contents were lower than in May. An ant colony needs sugars as a rapid energy source, in particular in early spring. Aphid numbers were apparently low in the beginning of the vegetation season and thus the ants probably took in sugars from other sources, for example from floral nectaries. Aphid tending may also take place on the nest trees, including the large crowns of these trees. The lowest amount of sugars found in July may have been related to a high production of new workers in the colonies, which were thus requiring protein-rich nutrition (VEPSÄLAINEN & SAVOLAINEN 1990). Another explanation is a decrease in aphid population

densities, which usually happens in mid and late summer with respect to their life cycles (JAROŠÍK & DIXON 1999). We confirmed that trophobiosis is important for the nutrition in L. microcephalum. It presents a major reason for which its workers visit their foraging trees. In the near future, I would like to apply my experience with prey DNA detection in predator guts to determine the intensity of predation in L. microcephalum and their potential preference for certain invertebrate prey types.

Eur. J. Entomol. 110(2): 231–239, 2013 http://www.eje.cz/pdfs/110/2/231 ISSN 1210-5759 (print), 1802-8829 (online)

Trophobiosis in the arboricolous ant Liometopum microcephalum (Hymenoptera: Formicidae: Dolichoderinae)

JIŘÍ SCHLAGHAMERSKÝ1, JAN KAŠPAR1, LENKA PETRÁKOVÁ1 and VLADIMÍR ŠUSTR2

1 Masaryk University, Faculty of Science, Department of Botany and Zoology, Kotlářská 2, 611 37 Brno, Czech Republic; e-mail: [email protected] 2 Biology Centre of the ASCR, Institute of Soil Biology, Na Sádkách 7, 370 05 České Budějovice, Czech Republic; e-mail: [email protected]

Key words. Hymenoptera, Formicidae, Dolichoderinae, Liometopum, arboricolous, ants, trophobiosis, foraging, honeydew

Abstract. The arboricolous dolichoderine ant Liometopum microcephalum (Panzer, 1798) is considered to be mainly predatory, although there are some reports of it tending aphids. The main objective of the present study was to confirm that this ant has a trophobiotic relationship with aphids and assess seasonal differences in its utilization of honeydew. We hypothesized that the worker ants on trees where they have their nest (nest tree) and trees where they are foraging (foraging trees) should differ in gaster mass and sugar content depending on their direction of movement, and that both should be highest in spring. From spring to summer 2009, ascending and descending workers were collected from nest and foraging trees at a locality in South Moravia, Czech Republic. Mass of their gasters and their content of total and reducing sugars were measured using chemical (photometric) methods. Differences in gaster mass confirmed the flow of liquid food from foraging to nest trees, but there were no significant between-month differences. Contents of total and reducing sugars were positively correlated with gaster mass. The gasters of workers descending from foraging trees contained significantly more reducing sugars than those of workers descending or ascending nest trees. The content of reducing sugars was lowest at the beginning of the ants’ activity period in April and highest in June, with a non-significant drop in July. Results for total sugars were similar, with the decrease in July being significant. The concentration of sugars in the gasters of workers ascending and descending nest trees did not differ significantly but the absolute content of total sugars was higher in the gasters of ascending ants. Results from foraging trees confirmed that the ants collected the honeydew from these trees. Possible reasons for the ambigous results for nest trees are discussed. We conclude that trophobiosis is an important component of the nutritional biology of L. microcephalum.

INTRODUCTION Wiest (1967) reports a territory size of up to 600 m2 and Trophobiosis, the mutualistic relationship between ants Makarevich (2003) a colony with several interconnected and insects (trophobionts, mainly some groups of Hemi- permanent and temporary nests and visiting 23 foraging ptera and Lepidoptera), which produce sugar-rich exu- trees. Due to the large size of Liometopum colonies, their dates that in most cases are termed honeydew, is impor- feeding behaviour has to affect substantially the nutrient tant in the nutritional biology of many species of ants flow and biotic community in their vicinity. Although (Hölldobler & Wilson, 1990). The nutrient flow in terres- Mayr (1885) assumed that the only reason for L. micro- trial ecosystems can be substantially shaped by such rela- cephalum to visit trees other than the one in which it nests tionships, if ants with their trophobionts are a dominant could be aphid tending, the species was considered to be element of the ecosystem: wood ants (Formica rufa mainly a predatory ant, living almost exclusively on group) in boreal forests, for instance, transport large animal food, by renowned myrmecologists: According to quantities of carbon, nitrogen and phosphorus into their Forel (1892) “plant-lice [aphids] (are) heartily despised nests, of which at least 54%, 70% and 57%, respectively, by the Liometopum” (translation by Wheeler, 1905); are collected in the form of honeydew. Emery (1891) held the same opinion based on his own The dolichoderine ant Liometopum microcephalum observations. As already pointed out by Wheeler (1905), (Panzer, 1798) is the only species of this genus living in this would be a substantial difference to North American Europe (Seifert, 2007). It is arboricolous, thermophilous species, which have trophobiotic relationships with and lives in the north of its range predominantly on flood- aphids and coccids. Velasco Corona et al. (2007) report plains (forests and open landscape) where it builds its that Liometopum appiculatum in Tlaxcala, Mexico, has a trophobiotic relationship with 14 species of Sternorrhyn- nests several metres above ground in the trunks of old but 15 mostly still healthy trees, mainly oaks (Schlaghamerský cha. The particularly high  N reported in L. micro- & Omelková, 2007). From the nest trees conspicuous cephalum (albeit based on a sample from a single colony) trails lead to other trees in the vicinity and up into their compared to other ant species would indeed indicate a crowns (“foraging trees”). Colonies can be very large, predominance of zoophagy in this species (Fiedler et al., consisting of probably hundreds of thousands of individu- 2007). However, contrary to the preceding claims of strict als, which guard territories that include many nest trees: zoophagy in L. microcephalum, Wiest (1967) reports

231 from a site in Austria that the trails of L. microcephalum also lead to bark aphids of the genus Stomaphis. She observed L. microcephalum workers tending these aphids and transporting them to their nest in November, presumably for overwintering in the ant nest. Makarevich (2003) reports the tending of Stomaphis longirostris by a small population of L. microcephalum in the region of the Dnieper river (Ukraine). These two papers are the only published primary sources of trophobiosis occurring in L. microcephalum. However, they report only observational data with little detail and no quantification. In South Moravia, we have also observed L. microcephalum Fig. 1. Directions of movement of Liometopum micro- tending Stomaphis quercus (unpubl. observ.). Thus we cephalum and transport of honeydew between nest and foraging reject the claim that Stomaphis quercus has an obligate trees. Movement between nest and crown of the nest tree is assumed (therefore the question mark) as it was not studied. relationship with Lasius fuliginosus (e.g. Goidanich, 1959; Hopkins & Thacker, 1999; Stadler & Dixon, 2008). We found also other aphids potentially utilized by L. in the colony. Honeydew contains not only various kinds microcephalum on its nest and foraging trees, in par- of sugars that can be utilized by ants and is a valuable ticular Lachnus roboris (unpubl. observ.). short-term energy source, but may also contain amino In the canopies of tropical forests, where many more acids, proteins and lipids suitable for ant nutrition (Boevé species of ants are truly arboricolous, tree-dwelling tro- & Wäckers, 2003; Blüthgen et al., 2004). Nevertheless, phobionts are monopolized by the most common species sugars are the main component of honeydew (90–95% by of ants in the canopy, with dominant colonies and species dry weight, Hölldobler & Wilson, 1990) and in the of ants maintaining mutually exclusive territories absence of any other potential source (floral or extrafloral (Blüthgen et al., 2000). Of all the European ants that are nectaries), large quantities in ant gasters could indicate high in the behavioural hierarchy of ant assemblages, L. that the ants, either collect it directly by “milking” the tro- microcephalum is probably the most arboricolous (and phobionts, or from the surface of leaves where it falls much more so than for instance its North American co- after being excreted. Non-reducing sugars such as melezi- geners that frequently nest on the ground); it competes for tose, saccharose and trehalose may comprise up to 90% foraging trees with other dominant ants, such as Lasius of total honeydew sugars but often substantially less fuliginosus and Formica rufa (Petráková & Schlagham- (Fischer et al., 2005). Trehalose is also synthesized and is erský, 2011). It is very questionable, if the large colonies an important “blood-sugar” of insects, as it is major form formed by L. microcephalum could be supported by pre- of energy storage (Hölldobler & Wilson, 1990; Schilman dation alone; thus a more direct link to primary produc- & Roces, 2008). As there is a high concentration of non- tion, via tree sap processed by aphids, seems probable, reducing trehalose in the haemolymph of well-fed ant which would have implications for energy flow in ecosys- workers (Schilman & Roces, 2008), this might reflect an tems where L. microcephalum occurs. overall nutritional status that is independent of the energy The objectives of the present study were to confirm that source (lipids, proteins or sugars). Reducing sugars are trophobiosis occurs in L. microcephalum and assess its rare in haemolymph and are therefore a good indicator of importance for this species. In particular, we wanted to honeydew uptake, regardless of their low proportion in confirm that foraging trees are visited (also) to collect honeydew [nevertheless sometimes exceeding 50% of all honeydew and assess whether there are seasonal differ- sugars, see for instance Fischer & Schingleton (2001) or ences in the use of trophobiosis. Envisaging a situation in Blüthgen et al. (2004)]. Total and reducing sugar concen- which workers with empty gasters leave nests to collect trations can be readily measured using chemical (pho- honeydew on foraging trees and then return to the nest tometric) methods, allowing the evaluation of nutritional trees with full gasters (Fig. 1), we formulated the fol- status and indication of honeydew utilization by L. micro- lowing hypotheses: (1) gasters of workers descending cephalum in a perhaps old-fashioned but efficient way. from a nest tree should contain less sugars, in particular MATERIAL AND METHODS reducing sugars, than those of workers ascending these trees; (2) gasters of workers ascending a foraging tree Ant collection should contain less (reducing) sugars than those of This study was conducted in Southern Moravia (south-eastern workers descending from these trees; (3) sugar content of Czech Republic), which is on the north-western margin of this Liometopum gasters should be highest in spring, when the species’ range but where there is a very large population of L. microcephalum (Schlaghamerský & Omelková, 2007). Ants colony needs “fast energy”; (4) honeydew is the major were collected in an old forest stand (Rendezvous National liquid transported in the crop (located within the gaster) Nature Monument) between the towns of Valtice and Břeclav in and thus gaster mass will change with sugar content. Southern Moravia, Czech Republic (48°44´52˝N, 16°47´33˝E), Liquid food collected by ant workers is stored in the where 42 colonies (nest trees) were recorded in a preceding crop situated in the gaster and is used by the individual study, which is the third largest subpopulation in the South ants as well as for the feeding of other adults and larvae Moravian population (Omelková, unpubl. data; see also Schlaghamerský & Omelková, 2007). In 2009, workers of L.

232 microcephalum ascending and descending nest and foraging the glycogen also present in the homogenized sample (Olson et trees were collected once per month on April 23, May 20, June al., 2000). The second step was centrifuging (12,000 g / 3 min) 24 and July 22. This includes the species’ main activity period the sample, which separated the sedimented glycogen from the as in August the activity of workers declined rapidly and for- soluble sugars in the supernatant. The supernatant was divided aging trees were hardly visited anymore. On each of the above between two microcentrifuge tubes for subsequent analyses of dates, the ants were collected from ca 9:00 to 11:30 and 13:00 to reducing sugars (600 µl) and total sugars (100 µl). These 16:30 (collecting was time consuming and we wanted to avoid analyses were done separately for individual gasters of the ants. the period of low activity around noon). Workers ascending and Total sugars were assessed using the hot-anthrone test (Olson descending were collected at a height of 1.5 m above the ground et al., 2000). The tube with 100 µl of supernatant was placed in from trunks of nest and foraging trees (both oak). Whereas nest a hot water bath (90°C) and the solution evaporated to 50 µl. trees, which were previously marked, can be also distinguished After evaporation the sample was cooled in cold water and 950 from foraging trees by their greater size and higher abundance µl of anthrone reagent (120 mg of anthrone in 100 ml 20% of Liometopum workers, it was in many cases not possible to H2SO4) was added. The solution was well mixed and again establish links between individual nests and foraging trees placed in the hot water bath (90°C) for 15 min. After cooling, mainly due to the dense undergrowth. In particular, workers the absorbance at 625 nm was recorded (Synergy 2 microplate ascending nest trees could have visited one of several foraging reader, BioTek Instruments, Inc.). trees monopolized by a particular colony. Several of the colo- Reducing sugars were assessed using the Somogyi-Nelson nies and their foraging trees were adjacent to each other but it method (Nelson, 1944). A microcentrifuge tube containing 600 was not always clear which foraging trees were visited by which µl of sample was placed in a hot water bath (90°C) and the solu- colony. Therefore, we did not use a paired sampling design, tion evaporated to 50 µl. 150 µl of distilled water was added to although we collected ants from foraging trees close to the nest the tube. To remove proteins we added 100 µl 0.3 N Ba(OH)2 trees selected. Even in the course of a single year L. micro- and 100 µl 5% ZnSO4. Deproteination took 10 min, after which cephalum colonies were observed to change foraging trees or the sample was centrifuged (6,000 g / 7 min). 200 µl of the the intensity with which individual trees were visited. For this sample and 200 µl of Somogyi-Nelson reagent were poured into reason, we randomly selected nest trees and nearby foraging a new micro-centrifuge tube and this was placed in the hot water trees currently well-visited by L. microcephalum from which to bath (90°C) for 15 min. After cooling, 200 µl of arsen- collect ants on each sampling date. Distances between nest trees molybden reagent was added, the solution was mixed and the at the study site ranged between 12 and 25 m, and that between absorbance at 670 nm recorded. nest and foraging trees of the same colony between 2 and 10 m. Photometric measurements yielded the concentrations of the From each tree 5 ascending and 5 descending workers were different sugars. By multiplying by gaster mass the absolute collected from the bark of the trunk with entomological forceps. sugar content of each gaster was obtained. The workers were placed in plastic tubes (pooled separately for Data analysis each combination of type of tree and direction of movement), which were immediately placed in an ice box with carbon- To model the relationships between the mean amount of sugar dioxide ice (–80°C) in order to prevent the ants from excreting and type of tree, direction of movement and month, we used or regurgitating their crop contents when stressed. Back in the Generalized Linear Models (GLM) in R 2.10.1 software (R laboratory, the tubes with the ants were stored in a freezer at Development Core Team, 2011). As the data did not fulfil the –20°C for up to four months. Ten nest trees and the same requirements for homogeneity of variances and normality of dis- number of foraging trees, randomly chosen anew on each sam- tribution, we corrected for gamma and log-normal distribution pling date, were sampled on each occasion. Thus 50 ascending (family = Gamma) using inverse or identity link. We modelled and 50 descending workers were collected on each date sampled three response variables: gaster mass, reducing sugar content from each type of tree (nest vs foraging tree), which resulted in and total sugar content. We used one model with sugar contents 200 specimens being collected on each date sampled and a total represented by concentrations for individual ant gasters (µg/mg of 800 for both types of trees on the four dates sampled. From fresh weight of gaster) and one model with the absolute values these, 20 workers for each combination of type of tree × direc- (µg/gaster). We used, “type of tree × direction” (nest tree up, tion of movement × date sampled were randomly selected for nest tree down, foraging tree up, foraging tree down) and analysis (80 specimens per date, 320 in total). “month” (IV–VII) as categorical explanatory variables. In the case of sugar concentrations we used “gaster mass” as another Analysis of gaster mass and sugar content explanatory variable (possible relationship between sugar con- The ants were taken out of the freezer and as soon as they centration and worker size). We also checked models using became supple, 20 specimens from a given tube were taken (see “direction” (up, down) and “type of tree” as separate variables, above) and the gaster of each ant cut off and placed on filter but – in agreement with our hypotheses – “direction” as such paper to absorb the water that condensed on its surface and sub- had no effect in any model (e.g., ants leaving the nest tree to sequently weighed on an analytical balance (Sartorius R 160P, forage and returning from foraging trees had the same precision to 0.02 mg). Gaster mass was recorded as fresh weight “direction” but on different types of trees). We always started (fw). We did not observe any loss of liquid when the gasters the modelling with the most complex models, that is, we tried to were cut off, as they were still partially frozen. determine the effects of all the explanatory variables and their As the composition of the honeydew collected by the ants and interactions and fitted models by backward selection. We used its potential variability in time were unknown, we analyzed both F-tests to test the significance of individual explanatory vari- the concentrations of total and reducing sugars in the ant ables and Akaike Information Criteria (AIC) to test the suit- gasters. Although reducing sugars were assumed to be a better ability of the model by comparing the AICs of several models, indicator of trophobiosis, see above, we were not sure if their the smaller the AIC the better the model, and diagnostic plots content would be sufficient to obtain meaningful results. Each with analyses of residuals and Cook’s distances to select the cut off and weighed gaster was homogenized in 1 ml of best model. The differences between values (“levels” in R) of chloroform-methanol (1 : 2). The solution of the sugars in response variables were assessed within the GLMs by t-tests, chloroform-methanol was the first step in separating them from using a treatment contrast matrix. In the models for total and

233 reducing sugars, one and four outliers (exceeding the other values two to three times), respectively, were excluded because we were not sure if these values were not the result of measure- ment errors. RESULTS We selected the most suitable model for each dependent variable based on the significance of the explanatory variables; for detailed characteristics see Table 1. In case of gaster mass only the explanatory variable “type of tree × direction” had a significant effect (Table 1). Pooled over all sampling dates, workers descending foraging trees had significantly greater gaster masses than those descending nest trees or ascending foraging trees (p < 0.05). Gasters of workers ascending nest trees had the greatest masses and differed significantly (p < 0.05) from Fig. 2. Differences in gaster mass of L. microcephalum those of the other three groups of workers (Fig. 2). workers associated with the direction of movement (up and Values for the different dates (months), showed a similar down) on nest and foraging trees (all dates sampled); predicted trend (Fig. 3). Between-month differences in gaster mass values [model formula: GLM (Total sugar concentration ~ type were not significant. of tree × direction, family = Gamma)]: mean values with 95% In the case of the total sugar concentrations in ant gas- confidence intervals; values with different letters are significantly different (t-test, p < 0.05 at least, details see text and ters, all explanatory variables had an important effect Table 1); n = 320. (Table 1). The mean value for the whole sampling season was significantly higher for workers descending foraging trees than those descending nest trees and also those By-far the lowest concentration of reducing sugars was ascending nest trees (p < 0.05 in both cases). The lowest recorded in April and the highest in June (difference concentration of total sugars was recorded in July (Fig. 4, between these months was significant at p < 0.05; Fig. 6). Fig. 5) and differed significantly (p < 0.05) from the high The absolute contents of total sugars and reducing concentration found in April and June. The concentration sugars in ant gasters were only associated with the of total sugars was positively related to worker gaster explanatory variable “type of tree × direction”. There mass (regression coefficient 3.72, p = 0.02). were no significant seasonal differences (Table 1). In the Also in the case of the concentrations of reducing sug- case of total sugars, their mean content was significantly ars, all explanatory variables had significant or, in the greater in workers descending than those ascending for- case of “month”, marginally significant effects (Table 1). aging trees (p < 0.05; Fig. 8) and also greater than in The mean value was again significantly higher (p < 0.05) those descending from nest trees (p < 0.0001). The mean for workers descending foraging trees than those either content of total sugars was also greater in workers descending or ascending nest trees (Fig. 6). The same ascending than descending from nest trees (p < 0.05). The trend was apparent for each of the dates (Fig. 7). The con- mean content of reducing sugars was greater in workers centration of reducing sugars was also correlated with descending foraging trees than in those descending from gaster mass (regression coefficient 5.18, p = 0.0367). nest trees (p < 0.0001), ascending foraging trees (p <

TABLE 1. Outputs of the best models obtained by the method of GLM, giving model formula (in R), Akaike information criterion (AIC), degrees of freedom (Df), explained variablity as decimal fraction of one (R2), explanatory variables, F statistics and respec- tive p-values. Model formula AIC Df R2 Predictor variable F-test p-value GLM [Gaster mass ~ typeoftree × direction, 534.42 5 0.1016 typeoftree × direction 11.39 < 0.0001 family = Gamma (link = identity)] typeoftree × direction 5.31 0.0014 GLM (Total sugar concentration ~ typeoftree × direction 2752.6 9 0.0906 month 3.8 0.0106 + month + gaster mass, family = Gamma) gaster mass 5.47 0.02 GLM (Total sugar amount ~ 3287.5 5 0.0573 typeoftree × direction 5.97 0.0006 typeoftree × direction, family = Gamma) typeoftree × direction 7.17 0.0001 GLM (Reducing sugar concentration ~ typeoftree × direction 2443.7 9 0.0862 gaster mass 4.41 0.0367 + gaster mass + month, family = Gamma) month 2.53 0.0571 GLM (Reducing sugar amount ~ 2889.1 3 0.0576 typeoftree × direction 7.69 < 0.0001 typeoftree × direction, family = Gamma)

234 Fig. 3. Gaster mass of L. microcephalum workers associated with the direction of movement on nest and foraging trees on four dates in the period April–July (IV–VII); measured values: medians (bars) with 95% confidence limits (notches), 25–75% quantils (boxes), 1.5 interquartile ranges (whiskers) and outliers (circles); n = 320.

0.05) and even than those ascending nest trees (p < 0.05; Fig. 9). DISCUSSION The observed differences in gaster mass support our hypothesis that liquid food is transported from foraging to nest trees. Differences between the gaster mass of workers were most pronounced in April and May between workers ascending and descending nest trees (Fig. 3). The difference in sugar content on foraging trees, where ascending workers contained less than descending workers, seems logical, because a worker ascending a foraging tree should have an empty crop and be hungry, thus contain less total sugars and in particular less reducing sugars than a worker leaving a foraging tree, presumably satiated after foraging. The observed pattern is in accordance with our expectation that differences between the ants’ directions of movement would be more reflected in Fig. 4. Differences in total sugar concentration in the gasters reducing than in total sugars. The comparison of sugar of L. microcephalum workers associated with the direction of movement (up and down) on nest and foraging trees (lower concentrations in workers descending and ascending the x-axis), sampling dates and gaster mass; predicted values nest tree did not confirm higher concentrations in gasters (model formula: GLM (Total sugar concentration ~ type of tree of in-coming than in those leaving nest trees. In contrast, × direction + month + gaster mass, family = Gamma), details the concentration of reducing sugars in ants ascending see Table 1): mean values with 95% confidence intervals; – seemed to be somewhat lower (not statistically signi- April, – May, – June, – July; regression line shows the ficant) than those descending from nest trees. The same relation between total sugar content and gaster mass (regression was true for the absolute content of reducing sugars. coefficient = 3.72), scale for gaster mass on upper x-axis applies.

235 Fig. 5. Total sugar concentration in the gasters of L. microcephalum workers associated with the direction of movement on nest and foraging trees on four dates in the period April–July (iv-vii); measured values: medians (bars) with 95% confidence limits (notches), 25–75% quantils (boxes), 1.5 interquartile ranges (whiskers) and outliers (circles); n = 320.

However, the absolute content of total sugars per gaster was significantly higher in workers ascending than descending from nest trees. The less convincing results obtained for nest trees might have had several reasons. Some of the returning foragers might not have collected any honeydew. There are many more workers moving around on the trunks of nest than of foraging trees, presumably doing different tasks, including guarding the territory (Petráková & Schlaghamerský, 2011), and the correct sampling of workers returning from foraging and those leaving to forage is more difficult than on foraging trees. Several outlying values of a high sugar content and the connected non-homogeneous variance support this notion. Some transfer of liquid food (trophallaxis) from in-coming workers to those on the lower part of the trunk of nest trees might take place. Furthermore, although the files of workers ascending to the crowns of trees indicate that a substantial portion of the foraging occurs there, we Fig. 6. Differences in the concentration of reducing sugars in have observed L. microcephalum tending Stomaphis the gasters of L. microcephalum workers associated with the quercus in bark fissures even on the lowest parts of oak direction of movement (up and down) on nest and foraging trees trunks, including those of nest trees. On the other hand, (lower x-axis), dates sampled and gaster mass; predicted values (model formula: GLM (Reducing sugar concentration ~ type of some of the ants collected descending nest trees might tree × direction + gaster mass + month, family = Gamma), have tended aphids further up the nest tree and by-passed details see Table 1): mean values with 95% confidence the nest. Due to the hidden position of a colony’s nest intervals; – April, – May, – June, – July; regression (often in the trunk several metres above ground) the flow line shows the relation between total sugar content and gaster of food towards the nest is difficult to study on the nest mass (regression coefficient = 5.18), scale for gaster mass on upper x-axis applies.

236 Fig. 7. Content of reducing sugars in the gasters of L. microcephalum workers associated with the direction of movement on nest and foraging trees on four dates in the period April–July (IV–VII); measured values: medians (bars) with 95% confidence limits (notches), 25–75% quantils (boxes), 1.5 interquartile ranges (whiskers) and outliers (circles); n = 320. tree. One of the reasons for L. microcephalum preferring to nest in live trees (Schlaghamerský & Omelková, 2007) might be the importance of a sufficient source of honeydew in the close vicinity of the nest because foraging on distant trees carries substantial additional costs, particularly for those colonies in solitary trees growing in meadows and similar habitats. The concentration of reducing sugars in ant gasters was lower in April than in May and June. The concentration of total sugars was somewhat lower in May for unknown reasons, but as explained above reducing sugars should be a better indicator of recent honeydew uptake. Thus our hypothesis that honeydew would be most important in spring was not confirmed. We assume this was due to the low population densities of aphids at the beginning of the observation period. We have no data on aphid dynamics in this area but substantial increases in the abundance of tree-dwelling aphids from April to May are recorded for Fig. 8. Differences in the absolute content of total sugars in temperate regions, for instance, by Sequeira & Dixon gasters of L. microcephalum workers associated with the direc- (1997), Jarošík & Dixon (1999), Molnár (2003) and tion of movement (up and down) on nest and foraging trees (all Durak (2008). However, we have no explanation for the sampling dates); predicted values [model formula: GLM (Total high gaster mass and low concentration of sugars in sugar ~ type of tree × direction, family = Gamma)]: mean workers returning to nest trees in April compared to the values with 95% confidence intervals; values not sharing the other groups of workers (Figs 3, 5, 7). Transport of water same letters are significantly different (t-test, p < 0.05 at least, in the crop would seem more probable during the summer details see text and Table 1); n = 320.

237 in the temperate zone but has behavioural traits similar to those of many arboricolous ant species in the tropics (agressiveness, territoriality, numerical dominance). These comparisons have shown that canopy dwelling ants are nitrogen-limited and epigeic species carbohydrate- limited and have different food preferences (protein-rich prey in the former vs. honeydew in the latter). However, this does not mean that the former group does not utilize trophobiosis and the latter predation, but that the scarcer food source is preferred when available. This is in line with the many reports about the predacious behaviour of L. microcephalum (see above) and does not contradict our present findings. ACKNOWLEDGEMENTS. The study was supported by the Ministry of Education, Youth and Sports of the Czech Republic, Fig. 9. Differences in the absolute content of reducing sugars Research Plan No. MSM0021622416. L. Petráková received in gasters of L. microcephalum workers associated with the additional funding from the Czech Science Foundation (grant direction of movement (up and down) on nest and foraging trees 526/09/H050). The Czech Ministry of the Environment granted (all sampling dates); predicted values [model formula: GLM access to the Rendezvous National Nature Monument and (Total sugar ~ type of tree × direction, family = Gamma)]: mean allowed us to collect specimens of L. microcephalum by values with 95% confidence intervals; values with different let- granting us Exception No. 8375/04-620/1377/04. The Master’s ters are significantly different (t-test, p < 0.05 at least, details thesis of M. Omelková (Masaryk University, Brno), provided see text and Table 1); n = 320. initial information on the feeding biology of L. microcephalum and a starting point for the present study. Two anonymous reviewers and J. Dauber (Thünen Institute, Braunschweig) pro- months. Moreover, as shown, at least for Formica vided valuable comments that were incorporated in the final polyctena, water probably passes quickly to the mid-gut version of the manuscript. Further advice on the statistical and is not stored in the crop (Schneider, 1966). Liquid analysis of the results was given by P. Drozd (Ostrava Univer- food other than honeydew could have confounded the sity). M. Barclay (Natural History Museum, London) and results. Possibly some nectar of low sugar content was A.F.G. Dixon kindly checked the English. collected elsewhere in the vicinity in April, but we have no observations that support this assumption. Many aphid REFERENCES populations decrease substantially in mid to late summer BLÜTHGEN N., VERHAAG M., GOITÍA W., JAFFÉ K., MORAWETZ K. (references see above), which might explain the slight & BARTHLOTT W. 2000: How plants shape the ant community decrease in total and reducing sugars (value for reducing in the Amazonian rainforest canopy: the key role of sugars not significantly different from that recorded on extrafloral nectaries and homopteran honeydew. — Oecologia any of the other dates) found in the ants in July (Figs 4, 6) 125: 229–240. and, in particular, the observed very low presence of L. BLÜTHGEN N., GOTTSBERGER G. & FIEDLER K. 2004: Sugar and amino acid composition of ant-attended nectar and honeydew microcephalum on foraging trees in August (not neces- sources from an Australian rainforest. — Austral. Ecol. 29: sarily due to trophobiosis as ants are also predators of 418–429. aphids – see for instance Novgorodova, 2005). Species of BOEVÉ J.-L. & WÄCKERS F.L. 2003: Gustatory perception and trees frequently visited by foraging L. microcephalum are metabolic utilization of sugars by Myrmica rubra ant oaks and poplars, both of which lack nectaries, so hon- workers. — Oecologia 136: 508–514. eydew is the only source of such food on these trees. We DURAK R. & WOJCIECHOWSKI W. 2008: Structure and dynamics conclude that trophobiosis is an important part of for- of aphid communities connected with trees in selected forest aging in L. microcephalum and a major reason why associations. — Pol. J. Entomol. 77: 79–92. workers visit foraging trees. Trees in which the colonies EMERY C. 1891: Zur Biologie der Ameisen. — Biol. Zentralbl. 11: 165–180. build their nests also seem to be important for the colo- FIEDLER K., KUHLMANN F., SCHLICK-STEINER B.C., STEINER F.M. nies’ nutrition, including trophobiosis. We are not yet & GEBAUER G. 2007: Stable N-isotope signatures in central able to give a quantitative assessment of the importance European ants – assessing positions in a trophic gradient. — of honeydew for L. microcephalum in relation to other Insectes Soc. 54: 393–402. types of food. A study of further aspects of foraging by FISCHER M.K. & SCHINGELTON A.W. 2001: Host plant and ants the South Moravian population of L. microcephalum is influence the honeydew sugar composition of aphids. — underway. In the light of the present data, however, we Funct. Ecol. 15: 544–550. assume that predation is not such a dominant part of the FISCHER M.K., VÖLKL W. & HOFFMANN K.H. 2005: Honeydew foraging behaviour of L. microcephalum as has formerly production and honeydew sugar composition of polyphagous been reported. This is also interesting in light of compari- black bean aphid, Aphis fabae (Hemiptera: Aphididae) on various host plants and implications for ant-attendance. — sons between canopy and litter dwelling ants based on Eur. J. Entomol. 102: 155–160. tropical assemblages (e.g. Hahn & Wheeler, 2002), as L. FOREL A. 1892: Die Ameisenfauna Bulgariens. — Verh. K.-K. microcephalum is one of the few truly arboricolous ants Zool.-Bot. Ges. Wien 42: 305–318.

238 GOIDANICH A. 1959: Le migrazioni coatte mirmecogene dello ant species of sympatric occurrence. — Commun. Ecol. 12: Stomaphis quercus Linnaeus, afide olociclico monoico omo- 9–17. topo. — Boll. Ist. Entomol. Univ. Studi Bologna 23: 93–131. R DEVELOPMENT CORE TEAM 2011: A Language and Environment HAHN D.A. & WHEELER D.E. 2002: Seasonal foraging activity for Statistical Computing. R Foundation for Statistical Com- and bait preferences of ants on Barro Colorado Island, Pan- puting, Vienna. URL: http://www.R-project.org ama. — Biotropica 34: 348–356. SCHILMAN P.E. & ROCES F. 2008: Haemolymph sugar levels in a HÖLLDOBLER B. & WILSON E.O. 1990: The Ants. Belknap Press nectar-feeding ant: dependence on metabolic expenditure and of Harvard University Press, Cambridge, MA, 732 pp. carbohydrate deprivation. — J. Comp. Physiol. (B) 178: HOPKINS G.W. & THACKER J.I. 1999: Ants and habitat specificity 157–165. in aphids. — J. Insect Conserv. 3: 25–31. SCHLAGHAMERSKÝ J. & OMELKOVÁ M. 2007: The present distribu- JAROŠÍK V. & DIXON A.F.G. 1999: Population dynamics of a tion and nest tree characteristics of Liometopum micro- tree-dwelling aphid: regulation and density-independent proc- cephalum (Panzer, 1798) (Hymenoptera: Formicidae) in esses. — J. Anim. Ecol. 68: 726–732. South Moravia. — Myrmecol. News 10: 85–90. MAKAREVICH O.N. 2003: Liometopum microcephalum (Hymeno- SCHNEIDER P. 1966: Versuche zur Frage der Futterverteilung und ptera, Formicidae) in the Lower Dnepr. — Vest. Zool. 37(4): Wasseraufnahme bei Formica polyctena (Först.). — Insectes 51–56 [in Ukrainian, Engl. abstr.]. Soc. 13: 297–304. MAYR G. 1856: Beiträge zur ungarischen Formicidenfauna. — SEIFERT B. 2007: Die Ameisen Mittel- und Nordeuropas. Lutra Verh. Zool.-Bot. Ges. Wien 6: 175–178. Verlags und Vertriebsgesellschaft, Görlitz / Tauer, 368 pp. MOLNÁR N. 2003: Population dynamics features of willow- SEQUEIRA R. & DIXON A.F.G. 1997: Population dynamics of tree- feeding aphids. — Acta Phytopathol. Entomol. Hung. 38: dwelling aphids: The importance of seasonality and time 125–135. scale. — Ecology 78: 2603–2610. NELSON N. 1944: A photometric adaptation of the Somogyi STADLER B. & DIXON A.F.G. 2008: Mutualism: Ants and their method for the determination of glucose. — J. Biol. Chem. Insect Partners. Cambridge University Press, New York, 219 153: 375–380. pp. NOVGORODOVA T.A. 2005: Ant-aphid interactions in multispecies VELASCO CORONA C., CORONA-VARGAS M.C. & PEŃA-MARTÍNEZ ant communities: Some ecological and ethological aspects. — R. 2007: Liometopum apiculatum (Formicidae: Dolichoderi- Eur. J. Entomol. 102: 495–501. nae) y su relacion trofobiotica con Hemiptera Sternorrhyncha OLSON D.M., FADAMIRO H., LUNDGREN J.G. & HEIMPEL G.E. en Tlaxco, Tlaxcala, México. — Acta Zool. Mexicana (n.s.) 2000: Effects of sugar feeding on carbohydrate and lipid 23(2): 31–42. metabolism in a parasitoid wasp. — Physiol. Entomol. 25: WHEELER W.M. 1905: The North American ants of the genus 17–26. Liometopum. — Bull. Am. Mus. Nat. Hist. 21: 321–333. PETRÁKOVÁ L. & SCHLAGHAMERSKÝ J. 2011: Interactions between WIEST L. 1967: Zur Biologie der Ameise Liometopum micro- Liometopum microcephalum (Formicidae) and other dominant cephalum Panz. — Wiss. Arb. Bgld 38: 136–144.

Received February 28, 2011; revised and accepted October 23, 2012

239 7. Study D: Genetic diversity of the South Moravian subpopulation

7. 1 Introduction

7.1.1 Colony structure and dispersal in ants The social organization of ants and cooperation between nest mates is usually given by the high relatedness of workers living in the same colony (e.g. HAMILTON 1964, FRUMHOFF & WARD 1992). The workers’ relatedness mostly reflects the number of reproducing queens inside the nest (e.g. BARGUM et al. 2007) ranging from one queen (= monogyny) to several or even many queens (= polygyny). About half of the European ant species are known to be polygynous (HANNONEN 2002). However, the polygyny of many of these species is known to be facultative (FRUMHOFF & WARD 1992, CHAPUISAT et al. 2004). The colony structure depends largely on environmental conditions (ROSENGREN & PAMILO 1983) or is closely tied to colony size (SCHMIDT et al. 2010). According to the habitat saturation hypothesis (e.g. CHAPUISAT et al. 2004) reduced dispersal leads to polygyny (SUNDSTRÖM et al. 2005). When a colony inhabits a site that allows long-range dispersal (i.e. there is no barrier preventing dispersal from that site) it usually tends to be monogynous. In general, polygynous colonies are larger and live much longer than monogynous ones (ROSSET & CHAPUISAT 2007). A special case of colony structure is called polydomy, which is defined as a one colony inhabiting at least two spatially separated nests. Even a monogynous colony can be polydomous. Polydomy has been often observed as a seasonal phenomenon, allowing to reduce costs of long-distance food transport and to decrease the risk of predation. Polydomy often facilitates ecological success, it may even lead to dominance of such colonies. In case of polygynous colonies, polydomy reduces conflicts among reproducing queens at sites with limited dispersal potential (DEBOUT et al. 2007). Dispersal in ants (inhabiting the temperate zone) is limited to the mating period, during which young females usually leave their natal nest and after mating try to found new colonies. Several ways of colony founding have been observed: semi-claustral founding (a young queen leaves the newly founded nest and forages until her brood hatches) was originally considered an ancestral strategy (HASKINS 1941). An increasing number of observations suggests that it is likely a facultative way of colony founding in some species (JOHNSON 2002, BROWN & BONHOEFFER 2003). In claustral colony founding the female does not leave the nest and feeds her brood using her metabolic reserves. A mated female can either found a new colony alone (haplometrosis), or together with several other females (pleometrosis). Collective colony founding increases the probability of survival of the female associations (WILSON 1966). The more females there are present at a site the more often they form associations (TSCHINKEL & HOWARD 1983). After the hatching of workers usually only a single queen survives (primary polygyny; MARKIN et al. 1972, HÖLLDOBLER & WILSON 1990). A mated female can also be accepted by an already existing mature colony (secondary polygyny). Queen adoption is a common phenomenon in some ants (EVANS 1996, SOUZA et al. 2005), although it seems to be more common in their parental nests (STUART et al. 1993). Mated females of some ant species enter a nest of a different species and thus force foreign workers to take care of their (the invader’s) brood. They use various strategies how to befool the original inhabitants of the invaded nests, mostly their pheromones and cuticular hydrocarbons play a key role

(HÖLLDOBLER & WILSON 1990). In contrast to permanent social parasitism, temporary social parasitism is often used as a strategy for colony founding. The budding (or fragmentation) of a part of the colony (when a female with a part of a colony leaves the main nest) is relatively rare in ants (TSCHINKEL & HOWARD 1983), in contrast to honeybees. It is known for example in the Argentine ant Linepithema humile or in Monomorium pharaonis (BUCZKOWSKI & BENNETT 2009). Another case of colony reproduction, observed for example in Ponerinae ants, is the laying of unfertilized male eggs by workers in colonies that have become queenless (D´ETTORRE et al. 2004) or their ability to mate with males and subsequently lay diploid eggs (HEINZE & TSUJI 1995). Generally, the dispersal range of females is shorter than that of males (ZINCK et al. 2007). If the habitat is saturated with nests or if it does not provide a sufficient amount of resources (mostly suitable nest sites - MCGLYNN 2010), colonies tend to produce small numbers of new queens and large numbers of males (PAMILO 1982, DEBOUT et al. 2007), which ensures the gene flow within the population. If there is a barrier preventing long-range dispersal and the site becomes saturated with nests of the same species, the probability of mating between related individuals (even between sexual individuals produced within the same nest) is higher and poses the threat of inbreeding. This may lead to a number of negative effects, such as changes in metabolism, defence and stress responses, decreased energy efficiency and pathogen resistance, reduced worker growth rate and queen longevity (VITIKAINEN et al. 2011), and the production of infertile diploid males (REES et al. 2010). Intranidal mating can be prevented by releasing males and females at different times (PAMILO 1982). Another possible mechanism preventing the negative effect of inbreeding, important mainly in monogynous colonies, is polyandry (= multiple mating by the queen), which increases the genetic diversity of her offspring and is known to occur, at least occasionally, in one third of the known ant species (HUGHES et al. 2008, HOLZER et al. 2009). The way how to definitely confirm polyandry is to genotype the DNA from spermathecae of freshly mated queens (CHAPUISAT 1998). However, it can be rather difficult to obtain freshly mated females, particularly in a species nesting inside large trees, with nest entrances several metres above ground. Moreover, WIEST (1966) proposed that mating in L. microcephalum might take place inside the nest. Populations of a species with specific habitat requirements, living at the margins of their distribution area, are often endangered by habitat fragmentation. Particularly in small populations the genetic structure changes rapidly (SUNDSTRÖM et al. 2005). Dispersal of sexuals over long distances is important for the maintenance of genetic variability within populations and to prevent negative effects of inbreeding. There is a lack of studies that would assess the possible dispersal distance of ant males. Studies on colony founding in Liometopum ants were only based on the observation of females: 2–40 queens were observed to found a single colony in L. luctuosum (pleometrosis), whereas in L. apiculatum a single female establishes a colony (CONCONI et al. 1987). The aim of the present study was (1) to assess the genetic structure of L. microcephalum colonies and (2) to specify gene flow within and among spatially isolated sites in South Moravia. Are the colonies mono- or polygynous? Colonies living in small and isolated habitat islands should have a low intercolonial genetic variability as the mating happens between relatives. Colonies with limited dispersal potential (at spatially isolated sites with a small area of suitable

96 habitat) should be polygynous because of the lack of suitable nest sites and the impossibility of long range dispersal. In a thermophilous species, colony founding at the northern border of its range is much more difficult than in the south because of the need to survive cold winters. The presence of several queens in one nest decreases the risk of colony extinction and increases the longevity of colonies.

7.1.2 Estimating relatedness using microsatellites Invasive methods, such as the destruction of a nest and searching for queens within, can be applied mainly in species that are abundant and form small colonies. Molecular markers allow us to study the colony structure and gene flow without any risk of disturbing colonies of the studied species. The use of those markers is advantageous in particular in species nesting in hardly accessible places. Moreover, in some species, several queens can be present inside the nest but not all of them actually have to reproduce (HANNONEN & SUNDSTRÖM 2003). Microsatellites are tandem repetitions of short nucleotide sequences (a motive usually consists of 2−8 nucleotides) with co-dominant inheritance. These regions are polymorphic in length − one repetition of a single motive corresponds to one allele. They have high mutation rates and thus present suitable tools for studying gene flow on a small geographic scale (GOLDSTEIN & SCHLÖTTERER 1999). Many papers assessing ant colony structure using microsatellites have been published (e.g. GYLLENSTRAND & SEPPÄ 2003, BERNASCONI et al. 2005, DALECKY et al. 2005, ZINCK et al. 2007, FOURNIER et al. 2008). The effective number of queens can be calculated from the relatedness between the workers sampled from a single colony (QUELLER & GOODNIGHT 1989). Unfortunately, the use of those markers has one constraint – due to the high mutation rate it is necessary to select primers that have been designed for closely related species, belonging at least to the same family.

7. 2 Methods

Liometopum microcephalum workers were collected at eight study sites: Křivé jezero, Bulhary game enclosure, Lednice, Kančí obora, Rendezvous, Pohansko, Ranšpurk, and Tvrdonice) in summer 2009 and spring/summer 2010. At each site five colonies were sampled: one colony in the centre of the site and four colonies situated approximately 20 m, 50 m, 100 m and more than 250 m away from the first colony. 25–50 workers were collected from each colony and immediately fixed in pure ethanol. As one colony may inhabit several interconnected nests, aggressiveness and tolerance tests (ASTRUC et al. 2001, SCHLICK-STEINER et al. 2005) were performed when I was not sure if a nest tree belonged to a single or to two distinct colonies. DNA was extracted from six individuals from each colony (representing minor, medium- sized and major workers when possible) using DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer´s protocol. As no markers designed for the studied species nor for any closely related species were available, I used primers that had been originally developed for Formica lugubris (Formicidae: Formicinae) and Azteca ulei var. cordiae (Formicidae: Dolichoderinae). The primers F20, F21 and F29 (CHAPUISAT 1996) had been widely used in related species, resulting in successful crossamplifications (KNADEN et al. 2005, HELANTERÄ & SUNDSTRÖM 2007). The primers Az002, Az004, Az014, Az016, Az022, Az025, Az035, Az042,

Az048, Az064, Az129, and Az171 had been designed by DEBOUT et al. (2006), who had tested the primers with nine species of the genus Azteca. Several optimisation steps of the PCR were performed (TouchDown, HotStart, temperature gradients, changing the concentration of PCR primers, BSA and MgCl2). The best results were obtained when using TouchDown PCR, under the following conditions: 95 °C for 15 min, 10 cycles at 95 °C for 15 s, 1 °C drop per cycle down to a final annealing temperature for 15s, 72 °C for 30 s, followed by 20 cycles at 95 °C for 15 s, annealing temperature (see Table 1 for details) for 30 s, 72 °C for 1 min.; 72 °C for 10 min. The reaction mixture consisted of 0.5 μL of DNA, 0.3 μL of each primer (4 μM), 0.2 μL of 10 mM dNTP´s, 0.8 μL of 25 mM MgCl2, 1 μL of 10x PCR buffer, 0.3 μL of bovine albumine serum (BSA), 0.1 μL of Taq Polymerase (5 U/μL), and 6.5 μL of DNA-free water. PCR products were detected by electrophoresis in 1% GoldView-stained agarose gels. Only five of these microsatellites (Az002, Az016, Az022, Az025, and Az171) were amplified successfully. After optimisation of the annealing temperature, fluorescent-labelled primers were used for PCR amplification. Fragment analysis was carried out on an ABI Prism 3130 Genetic Analyzer (Applied Biosystems). In total, 21 ant workers were genotyped: 15 workers from South Moravia (three workers from each colony, one colony from each of the following sites: Křivé jezero, Bulhary game enclosure, Lednice, Rendezvous, Pohansko), three workers from one colony sampled in central Italy and three workers from one colony sampled in south-eastern Bulgaria. The results were analysed in GeneMapper software.

7. 3 Results and Discussion

I was not able to amplify any DNA fragment using the primers F20-F29, despite trying several optimisations of PCR conditions. The main reason was probably the large evolutionary distance between the two subfamilies, that is Formicinae and Dolichoderinae. Although the amplification was successful in five microsatellite markers (Az002, Az016, Az022, Az025, and Az171), these fragments had extremely low variability (see Table 1) and thus they were not useful for further analyses. Two geographically remote samples (collected near Rome in Italy and in south-eastern Bulgaria) did not differ at all. Again, such a low variability can be explained as the consequence of evolutionary distance of Azteca ulei and Liometopum microcephalum. Azteca ulei, belonging to the tribe Leptomyrmecini, inhabits South America. According to WARD et al. (2010), the common ancestor of the two tribes, Leptomyrmecini and Tapinomini, was dated to approximately 67 million years ago. The fragments that I was able to amplify have been apparently conserved in other species whereas they mutated continuously in Azteca species. Up to 18 alleles were recognized in Azteca ulei var. cordiae.

Table 1: A comparison of size range (length of a repeating microsatellite motif, in base pairs) found and annealing temperatures applicable in Azteca ulei cordiae and in Liometopum microcephalum.

Azteca ulei cordiae Liometopum microcephalum

Locus Tann (ºC) Size range (bp) Nalleles Tann (ºC) Size range (bp) Nalleles

Az002 49.0 124-171 6 45.4 177-178 1

Az004 49.4 203-216 4 41.7 - -

Az014 54.4 218-262 10 - - -

Az016 55.1 326-408 16 44.2 178 1

Az022 49.5 114-198 11 42.9 111-112 1

Az025 54.3 182-275 18 45.4 126 1

Az035 48.4 136-210 14 - - -

Az042 49.6 193-247 18 - - -

Az048 46.2 288-328 12 39.9 - -

Az064 50.4 344-392 14 - - -

Az129 49.4 114-139 8 - - -

Az171 46.2 170-202 8 47.8 115 1

7. 4 Conclusions

The markers used in this study could not be used for any meaningful analysis due to their low variability. The design of more specific primers would be required. However, this would require to sequence the whole genome of several individuals of L. microcephalum, to select regions with a variable number of repetitions, and to test if the markers were applicable. That would have been rather expensive and time consuming, in view of the fact that several other questions remained to be answered during my PhD study. Therefore my supervisor and I decided to abandon this topic and focus on other research questions. One of the questions stated above (regarding the potential polygyny of colonies) was partly resolved within the study of the genetic variation across the entire species range. However, to confirm the exact number of queens inside a colony would require the use of microsatellite markers. As they have faster mutation rates than mitochondrial DNA they could help us to reveal the potential presence of closely related females within a colony. That could result in higher number of queens than those found using mitochondrial markers alone.

100

8. Study E: Phylogeography of the species

Population genetic structure in a species usually reflects its colonisation history. Past colonisation events may be affected by several processes having different impact on the populations, such as geological processes (tectonic and volcanic activity), climate changes (glaciations, sea level changes, aridification), or habitat fragmentation caused by humans (e.g. urbanisation, deforestation, intensive agriculture, river regulation). Populations may become extinct, migrate to more attractive areas or split due to forming geographic barriers that disable gene flow. Considering the scattered occurrence of L. microcephalum, its more or less spatially isolated populations, and the continuing decline in its colony numbers in many areas, we decided to study the genetic structure of individual populations to assess their origin, age, and to reconstruct their migration routes. Molecular markers are increasingly used to assess the relatedness of populations and species (e.g. PILOT et al. 2010, PELLEGRINO et al. 2014). Microsatellites evolve faster than mitochondrial DNA and thus they are more suitable for studies exploring populations on a small geographical range. Nuclear genes evolve much slower than mtDNA and their analysis may not reveal existing divergences among populations of a species. MtDNA is widely distributed in eukaryote genomes, occurring in many copies in every cell. Due to maternal inheritance and the lack of recombination, mitochondrial markers allow the reconstruction of female dispersal. In ants, female dispersal presents the crucial mechanism in colonization of new areas. Three mitochondrial markers (cytochrome c oxidase, cytochrome b and 12S rRNA genes) and one nuclear marker (28S rRNA gene) were used to assess the relatedness of populations across the distribution range of L. microcephalum. In total, 151 specimens collected at 40 sites within 16 countries were processed. Some of the collecting sites were situated relatively close to each other (ca 30 km in the case of two sites in Romania, almost all sampling sites in South Moravia were situated less than 20 km from each other). Some of the sampling sites within Italy, south-eastern Europe and the Middle East were separated by distances exceeding 200 km. This was partially caused by actual fragmentation of the species range, partially by the insufficient exploration of the ant fauna in some areas or our inability to obtain samples from certain regions. Seven subpopulations were delimited based on the geographic distribution of the sampled colonies (Black Sea, Balkan, Italian, Pannonian, Levantine, Western and Eastern Anatolian), taking into account the existence of geographic barriers that could prevent dispersal of the species. I extracted DNA from the specimens, amplified it using PCR, purified and sequenced the DNA fragments. I checked the sequences for potential presence of nuclear pseudogenes (according to BENSASSON et al. 2001, SONG et al. 2008) because of rather high variability of the mitochondrial markers. I constructed haplotype networks, calculated the genetic diversity, performed the analysis of molecular variance, and both mismatched distribution and neutrality tests for each population. Time estimations of the lineage divergence were calibrated based on published dating of fossil representatives of the subfamily Dolichoderinae and their molecular phylogeny (WARD et al. 2010). Sequences of 17 species from eight genera, all representatives of the subfamilies Dolichoderinae and Pseudomyrmecinae, were used for the primary calibration of phylogenetic trees (see the paper for details).

101

The mitochondrial markers were considered as a single locus for the analyses because the used genes were located on the same chromosome and no recombination occurs in the mitochondrial genome. Different genes often have different mutation rates, thus the haplotype networks for each gene fragment are presented separately in Appendix S3. Cytochrome b was the most variable gene from the studied mtDNA fragments, whereas 12S had the lowest nucleotide diversity. The 28S rRNA gene was found of no use for population genetic analyses due to its extremely low variation: A single haplotype occurred in all studied specimens; being identical with that of L. occidentale and L. luctuosum and differing only by one base from L. apiculatum (all three species occur in the western part of North America). In total, 36 mitochondrial haplotypes were distinguished. A high number of haplotypes has also been found in other arthropods: in the ant Myrmica rubra 40 haplotypes exist within Eurasia (LEPPÄNEN et al. 2011), in a trichopteran species, 23 haplotypes were found even in south-eastern Europe only (PAULS 2009). In COI within-species variation is mostly lower than 2%, whereas intraspecific variation exceeds 4%, but many exceptions to this rule have been found (HERBERT et al. 2003). Different haplotypes occurred even within single colonies. More than one mtDNA haplotype found within a single ant colony indicates the presence of several queens in that colony (SUNDSTRÖM et al. 2005). That confirms that that at least some L. microcephalum colonies are polygynous. Based on mtDNA divergences, the samples were divided into two distinct groups: European and Asian (the latter inhabiting the Levant and Western Anatolia). 41% of the total genetic variance were found between the two groups, 39% among populations within groups and 20% within individual populations. The common ancestor of all sampled L. microcephalum colonies lived most probably 3.88 million years ago. Colonies from the Levant differed distinctly from the rest, suggesting these populations evolved separately since the Pliocene. The Levantine colonies were more related to the last common ancestor of all extant L. microcephalum populations than to populations occurring in the western part of its distribution range. That suggest that the species colonized its present range from the east (most probably from Anatolia). This statement is also supported by the fact, that the species is entirely absent in Western Europe (west of the Alps). The phylogenetic analysis divided the sampled colonies into seven clades. The origin of individual European clades is apparently related to the Pleistocene climatic oscillations. The species had dispersed northwards and westwards through the Pannonian Basin, which was assigned as the ancestral area for all European colonies. The Balkan and Northern Clades shared their most recent common ancestor about 850 thousand years ago. The Northern Clade and the Black Sea Clade formed shortly before the Riss glaciation. Their ancestors probably lived in the Pannonian Basin but during the Riss glaciation one part of the ancestral population survived in the south (Balkan refugium), while another had to survive the subsequent glaciation in extra- Mediterranean refugia and evolved separately for a long time. High nucleotide diversity in the Balkan and Pannonian populations suggests a refugial character of those areas. The northernmost area (Czechia, Austria) was apparently colonized from two source populations in different time periods. The dispersal across the Pannonian Basin had happened before the species began to spread from the Apennine Peninsula. As the species prefers building nests in oak trees, I tried to find historical evidence of oak dispersal. However, there is a lack of information about its distribution earlier than 200 thousand

102 years ago. According to VAN ANDEL & TZEDAKIS (1996), most of Europe was covered with temperate deciduous forests (including oaks) about 125,000 years ago. After the cooling of the climate, deciduous forests were restricted to Mediterranean refugia and to the Black Sea coast (BHAGWAT & WILLIS 2008). Oaks were present in the south-eastern part of the present Czech Republic about 35−40 thousand years ago, shortly before the last glacial maximum (LGM), although conifers dominated there (WILLIS & VAN ANDEL 2004). During the LGM that area was covered by permafrost (VANDENBERGHE et al. 2012). Several authors (WILLIS et al. 2000, JUŘIČKOVÁ et al. 2014) proposed the existence of a broadleaf forest glacial refugium in the western part of the Carpathian Basin. Although hazel, alder, willow, pine, ash, beech, and birch prevailed in that refugium (BHAGWAT & WILLIS 2008), oaks were probably present as well (BREWER et al. 2002). Taking into account the fact that nests are also found in other trees than in oaks, L. microcephalum might have been well able to survive in such a refugium. The following manuscript had been reviewed by three referees, and was edited according to their comments. Data analyses have been improved based on the reviewers’ suggestions (BEAST analysis with lineage divergence time estimations, and ancestral area reconstruction were included). The manuscript was partially rewritten to reach a clearer structure of the text. After another revision it has been designated as “requiring minor revision” (Journal of Biogeography, 6th May 2016).

103

104

Phylogeography of the rare velvety tree ant Liometopum microcephalum (Formicidae: Dolichoderinae)

Lenka PETRÁKOVÁ1, Andrea TÓTHOVÁ1 & Jiří SCHLAGHAMERSKÝ1

1 Department of Botany and Zoology, Faculty of Science, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic

ABSTRACT

Aim: The European velvety tree ant (Liometopum microcephalum), forming large colonies defending territories, has specific habitat requirements and a scattered distribution. Its dispersal is slow and colony numbers have been declining in many areas. Our objectives were to assess the origin of populations across the species range and to reconstruct its colonization history. Location: From Italy to Israel in the south, from Czechia to Russia (Lower Volga) in the north. Methods: Three mitochondrial and one nuclear DNA marker were sequenced in 151 specimens collected at 40 sites. We constructed haplotype networks, calculated AMOVA and performed Bayesian evolutionary analysis with lineage divergence time estimations and ancestral area reconstructions. Results: We found 36 mitochondrial haplotypes, the nuclear marker had no variability. More than one haplotype was found within six colonies. Two groups, the European and the Asian, were clearly separated. Six main clades were delimited based on Bayesian evolutionary analysis: the Levantine, Western Anatolian, Black Sea, Northern, Balkan and Western. All populations had low nucleotide and high haplotype diversity values. The species arose 3.88 Ma and the youngest divergences within European clades ocurred 50 ka. The Levantine Clade is the basal one, the ancestral area of European populations is the Pannonian Basin. Main Conclusions: The Levantine population stayed isolated since the Pliocene, the European group reaches back to the beginning of the Quaternary. The European clades diverged during the Pleistocene climate oscillations, the youngest divergencies corresponding to the time before the last glacial maximum. The species survived the glaciations in the Mediterranean (Italian and Balkan Peninsulas) but also in extra-Mediterranean refugia (Carpathian Arc, Black Sea coast). The Italian Peninsula was colonized later than the eastern part of Europe. The north of the Pannonian Basin was colonized from at least two source populations.

Keywords: ancestral area; ant colonies; dispersal; divergence, Pleistocene; refugia

INTRODUCTION

The Eurasian continent has undergone many changes in the past, including climatic and geological ones, which had strong impacts on the population structure of many organisms (Hewitt, 1999). Ectothermic animals responded rapidly to climate changes, shrinking or expanding their distribution areas (Buse et al., 2013). At least during the last glacial maximum (LGM), a large part of western Europe was covered by ice or by permafrost (Vandenberghe et al., 2012) and deciduous 105

forests were restricted to refugia (Collins et al., 2012). Traditionally, three main European refugia were considered in the Mediterranean region: the Balkan, Italian and Iberian peninsulas. Recently, the existence of extra-Mediterranean refugia (e.g. the Carpathian Arc and Dinaric Alps) has been demonstrated (reviewed by Schmitt & Varga, 2012). Such refugia were located near mountain systems, which helped to maintain stable climatic conditions with higher precipitation levels in comparison with the surrounding dry steppes. Retrospectively, glacial refugia of individual species can be traced based on the presence of several distinct haplotypes within a small area (Fijarczyk et al., 2011). Long-term survival in distinct refugia led to genetic differentiation among isolated populations (Hewitt, 1996; Ursenbacher et al., 2006). Fast expansion to recently deglaciated areas and the associated recurrent genetic bottlenecks led to the loss of genetic variability within populations (Hewitt, 1999; Pusch et al., 2006; Horn et al., 2014). However, each species responded individually to the climate changes and thus different colonization routes are distinguished within Europe. Many studies trying to shed light on the origin of populations and reconstructing their migration routes have been published. Among animal groups, mammals, reptiles and amphibians have received most attention (e.g. Taberlet et al., 1998; Fijarczyk et al., 2011; Kuznetsova et al., 2012), fewer studies dealt with invertebrates (e.g. Rokas et al., 2003; Goropashnaya et al., 2004b; Schmitt et al., 2005; Schlick-Steiner et al., 2007; Leppänen et al., 2011, Leppänen et al., 2013). In ants, most studies have been done on a small spatial scale (Push et al., 2006; Mäki-Petäys & Breene, 2007; Wysocka et al., 2011) or compared a few distinct populations only (Goropashnaya et al., 2004a; Sanllorente et al., 2010). Results differed among species and strongly depended on their ecology and present range. The ant Liometopum microcephalum (Panzer, 1798) is the only species of its genus inhabiting Europe. It belongs to the subfamily Dolichoderinae, which dominates the fossil record of ants from the Palaeogene (Dlussky & Raznitsyn, 2002). The oldest representative of the genus Liometopum was found in Baltic amber from the mid-Eocene (Wheeler, 1915). At present, populations of L. microcephalum are scattered from Italy in the west to Russia’s Lower Volga region, western Iran and Israel in the east. The northern border of its range passes through Czechia (short geographic name for the Czech Republic), Slovakia, the Ukraine and European Russia. The southern border is delimited by the Mediterranean Sea. For most countries, published information on its present distribution is scarce and outdated. For example, in Italy the species was widely distributed until the mid-20th century (e.g. Menozzi, 1924; Menozzi, 1942; Consani & Zangheri, 1952) but does not occur anymore at many of these sites (pers. obs.). From Serbia, Bosnia-Herzegovina, Montenegro, Romania, Macedonia, Greece, Moldova, Turkey, and Israel there are only few records. Communication with local myrmecologists confirms that this is indeed due to its rare occurrence in these countries. In contrast, the northwestern-most population in south-eastern Czechia counts ca 1000 colonies and seems to be very vital (Schlaghamerský & Omelková, 2007). In Bulgaria the species is abundant at low altitudes (T. Ljubomirov, pers. comm.) and in Hungary the species still colonizes new sites nowadays (L. Gallé, pers. comm.). In many of the above-mentioned countries L. microcephalum has been red-listed. Existing populations are more or less spatially isolated due to the species’ specific habitat requirements. In the north of its range, the species inhabits mostly floodplains of large rivers or other humid and warm lowland sites, whereas in the south also higher elevations are colonized. The species is closely associated with large trees, in particular oaks, building carton nests in their trunks or large limbs several metres above ground (Petráková & Schlaghamerský, 2014). At least some colonies inhabit several interconnected nests (Makarevich, 2003). Mature colonies consist of hundreds of thousands of workers aggressively defending the nest trees and their surroundings. Thus they reach high ecological importance, affecting the composition and distribution of invertebrate fauna and

106

nutrient fluxes inside their territories (Petráková & Schlaghamerský, 2011; Schlaghamerský et al., 2013). Considering the scattered occurrence of L. microcephalum, fragmentation of its preferred habitat, and the continuing decline in colony numbers in many areas, we decided to study the genetic structure of individual populations to assess their origin and age. In ants, dispersal and gene flow are restricted to a short period between mating and colony founding (Pamilo et al., 1997). According to Wiest (1966) young queens of L. microcephalum are not able to establish new colonies without the help of workers or other queens. That could result in slow dispersal. However, our observation of winged males and females indicates that mated females might also disperse by flight. The objectives were (1) to assess the genetic diversity of populations across the species range, and (2) to reconstruct the colonization history of the species’ distribution area based on phylogenetic relationships among colonies. On the basis of known information about the species’ ecology we hypothesized that it had dispersed through lowlands and that high mountain ranges acted as dispersal barriers for the thermophilous species. We assumed that the present genetic structure reflected patterns of past colonization events and subsequent fragmentation of suitable habitat. We expected high genetic diversity in the Balkans and the Italian Penninsula as probable glacial refugia, and low genetic diversity at the northern sites.

MATERIALS AND METHODS

Workers of L. microcephalum from individual colonies were collected with the objective to cover the entire distribution area. When possible, the sampled ants were stored in vials with pure ethanol (the specimens from Lower Volga had been stored as dry material). In total, 40 sites in 16 countries were sampled (Table 1). Sampling sites were not spaced equally within the distribution area. This was partially caused by fragmentation of the range of the species, partially by the insufficient knowledge about its occurrence.

DNA sequencing At least three workers from each colony were sequenced. DNA was extracted using the DNeasy Blood & Tissue kit (Qiagen), following the manufacturer’s protocol. Fragments of mitochondrial cytochrome b (cyt b), cytochrome c oxidase subunit I (COI), 12S and nuclear 28S rRNA genes were PCR-amplified using primers listed in Table 2. The reaction mixture contained 5

μL of DNA, 1 μL of each primer (10 μM), 0.4 μL of 10 mM dNTPs, 2.2 μL of 25 mM MgCl2, 2 μL of 10x PCR buffer, 1 μL of bovine albumine serum (BSA), 0.3 μL of Taq Polymerase (5 u/μL), and 7.1 μL of DNA-free water. PCR cycle conditions in mitochondrial gene fragments were as follows: initial denaturation at 94 ºC for 3.5 min; 35 cycles of 94 ºC for 1 min, annealing temperature for 1 min (Table 2), 72 ºC for 1.5 min; a final extension at 72 ºC for 7 minutes. In the case of 28S rRNA gene, initial denaturation at 95 ºC for 3 min; 38 cycles of 95 ºC for 45 s, 53 ºC for 45 s and 72 ºC for 1 min; a final extension at 72 ºC for 7 minutes. PCR products were detected by electrophoresis in 2% GoodView-stained agarose gels. Amplified products were purified using the QIAquick PCR Purification Kit (Qiagen) and sequenced in both directions with BigDye Terminator v3.1 Sequence Kit (Applied Biosystems). Sequencing was carried out on an ABI Prism 3130 Genetic Analyzer (Applied Biosystems). All sequences were assembled in SEQUENCHER 4.8 (Gene Codes Corporation, Ann Arbor, MI). The sequences were checked for potential presence of NUMTs comparing to Liometopum luctuosum, L. sinense and L.occidentale (from GenBank). Neither ghost bands, double peaks,

107

Table 1: Overview of sampled localities of Liometopum microcephalum (Col. = number of sampled ant colonies, Ind. = number of sequenced ant workers, Czechia is the short, geographic name for the Czech Republic).

Site Country Location Coordinates Date Collected by Col. Ind. Sughereta N 41° 22' 47.58" 1 Italy 4/2010 N. Jansson 1 3 di San Vito E 13° 18' 58.33" Calabria, N 39° 59' 05.27" 2 Italy 6/1997 D. Hauck 1 3 Pollino E 16° 18' 25.72" N 45° 09' 15.9" 3 Italy Mantova 10/2011 L. Petráková 1 3 E 10° 46' 46.0" Gargano, N 41° 51' 48" 4 Italy 5/2012 J. Procházka 1 3 Umbra E 16° 02' 40" N 45° 46' 53.4" 5 Croatia Zagreb 9/2012 M. Zec 1 3 E 15° 56' 38.4" Dolina N 46° 32' 29.7" 6 Slovenia 6/2007 G. Bračko 1 3 pri Lendavi E 16° 30' 05.2" N 45° 11' 47.51" 7 Croatia Lonjsko Polje 7/2012 L. Čížek 1 3 E 17° 7' 39.26" N 48° 17' 7" 8 Austria Marchegg 8/2009 M. Tista 1 6 E 16° 53' 47" Sekulská N 48° 37' 38.6" 9 Czechia 7/2010 L. Petráková 1 3 Morava E 16° 57' 07.5" N 48° 49' 03" 10 Czechia Milovický les 7/2010 L. Petráková 1 3 E 16° 43' 05" N 48° 50' 50.3" 11 Czechia Křivé jezero 8/2010 L. Petráková 2 6 E 16° 43' 32.7" Rumunská N 49° 02'00.96" 12 Czechia 2013 D. Hauck 1 3 bažantnice E 16° 41' 56.28" N 48° 58' 28.98" 13 Czechia Uherčický les 2013 D. Hauck 1 3 E 16° 38' 55.14" Tvrdonice - N 48° 45' 13.18" 14 Czechia 8/2009 J. Schlaghamerský 1 3 Rýnava E 17° 01' 29.36" N 47° 46' 41.9" 15 Slovakia Dunajské luhy 8/2012 L. Petráková 1 3 E 17° 41' 39.8" Kamenice nad N 47° 49' 22.5" 16 Slovakia 7/2015 P. Průdek 1 3 Hronom E 18° 46' 47" N 48° 18' 46.2" 17 Slovakia Lučenec 8/2012 L. Petráková 1 4 E 19° 36' 58.6" N 42° 27' 41.7" 18 Montenegro Herceg Novi 5/2013 A. Vesnic 1 3 E 18° 31' 53.7" N 42° 23´ 45" 19 Montenegro Pikal, Cijevna 5/2013 G. Bračko 1 3 E 19° 22´49.2" N 41° 24´ 30" 20 Macedonia Demir Kapija 4/2010 G. Bračko 1 3 E 22° 17´0" N 38° 58´ 41.6" 21 Greece Agii Apostoli 7/2012 J. Procházka 1 3 E 20° 48´ 15.7" N 41° 5´ 36.6" 22 Greece Galani, Xanthi 5/2014 G. Bračko 1 3 E 24° 45´ 31.8" N 41° 9´ 25.2" 23 Greece Plagia, Arriana 4/2014 G. Bračko 1 3 E 25° 45´ 34.2" N 46° 41' 33.2" 24 Hungary Köres 8/2004 J. Schlaghamerský 1 3 E 21° 21' 05.5" N 47° 33' 10.66" A. Tartally, 25 Hungary Debrecen 7/2011 3 9 E 21° 37' 21.56" J. Schlaghamerský N 41° 44' 42.12" 26 Bulgaria Gorna Breznitsa 9/2012 T. Ljubomirov 1 3 E 23° 06' 55.45"

108

Milin Kamuk N 43°19'58.76" 27 Bulgaria 9/2012 T. Ljubomirov 1 7 hills, Banitsa E 23°39'55.76" Kardzhali, N 41º 39' 10.91" 28 Bulgaria 5/2010 D. Horal 1 3 Gnyazdovo E 25º 33' 3.53" Krumovgrad, N 41º 33' 10.72" 29 Bulgaria 5/2010 D. Horal 1 3 Dolna Kula E 25º 38' 45.43" N 45° 45' 05.9" 30 Romania Timişoara 6/2014 J. Procházka 1 3 E 21° 13' 11.4" N 44° 09' 36.63" 31 Romania Comana 6/2012 L. Čížek 1 3 E 26° 06' 02.84" N 44° 26' 10.38" 32 Romania Bucurest 6/2012 B. Markó 1 3 E 26° 05' 24.89" N 44° 20' 23.23" 33 Romania Cernavodă 6/2012 B. Markó 1 3 E 28° 01' 48.62" N 45° 57' 33.50" 34 Moldova Rosu 6/2014 V. Škorpíková 1 3 E 28° 12' 00.83" Edirne, N 41º 39' 45" 35 Turkey 10/2011 K. Kiran 1 3 Karaagac E 26º 32' 56" Balikesir, N 39° 40' 36.49" 36 Turkey 10/2011 N. Jansson 1 5 Bakacak Köyü E 27° 43' 19.66" Antakya, N 36 º 04' 39" 37 Turkey 3/2011 N. Jansson 1 8 Altinözü E 36 º 14' 20" N 37° 55' 54" 38 Turkey Adiyaman 3/2011 N. Jansson 1 5 E 38° 22' 20" N 33° 13' 17.16" 39 Israel Khorshat Tal 2012 E. Avital 1 5 E 35° 37' 45.37" N 48° 39' 00.04" 40 Russia Volgograd 3/1999 K. A. Grebennikov 1 5 E 44° 29' 39.87"

Table 2: Primer sequences used in PCR amplification. The first gene fragment is nuclear, the other four represent mitochondrial genes.

Gene Annealin Primer Primer sequence Published in fragment g temp. 28Sforward-F2 5’-AGAGAGAGAGTTCAAGAGTACGTG-3’ Belshaw 28S 53 °C 28Sreverse -3DR 5’-TAGTTCACCATCTTTCGGGTC-3’ et al. 2001 LCO 5´-GGTCAACAAATCATAAAGATATTGG- 3´ Folmer COIa 46-48 °C HCO 5´-TAAACTTCAGGGTGACCAAAAAATCA- 3´ et al. 1994 C1-J-2183 5´-CAACATTTATTTTGATTTTTTGG- 3´ Simon COIb 44-48 °C TL2-N-3014 5´-TCCAATGCACTAATCTGCCATATTA- 3´ et al. 1994 CytB CB-J-10933 5´-TATGTTTTATGAGGACAAATATC- 3´ 44-48 °C Su et al. 2008

nor stop codons or indels were found, especially in protein-coding genes (see Bensasson et al., 2001). Thus we consider the sequences to be true mitochondrial gene fragments. Whenever unique mutations were found in the aligned sequences, we checked raw sequences once again manually and sequenced additional workers (up to eight workers per colony). Altogether, 151 workers were sequenced. The obtained sequences, represening unique haplotypes of the single genes, are available via GenBank (Accession numbers: KT844569 – KT844626).

Population genetic analyses The mitochondrial markers have been considered as a single locus for the analyses. The genes are located close to each other on the same chromosome and no recombination occurs in the 109

mitochondrial genome. Different genes often have different mutation rates, thus we present the partial results (haplotype networks for each gene fragment) as supporting information. Samples from the Lower Volga could not be included in all of the analyses because of the poor quality of the DNA - only the 12S rRNA gene and a very short fragment of the cytochrome b gene were amplified. Haplotype networks were constructed in TCS 1.21 (Clement et al., 2000). To obtain correlation coefficients of genetic and geographical distances (r), the Mantel test was performed with 10 000 permutations using ALLELES IN SPACE (Miller, 2005). Subpopulations were delimited based on the geographic distribution of the sampled colonies, taking into account the existence of geographic barriers that could prevent dispersal of the ants. Five subpopulations were defined: (a) Black Sea population (sites 22, 23, 27-29, 31-35), delimited by the Black Sea in the east and by the Carpathians, Balkan and Rhodope Mountains in the west. The colony from the Lower Volga (40) was assigned to this population due to the lack of any geographical barrier other than distance; (b) Balkan population (sites 18-21, 26), assuming its refugial character during past climatic oscillations; (c) Italian population (sites 1-4; delimited by the Alps in the north and by the Mediterranean Sea); (d) Pannonian population (sites 5-17, 24, 25, 30; inhabiting the Pannonian Basin sensu lato, delimited by the Alps, Dinaric Alps, Balkan Mts., and the Carpathians); (e) Western Anatolian population (site 36); (f) Eastern Anatolian population (site 38); and (g) Levantine population (sites 37, 39). Haplotype (Hd) and nucleotide (π) diversities for each population were calculated in DNASP (Librado & Rozas, 2009). Analysis of molecular variance (AMOVA), mismatch distribution (estimated using the sudden expansion model) and neutrality tests (Tajima’s D, Fu’s F) were performed in ARLEQUIN 3.1 (Excoffier et al., 2005).

Lineage divergence

The lineage divergence was calculated using BEAST 2.3.1 (Bouckaert et al., 2014). The data were divided into partitions for each gene and codon position separately. The best model of nucleotide substitution was selected using PARTITION FINDER v1.1.1 (Lanfear et al., 2012). To assess the time of origin of L. microcephalum, three distinct haplotypes (h19, h29, h36) were compared to related species (primary calibration). Six calibration points (Liometopum stem, Tapinoma stem, Dolichoderus stem, Dolichoderinae, Pseudomyrmecinae and the root) were denoted according to Ward et al. (2010), who reconstructed the phylogeny of Dolichoderinae using molecular markers and fossils. The compared species were: Liometopum luctuosum, L. occidentale, L. sinense, Tapinoma erraticum, T. sessile (the genus Liometopum belongs to the tribe Tapinomini), Dolichoderus pilosus, D. quadripunctatus, D. taprobanae (Dolichoderus is considered to be the most ancient genus of the subfamily Dolichoderinae; Ward et al., 2010), Linepithema humile, L. oblonga, Technomyrmex albipes, T. antennus, Iridomyrmex sanguineus. Pseudomyrmecinae, closely related to Dolichoderinae (Moreau et al., 2006), were represented by Pseudomyrmex dendroicus, P. flavicornis, Tetraponera attenuata, and T. rufonigra. Tapinoma erraticum and D. quadripunctatus were sequenced for this study, the rest of the sequences were obtained from the GenBank database (accession numbers are provided in Appendix S1). The sequences were aligned using MAFFT (Katoh & Standley, 2013). The evolution of 28S and the first two codon positions of COI and cyt b genes was simulated under the GTR+I+G model, the third codon positions of the COI and cyt b genes ran under the HKY+G model of nucleotide substitution. The 12S rRNA gene was not involved in the primary calibration because sequences of this gene were not available for outgroup species. Log- normal relaxed molecular clock model and Yule speciation process were modelled with the Markov chain Monte Carlo for 25 million of generations. In the secondary calibration, only Liometopum and Tapinoma spp. were used as outgroups. We selected four calibration points: root (common ancestor 110

of the two genera), Tapinoma stem and Liometopum stem were set based on Ward (2010), L. microcephalum common ancestor divergence time was assessed based on the first calibration (mean 1.05, 0.4 SD, 1.4 offset). The evolution of 28S, 12S and the first two codon positions of the COI and cyt b genes was simulated under the HKY+G model, the third codon positions of the COI and cyt b genes under the TN93 model. The log-normal relaxed molecular clock model and the coalescent speciation process for constant population size were modelled with the MCMC chain for 25 million generations. The first 30 percent of the sampled trees were discarded as a burn-in. The runs were checked in TRACER v1.6.0 (Rambaut et al., 2014). Phylogenetic trees were edited using FIGTREE v1.4.2 (Rambaut, 2014). The phylogenetic tree used for the primary calibration is presented in Appendix S2. S-DIVA implemented in RASP software (Yan et al., 2010) was used to reconstruct the ancestral area of the individual clades based on the obtained phylogenetic trees. The distribution range was divided into 13 geographic areas: Levant, Western Anatolia, Eastern Anatolia, Thrace, Black Sea Coast, Eastern Balkans, Southern Balkans, Northern Balkans, Western Pannonia, Northern Pannonia, Dinaric Alps, Northern Italy and Southern Italy. Two ancestral areas were allowed as the maximum number in the analysis, excluding direct migration between the Levant and Northern Pannonia and between the Levant and Italy.

RESULTS

Genetic diversity Only haplotypes for the concatenated dataset (2378 nucleotides, 36 haplotypes) are presented. For detailed information about single genes see Appendix S3. Figure 1a and Table 3 show the distribution of the haplotypes across the species range and among the populations. At ten sites (8, 11, 16, 19, 22, 23, 25, 27, 36, 37), two or three different haplotypes occurred. Different haplotypes were found even within single colonies (sites 19, 22, 23, 27, 36, 37). The nuclear marker, 28S rRNA (611 bp), was conservative across the whole distribution area of L. microcephalum and was identical to that of L. occidentale and L. luctuosum from North America. Results of the Mantel test showed significant correlation between genetic and geographic distances (r = 0.6003, p = 0.001). Nucleotide diversities (π) varied from 0.0007 (Black Sea population) to 0.0067 (Levantine population), haplotype diversities (Hd) varied from 0.872 to 1 (for details see Table 3). Italian and Pannonian populations shared one haplotype and the Black Sea population shared one haplotype with the Balkan one. The remaining populations had unique haplotypes. Observed values of Tajima’s D and Fu’s F were significantly higher than the simulated ones in the Black Sea population only. Indices of mismatched nucleotide distribution were significantly different from the simulated values in all populations except the Black Sea one. The Black Sea population had a unimodal distribution, whereas the others had a multimodal mismatched distribution. In the hierarchical AMOVA, 41.4% of the total genetic variance were found between the two groups (European vs Asian), 39.1% among populations within groups and 19.5 % within populations (Table 4).

Lineage divergence The phylogenetic analyses resulted in seven clades (Fig. 2). Six of them were well supported (PP = 1), whereas the NW Balkan Clade (sites in Montenegro; h11-h13) had low nodal support (PP

111

Figure 1: The distribution of single haplotypes (concatenated dataset of three mitochondrial genes, 2378 bases: cytochrome c oxidase I, cytochrome b and 12S rRNA) in Liometopum microcephalum at the sampled sites (different colours correspond to different haplotypes, h1-h36; numbers correspond to the sites, 1-39; for details see Table 1) a) Map of the whole distribution range (Lambert Conformal Conic Projection); b) Haplotype network (circle sizes are proportional to haplotype frequencies, lines perpendicular to the lines connecting haplotypes represent mutation steps). Site 40 (Lower Volga) is shown with a smaller circle in the map and was not included into the haplotype network as only 12S rRNA and a short fragment of cyt b were amplified because of poor sample quality).

112

Figure 2: Phylogenetic relationships among analyzed colonies with lineage divergence time estimations based on DNA sequence data (12S, cytochrome c oxidase I, cytochrome b, 28S rRNA genes; 2981 characters) calculated in BEAST. Horizontal blue bars represent the 95% credibility limits for node ages; divergence time was assessed for nodes with posterior probability node support higher than 0.7.

113

= 0.58). The haplotype network showed seven groups of haplotypes corresponding to the clades (Fig. 1b). The Black Sea Clade (haplotypes h1-h10) included colonies from Romania, north- and southeastern Bulgaria, Macedonia, Moldova, Thrace, eastern Anatolia and most probably also the one sampled on the Lower Volga. The Western Clade (h14-23) included colonies from Italy, Croatia, Austria, and part of those from Czechia, eastern Hungary and south-western Romania. The Northern Clade (h24-h28) included samples from one part of the colonies from Czechia, those from Slovakia, north-eastern Slovenia and north-eastern Hungary. Samples from southern Greece and south-western Bulgaria formed the Balkan Clade (h29, h30). Western Anatolia (h31-33) and the Levant (h3-h36; Israel, southernmost Turkey) represented two separate clades. The colony sampled in western Anatolia differed from all the European colonies, but was closer to them than to the Levantine group. Over 40 mutation steps separated the Levantine Clade from the others. Taking related species into account, the Levantine Clade appears to be the basal one. Assessing the time of lineage divergence, the common ancestor of all sampled L. microcephalum colonies lived most probably 3.88 Ma (2.7-5.3 Myr; median with 95% highest posterior density interval). The common ancestor of the European and Anatolian colonies was dated to 2.2 Ma (1.35-3.25 Myr). The Balkan Clade and Northern Clade shared their ancestor 850 ka (0.41- 1.44 Myr). The common ancestor of the Black Sea Clade and NW Balkan Clade is dated to 500 ka (263-830 kyr). The Western Anatolian Clade formed 229 ka (43-540 kyr), the Black Sea Clade formed 268 ka (118-484 kyr) and the Northern Clade arose 236 ka (77-489 kyr). The ancestor of the Balkan Clade was dated to 181 ka (20-478 kyr). The common ancestor of the Western Clade was dated to 310 ka (133-577 kyr). However, colonies inhabiting the Italian Peninsula (including the Po Plain) shared their ancestor about 168 ka (62-34 kyr). The youngest divergence within individual European clades was dated to 50 ka (0.4-151 kyr). S-DIVA suggested W Anatolia and the Levant as the ancestral areas for L. microcephalum (92% probability). N Pannonia was assigned as the ancestral area for all European colonies (93%). The Black Sea Clade most probably dispersed from Thrace (74%). The Western Clade originated in Pannonia (82%). In total, 20 vicariance and 18 dispersal events were identified (for details see Fig. 3).

DISCUSSION

In ants, female dispersal presents the crucial mechanism of colonization of new areas. Therefore we have used mitochondrial markers to reconstruct colonisation events. The high number of mitochondrial haplotypes found across the distribution area of L. microcephalum is not surprising when compared to similar studies on arthropods (e.g. Leppänen et al., 2011; Pauls, 2009). In contrast, the nuclear marker was invariable across the whole species range. Mitochondrial and nuclear markers evolve with different rates. Papadopoulou et al. (2010) estimated the rate for insects to be 3.54% My- 1 in cytochrome c oxidase and 0.12% My-1 in the 28S rRNA gene. Several distinct haplotypes found within single colonies showed that at least those colonies included several queens. Colonies could have been founded by several females, even non-related ones, because cooperation among young queens decreases their mortality (Nonacs, 1990). Furthermore, young queens could have been adopted into a mature colony (Evans, 1996). Such associations often arise when the habitat is saturated (Chapuisat et al., 2004) or when mated queens have restricted dispersal ability (Hölldobler & Wilson, 1990). A significant correlation between genetic and geographic distance stands for isolation by distance. Indeed, the species inhabits isolated islands of suitable habitats, divided often by hundreds of kilometres. Colonies from the Levant differed distinctly from the rest of the samples, whereas the

114

Table 3: Haplotype frequencies in each population (defined based on the geographic location of sampled sites). The numbers of sites / sequenced specimens are given in parentheses. Presented are nucleotide diversity (π), haplotype diversity (Hd), mismatch distribution analysis (SSD = Sum of Squared Deviation, Raggedness = Harpending's Raggedness Index), and neutrality test outputs (Tajima’s D, Fu’s F); * and ** denote significant differences between simulated and observed values (sim. >= obs.) at the 0.05 and 0.01 level, respectively.

Population Western Eastern Black Sea Italian Pannonian Balkan Levantine Anatolian Anatolian (10/37) (4/12) (14/61) (5/15) (2/13) (1/5) (1/5) Haplotypes h1 (16) h14 (3) h15 (3) h1 (3) h31 (2) h9 (5) h34 (5) h2 (1) h19 (3) h16 (3) h11 (2) h32 (1) h35 (3) h3 (3) h21 (3) h17 (3) h12 (1) h33 (2) h36 (5) h4 (4) h22 (3) h18 (9) h13 (3) h5 (3) h19 (3) h29 (3) h6 (3) h20 (3) h30 (3) h7 (3) h22 (3) h8 (3) h23 (3) h10 (1) h24 (3) h25 (20) h26 (3) h27 (3) h28 (4) π 0.0007 0.00126 0.00376 0.00421 0.0011 0.00673 - (± SD) ± 0.0001 ± 0.0003 ± 0.0003 ± 0.0008 ± 0.0003 ± 0.0024 Hd 0.872 1 0.882 1 1 1 - (± SD) ± 0.091 ± 0.177 ± 0.064 ± 0.096 ± 0.272 ± 0.272 SSD 0.0019 0.1758* 0.0555* 0.056* 0.1849 0.2149** -

Raggedness 0.0501 0.7025** 0.042 0.1336* 0.6 0.4403** -

Tajima’s D -1.1951 0.8978 2.3467 1.6236 0.9571 2.2023 -

Fu’s Fs -3.0384* 1.4358 3.5041 4.3534 0.8036 10.9674 -

Table 4: Analysis of molecular variance (AMOVA) in Liometopum microcephalum among and within seven populations and two groups (European vs. Asian) calculated in ARLEQUIN with 10000 permutations.

Source of variation d.f. Sum of squares % of variation Fixation index P Among groups 1 352.987 41.39 F = 0.414 0.025 CT Among populations

within groups 5 608.481 39.11 FSC = 0.667 0

Within populations 141 439.769 19.50 FST = 0.805 0 Total 147 1401.236

115

Figure 3: Potential ancestral areas reconstructed in S-DIVA. Probabilities of ancestral areas are shown only for nodes with posterior probabilities (PP) exceeding 0.7, areas with probabilities lower than 1% are not listed. h1-h36 = haplotypes; 1-39 = sites, numbers next to nodes = PP; / = vicariance event, > = dispersal.

116

colony sampled in eastern Anatolia was surprisingly most similar to those from the European Black Sea coast. Anatolia is considered to be the centre of genetic diversity for several western Palearctic species and one of the source areas for the colonization of Europe (Franck et al., 2001; Rokas et al., 2003, Bilgin, 2011). This seems to be also true for L. microcephalum as colonies from the Levant seem to be more related to the common ancestor of L. microcephalum than to the “most derived” populations occurring in the western-most part of its distribution range. Similarly, Rokas et al. (2003) observed that populations of Andricus quercustozae (Hymenoptera: Cynipidae) from Europe and central Anatolia formed a single clade, which differed from samples collected in south-western and north-eastern Anatolia, suggesting that this divergence predated the Pleistocene. Considering the lineage divergence estimations, our calculations correspond to those published by Ward et al. (2010). Despite the long time spans, the median values allow for reasonable interpretation when compared to the time of geological and climatic events in the past. The Levant had high tectonic and volcanic activity from the late Miocene to the Holocene (Adiyaman & Chorowicz, 2002), thus the local L. microcephalum population could have been isolated since the Pliocene. The median value of divergence time between the western Anatolian colonies and the European colonies corresponds to the transition between the Tertiary and Quaternary. The end of the Tertiary was characterized by cooling of the climate. Thus the common ancestors of both groups had to survive in a warm area. Based on topology of basal nodes and ancestral area reconstruction, the species could have spread northwesternwards from Anatolia, reaching the Pannonian Basin. Several ant lineages diverged during the Pleistocene, eventually giving rise to new species (Goropashnaya et al., 2004a; Schlick-Steiner et al., 2007; Wysocka et al., 2011; Goropashnaya et al., 2012; Leppänen et al., 2013). The origin of individual European clades is apparently related to the Pleistocene climatic oscillations. Recurrent glaciations during the Pleistocene must have caused the species’ range to shrink and expand repeatedly, affecting the species’ genetic diversity. The youngest divergencies within the European clades were dated to the last glacial period (Würm, 116-12 kyr ago; van Andel & Tzedakis, 1996). The Balkan Clade arose at the end of the Riss Glacial (240-188 kyr ago). The Black Sea Clade, Northern Clade and Western Clade formed during the Mindel/Riss Interglacial (430- 240 kyr ago; Kukla, 2005). The Balkan and Northern Clades shared their origin. Their ancestors probably lived in the Pannonian Basin much earlier but during the Riss glaciation one part of the ancestral population retreated to the south (Balkan refugium), while another had to survive the subsequent glaciation (Würm) in extra-Mediterranean refugia and evolved separately for a long time. Low nucleotide diversities with relatively high haplotype diversities in each population imply small effective population sizes in the past, followed by subsequent expansion (Grant & Bowen, 1998). The highest π was found in Levantine, Balkan and Pannonian populations, suggesting a refugial character of those areas. Liometopum microcephalum is dependent on its nest trees, strongly preferring oaks. Quercus spp. were present in the south-eastern part of Czechia dating back to 38-27 kyr ago (Willis & van Andel, 2004), but this area was covered by permafrost during the LGM (Vandenberghe et al., 2012). Considering the geographical position of the mutually related colonies in the northwest, we suggest survival in a refugium within the Carpathian Arc. This area has been mentioned in recent studies as an important refugium for a number of species (e.g. Kotlík et al. 2006; Ursenbacher et al., 2006; Fijarczyk et al., 2011). Other studies (Willis et al., 2000; Juřičková et al., 2014) suggested even the survival of deciduous trees, including oaks, in this refugium (Brewer et al., 2002; Feuerdan et al., 2011). The Black Sea coast was proposed as another refugium of Quercus spp. (Bhagwat & Willis, 2008). The Black Sea population of L. microcephalum apparently underwent a recent population expansion as indicated by the unimodal mismatch distribution, significantly negative Fu’s Fs value, lowest π and Hd. The species spread from Thrace northeastward along the northern coast of the

117

Black Sea up to the Lower Volga River. It could disperse here easier than overcoming the Caucasus Mountains when coming from the south. The occurrence of the same haplotype in eastern Anatolia is rather intriguing because of the spatial isolation of that population with many geographic barriers in-between. This haplotype is thus either an ancient one, or its occurrence there might be the consequence of a rather recent human-mediated introduction. Mutual genetic similarity and the divergence time estimations indicate a rather late colonization of the Italian Peninsula. During glacial periods, the sea level of the Adriatic Sea was substantially lower, and the land connection between the Italian Peninsula and south-eastern Europe was much wider than nowadays (Taberlet et al., 1998). This corridor could have been used by L. microcephalum as suggested by identical haplotypes found in Italian, Slovenian and Croatian samples. The easternmost point of the Italian Peninsula is only 70 km away from the Balkan Peninsula and thus winged females could have crossed that distance carried by strong wind. Similarly, Satyrinae butterflies probably entered the peninsula from south-eastern Europe (Dapporto et al., 2012). Three different haplotypes found within a small geographic area covering eastern Slovenia and Croatia suggest that another refugium could have been situated at the south-eastern margin of the Alps. The Northern Clade is ancient compared to the other European Clades. The northern part of the Pannonian Basin seems to be the ancestral area for all European populations. It was apparently colonized from two source populations, in different time periods. The dispersal northwards from the south-east happened before the species began to spread from the Italian Peninsula. Different times of dispersal could be explained by larger geographical barriers in the western direction in comparison with the easier dispersal through the Pannonian lowlands in the north. Some authors (Surget-Groba et al., 2006; Velekei et al., 2014) proposed a migration corridor leading from the Balkans across eastern Hungary and western Romania up to the western Ukraine. Colonies of L. microcephalum occurring in the southern part of the Pannonian Basin are more related to those from the Italian Peninsula and surprisingly not to those living in the Balkans. Parts of the Pannonian Basin were possibly left unoccupied at the end of the last glaciation and later recolonized from the Italian Peninsula. The Alps, as an unsurmountable geographic barrier for this species, played an important role in channelling its expansion from the Italian Peninsula towards the east. The species is entirely absent westwards of the Alps as it apparently was not able to overcome the high elevations to the north-west of the Italian Peninsula. Our results show that the present distribution range of L. microcephalum is the outcome of a rather complex colonization history. While a centre of genetic diversity in Anatolia was to be expected, the proof of a rather old northern lineage that survived the last iceage within the Carpathian Arc, at the Black Sea coast and possibly also at the southeastern margin of the Alps is an important finding. As the species is heavily dependent on rather large deciduous trees, this is another proof of the importance of these extra-Mediterranean refugia. Also the rather late colonization of the Italian Peninsula and the very ancient character of the populations in the Levant (Hatay province of Turkey and Israel-Lebanon borderland) are remarkable outcomes. Both lineages are distinct from those in other areas of the species’s range and worth preserving. The remnant populations in the Levant, representing the most ancient lineage, have to be considered particularly at risk and deserve further study and strict protection. The European velvety tree ant is a conspicuous insect species with an interesting ecology (not fully explored yet) that could be well promoted as one of the flagship species for the conservation of veteran trees and macrohabitats such as floodplain forests. Although historic and current data probably allow to demonstrate considerable decline of the majority of its populations, we are not aware of any official conservation efforts on the national or multinational level exceeding the mere red-listing of the species.

118

ACKNOWLEDGEMENTS

We are grateful to the colleagues, who provided samples for genetic analyses and information about the present occurrence of L. microcephalum: Eygen Avital, Gregor Bračko, Lukáš Čížek, Dmitry A. Dubovikoff, Laszló Gallé, David Hauck, David Horal, Nicklas Jansson, Marko Karaman, Kadri Kiran, Toshko Ljubomirov, Bálint Markó, Omid Paknia, Irinel Eugen Popescu, Jiří Procházka, Pavel Průdek, Zhanna Reznikova, Vlasta Škorpíková, András Tartally, Melanie Tista, Adi Vesnic, and Maté Zec. Thanks to Tatiana Aghová for her advice concerning analyses using BEAST software. Keith Alexander kindly checked the English.

REFERENCES

Adiyaman, Ö. & Chorowicz, J. (2002) Late Cenozoic tectonic and volcanism in the northwestern corner of the Arabian plate: a consequence of the strike-slip Dead Sea fault zone and the lateral escape of Anatolia. Journal of Volcanology and Geothermal Research, 117, 327-345.

Belshaw, R., Lopez-Vaamonde, C., Degerli, N., Quicke, D.L.J. (2001) Paraphyletic taxa and taxonomic chaining: evaluation the classification of braconine wasps (Hymenoptera: Braconidae) using 28S D2-3 rDNA sequences and morphological characters. Biological Journal of the Linnean Society, 73 (4), 411-424.

Bensasson, D., Zhang, D.-X., Hartl, D.L. & Hewitt, G.M. (2001) Mitochodrial pseudogenes: Evolution´s misplaced witnesses. Trends in Ecology and Evolution, 16, 314-321.

Bhagwat, S.A. & Willis, K.J. (2008) Species persistence in northerly glacial refugia of Europe: a matter of chance or biogeographical traits? Journal of Biogeography, 35, 464-482.

Bilgin, R. (2011) Back to the suture: the distribution of intraspecific genetic diversity in and around Anatolia. International Journal of Molecular Sciences, 12 (6), 4080-4103.

Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C.-H., Xie, D., Suchard, M.A., Rambaut, A. & Drummond, A.J. (2014) BEAST 2: A software platform for Bayesian evolutionary analysis. PLoS Computational Biology, 10(4), e1003537.

Brewer, S., Cheddadi, R., de Beaulieu, J.L. & Reille, M. (2002) The spread of deciduous Quercus throughout Europe since the last glacial period. Forest Ecology and Management, 156, 27-48.

Buse, J., Griebeler, E.M. & Niehuis, M. (2013) Rising temperatures explain past immigration of the thermophilic oak-inhabiting beetle Coraebus florentinus (Coleoptera: Buprestidae) in south-west Germany. Biodiversity and Conservation, 22, 1115-1131.

Chapuisat, M., Bocherens, S. & Rosset, H. (2004) Variable queen number in ant colonies: No impact on queen turnover, inbreeding, and population genetic differentiation in the ant Formica selysi. Evolution, 58, 1064-1072.

Clement, M., Posada, D., & Crandall, K.A. (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology, 9 (10), 1657-1660.

Collins, P.M., Davis, B.A.S. & Kaplan, J.O. (2012) The mid-Holocene vegetation of the Mediterranean region and southern Europe, and comparison with the present day. Journal of 119

Biogeography, 39, 1848-1861.

Consani, M. & Zangheri, P. (1952) Fauna di Romagna. Imenotteri – Formicidi. Memorie della Societa Entomologica Italiana, 31, 38-48.

Cook, C.E., Austin, J.J. & Disney, H.L. (2004) A mitochondrial 12S and 16S rRNA phylogeny of critical genera of Phoridae (Diptera) and related families of Aschiza. Zootaxa, 593, 1-11.

Dapporto, L., Bruschini, C., Dincă, V., Vila, R. & Dennis, R.L.H. (2012) Identifying zones of phenetic compression in West Mediterranean butterflies (Satyriane): refugia, invasion and hybridization. Diversity and Distributions, 18, 1076-1086.

Dlussky, G.M. & Raznitsyn, A.P. (2002) Ants (Hymenoptera: Formicidae) of Formation Green River and some other Middle Eocene deposits of North America. Russian Entomological Journal, 11, 411-436.

Evans, J. (1996) Queen longevity, queen adoption and posthumous indirect fitness in the facultatively polygynous ant Myrmica tahoensis. Behavioral Ecology and Sociobiology, 39, 275-284.

Excoffier, L., Laval, G. & Schneider, S. (2005) Arlequin (version 3.0): An integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online, 1, 47-50.

Feurdean, A., Tanţǎu, I., Fǎrcaş, S. (2011) Holocene variability in the range distribution and abundance of Pinus, Picea abies and Quercus in Romania; implication for their current status. Quaternary Science Reviews, 30, 3060-3075.

Fijarczyk, A., Nadachowska, K., Hofman, S., Litvinchuk, S.N., Babik, W., Stuglik, M., Gollmann, G., Choleva, L., Cogălniceanu, D., Vukov, T., Džukić, G. & Szymura, J.M. (2011) Nuclear and mitochondrial phylogeography of the European fire-bellied toads Bombina bombina and Bombina variegata supports their independent histories. Molecular Ecology, 20, 3381-3398.

Folmer, O., Black, M., Hoeh, W., Lutz, R. & Vrijenhoek, R. (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3(5), 294-299.

Franck, P., Garnery, L., Loisseau, A., Oldroyd, B.P., Hepburn, H.R., Solignac, M. & Cornuet, J.M. (2001) Genetic diversity of the honeybee in Africa: microsatellite and mitochondrial data. Heredity, 86, 420-430.

Goropashnaya, A.V., Fedorov, V.B. & Pamilo, P. (2004a) Recent speciation in the Formica rufa group ants (Hymenoptera, Formicidae): interference from mitochondrial DNA phylogeny. Molecular Phylogenetics and Evolution, 32 (1), 198-206.

Goropashnaya, A.V., Fedorov, V.B., Seifert, B. & Pamilo, P. (2004b) Limited phylogeographical structure across Eurasia in two red wood ant species Formica pratensis and F. lugubris (Hymenoptera, Formicidae). Molecular Ecology, 13, 1849-1858.

Goropashnaya, A.V., Fedorov, V.B., Seifert, B. & Pamilo, P. (2012) Phylogenetic relationships of palaearctic Formica species (Hymenoptera: Formicidae) based on mitochondrial cytochrome b sequences. PLoS ONE, 7 (7), e41697.

120

Grant, W.S. & Bowen, B.W. (1998) Shallow population histories in deep evolutionary lineages of marine fishes: insights from sardines and anchovies and lessons for conservation. Journal of Heredity, 89 (5), 415-426.

Hewitt, G.M. (1996) Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society, 58, 247-276.

Hewitt, G.M. (1999) Post-glacial re-colonization of European biota. Biological Journal of the Linnean Society, 68, 87-112.

Hölldobler, B. & Wilson, E.O. (1990) The Ants. Harvard University Press, Cambridge, MA.

Horn, S., Prost, S., Stiller, M., Makowiecki, D., Kuznetsova, T., Benecke, N., Pucher, E., Hufthammer, A.K., Schouwenburg, C., Shapiro, B. & Hofreiter, M. (2014) Ancient mitochondrial DNA and the genetic history of Eurasian beaver (Castor fiber) in Europe. Molecular Ecology, 23, 1717-1729.

Juřičková, L., Horáčková, J. & Ložek, V. (2014) Direct evidence of central European forest refugia during the last glacial period based on mollusc fossils. Quaternary Research, 82, 222-228.

Katoh, K. & Standley, D.M. (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution, 30 (4), 772-780.

Kotlík, P., Deffontaine, V., Mascheretti, S., Zima, J., Michaux, J.R. & Searle, J.B. (2006) A northern glacial refugium for bank voles (Clethrionomys glareolus). Proceedings of the National Academy of Sciences of the United States of America, 103 (40), 14863.

Kukla, G. (2005) Saalian supercycle, Mindel / Riss interglacial and Milankovitch’s dating. Quaternary Science Reviews, 24, 1573-1583.

Kuznetsova, M.V., Danilkin, A.A. & Kholodova, M.V. (2012) Phylogeography of red deer (Cervus elaphus): analysis of mtDNA cytochrome b polymorphism. Biology Bulletin, 39 (4), 323-330.

Lanfear, R., Calcott, B., Ho, S.Y.W. & Guindon, S. (2012) PartitionFinder: Combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution, 29 (6), 1695-1701.

Leppänen, J., Vepsäläinen, K. & Savolainen, R. (2011) Phylogeography of the ant Myrmica rubra and its inquiline social parasite. Ecology and Evolution, 1 (1), 46-62.

Leppänen, J., Vepsäläinen, K., Anthoni, H. & Savolainen, R. (2013) Comparative phylogeography of the ants Myrmica ruginodis and Myrmica rubra. Journal of Biogeography, 40, 479-491.

Librado, P. & Rozas, J. (2009): DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25, 1451-1452.

Makarevich, O.N. (2003) Liometopum microcephalum on the Lower Dnepr. Vestnik Zoologii, 37, 51-56.

Mäki-Petäys, H. & Breene, J. (2007) Genetic vulnerability of a remnant ant population. Conservation Genetics, 8, 427-435.

121

Miller, M.P. (2005) Alleles in Space: Computer software for the joint analysis of interindividual spatial and genetic information. Journal of Heredity, 96, 722-724.

Menozzi, C. (1924). Res Mutinenses. Formicidae (Hymenoptera). Atti della Societa dei Naturalisti e Matematici di Modena, 6 (3), 22-47.

Menozzi, C. (1942) Esplorazione entomologica del Parco Nazionale del Circeo. Hymenoptera – Formicidae. M. Spadafora (ed.), Salerno, Italy.

Moreau, C.S, Bell, C.D., Vila, R., Archibald, S.B. & Pierce, N.E. (2006) Phylogeny of the ants: Diversification in the age of angiosperms. Science, 312, 101-104.

Nonacs, P. (1990): Size and kinship affect success of co-founding Lasius pallitarsis queens. Psyche, 97, 217-228.

Pamilo, P., Gertsch, P., Thoren, P. & Seppä, P. (1997) Molecular population genetics of social insects. Annual Review of Ecology and Systematics, 28, 1-25.

Papadopoulou, A., Anastasiou, I. & Vogler, A.P. (2010) Revisiting the insect mitochondrial molecular clock: the Mid-Aegean trench calibration. Molecular Biology and Evolution, 27 (7), 1659-1672.

Pauls, S.U., Theissinger, K., Ujvarosi, L., Balint, M. & Haase, P. (2009) Patterns of population structure in two closely related, partially sympatric caddisflies, in eastern Europe: historic introgression, limited dispersal, and cryptic diversity. Journal of the North American Benthological Society, 28, 517-536.

Petráková, L. & Schlaghamerský, J. (2011) Interaction between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence. Community Ecology, 12, 9-17.

Petráková, L. & Schlaghamerský, J. (2014) Worker polymorphism in the arboricolous ant Liometopum microcephalum (Hymenoptera: Formicidae: Dolichoderinae): Is it related to territory size? Myrmecological News, 20, 101-111.

Pusch, K., Seifert, B., Foitzik, S. & Heinze, J. (2006) Distribution and genetic divergence of two parapatric sibling ant species in Central Europe. Biological Journal of the Linnean Society, 88 (2), 223-234.

Rambaut, A. (2014) FigTree v1.4.2: Tree Figure Drawing Tool. Available from http://tree.bio.ed.ac.uk/software/figtree/

Rambaut A., Suchard M.A., Xie D. & Drummond A.J. (2014) Tracer v1.6, Available from http://beast.bio.ed.ac.uk/Tracer

Rokas, A., Atkinson, R.J., Webster, L.M.I., Csóka, G. & Stone, G.M. (2003) Out of Anatolia: longitudinal gradients in genetic diversity support an eastern origin for a circum-Mediterranean oak gallwasp Andricus quercustozae. Molecular Ecology, 12, 2153-2174.

Sanllorente, O., Hammond, R.L., Ruano, F., Keller, L. & Tinaut, A. (2010) Extreme population differentiation in a vulnerable slavemaking ant with a fragmented distribution. Conservation

122

Genetics, 11, 1701-1710.

Schlaghamerský, J. & Omelková, M. (2007) The present distribution and nest tree characteristics of Liometopum microcephalum (Panzer, 1798) (Hymenoptera: Formicidae) in South Moravia. Myrmecological News, 10, 85-90.

Schlaghamerský, J., Kašpar, J., Petráková, L. & Šustr, V. (2013) Trophobiosis in the arboricolous ant Liometopum microcephalum (Hymenoptera: Formicidae: Dolichoderinae). European Journal of Entomology, 110, 231-239.

Schlick-Steiner, B.C, Steiner, F.M., Sanetra, M., Seifert, B., Christian, E. & Stauffer, C. (2007) Lineage specific evolution of an alternative social strategy in Tetramorium ants (Hymenoptera: Formicidae). Biological Journal of the Linnean Society, 91, 247-255.

Schmitt, T., Röber, S. & Seitz, A. (2005) Is the last glaciation the only relevant event for the present genetic population structure of the meadow brown butterfly Maniola jurtina (Lepidoptera: Nymphalidae)? Biological Journal of the Linnean Society, 85, 419-431.

Schmitt, T. & Varga, Z. (2012) Extra-Mediterranean refugia: The rule and not the exception? Frontiers in Zoology, 9, 22.

Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & Flook, P. (1994) Evolution, weighting, and phylogenetic utility of mitochondrial gene-sequences and a compilation of conserved polymerase chain-reaction primers. Annals of the Entomological Society of America, 87, 651-701.

Su, K.F.I., Kutty, S.N. & Meier, R. (2008) Morphology versus molecules: the phylogenetic relationships of Sepsidae (Diptera: Cyclorrhapha) based on morphology and DNA sequences data from ten genes. Cladistics, 24, 902–916.

Surget-Groba, Y., Heulin, B., Guillaume, C.P., Puky, M., Semenov, D., Orlova, V., Kupriyanova, L., Ghira, I. & Smajda, B. (2006) Multiple origins of viviparity or reversal from viviparity to oviparity? The European common lizard (Zootoca vivipara, Lacertidae) and the evolution of parity. Biological Joournal of the Linnean Society, 87, 1-11.

Taberlet, P., Fumagalli, L., Wust-Saucy A.G. & Cosson J.F. (1998) Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology, 7, 453-464.

Ursenbacher, S., Carlsson, M., Helfer, V., Tegelström, H. & Fumagalli, L. (2006) Phylogeography and Pleistocene refugia of the adder (Vipera berus) as inferred from mitochondrial DNA sequence data. Molecular Ecology, 15, 3425-3437. van Andel, T.H. & Tzedakis, P.C. (1996) Palaeolithic landscapes of Europe and environs, 150,000- 25,000 years ago: an overview. Quaternary Science Review, 15, 481-500.

Vandenberghe, J., Renssen, H., Roche, D.M., Goosse, H., Velichko, A.A., Gorbunov, A. & Levavasseur, G. (2012) Eurasian permafrost instability constrained by reduced sea-ice cover. Quaternary Sciences Reviews, 34, 16-23.

Velekei, B., Lakatos, F., Bíró, P., Ács, É. & Puky, M. (2014) The genetic structure of Zootoca vivipara (Lichtenstein, 1823) populations did not support the existence of a north-south corridor of the VB haplogroup in eastern Hungary. North-Western Journal of Zoology, 10 (1), 187-189.

123

Ward, P.S., Brady, S.G., Fisher, B.L. & Schultz, T.R. (2010) Phylogeny and biogeography of dolichoderine ants: Effects of data partitioning and relict taxa on historical inference. Systematic Biology, 59, 1-21.

Wheeler, W.M. (1915) The ants of the Baltic amber. Schriften der physikalisch-ökonomischen Gesellschaft zu Königsberg, 55, 1-142.

Wiest, L. (1966) Über die Funktion des Gasters bei der Verständigung von Ameisen. PhD Thesis, University of Vienna, Wien.

Willis, K.J. & van Andel, T.H. (2004) Trees or no trees? The environments of central and eastern Europe during the last Glaciation. Quaternary Science Review, 23, 2369-2387.

Willis, K., Rudner, E. & Sümegi, P. (2000) The full-glacial forests of central and southeastern Europe. Quaternary Research, 53, 203-213.

Wysocka, A., Krzysztofiak, L., Krzysztofiak, A., Żołnierkiewicz, O., Ojdowska, E. & Sell, J. (2011) Low genetic diversity in Polish populations of sibling ant species: Lasius niger (L.) and Lasius platythorax Seifert (Hymenoptera: Formicidae). Insectes Sociaux, 58, 191-195.

Yan, Y., Harris, A.J. & Xingjing, H. (2011) RASP (Reconstructing Ancestral State in Phylogenies) 1.1 Available from http://mnh.scu.edu.cn/soft/blog/RASP.

BIOSKETCHES

Lenka Petráková has worked on the biology of Liometopum microcephalum for several years, peaking with her doctoral thesis. Currently she is also involved in molecular phylogeography, phylogeny and prey DNA detection in other arthropods. Andrea Tóthová works in taxonomy and molecular phylogenetics of insects, mainly Diptera. Jiří Schlaghamerský’s research interests lie in invertebrate ecology and soil biology, with focus on soil annelids and saproxylic invertebrates, including ants nesting in dead wood.

Author contributions: J.S. and L.P. designed the study and acquired the ant samples; A.T. specified the molecular analyses and supervised L.P. in the lab; L. P. performed the data analyses; L.P. drafted the manuscript with input from all authors, J.S. checked and corrected the final version.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article: Appendix S1 GenBank access numbers for the sequences used in phylogenetic analyses. Appendix S2 Phylogenetic relationships among Liometopum microcephalum and the related species used for the primary calibration of the molecular clock. Appendix S3 Haplotype networks for individual mitochondrial gene fragments in Liometopum microcephalum.

124

Journal of Biogeography

SUPPORTING INFORMATION

Phylogeography of the rare velvety tree ant Liometopum microcephalum (Formicidae: Dolichoderinae)

Lenka PETRÁKOVÁ, Andrea TÓTHOVÁ & Jiří SCHLAGHAMERSKÝ

Appendix S1: GenBank access numbers for the gene fragments used in phylogenetic analyses (for each single gene separately).

a) Accession numbers of related taxa Genus Species COI CytB 28S 12S Liometopum luctuosum DQ353394 - DQ353590 - occidentale - AF146721 AY867465 - sinense JQ913603 JQ913636 - - Tapinoma erraticum GU388394 KT844607 - KT844580 sessile FJ161757 - EF013066 - Linepithema humile FJ161726 AF146720 EF013003 - oblongum FJ496350 FJ496347 - - Iridomyrmex sanguineus JN134882 JN134772 FJ939799 - Azteca longiceps - AF146713 - - pittieri JQ867664 - - - Technomyrmex albipes JQ913601 JQ913623 DQ353667 - antennus JQ913599 JQ913625 - - Dolichoderus pilosus JQ913598 JQ913635 - - quadripunctatus KT844626 - - KT844581 taprobanae JQ913592 JQ913630 - - Aneuretus simonii DQ353354 - DQ353685 - Tetraponera attenuata JQ681096 JQ681151 KR828927 - rufonigra FJ436846 U02458 AY703582 - Pseudomyrmex flavicornis FJ436819 KC420094 AY703594 - dendroicus KP271184 - AY703590 -

125

b) Accession numbers of the unique haplotypes found in Liometopum microcephalum (see Table 1 and Appendix S3 for details)

site 12S rRNA cytochrome c oxidase I cytochrome b 28S 1 KT844569 KT844608 KT844582 - 2 - KT844609 KT844583 - 3 - - KT844584 - 4 - KT844610 KT844585 - 5 - KT844611 - - 6 - KT844612 KT844586 - 7 - - KT844587 - 10 - - KT844588 - 12 - - KT844589 - 13 - - KT844590 - 14 - - KT844591 - 16 KT844570 - - - 17 KT844571 KT844613 - - 18 - KT844614 KT844592 - 19 - - KT844593 - 21 - KT844615 KT844594 - 22 - KT844616 - - 23 - - KT844595 - 25 - KT844617 - - 26 - KT844618 KT844596 - 27 KT844572 KT844619 KT844597, - KT844598 28 - - KT844599 - 31 - KT844620 - - 32 KT844573 - - - 34 KT844574 - - - 35 KT844575 - - - 36 KT844577 KT844621, KT844600, - KT844622 KT844601, KT844602 37 KT844578 KT844623, KT844603, - KT844624 KT844604 38 - - KT844605 - 39 KT844579 KT844625 KT844606 KU521373 40 KT844576 - KU521374 -

126

Appendix S2: Phylogenetic relationships among Liometopum microcephalum and related taxa based on DNA sequence data (cytochrome c oxidase I, cytochrome b, 28S rRNA genes) and calculated inBEAST. Horizontal blue bars represent the 95% credibility limits for node ages, divergence time was assessed for nodes with a posterior probability node support higher than 0.7.

127

Appendix S3: Haplotype networks for individual mitochondrial genes in Liometopum microcephalum. Each circle represents a unique haplotype, circle size is proportional to haplotype frequency. Lines perpendicular to the lines connecting haplotypes present mutation steps, numbers inside the circles correspond to the sampling sites (for details see Table 1); a) 12S rRNA gene (380 bp), b) cytochrome b (684 bp), c) cytochrome c oxidase I (1314 bp).

a) b)

128

9. Conclusions

The aim of this thesis was to shed light on several questions regarding the biology and genetic diversity of the rare ant species Liometopum microcephalum. It has brought new and interesting facts about the species’ space partitioning, interactions with its competitors, worker size variability, aphid tending, relatedness of populations, and colonisation history of its present distribution area. Based on the observed defence of the nest trees, foraging trees and their surroundings, I consider L. microcephalum to be a territorial species, ranking at the top position in the hierarchy of ant communities. The combination of different approaches (pitfall trapping, mapping of territories, food bait experiments and direct observation of interactions) allowed to obtain insight into the way how the colonies interact with their neighbours. Situations where the studied species nests close to its competitors are rather sporadic, but the fact that L. microcephalum is able to coexist with strong competitors presents an important finding. The species takes advantage of worker cooperation when defending its territories. Space partitioning took place in early spring when the competitors aggressively defended large foraging trees. In spring, L. microcephalum workers were also searching for temporary food resources, moving densely on the soil surface. Later, L. microcephalum workers moved predominantly on trunk trails leading to foraging trees and all competitors avoided each other. The territory sizes changed in the course of time, being larger in spring than in summer. Regarding the worker size polymorphism, I cannot confirm the existence of clearly separated worker castes. The range of measured worker sizes was broad in all studied colonies, the sizes increased continuously from minor to major workers. Different allometric relations were found in the studied colonies, thus L.microcephalum is indeed a polymorphic species. However, I did not find any case of clear dimorphism, which would include total absence of intermediate workers. The found levels of polymorphism did not correlate with any of the studied factors but changed in the course of time. No differences in size were found in workers occupied by different task outside the nest, i.e. nest defence and foraging. Worker body size (represented by mean femur length) correlated, however, with territory area, which served as a measure of colony size and success. Workers collected in spring were bigger than those collected in summer. Large workers are generally more resistant to suboptimal conditions, thus these were probably individuals that had overwintered. The largest workers (and largest territories) were found in colonies neighbouring with colonies of the strong competitor Lasius fuliginosus. Perhaps selective pressure played a role in the production of large workers in that case. Honeydew intake was confirmed by measuring the content of reducing sugars in the gasters of workers sampled on their way to and from foraging trees. Aphid tending took place mainly in the foraging trees as was evident from the highest amount of reducing sugars in ants descending the foraging trees, but it is highly probable that it occurred also on the nest trees. The amount of both total and reducing sugars in the gasters varied between months. Sugar contents were lowest in April, peaking in June and dropping in July. The intensity of honeydew intake was not the highest at the beginning of the season (as we had expected) but increased gradually from April to June. It apparently corresponded to the aphids’ life cycles – their population densities usually increase from April to May and decrease in mid to late summer

129

(JAROŠÍK & DIXON 1999). To conclude, trophobiosis is important for the nutrition of L. microcephalum in contrast to the published statements of several renowned myrmecologists, who had considered the species to be a strict predator. However, we are still lacking data that would allow to quantify the importance of honeydew obtained by trophobiosis in comparison to the intake of other food. The used mitochondrial markers reflected the relatedness of the studied populations, allowing also time estimations of lineage divergences, whereas the used nuclear marker showed no variability across the species’ range. The common ancestor of the studied colonies was dated to the Pliocene epoch, nearly four million years ago. The Levantine population differed distinctly from the rest of the samples, being isolated probably due to high tectonic activity in that region. The species had obviously dispersed from Anatolia to the north-west, reaching the Pannonian Basin. After the cooling of the climate it was restricted to several refugia. The youngest divergences within the European clades were dated to the time shortly before the last glacial maximum. The species survived the Pleistocene glaciations not only in the two traditional refugia (Balkan and Apennine Peninsulas) but also at the Black Sea coast, within the Carpathian Arc and at the south-eastern margin of the Alps. This findings provide strong additional support to the notion that these extra-Mediterranean refugia were important for the survival of species adapted to temperate climate, including such thermophilous species as L. microcephalum, and deciduous trees (used as its nest trees) in those areas. Although we were not able to obtain samples from all sites with known occurrence of the studied species, we managed to reach a good coverage of the species’ range, including its northern, western, southern and western borders. Based on the number of haplotypes found within individual colonies, I confirmed that at least some colonies of L. microcephalum are polygynous. Although the above studies have clarified ambiguities or even uncovered unknown facts about the species’ biology, further questions have emerged or gained importance. Based on the known facts, the key factor for the species’ occurrence is the availability of nest sites in preferred types of habitat. However, to fully understand the process of colonisation of new sites, it would be necessary to know the distance that a mated female is able to overcome when dispersing from her natal nest. Similarly we know that trophobiosis is important for L. microcephalum, but it remains to quantify what importance honeydew intake has compared to other food components. We have characterized workers moving outside of the nest, but to the present day nobody has been able to study the ants inside the nest in sufficient detail. In regard to the ecological importance of this remarkable ant species, the sites where it occurs should be strictly protected to prevent further decrease in colony numbers, loss of entire populations or even its total extinction.

130

10. References

AGOSTI, D. & COLLINGWOOD, C. A. (1987): A provisional list of the Balkan ants (Hym. Formicidae) and a key to the worker caste. I. Synonymic list. Mitteilungen der Schweizerischen Entomologischen Gesellschaft 60: 51-62.

AKIMOVA, I.A. (Ed.) (2009): Chervona kniga Ukraini. Tvarinny Svit. [Red Book of Ukraine. Animals]. Globalkonsalting, Kiev, 600 pp. [in Ukrainian]

AKTAÇ, N. (1976): Studies on the myrmecofauna of Turkey I. Ants of Siirt, Bodrum and Trabzon. Istanbul Universitesi Fen Fakultesi Mecmuasi Seri B 41: 115-135.

AKTAÇ, N., ARAS, A. & ÇAMLITEPE, Y. (1994): Ants of Thracian part of Turkey. Bios 2: 203- 209.

AKTAÇ, N. & ÇAMLITEPE, Y. (1987): Bandırma kuş cenneti karınca faunası, pp. 175–179. In: II. Bandırma Kuş Cenneti ve Kuş Gölü Sempempozyumu Bildirileri Balıkesir-Bandırma, Turkey.

APOSTOLOV, L.G.  LIKHOVIDOV, V.E. (1973): Contribution to the fauna and ecology of Formicidae (Hymenoptera) from south-eastern Ukraine. Vestnik Zoologii 6: 60-66.

ARAKELIAN, G.R. (1994): Fauna Respubliki Armeniya Nasekomye Pereponchatokrylye Murav'i (Formicidae). Gitutium, Erevan, 153 pp.

ARAS, A. & AKTAÇ, N. (1987): Edirne yöresi çayır ve mera Karıncaları üzerinde faunistik araştırmalar, p. 695–703. In: Proceedings of the First Turkish National Congress of Entomology. İzmir, Turkey.

ARAS, A. & AKTAÇ, N. (1990): Trakya Bölgesi çayır ve mera karınca faunası, pp. 173–182. In: Proceeding of 17th National Congress of Biology. Atatürk Üniversity, Erzurum, Turkey.

ASTRUC, C., MALOSSE, C. & ERRARD, C. (2001): Lack of intraspecific aggression in the ant Tetramorium bicarinatum: a chemical hypothesis. Journal of Chemical Ecology 27: 1229–1248.

ATANASSOV, N. (1934): Contribution to a study of the ant fauna (Formicidae) of Bulgaria [in Bulgarian]. Izvestiya na Bulgarskoto Entomologichno Druzhestvo 8: 159–173.

ATANASSOV, N. (1936): Second contribution to a study of the ant fauna (Formicidae) of Bulgaria [in Bulgarian]. Izvestiya na Bulgarskoto Entomologichno Druzhestvo 9: 211– 236.

131

ATANASSOV, N. (1957): Recherches sur la biologie de Liometopum microcephalum (Hymenoptera, Formicidae) en Bulgarie. Bulletin de la Société Entomologique de Mulhouse 501: 43–47.

ATANASSOV, N. (1964): Studies on the systematics and ecology of ants (Formicidae, Hym.) from the Petrich region (south-west Bulgaria). [In Bulgarian]. Izvestiya na Zoologicheskiya Institut s Muzej. Bulgarska Akademiya na Naukite 15: 77–104.

ATANASSOV, N. & DLUSSKY G. M. (1992): Fauna of Bulgaria. Hymenoptera, Formicidae. Fauna Bulgarica 22: 1–310.

BARGUM, K., HELANTERÄ, H. & SUNDSTRÖM, L. (2007): Genetic population structure, queen supersedure and social polymorphism in a social Hymenoptera. Journal of Evolutionary Biology 20: 1351–1360.

BARONI URBANI, C. (1964): Formiche dell`Italia Appenninica (Studi sulla mirmecofauna d´Italia, III). Memorie del Museo Civico di Storia Naturale di Verona 12: 149–172.

BENSASSON, D., ZHANG, D.-X., HARTL, D.L. & HEWITT, G.M. (2001): Mitochondrial pseudogenes: Evolution´s misplaced witnesses. Trends in Ecology and Evolution 16: 314–321.

BERNASCONI, C., MAEDER, A., CHERIX, D. & PAMILO, P. (2005): Diversity and genetic structure of the wood ant Formica lugubris in unmanaged forests. Annales Zoologici Fennici 42: 189–199.

BERNARD, F. (1968): Les Fourmis (Hymenoptera Formicidae) d’Europe occidentale et septentrionale. Masson et Cie éditeurs, Paris, 411 pp.

BEZDĚČKA, P. (1995): Hymenoptera: Formicoidea. Folia Facultatis Scientiarium Naturalium Universitatis Masarykianae Brunensis, Biologia 93: 323–329.

BEZDĚČKA, P. (1996a): První příspěvek k poznání mravenců jihovýchodní Moravy. Sborník Přírodovědeckého klubu v Uherském Hradišti 1: 70–74.

BEZDĚČKA, P. (1996b): Mravenci Slovenska (Hymenoptera: Formicidae). Entomonofauna Carpathica 8: 108–114.

BHAGWAT, S.A. & WILLIS, K.J. (2008): Species persistence in northerly glacial refugia of Europe: a matter of chance or biogeographical traits? Journal of Biogeography 35: 464– 482.

BHARTI, H., GUÉNARD, B., BHARTI, M. & ECONOMO, E.P. (2016): An updated checklist of the ants of India with their specific distributions in Indian states (Hymenoptera, Formicidae). ZooKeys 51: 1–83.

132

BITSCH, J. & LECLERCQ, J. (1993): Hyménoptères Sphecidae d’Europe occidentale. Volume 1. Généralités – Crabroninae. Faune de France 79: 1–325.

BOLTON, B. (2014): An online catalog of the ants of the world. Available from http://antcat.org. (accessed on 25. 2. 2016)

BOROWIEC, L. (2014): Catalogue of ants of Europe, the Mediterranean Basin and adjacent regions (Hymenoptera: Formicidae). Genus 25: 1–340.

BOROWIEC, L. & SALATA, S. (2012): Ants of Greece – checklist, comments and new faunistic data (Hymenoptera: Formicidae). Genus 23: 461–563.

BRAČKO, G. (2006): Review of the ant fauna (Hymenoptera: Formicidae) of Croatia. Acta Entomologica Slovenica 14: 131–156.

BRAČKO, G. (2007): Checklist of the ants of Slovenia (Hymenoptera: Formicidae). Natura Sloveniae 9: 15–24.

BRAČKO, G., GOMBOC, M., LUPŠE, B., MARIĆ, R., PRISTOVŠEK, U. (2014a): New faunistic data on ants (Hymenoptera: Formicidae) of the southern part of Montenegro. Natura Sloveniae 16: 41–51.

BRAČKO, G., WAGNER, H. C., SCHULZ, A., GIOHAIN, E., MATIČIČ, J. & TRATNIK, A. (2014b): New investigation and a revised checklist of the ants (Hymenoptera: Formicidae) of the Republic of Macedonia. North-Western Journal of Zoology 10: 10–24.

BREWER, S., CHEDDADI, R., DE BEAULIEU, J.L. & REILLE, M. (2002): The spread of deciduous Quercus throughout Europe since the last glacial period. Forest Ecology and Management 156: 27–48.

BROWN, M.J.F. & BONHOEFFER, S. (2003): On the evolution of claustral colony founding in ants. Evolutionary Ecology Research 5: 305–313.

BUCZKOWSKI, G. & BENNET, G. (2009): Colony budding and its effects on the food allocation in the highly polygynous ant, Monomorium pharaonis. Ethology 115: 1091–1099.

ÇAMLITEPE, Y. & AKTAÇ, N. (1987): Trakya bölgesi orman karınca faunası üzerinde araştırmalar, p. 685–694. In: Proceedings of the First Turkish National Congress of Entomology, İzmir, Turkey.

CASTELLANI, O. (1937): Contributo alla conoscenza della fauna entomologica del Lazio. Hymenoptera – Formicidae. Bollettino della Societa Veneziana di Storia Naturale 1: 179– 183.

CECCONI, G. (1903): Illustrazioni di guasti operati da animali su piante legnose italiane. Le Stazione Sperimentali Agrarie Italiane 36: 1–37.

133

CHAPUISAT, M. (1996): Characterization of microsatellite loci in Formica lugubris B and their variability in other species. Molecular Ecology 5: 599–601.

CHAPUISAT, M. (1998): Mating frequency of ant queens with alternative dispersal strategies, as revealed by microsatellite analysis of sperm. Molecular Ecology 7: 1097–1105.

CHAPUISAT, M., BOCHERENS, S. & ROSSET, H. (2004): Variable queen number in ant colonies: no impact on queen turnover, inbreeding and population genetic differentiation in the ant Formica selysi. Evolution 58: 1064–1072.

COLLINGWOOD, C. A. (1978): A provisional list of Iberian Formicidae with a key to the worker caste (Hym. Aculeata). EOS Revista española de entomología 52: 65–95

CONCONI, J.R.E. DE, FLORES, P.M.J., PEREZ, J.A., CUEVAS, G.L., SANDOVAL, C.S., GARDUNO, C.E., PORTILLO, I., DELAGE-DARCHEN, T. & DELAGE-DARCHEN, B. (1987): Colony structure of Liometopum apiculatum and Liometopum occidentale var. luctuosum W. In: Eder, J. & Rembold, H. (Eds.), Chemistry and Biology of Social Insects, Verlag J. Peperny, München, Germany, pp. 671–673.

CONSANI, M. (1949): Formiche raccolte nell´Apennino Abruzzese dal Sig. Pio Bisleti. Bollettino dell´Associazone Romana di Entomologia 4: 11–12.

CRUZ-LABANA, J. D., TARANGO-ARAMBULA, L. A., ALCANTARA-CARBAJAL, J. L., PIMENTEL- LOPEZ, J., UGALDE-LEZAMA, S., RAMIREZ-VALVERDE, G., MENDEZ-GALLEGOS, S.J. (2014): Habitat use by the “escamolera” ant (Liometopum apiculatum Mayr) in central Mexico. Agrociencia 48: 569–582.

CSÖSZ, S. (2012): Nematode infection as significant source of unjustified taxonomic descriptions in ants (Hymenoptera: Formicidae). Myrmecological News 17: 27–31.

DALECKY, A., GAUME, L., SCHATZ, B., MCKEY, D. & KJELLBERG, F. (2005): Facultative polygyny in the plant-ant Petalomyrmex phylax (Hymenoptera: Formicinae): sociogenetic and ecological determinants of queen number. Biological Journal of the Linnean Society 86: 133–151.

DEBOUT, G., VENTELON-DEBOUT, M., EMERSON, B.C. & YU, D.W. (2006): PCR primers for polymorphic microsatellite loci in the plant-ant Azteca ulei cordiae (Formicidae: Dolichoderinae). Molecular Ecology Notes 7: 607–609.

DEBOUT, G., SCHATZ, B., ELIAS, M. & MCKEY, D. (2007): Polydomy in ants: what we know, what we think we know and what remains to be done. Biological Journal of the Linnean Society 90: 319–348.

DEL TORO, I., PACHECO, J.A. & MACKAY, W.P. (2009): Revision of the ant genus Liometopum (Hymenoptera: Formicidae). Sociobiology 53: 299–369.

DE STEFANI, T. (1888): Miscellanea imenoterrologica sicula. Naturalista Siciliano 8: 142–145.

134

D´ETTORE, P., HEINZE, J. & RATNIEKS, F.L. (2004): Worker policing by egg eating in the ponerine ant Pachycondyla inversa. Proceedings of the Royal Society - Biological Science 271: 1427–1434.

DLUSSKY, G.M. & PUTYATINA, T.S. (2014): Early Miocene ants (Hymenoptera, Formicidae) from Radoboj, Croatia. Neues Jahrbuch für Geologie and Paläontologie Abhandlungen 272: 237–285.

DOFLEIN, F. (1920): Mazedonische Ameisen, Beobachtungen über ihre Lebensweise. Gustav Fischer Verlag, Jena, pp. 158–205.

DONISTHORPE, H. (1926): The ants (Formicidae), and some myrmecophiles, of Sicily. Entomologist´s Record and Journal of Variation 38: 161–165.

EMERY C. (1869): Enumerazione dei Formicidi che rinvengonsi nei contorni di Napoli con descrizioni di specie nuove o meno conosciute. Annali dell´Accademia degli Aspiranti Naturalisti di Napoli 2: 1–26.

EMERY, C. (1878): Catalogo delle formiche esistenti nelle collezioni del Museo Civico di Genova. Parte seconda. Formiche dell´Europa e delle regioni limitrofe in Africa e in Asia. Annali del Museo Civico di Storia Naturale di Genova 12, 43–59.

EMERY, C. & CAVANNA, G. (1880): Escursione in Calabria (1877- 1878). Formicidei. Bullettino della Societa Entomologica Italiana 12: 123–126.

EMERY, C. (1891): Zur Biologie der Ameisen. Biologisches Zentralblatt 11: 165–180.

EMERY, C. (1912): Hymenoptera, fam. Formicidae, subfam. Dolichoderinae. In: WYTSMAN, P. (Ed.): Genera Insectorum 137. V. Verteneuil & L. Desmet, Bruxelles, Belgium, pp. 1– 50.

EMERY, C. (1915): Contributo alla conoscenza delle formiche delle isole italiani. Descrizioni di forme mediterranee nuove o critiche. Annali del Museo Civico di Storia Naturale Giacomo Doria (Genova) 46: 244–270.

EMERY, C. (1916): Fauna Entomologica Italiana. I. Hymenoptera - Formicidae. Bulletino della Societa Entomologica Italiana 47: 79–275.

EVANS, J. (1996) Queen longevity, queen adoption and posthumous indirect fitness in the facultatively polygynous ant Myrmica tahoensis. Behavioral Ecology and Sociobiology 39: 275–284.

FOREL, A. (1892): Die Ameisenfauna Bulgariens. (Nebst biologischen Beobachtungen.) Verhandlungen der Zoologisch-Botanischen Gesellschaft in Wien 42: 305–318.

135

FOREL, A. (1909): Etudes myrmécologiques en 1909. Fourmis de Barbarie et de Ceylan. Nidification des Polyrhachis, Bulletin de la Société Vaudoise des Sciences Naturelles 45: 369–407.

FOURNIER, D., BATTAILLE, G., TIMMERMANS, I. & ARON, S. (2008): Genetic diversity, worker size polymorphism and division of labour in the polyandrous ant Cataglyphis cursor. Animal Behaviour 75: 151–158.

FRUMHOFF, P. C. & WARD, P. S. (1992): Individual-level selection, colony-level selection, and the association between polygyny and worker monomorphism in ants. The American Naturalist 139: 559–590.

GALLÉ, L. (1981): The formicoid fauna of the Hortobágy. In: MAHUNKA, S. (ed.) The fauna of the Hortobágy National Park. Akadémiai Kiadó, Budapest, 415 pp.

GALLÉ, L., MARKÓ, B., KISS, K., KOVÁCS, É., DÜRGŐ, H., KŐVÁRY, K. & CSŐSZ, S. (2005): Ant fauna of Tisza river basin (Hymenoptera: Formicidae). In: GALLÉ, L. (Ed.): Vegetation and Fauna of Tisza River Basin I. Tiscia Monograph Series 7: 149-197.

GOLDSTEIN, D.B. & SCHLÖTTERER, C. (eds.) (1999): Microsatellites: Evolution and Applications. Oxford University Press, 368 pp.

GRABENWEGER, G., KEHRLI, P., SCHLICK-STEINER, B., STEINER, F., STOLZ, M. & BACHER, S. (2005): Predator complex of the horse chestnut leafminer Cameraria ohridella: identification and impact assessment. Journal of Applied Entomology 129: 353-362.

GRANDI, G. (1935): Contributi alla conoscenza degli Imenotteri Aculeati. XV. Bollettino dell`Istituto di Entomologia della Università di Bologna 8: 27-121.

GRATIASHVILI, N. & BARJADZE, S. (2008): Checklist of the ants (Formicidae Latreille, 1809) of Georgia. Proceedings of the Institute of Zoology, Tbilisi XXIII: 130-146.

GREBENNIKOV K. A., DUBOVIKOFF D. A. & SAVRANSKAYA Z. V. (2007): The ants (Hymenoptera: Formicidae) of the lower Volga region. retrieved on 16 August 2007

GUÉNARD, B. & DUNN, R. R. (2012): A checklist of the ants of China. Zootaxa 3358: 1–77.

GYLLENSTRAND, N. & SEPPÄ, P. (2003): Conservation genetics of the wood ant, Formica lugubris, in a fragmented landscape. Molecular Ecology 12: 2931–2940.

HALMÁGYI, L. (1968): Die Blattläuse (Aphidoidea) der Edelkastanie (Castanea sativa Mill.) in Ungarn. Opuscula Zoologica 8: 3-9.

HAMILTON, W. D. (1964): The genetical evolution of social behaviour. Journal of Theoretical Biology 7: 1–16.

136

HANNONEN, M. (2002): Proximate and ultimate determinants of reproductive skew in the polygyne ant Formica fusca. Ph.D. Thesis, University of Helsinki, Finland, 20 pp.

HANNONEN, M. & SUNDSTRÖM, L. (2003): Reproductive sharing among queens in the ant Formica fusca. Behavioral Ecology 14: 870–875.

HASKINS, C. P. (1941): Notes on the method of colony foundation of the Ponerine ant Bothroponera soros Emery. Journal of the New York Entomological Society 49: 211–216.

HERBERT, P.D.N., RATNASINGHAM, S. & DEWAARD, J.R. (2003): Barcoding animal life: cytochrome c oxidase subunit 1divergences among closely related species. Proceedings of the Royal Society of London B: Biological Sciences 270: S96–S99.

HEER, O. (1849): Die Insektenfauna der Tertiärgebilde von Oeningen und von Radoboj in Croatien. Zweiter Theil: Heuschrecken, Florfliegen, Aderflügler, Schmetterlinge und Fliegen. Leipzig: W. Engelmann, 154 pp.

HEER, O. (1865): Die Urwelt der Schweiz. Friedrich Schulthess, Zürich, pp. 289–496.

HEER, O. (1867): Fossile Hymenopteren aus Oeningen und Radoboj. Neuen Denkschriften der allgemeinem schweizerischen naturforschenden Gesellschaft für die gesamten Naturwissenschaften 22: 1–42.

HEINZE, J. & TSUJI, K. (1995): Ant reproductive strategies. Researches on Population Ecology 37: 135–149.

HELANTERÄ, H. & SUNDSTRÖM, L. (2007): Worker reproduction in Formica ants. American Naturalist 170: E15–E25.

HETERICK, B.E. & SHATTUCK, S.O. (2011): Revision of the ant genus Iridomyrmex (Hymenoptera: Formicidae). Zootaxa 2845: 1–174.

HOEY-CHAMBERLAIN, R., RUST, M.K. & KLOTZ, J.H. (2013): A Review of the biology, ecology and behavior of velvety tree ants of North America. Sociobiology 60: 1–10.

HOEY-CHAMBERLAIN, R. & RUST, M.K. (2014): Food and bait preferences of Liometopum occidentale (Hymenoptera: Formicidae). Journal of Entomological Science 49: 30–43.

HÖLLDOBLER, B. & WILSON, E. O. (1990): The ants. Harvard Univ. Press, Cambridge, MA, 732 pp.

HOLZER, B., KELLER, L. & CHAPUISAT, M. (2009): Genetic clusters and sex-biased gene flow in a unicolonial Formica ant. BMC Evolutionary Biology 31: 9–69.

HUGHES, W.O.H., RATNIEKS, F.L.W. & OLDROYD D.P. (2008): Multiple paternity or multiple queens: two routes to greater intracolonial genetic diversity in the eusocial Hymenoptera. Journal of Evolutionary Biology 21: 1090–1095.

137

JAROŠÍK, V. & DIXON, A.F.G. (1999): Population dynamics of a tree-dwelling aphid: regulation and density independent processes. Journal of Animal Ecology 68: 726-732.

JIJILASHVILI, T.I. (1964): Ecological-faunistic characteristic of the ant fauna of the steppes zone of Georgia [in Russian]. Soobshcheniya Akademii Nauk Gruzii 34: 651–657.

JOHNSON, R.A. (2002): Semi-claustral colony founding in the seed-harvester ant Pogonomyrmex californicus: a comparative analysis of colony founding strategies. Oecologia 132: 60–67.

JUŘIČKOVÁ, L., HORÁČKOVÁ, J. & LOŽEK, V. (2014): Direct evidence of central European forest refugia during the last glacial period based on mollusc fossils. Quaternary Research 82: 222–228.

KARAMAN, M. & KARAMAN, G. S. (2003): Contribution to the knowledge of the ants (Hymenoptera, Formicidae) from Serbia. The Montenegrin Academy of Sciences and Arts. Glasnik of the Section of Natural Sciences 15: 39-58.

KARAMAN, M. (2009): An introduction to the ant fauna of Macedonia (Balkan Peninsula), a check list (Hymenoptera, Formicidae). Natura Montenegrina 8: 151-162.

KARAMAN, M. G. (2011): A catalogue of the ants (Hymenoptera, Formicidae) of Montenegro [in Montenegrin]. Montenegrin Academy of Sciences and Arts, Podgorica, Montenegro, 140 pp.

KIRAN, K. & AKTAÇ, N. (2006): The Vertical Distribution of the Ant Fauna (Hymenoptera: Formicidae) of the Samanlı Mountains, Turkey. Linzer Biologische Beiträge 38: 1105– 1122.

KIRAN, K. & KARAMAN, M. (2012): First annotated checklist of the ant fauna of Turkey (Hymenoptera, Formicidae). Zootaxa 3548: 1–38.

KIRCHNER, L. (1867): Catalogus Hymenopterorum Europae. K. K. Zoologisch-Botanische Gesellschaft, Wien, 285 pp.

KLOTZ, J., HANSEN, L., FIELD, H., RUST, M., OI, D. & KUPFER, K. (2010): Urban pest management of ants in California. University of California, Agricultural and Natural Resources, Davis, CA, 72 pp.

KNADEN, M., TINAUT, A., CERDA, X., WEHNER, S. & WEHNER, R. (2005): Phylogeny of three parapatric species of desert ants, Cataglyphis bicolor, C. viatica and C. savignyi: A comparison of mitochondrial DNA, nuclear DNA, and morphological data. Zoology 108: 169–177.

KORLEVIĆ, A. (1890): Prilozi fauni hrvatskih opnokrilaca. Glasnik Hrvatskoga Naravnoslovnoga Družstva 5: 189–250.

138

KRÁSA, A. (2014): Nová lokalita mravence lužního. Živa 1: 33–34.

KRATOCHVÍL, J. (1936): Rozbor mravenčí zvířeny Pavlovských vrchů. Práce Moravské přírodovědecké společnosti 10: 1–30.

KRATOCHVÍL, J. (1938): Myrmekologické poznámky 1–2. Sborník klubu přírodovědného v Brně 2: 161.

KUGLER, J. (1988): The zoogeography of social insects of Israel and Sinai. Pp. 251–275. In: YOM-TOV, Y. & TCHERNOV, E. (eds.): The zoogeography of Israel. Dr W Junk Publishers, Dordrecht, 600 pp.

KUPYANSKAYA, A.N. (1988): Far Eastern representatives of the genus Liometopum (Hymenoptera, Formicidae). Vestnik Zoologii 1: 29–34.

LAPEVA-GJONOVA, A. (2004): Ants (Hymenoptera: Formicidae) from the Eastern Rhodopes (Bulgaria). In: BERON, P. & POPOV, A. (eds.) Biodiversity of Bulgaria. 2. Biodiversity of Eastern Rhodopes (Bulgaria and Greece). Pensoft Publishers & National Museum of Natural History, Sofia, pp. 507–513.

LAPEVA-GJONOVA, A., ANTONOVA, V., RADCHENKO, A. G. & ATANASOVA, M. (2010): Catalogue of the ants (Hymenoptera: Formicidae) of Bulgaria. ZooKeys 62: 1–124.

LAPEVA-GJONOVA, A. (2011): Ants (Hymenoptera, Formicidae) from Western Rhodope Мountains (Bulgaria). In: BERON P. (ed). Biodiversity of Bulgaria. 4. Biodiversity of Western Rhodopes (Bulgaria and Greece). Pensoft & National Museum of Natural History, Sofia, Bulgaria, pp. 287–298.

LAPEVA-GJONOVA, A. & KIRAN, K. (2012): Ant fauna (Hymenoptera, Formicidae) of Strandzha (Istranca) Mountain and adjacent Black Sea coast. North-Western Journal of Zoology 8: 72–84.

LAPOLLA, J.S., DLUSSKY, G.M. & PERRICHOT, V. (2013): Ants and the fossil record. Annual Review of Entomology 58: 609–630.

LARA-JUÁREZ, P., AGUIRRE RIVIERA, J.R., CASTILLO LARA, P. & REYES AGÜERO (2015): Biología y aprovechamiento de la hormiga de escamoles, Liometopum apiculatum Mayr (Hymenoptera: Formicidae). Acta Zoológica Mexicana 31: 251–264.

LEGAKIS, A. (1984): The Zoological Museum of the University of Athens 2. The collection of ants from Greece. Biologia Gallo-Hellenica 11: 85-87.

LEPPÄNEN, J., VEPSÄLÄINEN, K., ANTHONI, H. & SAVOLAINEN, R. (2013): Comparative phylo-geography of the ants Myrmica ruginodis and Myrmica rubra. Journal of Biogeography 40: 479–491.

139

MAKAREVICH, O. N. (2003): Liometopum microcephalum (Hymenoptera: Formicidae) in the Lower Dnepr. Vestnik Zoologici 37: 51–56.

MANTERO, G. (1889): Res Ligusticae XXX. Materiali per un catalogo degli Imenoterri liguri. Parte I. Formicidi. Annali del Museo Civico di Storia Naturale di Genova 19: 146–160.

MARKIN, G. P., COLLINS, G. J. & DILLIER, J. H. (1972): Colony founding by queens of the red imported fire ant Solenopsis invicta. Annals of the Entomological Society of America 65: 1053–1058.

MARKÓ, B. & CSÖSZ, S. (2002): Die europäischen Ameisenarten (Hymenoptera: Formicidae) des Hermannstädter (Sibiu, Romania) Naturkundemuseums I.: Unterfamilien Ponerinae, Myrmicinae und Dolichoderinae. Annales Historico-Naturales Musei Nationalis Hungarici 94: 109–121.

MARTÍNEZ, M. D. & TINAUT, A. (2001): Sobre la presencia en la Península Ibérica de Liometopum microcephalum (Panzer, 1798) (Hymenoptera. Formicidae). Boletín de la Asociación española de Entomología 25: 123–124.

MATHEW, R. & TIWARI, R.N. (2000): Insecta: Hymenoptera: Formicidae. Zoological Survey of India. Fauna of Meghalaya 7: 251–409.

MAYR, G. (1855): Formicina austriaca. Beschreibung der bisher im österreichischen Kaiserstaate aufgefundenen Ameisen, nebst Hinzufügung jener in Deutschland, in der Schweiz und in Italien vorkommenden Arten. Verhandlungen der Zoologisch- Botanischen Gesellschaft in Wien 5: 273–478.

MCGLYNN, T. P. (2010): Polygyny in thief ants responds to competition and nest limitation but not food resources. Insectes Sociaux 57: 23–28.

MCGLYNN, T.P. & OWEN, J.P. (2002): Food supplementation alters caste allocation in a natural population of Pheidole flavens, a dimorphic leaf-litter dwelling ant. Insectes Sociaux 49: 8–14.

MCGLYNN, T.P., DIAMOND, S.E. & DUNN, R.R. (2012): Tradeoffs in the evolution of caste and body size in the hyperdiverse ant genus Pheidole. PLoS ONE 7: e48202.

MEI, M. (2016): Records of Tracheliodes varus (Panzer) and T. curvitarsis (Herrich-Schaeffer) from Central Italy (Hymenoptera, Crabronidae). Ampulex 8: 16–19.

MEUNIER, F. (1917): Sur quelques insectes de l'Aquitainien de Rott, Sept Montagnes (Prusse rhénane). Verhandelingen der Koninklijke Akademie van Wetenschappen Amsterdam, Tweede Sectie 20: 3–17.

MENOZZI, C. (1918): Primo contributo alla conoscenza della fauna mirmecologica del Modenese. Atti della Societa dei Naturalisti e Matematici di Modena 4: 81–88.

140

MENOZZI, C. (1921): Formiche dei dintorni di Sambiase di Calabria. Bolletino del Laboratorio di Zoologia Generale e Agraria della Facolta Agraria in Portici 15: 24–32.

MENOZZI, C. (1924): Res Mutinenses. Formicidae (Hymenoptera). Atti della Societa dei Naturalisti e Matematici di Modena 3: 22–47.

MENOZZI, C. (1942): Esplorazione entomologica del Parco Nazionale del Circeo. Hymenoptera – Formicidae. M. Spadafora, Salerno, 8 pp.

MOCSÁRY, A. (1918): Ordo Hymenoptera In: PASZLAVSKY, J.: Fauna Regni Hungariae. Regia Societas Scientiarum Naturalium Hungarica, Budapest, pp. 7–113.

MOREAU, C.S., BELL, C.D., VILA, R., ARCHIBALD, S.B. & PIERCE, N.E. (2006): Phylogeny of the ants: Diversification in the age of angiosperms. Science 312: 101–104.

MÜLLER, G. (1921): Due nuove formiche della regione adriatica. Bolletino della Societa Adriatica di Scienze Naturale in Trieste 27: 46–49.

MÜLLER, G. (1923): Le formiche della Venezia Giulia e della Dalmazia. Bolletino della Societa Adriatica di Scienze Naturale in Trieste 28: 11–180.

NASONOV, N. V. (1889): Contribution to the natural history of the ants primarily of Russia. 1. Contribution to the ant fauna of Russia. Izvestiya Imperatorskago Obshchestva Lyubitelei Estestvoznaniya, Antropologii i Etnografii 58: 1–78 [in Russian].

OMELKOVÁ, M. (2005): Biologie a rozšíření mravence Liometopum microcephalum PANZER, 1798. Master Thesis, Department of Botany and Zoology, Faculty of Science, Masaryk University, Brno, 143 pp.

PAKNIA, O. & KAMI, H. G. (2007): New and additional record for the formicid (Hymenoptera: Insecta) fauna of Iran. Zoology in the Middle East 40: 85–90.

PAKNIA, O., RADCHENKO, A., ALIPANAH, H. & PFEIFFER, M. (2008): A preliminary checklist of the ants (Hymenoptera: Formicidae) of Iran. Myrmecological News 11: 151–159.

PAMILO, P. (1982): Genetic population structure in polygynous Formica ants. Heredity 48: 95– 106.

PASSERA, L., RONCIN, E., KAUFMANN, B. & KELLER, L. (1996): Increased soldier production in ant colonies exposed to intraspecific competition. Nature 379: 630–631.

PAULS, S.U., THEISSINGER, K., UJVAROSI, L., BALINT, M. & HAASE, P. (2009): Patterns of population structure in two closely related, partially sympatric caddisflies, in eastern Europe: historic introgression, limited dispersal, and cryptic diversity. Journal of the North American Benthological Society 28: 517–536.

141

PELLEGRINO, I., NEGRI, A., CUCCO, M., MUCCI, N., PAVIA, M., ŠÁLEK, M., BOANO, G. & RANDI, E. (2014): Phylogeography and Pleistocene refugia of the little owl Athene noctua inferred from mtDNA sequence data. International Journal of Avian Science 156: 639– 657.

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2011) Interaction between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence. Community Ecology 12: 9–17.

PETROV, Z.I. & COLLINGWOOD, C.A. (1992): Survey of the Myrmecofauna (Formicidae, Hymenoptera) of Yugoslavia. Archives of Biological Science Belgrade 44: 79–91.

PETROV, Z. I. (2002): Contribution to the myrmecofauna (Formicidae, Hymenoptera) of Vojvodina (Serbia). Archives of Biological Science Belgrade 54: 27–28.

PILOT, M., BRANICKI, W., JĘDRZEJEWSKI, W., GOSZCZYŃSKI, J., JĘDRZEJEWSKA, B., DYKYY, I., SHKVYRYA, M. & TSINGARSKA, E. (2010): Phylogeographic history of grey wolves in Europe. BMC Evolutionary Biology 10: 104.

PINHEIRO, J. C. & BATES, D. M. (2000): Mixed-Effects Models in S and S-PLUS. Statistics and Computing Series, Springer-Verlag, New York, NY, 530 pp.

POLDI, B., MEI, M., & RIGATO, F. (1995): Hymenoptera Formicidae, pp. 1–10. In: MINELLI, A., RUFFO, S. & LA POSTA, S. (eds.) Checklist delle Specie della Fauna Italiana. Calderini, Bologna.

POPESCU, I.E. & DAVIDEANU, A. (2009): Conservation status of protected or rare invertebrates from the border area Romania – Republic of Moldova. Advances in Environmental Sciences 1: 43–53.

QUELLER, D.C. & GOODNIGHT, K.F. (1989): Estimating relatedness using genetic markers. Evolution 43: 258–275.

RADCHENKO, A. (2005): Monographic revision of the ants (Hymenoptera: Formicidae) of North Korea. Annales Zoologici 55: 127–221.

RADCHENKO, A. (2011): Zonal and zoogeographic characteristics of the ant fauna (Hymenoptera, Formicidae) of Ukraine. Vestnik Zoologii 45: e30–e39.

REES, S. D., ORLEDGE, G. M., BRUFORD, M. W., BOURKE, A. F. G. (2010): Genetic structure of the black bog ant (Formica picea Nylander) in the United Kingdom. Conservation Genetics 11: 823–834.

ROSENGREN, R. & PAMILO, P. (1983): The evolution of polygyny and polydomy in mound- building Formica ants. Acta Entomologica Fennica 42: 65–77.

142

ROSSET, H. & CHAPUISAT, M. (2007): Alternative life-histories in a socially polymorphic ant. Evolutionary Ecology 21: 577–588.

RUZSKY, M. (1905): The ants of Russia. (Formicariae Imperii Rossici). Systematics, geography and data on the biology of Russian ants. Part I. Trudy Obshchestva Estestvoispytateley pri Imperatorskom Kazanskom Universitete 38: 1–800 [in Russian].

SAVRANSKAJA, Z. V. (1998): O nachodke Liometopum microcephalum na territorii Kalmykii. Problemy zochranenija bioraznoobrazija aridnych regionov Rosii. Volgogradskij Gosudarstvennyj Universitet, Volgograd, Russia, pp. 145–146.

SCHLAGHAMERSKÝ, J. & OMELKOVÁ, M. (2007): The present distribution and nest tree characteristics of Liometopum microcephalum (Panzer, 1798) (Hymenoptera: Formicidae) in South Moravia. Myrmecological News 10: 85–90.

SCHLAGHAMERSKÝ, J., KAŠPAR, J., PETRÁKOVÁ, L. & ŠUSTR, V. (2013): Trophobiosis in the arboricolous ant Liometopum microcephalum (Hymenoptera: Formicidae: Dolichoderinae). European Journal of Entomology 110: 231–239.

SCHLAGHAMERSKÝ, J. & PETRÁKOVÁ, L. (2014): Mravenec lužní – Mýty a fakta. Živa 5/2014: 230–233.

SCHLICK-STEINER B. C.  STEINER F. M.  SCHÖDL S. (2003): Ameisen (Hymenoptera: Formicidae). Eine Rote Liste der in Niederösterreich gefährdeten Arten. Amt der NÖ Landesregierung / Abteilung Naturschutz, St. Pölten, 75 pp.

SCHLICK-STEINER, B. C., STEINER, F. M., STAUFFER, C. & BUSCHINGER, A. (2005): Life history traits of a European Messor harvester ant. Insectes Sociaux 52: 360–365.

SCHMIDT, A. M., LINKSVAYER, T. A., BOOMSMA, J. J. & PEDERSEN, J. S. (2010): Queen-worker caste ratio depends on colony size in the pharaoh ant (Monomorium pharaonis). Insectes Sociaux 58: 139–144.

SHATTUCK, S.O. (1992): Generic revision of the ant subfamily Dolichoderinae (Hymenoptera: Formicidae). Sociobiology 21: 1–176.

SEIFERT, B. (2007): Die Ameisen Mittel- und Nordeuropas. Iutra – Verlags– und Vertriebsge- sellschaft, Görlitz / Tauer, 368 pp.

SONG, H., BUHAJ, J.E., WHITING, M.F. & CRANDALL, K.A. (2008): Many species in one: DNA baroding overestimates the number of species when nuclear mitochondrial pseudogenes are coamplified. Proceedings of the National Academy of Sciences of the United States of America 105: 13486–13491.

SOUDEK, Š. (1922): Mravenci, soustava, zeměpisné rozšíření, oekologie a určovací klíč mravenců žijících na území Československé republiky, Česká společnost entomologická, Praha, 98 pp.

143

SOUZA, D.J., DELLA LUCIA, T.M.C. & LIMA, E.R. (2005): Queen adoption in colonies of the leaf-cutting ant Acromyrmex subterraneus molestans (Hymenoptera: Formicidae). Behavioural Processes 70: 62-68.

STITZ, H. (1939): Hautflügler oder Hymenoptera I: Ameisen oder Formicidae. In DAHL, F., DAHL, M. & BISCHOF, H. (Eds.): Die Tierwelt Deutschlands und der angrenzenden Meersteile nach ihren Merkmalen und nach ihrer Lebensweise 37. Gustav Fischer, Jena, Germany, 428 pp.

STUART, R.J., GRESHAM-BISSETT, L. & ALLOWAY, T.M. (1993): Queen adoption in the polygynous and polydomous ant, Leptothorax curvispinosus. Behavioral Ecology 4: 276– 281.

SUNDSTRÖM, L., SEPPÄ, P. & PAMILO, P. (2005): Genetic population structure and dispersal patterns in Formica ants – a review. Annales Zoologici Fennici 42: 163–177.

TARTALLY, A. (2006): Long term expansion of a supercolony of the invasive garden ant, Lasius neglectus (Hymenoptera: Formicidae). Myrmecologische Nachrichten 9: 21–25.

TOHMÉ, G. & TOHMÉ, H. (2014): Nouvelle liste des espèces de fourmis du Liban (Hymenoptera, Formicoidea). Lebanese Science Journal 15: 133–141.

TRABALON, M., PLATEAUX, L., PÉRU, L., BAGNÈRES, A.G. & HARTMANN, N. (2000): Modification of morphological characters and cuticular compounds in worker ants Leptothorax nylanderi induced by endoparasites Anomotaenia brevis. Journal of Insect Physiology 46: 169–178.

TSCHINKEL, W. R. & HOWARD, D. F. (1983): Colony founding by pleometrosis on the fire ant, Solenopsis invicta. Behavioral Ecology and Sociobiology 12: 103–113.

TSCHINKEL, W.R., MIKHEYEV, A.S. & STORZ, S. (2003): Allometry of workers of the fire ant, Solenopsis invicta. Journal of Insect Science 3: 2.

VAN ANDEL T.H. & TZEDAKIS P.C. (1996): Palaeolithic landscapes of Europe and environs: 150,000-25,000 years ago: an overview. Quaternary Science Reviews 15: 481–500.

VANDENBERGHE, J., RENSSEN, H., ROCHE, D.M., GOOSSE, H., VELICHKO, A.A., GORBUNOV, A. & LEVAVASSEUR, G. (2012): Eurasian permafrost instability constrained by reduced sea-ice cover. Quaternary Sciences Reviews 34: 16–23.

VEPSÄLÄINEN, K. & SAVOLAINEN, R. (1990): The effect of interference by Formicinae ants on the foraging of Myrmica. Journal of Animal Ecology 59: 643–654.

VEPSÄLÄINEN, K. & PISARSKI, B. (1982): Assembly of island ant communities. Annales Zoologii Fennici 19: 327–335.

144

VITIKAINEN, E., HAAG-LIAUTAAD, C. & SUNDSTRÖM, L. (2011): Inbreeding and reproductive investment in the ant Formica exsecta. Evolution 65: 2026–2037.

VOGRIN, V. (1955): Prilog fauni Hymenoptera-Aculeata Jugoslavie. Zaštita Bilja 31 (Prilozi entomofauni Jugoslavie, dodatak): 1–74.

VONSHAK, M. & IONESCU-HIRSCH, A. (2009): A checklist of the ants of Israel (Hymenoptera: Formicidae). Israel Journal of Entomology 39: 33–55.

WANG, T.B., PATEL, A., VU, F. & NONACS, P. (2010): Natural history observations on the velvety tree ant (Liometopum occidentale): unicoloniality and mating flights. Sociobiology 55: 787–794.

WARD, P.S., BRADY, S.G., FISHER, B.L. & SCHULTZ, T.R. (2010): Phylogeny and biogeography of dolichoderine ants: Effects of data partitioning and relict taxa on historical inference. Systematic Biology 59: 1–21.

WHEELER, D.E. (1991): The developmental basis of worker caste polymorphism in ants. The American Naturalist 138: 1218–1238.

WHEELER, W. M. (1915): The ants of the Baltic Amber. Schriften der Physikalisch- Ökonomischen Gesellschaft zu Königsberg in Preussen 55: 1–142.

WHEELER, W.M. (1928): Mermis parasitism and intercastes among ants. Journal of Experimental Zoology 50: 165–237.

WIEST, L. (1966): Über die Funktion des Gasters bei der Verständigung von Ameisen. PhD Thesis, University of Vienna, Wien, 151 pp.

WIEST, L. (1967): Zur der Biologie der Ameise Liometopum microcephalum PANZ. Wissenschaftliche Arbeiten aus dem Burgenland 38: 136–144.

WILLIS, K.J. & VAN ANDEL, T.H. (2004): Trees or no trees? The environments of central and eastern Europe during the Last Glaciation. Quaternary Science Review 23: 2369–2387.

WILLIS, K., RUDNER, E. & SÜMEGI, P. (2000): The full-glacial forests of central and southeastern Europe. Quaternary Research 53: 203–213.

WILSON, E.O. (1953): The origin and evolution of polymorphism in ants. Quarterly Review of Biology 28: 136–156.

WILSON, E.O. (1966): Behaviour of social insects. Symposia of the Royal Entomological Society of London 3: 81–96.

ZÁLESKÝ, M. (1939): Prodromus našeho blanokřídlého hmyzu - pars III. Sborník entomologi- ckého odd. Národního Musea v Praze, XVII., 176: 219–220.

145

ZANGHERI, P. (1969): Repertorio sistematico e topografico della Flora e Fauna vivente e fossile della Romagna. Museo Civico di Storio Naturale di Verona, Mem. F.S.,1: 1415–1963.

ZDOBNITZKY, W. (1910): Beitrag zur Ameisenfauna Mährens. Mitteilungen der Kommission zur naturwissenschaftlichen Durchforschung Mährens. Zoologische Abteilung 15: 9–10

ZETTEL, H., LUBOMIROV, T., STEINER, F. M., SCHLICK-STEINER, B. C., GRABENWEGER, G., WIESBAUER, H. (2004): The European ant hunters Tracheliodes curvitarsus and T. varus (Hymenoptera: Crabronidae): taxonomy, species discrimination, distribution, and biology. Myrmecologische Nachrichten 6: 39–47.

ZIMMERMANN, S. (1934): Beitrag zur Kenntnis der Ameisenfauna Süddalmatiens. Verhandlungen der Zoologisch-Botanischen Gesellschaft in Wien 84: 5–65.

ZINCK, L., JAISSON, P., HORA, R. R., DENIS, D., POTEAUX, P. & DOUMS, C. (2007): The role of breeding system on ant ecological dominance: genetic analysis of Ectatomma tuberculatum. Behavioral Ecology 18: 701–708.

146

Author’s contributions to the papers

Paper 1:

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2011): Interactions between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence. Community Ecology 12 (1): 9–17.

LP and JS designed the study, LP collected the data in the field and performed all experiments and data analyses, LP and JS wrote the manuscript; contribution of LP was 65%.

Paper 2:

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2014): Worker polymorphism in the arboricolous ant Liometopum microcephalum (Hymenoptera: Formicidae: Dolichoderinae): Is it related to territory size? Myrmecological News 20: 101–111.

LP designed the study with advise from JS, collected the data in the field, measured the samples and performed data analyses, LP and JS participated in writing manuscript; contribution of LP was 80%.

Paper 3:

JS designed the study with advice from VS, JK collected the data in the field, JK conducted the chemical analyses under guidance by VS, LP performed the final data analyses; JS, VS and LP participated in writing the manuscript; contribution of LP was 20%.

Paper 4:

PETRÁKOVÁ, L., TÓTHOVÁ, A. & SCHLAGHAMERSKÝ, J.: Phylogeography of the rare velvety tree ant Liometopum microcephalum (Formicidae: Dolichoderinae). Journal of Biogeography (submitted manuscript).

JS, LP and AT designed the study, LP and JS collected the samples, LP processed the samples in the lab (under initial guidance by AT) and performed data analyses, all authors participated in writing the manuscript; contribution of LP was 65%.

147

148

List of publications

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2007): Preliminary results on the interaction of Liometopum microcephalum (Panzer, 1798) with other ants (Hymenoptera: Formicidae). Myrmecological News 10: 118.

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2011): Interactions between Liometopum microcephalum (Formicidae) and other dominant ant species of sympatric occurrence. Community Ecology 12: 9–17.

PETRÁKOVÁ, L., LÍZNAROVÁ, E., PEKÁR, S., HADDAD, C.R., SENTENSKÁ, L. & SYMONDSON, W.O.C (2015): Discovery of a monophagous true predator, a specialist termite-eating spider (Araneae: Ammoxenidae). Scientific Reports 5, doi:10.1038/srep14013.

PEKÁR, S., PETRÁKOVÁ, L., CORCOBADO, G. & WHYTE, R.: Revision of Australian ant- mimicking species of the genus Myrmarachne (Araneae, Salticidae) reveals a diversified complex of species, subspecies, and forms. Zoological Journal of the Linnean Society. In press.

Manuscripts under review:

PETRÁKOVÁ, L., TÓTHOVÁ, A. & SCHLAGHAMERSKÝ, J.: Phylogeography of the rare velvety tree ant Liometopum microcephalum (Formicidae: Dolichoderinae). Journal of Biogeography

PETRÁKOVÁ, L., MICHALKO, R., LOVERRE, P., SENTENSKÁ, L., KORENKO, S., PEKÁR, S.: Intraguild predation in a pear orchard in winter: quantification of predation on pear psylla and among winter-active spiders (Psyllidae, Araneae). Agriculture, Ecosystems & Environment

PEKÁR, S., PETRÁKOVÁ, L., BULBERT, M.W., WHITING, M.J. & HERBERSTEIN, M.: Less defended mimics enhance protection in a vast and diverse mimetic complex. Current Biology

Publications aimed at a wider public:

SCHLAGHAMERSKÝ, J. & PETRÁKOVÁ, L. (2014): Mravenec lužní – Mýty a fakta. Živa 5/2014: 230–233.

149

International conference contributions

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2009): Interactions between Liometopum microcephalum and other dominant species of sympatric occurrence. 3rd Central European Workshop of Myrmecology, Chiemsee, Germany. Oral presentation.

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2011): Polymorphism in Liometopum microcephalum (Formicidae: Dolichoderinae): Is it related to territory size? 4th Central European Workshop of Myrmecology, Cluj-Napoca, Romania. Oral presentation.

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2013): Phylogeography of the rare ant Liometopum microcephalum (Formicidae: Dolichoderinae): Preliminary results of a study on populations across the entire species range. 3rd Central European IUSSI Meeting, Cluj-Napoca, Romania. Oral presentation.

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2013): Phylogeography of the rare ant Liometopum microcephalum (Formicidae: Dolichoderinae): results of a study on populations across the entire species range. 5th Central European Workshop of Myrmecology, Innsbruck, Austria. Oral presentation.

SCHLAGHAMERSKÝ, J., KAŠPAR, J., OMELKOVÁ, M. & PETRÁKOVÁ, L. (2015): Foraging in Liometopum microcephalum (Formicidae: Dolichoderinae): territories, activity dynamics and food items. 6th Central European Workshop of Myrmecology, Debrecen, Hungary. Poster.

PETRÁKOVÁ, L., LÍZNAROVÁ, E., PEKÁR, S., HADDAD, C.R., SENTENSKÁ, L. & SYMONDSON, W.O.C (2015): Discovery of a monophagous true predator, a termite-eating spider specialist (Araneae: Ammoxenidae). European Congress of Arachnology 2015, Brno, Czech Republic. Oral presentation.

National conference contributions

PETRÁKOVÁ, L. & SCHLAGHAMERSKÝ, J. (2012): Polymorfismus a potravní teritoria u mravence lužního (Liometopum microcephalum). Zoologické dny 2012, Olomouc. Oral presentation.

SCHLAGHAMERSKÝ, J. & PETRÁKOVÁ, L. (2012): Co víme o mravenci lužním (Liometopum microcephalum): mýty a fakta. Zoologické dny 2012, Olomouc. Oral presentation.

PETRÁKOVÁ, L., LÍZNAROVÁ, E., PEKÁR, S., HADDAD, C.R., SENTENSKÁ, L. & SYMONDSON, W.O.C (2015): Objev monofágního pravého predátora, pavouka specializovaného na lov termitů. Zoologické dny 2016, České Budějovice. Oral presentation.

150