Structure Des Assemblages De Fourmis Dans Une Foret Naturellement Fragmentee Du Chaco Humide Argentin

Université Libre de Bruxelles Institut Royal des Sciences Naturelles de Belgique Faculté des sciences Section d'Evaluation Biologique Service d'Eco-Ethologie Evolutive

STRUCTURE DES ASSEMBLAGES DE FOURMIS DANS UNE FORET NATURELLEMENT FRAGMENTEE DU CHACO HUMIDE ARGENTIN.

Thèse présentée pour obtenir le grade de Docteur en Sciences Biologiques

Par

Laurence Theunis

2008

Soutenue devant le jury composé de:

Jacques M. PASTEELS Université Libre de Bruxelles Président

Jean-Christophe DE BISEAU D’HAUTEVILLE Université Libre de Bruxelles Secrétaire

Marius GILBERT Université Libre de Bruxelles Rapporteur

Bruno CORBARA Université Blaise Pascal de Clermont- Rapporteur Ferrand Yves ROISIN Université Libre de Bruxelles Promoteur

Maurice LEPONCE Institut Royal des Sciences Naturelles de Co-promoteur Belgique

À mes parents À Karim, Oscar et Victor

iii iv

REMERCIEMENTS

Vous n’imaginez pas combien de fois au cours de ces 9 dernières années j’ai rêvé cet instant où j’allais enfin écrire ces remerciements !! En 9 ans, de nombreuses personnes m’ont encouragée et soutenue, c’est ici pour moi l’occasion de les remercier.

Je commence évidemment par remercier le Dr Maurice Leponce, mon directeur de thèse pour l'aide et l'encouragement prodigués tout au long de ce travail de recherche. Il a su remarquablement me guider durant mes recherches. Maurice est toujours disponible et à l’écoute malgré la multitude de ses activités. Il est une de ces personnes rares qui possèdent de grandes qualités tant professionnelles qu’humaines.

Je remercie également le Prof. Yves Roisin, mon promoteur de thèse, pour m’avoir acceptée dans son unité de recherche afin que je puisse réaliser cette thèse. Yves est un chercheur exigent mais juste dont les critiques et conseils sont toujours pertinents. Mais sa plus grande qualité est, sans conteste, l’aisance dans le maniement de la machette.

Un chaleureux merci va au Prof. Jacques Pasteels qui m’a permis de débuter cette thèse sous sa direction et qui a nourri ma passion pour les insectes.

Je vous remercie tous les trois ainsi que le Dr Chantal De Ridder, en tant que membres de mon comité d’accompagnement, pour m’avoir permis d’achever ma thèse après tant d’années !

Je remercie Jean-Christophe de Biseau d’Hauteville, Marius Gilbert et Bruno Corbara d’avoir accepté d’évaluer ce travail.

Merci à l’équipe de la Section d’Evaluation Biologique de l’IRSNB. J’ai passé en votre compagnie 6 années (ben oui les 3 dernières fallait bien gagner des sous….) formidables, pleines de souvenirs, de fous rires et d’anecdotes ! Merci à Pierre Devillers, René-Marie Lafontaine, les sœurs Beudels, Didier Vangeluwe, Isabelle Bachy, Chris Kerwyn, Yves Laurent, Géraldine Kapfer, Géraldine de Montpellier, David Monticelli, Yves Braets, Thibaud Rigot, David Cammaerts.

Un merci spécial à Thibaut Delsinne, mon voisin de bino, pour sa bonne humeur, son humour, sa disponibilité, son aide en toutes circonstances. Je te dois quelques fières chandelles !!

Un merci tout particulier à Yves Laurent pour son aide indispensable lors de la dernière mission de terrain.

Merci également à Isabelle Bachy pour avoir mis ces talents de graphiste au service de cette thèse.

J'exprime ma profonde gratitude envers Julien Cillis (IRSNB, microscopie à balayage) pour l'énorme travail de qualité qu'il a réalisé en photographiant en microscopie électronique les

v fourmis collectées. Merci aussi à Jérôme Constant (IRSNB, entomologie) pour les dépannages en tubes, colle, etc.! Je remercie Patricia Féron (ULB - Service d'Eco-Ethologie Evolutive) pour son dévouement et sa grande gentillesse.

Cette thèse n’aurait pu exister sans le matériel collecté dans le Parc National Río Pilcomayo. Merci à l’« Administración de Parques Nacionales » d’Argentine pour nous avoir délivré les permis de récoltes. Merci à Nestor Sucunza ainsi qu’aux gardiens de parc pour nous avoir grandement faciliter le travail de terrain. Un tout grand merci à Cornélio pour son authenticité, sa connaissance du parc et ses magnifiques chevaux.

Merci à Gladys Torales et Enrique Laffont, Universidad Nacional del Nordeste, pour le support logistique. Merci à Fabiana Cuezzo et son équipe, Instituto Miguel Lillo, de m’avoir permis de consulter leur collection de référence de fourmis.

J'adresse mes vifs remerciements au Dr. Frieda Billiet (Jardin Botanique de Meise) pour avoir participé à l'identification des espèces d'arbres et d'arbustes chaquéens.

Enfin j'exprime toute ma gratitude au FNRS (Fonds national de la Recherche Scientifique), pour la bourse d’aspirant dont j'ai été la bénéficiaire durant 4 ans ainsi que pour le financement de deux missions qui m’a été accordé, et la Fondation Universitaire David et Alice Van Buuren pour le soutien financier qui m’a été octroyé en 5ème année de thèse. Que le Fonds Léopold III pour l'Exploration et la Conservation de la Nature trouve également ici toute ma reconnaissance pour avoir subventionné les frais de ma seconde mission de terrain.

Merci encore à mes chers parents de m’avoir encouragée et soutenue dans mes choix professionnels. Merci à vous pour les valeurs que vous m’avez transmises. Merci à ma soeurette Sandrine pour sa joie de vivre et son soutien moral. Merci à ma famille pour leurs affections et encouragements.

Merci à mes adorables beaux-parents, Nadia et Ali, pour leur aide er leur dévouement. Vous m’avez acceptée et aimée dès notre rencontre. Merci.

Merci à mes hommes… Oscar, Victor et Karim. Sans vous, je n’aurais pas trouvé le courage pour terminer ce travail. Désolée pour tous les moments ou les vacances qu’il nous aura volés. Même si je ne vous l’ai pas dit souvent ces derniers temps : Je vous aime plus que tout !

Merci à tous mes amis ! Qu’ils sachent ici combien ils m’ont toujours été précieux et combien leur présence à mes côtés m’est vital.

Je tiens néanmoins à NE PAS remercier ni les broméliacées, pour avoir lacérées mes jambes pendant la collecte de fourmis des litières, ni les taons, les moustiques et les tiques pour nous avoir presque rendus fous !

Je tiens pour finir à remercier les fourmis qui ont fait dont de leur corps pour la science. Qu’elles n’y voient rien de personnel !

vi vii viii

Résumé de la thèse

Contexte: La fragmentation des habitats induisant une diminution de leur surface, de leur connectivité et une augmentation de la zone de contact avec d’autres milieux constitue l’une des menaces majeures pour le maintien de la biodiversité. Les effets de la fragmentation ne doivent pas être confondus avec les perturbations transitoires liées à un morcellement de l’habitat par une déforestation récente. Les îlots forestiers du Chaco humide, situés sur des monticules légèrement surélevés par rapport à la savane environnante qui est régulièrement inondée et brûlée, constituent un système naturellement fragmenté propice à l’étude des effets de la fragmentation sensu stricto. Dans ces forêts subtropicales sèches, comme dans la plupart des écosystèmes terrestres, les fourmis constituent l’un des organismes les plus abondants. Objectif: Le but principal de la thèse a été de déterminer, à trois échelles spatiales, les facteurs influençant la structure des assemblages de fourmis terricoles : (1) à l’échelle du microhabitat constitué par la litière de feuilles et la couverture végétale dominée par des broméliacées terrestres ; (2) à l’échelle du fragment forestier dont la surface, la forme et l’isolement est variable ; (3) à l’échelle du paysage, constitué de forêt et de savane, soumis à des feux périodiques, et au niveau duquel nous nous sommes intéressés aux effets de bord se produisant à l’interface entre les deux milieux. Méthode: Le site d’étude est la forêt naturellement fragmentée du Parc national Rio Pilcomayo localisé dans le Chaco humide argentin. Onze fragments forestiers de taille (± 2.5ha, 25ha et 250ha), de forme et de degré d’isolation divers ont été échantillonnés ainsi que la savane environnante, récemment brûlée ou non. La diversité et la densité des fourmis a été quantifiée au moyen d’un protocole standardisé (« protocole A.L.L. ») qui a été préalablement calibré pour en définir la représentativité. Ce protocole consiste en un transect de 200m le long duquel sont placés, à intervalles de 10m, des pièges à fosse et des quadrats délimitant 1m² de litière de feuilles. La faune vivant dans la litière est ensuite extraite au moyen d’un dispositif appelé Winkler. Le calibrage du protocole a été réalisé en suréchantillonnant 8 fois le transect (160 points d’échantillonnage au lieu de 20). Cet échantillonnage quasi exhaustif de 200m² a permis de comparer l’estimation du nombre d’espèces obtenue par le transect standardisé ALL avec sa valeur réelle et d’étudier la distribution des espèces à l’échelle du mètre. Les facteurs du micro- habitat les plus susceptibles d’influencer la distribution des fourmis (quantité de litière et densité de broméliacées) ont été mesurés systématiquement le long des transects. Pour l’étude de la distribution des fourmis depuis le coeur d’un grand fragment jusque dans la savane, des transects de 500m ont été utilisés et ont permis de mesurer des effets de bords éventuels. Un total de 800 Winkler et 560 pièges à fosses ont été analysés lors de cette étude. Résultats: Un transect standardisé A.L.L. permet d’obtenir, à partir de 20 échantillons et de méthodes analytiques adéquate, une estimation fiable de la richesse locale au sein de 200m² mais n’est pas toujours représentatif de la fréquence relative des espèces. Au total, 150 espèces de fourmis ont été récoltées dont 130 en forêt et 79 en savane (dont 59 espèces communes aux deux milieux). Au niveau du micro-habitat, on observe pour certaines espèces des pics périodiques d’abondance

(maximum tous les 10m) correspondant vraisemblablement à l’emplacement des colonies qui s’espacent pour diminuer la compétition intraspécifique. Associé aux micrconvexités topographiques l’on observe également des pics de densité de broméliacées et de quantité de litière qui favorisent une grande densité d’espèces différentes de fourmis. À l’échelle de l’habitat, les îlots forestiers petits et isolés sont les moins riches, principalement en espèces typiquement forestières. Dans les larges fragments, les espèces typiquement forestières se distribuent indépendamment de la distance les séparant du bord. Quelques espèces typiques de savane pénètrent en bordure de forêt et provoquent une plus grande variabilité de la faune récoltée au sein des quadrats de litière situés à cet endroit. Cependant, aucun pic de diversité correspondant à une zone de superposition d’espèces de bord et de centre n’a été observé au sein des fragments forestiers. Les feux de savane modifient la fréquence relative des espèces les plus communes mais n’affectent pas la richesse globale du milieu et ne pénètrent pas dans la forêt. Conclusions: Le protocole standardisé ALL, utilisé couramment par de nombreuses équipes de chercheurs à travers le monde, mais qui n’avait encore jamais été réellement calibré avant notre étude, apparaît comme une méthode minimale mais suffisante pour déterminer la richesse locale en fourmis d’une forêt du Chaco humide. Ce calibrage a permis, en outre, de mettre en évidence un taux important de renouvellement des espèces à l’échelle du mètre carré. Nos résultats soutiennent l’idée que la disponibilité en ressources favorables, plus que la compétition interspécifique, est un mécanisme majeur structurant les assemblages de fourmis des litières. À l’échelle du micro-habitat, un grand nombre d’espèces de fourmis forestières coexistent dans les zones riches en matière organique associée à la présence de broméliacées qui apparaissent comme un facteur structurant majeur de la distribution des fourmis. Au niveau de la litière, les colonies de différentes espèces ont des aires de fourragement qui se superposent tandis que les colonies de même espèce ont tendance à s’espacer limitant la compétition pour les mêmes ressources. Un effet de bord, lié à des modifications locales des conditions climatiques et de la structure de la végétation, ne se marque pas au niveau de la myrmécofaune dans ce type de milieu, ce qui explique que l’on n’observe pas le traditionnel pic de diversité au niveau de la zone de transition entre deux milieux. Du point de vue de la conservation des espèces, des fragments forestiers de 15ha, bien connectés, apparaissent comme des conditions minimum pour conserver l’ensemble des espèces de fourmis de l’assemblage.

TABLE DES MATIERES

REMERCIEMENTS V

RESUME DE LA THESE IX

TABLE DES MATIERES XI

INTRODUCTION GENERALE 1

Chapitre 1: 33 SCALE DEPENDANCE OF DIVERSITY MEASURES IN A LEAF LITTER ANT ASSEMBLAGE

Chapitre 2: 49 SPATIAL STRUCTURE OF LITTER-DWELLING ANT DISTRIBUTION IN A SUBTROPICAL DRY FOREST Chapitre 3: 63 TERRESTRIAL BROMELIAD PRESENCE AFFECTS THE GROUND-DWELLING ANT DISTRIBUTION IN CHACOAN FORESTS Chapitre 4: 79 LONG-TERM EFFECTS OF FOREST FRAGMENTATION ON ANTS OF THE HUMID CHACO

Chapitre 5: 111 IMPACT OF FIRE ON THE ANT ASSEMBLAGE STRUCTURE IN GRASSLANDS OF THE HUMID CHACO

DISCUSSION GENERALE - PERSPECTIVES 131

Annexe 1: 151 RAINFALL INFLUENCES ANT SAMPLING IN DRY FORESTS

Annexe 2: 159 LISTE DES ESPÈCES

xii

INTRODUCTION

INTRODUCTION GENERALE

Cette introduction présentera, tout d’abord, les différents thèmes abordés dans cette thèse: 1- La biodiversité et la fragmentation des forêts tropicales et subtropicales, 2- Les fourmis et 3- La structure des assemblages de fourmis à différentes échelles spatiales. Ensuite, nous présenterons le site d’étude et les objectifs que nous nous sommes fixés dans ce travail. Nous développerons la méthodologie générale utilisée dans cette étude. L’introduction sera suivie par les différents chapitres, présentés sous format d’articles, et ensuite par une discussion générale.

1. Biodiversité et fragmentation des forêts tropicales et subtropicales

1.1 Biodiversité en danger !

La perte de biodiversité est une menace subtile mais grandissante dans les forêts tropicales. La perte de biodiversité due aux activités humaines n’a jamais été aussi rapide que lors de ces 50 dernières années (Millenium Ecosystem Assessment [MEA] 2005). Pimm et al. (1995) estiment que les taux de perte de biodiversité sont 100 à 1000 fois supérieurs à ce qu’ils étaient avant l’intervention de l’homme. Les perspectives pour le futur ne sont pas plus encourageantes : les taux actuels d’extinction des espèces devraient, selon le scénario développé par le MEA (MEA 2005) retenu, soit se maintenir soit s’accélérer.

Les causes de ce déclin sont la disparition, la fragmentation et la dégradation des habitats, leur surexploitation, les changements climatiques globaux, l’introduction d’espèces invasives et la pollution (MEA 2005). Ces facteurs interagissent et amplifient leurs effets sur la biodiversité (Woodruff 2001, Travis 2003, Wilson 2003, Thomas et al. 2004, Didham et al. 2005). La perte de biodiversité touche tous les écosystèmes et tous les groupes taxonomiques (McKinney 1999, Laurance et al. 2002). Cependant, la majorité des extinctions qui auraient eu lieu dans le passé et que l’on prédit dans le futur concernent des insectes (Dunn 2005). Malgré cela, l’étude des extinctions d’insectes a été négligée. Seulement 70 extinctions d’insectes contemporains ont été documentées, alors qu’elles sont estimées à plusieurs milliers (Dunn 2005). Ces chiffres faussement optimistes reflètent à la fois le manque d'études concernant la diversité des insectes et la difficulté de les réaliser (Kellert 1993, McKinney 1999, Dunn 2005,

1 INTRODUCTION

Samways 2007, Condon et al. 2008). Les taux d’extinctions des populations d'insectes sont égaux ou supérieurs à ceux qui sont observés chez des groupes plus couramment étudiés tels que les mammifères, les oiseaux ou les plantes vasculaires (McKinney 1999, Thomas et al. 2004). Beaucoup d'insectes ont un temps de génération bref, répondent finement aux conditions du milieu (Oliver et al. 1998, French 1999) et ont une répartition géographique restreinte (Yeates et al. 2002). Ces caractéristiques rendent les insectes particulièrement vulnérables aux changements environnementaux (Thomas et al. 2004, Dunn 2005).

1.2 Fragmentation des habitats forestiers

La fragmentation de l’habitat est généralement définie comme un processus au cours duquel « une large étendue d’habitat est transformée en un plus petit nombre de fragments de surface plus petite, isolés les uns des autres par une matrice d’habitats différents de l’original » (Wilcove et al. 1986). Dans la littérature, le terme « fragmentation des habitats » englobe deux processus différents mais souvent liés : le morcellement de l’habitat et la déforestation pratiquée par l’homme (perte d’habitat) (Fahrig 2003, Laurance 2008). Ces processus très distincts se déroulent à des échelles spatio-temporelles très différentes (processus naturel à long terme vs actions humaines à court terme) et dont les conséquences sur les assemblages d’espèces peuvent être très différentes (Fahrig 2003). Cela a pour conséquence d’augmenter la variabilité des conclusions des différentes études des effets de la fragmentation sur la dynamique des populations et d’en compliquer leurs interprétations (Debinski et Holt 2000).

Le morcellement et la perte des habitats sont des causes majeures de la modification de la biodiversité (Brooks et al. 2002, Laurance et al. 2002, Laurance et Vasconcelos 2004). Cependant, le dramatique déclin de la biodiversité rapporté dans différents paysages fragmentés doit probablement être plus largement attribué à la perte d’habitat qu’à la fragmentation de l’habitat (Fahrig 2003). L’appauvrissement biologique des forêts est un des changements globaux les plus significatifs de notre époque, au même titre que les changements environnementaux tels que le réchauffement climatique et la perturbation du cycle de l’azote. Leurs effets sont généralement étudiés dans les forêts à haute valeur biologique (richesse biologique et degré d’endémisme élevés) fortement dégradées par l’homme telles que les forêts tropicales humides d’Amazonie Centrale (Biological Dynamics of Forest Fragments Project, Lovejoy et al. 1983, 1986, Bierregaard et al. 1992, De Souza et

2 INTRODUCTION

Brown 1994, Carvalho et Vasconcelos 1999, Didham 1997 a et b, 1998, Laurance et al. 2002), la forêt humide Atlantique (Majer et al. 1997, Delabie et al. 2000), les forêts tempérées humides du Chili (Echeverria et al. 2006, 2007), ou la forêt humide de basse altitude en Malaisie (Brühl et al. 2003). Les études en Amazonie suggèrent que les effets de bord jouent un rôle clé dans la dynamique des fragments forestiers (Malcolm 1994, Brown et Hutchings 1997, Didham 1997b, Kapos et al. 1997, Laurance et al. 1998, Didham et Lawton 1999, Carvalho et Vasconcelos 1999) que la nature et la structure de la matrice environnante a une influence majeure sur la connectivité des fragments (Ricketts 2001) et que beaucoup d’espèces animales amazoniennes (mammifères, oiseaux, reptiles, insectes,..) évitent de traverser des clairières inférieures à 100m de large (Powell and Powell 1987, Klein 1989, Malcolm 1991, Stratford et Stouffer 1999, Laurance et al. 2002). Les effets de la fragmentation sont très nombreux ; ils modifient la richesse et l’abondance spécifique (Lovejoy et al. 1986, Vasconcelos 1988, Klein 1989, Didham 1997a, Brown et Hutchings 1997, Stratford et Stouffer 1999, Laurance et al. 2002), ils entraînent des invasions d’espèces (Gascon 1993, Brown et Hutchings 1997), ils perturbent la dynamique forestière, la structure trophique des communautés ainsi qu’une variété de processus écologiques (interactions biologiques tels que les symbioses, mutualismes, prédation, pollinisation, …) et écosystémiques (cycle du carbone et climat, cycle de l’azote, des nutriments, la photosynthèse, propriétés du sol...) (Suarez et al. 1998, Gibbs et Stanton 2001, Laurance et al. 2002, Fahrig 2003,). La chasse et les feux liés à l’exploitation forestière aggravent les effets de la fragmentation constituant ainsi une plus grande menace pour les biotopes forestiers (Laurance et al. 2002).

Bien que l’on considère les forêts subtropicales (sèches et humides) comme étant parmi les formations tropicales les plus menacées par la perte de biodiversité et nécessitant des mesures immédiates en matière de conservation (Janzen 1988, Redford et al. 1990, Olson et Dinnerstein 2002, Sánchez-Azofeifa et al. 2005a, b, Miles et al. 2006), peu d’articles scientifiques traitant de la fragmentation de ces habitats existent. La littérature actuelle est également pauvre en études sur les effets de la fragmentation naturelle (ou à long terme) des habitats forestiers sur la biodiversité. Kotze et Samways (2001) ont étudié les patrons de diversité d’invertébrés dans des forêts et savanes naturelles d’Afrique du Sud et nous commençons à avoir assez de recul pour des études de la fragmentation à long terme en Amazonie (Laurance et al. 2002, Vasconcelos et al. 2006). Etudier les patrons de diversité dans un habitat naturellement fragmenté (ou fragmenté depuis longtemps) devrait permettre de dégager les effets dus à la fragmentation sensu stricto et non ceux dus à la déforestation

3 INTRODUCTION récente (perte d’habitat). Fahrig (2003) suggère d’ailleurs de réserver le terme de « fragmentation » au morcellement naturel d’un habitat.

2. Les fourmis du sol

2.1 Avantages des fourmis pour les études en biologie de la conservation

Parmi les invertébrés, les fourmis possèdent de nombreux attributs les rendant particulièrement utiles pour les évaluations biologiques et les études de suivi (Majer 1983, New 1995). Avec plus de 12000 espèces décrites (Bolton 2003, Agosti et Johnson 2005, Bolton et al. 2007), elles ne constituent pas un groupe hyperdiversifié et leur taxonomie est relativement bien connue (Bolton 1994, 1995). Elles sont écologiquement importantes dans la plupart des écosystèmes terrestres (Wilson 1987, LaSalle et Gauld 1993). Dans la forêt humide d’Amazonie, elles constituent jusqu’à 15% de la biomasse animale totale (Fittkau et Klinge 1973). Elles se développent dans toutes les strates forestières (Brühl et al. 1998), utilisent une large variété de ressources alimentaires (Hölldobler et Wilson 1990, Kaspari 2000) et contribuent à la richesse biologique des forêts par leurs multiples interactions avec les autres organismes (Kaspari 1996a, Schultz et McGlynn 2000). Généralement 45 - 50% de la biomasse des macroinvertébrés des litières de feuilles sont des fourmis (Lavelle et Kohlmann 1984, Adis et al. 1989, Brühl 1996, Burghouts et al. 1994, Olson 1994). Les fourmis du sol et de la litière peuvent être considérées comme des ingénieurs des écosystèmes (Jones et al. 1994). Elles y jouent effectivement des fonctions importantes telles que le recyclage des nutriments, la dispersion des graines et la régulation des populations des autres insectes (Hölldobler et Wilson 1990, Folgarait 1998, Bestelmeyer et Wiens 2003). Étant donné leur importance écologique, le caractère sessile des colonies de la plupart des espèces et leur sensibilité aux conditions et changements environnementaux, les fourmis sont utilisées dans de nombreux programmes d’évaluation de la biodiversité (Majer 1983, Majer et al. 1984, Andersen 1995, 1997, Andersen et Sparling 1997, Peck et al. 1998, Vasconcelos 1999, Carvalho et Vasconcelos 1999, Whitford et al. 1999, Kaspari et Majer 2000, Agosti et al. 2000, Kalif et al. 2001, Vasconcelos et al. 2001, Underwood et Fisher 2006). Les communautés de fourmis sont des indicateurs utiles de perturbations (Majer 1983, Bestelmeyer et Wiens 1996, Andersen 1997, Majer et al. 1997, Suarez et al. 1998, Vasconcelos 1999, Carvalho et Vasconcelos 1999, Vasconcelos et Delabie 2000, Kaspari et Majer 2000, Soares et

4 INTRODUCTION

Schoereder 2001, Brühl et al. 2003) et de l’état de régénération des écosystèmes (Majer et al. 1984, Andersen et Sparling 1997, Alonso 2000, Alonso et Agosti 2000, Vasconcelos et al. 2001, Palladini et al. 2007, Silva et al. 2007). Elles sont facilement collectées selon un protocole standardisé adopté par différentes équipes de chercheurs au niveau mondial : le protocole « Ants of the Leaf Litter » (ALL protocol) (Agosti et Alonso 2000). Ce protocole a été mis au point dans les forêts tropicales humides et il reste à vérifier son applicabilité dans d’autres habitats tels que les forêts sèches. De plus, le calibrage de ce protocole permettrait des comparaisons fiables à larges échelles d’inventaires réalisés à différents endroits par différentes équipes de chercheurs. Pour optimiser le potentiel informatif des fourmis en biologie de la conservation, il est indispensable d’étudier les facteurs qui déterminent la distribution des espèces et la structure des assemblages à différentes échelles spatio-temporelles (Levin 1992).

5 INTRODUCTION

3. Structure des assemblages de fourmis à différentes échelles spatiales

Un assemblage d’espèces est un ensemble d’espèces liées phylogénétiquement, présentes au sein d’une même unité spatio-temporelle mais n’exploitant pas nécessairement les mêmes ressources (Fauth et al. 1996). La composition et l’abondance relative des espèces constituant un assemblage sont influencées par des facteurs abiotiques (facteurs climatiques, édaphiques, topographiques, …), biotiques (compétition inter- et intraspécifique, prédation, présence de parasites ou de mutualistes, …), historiques ou aléatoires. Les écologistes examinent les patrons spatiaux des espèces ou des assemblages dans le but de caractériser la structure des assemblages, d’optimiser les inventaires d’espèces et afin de comprendre les mécanismes qui contrôlent leur distribution. La perte de biodiversité est causée par des facteurs locaux, régionaux et globaux, de telle sorte qu’il est nécessaire d’étudier les assemblages d’espèces à différentes échelles spatiales afin de pouvoir mettre en place des stratégies de conservation efficaces (Levin 1992, Andersen 1997).

Les processus qui génèrent les structures des assemblages de fourmis peuvent être envisagés à 3 échelles spatiales complémentaires: 1- micro-habitat, 2- habitat et 3- paysage (Schluters et Ricklefs 1993, Williams et al. 2002).

3.1. Structuration des assemblages de fourmis à l’échelle du micro-habitat

La distribution des assemblages de fourmis dépend des contraintes biotiques (interactions entre organismes: compétition, prédation, parasitisme) et abiotiques (facteurs du milieu physique). Une importante hétérogénéité de la densité des colonies et de la distribution des espèces est une caractéristique commune de nombreux assemblages de fourmis tropicales (Wilson 1958, Levings et Franks 1982, Levings 1983, Benson et Brandão 1987, Kaspari 1996a, Vasconcelos et Delabie 2000). Les caractéristiques comportementales des fourmis, telles que la compétition interspécifique ou l’association d’espèces, sont considérées comme des mécanismes majeurs structurant les assemblages de fourmis (Levings et Franks 1982, Vepsälainen et Pisarki 1982, Hölldobler et Wilson 1990, Perfecto 1994, Sanders et al. 2003). Ces interactions sont parfois particulièrement marquées dans la canopée lors de la présence d’espèces territoriales responsables d’une distribution des espèces en mosaïque (« arboreal ant mosaic ») et agissent, dans une moindre mesure, sur les assemblages de fourmis qui nichent ou fourragent dans la litière et le sol (Levings et Traniello 1981, Levings et Franks 1982, Majer 1993, Delabie et al. 2000, Dejean et al. 2007).

6 INTRODUCTION

Contrairement aux espèces de la canopée, les fourmis des litières ne sont pas territoriales impliquant le recouvrement de leurs aires de fourragement. Un ensemble d’observations suggèrent que la disponibilité en ressources favorables, plus que la compétition, soit un mécanisme majeur structurant leurs assemblages (Franks 1982, Jackson 1984, Byrne 1994, Kaspari 1996a,b, Soares et Schoereder 2001). Ainsi, la distribution des colonies de fourmis des litières sont principalement influencées par la variabilité spatio-temporelle de l’humidité du sol (Levings 1983, Levings et Windsor 1984, Kaspari 1996a) et de la température (Bestelmeyer 2000, Delsinne et al. 2007), la topographie (Vasconcelos et al. 2003), la disponibilité en nid et en nourriture (Herbers 1989, Byrne 1994, Kaspari 1996b, Kaspari et Majer 2000), la quantité et la qualité de la litière de feuille (Vasconcelos 1990, Höfer et al. 1996, Kaspari 1996b, Carvalho et Vasconcelos 1999, Theunis et al. 2005) ainsi que la structure et la composition végétale (Wilson 1958, Gadagkar et al. 1993, Feener et Schupp 1998, Moutinho 1998, Retana et Cerdà 2000, Bestelmeyer et Wiens 2001, Vasconcelos et al. 2008). D’autres facteurs, tels que la prédation (Gotelli 1993, 1996) et le parasitisme (Adler et al. 2007, Lebrun et Feener 2007) ont montré leur impact sur la distribution des fourmis. En Amérique du Sud, par exemple, la prédation par les fourmis légionnaires a un impact considérable sur l’organisation spatiale des espèces (Franks et Bossert 1983, Kaspari 1996b, Hirosawa et al. 2000). La compréhension des effets de la compétition interspécifique et des facteurs du micro-habitat sur la distribution des fourmis à fine échelle dans les forêts tropicales et subtropicales est encore fragmentaire et nécessite des études complémentaires.

3.2. Structuration des assemblages de fourmis à l’échelle de l’habitat

La structure de la végétation est généralement un facteur important affectant les populations de fourmis dans les habitats tempérés et tropicaux (Retana et Cerda 2000, Ribas et al. 2003, Lassau et Hochulli 2004). La végétation est effectivement un régulateur majeur des conditions climatiques influençant l’activité des fourmis. En outre, la végétation peut affecter directement et indirectement la disponibilité en nourriture et sites de nidification, ainsi que les interactions entre espèces (Retana et Cerda 2000, Ribas et al. 2003, Lassau et Hochulli 2004). Ces dernières décennies, de nombreuses études ont cherché à déterminer les effets de la fragmentation des habitats sur la structure des assemblages de fourmis, principalement dans les forêts tropicales humides. Les études suggèrent que les effets de bord affectent, jusqu’à 200 m. à l’intérieur de la forêt, la richesse spécifique, la densité, la composition et l’abondance des fourmis (Didham 1997a, Carvalho et Vasconcelos 1999, Laurance 2004, Laurance et al.

7 INTRODUCTION

2002). Dans les ilôts les plus petits et isolés, la richesse et la densité spécifiques sont inférieures à celles mesurées en forêt continue (Vasconcelos 1988, Didham 1997b, Vasconcelos et Delabie 2000, Brühl et al. 2003). Les différences de composition des assemblages sont en partie dues à une quantité de litière de feuille plus importante près des bords (Carvalho et Vasconcelos 1999). Les assemblages de fourmis dans les petits fragments sont fortement influencés par la structure et la composition de la végétation de la matrice environnante (Vasconcelos 1999, Ambrecht et Ulloa-Chacón 1999). Plus la transition entre la végétation du fragment et de la matrice est nette, moins le nombre d’espèces peut se maintenir ou immigrer dans le fragment (Gascon et al. 1999, Vasconcelos 1999, Kotze et Samways 2001).

Comme nous l’avons exposé plus haut, la littérature est pauvre en études sur l’effet de la fragmentation naturelle (ou à long terme) sur les assemblages de fourmis. Or ces études devraient permettre d’évaluer le réel impact de la fragmentation de l’habitat indépendamment de celui de la déforestation.

3.3. Structuration des assemblages de fourmis à l’échelle du paysage

La distribution des fourmis au sein d’un paysage est déterminée par sa structure (Vasconcelos 1999, Deblauwe et Dekoninck 2007a et b, Vasconcelos et al. 2008), son degré de fragmentation (Majer et al. 1997, Suarez et al. 1998, Carvalho et Vasconcelos 1999, Brühl et al. 2003, Vasconcelos et al. 2006), la distance géographique séparant les fragments d’habitat (Vasconcelos et Delabie 2000), le régime des feux (York 2000, Farji-Brener et al. 2002, Hoffmann 2003, Andersen et al. 2006, Arnan et al. 2006, Santos et al. 2008), ou les perturbations d’origine anthropique (Bestelmeyer et Wiens 1996, 2001). Les formations végétales structurant le paysage sont forgées par les variations du relief et du sol. Les sols dépendent à la fois de la roche mère, du climat, de la morphologie (et de l’hydrographie), des agents biologiques et du temps (Duchaufour 2001). Dès lors, ces facteurs abiotiques influencent indirectement la distribution des fourmis à une large échelle. Les paysages fragmentés sont composés d’habitats de qualité différente pour la faune et constituent donc un système idéal pour l’étude des liens entre fourmis et habitats. Comprendre comment les espèces sont distribuées dans une forêt fragmentée requiert des informations sur leurs affinités vis-à-vis des habitats constituants le paysage (Malcolm 1991, Laurance 1994, Prance 2006). La distribution des fourmis dans un paysage fragmenté est influencée par les

8 INTRODUCTION perturbations naturelles ou anthropiques qui maintiennent les frontières entre les habitats (feux, inondations), la capacité de colonisation et les caractéristiques physiologiques des espèces (mode de dispersion, preferendums de température et d’humidité, besoins en ressources alimentaires et en sites de nidification), ainsi que les interactions biologiques entre espèces. La dynamique de formation et de maintien des frontières entre forêt et savane est une question clé en biogéographie tropicale (Furley et al. 1992) mais est restée relativement peu étudiée jusqu’à présent, principalement dans les régions sub-tropicales.

4. Site d’étude

4.1 Parc National Río Pilcomayo

Notre site d’étude, le Parc National Río Pilcomayo (PNRP), est situé au nord-est de l’Argentine, dans la province de Formosa, à la frontière paraguayenne (latitude S 25° 07’, longitude O 58° 10’) (Figure 0-1). Cette aire protégée crée en 1951 représente aujourd’hui une aire de 500km² délimitée par la rive sud du Río Pilcomayo et se trouve dans la partie humide, le plus à l’est, du Gran Chaco. Ce parc, comme le reste du Gran Chaco, a souffert de l’exploitation humaine indigène (chasse, exploitation forestière, pâturage,..). Cette pression anthropique a été malgré tout relativement faible et on considère que plus de 70% de la surface du biome est resté intact (Noss et al. 2002).

4.1.1 Origine, faune et flore

La pénéplaine que constitue le Chaco est d’origine alluviale (Ramella et Spichiger 1989, Pujalte et al. 1995). Le relief est formé de dépressions éparses résultant principalement du régime hydrologique. L’action de la géomorphologie sur la distribution de l’eau est à l’origine des formations édapho-climaciques, telles que les mosaïques du Chaco humide. Ces mosaïques, typiques du PNRP, sont constituées principalement 1- d’une savane-palmeraie à Copernicia australis Becc. et C. alba Morong. (« pastizal »), 2- d’une forêt xéromorphe (ou xérophile) naturellement fragmentée, jamais inondée, à Aspidosperma quebracho-blanco (Schltdl., 1861) (« monte fuerte »), avec une couverture au sol de Broméliacées (Aechmea distichantha Lemaire et Pseudananas sagenarius (Arruda) Camargo). 3- de zones herbacées périodiquement inondées où dominent Eleocharis elegans (Kunth) Roem. & Schult. et Sesbania virgata (Cav.) Poir. (« esteros ») et 4- d’une forêt galerie hygrophile bordant le fleuve où

9 INTRODUCTION dominent des espèces telles que Calycophyllum multiflorum (Griseb.) Castelo, Pisonia zapallo (Griseb.) ou Chlorophora tinctoria (L.) Benth. & Hook. (« bosque galeria ») (Ramella et Spichiger 1989, Pujalte et al. 1995, Devillers and Devillers-Terschuren 1996) (Figure 0-2). Les Broméliacées, associées aux zones convexes du relief produites par l’histoire géologique et fluviale du Chaco, (Morello et Adámoli 1974, Popolizio et al. 1980, Barberis et al. 1998) peuvent donc être distribuées de manière sporadique ou continue dans les fragments forestiers. On y a recensé 53 espèces de mammifères parmi lesquelles les tatous, le grand fourmilier, le jaguar, le puma, le koati, les opossums, le tapir, le pécari, le raton laveur, le singe hurleur ainsi que 15 espèces de chauves-souris (Heinonen Fortabat 2001). On y dénombre 191 espèces d’oiseaux (Lopez 1997), 30 espèces d’amphibiens, 35 espèces de poissons (Lanfiutti 2000). Sur l’ensemble de ces espèces, 22 sont protégées (Pujalte et al. 1995).

10 INTRODUCTION

Figure 0-1: Le Gran Chaco argentin comprend le Chaco humide plus à l’est (1000 à 1400mm de précipitation moyenne annuelle) et le Chaco sec (350 à 1000mm) (A). Le site d’étude, le Parc

National Rio Pilcomayo (B), est situé dans le Chaco humide au nord-est de l’Argentine dans la province de Formosa. Les fragments forestiers (en gris) sont entourés par la savane (en blanc).

11 INTRODUCTION

12 INTRODUCTION

Figure 0-2: Formations végétales du Parc National Río Pilcomayo

A) Pastizal: une savane-palmeraie à Copernicia australis Becc. et C. alba Morong., B) Monte Fuerte: une forêt xéromorphe (ou xérophile) naturellement fragmentée, jamais inondée, à Schinopsis balansae Engler et Aspidosperma quebracho-blanco (Schltdl., 1861), C) Esteros: zones herbacées périodiquement inondées où dominent Eleocharis elegans (Kunth) Roem. & Schult. et Sesbania virgata (Cav.) Poir., et D) Bosque Galeria: une forêt galerie hygrophile bordant le fleuve où dominent des espèces telles que Calycophyllum multiflorum (Griseb.) Castelo, Pisonia zapallo (Griseb.) ou Chlorophora tinctoria (L.) Benth. & Hook.

13 INTRODUCTION

4.2.2 Climat Le Parc est situé dans une zone de climat subtropical continental. La température moyenne annuelle est de 23C° et la moyenne annuelle des précipitations est de 1200 mm (Figure 0-3). En été (de décembre à mars), les températures peuvent grimper au delà de 40°C et les hivers ne sont pas exempts de jours avec des températures inférieures à 0°C. Les températures les plus chaudes coïncident avec la période des pluies, de décembre à avril (Ramella et Spichiger 1989).

180 90 Précipitation 160 80 Températures )

140 70 ) m C ° m ( (

s s

120 60 e e n n n n e e y

y 100 50 o o m m

s s e

n 80 40 r o u i t t a a r t i 60 30 é p p i c m é e r T

P 40 20

20 10

0 0 J F M A M J J A S O N D Mois

Figure 0-3: Températures et précipitations mensuelles moyennes relevées en 2001 (provenant de la station de mesure dans le Parc National Río Pilcomayo à Laguna Blanca).

14 INTRODUCTION

4.2.3 Feux et inondations : Les feux d’origine naturelle ou contrôlée représentent la principale perturbation dans le PN Río Pilcomayo, ils maintiennent la transition nette entre les fragments forestiers et la matrice environnante et empêchent la progression de la savane dans les autres habitats (forêt galerie, zones herbacées inondées, fragments forestiers). Cependant, les effets de ces feux sur la dynamique des espèces végétales et animales sont très peu connus. Seuls les aspects généraux à l’échelle régionale ont été développés par Morello (1970). Dans d’autres zones du Chaco humide, les conclusions d’études sur les effets des feux sur la faune sont très diverses, certains auteurs les considèrent extrêmement destructeurs tandis que d’autres ont démontré leurs effets bénéfiques sur la diversité des espèces. Les inondations constituent la deuxième perturbation naturelle survenant dans la savane du Parc. Certaines zones sont inondées en permanence, tandis que d’autres zones de savane sont inondées périodiquement suite à des pluies et des crues intenses des fleuves Pilcomayo et Paraguay. Peu d’informations existent sur les stratégies développées par la faune et la flore pour résister à ces inondations prolongées (2 à 3 mois). Les espèces mobiles et opportunistes peuvent retirer le bénéfices qu’apportent ces inondations sans que leur population ne déclinent, tandis que les espèces terricoles et sédentaires doivent être affectés par une réduction drastique en habitats favorables lors d’inondations pouvant durer plusieurs mois consécutifs. Leurs effectifs doivent probablement fortement osciller entre les périodes de sécheresse et d’inondation (Pujalte et al. 1995).

15 INTRODUCTION

5. Objectifs de la thèse

Le parc national Río Pilcomayo offre, par sa configuration, son histoire et sa localisation, l’opportunité d’étudier les effets de la fragmentation naturelle à long terme sur la distribution des fourmis du sol. Pour y parvenir, les patrons de distributions des fourmis du sol ont été étudiés à différentes échelles spatio-temporelles :

À l’échelle du microhabitat (quadrat de 1m²), nous avons échantillonné les fourmis du sol au sein d’un fragment forestier afin de comprendre les facteurs (biotiques et abiotiques) qui régissaient la structure de l’assemblage des fourmis à fine échelle. Plus spécifiquement, les trois premiers chapitres constituant cette thèse ont comme thème:

CH1 : “Scale dependence of diversity measures in a leaf litter ant assemblage”

Ce premier chapitre a pour but de calibrer le protocole d’échantillonnage standardisé (dit « ALL » ie “Ants of the Leaf Litter” Agosti et Alonso 2000) en conditions de forêt subtropicale sèche. L’utilisation d’un protocole standardisé et calibré est indispensable, avant tout, pour permettre la comparaison d’inventaires réalisés par différents chercheurs à des endroits différents. Ensuite, il permet de mener des analyses sur les patrons de distribution d’espèces et d’en comprendre la dynamique. Plus précisément, nous avons voulu répondre aux questions suivantes : - Quelle part de la faune locale est réellement collectée par ce protocole ALL ? - Toutes les espèces caractéristiques (espèces dominantes numériquement, groupes fonctionnels,....) sont-elles inclues dans un transect ALL ? - Quels sont les indices de diversité les plus appropriés? - Comment comparer des résultats obtenus à différents moments de l’année (effet saisonnier) ?

CH2 : “Spatial structure of litter-dwelling ant distribution in a subtropical dry forest”

Comprendre la distribution spatiale des espèces est essentiel pour caractériser la structure des assemblages, pour optimiser les inventaires d’espèces et pour évaluer l’impact des variables biotiques et abiotiques. Le but ici est d’étudier, à fine échelle spatiale, la structure de l’assemblage des fourmis terricoles et déterminer les facteurs biotiques et abiotiques majeurs structurant cet assemblage forestier.

16 INTRODUCTION

CH3 : “Terrestrial bromeliads affect the ground-dwelling ant distribution in a chacoan forests”.

Ce chapitre a pour but d’évaluer l’impact des broméliacées terrestres sur la distribution des fourmis. Ces broméliacées forment une couverture végétale au sol quasi continue. Cependant, certaines zones de la forêt en sont dépourvues. L’effet structurant de la présence et de la densité de cet élément majeur du micro-habitat a donc pu être étudié.

À l’échelle de l’habitat, nous avons comparé la distribution des fourmis du sol entre fragments de taille, forme et degré d’isolement divers. Plus précisément, le

CH4 : “Long-term effects of forest fragmentation on ants of the humid Chaco” a pour but de pallier au déficit de connaissances relatives à l’effet de la fragmentation naturelle sur la distribution des assemblages de fourmis en forêt sèche. Nous avons voulu répondre plus spécifiquement aux questions suivantes :

- Quelle est l’influence de la taille, la forme et du degré d’isolement des fragments forestiers sur la structure de l’assemblage des fourmis terricoles? - Quelle est l’intensité des effets de bord ? - Quelles sont les introgressions d’espèces entre les fragments forestiers et la savane ? - Les effets de la fragmentation sensu stricto (Fahrig 2003) sur les assemblages de fourmis sont-ils aussi important que les effets d’une déforestation. Quelles conclusions peut-on en tirer en matière de conservation ?

Finalement, dans le dernier chapitre, nous comparerons la distribution des fourmis à l’échelle du paysage:

CH5 : “Habitat type and fire effects on a ground-foraging ants assemblage in a fragmented forest of the humid Chaco”. Le but sera d’étudier les différences de structure entre assemblage vivant dans les fragments forestiers et dans la savane environnante. L’effet immédiat d’un feu de savane sur l’assemblage de fourmis sera également abordé.

17 INTRODUCTION

6. Méthodologie

Dans la forêt naturellement fragmentée du Parc National Río Pilcomayo, la myrmécofaune terricole a été échantillonnée lors de 4 missions s’échelonnant de 1999 à 2002, toujours entre juillet et octobre (à la période de transition entre la saison sèche et la saison des pluies (Figure 0-3). Les fourmis ont été récoltées le long de 15 transects localisés dans 11 fragments forestiers de taille, de forme et de degré d’isolement divers. Cinq transects ont été réalisés dans la savane. Les transects le long desquels les fourmis ont été collectées se basent sur le protocole "Ants of the Leaf Litter" (A.L.L, Agosti et Alonso 2000) : il s’agit d’un transect linéaire de 200m de long comprenant 20 points d’échantillonnage espacés entre eux de 10m. Cependant, la longueur des transects ainsi que l’intervalle entre les échantillons ont été modifiés pour répondre à certains objectifs particuliers (calibrage de l’échantillonnage (CH1), étude de la structure spatiale de la distribution des fourmis (CH2), étude de l’effet de la fragmentation du milieu (CH4). Les particularités méthodologiques propres à chaque objectif seront développées en détail dans les différents articles. À chaque point d’échantillonnage, les fourmis sont collectées par piège à fosse et par mini- Winkler. Au total, 540 échantillons obtenus par extraction mini-Winkler et 465 pièges à fosses ont été récoltés. Le mini-Winkler est un dispositif permettant d'extraire la faune des litières. Pour ce faire, la litière de feuilles mortes présente à l'intérieur d'un quadrat de 1m² est ramassée, tamisée et placée dans un sac en tulle suspendu dans un sac en coton (mini- Winkler) pendant 24h (Fisher, 1998). Les arthropodes se déplacent à l’intérieur du sac afin de tenter de fuir la dessiccation progressive de la litière et tombent dans un gobelet rempli d’alcool disposé en bas du mini-Winkler (Figure 0-4). Les échantillons de litière tamisée ont toujours été récoltés les jours sans pluie et l’après-midi (14h-17h) étant donné que l’activité des fourmis est fortement liée à la température. Les pièges à fosse ont, quant à eux, toujours été placés en matinée et laissés en place durant 24 heures. Deux autres méthodes d’échantillonnage ont été utilisées afin de compléter l’inventaire d’espèces et d’obtenir des informations sur la biologie des espèces. Nous avons étudié la myrmécofaune hypogée dans 100 échantillons de sol (cube de 2,25 dm3) et récolté par chasse à vue toutes les espèces trouvées dans des nids, du bois mort ou tout autre micro-habitat favorable.

18 INTRODUCTION

Figure 0-4: Protocole d’échantillonnage des fourmis terricoles utilise dans le cadre de cette thèse La myrmécofaune de 11 fragments forestiers et de la savane furent échantillonnées entre 1999 et 2002. L'échantillonnage se fit toujours en septembre ou novembre afin d'éviter les températures extrêmes connues pour modifier l'activité des fourmis dans le Gran Chaco (Bestelmeyer, 2000). Notre protocole se base sur le protocole "Ants of the Leaf Litter" (A.L.L.) qui correspond à 20 points d’échantillonnage espacés de 10m. Le nombre d’échantillons et l’étendue du transect varie en fonction des objectifs. Les fourmis sont collectées par piège à fosse ("24h-pitfall trap") et par Winkler. Le Winkler est un dispositif permettant d'extraire la faune des litières. Pour cela, la litière de feuilles mortes présente à l'intérieur d'un quadrat de 1m2 est ramassée, tamisée et placée dans un sac mini-Winkler pendant 24h (Fisher, 1998). L'extraction des insectes se fait ensuite de façon passive: ils se déplacent, notamment afin de tenter de fuir la dessiccation progressive de la litière, et tombent dans un sachet rempli d'alcool disposé au bas du sac mini-Winkler.

19 INTRODUCTION

Bibliographie Adis, J., Ribeiro, E.F., deMorais, J.W., Cavalcante, E.T.S. (1989). Vertical distribution and abundance of arthropods from white sand soil of a Neotropical Campinarana Forest during the dry season. Studies on Neotropical Fauna Environment 24: 201-211. Adler, F.R., Lebrun, E.G., Feener, D.H.Jr. (2007) Maintaining diversity in an ant community: modeling, extending, and testing the dominance-discovery trade-off. The American Naturalist 169: 323-333. Agosti, D., Alonso, L.E. (2000) The ALL protocol. A standard protocol for the collection of ground dwelling ants. In: Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R. (eds), Ants. Standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press, pp. 204-206. Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R. (2000) Ants. Standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press. Agosti, D., Johnson, N.F. (2005) Editors Antbase. www.antbase.org Version (05/2005). Alonso, L.E. (2000) Ants as indicators of diversity. In: Agosti, D. et al. (eds), Ants. Standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press, pp. 80- 88. Alonso, L.E., Agosti, D. (2000) Biodiversity studies, monitoring, and ants: an overview. In: Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R.(eds), Ants. Standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press, pp. 1-8. Ambrecht, I., Ulloa-Chacón, P. (1999) Rareza y diversidad des hormigas en fragmentos de bosque seco colombianos y sus matrices. Biotropica 31: 646-653. Andersen, A.N. (1995) A classification of Australian ant communities, based on functional groups which parallel plant life-forms in relation to stress and disturbance. Journal of Biogeography 22: 15-29. Andersen, A.N. (1997) Using ants as bioindicators: multiscale issues in ant community ecology. Conservation Ecology B www.consecol.org/Journal/vol1/iss1/art8/. Andersen, A.N., Sparling, G.P. (1997) Ants as indicators of restoration success: relationship with soil microbial biomass in the Australian seasonal tropics. Restoration Ecology 5: 109- 114. Andersen, A.N., Hertog, T., Woinarski, J.C.Z. (2006) Long-term fire exclusion and ant community structure in an Australian tropical savanna: congruence with vegetation succession. Journal of Biogeography 33: 823-832.

20 INTRODUCTION

Arnan, X. Rodrigo, A., Retana, J. (2006) Post-fire recovery of Mediterranean ground ant communities follows vegetation and dryness gradients. Journal of Biogeography 33: 1246- 1258. Barberis, I.M., Pire, E.F., Lewis, J.P. (1998) Spatial heterogeneity and woody species distribution in a Schinopsis balansae (Anacardiaceae) forest of the Southern Chaco, Argentina. Revista de Biologia Tropical 46: 515–524. Benson, W.W., Brandão, C.R.F. (1987) Pheidole diversity in the humid Tropics: a survey from Serra dos Carejas, Para, Brazil. - In: Eder, J., Rembold, H. (eds). Chemistry and biology of social insects, Munchen, Verlag J. Peperny, pp. 593-594. Bestelmeyer, B.T. (2000) The trade-off between thermal tolerance and behavioral dominance in a subtropical South American ant community. Journal of Animal Ecology 69: 998-1009. Bestelmeyer, B.T., Wiens, J.A. (1996) The effects of land use on the structure of ground- foraging ant communities in the Argentinean Chaco. Ecological Applications 6: 1225– 1240. Bestelmeyer, B.T., Wiens, J.A. (2001) Local and regional-scale responses of ant diversity to a semi-arid biome transition. Ecography 24: 381–392. Bestelmeyer, B.T., Wiens, J.A. (2003) Scavenging and foraging behavior and variation in the scale of nutrient redistribution among semi-arid grasslands. Journal of Arid Environments 53: 373-386. Bierregaard, R.O. Jr., Lovejoy, T.E., Kapos, V., dos Santos, A.A., Hutchings, R.W. (1992) The biological dynamics of tropical rainforest fragments. BioScience 42: 859-866. Bolton, B. (1994) Identification guide to the ant genera of the World. Harvard University Press. Bolton, B. (1995) A new general catalogue of the ants of the World. Harvard University Press. Bolton, B. (2003) Synopsis and classification of Formicidae. Memoirs of the American Entomological Institute Vol. 71, Gainesville, FL. 370pp. Bolton, B., Alpert, G., Ward, P.S., Naskrecki, P. (2007) Bolton's catalogue of ants of the world: 1758- 2005. Harvard University Press. Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G. & Hilton-Taylor, C. (2002) Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology 16: 909–923. Brown, K.S. Jr., Hutchings, R.W. (1997) Disturbance, fragmentation, and the dynamics of diversity in Amazonian forest butterflies. In: W.F. Laurance and R.O. Bierregaard Jr., (eds). Tropical forest remnants. Ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago. pp 91-110.

21 INTRODUCTION

Brühl, C.A. (1996). Ameisenzönosen in einem Primärwaldgradienten am Mount Kinabalu, Borneo: Höhenverbreitung und Stratifizierung. Diploma Thesis, Julius-Maximilians- Universität, Würzburg. Brühl, C.A., Gunsalam, G., Linsenmair, K.E. (1998) Stratification of ants (Hymenoptera, Formicidae) in a primary rain forest in Sabah, Borneo. Journal of Tropical Ecology 14: 285- 297. Brühl, C.A., Eltz, T., Linsenmair, K.E. (2003) Size does matter - effects of tropical rainforest fragmentation on the leaf litter ant community in Sabah, Malaysia. Biodiversity and Conservation 12: 1371-1389. Burghouts, T.B.A., Campbell, E.J.F., Kolderman, P.J. (1994) Effects of tree species heterogeneity on leaf fall in primary and logged dipterocarp forest in the Ulu Segama Forest Reserve, Sabah, Malaysia. Journal of Tropical Ecology 10: 1-26. Byrne, M.M. (1994) Ecology of twig-dwelling ants in a wet lowland tropical forest. Biotropica 26: 61–72. Carvalho, K.S., Vasconcelos, H.L. (1999) Forest fragmentation in central Amazonia et its effects on litter-dwelling ants. Biological Conservation 91: 151-157. Condon, M.A., Scheffer, S.J., Lewis, M.L., Swensen, S.M. (2008) Hidden neotropical diversity: greater than the sum of its parts. Science 320: 928-931. Debinski, D.M., Holt, R.D. (2000) A survey and overview of habitat fragmentation experiments. Conservation Biology 14: 342-355. Deblauwe, I., Dekoninck, W. (2007a) Diversity and distribution of ground-dwelling ants in a lowland rainforest in southeast Cameroon. Insectes Sociaux 54: 334-342. Deblauwe, I., Dekoninck, W. (2007b) Spatio-temporal patterns of ground-dwelling ants in a lowland rainforest in southeast Cameroon. Insectes Sociaux 54: 343-350. Dejean, A., Corbara, B., Leponce, M., Orivel, J. (2007) Ants of rainforest canopies: the implications of territoriality and predatory behavior. Functional Ecosystems and Communities 1: 105-120 Delabie, J.H.C., Agosti, D., do Nascimento, I.C. (2000) Litter ant communities of the Brazilian Atlantic rain forest region. In: D. Agosti, J. Majer, L.E. Alonso, Schultz, T. (eds) Sampling Ground-dwelling ants: case studies from the worlds’ rain forests. Curtin University, Australia, School of Environmental Biology, Bulletin n° 18: 1-17. Delsinne, T., Roisin, Y., Leponce, M. (2007) Spatial and temporal foraging overlaps in a Chacoan ground-foraging ant assemblage. Journal of Arid Environment 71: 29-44.

22 INTRODUCTION

De Souza O., Brown, V.K. (1994) Effects of habitat fragmentation on Amazonian termite communities. Journal of Tropical Ecology 10: 197–206. Devillers, P. and Devillers-Terschuren, J. 1996. A classification of South American habitats. IRSNB, Brussels, and Institute of Terrestrial Ecology, Huntingdon. Didham, R.K. (1997a) The influence of edge effects and forest fragmentation on leaf litter invertebrates in Central Amazonia. In Laurence, W.F.; Bierregaard, R.O. Jr., editors. Tropical forest remnants. Ecology, management, and conservation of fragmented communities. The University of Chicago Press Chicago. pp. 55-70. Didham, R.K. (1997b) An overview of invertebrate responses to forest fragmentation. In Watt, A.D.; Stork, N.E.; Hunter, M.D, editors. Forests and Insects. Chapman and Hall: 303-320. Didham, R.K. (1998). Altered leaf-litter decomposition rates in tropical forest fragment. Oecologia 116: 397-406. Didham, R.K., Lawton J.H. (1999) Edge structure determines the magnitude of changes in microclimate and vegetation structure in tropical forest fragments. Biotropica 31: 17-30. Didham, R.K., Tylianakis, J.M., Hutchison, M.A., Ewers, R.M., Gemmel, N.J. (2005) Are invasive species the drivers of ecological change? Trends in Ecology and Evolution 20: 470-474. Duchaufour, P. (2001) Introduction à la Science du Sol. Sol, végétation , environnement. Dunod, 331 p. Dunn, R.R. (2005) Modern insect extinctions, the neglected majority. Conservation Biology 19: 1030-1036. Echeverria, C., Coomes, D., Salas, J., Rey Benayas, J.M., Lara, A., Newton, A. (2006) Rapid deforestation and fragmentation of Chilean temperate forest. Biological Conservation 130: 481-494. Echeverria, C. , Newton, A.C., Lara, A., Rey Benayas, J.M., Coomes, D.A. (2007) Impacts of forest fragmentation on species composition and forest structure in the temperate landscape of southern Chile. Global Ecology and Biogeography 16: 426-439. Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual Reviews of Ecology and Systematics 34: 487-515. Farji-Brener A.G., Corley J.C., Bettinelli, J. (2002) The effects of fire on ant communities in north-western Patagonia: the importance of habitat structure and regional context. Diversity and Distribution 8: 235-243.

23 INTRODUCTION

Fauth, J.E., Bernardo, J., Camara, M., Resetarits, W.J.Jr., van Buskirk, J., McCollum, S.A. (1996) Simplifying the jargon of community ecology: a conceptual approach. The American Naturalist 147: 282-286. Feener, D.H., Schupp, E.W. (1998) Effect of treefall gaps on the patchiness and species richness of Neotropical ant assemblages. Oecologia 116: 191-201. Fittkau, E.J., Klinge, H. (1973) On biomass and trophic structure of the central Amazonian rain forest ecosystem. Biotropica 5: 2-14. Fisher, B.L., 1998. Ant diversity patterns along an elevational gradient in the Réserve Spéciale d'Anjanaharibe-Sud and on the Western Masoala Peninsula, Madagascar. Fieldiana Zoology 90: 39-67. Folgarait, P. (1998) Ant biodiversity and its relationship to ecosystem functioning: review. Biodiversity and Conservation 7: 1221-1244. Franks, N.R. (1982) Social insects in the aftermath of swarm raids of the army ant Eciton burchelli. In: Breed, M.D., C.D. Michener and H.E. Evans (eds).The Biology of Social Insects. Westview Press, Boulder, CO, pp. 275–279. Franks N.R., Bossert, W.H. (1983) The influence of swarm raiding army ants on the patchiness and diversity of tropical leaf litter ant community. In: Sutton, E.L., T.C. Whitmore and A.C. Chadwick (eds). Tropical Rain Forest: Ecology and Management Blackwell, Oxford, pp. 151–163. French, K. (1999) Spatial variability in species composition in birds and insects. Journal of Insect Conservation 3: 183-189. Furley, P.A., Proctor, J., Ratter, J.A. (1992) Nature and diversity of forest-savanna boundaries. Chapman & Hall, London, 644 p. Gadagkar, R., Nair, P. Chandrashekhar, K., Bhat, D. M. (1993) Ant species richness in some selected localities in Western Ghats, India. Hexapoda 5:70-94. Gascon, C. 1993. Breeding habitat use by Amazonian primary-forest frog species at the forest edge. Biodiversity and Conservation 2: 438-444. Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L., Laurance, W.F., Zimmerman, B., Tocher, M., Borges, S. (1999) Matrix habitat and species richness in tropical forest remnants. Biological Conservation 91: 223-229. Gibbs, JP, Stanton, EJ. (2001) Habitat fragmentation and arthropod community change: carrion beetles, phoretic mites, and flies. Ecological Application 11:79–85. Gotelli, N.J. (1993) Ant lion zones: causes of high-density predator aggregations. Ecology 74: 226-237.

24 INTRODUCTION

Gotelli, N.J. (1996) Ant community structure: effects of predatory ant lions. Ecology 77: 30-638 Heinonen-Fortabat, S. (2001) Los mamíferos en Parque Nacional Río Pilcomayo, Provincia de Formosa, Argentina. Facmua, Vol 17. Herbers, J.M. (1989) Community structure in north temperate ants: temporal and spatial variation. Oecologia 81: 201–211. Hirosawa, H., Higashi, S., Mohamed, M. (2000) Food habits of Aenictus army ants and their effects on the ant community in a rain forest Borneo. Insectes Sociaux 47: 42–49. Höfer, H., Martius, C., Beck, L. (1996) Decomposition in an Amazonian rain forest after experimental litter addition in small plots. Pedobiologia 40: 570–576. Hoffmann, B.D. (2003) Responses of ant communities to expermiental fire regimes on rangelands in the Victoria River District of the Northern territory. Austral Ecology 28: 182- 195. Hölldobler, B., Wilson, E.O. (1990) The Ants. Cambridge: Bellknap Press. 732pp. Janzen, D. (1988) Tropical dry forests. The most endangered major tropical system. In: Wilson, E.O., (ed). Biodiversity, National Academy of Sciences/Smithsonian institution, Washington, D.C, pp. 130-137. Jones, C.G., Lawton, J.H., Shachak, M. (1994) Organisms as ecosystem engineers. Oikos 69: 373-386. Kalif, K.A.B., Azevedo-Ramos, C., Moutinho, P., Malcher, S.A.O. (2001) The effect of logging on the ground foraging ant community in eastern Amazonia. Studies in Neotropical Fauna and Environment 36: 215-219. Kapos, V., Wandelli, E., Camargo, J.L., Ganade, G. (1997) Edge-related changes in environment and plant responses due to forest fragmentation in central Amazonia. In: Laurance, W.F., Bierregaard Jr., R.O. (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago. Pp. 33-44. Kaspari, M. (1996a) Litter ant patchiness at 1-m² scale: disturbance dynamics in three Neotropical forests. Oecologia 107: 265-173. Kaspari M. (1996b) Testing resource-based models of patchiness in four Neotropical litter ant assemblages. Oikos 76: 443–454. Kaspari, M. (2000) A primer on ant ecology. In: Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R.(eds). Ants. Standard methods for measuring and monitoring biodiversity. Smithsonian Institution Press, pp. 9-24.

25 INTRODUCTION

Kaspari M. and Majer J.D. (2000) Using ants to monitor environmental changes. In: Agosti, D., J. Majer, L.A. Alonso and T. Schultz (eds) Ants: Standard Methods for Measuring and Monitoring Biodiversity. Smithsonian Institution Press. Pp. 89–98. Kellert, S.R. (1993) Values and perceptions of invertebrates. Conservation Biology 7: 845-855. Klein, B.C. (1989) Effects of forest fragmentation on dung and carrion beetlle communities in central Amazonia. Ecology 70: 1715-1725. Kotze, D.J., Samways, M.J. (2001). No general edge effects for invertebrates at afromontane forest/grassland ecotones. Biodiversity and Conservation 10: 443-466. Lanfiutti A. (2000) Actualización del listado de peces, anfibios, reptiles y mamíferos del Parque Nacional Río Pilcomayo, APN. LaSalle, J. and Gauld, I.D. (1993) Hymenoptera: their diversity and their impact on the diversity of other organisms. pp. 1-27. In Hymenoptera and Biodiversity. LaSalle, J. and I. D. Gauld (editors) CAB International, Wallingford UK 368 Pp. Lassau, S.A., Hochulli, D.F. (2004) Effects of habitat complexity on ant assemblages. Ecography 27: 157-164. Laurance W.F (2004) Forest–climate interactions in fragmented tropical landscapes. Philosophical Transactions of the Royal Society B 359: 345-352. Laurance, W.F. (2008) Theory meets reality: How habitat fragmentation research has transcended island biogeographic theory. Biological Conservation 141: 1731-1744. Laurance, W.F., Ferreira, L.V., Rankin-de Merona, J.M., Laurance, S.G. (1998) Rain forest fragmentation and the dynamics of Amazonian tree communities. Ecology 79: 2032-2040. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L, Brunia, E.M, Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G., Sampaio, E. (2002) Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16: 605-618. Laurance, W.F., Vasconcelos, H.L. (2004) Ecological effects of habitat fragmentation in the tropics. In: Schroth, G., Fonseca, G.A.B., Harvey, C., Gascon, C., Vasconcelos, H.L., Izac, A.M. (eds.) Agroforestry and biodiversty conservation in tropical landscapes, pp. 33-49, Island Press, Washington, DC. Lavelle, P., Kohlmann, B. (1984) Étude quantitative de la macro faune du sol dans une forêt tropicale humide du Mexique (Bonampak, Chiapas). Pedobiologia 27: 377-393. Lebrun, E.G., Feener, D.H.Jr. (2007) When trade-offs interact: balance of terror enforces dominance discovery trade-off in a local ant assemblage. Journal of Animal Ecology 76: 58-64. Levin, S.A. (1992) The problem of pattern and scale in ecology. Ecology 73: 1943-1967.

26 INTRODUCTION

Levings, S.C. (1983) Seasonal, annual et among-site variation in the ground ant community of a deciduous tropical forest: some causes of patchy species distributions. Ecological Monographs 53: 435-455. Levings, S.C., Franks, N.R. (1982) Patterns of nest dispersion in a tropical ground ant community. Ecology 63: 338-344. Levings, S.C., Traniello, J.F.A. (1981) Territoriality, nest dispersion, and community structure in ants. Psyche 88: 265-319. Levings S.C., Windsor, D.M. (1984) Litter moisture content as a determinant of litter arthropod distribution and abundance during the dry season on Barro Colorado Island, Panama. Biotropica 16: 125–131. López, L. (1997) Inventario de las aves de Parque Nacional Río Pilcomayo. Lovejoy, T.E., Bierregaard, R.O. Jr., Rankin, J.M. and Schubart, H.O.R.. (1983) Ecological dynamics of forest fragments. In: Sutton, S.L, Whitmore, T.C., Chadwick, A.C. (eds). Tropical rain forest : ecology et management. Blackwell Scientific, Oxford, United Kingdom pp. 377-384. Lovejoy, T.E., Bierregaard, R.O. Jr., Rylands, A.B., Malcolm, J.R., Quintela, C.E., Harper, L.H., Brown, K.S.Jr., Powell, A.H., Powell, G.V.N., Schubart, H.O., Hays, M.B. (1986) In : M.E. Soulé (ed). Edge and other effects of isolation on Amazon forest fragments. pp. 257-285. Sinauer, Sunderland, Massachussets. Majer, J.D. (1983) Ants: bio-indicators of minesite rehabilitation, land-use, and land conservation. Environmental Management 7: 375-383. Majer, J.D. (1993). Comparison of the arboreal ant mosaic in Ghana, Brazil, Papua New Guinea and Australia - its structure and influence on arthropod diversity. In: La Salle, J., Gauld , I. D. (eds.) Hymenoptera and biodiversity, CAB International. pp. 115-141. Majer, J.D., Day, J.E., Kabay, E.D., Perriman, W.S. (1984) Recolonisation by ants in bauxite mines rehabilitated by a number of different methods. Journal of Applied Ecology 21: 355- 375. Majer, J.D., Delabie, J.H.C., Mckenzie, N.L. (1997) Ant litter fauna of forest, forest edges and adjacent grassland in the Atlantic rain forest region of Bahia, Brazil. Insectes Sociaux 44: 255-266. Malcolm, J.R. (1991) The small mammals of Amazonian forest fragments: pattern and process. PhD thesis. University of Florida, Gainesville. Malcolm, J.R. (1994) Edge effects in Central Amazonian forest fragments. Ecology 75: 2438- 2445.

27 INTRODUCTION

McKinney, M.L. (1999) High rates of extinction and threats in poorly studied taxa. Conservation Biology 13: 1273-1281. Miles, L., Newton, A.C., DeFries, R.S., Ravilious, C., May, I., Blyth, S., Kapos, V., Gordon, J.E. (2006) A global overview of the conservation status of tropical dry forests. Journal of Biogeography 33: 431-505. Millenium Ecosystem Assessment [MEA] (2005) Ecosystems and human well-being: biodiversity synthesis. World Resource Institute, Washington, D.C., 86 pp. Morello, J. (1970) Modelo de relaciones entre pastizales y leñosas colonizadoras en el Chaco Argentino. IDIA : 31-52. Morello, J., Adámoli, J. (1974) Las grandes Unidades de vegetación y ambiente del Chaco Argentino. 2° parte: Vegetación y ambiente en la provincia del Chaco. INTA. Serie fotogeográfica N° 13, pp. 130. Moutinho, P.R.S. (1998) Impactos do uso da terra sobre a fauna de formigas, consequências para recuperação florestal na Amazônia Oriental. In: Gascon, C., Moutinho, P. (Eds) Floresta Amazônica: dinâmica, regeneração e manejo, Manaus, MCT-INPA, pp. 155–170. New, T.R. (1995) Introduction to invertebrate conservation biology. Oxford Univ. Press. Noss, A., Mares, M.A., Diaz, M.M. (2002) The Chaco In: Mittermeier, R.A., Mittermeier, C.G., Robles-Gil, P., Pilgrim, J.D., Fonseca, Brooks, T.M., Konstant, W.R. (eds): Wilderness. Earth’s last wild places, pp. 165-172. Cemex, Mexico City. Oliver, I., Beattie, A.J., York, A. (1998) Spatial fidelity of plant, vertebrate, and invertebrate assemblages in multiple-use forest in Eastern Australia. Conservation Biology 12: 822-835. Olson, D.M. (1994) The distribution of leaf litter invertebrates along a Neotropical altitudinal gradient. Journal of tropical Ecology 10: 129-150. Olson, D.M., Dinerstein, E. (2002) The global 200: priority ecoregions for global conservation. Annals of the Missouri Botanical Garden 89: 199-224. Palladini, J.D., Jones, M.G., Sanders, N.J., Jules, E.S. (2007) The recovery of ant communities in regenerating temperate conifer forests. Forest Ecology and Management 242: 619-624. Peck, S.L., McQuaid, B., Campbell, C.L. (1998) Using ant species (Hymenoptera: Formicidae) as a biological indicator of agroecosystem condition. Environmental Entomology 27: 1102-1110. Perfecto, I. (1994) Foraging behavior as a determinant of asymmetric competitive interaction between two ant species in a tropical agroecosystem. Oecologia 98: 184-192. Pimm, S.L., Russell, G.J., Gittleman, J.L., Brooks, T.M. (1995) The future of biodiversity. Science 269: 347-350.

28 INTRODUCTION

Popolizio, E., Serra, P., Hortt, G. (1980) Domo Central de la Provincia del Chaco con Bosques y Sabanas Secos - Unidad 1.3.2, Centro de Geociencias Aplicadas: Investigación 3: 39-75. Powell, A.H., Powell, G.V.N. (1987) Population dynamics of male euglossine bees in Amazonian forest. Biotropica 19: 176,179. Prance, G.T. (2006) Tropical savannas and seasonally dry forests: an introduction. Journal of Biogeography 33: 385-386. Pujalte, J.C., Reca, A.R., Balabusic, A., Canevari, P., Cusato, L., Fleming, V.P. (1995) Anales de parques nacionales. Unidades Ecológicas del parque nacional Rio Pilcomayo. Administración de Parques Nacionales XVI: 1-185. Ramella L., Spichiger R. (1989) Interpretación preliminary del medio físico y de la vegetación del Chaco Boreal – Contribución al studio de la flora y de la vegetación del Chaco. I. Candollea 44: 639-680. Redford, K.H., Taber, A., Simonetti, J.A. (1990) There is more to biodiversity than the tropical rain forests. Conservation Biology 4: 328-330. Retana, J., Cerdà, X. (2000) Patterns of diversity and composition of Mediterranean ant communities tracking spatial and temporal variability in the thermal environment. Oecologia 123: 436-444. Ribas, C.R., Schoereder, J.H., Pic, M., Soares, S.M. (2003) Tree heterogeneity, resource availability, and larger scale processes regulating arboreal ant species richness. Austral ecology 28: 305-314. Ricketts, T.H. (2001) The matrix matters: effective isolation in fragmented landscapes. The American Naturalist 158: 87-99. Samways, M.J. (2007) Insect conservation: a synthetic management approach. Annual Review of Entomology 52: 465-487. Sánchez-Azofeifa, A.G., Kalacska, M., Quesada, M., Calvo-Alvarado, J.C., Nassar, J.M., Rodríguez, J.P. (2005a) Need for integrated research for a sustainable future in tropical dry forests. Conservation Biology 19: 285-286. Sánchez-Azofeifa, A.G., Quesada, M., Rodríguez, J.P., Nassar, J.M., Stoner, K.E., Castillo, A., Garvin, T., Zent, E.L., Calvo-Alvarado, J.C., Kalacska, M., Fajardo, L., Gamon, J.A., Cuevas-Reyes, P. (2005b) Research priorities for neotropical dry forests. Biotropica 27: 477-485. Sanders, NJ., Gotelli, N.J., Heller, N.E., Gordon, D.M. (2003) Community diassembly by an invasive species. The Proceedings of the National Academy of Sciences 100: 2474-2477.

29 INTRODUCTION

Schluter, D., and R. E. Ricklefs (1993) Species diversity: an introduction to the problem. In: Ricklefs, R. E., Schluter, D. (eds) Species diversity in ecological communities: historical et geographical perspectives. University of Chicago Press, Chicago, Illinois, USA, pp.1-10. Schultz, T.R., McGlynn, T.P. (2000) the interactions of ants with other organisms. In: Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R. (eds). Ants: standard methods for measuring and monitoring biodiversity, pp. 35-44. Smithsonian Institution Press, Washington, D.C. Silva, R.R., Machado Feitosa, R.S., Eberhardt, F. (2007) Reduced ant diversity along a habitat regeneration gradient in the southern Brazilian Atlantic Forest. Forest Ecology and Management 240: 61-69. Soares, S.M. and Schoereder J.H. (2001) Ant nest distribution in a remnant of tropical rainforest in southeastern Brazil. Insectes Sociaux 48: 280–286. Stratford, J.A., Stouffer, P.C.. (1999) Local extinctions of terrestrial insectivorous birds in a fragmented landscape near Manaus, Brazil. Conservation Biology 13: 1416-1423. Suarez, A.V., Bolger, D.T., Case, T.J. (1998) Effects of fragmentation and invasion on native ant communities in coastal southern California. Ecology 79: 2041-2056. Theunis, L., Gilbert, M., Roisin, Y., Leponce, M. (2005) Spatial structure of litter-dwelling ant distribution in a subtropical dry forest. Insectes Sociaux 52: 366-377. Thomas, C.D., Cameron, A., Green, R.E., Bakkenes, M., Beaumont, L.J., Collingham, Y.C., Erasmus, B.F.N., Ferreira de Siqueira, M., Grainger, A., Hannah, L., Hugues, L., Huntley, B., van Jaarsveld, A.S., Midgley, G.F., Miles, L., Ortega-Huerta, M.A., Peterson, A.T., Phillips, O.L., Williams, E. (2004) Extinction risk from climate change. Nature 427: 145- 148. Travis, J.M. (2003) Climate change and habitat destruction: a deadly anthropogenic cocktail. Proceedings of the Royal Society of London B. 270: 467-473. Underwood, E.C., Fisher, B.L. (2006) The role of ants in conservation monitoring: if, when, and how. Biological Conservation 132: 166-182. Vasconcelos, H.L. (1988) Distribution of Atta (Hymenoptera, Formicidea) in « terra-firme » rain forest of Central Amazonia : density, species composition et preliminary results on effects of forest fragmentation. Acta Amazônica 18: 309-315. Vasconcelos, H.L. (1990) Effects of litter collection by understory palms on the associated macroinvertebrate fauna in Central Amazonia. Pedobiologia 34: 157–160. Vasconcelos, H.L. (1999). Effects of forest disturbance on the structure of ground-foraging ant communities in central Amazonia. Biodiversity and Conservation 8: 409-420.

30 INTRODUCTION

Vasconcelos, H.L., Delabie, J.H.C. (2000) Ground ant communities from central Amazonia forest fragments. Sample Ground-dwelling ants: Case studies from the Worlds' Rain Forests. Curtin University School of Environmental Biology Bulletin, Perth, Australia 18: 59- 70. Vasconcelos, H.L., Carvalho, K.S., Delabie, J.H.C. (2001) Landscape modifications and ant communities. In: Bierregaard, R. O. Jr et al. (eds). Lessons from Amazonia. Yale Univ. Press, pp. 199-207. Vasconcelos H.L., Macedo, A.C.C., Vilhena, J.M.S. (2003) Influence of topography on the distribution of ground-dwelling ants in an Amazonian forest. Studies of Neotropical Fauna and Environment 38: 115–124. Vasconcelos, H.L., Vilhena, J.M.S., Magnusson, W.E., Albernaz, A.L.K.M. (2006) Long-term effects of forest fragmentation on Amazonian ant communities. Journal of Biogeography 33: 13448-1356. Vasconcelos, H.L., Leite, M.F., Vilhena, J.M.S., Lima, A.P., Magnusson, W.E. (2008) Ant diversity in an Amazonian savanna: relationship with vegetation structure, disturbance by fire, and dominant ants. Austral Ecology 33: 221-231. Vepsälainen, K., Pisarki, B. (1982) Assembly of island ant communities. Annales Zoologici Fennici 19: 327-335. Whitford, W.G., Van zee, J., Nash, M.S., Smith, W.E., Herrick J. E. (1999) Ants as indicators of exposure to environmental stressors in North American desert grasslands. Environmental Monitoring and Assessment 54: 143-171. Wilcove D.S., McLellan, C.H., Dobson, A.P. (1986) Habitat fragmentation in the temperate zone. In: ME Soul´e (ed). Conservation Biology, pp. 237–56. Sunderland, MA: Sinauer. Williams, S.E., Marsh, H. and Winter J. (2002) Spatial Scale, Species Diversity, et Habitat Structure: Small Mammals in Australian Tropical Rain Forest. Ecology 83: 1317-1329. Wilson, E.O. (1958) Patchy distribution of ant species in New Guinea rain forests. Psyche 65: 26-38. Wilson, E.O. (1987) Success and dominance in Ecosystems: the case of the social insects. Kinne, O. (ed.), Ecology Institute, Oldendorf, Germany, 104 pp. Wilson, E.O. (2003) L’avenir de la vie. Editions du Seuil, Paris, 285pp. Woodruff, D.S. (2001) Declines of biomes and biotas and the future of evolution. The Proceedings of the National Academy of Sciences 98: 5471-5476.

31 INTRODUCTION

Yeates, D.K., Bouchard, P., Monteith, G.B. (2002) Patterns and levels of endemism in the Australian wet tropics rainforest: evidence from flightless insects. Invertebrate Systematics 16: 605-619. York, A. (2000) Long-term effects of frequent low intensity burning on ant communities in coastal blackbutt forest of southeastern Australia. Austral Ecology 25: 83-98.

32 Chapitre 1: Scale dependence of diversity measures in a leaf-litter ant assemblage

SCALE DEPENDENCE OF DIVERSITY MEASURES IN A LEAF-LITTER ANT

ASSEMBLAGE

MAURICE LEPONCE 1, LAURENCE THEUNIS 1, 2, JACQUES H.C. DELABIE3 AND YVES ROISIN2

1. Biological Evaluation Section, Royal Belgian Institute of Natural Sciences, Rue Vautier 29, B-1000 Brussels, Belgium. 2. Behavioral and Evolutionary Ecology, CP 160/12, Université Libre de Bruxelles, Avenue F.D. Roosevelt 50, B-1050 Brussels, Belgium. 3. Laboratório de Mirmecologia, CEPLAC-UESC, Itabuna, BA, Brazil.

MAURICE LEPONCE

Phone +32 2 627.43.58 Fax +32 2 649.48.25 E-mail: [email protected]

33 ECOGRAPHY 27: 253Á/267, 2004

Scale dependence of diversity measures in a leaf-litter ant assemblage

Maurice Leponce, Laurence Theunis, Jacques H. C. Delabie and Yves Roisin

Leponce, M., Theunis, L., Delabie, J. H. C. and Roisin, Y. 2004. Scale dependence of diversity measures in a leaf-litter ant assemblage. Á/ Ecography 27: 253Á/267.

A reliable characterization of community diversity and composition, necessary to allow inter-site comparisons and to monitor changes, is especially difficult to reach in speciose invertebrate communities. Spatial components of the sampling design (sampling interval, extent and grain) as well as temporal variations of species density affect the measures of diversity (species richness S, Buzas and Gibson’s evenness E and Shannon’s heterogeneity H). Our aim was to document the small-scale spatial distribution of leaf litter ants in a subtropical dry forest of the Argentinian Chaco and analyze how the community characterization was best achieved with a minimal sampling effort. The work was based on the recent standardized protocol for collecting ants of the leaf litter (‘‘A.L.L.’’: 20 samples at intervals of 10 m). To evaluate the consistency of the sampling method in time and space, the selected site was first subject to a preliminary transect, then submitted after a 9-month interval to an 8-fold oversampling campaign (160 samples at interval of 1.25 m). Leaf litter ants were extracted from elementary 1 m2 quadrats with Winkler apparatus. An increase in the number of samples collected increased S and decreased E but did not affect much H. The sampling interval and extent did not affect S and H beyond a distance of 10 m between samples. An increase of the sampling grain had a similar effect on S than a corresponding increase of the number of samples collected, but caused a proportionaly greater increase of H. The density of species m2 varied twofold after a 9-month interval; the effect on S could only be partially corrected by rarefaction. The measure of species numerical dominance was little affected by the season. A single standardized A.L.L. transect with Winkler samples collected B/45% of the species present in the assemblage. All frequent species were included but their relative frequency was not always representative. A log series distribution of species occurrences was oberved. Fisher’s a and Shannon’s H were the most appropriate diversity indexes. The former was useful to rarefy or abundify S and the latter was robust against sample size effects. Both parametric and Sobero´n and Llorente extrapolation methods outperformed non- parametric methods and yielded a fair estimate of total species richness along the transect, a minimum value of S for the habitat sampled.

M. Leponce ([email protected]) and L. Theunis, Sect. of Conservation Biology, Royal Belgian Inst. of Natural Sciences, Rue Vautier 29, B-1000 Brussels, Belgium. Á/ J. H. C. Delabie, Lab. de mirmecologia, CEPLAC-UESC, Itabuna, BA, Brazil. Á/ Y. Roisin, Behavioral and Evolutionary Ecology, CP 160/12, Univ. Libre de Bruxelles, Avenue F.D. Roosevelt 50, B-1050 Brussels, Belgium.

Conservation biologists and environmental planners tories is that species diversity cannot be recorded without need reliable methods to evaluate the biological value reference to space, time and collection method. Compo- of sites and to monitor changes over time. A major nents of species diversity include species richness (S, the difficulty encountered when conducting diversity inven- number of species) and species evenness (E, equitability

Accepted 30 September 2003 Copyright # ECOGRAPHY 2004 ISSN 0906-7590

ECOGRAPHY 27:2 (2004) 253 of the distribution of species abundance). Species The standardized protocol for ground-dwelling ants, diversity can be summarized into a single index of abbreviated ‘‘A.L.L.’’ (Ants of the Leaf Litter), consists 2 heterogeneity such as the Shannon’s index H (H/ in a line-transect with an extent of 200 m and a ln S/ln E) (Hayek and Buzas 1997). To obtain the exact sampling interval of 10 m. The leaf litter fauna from 1 2 values of S, E, and H, complete diversity inventories m quadrats (/sampling grain) is extracted with a mini- should be conducted but are almost an unachievable Winkler apparatus (Fisher 1998, Bestelmeyer et al. goal for invertebrate taxa, especially under the tropics 2000). Other fractions of the local ant fauna can be (Longino and Colwell 1997, Lawton et al. 1998, Longino collected by complementary methods (pitfall activity et al. 2002). Alternatively, structured inventories based traps, soil samples, wood samples and visual search) on limited but well-quantified sampling effort aim to (Agosti et al. 2000). Data sets can be deposited in a reliably characterize communities (Longino and Colwell database available on-line (Agosti 2003). 1997) but depend on the spatial scale considered. The The use of the same sampling effort and methods spatial components of the sampling scheme can be should allow comparisons of inventories conducted by decomposed into sampling grain, interval and extent different research teams at different sites and lead to (Wiens 1989, Palmer and White 1994, Legendre and global analyses of species distribution. However, it may Legendre 1998). Sampling grain is the area concerned by still be difficult to interpret the results of a single each elementary sampling unit. Sampling interval is the inventory since several questions remain open: which distance between individual sampling units. Sampling proportion of the local fauna is really collected, are all extent is the total length, area or volume included in the characteristic species (dominant species, functional study. In the case of a line-transect, the number of groups, . . .) of the assemblage included, are the mea- samples correspond to the extent divided by the interval. sured species evenness and heterogeneity representative Adjacent sampling units are generally more similar in of the assemblage, would comparable results be obtained their fauna or flora than distant ones (Palmer 1995). at another time of the year? The answers depend on the In the case of social insects, if the sampling interval is spatio-temporal distribution of individuals. Ants are too short one can expect to collect individuals from the social insects and most of them have sessile colonies so same colony in contiguous samples. This should reduce that one may expect a fair consistence between con- the rate of species accumulation, in other terms the secutive measures of species richness. In tropical forests efficacy of the inventory and the shape of the species the heterogeneity in species distribution may be high, for abundance distribution. On the other hand, increasing example Kaspari (2000) observed from 1 to 17 ant the sampling interval requires more field work and species nesting in patches of 1 m2 of leaf litter. might also be unpractical when patches of habitat are Behavioral traits of ants, such as competition or species small. associations, have also an effect on species spatial To circumvent the difficulty of comparing structured distribution and result in a patchy distribution of inventories with various spatial designs, an effort to- colonies, particularly marked in the canopy (arboreal wards standardization of sampling protocols has been ant mosaics) and to some extent at the ground level realized for a variety of taxa such as mammals (Wilson et (Levings and Traniello 1981, Levings and Franks 1982, al. 1996), amphibians (Heyer et al. 1993), termites Majer 1993, Delabie et al. 2000a). Microenvironmental (Eggleton et al. 1995, Jones and Eggleton 2000) and factors, such as nest site availability, are also known to ants (Agosti and Alonso 2000). affect species distribution (Byrne 1994, Kaspari 2000). Among invertebrates, ants have numerous attributes Finally, despite the fact that most ants have sessile that make them useful for biological evaluation and colonies, their abundance and foraging activity depends monitoring (Majer 1983, New 1995). They are ecologi- on temperature and humidity (Levings 1983, Bestel- cally important in most terrestrial ecosystems. In the meyer and Wiens 1996). Amazonian rainforest they constitute up to 15% of the The aims of this paper were to study the effects of total animal biomass (Fittkau and Klinge 1973). With spatial components (number of samples, sampling inter- 11 000 described species (Agosti 2003), it is not an val, sampling extent, sampling grain) on diversity hyperdiverse group and their taxonomy is fairly well measures and to evaluate the performance of the known (Bolton 1994, 1995) and accessible to non- standardized A.L.L. protocol at documenting local ant specialists (Oliver and Beattie 1996). They are good diversity. In particular, we addressed the following indicators of environmental changes (Majer 1983, An- questions: is the sampling interval of 10 m optimal, are dersen 1997, Andersen and Sparling 1997, Peck et al. 20 samples enough to obtain a fair estimate of commu- 1998, Vasconcelos 1999, Carvalho and Vasconcelos nity diversity and to collect all numerically dominant 1999, Whitford et al. 1999, Kaspari and Majer 2000, species, what is the reproducibility of the diversity Kalif et al. 2001, Vasconcelos et al. 2001) and are useful measures? To answer these questions the selected site for conservation planning (Alonso 2000, Alonso and was first subject to a preliminary transect, then sub- Agosti 2000). mitted after a 9-month interval to an 8-fold oversam-

254 ECOGRAPHY 27:2 (2004) pling campaign. We focused on the most efficient can also be used for fitting species occurrences (Hayek collection method proposed in the A.L.L. protocol, the and Buzas 1997). mini-Winkler extraction of leaf-litter ants (Fisher 1999, Species richness (S), Shannon’s index of diversity (H) Delabie et al. 2000b). and Buzas and Gibson’s evenness E were calculated for an increasing number of samples taken in natural order to keep the sampling interval to a constant value of 1.25

m. H was obtained from the equation: H/ / p ln(p ) Material and methods S i× i where pi is the proportion of the ith species (Shannon Study site and Weaver 1949). Evenness (i.e. equitability or dom- H inance), E, was calculated with the equation: E/e / The study site was located inside the Rio Pilcomayo S(0B/E5/1). H was decomposed into H/ln S/ln E by National Park in northern Argentina (25804?06ƒS, Hayek and Buzas (1997) and these two latter values were 58805?36ƒW). The habitat, called ‘‘monte fuerte’’ is a calculated as well because the pattern of H, ln S, ln E, subtropical mesoxerophile oligarchic forest (Pujalte et al. and ln E/ln S during the accumulation of individuals are 1995, PHYSIS habitat unit 48.2412 of Devillers and characteristic of the underlying distribution of species Devillers-Terschuren 1996). abundance. For a broken stick distribution ln E remains constant, for a log normal distribution ln E/ln S remains constant, and for a log series H remains constant. The Sampling protocol, environmental measures evaluation of these patterns has been named ‘‘SHE analysis’’ by Hayek and Buzas (1997). A preliminary transect, following the A.L.L. protocol The effects of sampling interval and sampling extent, (Agosti et al. 2000) was conducted on 8 October 1999. A for fixed sample sizes (5, 10, 20, 40, 80, 160 quadrats) 200 m long line transect was traced and at intervals of 10 and for a sampling grain of 1 m2 corresponding to the m the leaf litter present inside a 1 m2 quadrat was size of the elementary quadrats, on SHE values were collected, sifted and put in a bag. The sifted material was explored by subsampling the pool of 160 quadrats. For brought back to the field laboratory and its fauna was each sample size and sampling interval (ranging from extracted with a mini-Winkler apparatus (Fisher 1998) 1.25 to 40 m), all possible groups of quadrats were drawn for 24 h. Temperature variations during the sampling from the pool of 160 quadrats. The effect of sampling period were recorded with a datalogger. A calibration grain was investigated by pooling 2, 4 or 8 contiguous transect was conducted 9 months later, between 23 and 31 July 2000, 2 m aside the preliminary transect. The quadrats yielding elongated sampling units. SHE values calibration transect followed the same protocol as the were then calculated according to a similar procedure as preliminary transect except that samples were collected for the analysis of the effect of interval and extent. No at an 8-fold increased density: central points of succes- attempt were made to test statistical hypotheses when sive quadrats were at intervals of 1.25 m instead of the 10 comparing groups of different interval, extent or grain m used in the preliminary transect. Temperature during because of the non-independence of data (nearby quad- that period fluctuated between 3.68C (at night) and rats more similar than distant ones, data present in one grain contribute to the data of the next highest grain, 27.68C with an average of 14.19/4.18C and was colder than during the preliminary sampling of October 1999 etc. The predictive performance, in terms of estimation of (values of 18.0Á/24.48C and 22.29/1.58C for the same parameters). local species richness, of a single A.L.L. transect was tested by decomposing the calibration transect (160 quadrats at intervals of 1.25 m) into eight A.L.L. transects (20 quadrats at intervals of 10 m). Each Data analysis data set was then extrapolated by three different All ants were identified to species or alternatively to approaches (Chazdon et al. 1998): 1) parametric morphospecies. The data concerning both specimens and methods; 2) non-parametric methods; 3) curve-fitting stations was input into a database (SIDbase, Leponce extrapolation. We tried one or several estimators and Vander Linden 1999) which allowed to retrieve the among the more commonly used in each category since data and perform all the subsequent diversity analyses. at this stage of knowledge it is very difficult to guess Species occurrence in samples (absence/presence data) a priori which estimator will work best for a given data was used as a surrogate of species abundance because set. ants are social insects, which implies that a single sample For parametric methods the observed relative abun- may contain an extreme abundance of a rare species dance species distribution was first compared to several (Longino 2000). Occurrences provide reliable informa- theoretical data distributions (log-normal, log series, tion on species proportions and the statistical distribu- broken-stick) (Preston 1948, Magurran 1988, Miller tions that are used to fit species abundance observations and Wiegert 1989, Hayek and Buzas 1997, Krebs

ECOGRAPHY 27:2 (2004) 255 1999). Our data fitted well a log series distribution (Fig. species richness for an equivalent diversity with a lower

4). For the logarithmic series the total number of species number of occurrences I1 (I1 B/I2) (eq. 12.12 in Hayek is given by: S/ /a×/ln(1/x) (eq. 9.12 in Hayek and and Buzas 1997). Buzas 1997) where a is a constant and x a number (0B/ x5/1) can be calculated with x/I/(I/a) where I is the number of species occurrences (eq. 9.14 in Hayek and Buzas 1997). Log series parameter a was calculated with the computer program EstimateS (Colwell 1997). Results The relationship between species occurrences (I) and Species spatial distribution sample size (A) was calculated by fitting a linear function to the points with Statistica 5.5 (Anon. 2000). Sixty-six species corresponding to 720 occurrences and To estimate S for 160 samples, the a value obtained with 10 554 individuals were found in the 160 quadrats of the 20 samples was used and the number of species calibration transect. A single quadrat contained between occurrences was extrapolated with the linear function 0 and 13 species (median value/4.0, n/160) and the linking I with A. Jaccard index of faunal similarity between quadrats was Among non-parametric estimators of species richness, 0.189/0.16 (avg9/SD, n/12 654). Species present in five common incidence-based estimators were compared only one (‘‘uniques’’) or two (‘‘duplicates’’) quadrats (see Colwell 1997, EstimateS user guide for their represented 44% of the 66 species collected. Only 11 description): Jackknife 1 and 2 (Heltsche and Forrester species were found in at least 10% of the quadrats, and 1983, Palmer 1991), Chao 2 (Chao 1987), bootstrap will hereafter be referred as ‘‘frequent species’’. An (Smith and van Belle 1984), and ICE (Lee and Chao interval of 1.25 m revealed that for most of them, except 1994). These estimators were calculated with EstimateS 5 the arboreal Crematogaster sp.2 and the Pheidole sp.1 (Colwell 1997). (species #6 and #7 on Fig. 1), occurrences are non- Among curve-fitting extrapolation methods two non- randomly clumped in adjacent quadrats (Runs tests, p5/ asymptotic models were chosen because the species 0.05). By contrast, less frequent species that occurred in accumulation curve failed to reach a plateau (Fig. 2): at least 4 quadrats were generally randomly distributed z a) the Arrhenius species-area model: S/c.A where z (14 of the 21 Runs tests performed on these species and c are curve-fitting parameters (Arrhenius 1921, yielded a p!/0.05). The randomness of the distribution Preston 1962a, b); b) the Sobero´n and Llorente model of the 34 remaining species was not tested due to

(Sobero´n and Llorente 1993, Fisher 1999): S(t)/ln(1/ insufficient data. z.a.t)/z which assumes that the probability of adding a new species depends on the current size of the species list. The parameter t represents the sampling effort (i.e. sampling time, number of samples, number of indivi- S, H, E vs number of samples duals), other parameters (z, a) are curve-fitting parameters. Species accumulation curves for each of the 8 The mean number of species collected in the calibration data sets were smoothed (sample-based rarefaction sensu transect, with a sampling interval of 1.25 m, could be Gotelli and Colwell 2001) by 500 random ordering of approximated by a logarithmic function of the number samples using EstimateS 5 (Colwell 1997). Models were of quadrats and of species occurrences (Fig. 2). After fit to the curves by the quasi-Newton method provided 160 quadrats no plateau was reached. The number of in Statistica 5.5 (Anon. 2000). uniques reached 20 and was still increasing. The number The reproducibility of diversity measures obtained of duplicates tended to level off to 9 species beyond 110 with a single A.L.L. transect and the mini-Winkler quadrats. Values of S, E, H for an increasing number of method was evaluated by comparing the values obtained quadrats, taken in natural order to keep a fixed sampling with the 8 A.L.L. transects constituting the calibration interval of 1.25 m, are plotted in Fig. 3. Evenness values transect to those obtained with the preliminary transect. decreased as quadrats accumulated; ln(E) decreased by Species richness was compared among transects by the same amount as ln(S) so that H (H/ln S/ln E) adjusting the series of samples to a common number changed little beyond 22 samples (around 70 species of occurrences, a procedure called rarefaction (Sanders occurrences). This pattern resembles the typical 1968, Krebs 1999, Gotelli and Colwell 2001). Rarefac- pattern obtained for a log series distribution of abun- tion curves were calculated with the Coleman method of dance (see Fig. 14.1 in Hayek and Buzas 1997). The EstimateS 5 (Colwell 1997). In addition, the rarefied species abundance distribution of the calibration trans- species richness for an equivalent a diversity was ect indeed conformed well to a log series distribution 2 calculated with the formula: S1 /a2×ln(1/(I1/a2)) where (goodness of fit test, x /6.46, DF/6, ns) (Fig. 4). a2 represents the parameter of the log series for I2 The corresponding log series Fisher’s a parameter was occurrences and S2 species and where S1 is the expected 17.7.

256 ECOGRAPHY 27:2 (2004) Fig. 1. Spatial distribution of the 46 ant species present in at least two samples along the 200 m calibration transect. Each square represent the species presence in a 1 m2 quadrat. Quadrats were separated by 0.25 m. Each row represents a different species. Species were sorted by decreasing occurrence in samples.

Fig. 2. Rarefaction curve representing the average number of species expected for a given number of quadrats (diamonds, sample-based rarefaction sensu Gotelli and Colwell 2001) or species occurrence (occurrence-based rarefaction, superimposed Fig. 3. SHE analysis of the calibration transect. Values for to the previous curve) and for a given number of species present species richness S, Shannon H, and Hayek and Gibson’s in only one (‘‘uniques’’, squares) or two quadrats (‘‘duplicates’’, evenness E are calculated for an increasing number of quadrats triangles). The abscissa is scaled logarithmically to reveal more taken in natural order of accumulation along the transect. H/ clearly the logarithmic nature of curves. ln S/ln E.

ECOGRAPHY 27:2 (2004) 257 sampling grain to 2, 4 and 8 m2. To avoid the depressing effect caused by small intervals, sampling intervals considered were equal or at least 10 m. Mean values of SHE for 5, 10 and 20 sampling units are presented in Fig. 6. For a given sampling effort, doubling the grain or doubling the number of samples yielded similar results in term of species richness but caused a proportionally greater increase of heterogeneity (H). Evenness values decreased much more when the number of samples was doubled than when the grain was doubled.

Performance of the standardized A.L.L. protocol Fig. 4. Species occurrence distribution (hatched) and predicted 2 values for a log series distribution with a/17.7 (solid). Inside the area sampled of 200 m it appeared that, with 20 samples at intervals of 10 m, a substantial variation of S, H, E vs sampling interval and extent the diversity measures was observed among the 8 A.L.L. transects constituting the calibration transect (Fig. 5). As Fig. 5 shows, with a sampling grain of 1 m2, the The variation range was 27Á/32 (average/309/2) for the number of samples had a stronger influence than species richness S, 910Á/1865 (average/13209/369) for sampling interval and extent on values of S, H and E. the number of individuals N, 77Á/104 (average/909/9) This is particularly true for evenness E, which was for the number of occurrences I, 0.59Á/0.69 (average/ almost constant for a given sample size but ranged from 0.649/0.03) for the evenness E, 2.87Á/3.03 for Shannon’s 0.87, with 5 samples, to 0.37, with 160 samples. S and H H (average/2.949/0.07), and 13.3Á/17.9 for Fisher’s a initially increased with the sampling interval and extent (average/15.79/1.8). Therefore, on the average, a single but reached a plateau for intervals over 10 m and extents A.L.L. transect collected /45% (30/66) of the species 2 B over 100 m . Shannon’s H varied between 2.3 (5 samples really present in the habitat. The average faunal similar- at intervals of 1.25 m) to 3.2 (160 samples). ity between the 8 A.L.L. transects, measured with the Jaccard index, was 0.489/0.06.

S, H, E vs sampling grain Extrapolations of S from 20 samples Contiguous quadrats were pooled to investigate the The performance of various extrapolation methods in effect on diversity measures of a doubling of the estimating species richness from the 8 A.L.L. subtran-

Fig. 5. Effect of sampling interval on mean values of species richness (S), evenness (E) and heterogeneity (H). Vertical bars indicate the standard error on the mean. The effect is presented by classes of sample sizes because of the predominent effect of this latter factor on S, H, E values. The sampling extent corresponding to each mean can be easily calculated by multiplying the sampling interval by the number of samples (on the x-axis, an extent of 200 m2 corresponds to an interval of 40, 20, 10, 5, 2.5, 1.25 m for 5, 10, 20, 40, 80, 160 samples respectively). The sampling grain was kept constant to 1 m2.

258 ECOGRAPHY 27:2 (2004) Fig. 6. Effect of sampling grain on mean values of species richness (S), heterogeneity (H) and evenness (E). Means were calculated for sampling intervals equal to or over 10 m to reduce the inﬂuence of this factor on S, H, E. Vertical bars indicate the standard error on the mean. Curves A, B, C correspond to 5, 10 and 20 sampling units respectively. A sampling unit comprises 2, 4 or 8 contiguous 1 m2 sampling quadrats that were pooled to increase the sampling grain over 1 m2.

sects was compared in Fig. 7. Parametric and curve- yielded both the closest estimates although they tended fitting extrapolation methods allow to calculate an to underestimate the true value by 10% on average. The estimate for a given number of samples. Since the species species-area model of Arrhenius tended to largely over- richness for 160 samples was known (66 species), estimate the true value. All non-parametric estimators extrapolations were performed for 160 samples with for incidence data (i.e. based on species occurrence) the parametric log series model and the curve-fitting which were tested (Jackknife 1 and 2, Chao 2, Bootstrap, species-area and Sobero´n and Llorente models (Fig. ICE) tended to underestimate the total species richness 7A). The log series and the Sobero´n and Llorente models which was at least 66 species (Fig. 7B). These five non-

Fig. 7. Extrapolation of species richness from the 8 standardized A.L.L. transects (20 samples at intervals of 10 m) composing the calibration transect. Comparison of the performance of (A) parametric, curve-ﬁtting and (B) non- parametric extrapolation methods. Observed species richness for 160 samples, a minimum value for the true total species richness, was 66 species (dotted line). All estimations were based on occurrence data.

ECOGRAPHY 27:2 (2004) 259 2 parametric estimators still steadily increased with sample m (Mann-Whitney rank sum test U/700.5, n1/20, size so that the value obtained with 20 samples cannot be n2/160, pB/0.001). As a consequence, the accumula- considered as a stable estimate of total species richness. tion of species was faster during the preliminary than Even with 160 samples all non-parametric estimators during the calibration transect and the number of failed to reach a stable value of total species richness. uniques began to level off. Despite an 8-fold more intensive sampling effort, 8 species found in the pre- Species proportions liminary transect were not collected in the calibration transect. All these 8 species were infrequent (species

In seven out of the eight A.L.L. transects extracted from occurrence: 1Á/3/20 samples, species abundance: 1Á/8 the calibration transect, all locally frequent species were individuals) (Appendix 1). All but one of the 11 frequent collected (Fig. 8). In one case, the frequent species species of the calibration transect were also found Crematogaster sp.2 was missed. among the 20 frequent species of the preliminary transect. However Hypoponera sp.4, present in 28/60 Temporal variation of S, H, E samples and ranked 8th most frequent species in the calibration transect was only found in 1/20 samples of A preliminary A.L.L. transect was performed 9 months the preliminary transect. earlier 2 m aside the calibration transect. Diversity The rarefaction technique allows to adjust a series of values obtained with this preliminary transect were samples to a common number of individuals so that S/45 species, E/0.66, H/3.39, I/161 occurrences, species richness can be compared among samples. To N/2316 individuals, Fisher’s a/20.7. Its average evaluate the performance of the rarefaction method to faunal similarity with the 8 A.L.L. transects was buffer the measure of S against seasonal variations, 0.449/0.04 (Jaccard index). The density of individuals rarefaction curves were calculated for both the prelimin- 2 collected and the number of species occurrences m ary transect and the calibration transect (decomposed were significantly higher in the preliminary transect than into 8 A.L.L. subtransects so that both the number of in the calibration transect: median values of 68.5 vs 33.0 samples and the sampling interval were identical to those 2 individuals m (Mann-Whitney rank sum test U/984, of the preliminary transect) (Fig. 9). The expected n1/20, n2/160, pB/0.005) and of 7.5 vs 4.0 species species richness for the largest common species occurrence (77) was 32.5 for the preliminary transect and 27.69/1.7 for the other 8 transects. When the a parameter of the log series is used to predict the rarefied species richness of the preliminary

transect for 77 occurrences one obtains a value of S/

20.7 ln(1/(77/20.7))/32.1 species. The same method applied to the other 8 transects yielded 27.89/1.7 species for 77 occurrences.

Discussion Composition and spatial structure of the leaf litter ant community The leaf litter ant community was composed of a few numerically dominant ants and of numerous rare species. Eighty-three percent of the 66 species encountered in the calibration transect were present in B/10% of the total area sampled. Even more, 44% of the species collected were known from only one or two samples (i.e. B/1.25% of the surface sampled). Eleven species were frequent (present in over 10% of the samples of the calibration Fig. 8. Sample-based rarefaction curves of the 11 frequent species (i.e. present in at least 10% of the whole 160 samples) for transect). These dominant species were predominantly each of the 8 A.L.L. standardized transects (20 samples at Myrmicinae belonging to genera such as Pheidole and intervals of 10 m) which composed the calibration transect (160 Solenopsis as it is often the case with Winkler extracts samples at intervals of 1.25 m). Each curve represents the (Ward 2000). Preliminary data suggest that many of the expected number of frequent species collected for a given sampling effort. Both axes were scaled logarithmically to better frequent species probably nest in the soil. Individuals of distinguish between curves. all 11 frequent species but Crematogaster sp.2 Á/ an

260 ECOGRAPHY 27:2 (2004) ecologically meaningful anyway. Unfrequent species showed little clumping in their spatial distribution (Fig. 1) suggesting that their colony size is small. Frequent species appeared generally clumped but it is unclear if the aggregates are constituted by one or several colonies.

Diversity estimates vs sampling design Diversity measures vs number of samples The number of samples collected had a much stronger influence than sampling interval and extent on measures of community species richness. At the scale of the 200 m transect, species richness increased logarithmically with the area sampled and the number of uniques was still rising (Fig. 2). When an inventory is nearly completed uniques decline (Longino 2000, Longino et al. 2002). At a larger scale, the slope of the species accumulation curve may change and result in a sigmoidal curve (Longino et al. 2002). Even if there is a limited number of species in the community considered, the species accumulation curve may hardly reach a plateau in a speciose environment like most tropical and subtropical forests because of rare species (Longino et al. 2002). These rare species may correspond to species normally not living in the habitat considered or to species generally not collected Fig. 9. Temporal variations of species richness between the with the method considered. 2 preliminary transect (A.L.L. protocol: 20 quadrats of 1 m at Species evenness was particularly sensitive to the intervals of 10 m) and the calibration transect (8-fold over- sampled A.L.L. protocol) sampled 9 months later at the same sampling effort. Values of 0.87 were obtained with 5 station but in colder weather conditions. Comparison of species samples, far from the 0.37 corresponding to 160 samples richness between the two period for an identical sampling (Fig. 5). It is a trivial consequence of the fact that the protocol (A.L.L.) and a common number of species occurrences commonness or rarity of species can not be assessed (dotted line). accurately with a few samples. With only 5 samples, for example, any species collected has a frequency at least arboreal species Á/ were found among 100 soil samples equal to 20%. With 160 samples species frequencies may collected at the same locality (unpubl.). Queens were range from 0.625 to 100%. Comparison of equitability of found among these soil samples for Solenopsis sp.1, species distribution among assemblages may thus be Brachymyrmex physogaster, Paratrechina sp.2, Octos- misleading especially if the number of samples consid- truma rugifera and Hypoponera sp. prox. trigona. ered is different. Nesting in the soil might be advantageous to buffer against extreme temperature variations experienced in the Chaco region (Burgos 1970). Diversity measures vs sampling interval, extent and grain Because ants live in colonies varying largely in size, it For a fixed number of samples, the influence of sampling is not possible to conclude that two occurrences of the interval was slight on species richness and was even less same species in neighbour quadrats correspond to two conspicuous on species evenness (Fig. 5). distinct colonies. This would require additional informa- Species richness tended to be lower when the sampling tion (such as presence of reproductives in nests, tests of interval was below 10 m (Fig. 5). This could be antagonism, genetic or chemical analyses) that is time- interpreted as the result of spatial autocorrelation for consuming and incompatible with a rapid evaluation of distances below 10 m, especially for the most frequent community diversity. In this respect the analysis of species since unfrequent species generally showed little community diversity of colonial organisms such as ants clumping in their spatial distribution (Fig. 1). Over 10 m differ from the analysis of diversity of non colonial the sampling interval did not affect much S. This result is organisms where what is taken into account is the consistent with the findings of Fisher (1999) who number of specimens of each species encountered during observed that in Madagascar forests, rates of species the inventory. In the context of the current diversity accumulation were not improved beyond an interval of 5 inventory ‘‘species frequency’’ should be understood as m. Species evenness was almost insensitive to the ‘‘species numerical dominance’’, a measure that is sampling interval probably because the number of

ECOGRAPHY 27:2 (2004) 261 samples is probably the major factor that affect evenness more than one colony is present) and some information as already discussed. The effect of sampling interval on on species spatial aggregation is lost. This bias can be H (/ln S/ln E) is almost exclusively explained by the reduced by using a sampling interval that is large and a response of S since E was nearly constant for a fixed sampling grain that is small in comparison to colony number of samples. In other words, H was lower for size. By contrast to individual-based rarefaction curves, sampling intervals below 10 m. which generally lie under sample-based rarefaction Sampling extent followed the same general pattern as curves because of the spatial aggregation of individuals sampling interval (Fig. 5): lower species richness and (Gotelli and Colwell 2001), the occurrence-based rar- heterogeneity for short sampling extent, evenness inde- efaction curve is generally very slightly above the pendent of extent for a given sample size. This is easily sample-based rarefaction curve. On Fig. 2 the sample- understood since extent is closely related to both interval based rarefaction curve was superimposed on the 2 and number of samples. Beyond an extent of 100 m , occurrence-based rarefaction curve. doubling the extent did not allow to collect more species In our study, occurrence-based rarefaction only par- (considering a given number of samples). This result has tially compensated for the variations of species density to be considered for a local scale. At regional or (32.5 instead of 27.6 species obtained for 77 cumulated geographic scale a high species turnover between sam- occurrences) and yielded a result very close to the one ples should occur. obtained by using Fisher’s a to rarefy (32.1 species). The An increase of sampling grain had a similar effect than rarefaction curve for the preliminary transect was above, a corresponding increase of the number of samples rather than between, individual rarefaction curves of the collected on species richness. Evenness decreased pro- 8 A.L.L. transects performed 9 months later. A possible portionaly less for an increase of grain than for a explanation is that the sub-community sampled with corresponding increase of the number of samples. As a Winkler extracts is larger during warm weather condi- result, an increase of grain caused a proportionally tions. On the one hand, the probability to collect species greater increase of heterogeneity (H) than a correspond- from other strata than the leaf litter (i.e. species nesting ing increase of the number of samples (Fig. 6). The in the soil or in trees) should increase with the higher accrual of species for a given area sampled is faster for a number of foragers often associated with warmer small compared to a large grain due to spatial depen- temperatures (Levings 1983). On the other hand, an dence (Malsch 2000) and to the use of species occur- increased number of foragers should also increase the rence. Compared to sampling interval and extent, it was probability to collect species living in small colonies and the grain that had the stronger effect on S, a result present in low numbers in the sampling unit or species consistent with those of Palmer and White (1994). present in its surroundings. After a 9-month interval numerically dominant leaf litter ant species were not different. With activity traps (pitfalls), more marked seasonal differences among foraging ants may be ob- Reproducibility of diversity measures served in the Argentinian Chaco (Bestelmeyer and Wiens The density of species m2 varied two-fold at an interval 1996). 2 of 9 months (median/4.0 vs 7.5 species m ). During the cold season significantly fewer species were collected with an A.L.L. transect than during warmer weather (30 vs 45 species). Individual-based rarefaction (based on Representativeness of a single A.L.L. transect species abundance data) is commonly used to compen- sate for variations of species density and to compare the A single A.L.L. transect with Winkler samples collected species richness among communities of similar taxo- on average B/45% of the species present in the leaf litter nomic composition and coming from similar habitats ant community. In a cocoa plantation in Brasil, a similar (the smaller sample is supposed to be a random sample proportion of the leaf litter ant fauna present in 0.87 ha of the larger set) (Sanders 1968, Simberloff 1972, Krebs was captured on average with 20 Winkler samples (50/ 1999). The rarefaction method is based on the frequency 106/47%, Delabie et al. 2000b). of each species (see Hayek and Buzas 1997 for details). All frequent species were collected with a single A.L.L. The rarefaction of data sets corresponding to colonial transect. Nevertheless the estimation of species propor- organisms such as ants inventoried without clear identi- tion remained approximative because of the limited fication of colonies has some peculiarities for two main number of samples. With 20 samples, unfrequent species reasons. First, as already discussed above, species that occur by chance in 2 samples earn a frequency of numerical dominance rather than species frequency 10% whereas a higher sampling effort would reveal that (calculated with the number of colonies present) is they are in fact rare. Conversely, common species may measured. Second, no more than one single species appear uncommon even though they are generally occurrence can be counted in a sampling unit (even if collected in at least 1 of the 20 samples.

262 ECOGRAPHY 27:2 (2004) Species richness for an increased sampling effort could a Costa Rican rainforest, Longino et al. (2002) obtained be inferred with little error (average underestimation of a log normal rather than a log series distribution of 10%) by either a parametric log series or a curve-fitting species occurrence. Collection methods taken individu- extrapolation model (Sobero´n and Llorente 1993). Both ally yielded distributions close to a log series. It should approaches outperformed non-parametric methods (Fig. also be noted that vascular plants may exhibit species 7). With 20 samples and even with 160 samples, non- spatial distributions very similar to those observed in the parametric estimators were far from reaching a stable leaf litter ant community studied here (see Fig. 2 in estimate of total species richness. Other studies also Palmer 1995). obtained poor results with non-parametric methods. The analysis of 500 Winkler samples taken from 0.87 ha of cocoa plantation revealed that Jackknife 1 and ICE estimators did not reach an asymptote before 300 Conclusions samples (Delabie et al. 2000b). The strong dependence The standardization of sampling protocols is an im- of non-parametric estimators to sample size has also portant step to allow quantitative comparisons between been observed in other ant inventories (Fisher 1996, communities in space and time. A single standardized 1998, 1999, Longino et al. 2002). Non-parametric A.L.L. transect with Winkler samples appears as the estimators can be considered as the minimum richness minimum sampling effort necessary for characterizing in the habitat (Longino et al. 2002). the leaf litter ant assemblage studied. Indeed, with 20 The oversampling of the A.L.L. protocol demon- samples: 1) all frequent species were included; 2) the strated that Shannon’s index H was very similar among Shannon’s index of diversity became little dependent of replicates whereas species richness, species occurrence sample size; 3) the species accumulation curve entered in and abundance varied substantially from one replicate to a stable logarithmic phase and species richness for an 8- another. Beyond 22 samples (ca 70 species occurrences), fold increased sampling effort could be inferred with a ln E decreased by the same amount as ln S so that H precision of ca 10%. As stressed by Cao et al. (2002), (H/ln S/ln E) changed only from 2.9 to 3.2 (10%) equal-sized samples may however differentially represent when the number of samples considered changed from the communities from which they are drawn. The 22 to 160. The stability of H is a characteristic of the log autosimilarity between replicated A.L.L. transects series distribution (Hayek and Buzas 1997). H was drawn from the community sampled was near 50% affected by temporal variations of species density and (Jaccard Index), a value that guarantees some degree of a value of 3.4 was obtained for the preliminary A.L.L. representativeness and should allow to measure between- transect during warmer weather conditions. This higher site complementarity (Cao et al. 2002). With a single value was the reflect of differences in species richness (45 transect B/45% of the local ant fauna was collected and vs 309/2 species) rather than of species evenness (0.66 vs the relative frequency of species was not always repre- 0.649/0.03). sentative. One or two additional transects allowed to The heterogeneity of species spatial distribution was collect respectively /60% and /72% of the local ant 2 B B high (0Á/13 species m ) and the average faunal fauna (Fig. 2) and are probably preferable to a single similarity between quadrats was low (Jaccard index/ transect in most situations, especially in the case of 0.18). This explains why diversity results obtained with assemblages more diverse than the one studied. With replicated A.L.L. transects inside the same sampling eight transects, the species accumulation curve was not extent were variable. As already clearly demonstrated by asymptotic yet, indicating that a higher sampling effort of Palmer and White (1994), no single species accumula- is required to estimate the total species richness of the tion curve exists for a habitat, but instead a collection of assemblage (Leponce et al. 2003b). The density of curves can be drawn and their extrapolation may lead to species m2 varied twofold after a 9-month interval. quite variable results. As a result, despite the use of identical sampling How these results could be generalized to other protocols, measures of species richness were on average communities is still speculative because only a few 50% (45 vs 30 species) higher during warmer weather communities have been inventoried intensively. A log conditions and could only be partially corrected by series distribution of species occurrence was also ob- rarefaction. Our results emphasize the need to compare served in a leaf litter ant community from a Brazilian diversity among communities for a similar number of cocoa plantation (Delabie et al. 2000b, Leponce et al. species occurrences and whenever possible to conduct 2003b) and from the Amazonian forest (Leponce et al. inventories in similar weather conditions and at a period 2003a). In the former study the parametric log series and where most species are active in order to maximize the the Sobero´n and Llorente models also yielded the best number of species collected per sampling effort. In the estimates of species richness for 500 samples with the case of a log series distribution, the widespread Fisher’s data from 25 samples. By combining several collection log series a (Fisher et al. 1943) and Shannon’s index were methods in order to inventory the complete ant fauna of the most appropriate diversity indexes. The former was

ECOGRAPHY 27:2 (2004) 263 useful to rarefy or abundify species richness and the Carvalho, K. S. and Vasconcelos, H. L. 1999. Forest fragmenta- latter was robust against sample size effects. Finally, tion in central Amazonia and its effects on litter-dwelling ants. Á/ Biol. Conserv. 91: 151Á/157. both parametric and Soberoń and Llorente extrapola- Chazdon, R. L. et al. 1998. Statistical methods for estimating tion methods yielded a fair estimate of total species species richness of woody regeneration in primary and richness along the transect, a minimum value of species secondary rain forests of NE Costa Rica. Á/ In: Dallmeier, richness for the assemblage sampled. F. and Comiskey, J. A. (eds), Forest biodiversity research, monitoring and modeling: conceptual background and Old World case studies. Parthenon Publishing, Paris, pp. 285Á/ Acknowledgements Á/ We thank the Administracioń de Parques 309. Nacionales, Buenos Aires, Argentina, for allowing us to Chao, A. 1987. Estimating the population size for capture- conduct research in P.N. R´ıo Pilcomayo. Ne´stor Sucunza, the recapture data with unequal catchability. Á/ Biometrics 43: guardaparques and Cornelio Paredes greatly facilitated our 783Á/791. work in the park. Thanks to G. J. Torales and E. R. Laffont, Colwell, R. K. 1997. EstimateS: Statistical estimation of species Univ. Nacional del Nordeste, for logistic support. This work richness and shared species from samples. Ver. 5. Á/ User’s was supported by the National Fund for Scientific Research guide and application published at: B/http://viceroy.eeb. (FNRS, Belgium) to YR (‘‘senior research associate’’ position uconn.edu/estimates!/. and grant F.R.F.C. no. 2.4519.00) and to LT (PhD grant) who Delabie, J. H. C., Agosti, D. and do Nascimento, I. C. 2000a. also received a grant from the ‘‘Fonds Leópold III pour Litter ant communities of the Brazilian Atlantic rain forest l’Exploration et la Conservation de la Nature’’ to conduct the region. Á/ In: Agosti, D. et al. (eds), Sampling ground- collection of samples in Argentina. dwelling ants: case studies from the worlds’ rain forests. Curtin Univ., Australia, School of Environmental Biology, Bull. no. 18, pp. 1Á/17. Delabie, J. H. C. et al. 2000b. Sampling effort and choice of methods. Á/ In: Agosti, D. et al. (eds), Ants. Standard References methods for measuring and monitoring biodiversity. Smith- sonian Institution Press, pp. 145Á/154. Agosti, D. 2003. Antbase. Á/ B/http://www.antbase.org!/. Devillers, P. and Devillers-Terschuren, J. 1996. A classification Agosti, D. and Alonso, L. E. 2000. The ALL protocol. A of South American habitats. Á/ IRSNB, Brussels, and standard protocol for the collection of ground dwelling- Institute of Terrestrial Ecology, Huntingdon. ants. Á/ In: Agosti, D. et al. (eds), Ants. Standard methods Eggleton, P. et al. 1995. The species richness of termites for measuring and monitoring biodiversity. Smithsonian (Isoptera) under differing levels of forest disturbance in Institution Press, pp. 204Á/206. the Mbalmayo Forest Reserve, southern Cameroon. Á/ J. Agosti, D. et al. 2000. Ants. Standard methods for measuring Trop. Ecol. 11: 85Á/98. and monitoring biodiversity. Á/ Smithsonian Institution Fisher, B. L. 1996. Ant diversity patterns along an elevational Press. gradient in the Re´serve Naturelle Inte´grale d’Andringitra, Alonso, L. E. 2000. Ants as indicators of diversity. Á/ In: Agosti, Madagascar. Á/ Fieldiana Zool. (n.s.) 85: 93Á/108. D. et al. (eds), Ants. Standard methods for measuring and Fisher, B. L. 1998. Ant diversity patterns along an elevational monitoring biodiversity. Smithsonian Institution Press, pp. gradient in the Re´serve Spećiale d’Anjanaharibe-Sud and on 80Á/88. the Western Masoala Peninsula, Madagascar. Á/ Fieldiana Alonso, L. E. and Agosti, D. 2000. Biodiversity studies, Zool. (n.s.) 90: 39Á/67. monitoring, and ants: an overview. Á/ In: Agosti, D. et al. Fisher, B. L. 1999. Improving inventory efficiency: a case study (eds), Ants. Standard methods for measuring and monitor- of leaf litter ant diversity in Madagascar. Á/ Ecol. Appl. 9: ing biodiversity. Smithsonian Institution Press, pp. 1Á/8. Á/ Andersen, A. 1997. Using ants as bioindicators: multiscale 714 731. Fisher, R. A., Corbet, A. S. and Williams, C. B. 1943. The issues in ant community ecology. Á/ Conserv. Ecol. B/www. relation between the number of species and the number of consecol.org/Journal/vol1/iss1/art8/!/. Andersen, A. N. and Sparling, G. P. 1997. Ants as indicators of individuals in a random sample of an animal population. restoration success: relationship with soil microbial biomass Á/ J. Anim. Ecol. 12: 42Á/58. Fittkau, E. J. and Klinge, H. 1973. On biomass and trophic in the Australian seasonal tropics. Á/ Rest. Ecol. 5: 109Á/114. Anon. 2000. STATISTICA for Windows. Computer program structure of the central Amazonian rain forest ecosystem. Á/ Biotropica 5: 2Á/14. manual. Á/ StatSoft, Tulsa, OK. Gotelli, N. J. and Colwell, R. K. 2001. Quantifying biodiversity: Arrhenius, O. 1921. Species and area. Á/ J. Ecol. 9: 95Á/99. Bestelmeyer, B. T. and Wiens, J. A. 1996. The effects of land use procedures and pitfalls in the measurement and comparison on the structure of ground-foraging ant communities in the of species richness. Á/ Ecol. Lett. 4: 379Á/391. Argentine Chaco. Á/ Ecol. Appl. 6: 1225Á/1240. Hayek, L. A. and Buzas, M. A. 1997. Surveying natural Bestelmeyer, B. T. et al. 2000. Field techniques for the study of populations. Á/ Columbia Univ. Press. ground-dwelling ants. Á/ In: Agosti, D. et al. (eds), Ants. Heltsche, J. F. and Forrester, N. E. 1983. Estimating species Standard methods for measuring and monitoring biodiver- richness using the jackknife procedure. Á/ Biometrics 39: 1Á/ sity. Smithsonian Institution Press, pp. 122Á/144. 11. Bolton, B. 1994. Identification guide to the ant genera of the Heyer, W. R. et al. (eds) 1993. Measuring and monitoring world. Á/ Harvard Univ. Press. biological diversity: standard methods for amphibians. Bolton, B. 1995. A new general catalogue of the ants of the Á/ Smithsonian Institution Press. world. Á/ Harvard Univ. Press. Jones, D. T. and Eggleton, P. 2000. Sampling termite assem- Burgos, J. J. 1970. El clima de la region noreste de la Republica blages in tropical forests: testing a rapid biodiversity Argentina en relacion con la vegetation natural y el suelo. assessment protocol. Á/ J. Appl. Ecol. 37: 191Á/203. Á/ Boletin de la Sociedad Argentina de Botanica 11: 37Á/ Kalif, K. A. B. et al. 2001. The effect of logging on the ground- 102. foraging ant community in eastern Amazonia. Á/ Stud. Byrne, M. M. 1994. Ecology of twig-dwelling ants in a wet Neotropical Fauna Environ. 36: 215Á/219. lowland tropical forest. Á/ Biotropica 26: 61Á/72. Kaspari, M. 2000. A primer on ant ecology. Á/ In: Agosti, D. et Cao, Y. D., Williams, D. and Larsen, D. P. 2002. Comparison of al. (eds), Ants. Standard methods for measuring and ecological communities: the problem of sample representa- monitoring biodiversity. Smithsonian Institution Press, pp. tiveness. Á/ Ecol. Monogr. 71: 42Á/56. 9Á/24.

264 ECOGRAPHY 27:2 (2004) Kaspari, M. and Majer, J. D. 2000. Using ants to monitor LaSalle, J. and Gauld, I. (eds), Hymenoptera and biodiver- environmental changes. Á/ In: Agosti, D. et al. (eds), Ants. sity. CAB International, pp. 115Á/141. Standard methods for measuring and monitoring biodiver- Miller, R. J. and Wiegert, R. G. 1989. Documenting complete- sity. Smithsonian Institution Press, pp. 89Á/98. ness, species-area relations, and species-abundance distribu- Krebs, C. J. 1999. Ecological methodology. Á/ Addison Wesley tion of a regional flora. Á/ Ecology 70: 16Á/22. Longman. New, T. R. 1995. Introduction to invertebrate conservation Lawton, J. H. et al. 1998. Biodiversity inventories, indicator taxa biology. Á/ Oxford Univ. Press. and effects of habitat modification in tropical forest. Oliver, I. and Beattie, A. J. 1996. Designing a cost-effective Á/ Nature 391: 72Á/76. invertebrate survey: a test of methods for rapid assessment Lee, S.-M. and Chao, A. 1994. Estimating population size via of biodiversity. Á/ Ecol. Appl. 6: 594Á/607. sample coverage for closed capture-recapture models. Palmer, M. W. 1991. Estimating species richness: the second- Á/ Biometrics 50: 88Á/97. order jackknife reconsidered. Á/ Ecology 72: 1512Á/1513. Legendre, P. and Legendre, L. 1998. Numerical ecology, 2nd ed. Palmer, M. W. 1995. How should one count species? Á/ Nat. Á/ Development in environmental modelling, Elsevier, p. 20. Areas J. 15: 124Á/135. Leponce, M. and Vander Linden, C. 1999. SIDbase: a database Palmer, M. W. and White, P. S. 1994. Scale dependance and the built for the management of social insects diversity inven- species-area relationship. Á/ Am. Nat. 144: 717Á/740. tories. Á/ In: Colloque de la Section Française de l’IUSSI, Peck, S. L., McQuaid, B. and Campbell, C. L. 1998. Using ant Tours, France, 1Á/3 Sept. 1999, p. 38. species (Hymenoptera: Formicidae) as a biological indicator Leponce, M., Delabie, J. H. C. and Vasconcelos, H. L. 2003a. of agroecosystem condition. Á/ Environ. Entomol. 27: 1102Á/ Diversity and structure of ant communities in central 1110. Amazonia. Á/ In: Biotic Interactions in The Tropics: A Preston, F. W. 1948. The commoness and rarity of species. Special Symposium of the British Ecological Society and Á/ Ecology 29: 254Á/283. The Annual Meeting of the Association for Tropical Biology Preston, F. W. 1962a. The canonical distribution of common- and Conservation. Univ. of Aberdeen, 7Á/10 July 2003, p. 10. ness and rarity: part I. Á/ Ecology 43: 185Á/215. Leponce, M., Missa, O. and Delabie, J. H. C. 2003b. Perfor- Preston, F. W. 1962b. The canonical distribution of common- mance of estimators of species richness in a leaf-litter ant ness and rarity: part II. Á/ Ecology 43: 410Á/432. assemblage. Á/ In: Colloque de la Section Française de Pujalte, J. C. et al. 1995. Anales de parques nacionales. l’IUSSI, Bruxelles, Belgique, 1Á/3 Sept. 2003, p. 21. Unidades Ecolo´gicas del parque nacional Rio Pilcomayo. Levings, S. C. 1983. Seasonal, annual, and among-site variation Á/ Administracioń de Parques Nacionales, Tomo XVI, in the ground ant community of a deciduous tropical forest: Buenos Aires, Argentina. Á/ some causes of patchy species distributions. Ecol. Monogr. Sanders, H. 1968. Marine benthic diversity: a comparative 53: 435Á/455. study. Á/ Am. Nat. 102: 243Á/282. Levings, S. C. and Traniello, J. F. A. 1981. Territoriality, nest Shannon, C. E. and Weaver, W. 1949. The mathematical theory dispersion, and community structure in ants. Á/ Psyche 88: of communication. Á/ Univ. of Illinois Press. 265Á/319. Simberloff, D. 1972. Properties of the rarefaction diversity Levings, S. C. and Franks, N. R. 1982. Patterns of nest measurement. Á/ Am. Nat. 106: 414Á/418. dispersion in a tropical ground ant community. Á/ Ecology Smith, E. P. and van Belle, G. 1984. Nonparametric estimation 63: 338Á/344. of species richness. Á/ Biometrics 40: 119Á/129. Longino, J. T. 2000. What to do with the data? Á/ In: Agosti, D. et al. (eds), Ants. Standard methods for measuring and Soberoń, J. and Llorente, J. 1993. The use of species accumula- monitoring biodiversity. Smithsonian Institution Press, pp. tion functions for the prediction of species richness. Á/ Conserv. Biol. 7: 480Á/488. 186Á/203. Longino, J. T. and Colwell, R. K. 1997. Biodiversity assessment Vasconcelos, H. I., Carvalho, K. S. and Delabie, J. H. C. 2001. using structured inventory: capturing the ant fauna of a Landscape modifications and ant communities. Á/ In: Bierregaard, R. O. Jr et al. (eds), Lessons from Amazonia. tropical rainforest. Á/ Ecol. Appl. 7: 1263Á/1277. Longino, J. T., Coddington, J. and Colwell, R. K. 2002. The ant Yale Univ. Press, pp. 199Á/207. fauna of a tropical rain forest: estimating species richness Vasconcelos, H. L. 1999. Effects of forest disturbance on the three different ways. Á/ Ecology 83: 689Á/702. structure of ground-foraging ant communities in central Magurran, A. 1988. Ecological diversity and its measurement. Amazonia. Á/ Biodiv. Conserv. 8: 409Á/420. Á/ Chapman and Hall. Ward, P. S. 2000. Broad-scale patterns of diversity in leaf-litter Malsch, A. 2000. Investigation of the diversity of leaf-litter ant communities. Á/ In: Agosti, D. et al. (eds), Ants. inhabiting ants in Pasoh, Malaysia. Á/ In: Agosti, D. et al. Standard methods for measuring and monitoring biodiver- (eds), Sampling ground-dwelling ants: case studies from the sity. Smithsonian Institution Press, pp. 99Á/121. worlds’ rain forests. Á/ Curtin Univ., Australia, School of Whitford, W. G. et al. 1999. Ants as indicators of exposure to Environmental Biology, Bull. no. 18, pp. 31Á/40. environmental stressors in North American desert grass- Majer, J. D. 1983. Ants: bio-indicators of minesite rehabilita- lands. Á/ Environ. Mon. Assess. 54: 143Á/171. tion, land-use, and land conservation. Á/ Environ. Manage. Wiens, J. A. 1989. Spatial scaling in ecology. Á/ Funct. Ecol. 3: 7: 375Á/383. 385Á/397. Majer, J. D. 1993. Comparison of the arboreal ant mosaic in Wilson, D. et al. 1996. Measuring and monitoring biological Ghana, Brazil, Papua New Guinea and Australia Á/ its diversity. Standard methods for mammals. Á/ Smithsonian structure and influence on arthropod diversity. Á/ In: Institution Press.

ECOGRAPHY 27:2 (2004) 265 Appendix 1. Species found in the preliminary and in the calibration transect. Frequency of occurrence in the 20 and 160 samples respectively.

Subfamily Á/ Species Preliminary transect % Calibration transect %

Formicinae Brachymyrmex physogaster 45.0 56.9 Camponotus (Myrmosphincta ) sp.11 0.0 1.9 Camponotus (Myrmaphaenus ) sp.13 5.0 0.0 Camponotus (Pseudocolobopsis ) sp.17 0.0 0.6 Camponotus (Myrmothrix ) renggeri 0.0 3.1 Camponotus arborens 0.0 1.3 Camponotus crassus 15.0 9.4 Myrmelachista sp.2 0.0 0.6 Paratrechina pubens 20.0 2.5 Paratrechina sp.2 50.0 30.0

Myrmicinae Acromyrmex hispidus fallax 10.0 1.3 Apterostigma sp. complex pilosum 0.0 1.9 Cephalotes minutus 5.0 3.8 Crematogaster corticicola 5.0 3.1 Crematogaster euterpe 5.0 0.0 Crematogaster montezumia 5.0 1.3 Crematogaster sp.11 0.0 0.6 Crematogaster sp.14 0.0 1.3 Crematogaster sp.2 45.0 17.5 Crematogaster sp.5 15.0 0.0 Crematogaster sp.7 0.0 0.6 Cyphomyrmex rimosus 25.0 5.6 Megalomyrmex drifti 0.0 0.6 Myrmicocrypta foreli 5.0 0.0 Octostruma rugifera 60.0 23.8 Oxyepoecus sp.1 0.0 0.6 Pheidole aberrans 0.0 6.9 Pheidole radoszkowskii reflexans 20.0 10.6 Pheidole sp.30 0.0 0.6 Pheidole sp.1 45.0 17.5 Pheidole sp.21 5.0 0.0 Pheidole sp.22 15.0 7.5 Pheidole sp.4 5.0 5.0 Pyramica crassicornis 5.0 1.3 Pyramica denticulata 15.0 13.1 Pyramica gr. appretiatus sp.1 0.0 1.3 Pyramica gr. appretiatus sp.2 10.0 0.0 Pyramica sp.2 5.0 4.4 Rogeria scobinata 20.0 6.3 Solenopsis sp.1 95.0 76.9 Solenopsis sp.13 0.0 0.6 Solenopsis sp.15 0.0 1.3 Solenopsis sp.2 10.0 6.9 Solenopsis sp.5 5.0 0.0 Solenopsis sp.7 0.0 0.6 Solenopsis sp.8 25.0 1.3 Strumigenys ogloblini 15.0 0.6 Strumigenys sp. prox. elongata 0.0 0.6 Wasmannia sp.1 50.0 41.3 Wasmannia sp.3 0.0 2.5

Ponerinae Amblyopone sp.1 0.0 0.6 Anochetus diegensis 0.0 2.5 Discothyrea neotropica 5.0 5.0 Ectatomma edentatum 10.0 8.1 Ectatomma permagnum 10.0 0.0 Gnamptogenys striatula 25.0 2.5 Heteroponera sp.1 0.0 1.9

266 ECOGRAPHY 27:2 (2004) Appendix 1. (Continued).

Subfamily Á/ Species Preliminary transect % Calibration transect %

Hypoponera clavatula 5.0 0.6 Hypoponera opaciceps 0.0 2.5 Hypoponera opacior 15.0 2.5 Hypoponera sp.4 5.0 17.5 Hypoponera sp.5 0.0 0.6 Hypoponera sp. prox. opaciceps 5.0 1.9 Hypoponera sp. prox. trigona 30.0 16.9 Hypoponera sp.1 0.0 0.6 Leptogenys consanguinea 5.0 1.3 Odontomachus bauri 5.0 2.5 Odontomachus meinerti 0.0 0.6 Pachycondyla ferruginea 5.0 0.6 Pachycondyla harpax 10.0 2.5 Pachycondyla gr. villosa sp.1 0.0 0.6 Prionopelta punctulata 10.0 0.6 Typhlomyrmex pusillus 0.0 0.6

Pseudomyrmecinae Pseudomyrmex gracilis 0.0 1.9

ECOGRAPHY 27:2 (2004) 267 Chapitre 2: Spatial structure of litter-dwelling ant distribution in a subtropical dry forest

SPATIAL STRUCTURE OF LITTER-DWELLING ANT DISTRIBUTION IN A

SUBTROPICAL DRY FOREST

LAURENCE THEUNIS 1, 2, MARIUS GILBERT2, YVES ROISIN 2, AND MAURICE LEPONCE 1.

LAURENCE THEUNIS

Phone +32 2 627.43.64 Fax +32 2 649.48.25 E-mail: [email protected]

49 Insect. Soc. 52 (2005) 366–377 0020-1812/05/040366-12 Insectes Sociaux DOI 10.1007/s00040-005-0822-0 © Birkhäuser Verlag, Basel, 2005

Research article

Spatial structure of litter-dwelling ant distribution in a subtropical dry forest

L. Theunis1, 2, M. Gilbert3, Y. Roisin2 and M. Leponce1

1 Section of Conservation Biology, Royal Belgian Institute of Natural Sciences, Rue Vautier 29, 1000 Brussels, Belgium, e-mail: [email protected] 2 Behavioral and Evolutionary Ecology, CP 160/12, Université Libre de Bruxelles, Avenue F.D. Roosevelt 50, 1050 Brussels, Belgium 3 Biological Control and Spatial Ecology, CP 160/12, Université Libre de Bruxelles, Avenue F.D. Roosevelt 50, 1050 Brussels, Belgium

Received 14 March 2005; revised 26 April 2005; accepted 10 May 2005. Published Online First

Abstract. Understanding the spatial patterns of species Introduction distribution is essential to characterize the structure of communities, to optimize species inventories and to evaluate Understanding the spatial patterns of species distribution the impact of biotic and abiotic variables. Here we describe is essential to characterize the structure of communities, to the spatial structure of the distribution of leaf litter ant spe- optimize species inventories (Leponce et al., 2004) and to cies, and of biotic factors that could explain it, in a subtropi- evaluate the impact of biotic and abiotic variables. Little is cal semi-deciduous forest of the Argentinian Chaco, char- known about the fi ne spatial scaling of the majority of spe- acterized by a dense understorey of shrubs and terrestrial cies assemblages including leaf litter ants. Tropical ant as- bromeliads. Environmental variables (leaf litter quantity and semblages show a high species richness and a patchy distri- ground bromeliad density) were measured and ants were col- bution of colonies (Wilson, 1958; Levings and Franks, 1982; lected in 1 m2 quadrats distributed along two 200 m transects Levings, 1983; Benson and Brandão, 1987; Kaspari, 1996a; at intervals of 1.25 m. Overall 87 species were collected. Six- Vasconcelos and Delabie, 2000) which depends on biotic teen positive associations and a single negative association and abiotic constraints. Leaf litter ants are not territorial and were observed between the 11 most frequent species taken a considerable amount of evidence suggests that favourable pair-wise. Our results suggest that the spatial distribution of resource availability, rather than competition, is a major leaf litter ants was determined at two different scales. At a force structuring tropical leaf litter ant assemblages (Franks, small scale (period below 10 m) a periodic spatial structure, 1982; Byrne, 1994; Kaspari, 1996a,b; Soares and Schoere- likely due to intraspecifi c competition, produced a succes- der, 2001) involving overlapping foraging areas (Jackson, sion of peaks of abundance separated by gaps. At a larger 1984; Byrne, 1994). For ground-dwelling ants, causes of scale (period around 50 m), periodically distributed envi- patchiness include predation by swarm-raiding army ants ronmental factors induced aggregates of colonies of species (Franks and Bossert, 1983; Kaspari, 1996b; Hirosawa et al., responding positively to these factors. A high quantity of leaf 2000), moisture content preferences (Levings, 1983; Levings litter and, to a lesser extent, a high density of ground brome- and Windsor, 1984; Kaspari, 1996a), temperature prefer- liads promoted a high density and a high species richness of ences (Bestelmeyer, 2000), topography (Vasconcelos et al., ants. Numerically dominant ants being generally positively 2003), nest-site and food availability (Herbers, 1989; Byrne, associated, interspecifi c competition was apparently weak. 1994; Kaspari, 1996b; Kaspari and Majer, 2000), leaf litter All ant species whose abundance was correlated with an en- quantity and quality (Vasconcelos, 1990; Höfer et al., 1996; vironmental factor were not completely spatially structured Kaspari, 1996a; Carvalho and Vasconcelos, 1999) and both by it. This suggests that some other factors, such as intraspe- vegetation structure and composition (Wilson, 1958; Gadag- cifi c competition, may have counter-effects. kar et al., 1993; Feener and Schupp, 1998; Moutinho, 1998; Retana and Cerdà, 2000; Bestelmeyer and Wiens, 2001). Keywords: Spatial pattern, ant distribution, geostatistics, In a previous study carried out at a high resolution and Chaco. based on a nearly exhaustive sampling of a strip of 200 m2 in a subtropical semi-deciduous forest of the Argentinean Chaco, we demonstrated the highly heterogeneous distribu- Insect. Soc. Vol. 52, 2005 Research article 367 tion of leaf litter ant species and evaluated its consequences Faunal similarity between transect A and B was estimated using on diversity estimates (Leponce et al., 2004). The present Jaccard’s index (Jaccard, 1912; Wilson and Schmida, 1984) calculated study aimed at extending this work by the spatial analysis as follows: c of the ant species distribution and of the biotic factors that S = could explain it. To achieve this objective, we measured j a + b – c conspicuous environmental variables likely to affect ant (where a = total number of species in sample A, b = total number of spe- distribution and measured the nature of interactions between cies in sample B, c = number of common species to samples A and B). numerically dominant ants. Species associations and correlations between environmental factors and

ant abundance were evaluated on the log10 (n + 1)-transformed abundance in order to limit the weight of samples collected around nests, trails and Methods exploited resources. Standard parametric tests of signifi cance could not be used here because of spatial autocorrelation (SA), which represents a Study site bias to the assumption of independence among samples (Lennon, 2000; Legendre et al., 2002). Using simulation data, Legendre et al. (2002) The study site was located in Río Pilcomayo National Park, northern showed that Dutilleul’s modifi ed t-test (Dutilleul, 1993) constitutes an Argentina, in the wet Chaco region (25°04’06’’ S, 58°05’36’’ W). The effi cient method to account for SA in estimating the signifi cance of the habitat, called “monte fuerte” is a subtropical mesoxerophile oligarchic correlation between two autocorrelated variables, and this method was forest (Pujalte et al., 1995; habitat unit PHYSIS 48.2412 of Devillers used here to test the signifi cance of all bivariate correlations. We re-ad- and Devillers-Terschuren, 1996) dominated by Schinopsis balansae justed the p-values for statistical acceptance with the Holm procedure Engl., Astronium balansae Engl. and Aspidosperma quebracho-blanco (1979) (Legendre and Legendre, 1998) because the probability of a type Schlecht. and by a ground strata of bromeliads (Aechmea distichantha I error becomes larger than the nominal value of a when several tests of Lemaire and Pseudananas sagenarius (Arruda) Camargo) (Pujalte et signifi cance are carried out simultaneously (i.e. in a correlation matrix). al., 1995). Spatial analysis: autocorrelation and periodicity Two methods were used to explore spatial patterns in environmental fac- Sampling design tors and ant species distributions. First, spatial correlograms were used to quantify the level of spatial dependence, i.e. the tendency of points Ant sampling protocol close together to have more similar values than points farther apart. Two 200 m-long transects (A and B) located 400 m apart were sampled Spatial correlograms plot the values of the spatial correlations between between July 23 and August 8, 2000 in a 16 ha forest fragment. Each observations separated by increasing distance classes, and allow describ- transect consisted of 160 quadrats of 1 m2 separated by 1.25 m intervals ing the extent (distance over which no SA is measured), and intensity (transect A is extensively described in Leponce et al., 2004). At each (when autocorrelation is strong, points separated by close distances have sampling point, the leaf litter found inside the 1 m2 quadrat was collect- strongly correlated values) of SA (Rossi et al., 1992; Liebhold et al., ed, sifted and put in a cotton bag. The sifted material was brought back 1993, Legendre et al., 2002; Liebhold and Gurevitch, 2002; Perry et al., to the fi eld laboratory and its fauna was extracted with a mini-Winkler 2002). Correlogram values range from –1 to +1 (Rho(h)) and can be apparatus (Fisher, 1998) for 24 hours. Temperature, recorded every 10 interpreted as indicating negative or positive correlations in the same minutes, ranged between 3.6 °C (at night) and 27.6 °C with an average way as simple correlation coeffi cients. Second, periodograms were of 14.1 ± 4.1 °C during the sampling session of transect A and between used to quantify the presence of periodic patterns in the transect data. 10.6 and 30.2 °C (18.5 ± 4.2 °C) during the sampling of transect B. Aver- Periodograms resulted from a Fourier-transformation decomposing the age temperatures were lower (14.1 °C) during the sampling of transect A observed transect data into a sum of periodic terms, and plotting the than during the sampling of transect B (18.5 °C) (t-test, p < 0.001). The intensity (as measured by the amplitude) as a function of the period of weather was dry during the 17 days sampling campaign (only three short each term (Shumway, 1988; Legendre and Legendre, 1998). We ranked and light rains occurred). the level of periodicity in our transect data according to three arbitrary classes of amplitude: strong periodicity (highest peak > 6), intermediate Environmental measures (highest peak is < 6 and > 1) and low (highest peak < 1). Correlograms In order to interpret the pattern of species distribution, we measured and periodograms were calculated using Statistica 6.0 software (StatSoft three conspicuous environmental variables at each 1 m2 quadrat: (1) the Inc, 2004). sifted litter weight (which integrates factors such as food, nest, tempera- Studies of spatial patterns along transects (representing a single ture and moisture availability) (Levings, 1983) (2) the density of ground dimension) allow obtaining fi ne SA coeffi cients and periodogram values bromeliads (omnipresent in the habitat and affecting ant species density (Legendre and Fortin, 1989). and composition (unpublished results)), (3) canopy openness (infl uenc- ing the temperature and dryness at ground level). The percentage of Structuring effects canopy openness was estimated from hemispherical photographs, shot We aimed at exploring whether spatial periodicity observed in species 1.5 m above ground level and quantifi ed with the Gap Light Analyzer 2.0 distribution could be attributed to one or several environmental covari- program (Frazer et al., 1999). ates. The periodograms and correlograms of the residuals from the linear regression between species abundance and a microhabitat factor were therefore estimated, and compared to those of the species abundance Data analysis data. A strong structuring effect of a microhabitat factor on the species distribution would result in a substantial reduction in the amplitude of All ants were determined to species or morphospecies level. In order to the highest peak in the periodogram of regression residuals once the assess the impact of environmental variations on ant density and species variability related to microhabitat factor has been removed. This reduc- composition, we pooled the data from the two transects. By contrast, tion was quantifi ed and used as an estimate of the spatial structuring the two transects were considered separately for the analysis of spatial effect of the environmental variable: decreases over 50 %, between 50 % structure. Numerically dominant ant species were defi ned as species and 20 %, and lower than 20 % were considered as strong, intermediate found in at least 10 % of the samples, and will be hereafter referred as or low structuring effects, respectively. A similar approach was used to “frequent species”. explore the relationship between the environmental factors. 368 L. Theunis et al. Spatial structure of ant distribution

Results Relationships between ant species composition and environmental factors Eighty-seven species corresponding to 1880 occurrences and 24114 individuals were found in the 320 quadrats of the two The abundance of eight frequent species was positively transects (species list in Appendix 1). Both transects had 11 correlated with leaf litter weight and that of two species frequent species in common which occupied similar ranks with bromeliad density (Table 2). Solenopsis sp.01 and of occurrence (Spearman rank order correlation coefﬁ cient, r = 0.691, p < 0.05). Sixteen positive and a single negative associations were observed between these 11 frequent species (species found in at least 10 % of the samples) (Table 1).

Relationships between ant density and environmental factors

Median leaf litter weight was 357 g (quartiles: 212–524), bromeliad density 2 plants/m2 (0–4) and canopy openness 18.4 % (16.9–19.9) (N = 320) along transects A and B. Be- cause values of canopy openness varied very little (variation of ± 5 %) (Fig. 1), we did not undertake further investiga- tions of its effects on ant distribution. Litter quantity varied considerably, up to 25 fold between contiguous quadrats. Species density (number of species/m2) was positively correlated (Pearson’s correlation) with leaf litter weight (N = 320, r = 0.71, p < 0.05) and with bromeliad density (N = 320, r = 0.31, p < 0.05). Leaf litter weight was also positively correlated with bromeliad density (N = 320, r = 0.27, p < 0.05). Quadrats devoid of bromeliads (N = 85 out of 320) had signiﬁ cantly less leaf litter (Mann-Whitney rank sum test U Fig. 1. Spatial distribution of canopy openness (CO), bromeliad density = 5091, p < 0.001), a lower ant species density (M-W rank (BD), leaf litter weight (LW) and distribution of abundance of the four most frequent species (Solenopsis sp.01, Brachymyrmex physogaster, sum test: U = 5150, p < 0.001) and a lower species richness Wasmannia sp. prox. auropunctata and Crematogaster sp.02) along (47 vs. 55 species for 314 occurrences) than quadrats with transect A. Black line corresponds to smoothed curves calculating mo- bromeliads (N = 235). bile mean of data.

Table 1. Square matrix with Pearson’s correlation coeffi cients between abundance of individuals (log10(ur1)-transformed) of the eleven species taken by pair (transects pooled, N = 320). Statistically signifi cant positive or negative associations between species are greyed or blackened respectively. Levels of signifi cance were adjusted fi rst using Dutilleul’s modifi ed t-test and then using Holm’s procedures. Infrequent species were discarded because too little data was available to draw conclusions.

Occurrence Species 1 2 3 4 5 6 7 8 9 10 11 Rank

1 Solenopsis sp.01 1 2 Brachymyrmex physogaster 0.49 1 3 Wasmannia sp. prox. auropunctata 0.37 0.32 1 4 Crematogaster sp.02 0.07 0.07 0.06 1 5 Octostruma rugifera 0.35 0.29 0.15 0.09 1 6 Hypoponera sp. prox. trigona 0.32 0.33 0.26 0.08 0.26 1 7 Paratrechina sp.02 0.26 0.08 0.16 0.02 0.16 0.06 1 8 Pyramica denticulata 0.21 0.16 0.16 0.25 0.29 0.37 0.15 1 9 Solenopsis sp. 17 –0.30 0.12 –0.01 0.03 0.10 0.16 0.04 0.10 1 10 Pheidole ﬂ avens 0.07 –0.01 0.07 0.17 0.07 –0.01 0.02 0.01 0.04 1 11 Pheidole nubila 0.07 0.15 –0.01 0.08 0.24 0.31 0.02 0.22 0.24 0 1 Insect. Soc. Vol. 52, 2005 Research article 369

Table 2. Effect of environmental factors on the distribution of frequent ant species (LW = litter weight and BD = bromeliad density). Pearson’s correlation coeffi cients between the abundance of frequent species (log10(ur1)-transformed) and environmental factors (raw data) are indicated for pooled quadrats from the 2 transects (N = 320). Levels of signifi cance were adjusted using Dutilleul’s modifi ed t-test and Holm’s procedures. Species were sorted by decreasing rank of occurrence in the two pooled transects. Periodicity in species spatial distribution were measured on periodograms and were categorized as either none (highest peak < 1), intermediate (highest peak 1–6) or strong (highest peak > 6). Between brackets, the periodicity of highest peak was noted. The structuring effect of leaf litter weight and bromeliad density on species spatial distribution corresponded to the percentage of decrease of the 50 m (LW) and 66.7 m and 100 m (BD) periodical peak and was categorized as either none (decrease < 20 %), intermediate (decrease 20–50 %) or strong (decrease > 50 %) and were analysed for each transect separately (N = 160).

Frequent species Environmental factors Transect A Transect B

LW BD Periodicity Structuring effect Periodicity Structuring effect

LW (50 m) BD (66.7 m) LW (50 m) BD (100 m)

Solenopsis sp.01 0.49 *** 0.31 *** strong strong strong strong strong intermedi- (50 m) (66 m) ate Brachymyrmex 0.51 *** 0.16 strong strong intermedi- strong strong none physogaster (50 m) ate (50 m) Wasmannia sp. prox. 0.33 *** 0.06 strong intermedi- none strong strong none auropunctata (50 m) ate (18 m) Crematogaster sp.02 0.17 0.06 none intermedi- strong none ate (5.25 m) Octostruma rugifera 0.47 *** 0.11 intermedi- none none intermedi- strong none ate (12.5 m) ate (50 m) Hypoponera sp. prox. 0.47 *** 0.09 none none none strong strong none trigona (100 m) Paratrechina sp.02 0.27 *** 0.28 *** intermedi- none none intermedi- none none ate (13.3 m) ate (10 m) Pyramica denticulata 0.33 ** –0.02 none intermedi- strong none ate (50 m) Solenopsis sp. 17 0.23 0.07 intermedi- none none intermedi- none none ate (22 m) ate (66 m) Pheidole ﬂ avens 0.10 0.07 none none Pheidole nubila 0.34 *** 0.00 none intermedi- none none ate (33 m)

Paratrechina sp.02 were positively correlated with both leaf (Fig. 2). Leaf litter quantity correlogram indicated evidence litter weight and bromeliad density. Crematogaster sp.02, of a periodic spatial distribution along both transects (Fig. Solenopsis sp.17 and Pheidole fl avens did not show any sig- 2A, B). Positive autocorrelations (peaks) were observed at nifi cative correlation with either litter weight or bromeliad distances below 20 m, between 45 and 65 m and over 90 m. density. At other lag distances, samples were negatively autocorrelated (troughs). The distance between successive peaks (period) was thus T = 50 m as indicated by the highest peak Spatial pattern of environmental factors in corresponding periodograms (Fig. 2C, D). In transect B, and ant distribution a second large peak was observed at T = 100 m. Brome- liad density periodograms showed a different periodicity in The spatial distribution of the environmental factors and of transects A and B (Fig. 2G, H). In transect A, we observed the 4 most frequent species (present in at least 1/3 of sam- four large peaks corresponding to periods of 66.6 m, 22.2 m, ples) along transect A is presented in Fig. 1. All variables ex- 16.6 m and 11.8 m. In transect B, we observed a single peak cept canopy openness varied signifi cantly along the transect, corresponding to a period of 100 m. The shape of bromeliad with a succession of peaks and gaps. Similar results were density correlogram of transect B corresponded to a gradient obtained for transect B, except around a depressed zone of spatial structure, i.e. autocorrelation values decreased with 15 m long that was temporarily fl ooded and devoid of both increasing intervals. bromeliads and leaf-litter. Periodic spatial structures were observed in the distribu- Leaf litter weight and bromeliad density showed a strong tion of 10 out of 11 frequent ant species (all but Pheidole fl a- spatial structure in their distribution along transects A and B vens, Fig. 3) (Table 2). A strong (example of B. physogaster; 370 L. Theunis et al. Spatial structure of ant distribution

Fig. 2. Spatial analysis (correlograms and periodograms) of litter weight (above: A, B, C, D) and bromeliad density (below: E, F, G, H) for transects A (A, C, E, G) and B (B, D, F, H). Highest peaks in periodograms indicate a periodicity of environmental variables. Litter weight was distributed with a 50 m period in each transect. Bromeliad density showed different periodicity in his spatial distribution between transects (see text for more details). Rho (h) is the coefﬁ cient of autocorrelation varying between –1 and +1. Insect. Soc. Vol. 52, 2005 Research article 371

Fig. 3. Periodicity categories of spatial distribution of frequent ant species. Example of correlograms (above A, B, C) and periodograms (below F, G, H) of species showing a strong (A, D), an intermediate (B, E) and a lack of periodicity (C, F) in their spatial distribution along the transect B. The degree of periodicity was estimated according to the amplitude of the highest principal peak of the periodogram and was categorized as either strong (highest peak > 6), intermediate (highest peak > 1) or none (highest peak < 1). Rho (h) is the coefﬁ cient of autocorrelation varying between –1 and +1.

Fig. 3A, D) and an intermediate periodicity (example of no effect on the highest peak of periodicity at T = 50 m (for Solenopsis sp. 17; Fig. 3B, E) were observed in the spatial both transects). distribution of four and six species respectively. Solenopsis In a second step, we evaluated the structuring effect sp.01, Brachymyrmex physogaster, Wasmannia sp. prox. au- of litter weight and bromeliad density on frequent species ropunctata, Octostruma rugifera and Pyramica denticulata distribution with correlograms and periodograms of residu- showed the same periodicity (example of B. physogaster on als (e.g. Fig. 4A, B) obtained from the regression between Fig. 3D) as litter weight (Fig. 2C, D) with the highest peak species abundance and the environmental factor consid- at a period of 50 m. All frequent species but Crematogaster ered. Correlograms allowed a visualisation of the decrease sp.02 and Paratrechina sp. 02 showed a positive autocor- of periodicity and periodograms allowed us to quantify it. relation for distance lags below 10 m (Fig. 3, example for 3 The percentage of decrease of periodogram values were species). measured relative to periods where environmental variables showed the highest peak of periodicity, i.e. for litter weight at 50 m (transect A and B, Fig. 2C, D) and for bromeliad Environmental variation and spatial structure density at 66.7 m (transect A, Fig. 2G) and 100 m (transect of ant species distribution B, Fig. 2H). The litter weight was a structuring factor for seven fre- First, we veriﬁ ed whether the periodicity of leaf litter weight quent ant species as well as the bromeliad density for two of distribution could be related to bromeliad density and vice them (Table 2). versa, since the two environmental factors were correlated. Correlograms and periodograms of standardised residuals from the regression between these two factors showed the Litter quantity and bromeliad density as strong structuring same highest peak(s) as the initial ones (as in Fig. 2C, D, G, factors of ant spatial distribution H) although weak variations in periodogram values could be The structuring effect of environmental variable on each observed. Indeed, we observed no effect of litter weight on species spatial distribution was explored by inspecting the the periodicity of bromeliad density. In contrast, bromeliad periodograms of standardised residuals between the abun- density inﬂ uenced periodogram values of litter weight at dance of a species and the value of the variable. A strong T = 66.7 m (transect A) and at T = 100 m (transect B) but had structuring effect was evident when a peak of abundance of 372 L. Theunis et al. Spatial structure of ant distribution

Fig. 4. Measure of structuring effect intensity of environmental factors (leaf litter quantity and bromeliad density) on frequent ant species distribution.

Correlogram (A) and periodogram (B) of residuals from the linear regression between Brachymyrmex physogaster abundance (log10 (n + 1)-transformed) and leaf litter weight in transect B. Periodic spatial structure of species distribution disappeared after removing (by regression) leaf litter effects. Rho (h) is the coefﬁ cient of autocorrelation varying between –1 and +1.

a species experienced a decrease in amplitude over 50 %. For onstrating a strong or intermediate structuring effect of leaf example, the peak at a period of 50 m in the periodogram litter weight or bromeliad density on species distribution of Brachymyrmex physogaster abundance decreased from (Fig. 4). 7.28 (Fig. 3 D) to 1.89 in the periodogram of residuals (74 % decrease) (Fig. 4B) indicating that leaf litter quantity had a strong structuring effect on Brachymyrmex physogaster Species not structured by litter quantity or bromeliad distribution. density The comparison between the correlogram of a species Two frequent species (Pa. sp. 02 and Ph. nubila) were not and of the residuals allowed assessing the structuring effect structured by the leaf litter weight as determined by a residu- of an environmental variable (e.g. Fig. 3A and Fig. 4A). al analysis although they were correlated to this factor. In the The structuring effect of litter weight on the distribu- same way, Pa. sp.02 abundance was correlated to, although tion of the two most frequent ant species was strong in both not spatially structured by, bromeliad density. transects (Table 2). In contrast, a strong structuring effect of litter weight was only observed in transect B for Wasman- nia sp. prox. auropunctata, Crematogaster sp. 02, Octostru- Discussion ma rugifera, Hypoponera sp. prox. trigona and Pyramica denticulata. The structuring effect of bromeliad density Effects of environmental factors vs. interspecifi c interaction on the distribution of Solenopsis sp. 01 was only strong in on ant species density and composition transect A. Our results suggest that most of the frequent ant species coexist in leaf litter. Indeed, numerous species foraged in the Litter quantity and/or bromeliad density as intermediate same quadrat (up to 16 species m–2) and 16 positive vs. a sin- structuring factors of ant spatial pattern gle negative associations between frequent species suggested The analysis of correlograms and periodograms of residuals low interspecifi c competition in our assemblage where for- showed that litter weight had an intermediate spatial structur- aging ranges may overlap considerably. These results are in ing effect (20–50 % decrease of peaks) on the distribution of agreement with those of previous works (Levings, 1983; Wasmannia sp. prox. auropunctata in transect A. Bromeliad Levings and Windsor, 1984; Byrne, 1994; Kaspari, 1996a, density had an intermediate spatial structuring effect on the b). Moreover, the only negative association was found be- spatial distribution of Solenopsis sp.01 (transect B) and Bra- tween two Solenopsis species, which probably occupied chymyrmex physogaster (transect A), although the latter was very close ecological niches. Weak interspecifi c competi- not signifi cantly correlated to bromeliad density (Table 2). tion could be explained by suffi ciency of nesting sites and In addition, a positive autocorrelation remains at short food (Herbers, 1989; Kaspari, 1996b; Soares and Shoereder, distance (below 10 m) in the correlograms of residuals dem- 2001) or by avoidance behaviours between heterospecifi c in- Insect. Soc. Vol. 52, 2005 Research article 373 dividuals allowing a high overlap in food utilisation (Byrne, obtained elsewhere in the tropics (Kaspari, 1996b). The 1994). On the ground, as opposed to the canopy, numerically present study suggests a periodic distribution of the leaf lit- dominant ants (mostly generalist in our study) do not form a ter. A possible explanation for this phenomenon would be mosaic of non-overlapping territories. related to topographic differences (microrelief). In another The distribution of frequent species of our assemblage Chacoan Schinopsis balansae forest, Barberis et al. (1998) was principally associated to leaf litter quantity, rather than have demonstrated that most woody species and bromeliads competition. Several studies have highlighted the dominant grow preferentially on well-drained convex zones of the soil. infl uence of such environmental factors on tropical litter ant The clumped distribution of trees would induce an accumula- assemblages (Franks, 1982; Byrne, 1994; Kaspari, 1996a, tion of leaf litter on the slightly higher zones whereas the leaf b; Soares and Shoereder, 2001). Leaf litter provides nesting litter would tend to be carried away by temporary inundations sites (Vasconcelos, 1990; Didham, 1998), favorable moisture in the depressed zones of the forest. Bromeliads, preferring content (Levings, 1983; Vasconcelos, 1990; Bestelmeyer, convex zones, tend to increase the quantity of litter possibly 1997), and food resources (Andersen, 1983) for ants and because they affect the litter composition, adding their own other arthropods (Bestelmeyer and Schooley, 1999a). We dead material, and accumulation, due to their root network observed, as in other studies, a positive correlation between (Benzing, 1980). This might explain why we observed that the litter quantity and ant density (Vasconcelos, 1990; some peaks of periodicity of litter quantity could be attributed Kaspari, 1996b) and composition (Kaspari, 1996b; Carvalho to bromeliad density. Unfortunately, the periodicity of convex and Vasconcelos, 1999). However several studies did not fi nd zones remains to be demonstrated. Nevertheless it seems a an effect of the leaf litter quantity on ant species density and reasonable hypothesis since periodic patterns of vegetation species abundance (Soares and Shoereder, 2001; Delabie are sometimes observed (e.g. tiger bush in semi-arid African and Fowler, 1995). Litter quantity was found to be positively landscapes, Couteron and Lejeune, 2001). related to litter structural complexity, because of vertical lay- The bromeliad density was also spatially structured ering (Vasconcelos, 1990). Litter samples displayed variable but differently so in each transect. We observed a periodic vertical stratifi cation, some being mainly composed of intact structure in transect A with a period (T = 66.6 m) close to leaves, others of leaves at more advanced stages of decom- that of litter quantity. A gradient was observed in transect B position. Vertical litter stratifi cation may allow an increase (T = 100 m). Gradient structure (Legendre and Fortin, 1989; in the number of coexisting species of ground-dwelling Legendre and Legendre, 1998) was probably an artefact arthropods through habitat partitioning (Anderson, 1978; (false gradient) caused by the presence of a gap, deprived Vasconcelos, 1990) and by limiting competition (Yanoviak of bromeliads, inside transect B. This trend was also weakly and Kaspari, 2000). expressed in the leaf litter correlogram (Fig. 2B). Bromeliads Seventy percent (8 out of 11) of frequent species were showed strong SA below 5 m in both transects. This could positively correlated with litter weight. These species could be a consequence of the asexual reproduction by rhizomes occupy sub-layer(s) of litter composed of decayed leaves (Benzing, 1980). and might be specialized to exploit a thick cover of leaf litter. Among species not correlated with litter quantity, we found Crematogaster sp.02 which is arboreal, Pheidole Structuring effect of environmental variables on the spatial fl avens which has the ability to use different microhabitats distribution of ants as nesting sites with some preference for pieces of wood (Wilson, 2003) and Solenopsis sp.17 whose biology is un- Among the eight species whose abundance was correlated known. to leaf litter weight, six were strongly spatially structured Bromeliad density was also related to species density and (period around 50 m) by this environmental factor in at least abundance of several ant species but on the whole the impact one transect. Structuring effects were generally more appar- of bromeliads on the ant assemblage was more limited than ent in transect B because the ant activity was increased by the effect of leaf litter quantity (Table 2). Bromeliad leaves more favourable temperature conditions. Solenopsis sp. 01 form a rosette accumulating rain and litter, and contribute was correlated to and structured by bromeliad density in both to favorable moisture and temperature conditions for most transects. arthropods (Benzing 1980). Moreover, their spiny leaves The correlation between species abundance and a factor provide protection against predators such as opossums, giant is not necessarily spatial, and may be observed at the quadrat anteaters, tamanduas or armadillos (Pujalte et al., 1995; Ei- scale without implying a structuring spatial effect of the fac- senberg and Redford, 1999). In the same habitat, soil termite tor at a larger scale. Conversely, the presence of a structuring diversity is also positively correlated to bromeliad density effect does not necessarily imply a strong local correlation: (Roisin and Leponce, 2004). species abundance and a factor can fl uctuate together at large scale (when the whole transect is considered), but still show a loose association when observed for each quadrat. Two Spatial pattern of environmental variables examples illustrate this observation. First, Crematogaster sp.02 (in transect B) was not correlated to litter weight and We observed variation in litter quantity between contiguous was found to be distributed with a period of 50 m. Its highest quadrats up to 25 fold, which is consistent with the results peaks of abundance occurred in zones of high litter quantity 374 L. Theunis et al. Spatial structure of ant distribution so that a structuring effect of this factor was detected. Sec- Conclusions ond, Ph. nubila was locally correlated to leaf litter quantity, but not structured by this factor along the whole transect: this Our results suggest that in the subtropical forest studied, species was concentrated mostly at the end of the transects the spatial distribution of leaf litter ants is determined at and thus could not be spatially structured by the leaf litter two different scales. At a small scale (period below 10 m) a quantity with a 50 m period. periodic spatial structure is likely to be related to intraspe- After removing the structuring effect of the environmen- cifi c competition since we observed, for the most frequent tal factors (with residual analysis), some peaks of periodic- species, a succession of peaks of abundance separated by ity (at periods different from those that corresponded to our gaps reducing aggression between allocolonial individuals. environmental factor effects) persisted indicating that other At a larger scale (period around 50 m), environmental fac- factors structured the species distribution. These factors tors, also periodically distributed, may induce aggregates of could be predation by army ants (Franks and Bossert, 1983; colonies of species responding positively to these factors. Kaspari, 1996b), other biotic factors (e.g. competition, prey A high quantity of leaf litter and, to a lesser extent, a high availability), abiotic factors (e.g. soil characteristics, nest- density of bromeliads promoted a high density and a high site availability), or stochastic events. species richness of ants. Interspecifi c competition, even be- Nine out of the 11 frequent species showed a strong spa- tween numerically dominant ants, was weak. All ant species tial structure in their distribution below 10 metres (as shown correlated to an environmental factor were not obligatorily in Figs. 3 and 4). In other words, species displayed a clumped spatially structured by it, suggesting that some other factors, distribution. The correlogram of residuals (Fig. 4A), indi- such as intraspecifi c competition, dipersal and/or environ- cated that leaf litter quantity was not the cause of this pattern, mental factors not measured may have counter-effects. even for species strongly structured by litter quantity. It is likely that this pattern would be related to the size of the foraging area of individual colonies (Brühl et al., 2003; Delabie Acknowledgments et al., 2000b; Kaspari, 1993, 1996b) or to nest aggregation in suitable zones (Herbers, 1989; Soares and Shoereder, 2001). We thank the Administración de Parques Nacionales, Buenos Aires, Ar- Peaks of species abundance represented in Fig. 1 may indi- gentina, for allowing us to collect in P.N. Río Pilcomayo. Nestor Sucun- cate the location of nests and gaps between them could refl ect za, the guardaparques and Cornelio Paredes greatly facilitated our work in the park. Thanks to G.J. Torales and E.R. Laffont, Univ. Nacional intraspecifi c competition. This would be in agreement with del Nordeste, for logistic support. This work was supported by fellow- several studies showing that intraspecifi c interactions affect ships from the ‘Fonds National de la Recherche Scientifi que’ (FNRS, nest spacing (Levings and Franks, 1982; Ryti and Case, 1984, Belgium) to MG and to LT (PhD Grant). A grant to LT from the ‘Fonds 1986, 1988, 1992). Dispersal or other environmental factors Léopold III pour l’Exploration et la Conservation de la Nature’ allowed a may also be partly responsible of this pattern. fi eld campaign in Argentina. We would like to thank also J.H.C. Delabie and I.C. do Nascimiento (CEPEC, Brasil) for help in ant identifi cation, Species showing no spatial structure could be either I. Bachy (RBINS) for help in image treatment, A. Franklin (RBINS) for randomly distributed (Leponce et al., 2004) or could be useful advice in geostatistics, Prof. J.M. Pasteels for critical reading and submitted to several structuring factors with opposing forces. improvement of the manuscript.

Appendix 1. List of species found in transect A and B. Numbers represent their occurrences in the 160 samples collected in each transect.

Subfamily Species Transect A Transect B

DOLICHODERINAE Linepithema group humile sp.2 0 9 ECITONINAE Eciton vagans 0 3 Labidus coecus 0 3 FORMICINAE Brachymyrmex physogaster 89 95 Brachymyrmex sp.05 2 20 Camponotus (Myrmothrix) renggeri 5 0 Camponotus arboreus 2 0 Camponotus crassus 15 23 Camponotus sp. 19 0 1 Camponotus sp.11 (Myrmosphincta) 3 1 Camponotus sp.13 (?Myrmaphaenus) 0 1 Camponotus sp.14 0 2 Camponotus sp.17 (Pseudocolobopsis) 1 0 Myrmelachista sp.02 1 7 Insect. Soc. Vol. 52, 2005 Research article 375

Subfamily Species Transect A Transect B

FORMICINAE Paratrechina pubens 4 5 Paratrechina sp.02 48 47 MYRMICINAE Acromyrmex hispidus fallax 2 13 Apterostigma sp.complex pilosum 3 10 Carebarella bicolor 3 3 Cephalotes minutus 6 13 Crematogaster corticicola 5 2 Crematogaster euterpe 0 6 Crematogaster montezumia 2 0 Crematogaster sp.02 28 78 Crematogaster sp.07 1 1 Crematogaster sp.11 1 6 Crematogaster sp.14 2 0 Crematogaster sp.16 0 2 Cyphomyrmex rimosus 10 13 Leptothorax sp.01 0 8 Leptothorax sp.02 0 2 Megalomyrmex drifti 1 5 Mycocepurus goeldii 0 2 Myrmicocrypta foreli 0 2 Octostruma rugifera 39 66 Oxyepoecus reticulatus 1 1 Pheidole aberrans 11 2 Pheidole nubila 17 37 Pheidole sp.01 23 34 Pheidole sp.04 9 0 Pheidole sp.21 0 2 Pheidole sp.22 12 47 Pheidole sp.30 7 40 Pyramica crassicornis 2 1 Pyramica denticulata 21 81 Pyramica gr. appretiata sp.01 2 0 Pyramica gr. appretiata sp.02 0 2 Pyramica sp.02 8 4 Rogeria scobinata 10 24 Solenopsis clytemnestra bruchi 0 1 Solenopsis sp. 17 20 46 Solenopsis sp. 18 5 6 Solenopsis sp.01 101 98 Solenopsis sp.02 11 15 Solenopsis sp.10 0 6 Solenopsis sp.13 5 1 Solenopsis sp.15 3 5 Strumigenys louisianae 0 2 Strumigenys ogloblini 1 0 Strumigenys sp. prox. elongata 1 1 4 Trachymyrmex sp.01 0 7 Wasmannia sp. prox. auropunctata 66 102 Wasmannia sp.03 4 4 PONERINAE Amblyopone sp.01 1 0 Anochetus diegensis 4 2 Discothyrea neotropica 8 0 Ectatomma edentatum 13 19 Ectatomma permagnum 0 1 Gnamptogenys striatula 4 6 Heteroponera sp.01 3 0 Hypoponera clavatula 1 0 376 L. Theunis et al. Spatial structure of ant distribution

Subfamily Species Transect A Transect B

PONERINAE Hypoponera opaciceps 4 12 Hypoponera opacior 6 18 Hypoponera sp. 09 0 2 Hypoponera sp. prox. opaciceps 1 1 5 Hypoponera sp. prox. trigona 29 66 Hypoponera sp.04 29 3 Hypoponera sp.05 1 1 Hypoponera sp.07 0 1 Leptogenys consanguinea 2 1 Odontomachus chelifer 4 5 Odontomachus meinerti 1 0 Pachycondyla ferruginea 1 1 Pachycondyla harpax 4 9 Pachycondyla obscuricornis 0 5 Pachycondyla villosa 1 0 Prionopelta punctulata 1 2 Typhlomyrmex pusillus 1 1 PSEUDOMYRMECINAE Pseudomyrmex gracilis 3 0

References Couteron P. and Lejeune O. 2001. Periodic spotted patterns in semi-arid vegetation explained by a propagation-inhibition model. J. Ecol. 89: 616–628 Adams E.S. 1994. Territory defence by the ant Azteca trigona – Delabie J.H.C. and Fowler H.G. 1995. Soil and litter cryptic assemblag- maintenance of an arboreal ant mosaic. Oecologia 97: 202– es of Bahian cocoa plantations. Pedobiologia 39: 423–433 208 Delabie J.H.C., Agosti D. and do Nascimento I.C. 2000a. Litter ant com- Andersen A.N. 1983. Species diversity and temporal distribution of ants munities of the Brazilian Atlantic rain forest region. In: Sampling in the semi-arid mallee region of northwestern Victoria. Aust. J. Ground-dwelling Ants: Case Studies from the Worlds’ Rain Forests Ecol. 8: 127–137 (Agosti, D., J. Majer, L.E. Alonso and T. Schultz, Eds), Curtin Anderson J.M. 1978. Inter- and intra-habitat relationships between University, Australia, School of Environmental Biology, Bulletin woodland cryptostigmata species diversity and the diversity of soil no 18, pp1–17 and litter microhabitats. Oecologia 32: 341–348 Delabie J.H.C., Campiolo S. and Fresneau D. 2000b. Etude comparative Barberis I.M., Pire E.F. and Lewis J.P. 1998. Spatial heterogeneity de la saturation des communautés de fourmis des litières sous lati- and woody species distribution in a Schinopsis balansae (Anacar- tudes tropicales et tempérées. Actes Coll. Insectes Soc. 13: 71–75 diaceae) forest of the Southern Chaco, Argentina. Rev. Biol. Trop. Devillers P. and Devillers-Terschuren J. 1996. A Classiﬁ cation of South 46: 515–524 American Habitats. IRSNB, Brussels, and Institute of Terrestrial Benson W.W. and Brandão C.R.F. 1987. Pheidole diversity in the humid Ecology, Huntingdon, 419 pp Tropics: a survey from Serra dos Carejas, Para, Brazil. In: Chem- Didham R.K. 1998. Altered leaf-litter decomposition rates in tropical istry and Biology of Social Insects (J. Eder and H. Rembold, Eds), forest fragments. Oecologia 116: 397–406 München, Verlag J. Peperny, pp 593–594 Dutilleul P. 1993. Modifying the t test for assessing the correlation be- Benzing D.H. 1980. The Biology of the Bromeliaceae. Mad River Press tween two spatial processes. Biometrics 49: 305–314 Inc., Eureka, California, 305 pp Eisenberg J.F. and Redford K.H. 1999. Mammals of the Neotropics: the Bestelmeyer B.T. 1997. Stress tolerance in some Chacoan dolichoderine central Neotropics, Volume 3. The University of Chicago Press: ants, implications for community organisation and distribution. J. 609 pp Arid Envir. 35: 297–310 Feener D.H. and Schupp E.W. 1998. Effect of treefall gaps on the Bestelmeyer B.T. 2000. The trade-off between thermal tolerance and patchiness and species richness of Neotropical ant assemblages. behavioural dominance in a subtropical South American ant com- Oecologia 116: 191–201 munity. J. Anim. Ecol. 69: 998–1009 Fisher B.L. 1998. Ant diversity patterns along an elevational gradient Bestelmeyer B.T. and Schooley R.L. 1999. The ants of the southern So- in the Réserve Spéciale d’Anjanaharibe-Sud and on the Western noran desert: community structure and the role of trees. Biodivers. Masoala Peninsula, Madagascar. Fieldiana Zool. (n.s.) 90: 39– Conserv. 8: 643–657 67 Bestelmeyer B. and Wiens J.A. 2001. Local and regional-scale re- Franks N.R. 1982. Social insects in the aftermath of swarm raids of the sponses of ant diversity to a semi-arid biome transition. Ecography army ant Eciton burchelli. In: The Biology of Social Insects (Breed, 24: 381–392 M.D., C.D. Michener and H.E. Evans, Eds). Westview Press, Boul- Brühl C.A., Eltz T. and Linsenmair K.E. 2003. Size does matter – effects der, CO, pp 275–279 of tropical rainforest fragmentation on the leaf litter ant community Franks N.R. and Bossert W.H. 1983. The inﬂ uence of swarm raiding in Sabah, Malaysia. Biodivers. Conserv. 12: 1371–1389 army ants on the patchiness and diversity of tropical leaf litter ant Byrne M.M. 1994. Ecology of twig-dwelling ants in a wet lowland tropi- community. In: Tropical Rain Forest: Ecology and Management cal forest. Biotropica 26: 61–72 (Sutton, E.L., T.C. Whitmore and A.C. Chadwick, Eds), Blackwell, Carvalho K.S. and Vasconcelos H.L. 1999. Forest fragmentation in cen- Oxford, pp 151–163 tral Amazonia and its effects on litter-dwelling ants. Biol. Conserv. Frazer G.W., Canham C.D. and Lertzman K.P. 1999. Gap Light Ana- 91: 151–157 lyzer (GLA), Version 2.0: Imaging Software to Extract Canopy Insect. Soc. Vol. 52, 2005 Research article 377

Structure and Gap Light Transmission Indices from True-color Moutinho P.R.S. 1998. Impactos do uso da terra sobre a fauna de formi- Fisheye Photographs, Users Manual and Program Documentation. gas, consequências para recuperação fl orestal na Amazônia Oriental. Copyright 1999: Simon Fraser University/Institute of Ecosystem In: Floresta Amazônica: dinâmica, regeneração e manejo (Gascon, Studies, Burnaby, BC/ Millbrook/NY C. and P. Moutinho, Eds), Manaus, MCT-INPA, pp 155–170 Gadakgar R., Nair P., Chandrashekara K. and Bhat D.M. 1993. Ant Perry J.N., Liebhold A.M., Rosenberg M.S., Dungan J., Miriti M., Ja- species richness and diversity in some selected localities in western komulska A. and Citron-Pousty S. 2002. Illustrations and guidelines Ghats, India. Hexapoda 5: 79–94 for selecting statistical methods for quantifying spatial pattern in Herbers J.M. 1989. Community structure in north temperate ants: tem- ecological data. Ecography 25: 578–600 poral and spatial variation. Oecologia 81: 201–211 Pujalte J.C., Reca A.R., Balabusic A., Canevari P., Cusato L. and Flem- Hirosawa H., Higashi S. and Mohamed M. 2000. Food habits of Aenic- ing V.P. 1995. Anales de parques nacionales. Unidades Ecológicas tus army ants and their effects on the ant community in a rain forest del Parque Nacional Río Pilcomayo. Administración de Parques Borneo. Insect. Soc. 47: 42–49 Nacionales XVI: 1–185 Höfer H., Martius C. and Beck L. 1996. Decomposition in an Ama- Retana J. and Cerdà X. 2000. Patterns of diversity and composition of zonian rain forest after experimental litter addition in small plots. Mediterranean ant communities tracking spatial and temporal vari- Pedobiologia 40: 570–576 ability in the thermal environment. Oecologia 123: 436–444 Holm S. 1979. A simple sequentially rejective multiple test procedure. Roisin Y. and Leponce M. 2004. Characterizing termite assemblages in Scand. J. Stat. 6: 65–70 fragmented forests: a test case in the Argentinian Chaco. Austral. Jaccard P. 1912. The distribution of fl ora in the alpine zone. New Phytol. Ecol. 29: 637–646 11: 37–50 Rossi R.E., Mulla D.J., Journel A.G. and Franz E.H. 1992. Geostatistical Jackson D.A. 1984. Ant distribution patterns in a Cameroonian cocoa tools for modelling and interpreting ecological spatial dependence. plantation: investigation of the ant mosaic hypothesis. Oecologia Ecol. Monogr. 62: 277–314 62: 318–324 Ryti R.T. and Case T.J. 1984. Spatial arrangements and diet overlap Kaspari M. 1993. Body size and microclimate use in Neotropical between colonies of desert ants. Oecologia 62: 401–404 granivorous ants. Oecologia 96: 500–507 Ryti R.T. and Case T.J. 1986. Overdispersion of ant colonies: a test of Kaspari M. 1996a. Litter ant patchiness at 1-m2 scale: disturbance dy- hypotheses. Oecologia 69: 446–453 namics in three Neotropical forests. Oecologia 107: 265–173 Ryti R.T. and Case T.J. 1988. The regeneration niche of desert ants: ef- Kaspari M. 1996b. Testing resource-based models of patchiness in four fects of established colonies. Oecologia 75: 303–306 Neotropical litter ant assemblages. Oikos 76: 443–454 Ryti R.T. and Case T.J. 1992. The role of neighbourhood competition Kaspari M. and Majer J.D. 2000. Using ants to monitor environmental in the spacing and diversity of ant communities. Am. Nat. 139: changes. In: Ants: Standard Methods for Measuring and Monitoring 355–374 Biodiversity (Agosti, D., J. Majer, L.A. Alonso and T. Schultz, Eds), Shumway R.H. 1988. Applied Statistical Time Series Analysis. – Engle- Smithsonian Institution Press. pp 89–98 wood Cliffs, NJ: Prentice Hall, pp 117–165 Legendre P. and Fortin M.-J. 1989. Spatial pattern and ecological analy- Soares S.M. and Schoereder J.H. 2001. Ant nest distribution in a rem- sis. Vegetatio 80: 107–138 nant of tropical rainforest in southeastern Brazil. Insect. Soc. 48: Legendre P. and Legendre L. 1998. Numerical Ecology. Second English 280–286 edition. Elsevier, Development in environmental modelling, 20: 853 StatSoft Inc. 2004. STATISTICA (data analysis software system), ver- pp sion 6. www.statsoft.com Legendre P., Dale M.R.T., Fortin M.-J., Gurevitch J., Hohn M. and Myers Vasconcelos H.L. 1990. Effects of litter collection by understory palms D. 2002. The consequences of spatial structure for the design and on the associated macroinvertebrate fauna in Central Amazonia. analysis of ecological fi eld surveys. Ecography 25: 601–615 Pedobiologia 34: 157–160 Lennon J.J. 2000. Red-shifts and red herrings in geographical ecology. Vasconcelos H.L. and Delabie J.H.C. 2000. Ground ant communities Ecography 23: 101–113 from central Amazonia forest fragments. Sample Ground-dwelling Leponce M., Theunis L., Delabie J.H.C. and Roisin Y. 2004. Scale ants: Case studies from the Worlds’ Rain Forests. Curtin Univer- dependence of diversity measures in a leaf-litter ant assemblage. sity School of Environmental Biology Bulletin, Perth, Australia 18: Ecography, 27: 253–267 59–70 Levings S.C. 1983. Seasonal, annual and among-site variation in the Vasconcelos H.L., Macedo A.C.C. and Vilhena J.M.S. 2003. Infl uence ground ant community of a deciduous tropical forest: some causes of topography on the distribution of ground-dwelling ants in an of patchy species distributions. Ecol. Monogr. 53: 435–455 Amazonian forest. Stud. Neotrop. Fauna Environ. 38: 115–124 Levings S.C. and Franks N.R., 1982. Patterns of nest dispersion in a Wilson E.O. 1958. Patchy distribution of ant species in New Guinea rain tropical ground ant community. Ecology 63: 338–344 forests. Psyche 65: 26–38 Levings S.C. and Windsor D.M. 1984. Litter moisture content as a Wilson E.O. 2003. Pheidole in the New World: A dominant, hyperdiverse determinant of litter arthropod distribution and abundance during Ant Genus. Harvard University Press, Cambridge, Massachusetts, the dry season on Barro Colorado Island, Panama. Biotropica 16: 794 pp 125–131 Wilson M.V. and Shmida A. 1984. Measuring beta diversity with pres- Liebhold A.M. and Gurevitch J. 2002. Integrating the statistical analysis ence-absence data. J. Ecol. 72: 1055–1064 of spatial data in ecology. Ecography 25: 553–557 Yanoviak S. and Kaspari M. 2000. Community structure and the habi- Liebhold A.M., Rossi R.E. and Kemp W.P. 1993. Geostatistics and tat templet: ants in the tropical forest canopy and litter. Oikos 89: geographic information systems in applied ecology. Annu. Rev. 259–266 Entomol. 38: 303–327 62 Chapitre 3: Bromeliads affect the ant distribution

TERRESTRIAL BROMELIAD PRESENCE AFFECTS THE GROUND-

DWELLING ANT DISTRIBUTION IN CHACOAN FORESTS

LAURENCE THEUNIS 1, 2 YVES ROISIN 2, AND MAURICE LEPONCE 1.

LAURENCE THEUNIS

Phone +32 2 627.43.64 Fax +32 2 649.48.25 E-mail: [email protected]

Running head: Bromeliads affect the ant distribution

63 Chapitre 3: Bromeliads affect the ant distribution

ABSTRACT

Terrestrial bromeliads are a characteristic feature of the monte fuerte, a subtropical semideciduous forest of the humid Chaco in northern Argentina. The bromeliad distribution is affected by the soil microrelief. In forest zones with a continuous cover of terrestrial bromeliads, we previously demonstrated that, at a 1m² scale, the bromeliad density structured the spatial distribution of some ant species and affected positively the overall ant species density and richness. To complete these observations, we measured here the ant species relative frequency, density, turnover and richness in large zones of the same forest devoid of bromeliads. The presence of bromeliads doubled the ant species density per m², increased the species richness and turnover. Most ant species were common inside and outside bromeliad zones but the dominant species differ between the two habitats. Among 72 species collected in the monte fuerte (in 200m² of leaf-litter), 14 were positively and 2 were negatively associated to the bromeliad presence. Overall the total number of species was higher in bromeliad zones compared to zones deprived of them (62 vs 47 species) because of a higher species density and turnover. Our results stress the importance of considering the ground cover vegetation when studying the diversity and distribution of ground-dwelling ant species. In the humid Chaco the presence of terrestrial bromeliads clearly promoted the ant diversity and density. The patchily distribution of bromeliads in forests also acted as a factor increasing the spatial heterogeneity of ant distribution.

64 Chapitre 3: Bromeliads affect the ant distribution

INTRODUCTION

Vegetation structure and composition influence the distribution of ground-dwelling ant colonies and can induce patchiness in their distribution (Wilson 1958, Gadagkar et al. 1993, Feener and Schupp 1998, Moutinho 1998, Retana and Cerdà 2000, Bestelmeyer and Wiens 2001). In particular, the presence of terrestrial Bromeliaceae favours the diversity of several arthropods taxa (Benzing 1986, 2000). For example, they increase the diversity and the density of the jumping spiders (Romero and Vasconcellos-Neto 2004) which in a mutualistic relation supply nutrients to the plants (Romero et al. 2006). We previously investigated the ant species distribution in a forest with a continuous ground cover of terrestrial bromeliads and observed that the ant density and richness were positively correlated to bromeliad density (Theunis et al. 2005). Bromeliad density increased the abundance of two common ant species and structured their spatial distribution. In complement to our previous results in zones of continuous bromeliad cover, our goal here was to further evaluate the impact of the bromeliad presence/absence on the structure of the ant assemblage. This comparison was made possible by the presence in the same forest of large contiguous zones with and without terrestrial bromeliads.

65 Chapitre 3: Bromeliads affect the ant distribution

MATERIAL AND METHODS

Study Area

We carried out the study in the Rio Pilcomayo National Park situated in the humid Chaco of the northeastern Argentina (25°04’06’’ S, 58°05’36’’ W). The park is a mosaic of vegetation types, depending primarily on inundation frequency. The present study was limited to the semideciduous forest (Monte Fuerte), which occupies 20-22% of the park area (Pujalte et al. 1995) and displays a high degree of natural fragmentation amidst the grassland. This forest was dominated by Schinopsis balansae Engl., Astronium balansae Engl. and Aspidosperma quebracho-blanco Schlecht. and by a ground strata of bromeliads (Aechmea distichantha Lemaire and Pseudananas sagenarius (Arruda) Camargo) (Pujalte et al. 1995) (Fig. 3-1).

Ant sampling protocol

Leaf-litter ants were collected along 500m linear transects perpendicular to the edge of large forest fragments of approximately 250ha. Three parallel transects, distant of 10m, were conducted in October 2001 in a forest fragment where the ground cover of bromeliads was patchily distributed. Another transect was conducted in October 2002 in a large fragment with a continuous bromeliad cover. This constitutes additional data compared to our previous study (Theunis et al. 2005). Fifty samples, spaced at 10m intervals, were collected along each transect. At each sampling point, the leaf litter found inside a 1m²-quadrat was collected, sifted and put in a cotton bag. The arthropods were then extracted from the sifted leaf litter with a mini-Winkler apparatus (Fisher, 1998) during 24 hours. Altogether 115 out of the 200 samples were collected outside bromeliad patches. Inside bromeliad patches, the data sets collected in 2001 (n=35 samples) and 2002 (n=50) did not differ in species density, composition and accumulation (same occurrence-based rarefaction curves), and were pooled (n=85).

Data Analyses

Ant assemblages inside or outside bromeliad zones were compared on basis of their species richness, density and composition. Species richness was standardized by sample-based rarefaction (Coleman method of EstimateS 7.5 (Colwell 2004)) and compared by bootstrapping (PAST software, Hammer et al. 2001). Species density

66 Chapitre 3: Bromeliads affect the ant distribution

(number of species/m²) and the turnover of species (β diversity - Jaccard index) were compared between the two forest zones using a Mann-Whitney U test. Faunal similarity of ant assemblages between the two zones was expressed by the NNESS index (Grassle and Smith 1976), which is given by:

NNESSij / k =

1 - {ESSij / k / [(ESSii / k + ESSjj / k)/2]} where ESSij / k is the expected number of species shared for random draws (without replacement) of k occurrences from localities i and j. When k is small, the index is highly sensitive to the occurrences of the most frequent species. When k increases, the influence of rarer species is emphasized. If result patterns change when k varies, it could mean that different processes structure the diversity of common and rarer species. Similarity of ant assemblages was calculated for k=1 (identical to the Morisita- Horn index), k=128 and k=256. Values range from 0, if species collected in localities i and j are totally different to 1, if localities host the same species. The software program BiodivR 1.0 (Hardy 2005) was used to compute the NNESS index. We evaluated the association between the frequent species (defined as species present in more than 10% of samples) and the bromeliads by conducting Chi² tests.

67 Chapitre 3: Bromeliads affect the ant distribution

RESULTS

Ant species richness was higher inside than outside bromeliad patches (observed species richness: 62 vs. 52 species; total species richness in 85m2: 62 vs. 47 species p= 0.05, standardized species richness for 396 occurrences= 57 vs. 52 species) (Fig. 3-2). The density of species was two-fold greater inside than outside bromeliad patches (Median (min-max): 6 (0-14) vs. 3 (0-13), Mann-Whitney U test: U= 2627.5, p<0.001). The turnover of species (Jaccard index) between samples in bromeliad patches was higher (i.e. less similar) than outside bromeliad patches (Mean ± SE: 0.114 ± 0.002 (n=2605) vs. 0.129 ± 0.003 (n=3003), t-test= -4.03, p<0.001). The ant faunal similarity (NNESS index) between the two zones increased with increasing k (k=1, NNESS=0,663; k=128, NNESS= 0,809 and for k= 256, NNESS= 0,804) indicating that the two groups are more similar when the rarer species are considered. Indeed, we found 22 and 9 frequent species in samples with and without bromeliads, respectively; eight of them were shared (Table 3-1). Two species were negatively associated with the bromeliad presence: Crematogaster sp.17 (Chi²=8.21, df=1, p<0.01) and Solenopsis sp. 18 (Chi²=6.56, df= 1, p<0.01). In contrast, 14 species were positively associated to the bromeliad presence (Table 3-1). Among them, Rogeria scobinata was absent in zones devoid of bromeliads and 10 others species, rare outside bromeliad patches, were collected in quadrats adjacent to a bromeliads zone.

68 Chapitre 3: Bromeliads affect the ant distribution

DISCUSSION

We previously showed that inside forest zones with a continuous bromeliads cover, a high bromeliad density favored a high abundance of some ant species (Solenopsis sp. 01 and Brachymyrmex physogaster) and/or spatially structured the distribution of Solenopsis sp. 01 and Paratrechina sp. 02 (Theunis et al., 2005). The present work further demonstrates that the bromeliad presence/absence influences the structure of the ant assemblage. More than 60% of the common species were positively associated to the bromeliad presence. The total number of species is higher in bromeliad zones compared to zones deprived of them because both the species density and turnover of ants are higher when bromeliads are present. Bromeliad presence also tend to increase the number of frequent species in the ant assemblage. Bromeliad leaves form a rosette accumulating rain and litter, and contribute to favourable moisture and temperature conditions for most arthropods (Benzing 1980, Romero and Vasconcellos-Neto 2004, Barberis and Lewis 2005). Bromeliads cover associated to convex, well drained, soil zones in the forest fragments (Barberis et al. 1998) favoured the leaf-litter ant assemblage probably by supplying a favourable environment (food, nest and shelter). The prey availability might be a factor attracting predator ant species near bromeliads (e.g. Gnamptogenys striatula, Octostruma rugifera, Pyramica denticulata and Hypoponera species). Soil- and litter-nesting species (e.g. Solenopsis, Pheidole, Brachymyrmex and Paratrechina species) might be attracted by the abundant organic matter found around bromeliad roots.

The spiny leaves of bromeliads also provide ant protection against their large predators such as giant anteaters, tamanduas or armadillos (Pujalte et al. 1995, Eisenberg and Redford 1999). Two species were negatively associated with bromeliad presence (Crematogaster sp. 17 and Solenopsis sp.18), this is possibly an effect of competitive displacement by ants associated to bromeliads. Crematogaster species in neotropics are often arboreal species foraging on the floor and Solenopsis sp. are litter omnivores or seed specialists (J.H.C. Delabie, pers.com.) Seven other species were not influenced by the bromeliad presence. Among them, we found generalist species using various micro-habitats (pieces of wood, twigs, crotches of trees,...) and food sources as Solenopsis sp. 02, S. sp17, Carebarella bicolor, Pheidole

69 Chapitre 3: Bromeliads affect the ant distribution

flavens (Wilson 2003) and Wasmannia sp.01 or arboreal species as Crematogaster sp.02 and Cephalotes minutus (Andrade and Baroni Urbani 1999). Bromeliads also favoured a higher soil termite species diversity and abundance (Roisin and Leponce 2004). Terrestrial bromeliads also affect the distribution of other arthropods such as jumping spiders as shown by Romero and Vasconcellos-Neto (2004) in grasslands and forest of the southeastern Brazil. Our results stress the importance of considering the ground cover vegetation when studying the diversity and distribution of ground-dwelling ant species. In the humid Chaco the presence of terrestrial bromeliads clearly promotes the ant diversity and density. The patchily distribution of bromeliads in forest islets also acts as a factor increasing the spatial heterogeneity of ground-dwelling ant distribution.

70 Chapitre 3: Bromeliads affect the ant distribution

Acknowledgments – We thank the Administración de Parques Nacionales, Buenos Aires, Argentina, for allowing us to collect in P.N. Río Pilcomayo. Nestor Sucunza, the guardaparques and Cornelio Pares greatly facilitated our work in the park. Thanks to G.J. Torales and E.R. Laffont, Univ. Nacional del Nordeste, for logistic support. This work was supported by the ‘Fonds National pour la Recherche Scientifique’ (FNRS, Belgium) to LT (PhD Grant). A grant to LT from the ‘Fonds Léopold III pour l’Exploration et la Conservation de la Nature’ allowed a field campaign in Argentina. We would like to thank also Prof. J. H. C. Delabie and Dr. I. C. do Nascimiento (CEPEC, Brasil) for help in ant identification. Isabelle Bachy for her GIS software support and Julien Cillis for the SEM images of ant species. The manuscript benefited from comments by Dr. T. Delsinne (RBINS) and Dr. A-C. Mailleux (ULB).

71 Chapitre 3: Bromeliads affect the ant distribution

REFERENCES Andrade, M. L. de, and Baroni Urbani, C. (1999) Diversity and adaptation in the ant genus Cephalotes, past and present (Hymenoptera, Formicidae). Stuttgarter Beiträge zur Naturkunde Serie B (Geologie und Paläontologie) 271: 1-889. Barberis, I.M., Pire, E.F. and Lewis, J.P. (1998). Spatial heterogeneity and woody species distribution in a Schinopsis balansae (Anacardiaceae) forest of the Southern Chaco, Argentina. Revista de Biologia Tropical 46: 515-524. Barberis, I.M. and Lewis, J.P. (2005) Heterogeneity of terrestrial bromeliad colonies and regeneration of Acacia praecox (Fabaceae) in a humid-subtropical-Chaco forest, Argentina. Revista de Biologia Tropical 53: 377-385. Benzing, D.H. (1980) The Biology of the Bromeliaceae. Mad River Press Inc., Eureka, California: 305 p. Benzing, D.H. (1986) Foliar specialization for animal-assisted nutrition in Bromeliaceae. – In: Juniper, B. and Southwood, R. (eds). Insects and the plant surface. Edward Arnold, London, UK. pp. 235-256. Benzing, D.H. (2000) Bromeliaceae: Profile of an Adaptive Radiation. Bennett, B., Brown, G., Dimmitt, M., Luther, H., Ramirez, I., Terry, R. and Till, W. (Eds). Cambridge University Press: 708 p. Bestelmeyer, B. and Wiens, J.A. (2001) Local and regional-scale responses of ant diversity to a semi-arid biome transition. Ecography 24: 381-392. Colwell, R. K. (2004) EstimateS: Statistical estimation of species richness and shared species from samples. Version 7.5. Persistent URL: . Eisenberg, J.F. and Redford, K.H. (1999) Mammals of the Neotropics: the central Neotropics, Volume 3. The University of Chicago Press: 609 p. Feener, D.H. and Schupp, E.W. (1998) Effect of treefall gaps on the patchiness and species richness of Neotropical ant assemblages. Oecologia 116: 191-201. Fisher, B.L. (1998) Ant diversity patterns along an elevational gradient in the Réserve Spéciale d'Anjanaharibe-Sud and on the Western Masoala Peninsula, Madagascar. Fieldiana Zool. (n.s.) 90: 39-67. Gadakgar, R., Nair, P., Chandrashekara, K. and Bhat, D.M. (1993) Ant species richness and diversity in some selected localities in Western Ghats, India. Hexapoda 5: 79-94. Grassle, J. F. and Smith, W. (1976) A similarity measure sensitive to the contribution of rare species and its use in investigation of variation in marine benthic communities. Oecologia 25: 13-22.

72 Chapitre 3: Bromeliads affect the ant distribution

Hammer, Ø., Harper, D.A.T., and Ryan, P.D. (2001) PAST: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica 4: 1-9. Hardy, O. (2005) BiodivR 1.0. A program to compute statistically unbiased indices of species diversity within samples and species similarity between samples using rarefaction principles. Université Libre de Bruxelles. Persistent URL: Moutinho, P.R.S. (1998) Impactos do uso da terra sobre a fauna de formigas, consequências para recuperação florestal na Amazônia Oriental. - In: Gascon, C. and Moutinho, P. (Eds). Floresta Amazônica: dinâmica, regeneração e manejo. Manaus, MCT-INPA, pp. 155-170. Pujalte, J.C., Reca, A.R., Balabusic, A., Canevari, P. Cusato, L. and Fleming, V.P. (1995) Anales de parques nacionales. Unidades Ecológicas del Parque Nacional Río Pilcomayo. Administración de Parques Nacionales XVI: 1-185. Retana, J. and Cerdà, X. (2000) Patterns of diversity and composition of Mediterranean ant communities tracking spatial and temporal variability in the thermal environment. Oecologia 123: 436-444. Roisin, Y. and Leponce, M. (2004) Characterizing termite assemblages in fragmented forests: a test case in the Argentinian Chaco. Austral Ecology 29: 637-646. Romero, G.Q. and Vasconcellos-Neto, J. (2004) Spatial distribution patterns of jumping spiders associated with terrestrial bromeliads. Biotropica 36: 596-601. Romero, G.Q., Mazzafera, P., Vasconcellos-Neto, J. and Trivelin, P.C.O. (2006) Bromeliad-living spiders improve host plant nutrition and growth. Ecology 87: 803- 808. Theunis, L., Gilbert, M., Roisin, Y. and Leponce, M. (2005) Spatial structure of litter- dwelling ant distribution in a subtropical dry forest. Insectes Sociaux 52: 366-377. Wilson, E.O. (1958) Patchy distribution of ant species in New Guinea rain forests. Psyche 65: 26-38. Wilson, E.O. (2003) Pheidole in the New World: A dominant, hyperdiverse Ant Genus. Harvard University Press, Cambridge, Massachusetts: 794 p.

73 Chapitre 3: Bromeliads affect the ant distribution

TABLE AND FIGURE LEGENDS

Table 3-1: Proportion (%) of samples occupied by each species inside (n= 85) or outside (n=115) bromeliad patches and associations between ants and bromeliads (Chi² test, 1 df). Twenty-three species, which were present in >10% of the samples collected in any of the two groups compared, were considered here (bold values). Species in bold correspond to species collected in a quadrat devoid of bromeliads but near (at less than 10m) a bromeliad zone. Degrees of significance are: * = p<0.05, ** = p<0.01 and ***=p<0.001 and ns for non significant.

Figure 3-1: Terrestrial Bromeliad cover in “Monte Fuerte” forests of the argentinian humid Chaco.

Figure 3-2: Sample-based rarefaction curves [Coleman method of EstimateS 7.5, Colwell 2004] of species richness (mean ± SD) and unique (species found in one sample) inside and outside bromeliads.

74 Chapitre 3: Bromeliads affect the ant distribution

Table 3-1

Association with Bromeliads Species bromeliad presence present absent Crematogaster sp.02 53 44 no Wasmannia sp.01 36 25 no Pheidole flavens 22 31 no Solenopsis sp.02 22 27 no Cephalotes minutus 21 13 no Solenopsis sp. 17 21 15 no Solenopsis sp. 18 18 34 - ** Carebarella bicolor 12 20 no Rogeria scobinata 16 0 + *** Solenopsis sp.01 47 6 + *** Pyramica denticulata 44 1 + *** Pheidole sp.30 11 2 + ** Brachymyrmex physogaster 25 2 + *** Brachymyrmex sp.05 25 1 + *** Octostruma rugifera 22 2 + *** Pheidole nubila 20 1 + *** Pheidole sp.22 20 4 + *** Hypoponera sp. prox. trigona 14 1 + *** Gnamptogenys striatula 12 3 + ** Hypoponera opacior 12 4 + * Hypoponera sp. prox. opaciceps 01 12 3 + ** Paratrechina sp.02 35 5 + *** Crematogaster sp.17 5 18 - **

75 Chapitre 3: Bromeliads affect the ant distribution

Figure 3-1

76 Chapitre 3: Bromeliads affect the ant distribution

Figure 3-2

77 78 Chapitre 4: Ants in old forest fragments

LONG-TERM EFFECTS OF FOREST FRAGMENTATION ON ANTS OF THE

HUMID CHACO.

LAURENCE THEUNIS 1, 2, YVES ROISIN 2 AND MAURICE LEPONCE 1

LAURENCE THEUNIS

Phone +32 2 627.43.64 Fax +32 2 649.48.25 E-mail: [email protected]

Running head: Ants in old forest fragments

79 Chapitre 4: Ants in old forest fragments

ABSTRACT

Naturally fragmented forests offer the opportunity to study the effects of the area, shape and isolation of the fragments on the species richness, density and composition of its inhabitants without confounding effects of the disturbance caused by a recent deforestation. Such system was found in the humid Chaco were patches of dry subtropical forest was spread among a palm savanna periodically inundated and burnt. Ground-dwelling ants living in these two habitats were collected every 10m along standardized “A.L.L.” transects by pitfall traps and mini- Winkler extracts of the leaf-litter fauna. The size of forest fragments ranged between small (<4ha), medium (15-30ha) and large (ca. 250ha). Isolation ranged from 20m up to 600m to the nearest fragment. A total of 118 species were collected, 114 in forest, 13 in grassland. Species richness was positively correlated to the fragment size and edge/area ratio (i.e. shape index). Species density was positively correlated with the shape index. Species composition was influenced by habitat type. Species common in forests (n= 26) were generally not found in grassland. Species common in grassland were grassland specialists (n=4), forest edge species (n=3) or ubiquitous species (n=4). At edges, a higher leaf-litter quantity and a higher shrub density were sometimes observed together with a higher ant density but slightly higher species richness. A higher ant species turnover was also observed at edges which were more occupied by ubiquist and savanna species. By contrast, most forest specialists were indifferent to the distance from the fragment edge. These factors interacted and small, isolated fragments were the most affected by fragmentation because of a high edge proportion and significantly lower species richness, in particular of true forest species. By contrast a few species were more frequent in these fragments than elsewhere. In comparison to studies combining fragmentation and deforestation effects, the edge effect appeared moderate, in particular with no clear peak of species richness near the edge. Furthermore it seems that non-isolated 20ha fragments were sufficient to sustain the whole forest ant assemblage.

80 Chapitre 4: Ants in old forest fragments

INTRODUCTION

Habitat fragmentation is one of the major threats for the maintenance of biodiversity (Brooks et al. 2002, Wilson 2002, Millenium Ecosystem Assessment [MEA] 2005). It affects the habitat size, increases its isolation and changes the environmental conditions especially at the edge of the fragment (MacArthur and Wilson 1967, Andrèn 1994, Laurance et al. 1998, Laurance et al. 2002, Laurance and Vasconcelos 2004). Ants with their sessile colonies and diversified ecological niches (in terms of food resources, nesting sites, interaction with plants and others organisms) are early indicators of small-scale habitat changes (Majer 1983, Majer et al. 1984, Bestelmeyer and Wiens 1996, Andersen 1997, Vasconcelos 1999, Carvalho and Vasconcelos 1999, Agosti et al. 2000, Vasconcelos et al. 1998, Vasconcelos et al. 2000, Brühl et al. 2003, Silva et al. 2007). The effects of forest fragmentation on ants have been well documented since the 80’s in the Biological Dynamic of Forest Fragment Project (Laurance et al. 2002) a large-scale and long- term experiment in Central Amazonia. During this project, fragments of 1, 10 and 100 ha were isolated by distances of 80-650 m from surrounding forest by clearing the vegetation to establish cattle pastures. Others fragmented forests as the Atlantic rain forest of Bahia (Majer et al. 1997, Delabie et al. 2000) and the forests of Borneo (Brühl et al. 2003), particularly exploited during the second part of the 20th century, constituted also model landscapes to study the effects of forest degradation and fragmentation on ants. These studies showed that small (ca. 1 ha), isolated (several hundreds of metre), forest fragments both species richness and species density were lower than in continuous forests (Vasconcelos 1988, Didham 1997b, Vasconcelos and Delabie 1998, Brühl et al. 2003, Schoereder et al. 2004, Vasconcelos et al. 2006). Ant species composition of small fragments appeared to be much influenced by the structure and the composition of the vegetation surrounding them (Vasconcelos and Delabie 1998, Armbrecht and Ulloa-Chacon 1999). It was found that edge effects affected the ant species richness, density, composition and abundance up to 200m inside the forest (Didham 1997a, 1998b, Majer et al. 1997, Vasconcelos et al. 1998, Carvalho and Vasconcelos 1999, Laurance et al. 2002). In large fragments, differences in ant composition were partly due to a higher quantity of leaf litter near edges (Carvalho and Vasconcelos 1999). It has been shown that the sharper the transition is between the fragment and the matrix vegetation, the lower the number of ant species could immigrated in the fragment (Gascon et al. 1999, Vasconcelos 1999, Kotze and Samways 2001). The long-term viability of fragments and the nature of the equilibrium point that they reach constitute the information of greatest value in conservation planning (Turner 1996). We were 81 Chapitre 4: Ants in old forest fragments interested here to study the long-term effects of natural fragmentation (i.e. the breaking apart as defined by Fahrig (2003)) on the structure of the ant assemblage. We took advantage of a naturally fragmented forest system present in the humid Chaco where islets of forest are present on slight mounds and are surrounded by grassland periodically inundated or burnt. Our aim was to evaluate to which degree species richness, species density, species turnover and species composition were influenced at long-term by the size, shape and isolation of the fragments. Additionally we evaluated the edge effects, in large fragments, on ant species density and composition and the introgression of species from the grassland surrounding the forest fragments.

82 Chapitre 4: Ants in old forest fragments

MATERIAL AND METHODS

Study Area

The study was carried out in the Rìo Pilcomayo National Park (25°05’ S, 58°08’ W) in the Argentinean wet Chaco region characterized by a rainfall regime >1000 mm and periodic flooding of the grassland surrounding forest islets. The park was created in 1951 and left undisturbed since then. Anthropogenic disturbance before 1951 consisted in wood exploitation and cattle rearing by native population. The park was composed of a mosaic of vegetation types, permanently (“estero”), periodically (“pastizal”) or never (“monte fuerte”) flooded. The “monte fuerte”, a dense (“impenetrable”) subtropical mesoxerophile oligarchic forest (Pujalte et al. 1995, habitat unit PHYSIS 48.2412 of Devillers and Devillers-Terschuren 1996), occupies 20-22% of the park area (Pujalte et al. 1995), and displays a considerable degree of natural fragmentation amidst the pastizal. The forest fragments (“monte fuerte”) were located on slight mounds and both natural fires and floods occurring regularly in the surrounding grassland maintain their edges. The forest was characterised by a tree storey composed of Schinopsis balansae Engl., Astronium balansae Engl. and Aspidosperma quebracho-blanco Schlecht. and by a ground stratum of terrestrial bromeliads (Aechmea distichantha Lemaire and Pseudoananas sagenarius (Arruda) Camargo) (Pujalte et al. 1995). The savanna was dominated by herbaceous vegetation as Setaria sp., Luziola peruviana and palm trees as Copernicia alba Morong. (Pujalte et al. 1995).

Ant sampling protocol

To collect ants, we followed the methods and the spatial design of the standardized ALL protocol (Agosti and Alonso 2000). Sampling points were distant of 10m along linear transects running in the monte fuerte or in the pastizal. Pitfall traps were only used to compare the ant fauna inhabiting the forest and the grassland. In all other transects we used the Winkler method which consists in collecting the leaf-litter present inside 1m²-quadrats and in extracting the ant fauna with a mini-Winkler apparatus (Fisher 1998) during 24 hours (instead of the standard 48 hours which don’t improve the extraction results in this habitat type (Delsinne et al. 2008). The sifted leaf-litter was weighted before extraction. Most of the Winkler transects we conducted correspond to the standardized 200m long transect designed to collect ants of the leaf-litter (“A.L.L.” transect of Agosti and Alonso 2000). In the habitat studied, a single standardized A.L.L. transect with 20 Winkler samples collects less than 45% of the ant species present in the forest fragment (Leponce et al. 2004). All frequent species are included but their relative

83 Chapitre 4: Ants in old forest fragments frequency is not always representative. Ants were collected in September and October during four consecutive years (1999-2002). Temperature during the sampling sessions was on average of 27.7 ± 5.1°C. A total of 540 leaf litter and 465 pitfall traps samples were collected. Effects of size, shape and isolation of the forest fragments We used a SPOT satellite image to select eleven accessible forest fragments belonging to three size categories: small (< 4ha, n=5 fragments, “S1”-“S5”), medium (between 15 and 30ha, n=4 fragments, “M1”-“M4”) and large (around 250ha, n=2 fragments, “L1” and “L2”) in three localities (“Esteros”, “Fonzo” and “la Angela”) (Fig. 4-1). In small and medium forest fragments, standardized 200m long A.L.L. transects were conducted along the longest axis of the fragment. In large fragments, transects were 500 m long and perpendicular to one of the fragment edge.

Three parallel 500m transects L1A, L1B and L1C, 10m apart, were conducted in “L1” and a single 500m transect was conducted in “L2”. To study the effects of fragment size we only considered samples located further than 200m from the edge. Two hundred meters is the maximal distance at which edge effects on invertebrate assemblages (including ground- dwelling ants) have been detected (Laurance et al. 2002, Carvalho and Vasconcelos 1999, Didham 1997a, 1997b). All forest fragments had a dense and continuous ground stratum of bromeliads, except “L1” where these plants were patchily distributed. The bromeliad density affects the ant and termite species distribution (Roisin and Leponce 2004, Theunis et al., 2005). In order to calibrate our measures, the variability in species richness, density and composition measured between fragments of the three size classes was compared to the intra-fragment variability assessed by conducting 8 A.L.L transects (“M*1”-“M*8”) in the medium fragment “M3” (see Leponce et al. 2004 for details). Edge effects on ant species density and composition Edge effects were studied in the large fragments “L1” and “L2”. The variation of species density was studied along L1A, L1B and L1C. Only quadrats devoid of terrestrial bromeliads (n= 114) were considered in the analysis to avoid any confounding effect of bromeliad density (Theunis et al. 2005, Roisin and Leponce 2004).

The introgression of grassland ant species into the forest was studied along transects L1B and L2, which were extended up to 50m in the grassland. Pitfall traps were placed every 10m (next to the 1m² leaf-litter quadrats) along each 550m transect (n=110 pitfalls) and left during 6 days in the field. The variation of 4 environmental variables from the edge towards the centre of the forest fragments L1B and L2 was evaluated in a 500m long by 2m wide strip (also used to study the termite distribution, see Roisin & Leponce 2004). The strip was divided in 100 quadrats of 5m x 2m. The number of trees, their diameter at breast height, the number of shrubs, the number of 84 Chapitre 4: Ants in old forest fragments bromeliad rosettes were measured in these 10m² quadrats. In addition, the leaf-litter collected in 1m² quadrats every 10m was weighted before Winkler extraction (along L1B only). Changes in species composition of ants (Carvahlo and Vasconcelos 1999) and other litter invertebrates (Didham 1997b, Brown and Hutchings 1997, Didham et al. 1998b) have been detected up to 200m into the forest. Species preference in terms of environmental conditions (i.e. forest edge, 0-200m in the forest (n=120), forest center, 200-500m (n=180), grassland (n=10)) was studied by pooling Winkler and pitfall catches from “L1” (150 Winkler and 55 pitfalls samples) and “L2” (50 Winkler and 55 pitfalls samples). Only frequent species (defined as occurring in >10% of the samples collected in each zone defined above) were considered for this analysis. Species occurrence data were used instead of species abundance data in order to limit the weight of samples collected around nests, trails and resources exploited by ants.

Fragmentation indices

Seven fragmentation indices were calculated. The area (A) and perimeter (P) of fragments were calculated from the satellite image with MapInfo Professional 6.5 (MapInfo Corporation 2001). The fragment shape was characterized by the ratio logP/logA and by the Shape Index defined by Patton (1975) as P/(200*(π * A)1/2), where P is the perimeter of the forest patch in metres, and A is its area in hectares. Its value is 1 for a circle. Values >1 represent deviations from circularity (Laurance and Yensen 1991, Schumaker 1996). The fragment isolation was measured by the distance to the nearest fragment (NF) and by the mean nearest fragment distance (MNFD) calculated among all neighbour patches within 600 m around the fragment considered. The radius of 600 m was chosen because very few ant species can disperse, by nuptial flights, over greater distances (Hölldobler and Wilson 1990, Holway 1998, Holway et al. 2002, Hoffman et al. 1999, Ingram 2002). The isolation of a habitat patch depends not only of the distance to the nearest patch, but also of the area of the nearest patch. For this reason, a third isolation measure, called the Proximity Index (PI), was used (Gustafson and Parker 1994). It was calculated, for all patches within the 600m radius from edges, by first calculating PX= ∑

2 An/dn , where An is the area of the neighbour patch and dn is the distance between the patch and its neighbour. PX is then divided by the mean distance between the patch considered and all its neighbours to obtain PI.

85 Chapitre 4: Ants in old forest fragments

Data Analyses

Size, shape and isolation of forest fragments For all forest fragments, the Pearson correlation coefficient was calculated between species density, species richness and the 7 fragmentation indices (see Table 4-2). We then evaluated the impact of fragmentation on ant species composition. Species richness The measure of species richness is strongly dependent of the number of samples collected and of the species density during the sampling period (Gotelli and Colwell 2001, Leponce et al. 2004). To standardize and compare species richness among the different fragments we relied on Melo’s method (Melo et al. 2003) that compares the values of species richness expected for the largest occurrence (232 occurrences in our case) among all transects. This method avoids losing information by comparing the species richness expected for the smallest occurrence (58 occurrences in our case) among all transects with the traditional rarefaction method of Sanders (1968). In Melo’s method occurrence-poor datasets are extrapolated with an appropriate curve- fitting extrapolation model. In a previous study, some of us (Leponce et al., 2004) observed that the Soberón and Llorente logarithmic model (S=ln (1+z*a*x)/z where S is the number of species and a, b, z are fitted coefficients) performed well (Soberón and Llorente 1993, Fisher 1999). We also rarefied the species richness following the rarefaction method (Sanders 1968). Since the two methods gave consistent results, we only present here the results obtained by Melo’s method. Species richness values obtained with this method will be called hereafter “standardized species richness”. Species accumulation curves were obtained by the Mau-Tau method of EstimateS 8.0 program (Colwell 2006). Curve-fitting models were applied on accumulation curves with the non-linear estimation procedure and the quasi-Newton estimation method of Statistica 6.0 (StatSoft Inc 2001).

Species composition

Non-Metric Multidimensional Scaling (NMDS) was used to compare ant species composition between forest fragments. This technique preserves as well as possible the distance relationships among object and can produce ordinations of objects from any distance matrix (Legendre and Legendre 1998). The Bray-Curtis faunal similarity index, calculated on species frequencies, was used in the NMDS. NMDS was computed with PAST version 1.29 (Hammer et al. 2001). NMDS has already been applied in ant studies and produced robust results (Carvalho and Vasconcelos 1998, 1999, Golden and Crist 2000, Vasconcelos 1999, Vasconcelos et al.

86 Chapitre 4: Ants in old forest fragments

2000, Brühl 2003, Vasconcelos et al. 2006, Silva et al., 2007). A cluster analysis (UPGMA) was then superimposed upon the resulting spatial map to group the most similar fragments in terms of ant species composition. We tested the differences between clusters using an ANOVA performed on the ordination scores obtained with the NMDS.

87 Chapitre 4: Ants in old forest fragments

RESULTS

Effects of fragmentation on species richness and species density

Fragmentation indices, sampling procedure and diversity results are summarized in Table 4-1. Standardized species richness was positively correlated to the area (r= 0.60, p <0.05), perimeter (r= 0.66, p< 0.05), Shape Index (r= 0.77, p< 0.001) and marginally to the Proximity Index (r= 0.56, p= 0.07) (Table 4-2). A Mann-Whitney test showed a significant difference between the standardized species richness in small and medium fragments (U= 4, p=0.004) (Fig. 4-2). The difference of the standardized species richness between small and large fragments was marginal (U= 10, p =0.09) but we had only two measures for large fragments because of the difficulty of access to the sites. Differences of species density among transects were correlated with the Shape Index (r= 0.57, p<0,01) (Table 4-2) and were not related to the average temperature during sampling (Pearson’s r= 0.33, p= 0.16). Species density was compared between all fragments and the 8 transects ALL from the calibration transect (Table 4-1). An ANOVA revealed a significant difference in species number per m² (species density) between the 11 fragments (df = 10, F = 9.515, p < 0.001).

Species density in S1, S2, S5 and M1 was significantly lower than in S4, M2, M3, L2 and M*1-8 (Tukey-Kramer post-hoc tests: p< 0.05). The ant species density in S3, M4 and L1 was not significantly different from the one in other fragments.

Effect of fragmentation on ant species turnover and species composition

The Bray-Curtis faunal similarity index (the turnover) between quadrats belonging to the same transect (small: Mean= 0.2180 ± 0.0065, n= 946, medium: Mean= 0.2559 ± 0.0065, n= 760 and large: Mean= 0.2411 ± 0.0066, n=870) was significantly lower in small fragments (ANOVA: df= 2, F= 8.188, p<0.001; Student-Newman-Keuls multiple comparisons test: small- medium: p<.001, small-large: p<0.05 and medium-large: ns). The NMDS ordination of faunal similarities between fragments is presented in Figure 4-3. The NMDS results appeared reliable since point corresponding to the calibration transect were closely aggregated and close to the point corresponding to the transect conducted 10 month later in the same fragment M3. The distance between the eight M* transects and M3 corresponded to a difference in ant species activity after a 10-month interval. This effect was relatively weak because the species density between M*1-8 and M3 was very close (Table 4-1) and the composition highly similar. Temperature conditions during the two sampling periods were similar (30.3°C ± 1.9 (n=40) pour M1-8 vs. 31.1°C ± 2.7 (n=30) pour M3, unpaired t test, ns). 88 Chapitre 4: Ants in old forest fragments

Ant composition split into three clusters at 50 and 60 % of similarity: cluster I with S1, S2, M1 and M4. Cluster II with S3, S4 and S5 while M*, M2, M3, L1 and L2 were grouped into cluster III. The three clusters were significantly different from each other along axis 1 (ANOVA: df= 2, F= 22,882, p< 0.0001; Student-Newman-Keuls multiple comparisons test: cluster I – II, p<0.001, cluster I – III, p<0.05 and cluster II - III, p<0.001) and along axis 2 (ANOVA: df= 2, F= 58, 504, p< 0.0001; Student-Newman-Keuls multiple comparisons test: cluster I – II: p<0.001, cluster I - III: p<0.001 and cluster II – III: ns). Ant species composition was affected by fragment isolation. This was shown by the significant correlation between the scores produced by the first axis of the ordination analysis (NMDS) and the nearest fragment distance (Pearson’s r= -0.46, p< 0.05) and the mean nearest fragment distance (r= -0.47, p< 0.05). NMDS scores for the second axis of the ordination analysis were correlated with the Proximity Index (r= -0.46, p< 0.05). Differences in ant frequent species composition were also related to fragment size. Small fragments were rich in ubiquist species (11 species)(i.e. living in both the forest and grassland) as Camponotus sp.14, C. renggeri, C. rufipes, Odontomachus bauri, Paratrechina pubens and Solenopsis sp. 14. Medium and large fragments contained 11 forest species specialist as Crematogaster sp.17, Hypoponera sp. prox. opaciceps 1, Pachycondyla obscuricornis, Pheidole sp. 04, Ph. sp. 20, Rogeria scobinata, Solenopsis clytemnestra bruchi, S. sp. 18 and Strumigenys oglobini. Regarding species composition in the smallest and most isolated fragments (S3, S4 and S5), we noted that 3 species were found only in these fragments: Camponotus renggeri, Odontomachus bauri and Solenopsis sp.14.

Edge effects

Species density, measured in the 3 parallel Winkler transects performed in L1, was negatively correlated with the distance to the edge of the fragment (Pearson’s r correlation= -0.31, p= 0.034) (Fig. 4-4 A). The number of species collected by pitfall tended to be negatively correlated with the distance from the edge in the first large fragment (r= -0.44, p=0.02) but not in the second one (r= 0.22, p= 0.13) (Fig. 4-4 B&C). The density of trees, their diameter at breast height and the density of bromeliad rosettes did not vary significantly between the edge and the centre of the large fragment L1. However, the shrub density and the leaf litter quantity were negatively correlated with the distance from the fragment edge (r= -0.31, p< 0.05; r= -0.53, p<0.001, respectively). The number of species collected per pitfall was positively correlated with the leaf litter quantity in a radius of 1m around the pitfall (r= 0.33, p=0.023) but not with the shrub density (r= -0.10, p=0.69).

89 Chapitre 4: Ants in old forest fragments

The Bray-Curtis faunal similarity index was lower (Mann-Whitney Test: U=33005, p<0,001) between quadrats at edge (0 to 200m, mean= 0,18, n=210) than in the centre (210 to 500m, n=406) of the fragment L2. Thirty species were frequent either in the large forest fragments or in the grassland (Table 4-3). Three species were found in the grassland and at the edge of the forest fragment (Crematogaster sp.18, Paratrechina pubens and Pheidole aberrans). Four were encountered only in the grassland (Hypoponera sp.08, Paratrechina sp.01, Solenopsis sp. 14 and S. sp. 16) and three others preferentially (but not exlusively) in the center of the forest fragment (Brachymyrmex sp. 05, Paratrechina sp. 02 and Pyramica denticulata). More than half of the frequent species (16/30) were exclusively forest specialists species without preference for the edge or the center of the forest. Four were ubiquist species: Camponotus crassus, Labidus preadator, Pheidole sp.21 and Wasmannia sp.01.

90 Chapitre 4: Ants in old forest fragments

DISCUSSION

Our results suggest that the physical characteristics of the forest fragments influenced the structure and composition of ant communities in this naturally fragmented forest. In particular, species richness was correlated to the size, shape and isolation degree of the fragments. Species density was higher in fragments with greater proportion of edges. A greater activity of ground- foraging ant at the edge was observed in the two large forest fragments, possibly partly in response to higher litter quantity at the edge. Species composition was influenced principally by fragment isolation. Finally, the ant species found in the forest/grassland matrix were composed of grassland specialists, ubiquists and a majority of forest specialists whose distribution was independent of the distance from the fragment edge.

Effects of fragment size

Below 20ha, fragments contained a lower ant species richness. Area effects on ant species richness have been detected in other studies in both tropical and temperate zones (Dean and Bond 1990, Terayama and Murata 1990, Vasconcelos et al. 1998, Brühl et al. 2003, Schoereder et al. 2004). Despite an area several time larger in comparison to medium fragments, the large fragments did not support a higher ant species richness. Similarly, in South Africa, area effects were only apparent for fragments smaller than 20 ha (Dean and Bond 1990) and, at Alter do Chão (Brazil), area effects were more pronounced for fragments smaller than 15 ha (Vasconcelos et al. 2006). Schoereder et al. (2004) showed that smaller and more isolated fragments at Viçosa (SE Brazil) lost more ant forest specialists and were invaded more often by ubiquists (forest and grassland colonizers), which suffer higher extinction inside fragments. Consequently, small fragments presented, in their study, a higher turnover. We could not evaluate in our study the colonization or extinction rate of ant species because we did not sample the same sites during consecutive years. However, small fragments were rich in ubiquist species (6 species), presented a higher turnover and medium and large fragments contained a large proportion of forest species (9 species). Ant composition was thus also affected by fragment size. Our results were therefore in concordance with the observations of Schoereder et al. (2004). A lower carrying capacity of the habitat (and a lower availability in resources) would explain why more ubiquists and less forest specialists (more extinction prone) are sampled in small fragments in comparison with larger fragments.

91 Chapitre 4: Ants in old forest fragments

Effects of fragment isolation

The fragment isolation has a marked impact on species composition (Vasconcelos et al. 1998, Carvalho and Vasconcelos 1999, Schoereder et al. 2004, Brühl et al. 2003, Laurance et al. 2002, Didham 1996). Here, isolation indices are not related to a species depression. Noteworthingly, according to our NMDS study, smaller and more isolated fragments had a particular composition with 3 typical species. It is possible that fragment isolation constraints species dispersal especially in grassland that are regularly burned and/or flooded. Indeed, isolation might be seen as a filter selecting species which can live in small and isolated fragments with limited resources. Small areas of savanna might represent a barrier to species that reproduce by colony budding and/or show a strong association with the forest habitat. For example, forest species as some Solenopsis and Pheidole species (Porter et al. 1988, Wilson 2003) spread off by budding and found new colonies in the same fragment. Other species, well-adapted to the forest and matrix habitat, reproduce by nuptial flights and colonize neighbor forest fragments. Linepithema humile is able to cover a distance of 450 meters (Holway 1998, Ingram 2002), Wasmannia auropunctata up to 500 meters (Hoffmann et al. 1999) and very few species as Solenopsis invicta can fly over few kilometers (Porter et al. 1988, Hölldobler and Wilson 1990).

Effects of geographical distances

The decrease in species number and the variation in communities composition related to isolation degree of fragments most likely represent the effect of fragmentation rather than an eventual geographic effect as reported in literature (Majer 1983, Vasconcelos and Delabie 1998). Indeed, the leaf litter ant population can be assumed to be drawn from an identical local species assemblage because (1) the soils and forest types of the fragments (monte fuerte) are similar (Pujalte et al. 1995); and (2) the distances between the forests did not translate into proportional distances in NMDS.

Shape and edge effects

Interestingly, species richness and density was higher in fragments with a high Shape Index (ratio edge by area). Edges showed a higher shrubs density and greater leaf litter quantity increasing the structural heterogeneity of the ground vegetation in the fragment and consequently providing more ecological niches as demonstrated in other studies (Didham 1997a and b, Didham and Lawton 1999, Majer et al. 1997, Kotze and Samways 2001, Williams-Linera 1990, Carvalho and Vasconcelos 1999). Edges are more susceptible to be invaded by savanna and ubiquist species (Ewers and Didham 2006), increasing the turnover 92 Chapitre 4: Ants in old forest fragments between samples near edge. A high species turnover (because of much edge) and a high species density (because of more resources) induce a high species richness in fragments. Ant species density and ant activity were related to the distance to the forest edge as it has been observed in Amazonian forest fragmentation studies (Didham 1997a, Carvalho and Vasconcelos 1999). A greater density of shrubs and a greater quantity of litter could account for the increase of pitfall efficiency at the edge of the large fragment. We previously demonstrated the strong positive correlation between the leaf litter quantity and ant species richness (Theunis et al. 2005). However, we did not observe an “ecotone” (narrow zone with a peak of richness where edge species and forest centre species cohabit, Didham 1997a, Didham et al. 1998b). Our results suggest that the ant community composition within 200 m from edges differ little from the one deeper in the forest. More than half of the 30 frequent species (either in large forest and/or in grassland), including the 10 most frequent species, were forest species not affected by the distance from the edge. Among the other species, we found some ubiquists and some savanna species collected at the forest edge. Interestingly, no species were exclusively collected in the forest center. Many edge effects, including microclimatic effects (Camargo and Kapos 1995, Kapos et al. 1997), on leaf-litter decomposition (Rubinstein and Vasconcelos 2005), and on communities of dung beetles (Magurra et al. 2001, Quintero and Roslin 2005) have been found to be transitory in nature, and this may also apply for ant communities. A study conducted in South Africa with “old” edges (i.e. sharp edges between grassland and forest fragments) did not find evidence either of an edge effect on forest ant assemblages (Kotze and Samways 2001). Strong edge effects are always observed in fragments recently damaged. Then, forest regrowth (forming new edges) allows species to recolonize the edge. The more closely the regrowth habitat approximates the structure and microclimate of the forest, the more likely the fragmentation- sensitive species are able to use it (Laurance et al. 2002). In the NPRP, forest fragments are natural formed and maintained by natural fires and inundations preventing second regrowth of forest. These old forest edges resulted in dense vegetation that protects forest just inside it. The variation in vegetation structure at edges was too weak to modify the ant assemblage composition between the edge and the centre forest.

CONCLUSIONS

In comparison to studies combining fragmentation and deforestation effects, the edge effect appeared moderate, in particular with no clear peak of species richness near the edge.

93 Chapitre 4: Ants in old forest fragments

Furthermore it seems that non-isolated 20ha fragments were sufficient to sustain the whole forest ant assemblage.

94 Chapitre 4: Ants in old forest fragments

Acknowledgments - This study was made possible by financial support of the “Fonds National de le Recherche Scientifique” and the “Fonds Léopold III pour l’Exploration et la Conservation de la Nature”. We thank the Guards, Sr Sucunza, Cornélio Paredes for they help during the fieldwork or administrative procedures. Jacques Delabie and Ivan Nascimiento confirmed the ant identifications. Isabelle Bachy for her GIS software support and Julien Cillis for the SEM images of ant species. The manuscript benefited from comments by Dr Thibaut Delsinne, Dr Géraldine Kapfer (RBINS), Dr Anne-Catherine Mailleux (ULB), Dr Mohamed El Aydam (ULB) and Dr Jesus Millor (ULB).

95 Chapitre 4: Ants in old forest fragments

REFERENCES

Agosti, D. and Alonso, L.E. (2000) The ALL protocol. A standard protocol for the collection of ground dwelling-ants. - In: Agosti, D., Majer, J., Alonso, L. and Schultz, T. (eds) The ALL protocol. A standard protocol for the collection of ground dwelling-ants. Smithsonian Institution Press. pp. 204-206. Agosti, D., Majer, J., Alonso, L. and Schultz, T. (2000) Ants - Standard methods for measuring and monitoring biodiversity. (Eds) Agosti, D.; Majer, J. Alonso, L.; Schultz, T. Smithsonian Institution Press: 280 p. Andersen, A.N. (1997) Using ants as bioindicators: multiscale issues in ant community ecology. Conservation Ecology [online] 1: 8. http:// www.consecol.org/vol1/iss1/art8/. Armbrecht, I. and Ulloa-Chacón, P. (1999) Rareza y diversidad de hormigas en fragmentos de bosque seco colombianos y sus matrices. Biotropica 31: 546-653. Andrén, H. (1994) Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. Oikos 71: 355–366. Bestelmeyer, B.T., Wiens, J.A. (1996) The effects of land use on the structure of ground-foraging ant communities in the Argentinean Chaco. Ecological Applications 6: 1225–1240. Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G. and Hilton-Taylor, C. (2002) Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology 16: 909–923. Brown, K.S., Jr. and Hutchings, R.W. (1997) Disturbance, fragmentation, and the dynamics of diversity in Amazonian forest butterflies. In: Laurance, W.F. and Bierregaard, R.O. (eds). Tropical forest remnants: ecology, management, and conservation of fragmented communities. Chicago: University of Chicago Press. pp. 91-110. Brühl, C.A., Eltz, T. and Linsenmair, K.E. (2003) Size does matter - effects of tropical rainforest fragmentation on the leaf litter ant community in Sabah, Malaysia. Biodiversity and Conservation 12: 1371-1389. Camargo, J.L.C. and Kapos, V. (1995) Complex edge effects on soil moisture and microclimate in Central Amazonian forest. Journal of Tropical Ecology 11: 205–221. Carvalho, K.S. and Vasconcelos, H.L. (1998) Forest fragmentation in central Amazonia and its effects on leaf-litter ants. Biotropica 30 (2 suppl.) Carvalho, K.S. and Vasconcelos, H.L. (1999) Forest fragmentation in central Amazonia and its effects on litter-dwelling ants. Biological Conservation 91: 151-157. Colwell, R.K. (2006) EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.0. Persistent URL .

96 Chapitre 4: Ants in old forest fragments

Dean, W.R.J. and Bond, W.J. (1990) Evidence for rapid faunal changes on islands in a man- made lake. Oecologia 83: 388-391. Delabie, J.H.C., Agosti, D. and do Nascimento, I.C. (2000) Litter ant communities of the Brazilian Atlantic rain forest region. D. Agosti, J. Majer, L.E. Alonso and T. Schultz Sampling Ground-dwelling ants: case studies from the worlds'rain forests. Curtin University, Australia, School of Environmental Biology, Bulletin no 18:1-17. Delsinne, T., Leponce, M., Theunis, L., Braet, Y. and Roisin, Y. (2008) Rainfall influences ant sampling in dry forests. Biotropica (online published 6 may 2008). Devillers, P. and Devillers-Terschuren, J. (1996) A classification of South American habitats. IRSNB, Brussels, and Institute of Terrestrial Ecology, Huntinggdon: 419 p. Didham, R.K. (1996) Insects in fragmented forests: a functional approach. Trends in Ecology and Evolution 11: 255-260. Didham, R.K. (1997a) The influence of edge effects and forest fragmentation on leaf litter invertebrates in Central Amazonia. – In Laurance, W.F., Bierregaard, R.O. Jr (eds) Tropical forest remnants. Ecology, management, and conservation of fragmented communities. The University of Chicago Press: 55-70. Didham, R.K. (1997b) An overview of invertebrate responses to forest fragmentation. – In Watt, A.D.; Stork, N.E.; Hunter, M.D (eds). Forests and Insects Chapman & Hall, London: 303- 320. Didham, R.K. and Lawton J.H. (1999) Edge structure determines the magnitude of changes in microclimate and vegetation structure in tropical forest fragments. Biotropica 31: 17-30. Didham, R.K., Lawton, J.H., Hammond, P.M., Eggleton, P. (1998a) Trophic structure stability and extinction dynamics of beetles (Coleoptera) in tropical forest fragments. Philosophical Transactions - The Royal Society of London Biological Sciences 353: 437-451. Didham, R.K., Hammond, P.M., Lawton, J.H., Eggleton, P., Stork, N.E. (1998b) Beetle species responses to tropical forest fragmentation. Ecological Monographs 68: 295-303. Ewers, R.M. and Didham, R.K. (2006) Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews 81: 117–142. Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual Reviews of Ecology and Systematics 34: 487-515. Fisher, B.L. (1998) Ant diversity patterns along an elevational gradient in the Réserve Spéciale d'Anjanaharibe-Sud and on the Western Masoala Peninsula, Madagascar. Fieldiana Zoology (n.s.) 90: 39-67. Fisher, B.L. (1999) Improving inventory efficiency: a case study of leaf litter ant diversity in Madagascar. Ecological Applications 9: 714-731. 97 Chapitre 4: Ants in old forest fragments

Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L., Laurance, W.F., Zimmerman, B., Tocher, M. and Borges, S. (1999) Matrix habitat and species richness in tropical forest remnants. Biological Conservation 91: 223-229. Golden, D.M. and Crist, T.O. (2000) Experimental effects of habitat fragmentation on rove beetles and ants: patch area or edge? Oikos 90: 525-538. Gotelli, N.J. and Colwell, R.K. (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology Letters 4:379-391. Gustafson, E.J. and Parker, G.R. (1994) Using an index of habitat patch proximity for landscape design. Landscape and Urban Planning 29: 117-130. Hammer, Ø., Harper, D.A.T., and Paul D.R. (2001) Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica 4: 1-9. . Hoffmann, B.D., Andersen, A.N. and Hill G.J.E. (1999) Impact of an introduced ant on native forest invertebrates: Pheidole megacephala in monsoonal Australia. Oecologia 120: 595– 604. Hölldobler, B. and Wilson, E.O. (1990) The ants. Cambridge University Press, Cambridge. Holway, D.A. (1998) Effects of Argentine ant invasions on ground-dwelling arthropods in northern California riparian woodlands. Oecologia 116: 252–258. Holway, D.A., Lach, L., Suarez, A.V., Tsutsui, N.D. and Case, T.J. (2002) The causes and consequences of ant invasions. Ann. Rev. Ecol. Syst. 33: 181–233. Ingram, K.K. (2002) Flexibility in nest density and social structure in invasive populations of the Argentine ant, Linepithema humile. Oecologia 133: 492-500. Kapos, V., Wandelli, E.V., Camargo, J.L.C. and Ganade, G.M.S. (1997). Edge-related changes in environment and plant responses due to forest fragmentation in central Amazonia. pp. 33- 44. In Tropical Forest Remnants: Ecology, Management , and Conservation of Fragmented Communities. Laurance, W.F. and Bierregaard, R.O., eds. University of Chicago Press, Chicago, Ill., USA. Kotze, D.J. and Samways, M.J. (2001) No general edge effects for invertebrates at afromontane forest/grassland ecotones. Biodiversity and Conservation 10: 443-466. Laurance, W.F. and Yensen, E. (1991) Predicting the impacts of edge effects in fragmented habitats. Biological Conservation 55: 77-92. Laurance, W.F. and Vasconcelos, H.L. (2004) Ecological effects of habitat fragmentation in the tropics. In: Schroth, G., Fonseca, G.A.B., Harvey, C., Gascon, C., Vasconcelos, H.L., Izac, A.M. (eds.) Agroforestry and biodiversty conservation in tropical landscapes, pp 33-49, Island Press, Washington, DC. 98 Chapitre 4: Ants in old forest fragments

Laurance, W.F.,Ferreira, L.V., Rankin-de Merona, J.M., Laurance, S.G. (1998) Rain forest fragmentation and the dynamics of Amazonian tree communities. Ecology 79: 2032-2040. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L., Brunia, E.M., Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G. and Sampaio, E. (2002) Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16: 605-618.

nd Legendre, P. and Legendre, L. (1998). Numerical Ecology, 2 engl. edition. Elsevier, Amsterdam, XV+853p. Leponce, M., Theunis, L., Delabie, J.H.C. and Roisin, Y. (2004) Scale dependence of diversity measures in a leaf-litter ant assemblage. Ecography 27: 253-267. MacArthur, R.H. and Wilson, E.O. (1967) The theory of island biogeography. Princeton University Press, Princeton. Magura, T., Ködöböcz, V. and Thóthmérész, B. (2001) Effects of habitat fragmentation on carabids in forest patches. Journal of Biogeography 28: 129-138. Majer, J.D. (1983). Ants - useful bioindicators of minesite rehabilitation, land use and land conservation status. Environmental Management 7: 375-383. Majer, J.D., Day, J.E., Kabay, E.D. and Perriman, W.S. (1984) Recolonisation by ants in bauxite mines rehabilitated by a number of different methods. Journal of Applied Ecology 21: 355- 375. Majer, J.D., Delabie, J.H.C. and Mckenzie, N.L. (1997) Ant litter fauna of forest, forest edges and adjacent grassland in the Atlantic rain forest region of Bahia, Brazil. Insectes Sociaux 44: 255-266. Mapinfo Melo, A.S., Pereira, R.A.S., Santos, A.J., Shepherd, G.J., Machado, G., Medeiros, H.F. and Sawaya, R.J. (2003) Comparing species richness among assemblages using sample units: why not use extrapolation methods to standardize different sample sizes? Oikos 101: 398- 410. Millenium Ecosystem Assessment [MEA] (2005) Ecosystems and human well-being: biodiversity synthesis. World Resource Institute, Washington, D.C., 86 pp. Patton, D.R. (1975) A diversity index for quantifying habitat edge. Wildlife Society of Bulletin 3: 171-173. Porter, S.D., Van Eimeren, B. and Gilbert, L.E. (1988) Invasion of red imported fire ants (Hymenoptera: Formicidae): microgeography of competitive displacement. Ann. Entomol. Soc. Am. 81: 913–18.

99 Chapitre 4: Ants in old forest fragments

Pujalte, J.C., Reca, A.R., Balabusic, A., Canevari, P., Cusato, L. and Fleming, V.P. (1995) Anales de parques nacionales. Unidades Ecológicas del parque nacional Rio Pilcomayo. Administración de Parques Nacionales XVI: 1-185. Quintero, I. and Roslin, T. (2005) Rapid recovery of dung beetle communities following habitat fragmentation in Central Amazonia. Ecology 86: 3303–3311. Roisin, Y. and Leponce, M. (2004) Characterizing termite assemblages in fragmented forests: a test case in the Argentinian Chaco. Austral Ecology 29: 637-646. Rubinstein, A. and Vasconcelos, H.L. (2005) Leaf-litter decomposition in Amazonian forest fragments. Journal of Tropical Ecology 21: 699–702. Schoereder, J.H., Sobrinho,T.G., Ribas, C.R. and Campos, R.B.F. (2004) Colonization and extinction of ant communities in a fragmented landscape. Austral Ecology 29: 391-398. Sanders, H. (1968) Marine benthic diversity: a comparative study. American Naturalist 102: 243-282. Schumaker, N.H. (1996) Using landscape indices to predict habitat connectivity. Ecology 77: 1210-1225. Silva, R.R., Machado Feitosa, R.S. and Eberhardt, F. (2007) Reduced ant diversity along a habitat regeneration gradient in the southern Brazilian Atlantic Forest. Forest Ecology and Management 240: 61-69. Soberon, J. and Llorente, J. (1993) The use of species accumulation functions for the prediction of species richness. Conservation Biology 7: 480-488. StatSoft, Inc. (2001). STATISTICA (data analysis software system), version 6. www.statsoft.com. Terayama, M. and Murata, K. (1990) Effects of area and fragmentation of forests for nature conservation: analysis by ant communities. Bulletin of the Biogeographical Society of Japan 45: 11-18. Theunis, L., Gilbert, M., Roisin, Y. and Leponce, M. (2005) Spatial structure of litter-dwelling ant distribution in a subtropical dry forest. Insectes Sociaux 52: 366-377. Turner, I.M. (1996) Species loss in fragments of tropical forest: a review of the evidence. Journal of Applied Ecology 33: 200-209. Vasconcelos, H.L. (1988). Distribution of Atta (Hymenoptera, Formicidae) in “terra-firme” rain forest of Central Amazonia: density, species composition and preliminary results on effects of forest fragmentation. Acta Amazônica 18: 309-315. Vasconcelos, H.L. (1999). Effects of forest disturbance on the structure of ground-foraging ant communities in central Amazonia. Biodiversity and Conservation 8: 409-420. Vasconcelos, H.L., Carvalho, K.S. and Delabie, J.H.C. (1998). Landscapes modifications and ant communities. - In the Ecology and Conservation of a Fragmented Forest: lessons from 100 Chapitre 4: Ants in old forest fragments

Amazonia. Bierregaard Jr, R.O., Gascon, C., Lovejoy, T.E. and Santos, A.A. (eds). Yale University Press. Vasconcelos, H.L. and Delabie, J.H.C. (1998) A study of forest fragmentation near Manaus, Brazil. – In Measuring and monitoring biodiversity: standard methods for ground-dwelling ants. Agosti, D., Majer, J., Alonso, Schultz, T. and L. Tennant (Eds). Washington, DC: Smithsonian Institution Press. Vasconcelos, H.L., Vilhena, J.M.S. and Caliri, G.J.A. (2000). Responses of ants to selective logging of a central Amazonian forest. Journal of Applied Ecology 37: 508-514. Vasconcelos, H.L., Vilhena, J.M.S., Magnusson, W.E. and Albernaz, A.L.K.M. (2006) Long-term effects of forest fragmentation on Amazonian ant communities. Journal of Biogeography 33: 13448-1356. Williams-Linera, G. (1990). Vegetation structure and environmental conditions of forest edges in Panama. Journal of Ecology 78: 356-373. Wilson, E.O. (2002) The future of life. Alfred A. Knopf, New York. Wilson, E.O. (2003) Pheidole in the New World: A dominant, hyperdiverse ant genus. Cambridge, MA, Harvard University Press, 794 p.

101 Chapitre 4: Ants in old forest fragments

TABLE AND FIGURE LEGENDS

Table 4-1: Fragmentation indices, edge effects, species introgression, sampling procedure and diversity in each forest fragment. Log P= perimeter (log-transformed), SI= Shape Index, PI=

Proximity index, NF= nearest fragment, MNFD= Mean nearest fragment distance, S = species obs richness, I= number of species occurrences, SSR= Standardized species richness (for I= 232), SD= Species density (Number of species/m2).

Table 4-2: Effect of fragment size, shape and isolation on ant species richness and density. Pearson’s correlation coefficients and degree of significance p (ns= non significant, *< 0.05, **<0,01 and ***< 0.001) between the mean species density (number of species/m²), standardized species richness (for 232 occurrences) and the fragmentation indices of each forest fragment.

Table 4-3: Habitat preferences of frequent species (i.e. species present in >10% of the samples in the considered habitat). Ants were collected in large forest fragments (L1 and L2) by both collection methods (Winkler and pitfall traps) and in the grassland only by pitfall traps. In the forest fragment “L1”, samples in bromeliads were discarded to avoid any confounding effect of bromeliads. A significantly higher occurrence of ant species in the forest or in the grassland, in the centre or at the edge is indicated by a cross in the corresponded column (χ²-tests). A dot indicates the presence of the species in the habitat.

Figure 4-1: Map of the Rio Pilcomayo National Park, Argentina, showing the location of the eleven forest fragments where transects were performed. Fragments belonged to three size categories: small (S, < 4ha, n=5), medium (M, between 15 and 25ha, n=4) and large (L= 250ha, n=2). Fragments (in grey) chosen in our study are situated in three different localities (Esteros, Fonzo and la Angela) and are surrounded by the grassland (in white).

Figure 4-2: Comparison of standardized species richness in small (S, n=5), medium (M, n=12) and large (L, n=2) fragments. Species richness was measured with A.L.L. Winkler transects in bromeliad zones and was standardized for 232 occurrences by Melo’s method (Melo et al. 2003). A Mann-Whitney test showed that the standardized species richness was significantly different (p<0,01) between the small (S) and the medium forest fragments (M).

Figure 4-3: Comparison of ant faunal similarity (standardized Bray-Curtis index of similarity) between forest fragments of different size (S= small, M= medium and L= large, see table 4-1 for 102 Chapitre 4: Ants in old forest fragments details about fragments) by a Non-metric multidimensional scaling NMDS. M* represents the intra-fragment variability of this measure. An UPGMA analysis reveals three main groups of species (I, II and III).

Figure 4-4: Species density from the surrounding grassland towards the centre of large forest fragments. (A) Average species density along the 3 parallel transects in L1 (n= 115 quadrats devoid of Bromeliaceae, Winkler extraction); (B) number of species collected by pitfall traps along L1B; (C) number of species collected by pitfall traps along the transect performed in L2.

103 Chapitre 4: Ants in old forest fragments

Table 4-1

1. Study of fragmentation effects Fragmentation indices Sampling procedure Diversity results Number of Transects* Distribution Area Log P log P/ PI NF MNFD Collection samples SD (mean Fragment SI samples per of S I SSR (ha) (m) log A (m-1) (m) (m) method Grass obs ±SD) transect Forest bromeliads -land S1 1.6 2.85 1.56 0.676 1.45 20 218 1*20 20 20 58 28.6 2.9 (± 1.8) S2 1.1 2.80 1.66 0.689 0.79 100 330 1*20 20 23 78 34.2 3.25 (± 1.5) S3 3.5 3.06 1.74 0.674 0.02 30 353 1*20 20 35 125 39.6 6.25 (± 2.9) S4 2.2 2.90 1.52 0.669 0.22 20 297 1*20 20 40 157 43.1 7.85( ± 3.3) S5 3.0 2.95 1.44 0.659 0.03 580 610 1*20 20 continuous 25 71 39.0 3.6 (± 3.2) M1 20.5 3.48 1.88 0.655 0.71 20 287 1*20 20 28 82 37.0 4.1 (± 2.9) M2 27.9 3.54 1.85 0.650 0.06 60 180 1*20 20 34 143 40.0 7.2 (± 3.0) M3 16.1 3.56 2.55 0.684 11.93 20 152 1*20 Winkler 20 0 46 159 53.0 8.0 (± 3.6) M4 28.8 3.65 2.35 0.668 0.05 40 356 1*20 20 34 90 50.2 4.5 (± 2.4)

L1 256.2 4.09 2.19 0.639 13.92 10 199 3*50 30 * patchy 44 178 48.6 5.9 (± 3.2)

L2 254.4 4.04 1.94 0.631 0.09 120 321 1*50 30 * continuous 44 232 44.0 7.7 (± 3.7)

40 146 M* 16.1 3.56 2.55 0.684 11.93 20 152 8*20 160 continuous 46.6 (± 2.2) 7.3 (± 3.7) 1-8 (±1.2) (±11.2) 2. Study of edge effects L1 256.2 4.09 2.19 0.639 13.92 10 199 3*50 114 ** 0 patchy 55 418 - 3.6 (± 2.5) Winkler L2 254.4 4.04 1.94 0.631 0.09 120 321 1*50 50 0 continuous 44 232 - 7.7 (± 3.7) 3. Study of species introgression L1B 256.2 4.09 2.19 0.639 13.92 10 199 1*55 Pitfalls 50 5 patchy 51 261 - 5 (0 - 11) L2 254.4 4.04 1.94 0.631 0.09 120 321 1*55 Pitfalls 50 5 continuous 47 182 - 3 (0 - 9) * samples between 200 and 500 meters from the fragment edge and only in bromeliads ** samples in bromeliads were discarded to avoid confounding effects of bromeliads density (Theunis et al. 2005)

104 Chapitre 4: Ants in old forest fragments

Table 4-2

Fragmentation Index Standardized species Species p p richness density 1 Area (log) 0.60 * 0.45 ns 2 Perimeter (log) 0.66 * 0.46 ns 3 Shape Index 0.77 *** 0.57 ** 4 log P/ log A -0.21 ns -0.31 ns 5 Proximity Index 0.56 0.07 0.27 ns 6 Nearest Fragment -0.15 ns -0.31 ns 7 Mean Nearest Fragment Distance -0.19 ns -0.41 ns

105 Chapitre 4: Ants in old forest fragments

Table 4-3

Forest Grassland Species preference Centre Edge (-50 to 0m) (200 to 500m) (0 to 200m) Edge species Crematogaster sp.18 X • Paratrechina pubens X • Pheidole gr. aberrans sp.01 X • Grassland species Hypoponera sp.08 X Paratrechina sp.01 X Solenopsis sp.14 X Solenopsis sp.16 X Forest species Brachymyrmex sp.05 X • Paratrechina sp.02 X • Pyramica denticulata X • Acromyrmex hispidus fallax X X Carebarella bicolor X X Crematogaster sp.02 X X Crematogaster sp.11 X X Crematogaster sp.17 X X Gnamptogenys striatula X X Hypoponera sp. prox. opaciceps 1 X X Octostruma rugifera X X Pachycondyla harpax X X Pheidole flavens X X Pheidole nubila X X Rogeria scobinata X X Solenopsis sp.01 X X Solenopsis sp.02 X X Solenopsis Sp.17 X X Solenopsis Sp.18 X X Ubiquist species Camponotus crassus • • • Labidus praedator • • • Pheidole sp.21 • • • Wasmannia auropunctata • • •

106 Chapitre 4: Ants in old forest fragments

Figure 4-1

107 Chapitre 4: Ants in old forest fragments

Figure 4-2

108 Chapitre 4: Ants in old forest fragments

Figure 4-3

109 Chapitre 4: Ants in old forest fragments

Figure 4-4

110 Chapitre 5: Impact of fire on grassland ants

IMPACT OF FIRE ON THE ANT ASSEMBLAGE STRUCTURE IN

GRASSLANDS OF THE HUMID CHACO

LAURENCE THEUNIS 1, 2, YVES ROISIN 2 AND MAURICE LEPONCE 1.

LAURENCE THEUNIS

Phone +32 2 627.43.64 Fax +32 2 649.48.25 E-mail: [email protected]

Running head: Impact of fire on grassland ants.

111 Chapitre 5: Impact of fire on grassland ants

ABSTRACT

In the humid Chaco regular fires and floodings occur that affect the grassland but not the forest islets spread among it. The aim of the present study was to investigate the effects of fire on the grassland ant assemblage structure and composition and to investigate the role of forest islets as potential species reservoir. Ground-foraging ants were collected by pitfall traps placed at 10m intervals along transects. Transects, 200m long, were sampled in the grassland before (n=2) and 5 to 15 days after a fire (n=3). Ground foraging ants living in forests were sampled along two 500m long transects. Recent fire did neither affect the species richness nor the ant activity but modified the species numerical dominance. Epigeaic and omnivore ants were more present in the unburned grassland. Predators were more active in the burned grassland. Seventy seven percent of the species collected in the grassland were also found in the forest but with a different relative frequency. Our results suggest that in the humid chacoan savanna a fire modify the ant relative numerical dominance but little affect the ant composition and overall diversity and that forest islets may act as a species reservoir.

112 Chapitre 5: Impact of fire on grassland ants

INTRODUCTION

Fire is one of the most common natural disturbances occurring in grasslands and has been shown to affect differently arthropod diversity (Abbott 1984, McCullogh et al. 1998, Folgarait 1998, Hoffman 2003, Parr et al. 2004, Sackmann and Farji-Brener 2006). Species assemblages modifications after a fire depend on the characteristics of the fire (frequency), the assemblage structure (competitive hierarchies), and the habitat structure (Andersen 1991, Orgeas and Andersen 2001, Farji-Brener et al. 2002, Moretti et al. 2002, Hoffman 2003, Parr et al. 2004, Sackmann and Farji-Brener 2006, Andersen et al. 2006, Santos et al. 2008, Vasconcelos et al. 2008). Ants are one of the most abundant organisms in grassland and play key roles in ecosystem processes (nutrient cycling, seed dispersal, regulation of population insects). Furthermore, ant assemblages are structured through competitive dominance (Hölldobler and Wilson 1990, Folgarait 1998, Bestelmeyer and Wiens 2003). For these reasons, ants are an ideal faunal group to study the effects of fire disturbance on insect assemblage in savanna and evaluate the species reservoir, which constitute forest fragments. Fire can have direct effect on ants by destroying their nests (Parr et al. 2004). However, the effect is usually indirect through the modifications of vegetation structure induced by fire (Hoffmann 2003, Parr et al. 2004, Sackmann and Farji-Brener 2006, Andersen et al. 2006, Santos et al. 2008, Vasconcelos et al. 2008). Therefore, the effect of fire on the ant assemblage may depend on the magnitude of changes and the speed of regeneration of the vegetation structure (Farji-Brener et al. 2002, Parr et al. 2004). Changes in vegetation cover due to fire affect often ant species composition and sometimes species richness. Ant species hierarchies may be modified by a reduction of rare species associated to woody vegetation and by an increase of dominant species that tolerate hot temperature conditions (Farji-Brener et al. 2002). The Argentinian humid Chaco is a large area (480 km²) where grasslands are regularly submitted to natural floods and fires and scattered with fire-resistant forest islets growing on slight mounds (Morello 1970). Ant savanna species may display a considerable resistance to burning because of their long association with the fire regime (Andersen and Müller 2000, Orgeas and Andersen 2001, Parr et al. 2004). Farji-Brener et al. (2002) showed that fire-related changes for ants were small and short-term because of the rapid regeneration of the herbaceous vegetation of the savanna. Fire possibly can play a beneficial role in ecosystem biodiversity allowing to species,

113 Chapitre 5: Impact of fire on grassland ants

normally excluded by dominant species, to access to ressources (Panzer 1988, Connell 1978). The aim of the present study was to investigate the effects of fire on the grassland ant assemblage structure and composition and to investigate the role of forest islets as potential species reservoir

114 Chapitre 5: Impact of fire on grassland ants

MATERIAL AND METHODS

Study Area

We carried out the study in the Río Pilcomayo National Park situated in the humid Chaco of the northeastern Argentina (25°04’06’’ S, 58°05’36’’ W). Average annual rainfall in the park is about 1200mm, with a short dry period (0-3 months) in the southern hemisphere winter, between June and September. Temperature fluctuates broadly, with an annual average of 22-24°C and occasional winter frost (Pujalte et al. 1995). The grassland (“pastizal”) was dominated by herbaceous vegetation as Setaria sp., Luziola peruviana and palm trees as Copernicia alba (Pujalte et al. 1995, Mereles 2005). The fire-resistant forest islets (“monte fuerte”) were dominated by Schinopsis balansae Engl., Astronium balansae Engl. and Aspidosperma quebracho-blanco Schlecht. and by a ground strata of bromeliads (Aechmea distichantha Lemaire and Pseudananas sagenarius (Arruda) Camargo) (Ramella and Spichiger 1989, Pujalte et al. 1995).

Sampling procedure

The sampling was carried out at the beginning of the wet season, in Sept-Oct. 2001 and 2002. Ground-dwelling ants were captured with pitfall traps (300ml drinking cups filled with water) left during 6 days and distant of at least 10m to decrease the risks of spatial autocorrelation (see Theunis et al 2005). One hundred traps were placed in the grassland before (n=60) or 5-15 days after a fire (n=40). One hundred traps were simultaneously placed in adjacent fire-resistant forest islets (2 transects 500m long).

Data analyses

Species richness, number of species by pitfall and species composition were estimated to compare ant assemblages between the two habitats. Because of its strong dependence on sample size and species density, species richness must be standardized before comparisons are conducted. Species richness difference between assemblages (using the same number of samples) was thus calculated by bootstrapping using the PAST software (Hammer et al. 2004).

115 Chapitre 5: Impact of fire on grassland ants

Faunal similarity of ant assemblages between habitats was expressed by the NNESS index (Grassle and Smith 1976), which is given by:

NNESSij / k =

1 - {ESSij / k / [(ESSii / k + ESSjj / k)/2]} where ESSij / k is the expected number of species shared for random draws (without replacement) of k occurrences from localities i and j. When k is small, the index is highly sensitive to the occurrences of the most frequent species. When k increases, the influence of rarer species is emphasized. If result patterns change when k varies, it could mean that different processes structure the diversity of common and rarer species. Similarity of ant assemblages was calculated for k=1 (identical to the Morisita- Horn index), k=128 and k=256 (only for comparison between the forest and the grassland). Values range from 0, if species collected in localities i and j are totally different to 1, if localities host the same species. The software program BiodivR 1.0 (Hardy 2005) was used to compute the NNESS index. Frequent species were defined as species present in at least 10% of the samples and were classified in functional groups (Andersen 1995, Delabie et al. 2000).

116 Chapitre 5: Impact of fire on grassland ants

RESULTS

Post-fire ant activity

The number of species collected by individual pitfall traps, representative of the ant activity, was not different between the grassland burned (n=60) or unburned (n=40) (median [min-max]: 5.4 [0 – 13] vs 4.5 [0 – 9] respectively, Mann-Whitney U= 1058.5, p= 0.3208). The rarefied species richness was similar in the burned (S= 37 species for 180 occurrences) and unburned (S= 38) grassland (Fig.5-1). Dominance in unburned sites was composed of some highly frequent species. In contrast, in burned sites, the dominance was represented by a greater number of species with moderate frequencies (Fig. 5-2). Seven species were frequent in the burned grassland and totally absent from the unburned sites (Pheidole nubila, Labidus preadator, Brachymyrmex sp.05, Linepithema gr. humile sp. 02, Odontomachus meinerti, Pogonomyrmex sp.01 and Solenopsis sp.19) (Fig. 5-2). Six species were significantly more frequent in the grassland burned (Pheidole flavens, Ectatomma edentatum, E. Permagnum, Camponotus crassus, Paratrechina pubens and Solenopsis sp.10) (paired t-test, p<0.05) (Fig. 5-2). Five others species were significantly (more frequent in the unburned grassland (Wasmannia sp.01, Pheidole gr. aberrans sp.01, Camponotus sp.15, Crematogaster sp.18 and Solenopsis sp.14) paired t-test, p<0.05) (Fig.5-2). Twelve species were frequent (i.e. found in at least in 10% of samples) in both habitats (Wasmannia sp.01, Pheidole flavens, Ectatomma edentatum, E. permagnum, Camponotus crassus, C. sp.14, C. sp.15, Paratrechina pubens, Pheidole gr. aberrans sp.01, Brachymyrmex sp.07, Solenopsis sp.14 and Solenopsis sp.16). The faunal similarity (NNESS index) between the ant assemblage found in burned and unburned grassland was stable for an increasing k value (k=1, NNESS=0,626; k=64, NNESS= 0,678 and for k= 128, NNESS= 0,632) indicating that fire evenly affected dominant and rare species. Considering the ant functional group, we noted that the epigeaic and omnivores ants foraging at the ground level, such as Crematogaster sp.18, Wasmannia sp.01, Pheidole gr. aberrans sp.01, Camponotus sp.15 and Solenopsis sp.14 were preferentially collected in the unburned grassland. Predators species, such as Labidus preadator, Pachycondyla striata, Odontomachus meinerti, Ectatomma edentatum and E. permagnum, were more frequent in the burned than unburned grassland.

117 Chapitre 5: Impact of fire on grassland ants

Habitat specificity of grassland ants

The rarefied species richness was close between the grassland (S=54 ant species for 413 occurrences) and the adjacent fire-resistant forest (S=59) (Fig.5-1). The number of species collected by individual pitfall traps was larger in the grassland than in the forest (median [min-max]: 5.0 [1 – 13] (n=100) vs 4.0 [0 – 11] (n=100) respectively, Mann- Whitney U= 3486, p< 0.01). The faunal similarity (NNESS index) between the ant assemblage found in the forest and the grassland increased with an increasing k value (k=1, NNESS=0,429; k=128, NNESS= 0,547 and for k= 256, NNESS= 0,537) indicating that some dominant species are very different between the forest and the grassland. The frequencies of species frequent in both habitats are shown in Figure 5-2. Five species were frequent in both habitats (Wasmannia sp. 01, Pheidole flavens, Ph. nubila, Ectatomma edentatum and Solenopsis sp.10). Seven species were frequent in the grassland and scarce in the forest (Ectatomma permagnum, Camponotus crassus, Paratrechina pubens, Brachymyrmex sp.05, Camponotus sp.14, Linepithema gr. humile sp.02, Pheidole gr. aberrans sp.01). Five other species were frequent in the grassland (burned and unburned pooled) and totally absent from the forest fragments (Brachymyrmex sp.07, Camponotus sp.15, Crematogaster sp.18, Pogonomyrmex sp.01 and Solenopsis sp.16). Finally, five species were frequent in the forest and rarer in the grassland (Pachycondyla striata, Solenopsis sp.01, S. sp.18 Acromyrmex hispidus fallax and Gnamptogenys striatula). The fungus-grower Acromyrmex hispidus fallax or litter predators such as Pachycondyla striata and Gnamptogenys striatula were more frequent in the forest habitat. In the grassland, we collected lots of omnivore species foraging on the vegetation or at the ground level such as Brachymyrmex sp.05, B. sp.07, Camponotus crassus, C. sp.14, C. sp.15, Crematogaster sp.18 Linepithema gr. humile sp.02 and Paratrechina pubens.

118 Chapitre 5: Impact of fire on grassland ants

DISCUSSION

Effect of a recent fire on the grassland ant assemblage

A recent fire did not modify the ant species richness nor the ant activity but affected species dominance in the grassland. Most ant colonies nesting in the soil are able to withstand a rapid fire occurring in a dry grassland (Schoereder et al. 2004). The herbaceous vegetation in the grassland recovers quickly after a fire and supply food for herbivores (Farji-Brener et al. 2002, Hoffmann 2003, Parr et al. 2004). Farji-Brener et al. (2002) observed also low effects of fire on ant species richness in the northeastern Patagonian steppes. In savannas ant assemblages often display considerable resistance and resilience to burning because of their long association with the fire regime (Andersen and Müller 2000, Orgeas and Andersen 2001, Parr et al. 2004). Nevertheless, we cannot exclude the possibility that some ant species, foraging in the grassland, survive to the fire because their nest is located in palm-trees. Fire affected the ant assemblage composition principally by changing dominance relationships as it was reported in several other studies (Andersen 1991, MacKay et al. 1991, Andrew et al. 2000, York 2000, Farji-Brener et al. 2002). Indeed, the ground- foraging ant assemblage evenness is higher after a fire (i.e. there are more frequent species (19 species) with moderate (10-40%) frequencies). Particularly, the five most frequent species (present in more than 30% of samples) in the unburned grassland (Wasmannia sp.01, Pheidole gr. aberrans sp.01, Camponotus sp.15, Crematogaster sp.18 and Solenopsis sp.14) were affected by fire. Such a decrease in population of dominants after a fire, allows others species (sometimes pioneers) to recolonize the habitat provisionally (Connell 1978, Panzer 1988, Farji-Brener et al. 2002). This decrease would attenuate the competitive pressure of congeneric species for the same resources. In the dry Chaco, Delsinne et al. (2007) observed few interspecific conflicts. Avoidance or indifference between ant species seemed to be a frequent behavior in (semi-) arid region preventing an high energetic cost in fighting (Fellers 1987, Yanoviak and Kaspari 2000, Delsinne et al. 2007). For example, the higher activity of Solenopsis sp. 19 might be related to the decrease of S. sp.14 and S. sp.16; the increase of Pheidole flavens, P. nubila and P. sp. 21 to the decrease of Pheidole gr. aberrans sp.01 and the increase of Camponotus crassus to the decrease of Camponotus sp.15. Moreover, some species, pioneers, emerged only after a fire (Connell 1978, Panzer 1998, Farji-Brener et al. 2002). These pioneers species nest probably in the soil as

119 Chapitre 5: Impact of fire on grassland ants

Odontomachus meinerti (Obs. by Jochen Bihn), Pogonomyrmex sp. 01 (Tschinkle 2004), Labidus preadator (Obs. by John Longino), Pheidole nubila (Wilson 2003) and Solenopsis sp.10 (pers. obs.). Betch and Cancela de Fonseca (1995) have shown that nutrients are often released after burning (young grass) and attract arthropods preyed by ants. A greater production of green biomass is observed in burned savanna (Sarmiento 1972). It is possible that predator species, such as Odontomachus meinerti, Pachycondyla striata, Gnamptogenys striatula, Labidus praedator, Ectatomma edentatum and E. permagnum, might be more active after a fire because of an increased abundance of prey in a less complex environment (Castaño-Meneses and Palacios-Vargas 2003, Kaspari 2001). Reduction of vegetation cover is often followed by an increased dominance of stress- tolerant species (MacKay et al. 1991, Farji-Brener et al. 2002). More frequent species could coexist and share resources of the grassland probably through mechanisms limiting temporal overlaps of species activity rhythms (Bestelmeyer 2000, Andersen 1991, 1995, Delsinne et al. 2007). In the dry Chaco, it appears that the coexistence of numerous frequent species is allowed by the coexistence of two guilds exploiting food resources with a spatio-temporal lag: competitive non-thermophilic species (as Pheidole spp., Brachymyrmex spp., Wasmannia auropunctata) and subordinate thermophilic ones (Bestelmeyer 2000, Delsinne et al. 2007). Here, no thermophilic species were found. Nonetheless, given that the ground-dwelling ant species of the humid Chaco live in xeric climate, they may be better pre-adapted to the structure of the burned grassland (Kusnezov, 1953). Species may have physiological (heat tolerant species, deep nesting) or behavioural adaptations (shift of activity rhythm) to live and/or to forage in the grassland.

Ground-foraging ant assemblage structure in the grassland and in the forest

Most of the ground-foraging ants are ubiquist nesting and foraging in very contrasted habitat as the forest and the savanna. Although, 77% of species were common to both habitat, their relative frequency varies related to the habitat type. Nevertheless, 7 species were exclusively collected in the grassland (Solenopsis sp.14 was collected in forest during another sampling). Genera as Camponotus, Brachymyrmex, Pogonomyrmex and Linepithema were well represented in the grassland. In the literature, species of Pogonomyrmex, Camponotus and Linepithema genera are often reported as good colonizers (Holway 1998, Ingram 2002). Pogonomyrmex are

120 Chapitre 5: Impact of fire on grassland ants

harvester ants feeding on graminaceous seeds (MacMahon et al. 2000). Camponotus sp. 15 dominated numerically the grassland with a frequency of catches of 54% of the pitfall traps. Although the genus Camponotus is called carpenter ants, some species do not nest inside wood but rather build nest on the ground in forests or savannas (Mody and Linsenmair 2003, Yamamoto and Del-Claro 2008). Less than 9% of the total ant occurrences in forest was collected by pitfalls. Ground foraging ant species richness in forest islets represented a relatively low part of the entire ground-dwelling ant assemblage of the forest (see annexe 2). Pitfall traps allowed to capture mainly species foraging at the ground level but missed a lot of small and cryptic leaf-litter ant species (e.g. Solenopsis sp. 01 and S. sp 18) collected by other methods such as the Winkler (Leponce et al. 2004, Theunis et al. 2005). Predator ant such as Pachycondyla striata and Gnamptogenys striatula were possibly favoured by a high biomass of their prey in the forest. Acromyrmex hispidus fallax was also typical of the forest habitat where they find the leaves necessary to the growth of their fungus.

In conclusion, our results suggest that in the humid chacoan savanna a fire modify the ant relative numerical dominance but little affect the ant composition and overall diversity and that forest islets may act as a species reservoir.

121 Chapitre 5: Impact of fire on grassland ants

Acknowledgments – We thank the Administración de Parques Nacionales, Buenos Aires, Argentina, for allowing us to collect in P.N. Río Pilcomayo. Nestor Sucunza, the guardaparques and Cornelio Pares greatly facilitated our work in the park. Thanks to G.J. Torales and E.R. Laffont, Univ. Nacional del Nordeste, for logistic support and to Yves Laurent for field assistance. This work was supported by the ‘Fonds National pour la Recherche Scientifique’ (FNRS, Belgium) to LT (PhD Grant). A grant to LT from the ‘Fonds Léopold III pour l’Exploration et la Conservation de la Nature’ allowed a field campaign in Argentina. We would like to thank also Prof. J. H. C. Delabie and Dr. I. C. do Nascimiento (CEPEC, Brasil) for help in ant identification. Isabelle Bachy for her GIS software support and Julien Cillis for the SEM images of ant species. The manuscript benefited from comments by Dr. T. Delsinne (RBINS), Dr.G. Kapfer (RBINS), and Dr. A-C. Mailleux (ULB).

122 Chapitre 5: Impact of fire on grassland ants

REFERENCES

Abbott, I. (1984) Changes in the abundance and activity of certain soil and litter fauna of the Jarrah forest of Western Australia after moderate intensity fire. Australian Journal of Soil Research 22: 463-470.

Andersen, A.N. (1991) Responses of ground-foraging ant communities to three experimental fire regimes in a savanna forest of tropical Australia. Biotropica 23: 575-585.

Andersen, A.N. (1995) A classification of Australian ant communities, based on functional groups which parallel plant like-forms in relation to stress and disturbance. Journal of biogeography 22: 15-29.

Andersen, A.N. and Müller, W.J. (2000) Arthropod responses to experimental fire regimes in an Australian tropical savannah: ordinal level analyses. Austral Ecology 25: 199–209.

Andersen, A.N., Hertog, T. and Woinarski, J.C.Z. ( 2006) Long-term fire exclusion and ant community structure in an Australian tropical savanna: congruence with vegetation succession. Journal of Biogeography 33: 823-832.

Andrew, N., Rodgerson, L. and York, A. (2000) Frequent fuel-reduction burning: the role of logs and associated leaf litter in the conservation of ant biodiversity. Australian Ecology 25: 99-107. Bestelmeyer, B. (2000) The trade-off between thermal tolerance and behavioral dominance in a subtropical South American ant community. Journal of Animal Ecology 69: 998-1009. Bestelmeyer, B., Wiens J. (2003) Scavenging ant foraging behavior and variation in the scale of nutrient redistribution in semiarid grasslands. Journal of Arid Environments 53: 373-386.

Betch, J.M. and Cancela da Fonseca, P. (1995) Changes in edaphic factors and microarthropod communities after clearing and burning in a tropical rain forest in French Guyana. Acta Zoologica Fennica 196: 142-145.

Castaño-Meneses, G. and Palacios-Vargas, J.G. (2003) Effects of fire and agricultural practices on neotropical ant communities. Biodiversity and Conservation 12: 1913- 1919.

123 Chapitre 5: Impact of fire on grassland ants

Colwell, R. K. (2004) EstimateS: Statistical estimation of species richness and shared species from samples. Version 7.5. Persistent URL: . Connell, J.H. (1978) Diversity in tropical rain forest and coral reefs. Sciences 199: 1302-1310. Delabie, J.H.C., Agosti, D. and do Nascimento, I.C. (2000). Litter ant communities of the Brazilian Atlantic rain forest region. - In: Agosti, D., Majer, J., Alonso, L.E. and Schultz, T. (Eds). Sampling Ground-dwelling Ants: Case Studies from the Worlds' Rain Forests. Curtin University, Australia, School of Environmental Biology, Bulletin no 18, pp.1-17. Delsinne, T., Roisin, Y. and Leponce, M. (2007) Spatial and temporal foraging overlaps in a Chacoan ground-foraging ant assemblage. Journal of arid environments 71: 29- 44. Farji-Brener, A.G., Corley, J.C. and Bettinelli, J. (2002) The effects of fire on ant communities in north-western Patagonia: the importance of habitat structure and regional context. Diversity and Distribution 8: 235-243. Fellers, J.H. (1987) Interference and exploitation in a guild of woodland ants. Ecology 68: 1466–1478. Folgarait, P. (1998) Ant biodiversity and its relationship to ecosystem functioning: review. Biodiversity and Conservation 7: 1221-1244. Grassle, J. F. and Smith, W. (1976) A similarity measure sensitive to the contribution of rare species and its use in investigation of variation in marine benthic communities. Oecologia 25: 13-22. Hammer, Ø., Harper, D.A.T., and Paul, D.R. (2001) Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica 4: 1- 9. http://folk.uio.no/ohammer/past/>.

Hardy, O. (2005) BiodivR 1.0. A program to compute statistically unbiased indices of species diversity within samples and species similarity between samples using rarefaction principles. Université Libre de Bruxelles. Persistent URL: Hayek, L. A. and Buzas, M. A. (1997) Surveying natural populations. Columbia University Press. Hoffmann, B.D. (2003) Responses of ant communities to experimental fire regimes on rangelands in the Victoria River District of the Northern Territory. Austral Ecology 28: 182–195.

124 Chapitre 5: Impact of fire on grassland ants

Hölldobler, B. and Wilson, E.O. (1990) The ants. Cambridge University Press, Cambridge. Holway, D.A. (1998) Effects of Argentine ant invasions on ground-dwelling arthropods in northern California riparian woodlands. Oecologia 116: 252–258. Ingram, K.K. (2002) Flexibility in nest density and social structure in invasive populations of the Argentine ant, Linepithema humile. Oecologia 133: 492-500. Kaspari, M. (2001) Taxonomic level, trophic biology and the regulation of local abundance. Global Ecology and Biogeography 10: 229-244. Kusnezov, N. (1953) Las hormigas de los Parque Nacionales de la Patagonia y los problemas relacionados.

Leponce, M., Theunis, L., Delabie, J.H.C. and Roisin, Y. (2004) Scale dependence of diversity measures in a leaf-litter ant assemblage. Ecography 27: 253-267.

MacKay, W., Rebeles, A., Arredondo, H., Rodriguez, A., González, D. and Vinson, B. (1991) Impact of slashing and burning of a tropical rain forest on the native ant fauna (Hymenoptera: Formicidae). Sociobiology 18: 257-268.

MacMahon, J.A., Mull, J.F., Crist, T.O. (2000) Harvester ants (Pogonomyrmex SPP.): their community and ecosystem influences. Annual Review of Ecology and Systematics 31: 265-291. McCullogh, D.G., Werner, R.A. and Neumann, D. (1998) Fire and insect in northern and boreal forest ecosystems of North America. Annual Review of Entomology 43: 107-127.

Mereles, M.F. (2005) Una aproximación al conocimiento de las formaciones vegetales del Chaco boreal, Paraguay. Rojasiana 6: 5-48. Mody, K. and Linsenmair, K.E. (2003) Finding its place in a competitive ant community: leaf fidelity of Camponotus sericeus. Insectes Sociaux 50: 191–198. Morello, J. (1970) Ecología del Chaco. Boletín de la Sociedad Argentina de Botánica Vol XI (supl.): 161-174. Orgeas, J. and Andersen, A.N. (2001) Fire and biodiversity: responses of grass-layer beetles to experimental fire regimes in an Australian tropical savanna. Journal of Applied Ecology 38: 49–62. Parr, C.L., Robertson, H.G., Biggs, H.C. and Chown, S.L. (2004) response of African savanna ants to long-term fire regimes. Journal of Applied Ecology 41: 630-642. Panzer, R. (1988) Management of prairie-remnants for insect conservation. Natural Areas Journal 8: 83-90.

125 Chapitre 5: Impact of fire on grassland ants

Pujalte, J.C., A.R., Reca, A., Balabusic, P., Canevari, Cusato, L. and Fleming, V.P. (1995) Anales de parques nacionales. Unidades Ecológicas del parque nacional Rio Pilcomayo. Administración de Parques Nacionales XVI: 1-185. Ramella, L. and Spichiger, R. (1989) Interpretación preliminary del medio físico y de la vegetación del Chaco Boreal – Contribución al studio de la flora y de la vegetación del Chaco. I. Candollea 44 : 639-680. Sackmann, P. and Farji-Brener, A. (2006) Effect of fire on ground beetles and ant assemblages along an environmental gradient in PW Patagonia: Does habitat type matter? Ecosciences 13: 360-371. Sarmiento, G. (1972) Ecological and floristic convergences between seasonal plant formations of tropical and subtropical South America. The Journal of Ecology 60: 367-410. Shoereder, J.H., Sobrinho, T.G., Ribas C.R. and Campos R.B.F. (2004) Colonization and extinction of ant communities in a fragmented landscape. Austral Ecology 29: 391- 398. Santos, J.C., Delabie, J.H.C. and Fernandes, G.W. (2008) A 15-year post evaluation of the fire on ant community in an area of Amazonian forest. Revista Brasileira de Entomologia 52: 82-87. Theunis, L., Gilbert, M., Roisin, Y., Leponce, M. (2005) Spatial structure of litter- dwelling ant distribution in a subtropical dry forest. Insectes Sociaux 52: 366-377. Tschinkel, W.R. (2004) Nest architecture of the Florida harvester ant, Pogonomyrmex badius. Journal of Insect Science 4, Article number 21. Vasconcelos, H.L., Leite, M.F., Vilhena, J.M.S., Lima, A.P. and Mgnusson, W.E. (2008) Ant diversity in an Amazonian savanna: relationship with vegetation structure, disturbance by fire, and dominant ants. Austral Ecology 33: 221-231. Wilson, E.O. (2003) Pheidole in the New Worlds: A dominant, hyperdiverse ant genus. Harvard University Press. pp. 818. Yamamoto, M. and Del-Claro, K. (2008) Natural history and foraging behavior of the carpenter ant Camponotus sericeiventris Guérin, 1838 (Formicinae, Camponotini) in the Brazilian tropical savanna. Acta Ethologica. Online published: http://www.springerlink.com/content/b452682r427107q7/fulltext.pdf . Yanoviak, S.P. and Kaspari, M. (2000) Community structure and the habitat templet: ants in the tropical forest canopy and litter. Oikos 89: 259–266. York, A. (2000) Long-term effects of frequent low-intensity burning on ant communities in coastal blackbutt forests of southeastern Australia. Austral Ecology 25: 83-98.

126 Chapitre 5: Impact of fire on grassland ants

FIGURE LEGENDS

Figure 5-1: Occurrence-based rarefaction curves [Coleman method of EstimateS 7.5, Colwell 2004] of ant species richness in the grassland (both burned and unburned: empty squares, burned: grey stars, unburned: grey triangles) and the forest islets spread inside the savanna (black squares).

Figure 5-2: Relative frequency in samples (pitfall traps) of species found in forest islets (n= 100) and in the grassland either burned (n=60) or not (n=40). Only frequent species (i.e. present in at least 10% of the samples in at least one habitat type) are shown. Species are sorted by decreasing occurrences in the forest habitat. Stars indicate the level of significance of paired t-test of species abundance (log (x+1) transformed) between the grassland burned and unburned: * = p<0.05, ** = p<0.01 and *** = p<0.001.

Figure 5-3: Proportion of each functional group (following the classification of Delabie et al. 2000) in the savanna before and after a fire and in the forest islets spread inside this savanna.

127 Chapitre 5: Impact of fire on grassland ants

Figure 5-1

128 Chapitre 5: Impact of fire on grassland ants

Figure 5-2

129 Chapitre 5: Impact of fire on grassland ants

Figure 5-3:

130 DISCUSSION ET PERSPECTIVES

DISCUSSION GENERALE – PERSPECTIVES

1. Principaux résultats – Originalités de l’étude

Au sein de cette forêt naturellement fragmentée du Chaco humide argentin, nous avons montré que :

1- Au niveau du micro-habitat, on observe pour certaines espèces des pics périodiques d’abondance (maximum tous les 10m) correspondant vraisemblablement à l’emplacement des colonies qui s’espacent pour diminuer la compétition intraspécifique (Ch. 2). Associé aux microconvexités topographiques l’on observe également des pics de densité de broméliacées et de quantité de litière qui favorisent une grande densité d’espèces différentes de fourmis (Ch. 2 et Ch. 3). 2- À l’échelle de l’habitat, les petits fragments forestiers les plus isolés sont les moins riches en fourmis, particulièrement en espèces typiquement forestières (Ch. 4). 3- Dans les larges fragments, les espèces typiquement forestières se distribuent indépendamment de la distance les séparant du bord. Quelques espèces typiques de savane pénètrent en bordure de forêt et provoquent une plus grande variabilité de la faune récoltée au sein des quadrats de litière situés à cet endroit. Cependant, aucun pic de diversité correspondant à une zone de superposition d’espèces de bord et de centre n’a été observé au sein des fragments forestiers. (Ch. 4). 4- Le type d’habitat (savane vs forêt) influence la composition de l’assemblage (Fig 5.2) des fourmis terricoles (Ch.5). Au total, 79 espèces ont été collectées par pièges à fosse dans la savane, 130 espèces dans la forêt dont 83 récoltées par pièges à fosse et 112 par Winkler. Au total, 47% des espèces collectées sont forestières, 14% de la savane et 39% sont communes aux deux habitats. 5- Un feu récent modifie la fréquence relative des espèces les plus communes mais n’affecte pas la richesse globale du milieu et ne pénètre pas dans la forêt (Ch. 5).

131 DISCUSSION ET PERSPECTIVES

2. Facteurs structurant les assemblages de fourmis du Parc Rìo Pilcomayo

2.1. Type d’habitat Le système hydrographique propre au Chaco humide et plus particulièrement du Parc National Rìo Pilcomayo, situé à proximité du croisement des fleuves Pilcomayo et Paraguay, est à l’origine de la genèse de la mosaïque de végétations observée dans ce parc (Ramella et Spichiger 1989, Pujalte et al. 1995). De légères variations topographiques permettent de différencier des milieux avec différentes probabilités d'inondation: le Monte Fuerte (topographiquement surélevé, jamais inondé) ; la savane- palmeraie (inondée périodiquement) ; les esteros (inondés en permanence) et la forêt galerie (qui longe le fleuve Pilcomayo).

Dans la savane-palmeraie, 79 espèces de fourmis terricoles ont été collectées dont seulement 20 sont typiques de cet habitat. La savane-palmeraie a la particularité d’être soumise à des régimes réguliers de feux en plus des inondations. Un feu modifie la fréquence relative des espèces en défavorisant les 5 espèces les plus fréquentes et en favorisant d’autres espèces communes. Dans le chapitre 5, nous émettions l’hypothèse que la plupart des espèces nichent dans le sol permettant aux colonies de survivre lors d’un feu. Ryder Wilkie et al. (2007) a montré que la composition de l’assemblage de fourmis nichant dans le sol d’une forêt amazonienne, varie en fonction de la profondeur (de 0 à 50cm). La profondeur à laquelle nichent les différentes espèces de la savane pourrait expliquer pourquoi les espèces sont différemment affectées par les feux. Les espèces les plus fréquentes dans la savane nichent peut-être à proximité de la surface, les rendant plus vulnérables lors d’un feu, tandis que d’autres espèces mènent une vie plus hypogée mais profiteraient de la diminution de fréquence des espèces dominantes pour coloniser temporairement la savane. En outre, dans le Chaco humide des crues prolongées (2-3 mois), inondant la savane-palmeraie, surviennent pendant la saison des pluies (décembre à mars). Ces inondations couvrent la savane-palmeraie dont le sol, riche en argile, est très imperméable (Pujalte et al. 1995, Mereles 2005). Adis et Junk (2002) ont décrit différentes adaptations morphologiques, phénologiques, physiologiques et comportementales observées chez divers invertébrés terrestres leur permettant de survivre lors d’inondations dans les plaines d’Amazonie. Certaines

132 DISCUSSION ET PERSPECTIVES

fourmis de la savane du Chaco humide pourraient mener, pendant les périodes d’inondations, une vie strictement hypogée. Gryllenberg et Rosengren (1984) ont découvert qu’une espèce de Formica pouvait survivre pendant plusieurs semaines sous la terre inondée, en diminuant sa consommation d’oxygène à 5% à 20% du taux normal. Similairement une fourmi des régions tempérées, Cardiocondyla elegans, survit lors de crues de la Loire grâce à des poches d’air piégées dans la terre (Lenoir 2006). Dans une forêt semi-aride du sud de l’Australie soumise à des crues annuelles, Ballinger et al. (2007) démontrent que les fourmis persistent en grand nombre dans les plaines inondées. Néanmoins, certaines espèces, appartenant à des genres potentiellement arboricoles (Crematogaster, Camponotus) (Hölldobler et Wilson 1990) peuvent vivre dans la savane inondée en nidifiant dans les arbres, de manière temporaire ou permanente (Majer et Delabie 1994, Ellis et al. 2001, Adis et Junk 2002, Ballinger et al. 2007). Echantillonner les fourmis à différentes profondeur du sol et dans les arbres de la savane permettrait de mieux comprendre comment les colonies résistent aux inondations régulières.

Dans les fragments forestiers, sur les 130 espèces collectées, 71 étaient typiques de cet habitat dont Acromyrmex hispidus fallax, Brachymyrmex physogaster, Crematogaster sp.02, C. sp.17, Octostruma rugifera, Myrmicocrypta foreli, Pachycondyla striata et Paratrechina sp.02. La coupure nette entre forêt et savane est maintenue par les feux et les crues. Ils ne pénètrent pas dans les forêts et empêchent une formation secondaire de se développer dans la savane. Un effet de bord, lié à des modifications locales des conditions climatiques et de la structure de la végétation, ne se marque pas au niveau de la myrmécofaune dans ce type de milieu. Dans une forêt d’altitude d’Afrique du Sud, naturellement fragmentée et entourée d’une savane régulièrement brûlée, Kotze et Samways (2001) ont également montré l’absence d’effet de bord sur l’assemblage de fourmis.

2.2. Taille, forme et isolement des fragments forestiers

A l’échelle de l’habitat, la taille et la forme des fragments influencent la richesse spécifique en fourmis tandis que l’isolement influence la composition de l’assemblage. Dans notre site d’étude, seuls les petits fragments (± 2,5ha) étaient appauvris en espèces. Les fragments de 250ha (larges) ne comportaient pas un plus grand nombre d’espèces que les fragments de 25ha (moyens). Ce résultat va tout à fait dans le sens

133 DISCUSSION ET PERSPECTIVES

des conclusions récentes issues des révisions sur le sujet (Fahrig 2003, Laurance et al. 2002, Laurance 2008). Les effets de la fragmentation sensu stricto sont généralement faibles par rapport à ceux de la déforestation. Généralement, le déclin d’espèces est lié à la perte d’habitat plutôt qu’à la superficie du fragment. Par contre, l’isolement influence la composition spécifique probablement en limitant la (re)colonisation des fragments isolés, principalement par les espèces forestières.

2.3. Variations du microhabitat

Dans les fragments forestiers du Chaco humide, la structure de la végétation est contrôlée, à fine échelle, par les variations édaphiques (micro-relief et humidité) (Barberis et al. 1998, 2002, Barberis et Lewis 2005). La majorité des arbres et des broméliacées est concentrée sur les monticules bien drainés (zones convexes) (Barberis et Lewis 2005). La densité des broméliacées et la quantité de matière organique (litière) sont les facteurs principaux qui favorisent à l’échelle locale une grande densité et diversité de fourmis (ch.2 et ch.3). La disponibilité en ressources favorables, principalement en sites de nidification et en nourriture (Herbers 1989, Kaspari 1996b, Soares and Shoereder 2001), plus que la compétition, apparaît comme un mécanisme majeur structurant l’assemblage de fourmis terricoles. En effet, une grande densité d’espèce de fourmis (jusqu’à 20 espèces/m²) est associée à une faible compétition interspécifique (Levings 1983, Byrne 1994, Kaspari 1996a). Une faible compétition, permettant la coexistence d’un grand nombre d’espèces dans le même mètre carré, est rendue possible par différents mécanismes tels que des décalages dans les rythmes d’activités (Herbers 1989, Levings 1983, Verspäläinen et Savolainen 1990, Delsinne et al. 2007), l’occupation de niches écologiques différentes, une distribution verticale différenciée à travers l’épaisseur de la litière (Vasconcelos 1990), des besoins en ressources différents (taille des espèces) (Kaspari 1993). Notons d’ailleurs que les seules associations négatives mesurées concernent des espèces congénériques présentant probablement des exigences écologiques proches (Solenopsis sp. 01 et Sol. Sp. 17 (ch. 2)). Les colonies de même espèce s’espacent probablement pour restreindre la pression de compétition entre individus allocoloniaux (Levings et Franks 1982). Les autres interactions entre espèces telles que la prédation (Gotelli 1993, 1996) ou le parasitisme (Adler et al. 2007, Lebrun et Feener 2007) peuvent constituer d’autres éléments dans la structuration spatiale des espèces. Les fourmis légionnaires peuvent parcourir

134 DISCUSSION ET PERSPECTIVES

quotidiennement de larges superficies de forêt (Solé et al. 2000) et prélever une importante quantité de proies, dont d’autres fourmis pour lesquelles elles constituent localement une source de perturbation majeure (Franks et Bossert 1983, Kaspari 1996b, 2003, Hirosawa et al. 2000). Dans le Chaco humide 4 espèces de fourmis légionnaires vivent dans les fragments forestiers : Eciton vagans, Labidus praedator, Labidus coecus et Neivamyrmex pertyi. Les fourmis légionnaires désorganisent les colonies de fourmis en forçant certaines colonies à changer de site de nidification et en entrainant occasionnellement leur disparition (Kaspari 1996b, 2003). Leur action peut accentuer l’hétérogénéité de la distribution spatiale des espèces.

La relation qui lie les fourmis et les broméliacées terrestres mériterait d’être approfondie. Les rosettes des broméliacées sont régulièrement habitées par un nombre important d’espèces animales (Benzing 1986, 2000) dont certaines sont fortement dépendantes des plantes pour se nourrir et se reproduire (Benzing 2000). Une association stricte (mutualisme) entre des araignées sauteuses et des broméliacées terrestres a été décrite (Romero and Vasconcellos-Neto 2004, 2005a, b). Les araignées peuvent s’y cacher, y rechercher leur nourriture (et de l’eau) et s’y reproduire (Benzing 2000, Romero and Vasconcellos-Neto 2004, 2005a, b). En contrepartie, elles fertilisent les plantes avec différents débris (matières fécales, restes de proies,..) qu’elles laissent dans la rosette et favorisent ainsi leur croissance (Romero et al. 2006). Lorsqu’il s’agit de fourmis qui fournit des nutriments à sa plante-hôte, on parlera alors de myrmécotrophie (Benzing 1991, Solano et Dejean 2004). La myrmécotrophie est fréquente chez les mymécophytes épiphytes, posées sur un support pauvre en nutriments, telles que les broméliacées du genre Tillandsia (Benzing 1991, 2000). Les fourmis représentent une grande fraction de la biomasse et agissent comme des ingénieurs des écosystèmes (processus de décomposition, recyclage de la matière organique) (Folgarait 1998). Il serait intéressant de mesurer la quantité d’azote, ingéré par les broméliacées, qui pourrait provenir de l’activité des fourmis. Peut-on parler d’un mutualisme entre les broméliacées et certaines fourmis des litières dans le « Monte fuerte »? Les fourmis y trouvent, certes, un microhabitat favorable mais apportent-elles un bénéfice quelconque aux plantes ?

135 DISCUSSION ET PERSPECTIVES

3. Apports méthodologiques

L'un des objectifs de l’étude était de vérifier l’efficacité et de calibrer le protocole "Ants of the Leaf Litter" (ALL) (Agosti et Alonso, 2000) dans une forêt tropicale sèche. Premièrement, un transect standardisé ALL, avec échantillonnage par Winkler, apparaît donc comme étant l’effort d’échantillonnage minimum nécessaire pour une caractérisation rapide de l’assemblage du Chaco humide. Il permet de collecter <45% des espèces présentes dans l’assemblage. Toutes les espèces fréquentes étaient incluses mais leurs fréquences relatives n’étaient pas toujours représentatives (Chapitre 1). Selon les objectifs visés par les futures études s'intéressant aux assemblages de fourmis en forêt mésoxéromorphe, il sera probablement souvent préférable d'échantillonner le long de deux ou trois transects afin de mieux caractériser l'assemblage (collectent 60 % et 72 % de la faune locale, respectivement). Ensuite, les méthodes d’extrapolation (paramétriques et Soberón & Llorente) permettent d’obtenir une estimation fiable de la richesse spécifique totale le long de transect. Les méthodes de raréfaction permettent, quant à elles, de corriger partiellement les effets climatiques sur l’activité des fourmis et de ce fait sur l’estimation de la richesse spécifique. Nos résultats ont soulignés l’importance de comparer la diversité entre communautés pour un même nombre d’occurrences et de réaliser les inventaires, dans la mesure du possible, à un moment où la majorité des espèces sont actives. Ces calibrages nous ont permis d’évaluer la représentativité et la fiabilité des mesures effectuées ultérieurement lors de plusieurs campagnes d’échantillonnages.

Ce calibrage a permis de mettre en évidence, pour la première fois, l’importance du taux de renouvellement des espèces à petite échelle (entre mètres carrés contigus). En outre, nous avons pu mettre en évidence l’impact du régime de pluviométrie sur l’efficacité relative des méthodes Winkler et pièges à fosse dans le Chaco sec et humide. Il s’avère que dans le Chaco humide la méthode Winkler permet de collecter un plus grand nombre d’espèces que les pièges à fosse (voir Annexe 1, Delsinne et al. 2008). Cependant, les pièges à fosse restent indispensables pour échantillonner la myrmécofaune de la savane où il n’y a pas de litière de feuilles. Dans le Chaco humide, l’extraction de la faune pendant 24h s’est avérée suffisante pour récolter toutes les espèces présentent dans l’échantillon (Delsinne et al. 2008, voir Annexe 1).

136 DISCUSSION ET PERSPECTIVES

4. Apports faunistiques

Notre échantillonnage de fourmis terricoles dans le PN Rio Pilcomayo est le premier réalisé depuis sa création en 1951. Nous avons récolté près de 50 000 spécimens appartenant à 150 espèces ou morpho-espèces et 48 genres différents de fourmis terricoles au sein des 11 fragments forestiers et de la savane environnante. A titre de comparaison, dans 4 km2 de forêt primaire de Bornéo, 524 espèces ont été collectées au niveau du sol et de la canopée (Brühl et al. 1998), 268 espèces au niveau du sol dans 34 sites répartis sur 260 km2 en Amazonie (Vasconcelos et al. 2006) et 82 espèces sont recensées en Belgique (Dekoninck et al. 2006). Delsinne (2007) a récemment mis en évidence que l’aridité n’influence pas significativement la diversité locale α dans le Chaco sec paraguayen. Il est probable que ses conclusions soient également extensibles au Chaco humide. En effet, les courbes de raréfaction du nombre d’espèces collectées dans le Chaco humide et dans la station de référence dans le Chaco sec paraguayen se superposent (Figure 6-1). Cependant, bien que la richesse spécifique locale ne soit pas plus importante dans le Chaco humide, la richesse générique semble augmentée. Dans notre localité du Chaco humide 48 genres de fourmis ont été collectées (5257 occurrences) 27 genres dans le PN Chaco (446 occurrences) appartenant aussi au Chaco humide et 35 genres dans le PN Teniente Enciso du Chaco sec paraguayen (2639 occurrences) (Table 6-1Peut-être l’aridité limite t’elle la diversité générique ? Dans le Chaco sec, les genres typiquement forestiers collectés dans la Chaco humide sont sous-représentés en espèces ? (Rogeria, Hypoponera, Cyphomymrex, Paratrechina, Mycocepurus, Myrmicocrypta, Carebarella, Megalomyrmex, Pachycondyla) par contre les genres adaptés à la sécheresse étaient plus diversifiés (8 espèces de Forelius, 7 de Dorymyrmex). Ces genres typiques des zones arides existent dans la savane du Chaco humide. Notre collection est donc pour l’instant la plus importante pour le Chaco humide argentin. Une collection de référence est disponible à l’IRSNB, une autre sera envoyée à l’ « Instituto Miguel Lillo » de Tucuman où notre collection a été comparée en 2000 à leur collection de référence (initiée par Dr Kusnezov et poursuivie par le Dr F. Cuezzo). Une partie de nos identifications a été confirmée par les Dr Jacques Delabie, Dr Sébastien Lacau et Dr do Nascimiento (Brésil). De nombreuses espèces sont nouvelles pour la science (la moitié de notre collection sont des morpho-espèces). L'identification des morpho-espèces est fondamentale car elle permettra d'utiliser notre collection

137 DISCUSSION ET PERSPECTIVES

comme base de référence pour de futures études consacrées aux fourmis de cette région néotropicale. En collaboration avec Julien Cillis (I.R.Sc.N.B., microscopie à balayage), un projet de documentation, à l'aide de photographies prises en microscopie électronique, de toutes les espèces dont la taille est inférieure au centimètre est en cours. Jusqu'à présent, plus de 100 spécimens appartenant à 74 morpho-espèces ont ainsi été photographiés. Les espèces plus larges seront documentées à l'aide de photographies, à haute résolution, réalisées en microscopie optique. Afin d'accélérer l'identification des spécimens et d’encourager l'intégration des fourmis dans des programmes d'évaluations biologiques ou de suivi de la diversité des assemblages dans le Chaco humide, un site internet interactif présentant les images de ce "musée virtuel" en projet. Ce site sera en lien avec le "AntWeb" de la " California Academy of Science" (http://www.antweb.org) qui constitue un outil de référence pour les myrmécologistes.

138 DISCUSSION ET PERSPECTIVES

Table 6-1 : Espèces collectées dans le PN Rio Pilcomayo, le PN Chaco situés tous les deux dans le Chaco humide argentin et le PN Teniente Enciso situé dans le Chaco sec.

Chaco humide Chaco sec

PN Rio PN Sous-famille Tribu Genre Pilcomayo Chaco PN Enciso AMBLYOPONINAE Amblyoponini Amblyopone 1 Prionopelta 1 1 DOLICHODERINAE Dolichoderini Azteca 1 1 Dolichoderus 1 Dorymyrmex 1 7 Forelius 1 8 Gracilidris 1 Linepithema 2 1 Tapinoma 1 ECITONINAE Ecitonini Eciton 1 Labidus 2 1 1 Neivamyrmex 1 1 ECTATOMMINAE Ectatommini Ectatomma 2 3 4 Gnamptogenys 2 1 1 Typhlomyrmecini Typhlomyrmex 1 FORMICINAE Camponotini Camponotus 15 2 15 Plagiolepidini Brachymyrmex 6 3 6 Myrmelachista 1 1 Paratrechina 4 1 HETEROPONERINAE Heteroponerini Heteroponera 1 MYRMICINAE Attini Acromyrmex 2 1 1 Apterostigma 1 1 Atta 2 Cyphomyrmex 3 1 5 Mycetoplylax 2 Mycetosoritis 1 Mycocepurus 1 1 Myrmicocrypta 1 1 Nesomyrmex 1 Trachymyrmex 1 2 3 Basicerotini Eurhopalothrix 1 Octostruma 1 1 Blepharidattini Wasmannia 3 3 2 Cephalotini Cephalotes 6 2 5 Crematogastrini Crematogaster 12 5 6 Dacetini Pyramica 5 5 1 Strumigenys 3 4 Formicoxenini Leptothorax 2 1 Myrmecini Pogonomyrmex 1 1 Pheidolini Pheidole 18 12 18 Solenopsidini Carebarella 1 1 Megalomyrmex 1 Oxyepoecus 1 4 Solenopsis 14 4 17 Stenammini Rogeria 1 3 PONERINAE Ponerini Anochetus 1 1 Dinoponera 1 Hypoponera 9 5 1 Leptogenys 1

139 DISCUSSION ET PERSPECTIVES

Odontomachus 3 1 2 Pachycondyla 4 2 1 Thaumatomyrmecini Thaumatomyrmex 1 1 PROCERATIINAE Proceratiini Discothyrea 1 PSEUDOMYRMECINAE Pseudomyrmecini Pseudomyrmex 5 2 2 Total 150 65 131 Occurrences 5257 446 2639

140 DISCUSSION ET PERSPECTIVES

Figure 6-1: Courbes de raréfaction basées sur le nombre d’occurrences [Méthode Coleman d’EstimateS 7.5, Colwell 2004] pour les espèces de fourmis terricoles collectées dans 4 localités différentes du Chaco : le PN Rio Pilcomayo (Argentine) - 1200mm de pluviométrie par an ; le PN Terniente Enciso (Paraguay) – 500 mm de pluviométrie ; Rio verde (Paraguay) – 1000mm et le PN Chaco (Argentine) – 1100mm.

141 DISCUSSION ET PERSPECTIVES

5. Apports en biologie de la conservation

Le PN Rio Pilcomayo a été fondé en 1951 dans le but de protéger la biodiversité de cette forêt mosaïque typique du Chaco humide. Le parc national est inclus dans la liste des zones humides d’importance internationale suivant la Convention Ramsar (2006). Bien que l’endémisme du Gran Chaco soit relativement faible (Noss et al. 2002) la richesse des forêts sèches représente un patrimoine important pour la biodiversité mondiale.

L’abondance, l’ubiquité et l’importance écologique des fourmis justifient qu’elles soient utilisées comme taxon clé de voûte dans des plans de conservation (Underwood et Fisher, 2006). Dans le PN Rìo Pilcomayo, la majorité des espèces vivent dans les fragments forestiers mais la savane comporte néanmoins 20 espèces qui lui sont propres. Quelques fragments forestiers de 20ha, bien connectés, et la savane environnante devraient permettre de conserver l’assemblage local de fourmis (Chapitre 4). Empêcher les feux ne semble pas être essentiel pour conserver la myrmécofaune du Chaco humide. Un feu de savane semble en effet ne pas induire de modification de la composition de l’assemblage Une stratégie de conservation basée sur les formations végétales est souvent utilisée en biologie de la conservation. Les plantes, qui généralement répondent plus finement aux conditions du milieu, se sont révélées être des groupes « parapluie » pour la diversité et la conservation de la diversité des fourmis (Majer et Delabie 1994, Vasconcelos et Vihlena 2006, Delsinne 2007). Dans les fragments forestiers, la distribution des broméliacées est influencée par les variations microtopographiques et édaphiques. L’absence de broméliacées diminue la richesse et la diversité de l’assemblage (Ch.3). Des plans de conservation visant les fourmis devraient donc privilégier des fragments forestiers possédant une couverture dense de broméliacées.

142 DISCUSSION ET PERSPECTIVES

Bibliographie

Adis, J., Junk, W.J. (2002) Terrestrial invertebrates inhabiting lowland river floodplains of Central Amazonia and Central Europe: a review. Freshwater Biology 47:711- 731. Adler, F.R., Lebrun, E.G., Feener, D.H.Jr. (2007) Maintaining diversity in an ant community: modeling, extending, and testing the dominance-discovery trade-off. The American Naturalist 169: 323-333. Ballinger, A., Lake, P.S., Mac Nally, R. (2007) Do terrestrial invertebrates experience floodplains as landscape mosaics? Immediate and long-term effects of flooding on ant assemblages in a floodplain forest. Oecologia 152: 227-238. Barberis, I.M., Lewis, J.P. (2005) Heterogeneity of terrestrial bromeliad colonies and regeneration of Acacia praecox (Fabaceae) in a humid-subtropical-Chaco forest, Argentina. Revista de Biologia Tropical 53: 377-385. Barberis, I.M., Pire, E.F., Lewis, J.P. (1998) Spatial heterogeneity and woody species distribution in a Schinopsis balansae (Anacardiaceae) forest of the Southern Chaco, Argentina. Revista de Biologia Tropical 46: 515–524. Barberis, I.M., Batista, W.B., Pire, E.F., Lewis, J.P., León, R.J.C. (2002) Spatial distribution of woody populations in relation to local environmental heterogeneity in a humid-subtropical-Chaco forest, Argentina. Journal of Vegetation Sciences 13: 607-614. Benson, W.W., Brandão, C.R.F. (1987) Pheidole diversity in the humid Tropics: a survey from Serra dos Carejas, Para, Brazil. - Dans: Eder, J., Rembold, H. (eds). Chemistry and biology of social insects, Munchen, Verlag J. Peperny, pp. 593-594. Benzing, D.H. (1986) Foliar specialization for animal-assisted nutrition in Bromeliaceae. Dans: Juniper, B., Southwood, R. (eds). Insects and the plant surface. Edward Arnold, London, UK. pp. 235-256. Benzing, D.H. (1991) Myrmecotrophy: origins, operation, and importance. - Dans: Huxley CR, Cutler DF, (eds) Ant–plant interactions. Oxford: Oxford University Press, 353-373. Benzing, D.H. (2000) Bromeliaceae: Profile of an Adaptive Radiation. Bennett, B., Brown, G., Dimmitt, M., Luther, H., Ramirez, I., Terry, R., Till, W. (Eds). Cambridge University Press, 708 pp.

143 DISCUSSION ET PERSPECTIVES

Brühl, C.A., Gunsalam, G., Linsenmair, K.E. (1998) Stratification of ants (Hymenoptera, Formicidae) in a primary rain forest in Sabah, Borneo. Journal of Tropical Ecology 14: 285-297. Byrne, M.M. (1994) Ecology of twig-dwelling ants in a wet lowland tropical forest. Biotropica 26: 61-72. Colwell, R. K. (2004) EstimateS: Statistical estimation of species richness and shared species from samples. Version 7.5. Persistent URL: . Dekoninck, W., Maelfait, J.-P., Vankerkhoven, F., Baugnée, J.-Y., Grootaert, P., 2006. An update of the checklist of the Belgian ant fauna with comments on new species for the country (Hymenoptera, Formicidae). Belgian Journal of Entomology 8: 27- 41. Delsinne, T., Roisin, Y., Leponce, Y. (2007) Spatial and temporal foraging overlaps in a Chacoan ground-foraging ant assemblage. Journal of Arid Environment 71: 29-44. Delsinne, T., Leponce, M., Theunis, L., Braet, Y., Roisin, Y. (2008) Rainfall influences ant sampling in dry forests. Biotropica 40: 590-596. Ellis, L.M., Crawford, C.S., Molles Jr, M.C. (2001) Influence of annual flooding on terrestrial arthropod assemblages of a Rio Grande riparian forest. Regulated Rivers: Research and Management 17: 1-20. Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual Reviews of Ecology and Systematics 34: 487-515. Folgarait, P. (1998) Ant biodiversity and its relationship to ecosystem functioning: review. Biodiversity and Conservation 7: 1221-1244. Franks, N.R., Bossert, W.H. (1983) The influence of swarm raiding army ants on the patchiness and diversity of a tropical leaf litter ant community. In: Chadwick, A.C. (Ed) Tropical rain forest: Ecology and management. Special publication of the British Ecological Society 2, 1983, Blackwell: Oxford. xii + 498 pp. 151-163. Gotelli, N.J. (1993) Ant lion zones: causes of high-density predator aggregations. Ecology 74: 226-237. Gotelli, N.J. (1996) Ant community structure: effects of predatory ant lions. Ecology 77: 30-638 Gryllenberg, G. and Rosengren, R. (1984) The oxygen consumption of submerged Formica queens (Hymenoptera, Formicidae) as related to habitat and hydrochoric transport. Annales Entomologici Fennici 50: 76-80.

144 DISCUSSION ET PERSPECTIVES

Herbers, J.M. (1989) Community structure in north temperate ants: temporal and spatial variation. Oecologia 81, 201-211. Hirosawa, H., Higashi, S., Mohamed, M. (2000) Food habits of Aenictus army ants and their effects on the ant community in a rain forest Borneo. Insectes Sociaux 47: 42– 49. Hölldobler, B., Wilson, E.O. (1990) The Ants. Cambridge: Bellknap Press. 732pp. Kaspari, M. (1993) Body size and microclimate use in Neotropical granivorous ants. Oecologia 96: 500-507. Kaspari, M. (1996a) Litter ant patchiness at 1-m2 scale: disturbance dynamics in three Neotropical forests. Oecologia 107: 265-173. Kaspari, M. (1996b) Testing resource-based models of patchiness in four Neotropical litter ant assemblages. Oikos 76: 443-454. Kaspari, M (2003) High rates of army ant raids in the Neotropics and implications for ant colony and community structure. Evolutionary Ecology Research 5: 933. Kotze, D.J., Samways, M.J. (2001). No general edge effects for invertebrates at afromontane forest/grassland ecotones. Biodiversity and Conservation 10: 443-466. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L, Brunia, E.M, Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G., Sampaio, E. (2002) Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16: 605-618. Laurance, W.F. (2008) Theory meets reality: How habitat fragmentation research has transcended island biogeographic theory. Biological Conservation 141: 1731-1744. Lebrun, E.G., Feener, D.H.Jr. (2007) When trade-offs interact: balance of terror enforces dominance discovery trade-off in a local ant assemblage. Journal of Animal Ecology 76: 58-64. Lenoir, J.-C. (2006) Structure sociale et stratégie de reproduction chez Cardiocondyla elegans. Thèse de doctorat, Tours, France. 115p. Levings, S.C., Franks, N.R. (1982) Patterns of nest dispersion in a tropical ground ant community. Ecology 63: 338-344. Levings, S.C. (1983) Seasonal, annual, and among-site variation in the ground ant community of a deciduous tropical forest: some causes of patchy species distributions. Ecological Monographs 53: 435-455.

145 DISCUSSION ET PERSPECTIVES

Majer, J.D., Delabie, J.H.C. (1994) Comparison of the ant communities of annually inundated and terra firme forest at Trombetas in the Brazilian Amazon. Insectes Sociaux 41: 343-359. Mereles, M.F. (2005) Una aproximación al conocimiento de las formaciones vegetales del Chaco boreal, Paraguay. Rojasiana 6: 5-48. Noss, A., Mares, M.A., Diaz, M.M. (2002) The Chaco. In: Mittermeier, R.A., Mittermeier, C.G., Robles-Gil, P., Pilgrim, J.D., Fonseca, G., Brooks, T.M., Konstant, W.R. (eds): Wilderness. Earth's last wild places, pp. 165-172. Cemex, Mexico City. Parr, C.L., Robertson, H.G., Biggs, H.C., Chown, S.L. (2004) Response of African savanna ants to long-term fire regimes. Journal of Applied Ecology 41: 630-642. Pujalte, J.C., Reca, A.R., Balabusic, A., Canevari, P., Cusato, L., Fleming, V.P. (1995) Anales de parques nacionales. Unidades Ecológicas del parque nacional Rio Pilcomayo. Administración de Parques Nacionales XVI: 1-185. Ramella, L., Spichiger, R. (1989) Interpretación preliminary del medio físico y de la vegetación del Chaco Boreal Contribución al studio de la flora y de la vegetación del Chaco. I. Candollea 44: 639-680. Ramsar Convention Secretariat (2006) The Ramsar Convention Manual: a guide to the Convention on Wetlands (Ramsar, Iran, 1971), 4th ed. Ramsar Convention Secretariat, Gland, Switzerland. Romero, G.Q., Vasconcellos-Neto, J. (2004) Spatial distribution patterns of jumping spiders associated with terrestrial bromeliads. Biotropica 36: 596-601. Romero, G.Q., Vasconcellos-Neto, J. (2005a) The effects of plant structure on the spatial and microspatial distribution of a bromeliad-living jumping spider (Salticidae). Journal of Animal Ecology 74:12-21. Romero, G.Q., and Vasconcellos,Neto J. (2005b) Population dynamics, age structure and sex ratio of the bromeliad-dwelling jumping spider, Psecas chapoda (Salticidae). Journal of Natural History 39:153-163. Romero, G.Q., Mazzafera, P., Vasconcellos-Neto, J., Trivelin, P.C.O. (2006) Bromeliad- living spiders improve host plant nutrition and growth. Ecology 87: 803-808. Ryder Wilkie, K., Mertl, A.L., Traniello, J.F.A. (2007). Biodiversity below ground: probing the subterranean ant fauna of Amazonia. Naturwissenschaften 94: 725- 731.

146 DISCUSSION ET PERSPECTIVES

Soares S.M. and Schoereder J.H. (2001) Ant nest distribution in a remnant of tropical rainforest in southeastern Brazil. Insectes Sociaux 48: 280-286. Solano, P.J., Dejean. A. (2004) Ant-fed plants: comparison between three geophytic myrmecophytes. Biological Journal of the Linnean Society 83: 433-439. Sole, R.V., Bonabeau, E., Delgado, J., Fernandez, P., Marin, J. (2000) Pattern formation and optimization in army ant raids. Artificial Life 6: 219-226. Vasconcelos H.L. (1990) Effects of litter collection by understory palms on the associated macroinvertebrate fauna in Central Amazonia. Pedobiologia 34: 157- 160. Vasconcelos, H.L. (1999). Effects of forest disturbance on the structure of ground- foraging ant communities in central Amazonia. Biodiversity and Conservation 8: 409-420. Vasconcelos, H.L., Delabie, J.H.C. (2000) Ground ant communities from central Amazonia forest fragments. Sample Ground-dwelling ants: Case studies from the Worlds' Rain Forests. Curtin University School of Environmental Biology Bulletin, Perth, et Australia 18: 59-70. Vasconcelos, H.L., Vilhena, J.M.S. (2006) Species turnover and vertical partitioning of ant assemblages in the Brazilian Amazon: a comparison of forests and savannas. Biotropica 38: 100-106. Vasconcelos, H.L., Vilhena, J.M.S., Magnusson, W.E., Albernaz, A.L.K.M. (2006) Long- term effects of forest fragmentation on Amazonian ant communities. Journal of Biogeography 33: 1348-1356. Verspäläinen, K., Savolainen, R. (1990) The effect of interference by Formicine ants on the foraging of Myrmica. Journal of Animal Ecology 59, 643–654. Wilson, E.O. (1958) Patchy distribution of ant species in New Guinea rain forests. Psyche 65: 26-38.

147

148

ANNEXES

149 150 Annexe 1: Rainfall influences ant sampling in dry forests

RAINFALL INFLUENCES ANT SAMPLING IN DRY FORESTS

THIBAUT DELSINNE 1,2, MAURICE LEPONCE 1, LAURENCE THEUNIS 1,2, YVES BRAET1AND YVES ROISIN 2

THIBAUT DELSINNE

Phone +32 2 627.43.64 Fax +32 2 649.48.25 E-mail: [email protected]

151 BIOTROPICA 40(5): 590–596 2008 10.1111/j.1744-7429.2008.00414.x

Rainfall Inﬂuences Ant Sampling in Dry Forests

Thibaut Delsinne1,2,3, Maurice Leponce1, Laurence Theunis1,2, Yves Braet1, and Yves Roisin2

1Royal Belgian Institute of Natural Sciences, 29 rue Vautier, Biological Evaluation Section, 1000 Brussels, Belgium

2Universite´ Libre de Bruxelles, 50 avenue Roosevelt, Behavioral and Evolutionary Ecology – CP 160/12, 1050 Brussels, Belgium

ABSTRACT

The standardized ‘Ants of the Leaf Litter’ protocol aims to facilitate the use of ground-foraging and litter-dwelling ants in biodiversity assessment and monitoring programs. It was initially developed to characterize assemblages from tropical rain forests and is based on two main techniques: Winkler extractions and pitfall traps. Here, we tested to what extent this protocol was adapted to tropical dry forests and affected by the rainfall regime. Our 10 study sites were located along an aridity gradient (average annual rainfall: 350–1300mm) in the Gran Chaco. The number of species collected per sampling effort increased with aridity for pitfalls but followed an opposite trend for Winkler samples. This trend could be explained by the low daytime foraging activity in the leaf litter during drought periods. In arid and semiarid regions the good performance of pitfalls was probably related to their 24-h operation and to the attractiveness of the water they contained. Our results stress that the Winkler method used in the Ants of the Leaf Litter protocol may not be cost-effective during periods of drought and may lead to severe underestimations of litter ant diversity in tropical dry forests.

Abstract in Spanish is available at http://www.blackwell-synergy.com/loi/btp.

Key words: A.L.L. protocol; conservation; Formicidae; Gran Chaco; rapid biodiversity assessment; sampling method evaluation.

AMONG INSECTS, ANTS APPEAR TO BE ONE OF THE MOST INFORMA- in terms of species richness and species occurrences collected; (2) TIVE AND TRACTABLE GROUP for biodiversity evaluation and moni- their faunal complementarity; and (3) the fraction of the whole toring because of their ecological and numerical dominance (Folgar- assemblage captured per A.L.L. transect. The Gran Chaco plains, ait 1998, Underwood & Fisher 2006), their perennial nests (Alonso divided into the wet Chaco (1000–1400 mm of average annual & Agosti 2000), their quick response to environmental changes rainfall) and the dry Chaco (350–1000 mm), appear as ideal for (Kaspari & Majer 2000), and their relative ease of identification such a study because of their flat topography and wide rainfall (Brown 2000). The standardized ‘Ants of the Leaf Litter’ proto- range spread along a regular gradient (Fig. 1). col (hereafter called A.L.L. protocol) for collecting ground-dwelling In arid and semiarid ecosystems, rainfall occurs as sporadic ants (Agosti & Alonso 2000) was established to allow a qualitative pulse events separated by drought periods of variable duration comparison of ant community structure at different localities, the (Reynolds et al. 2004, Schwinning et al. 2004). As rainfall is known basic information needed to design conservation policies (Fisher to influence the activity of ant assemblages on a seasonal scale (Whit- 1999, 2005). This protocol relies on 200-m transects, along which ford 1978, Reddy & Venkataiah 1990, Lindsey & Skinner 2001), 1-m2 litter samples are taken for Winkler extraction and pitfall we hypothesized that rainfall pulses could have an immediate effect traps are placed at 10-m intervals. These two methods are the most on ant foraging and, consequently, on the efficiency of the sampling efficient for sampling litter-dwelling and ground-foraging ants in protocol. In this paper, we tested this hypothesis by comparing rain forest ecosystems (Delabie et al. 2000). Further techniques samples taken during drought or after rainfall, and by simulating such as dead wood inspection or baiting may be added to collect rainfall in experimental quadrats. more species, depending on the survey objectives (Agosti & Alonso 2000). The A.L.L. protocol has mainly been developed for tropical rain forests (Agosti et al. 2000) and it is still necessary to verify its METHODS applicability in a larger set of climates. To this purpose, Winkler and pitfall samples were recently compared in forests of Tennessee STUDY SITES.—The study was conducted between 1998 and 2004 (Martelli et al. 2004), Florida (King & Porter 2005), and France in the dry and wet Paraguayan Chaco and in the wet Argentinean (Groc 2006). In all these studies, the authors collected more species Chaco. Ten localities, 20–800 km from each other and distributed with Winkler than with pitfall traps, and advised to combine both along the aridity gradient (Fig. 1), were sampled at the end of methods to sample the ant assemblage. the dry season, in September and October, when temperature and Here, the A.L.L. protocol based on Winkler extractions and rainfall increased. This sampling period was selected in order to pitfall traps was carried out in dry forests along a gradient of aridity avoid the extremely high or low temperatures, which may occur to test the effect of rainfall on: (1) the efficiency of both methods during the dry or wet season, respectively (Ramella & Spichiger Received 6 March 2007; revision accepted 21 January 2008. 1989), because extreme temperatures limit the foraging of some 3Corresponding author; e-mail: [email protected] Chacoan ant species (Bestelmeyer 2000, Delsinne et al. 2007). The

C 590 ! 2008 The Author(s) C Journal compilation ! 2008 by The Association for Tropical Biology and Conservation Rainfall Inﬂuences Ant Sampling 591

ever, a preliminary test, consisting of the extraction of 100 Winkler samples for both 24 and 48 h, was conducted in the wettest locality (1 in Fig. 1) to determine if the extraction time could be reduced to 24 h. Pitfall traps consisted in 70-mm diameter drinking cups, containing water and a drop of detergent, operating for 24 h.

INFLUENCE OF A NATURAL RAINFALL PULSE.—In reference locality 7, three transects were sampled in September 2001 and three in September 2004. The data from this locality were used to study the inﬂuence of a recent rainfall event on the efﬁciency of the collection method. In September 2001, the mean ambient temperature at this locality was 28◦C and six rainfall pulses (totaling 147 mm of precipitation) had occurred in the 40 d preceding sampling, one of them (a 15 mm pulse) occurring the day before. In September 2004, the mean ambient temperature was 26.5◦C and no rainfall had occurred in the preceding 4 mo. In all other localities, no rainfall event occurred the day before sampling.

INFLUENCE OF SIMULATED RAINFALL PULSES.—In September 2004, 23 pairs of 1-m2 quadrats were randomly selected in the dry Chaco (10, 8, and 5 samples in localities 7, 6, and 10, respectively; Fig. 1). In each pair, one quadrat was sprinkled with 2 L of water while the other, 1 m away, was used as a control. The leaf litter was collected and sifted 90 min after the beginning of the experiment and its fauna was extracted with a mini-Winkler apparatus for 24 h.

ANALYSES.—All workers were identified to species or morphospecies. Reproductives were excluded from the analyses because only workers certify the presence of an established colony (Longino et al. 2002). Species occurrences (i.e., presence/absence) in samples FIGURE 1. Study sites distributed along the aridity gradient of the Gran were used as surrogate of abundance (i.e., number of workers) be- Chaco plains. Dotted lines are isolines of mean annual rainfall. Localities are: cause ants are spatially aggregated due to their sociality (Longino 1: Pilcomayo N.P.; 2: Chaco N.P.; 3: R´ıo Verde; 4: Cruce de Los Pioneros; 5: 2000) and because colony sizes may strongly vary among species Garrapatal; 6: Estancia Mar´ıa Vicenta; 7: Teniente Enciso N.P.; 8: Siracua; 9: (Holldobler¨ & Wilson 1990). As a consequence, a single trap may Nueva Asuncion;´ 10: Fort´ın Mayor Infante Rivarola. collect numerous individuals of a rare species (Longino 2000). For each reference locality, a matched rank-occurrence plot (Longino & Colwell 1997) was computed to study the complementarity between pitfall and Winkler samples. The graph consisted of a habitats are a continuum of xeromorphic forests in the dry Chaco standard rank-occurrence plot for pitfall traps (the reference method and mesoxeromorphic forests in the wet Chaco (Mereles 2005). sensu Longino & Colwell 1997) and the corresponding occurrences of the species collected with Winkler extractors. This graphical SAMPLING PROTOCOL.—Ants were sampled according to the stan- method is efficient to reveal the degree of correspondence between dardized A.L.L. protocol (Agosti & Alonso 2000). Overall, 26 two sampling methods and to detect species that are rare in samples A.L.L. transects were carried out: six in localities 1 and 7, which from one method but common in samples from the other (Longino were our reference localities for the wet and dry Chaco, respec- & Colwell 1997). tively, three in localities 2, 5, and 10 and one in the other localities In addition, complementarity between pitfall and Winkler (Fig. 1). samplings was expressed by the complement of the NNESS index The A.L.L. protocol consists of a line-transect of 20 sampling (Trueblood et al. 1994) points spaced at 10-m intervals. One Winkler and one pitfall sam- Complementarity 1 NNESS / ple are taken at each point. For Winkler extractions, the leaf litter P W/k = − P W k 2 1 ESS / /[(ESS / ESS / )/2] present inside a 1-m quadrat is collected and sifted and its fauna is = − { PW k P P k + WW k } extracted with a mini-Winkler apparatus (Fisher 1998). The Win- kler extraction is based on the passive desiccation of the leaf litter, where ESSPW /k is the expected number of species shared for ran- forcing the ants to find a more favorable environment and their nest dom draws (without replacement) of k occurrences from pitfall mates (Krell et al. 2005). Standard extraction time is 48 h. How- samples (P) and k occurrences from Winkler samples (W). When 592 Delsinne, Leponce, Theunis, Braet, and Roisin

k is small, the index is highly sensitive to the occurrences of the most frequent ant species. When k increases, the inﬂuence of rarer species is emphasized. Complementarity was calculated for k 1, = k 64, and k 128. Complementarity ranges from 0, if pitfall and = = Winkler samples are not complementary (i.e., they collect the same species) to 1, if species collected by the two methods are totally different. The software program BiodivR 1.0 (Hardy 2005) was used to compute the indices. To investigate the sampling completeness of the A.L.L. transects in both the wet and dry Chaco, the number of species collected with 1–6 A.L.L. transects was compared with the total number of species recorded in the reference localities by a larger set of methods and at other sampling dates. These records corresponded to the minimum total number of species present at the reference sites and represented a thorough inventory based on a high number of samples (1943 samples in the reference locality of the wet Chaco and 3458 samples in the reference locality of the dry Chaco) from diverse methods (pitfall traps, Winkler samples, dead wood and soil inspections, hand collection, carbohydrate and protein baits).

RESULTS

CALIBRATION OF THE WINKLER EXTRACTION TIME.—A preliminary test showed that 93 percent (7048) of the individuals, 98 percent (532) of the species occurrences, and 100 percent (67) of the species collected after 48 h were collected in the ﬁrst 24 h. After these results, a 24-h extraction time was adopted throughout the sampling program.

COMPARATIVE METHOD EFFICIENCY.—At the end of the dry season, the efficiency of both Winkler and pitfall samplings was related to the aridity (Fig. 2). A positive linear correlation existed between the number of species collected per Winkler transect and the annual mean precipitation (Pearson product moment correlation analysis; N 26 transects, r 0.70, P < 0.0001). The opposite trend = = was observed for pitfall transects (r 0.76, P < 0.00001) while = − no correlation was obtained when both collection methods were combined (r 0.24, P 0.24). A similar result was obtained = − = for species occurrences against rainfall (with Winkler transects: r = 0.73, P < 0.0001; with pitfall transects: r 0.74, P < 0.00001; FIGURE 2. Number of species collected per A.L.L. transect by (A) pitfall = − and with both methods combined: r 0.31, P 0.12). traps, (B) by Winkler extractions, and (C) by both methods combined along = − = the aridity gradient. Lines are the linear fitting of the plots. Empty symbols correspond to the three transects collected after a rainfall event at the dry Chaco INFLUENCE OF A RAINFALL PULSE ON METHOD EFFICIENCY.—In the reference locality. dry Chaco (locality 7 in Fig. 1), when no rainfall occurred during the preceding 4 mo, a single A.L.L. Winkler transect collected only 5 7 (mean SD) species and 9 13 species occurrences. This ± ± ± was significantly less than previous collections after rainfall in 2001, where 30 4 species and 87 18 species occurrences were recorded (224 44; 179 14; df 4, t 0.07 and 1.69, P 0.95 and ± ± ± ± = = − = (df 4, t 5.41 and 5.99, P < 0.006 and 0.004 for species and 0.17 for species and species occurrences, respectively). = = species occurrences, respectively). More species were collected with 60 pitfalls than with 60 Win- In contrast, the rainfall pulse did not significantly influence kler but, for a similar common number of occurrences, Winkler the efficiency of a single A.L.L. pitfall transect in terms of species extractions were as efficient as pitfall traps in terms of species col- richness (wet year: 45 5; dry year: 45 7) and species occurrences lected (Fig. 3). ± ± Rainfall Influences Ant Sampling 593

In the dry Chaco, after rainfall, three A.L.L. transects yielded 91 ant species (Fig. 4B). Pitfall traps were more efﬁcient (64 species) than Winkler extractions (51 species). Pitfall and Winkler complementarity was higher in the dry Chaco than in the wet Chaco (26 percent and 35 percent of shared species, respectively; Fig. 4A, B; Table 2). In the wet Chaco, frequent ants were globally the same by either method, and rare species collected in pitfall traps were also found in Winkler samples (Fig. 4A). However, frequent ant species were different according to the sampling method in the dry Chaco (Fig. 4B).

SAMPLING COMPLETENESS.—To collect around 50 percent of the total ant fauna known from the reference locality so far, the minimum sampling effort is one A.L.L. transect combining both Winkler and FIGURE 3. Occurrence-based rarefaction curves (Coleman method of Esti- pitfall traps in the dry Chaco during a rainy year. Two such transects mateS 7.5; Colwell 2004) of three pooled A.L.L. transects for Winkler and would be necessary in the dry Chaco during a period of drought pitfall catches after a rainfall event (2001) and during a dry year (2004) at the and up to ﬁve such transects in the wet Chaco (Table 2). dry Chaco reference locality (locality 7 in Fig. 1).

INFLUENCE OF A SIMULATED RAINFALL PULSE ON WINKLER DISCUSSION EFFICIENCY.—For an identical number of samples, wet 1-m2 quadrats collected nine times more species and 21-fold more oc- EFFECTIVENESS OF THE COLLECTION METHODS.—Winkler and pit- currences than control quadrats (Table 1). Both the mean number fall sampling efficiency, in terms of both species and species occur- of workers and occurrences per sample in the watered quadrats were rences per A.L.L. transect, was dependent of the rainfall regime with higher than in the control quadrats (Table 1; Mann–Whitney rank opposite trends for the two methods. sum test, P < 0.001). Ant species density is generally positively correlated with the leaf litter thickness (Kaspari 1996, Theunis et al. 2005). Although PITFALL AND WINKLER COMPLEMENTARITY.—These analyses were this factor could have influenced the Winkler sampling in this restricted to the data sets of the reference localities 1 and 7. For study, the drought occurring in the driest localities during and the latter, only data from the rainy year were analyzed. In the wet before the sampling campaigns seems to be the main cause of the Chaco, six A.L.L. transects yielded 90 ant species (Fig. 4A). Winkler Winkler inefficiency. This hypothesis is supported by the increased samples alone recorded 76 species and pitfall traps alone 48 species. efficiency of the Winkler extraction in the dry Chaco reference locality after a natural or artificial rainfall. The absence of leaf litter nesting specialists may be an additional explanation. Indeed, in the TABLE 1. Results of the artificial rainfall pulse experiment. dry Chaco reference locality, no ant colonies were encountered in the leaf litter even after careful searches (T. Delsinne, pers. obs.), Watered Control Mann– probably because the microclimatic conditions inside the leaf litter quadrats quadrats Whitney were too stressful. All discovered nests were subterranean. (N 23) (N 23) rank sum test = = The overall better performance of pitfalls in dry environments Mean number of 29.6 41.8 0.2 0.5 P < 0.001 may be explained by the fact that these traps also run during the ± ± workers/sample night when temperature is less stressful. In addition, the water con- ( SD) tained in the pitfall may actually attract workers. This would explain ± Total number of 681 5 the strong negative correlation between pitfall efficiency and rainfall. In the Mexican Sonoran Desert (mean annual rainfall 346 workers = Mean number of oc- 3.6 2.0 0.2 0.5 P < 0.001 mm; Bestelmeyer & Schooley 1999), in the Argentinean dry Chaco ± ± currences/sample (500 mm; Bestelmeyer & Wiens 1996), and in the Australian Gib- ( SD) son Desert (220 mm; Gunawardene & Majer 2005) ant assemblages ± Total number of 83 4 were also collected by pitfall traps but their efficiency was not as species noteworthy as in our study, possibly because the preservative used in occurrences these studies (i.e., propylene glycol, a 70 percent mixture of ethylene Total number of 37 4 glycol/ethanol, and ethylene glycol, respectively) neither attracts nor species repels ants (Bestelmeyer et al. 2000). To our knowledge, whether ants possess the capacity to detect water sources from a distance 594 Delsinne, Leponce, Theunis, Braet, and Roisin

FIGURE 4. Matched rank/occurrence plots of ants collected with pitfall and Winkler samples (A) at the wet Chaco reference locality (mean annual precipitation 1300 mm; N 6 transects; 90 species collected) and (B) at the dry Chaco reference locality after a rainfall (mean annual precipitation 500 mm; N 3 transects; 91 = = species collected).

(e.g., by detecting water vapor in a dry environment) is still un- METHOD COMPLEMENTARITY.—Pitfall and Winkler complemen- known. tarity was higher in the dry Chaco than in the wet Chaco. Diurnal Pitfall traps generally perform better in habitats with a low and nocturnal differences in ant activity and ant assemblage com- leaf litter cover (Melbourne 1999) and in open rather than closed position may be more pronounced in the dry Chaco than in the habitats (Parr & Chown 2001, Fisher & Robertson 2002). Dry Cha- wet Chaco due to stressful environmental conditions. It would be coan forests are seasonally deciduous (Ramella & Spichiger 1989). interesting to collect the leaf litter at night in the dry Chaco. We At the end of the dry season (i.e., during the sampling period), predict that the complementarity between methods would decrease. the majority of trees and shrubs are without leaves. Consequently, In a similar way, pitfalls could be operated at different times of the at that time the leaf litter depth and the canopy openness were day to document the daily pattern of ant activity. maximal. These environmental conditions may have influenced the pitfall sampling and may have contributed to the observed higher INVENTORY COMPLETENESS.—There are 150 ant species currently pitfall efficiency in the dry Chaco compared to the wet Chaco. known from the dry subtropical forests of the wet Chaco reference Rainfall Influences Ant Sampling 595

tarity of methods and on the completeness of standardized A.L.L. TABLE 2. Pitfall and Winkler complementarity and A.L.L. transect completeness transects. During drought periods, Winkler extractions are not cost- for reference localities of the wet and dry Chaco. effective and pitfalls traps should be used in the context of rapid Dry Chaco Dry Chaco assessments of the ant fauna. Wet Chaco Locality 7, Locality 7, Locality 1 drought after rainfall ACKNOWLEDGMENTS Winkler/Pitfall complementaritya: 1 NNESS (k 1) 0.37 – 0.76 − = We thank J. Kochalka, B. Garcete-Barrett, V. Filippi, J. Jara, C. 1 NNESS (k 64) 0.32 – 0.59 − = Benilez,´ C. Agu´ılar Julio, from the Museo Nacional de Historia 1 NNESS (k 128) 0.33 – 0.56 − = Natural del Paraguay, the guards from Teniente Enciso National Total species richnessb: 150 126 126 ≥ ≥ ≥ Park, and V. Olivier for their help during fieldwork or administrative Mean percent of species collected with: procedures. A. Wild and J. H .C. Delabie supervised a part of the ant 1 A.L.L. transect 29 39 52 identifications. This study was made possible by financial support 2 A.L.L. transects 38 52 64 from the Fonds pour la Formation a` la Recherche dans l’Industrie et 3 A.L.L. transects 43 59 72 l’Agriculture and the Fonds de la Recherche Scientifique — FNRS 4 A.L.L. transects 48 – – (Belgium). N. Gunawardene greatly improved the English version. 5 A.L.L. transects 51 – – T. M. Arias Penna translated the abstract into Spanish. 6 A.L.L. transects 54 – – aFor the wet Chaco, six transects were used to compute the complementarity LITERATURE CITED indices. For the dry Chaco after a rainfall, three transects were used. No indices were calculated for the dry Chaco during a drought because Winkler collected AGOSTI, D., AND L. E. ALONSO. 2000. The A.L.L. protocol. A standard protocol for the collection of ground-dwelling ants. In D. Agosti, J. D. Majer, L. very few ants. E. Alonso, and T. R. Schultz (Eds.). Ants: Standard methods for measur- b Total species richness values represent the total number of species collected at ing and monitoring biodiversity, pp. 204–206. Smithsonian Institution the reference localities so far. Hence, they correspond to a minimal value of the Press, Washington, DC. ant fauna present in the reference localities. AGOSTI, D., J. MAJER, L. E. ALONSO, AND T. R. SCHULTZ (Eds.). 2000. Sampling ground-dwelling ants: Case studies from the world’s rain forests. Curtin University of Technology, Perth, Australia. ALONSO, L. E., AND D. AGOSTI. 2000. Biodiversity studies, monitoring, and locality 1. A single A.L.L. Winkler/pitfall transect collected 29 per- ants: an overview. In D. Agosti, J. D. Majer, L. E. Alonso, and T. R. cent of them. This is slightly lower than the 36 percent obtained in Schultz (Eds.). Ants: Standard methods for measuring and monitor- the Atlantic region with the same methods (J. H. C. Delabie, pers. ing biodiversity, pp. 1–8. Smithsonian Institution Press, Washington, comm.; result obtained in 1 ha of cocoa plantation where 134 species DC. BESTELMEYER, B. T. 2000. The trade-off between thermal tolerance and be- were collected by a combination of 17 methods). By contrast, de- havioural dominance in a subtropical South American ant community. pending of the local rainfall history, a single A.L.L. Winkler/pitfall J. Anim. Ecol. 69: 998–1009. transect collected between 39 percent and 52 percent of the cur- BESTELMEYER, B. T., AND R. L. SCHOOLEY. 1999. The ants of the southern rently documented ant fauna of the dry Chaco reference locality Sonoran desert: Community structure and the role of trees. Biodivers. 7. This good performance could illustrate the high complementar- Conserv. 8: 643–657. BESTELMEYER, B. T., AND J. A. WIENS. 1996. The effects of land use on the ity between the fauna captured by pitfalls (active ground-foraging, structure of ground-foraging ant communities in the Argentine Chaco. thermophilic, or nocturnal species and also possibly ants attracted Ecol. Appl. 6: 1225–1240. by water) and by Winkler after rainfall, when cryptic subterranean BESTELMEYER, B. T., D. AGOSTI, L. E. ALONSO, R. F. BRANDAO˜ , W. L. BROWN nesting species forage in the leaf litter. Nevertheless, this apparently JR., J. H. C. DELABIE, AND R. SILVESTRE. 2000. Field techniques for the good sampling performance could also be the result of underesti- study of ground-dwelling ants: An overview, description, and evaluation. In D. Agosti, J. D. Majer, L. E. Alonso, and T. R. Schultz (Eds.). Ants: mating local species richness and a longer-term collection effort at Standard methods for measuring and monitoring biodiversity, pp. 122– the reference locality would be required to discard this eventuality. 144. Smithsonian Institution Press, Washington, DC. BROWN, W. L. JR. 2000. Diversity of ants. In D. Agosti, J. D. Majer, L. E. Alonso, AND T. R. Schultz (Eds.). Ants: Standard methods for measuring CONCLUSION.—Every sampling technique suffers from biases and monitoring biodiversity, pp. 45–79. Smithsonian Institution Press, (Bestelmeyer et al. 2000, Longino 2000) and none allows the col- Washington, DC. lection of the whole ant fauna locally present (Longino et al. 2002). COLWELL, R. K. 2004. EstimateS: Statistical estimation of species rich- Here, a single pitfall/Winkler A.L.L. transect collected at least one- ness and shared species from samples. Version 7.5. Persistent URL: third of the local assemblage. The A.L.L. protocol, which was ini- purl.oclc.org/estimates. tially developed for tropical rain forests, also seemed efficient for DELABIE, J. H. C., B. L. FISHER, J. D. MAJER, AND I. W. WRIGHT. 2000. Sampling effort and choice of methods. In D. Agosti, J. D. Majer, L. E. Alonso, sampling ant assemblages in dry forests. Nevertheless, we have un- and T. R. Schultz (Eds.). Ants: Standard methods for measuring and derlined the strong influence of rainfall on the efficiency of Winkler monitoring biodiversity, pp. 145–154. Smithsonian Institution Press, sampling in dry forests, and its consequences on the complemen- Washington, DC. 596 Delsinne, Leponce, Theunis, Braet, and Roisin

DELSINNE T., Y. ROISIN, AND M. LEPONCE. 2007. Spatial and temporal foraging LONGINO, J. T. 2000. What to do with the data? In D. Agosti, J. D. Majer, L. E. overlaps in a Chacoan ground-foraging ant assemblage. J. Arid Environ. Alonso, and T. R. Schultz (Eds.). Ants: Standard methods for measur- 71: 29–44. ing and monitoring biodiversity, pp. 186–203. Smithsonian Institution FISHER, B. L. 1998. Ant diversity patterns along an elevational gradient in the Press, Washington, DC. Reser´ ve Speciale´ Naturelle Integrale´ d’Andringitra, Madagascar. Fiel- LONGINO, J. T., AND R. K. COLWELL. 1997. Biodiversity assessment using struc- diana Zool. (n.s.) 85: 93–108. tured inventory: Capturing the ant fauna of a tropical rain forest. Ecol. FISHER, B. L. 1999. Improving inventory efficiency: A case study of leaf litter Appl. 7: 1263–1277. ant diversity in Madagascar. Ecol. Appl. 9: 714–731. LONGINO, J. T., J. CODDINGTON, AND R. K. COLWELL. 2002. The ant fauna FISHER, B. L. 2005. A model for a global inventory of ants: A case study in of a tropical rain forest: Estimating species richness three different ways. Madagascar. Proc. Calif. Acad. Sci. 56: 86–97. Ecology 83: 689–702. FISHER, B. L., AND H. G. ROBERTSON. 2002. Comparison and origin of for- MARTELLI, M. G., M. M. WARD, AND A. M. FRASER. 2004. Ant diversity est and grassland ant assemblages in the high plateau of Madagascar sampling on the southern Cumberland Plateau: A comparison of litter (Hymenoptera: Formicidae). Biotropica 34: 155–167. sifting and pitfall trapping. Southeast. Nat. 3: 113–126. FOLGARAIT, P. J. 1998. Ant biodiversity and its relationship to ecosystem func- MELBOURNE, B. A. 1999. Bias in the effect of habitat structure on pitfall traps: tioning: A review. Biodivers. Conserv. 7: 1221–1244. An experimental evaluation. Aust. J. Ecol. 24: 228–239. GROC, S. 2006. Diversite´ de la myrmecofaune´ des Causses avey- MERELES, M. F. 2005. Una aproximacion´ al conocimiento de las formaciones ronnais. Comparaison de differ´ entes methodes´ d’echantillonnage.´ vegetales del Chaco boreal, Paraguay. Rojasiana 6: 5–48. DESUPS Diplomeˆ d’Etudes´ Superieur´ es de l’Universite´ Paul PARR, C. L., AND S. L. CHOWN. 2001. Inventory and bioindicator sampling: Sabatier. Testing pitfall and Winkler methods with ants in a South African sa- GUNAWARDENE, N. R., AND J. D. MAJER. 2005. The effect of fire on ant assem- vanna. J. Insect. Conserv. 5: 27–36. blages in the Gibson Desert Nature Reserve, Western Australia. J. Arid RAMELLA, L., AND R. SPICHIGER. 1989. Interpretacion´ preliminar del medio Environ. 63: 725–739. f´ısico y de la vegetacion´ del Chaco Boreal. Contribucion´ al estu- HARDY, O. 2005. BiodivR 1.0. A program to compute statistically unbiases dio de la flora y de la vegetacion´ del Chaco. I. Candollea 44: indices of species diversity within samples and species similarity between 639–680. samples using rarefaction principles. Universite´ Libre de Bruxelles, Bel- REDDY, M. V., AND B. VENKATAIAH. 1990. Seasonal abundance of soil-surface gium. arthropods in relation to some meteorological and edaphic variables of HOLLDOBLER¨ , B., AND E. O. WILSON. 1990. The ants. Bellknap Press, Cam- the grassland and tree-planted areas in a tropical semi-arid savanna. Int. bridge, Massachusetts. J. Biometeorol. 34: 49–59. KASPARI, M. 1996. Testing resource-based models of patchiness in four Neotro- REYNOLDS, J. F., P. R. KEMP, K. OGLE, AND R. J. FERNANDEZ´ . 2004. Modify- pical litter ant assemblages. Oikos 76: 443–454. ing the “pulse-reserve” paradigm for deserts of North America: Pre- KASPARI, M., AND J. D. MAJER. 2000. Using ants to monitor environmental cipitation pulses, soil water, and plant responses. Oecologia 141: change. In D. Agosti, J. D. Majer, L. E. Alonso, and T. R. Schultz (Eds.). 194–210. Ants: Standard methods for measuring and monitoring biodiversity, pp. SCHWINNING, S., O. E. SALA, M. E. LOIK, AND J. R. EHLERINGER. 2004. 89–98. Smithsonian Institution Press, Washington, DC. Thresholds, memory, and seasonality: Understanding pulse dynamics KING, J. R., AND S. D. PORTER. 2005. Evaluation of sampling methods and in arid/semi-arid ecosystems. Oecologia 141: 191–193. species richness estimators for ants in upland ecosystems in Florida. THEUNIS, L., M. GILBERT, Y. ROISIN, AND M. LEPONCE. 2005. Spatial structure Environ. Entomol. 34: 1566–1578. of litter-dwelling ant distribution in a subtropical dry forest. Insect. Soc. KRELL, F.-T., A. Y. C. CHUNG, E. DEBOISE, P. EGGLETON, A. GIUSTI, K. IN- 52: 366–377. WARD, AND S. KRELL-WESTERWALBESLOH. 2005. Quantitative extraction TRUEBLOOD, D. D., E. D. GALLAGHER, AND D. M. GOULD. 1994. Three stages of macro-invertebrates from temperate and tropical leaf litter and soil: of seasonal succession on the Savin Hill Cove mudflat, Boston Harbor. Efficiency and time-dependent taxonomic biases of the Winkler extrac- Limnol. Oceanogr. 39: 1440–1454. tion. Pedobiologia 49: 175–186. UNDERWOOD, E. C., AND B. L. FISHER. 2006. The role of ants in conservation LINDSEY, P. A., AND J. D. SKINNER. 2001. Ant composition and activity patterns monitoring: If, when, and how. Biol. Conserv. 132: 166–182. as determined by pitfall trapping and other methods in three habitats in WHITFORD, W. G. 1978. Structure and seasonal activity of Chihuahua Desert the semi-arid Karoo. J. Arid Environ. 48: 551–568. ant communities. Insect. Soc. 25: 79–88. Annexe 2: Espèces de fourmis collectées

ANNEXE 2 LISTE DES ESPECES DE FOURMIS COLLECTEES DANS LE PARC NATIONAL RIO PILCOMAYO

Abondance des espèces de fourmis collectées dans le Chaco humide argentin à l’aide de 5 méthodes d’échantillonnage : 1- Extraction Winkler pendant 24 et 48h 2- Pièges à fosse 3- Échantillons de sol 4- Bois mort 5- Chasse à vue Au total, nous avons collecté 150 espèces de fourmis dans 925 échantillons. Cela correspond, au total, à 49409 spécimens et 5257 occurrences.

Abondance Occurence Winkler Winkler Piège Chasse Espèce Sol Bois totale totale 24h 48h à fosse à vue Acromyrmex hispidus 294 87 64 220 10 fallax Acromyrmex sp. 1 11 6 11 Amblyopone sp.01 7 6 7 Anochetus diegensis 60 20 57 3 Apterostigma sp.complex 49 22 34 6 9 pilosum Atta sexdens ?subsp 18 3 18 Atta vollenweideri 3 3 3 Azteca sp.01 70 11 17 1 47 5 Brachymyrmex 5436 329 5140 44 119 86 47 physogaster Brachymyrmex sp. 06 44 5 41 3 Brachymyrmex sp. 07 20 13 20 Brachymyrmex sp.01 2 1 2 Brachymyrmex sp.04 5 2 5 Brachymyrmex sp.05 467 79 384 83 Camponotus 49 6 48 1 (Myrmobrachys) scissus Camponotus 111 32 40 24 47 (Myrmothrix) renggeri Camponotus arboreus 132 9 4 128 Camponotus crassus 584 134 407 58 78 41 Camponotus rufipes 27 9 9 18 Camponotus sp. 19 3 3 3 Camponotus sp. 21 1 1 1 Camponotus sp.09 1 1 1 Camponotus sp.11 51 6 51

159 Annexe 2: Espèces de fourmis collectées

(Myrmosphincta) Camponotus sp.12 2 2 2 Camponotus sp.13 2 2 2 (?Myrmaphaenus) Camponotus sp.14 96 38 3 93 Camponotus sp.15 356 54 356 Camponotus sp.16 22 1 22 Camponotus sp.17 19 9 6 13 (Pseudocolobopsis) Carebarella bicolor 571 72 352 3 91 102 23 Cephalotes atratus 1 1 1 Cephalotes clypeatus 46 3 46 Cephalotes liogaster 1 1 1 Cephalotes minutus 110 75 90 3 15 2 Cephalotes pellans 13 10 10 1 1 1 Cephalotes persimilis 2 1 2 Crematogaster corticicola 24 18 21 1 2 Crematogaster euterpe 31 10 31 Crematogaster iheringi 2 1 2 Crematogaster 29 11 23 6 montezumia Crematogaster sp.02 1572 300 1501 11 40 14 6 Crematogaster sp.05 303 50 65 18 194 26 Crematogaster sp.07 22 13 20 2 Crematogaster sp.11 153 42 53 29 71 Crematogaster sp.13 1 1 1 0 Crematogaster sp.14 5 3 5 0 Crematogaster sp.17 125 32 88 37 Crematogaster sp.18 92 20 1 91 Cyphomyrmex rimosus 165 49 114 1 13 37 Cyphomyrmex sp. 02 1 1 1 Cyphomyrmex sp.01 3 3 3 Discothyrea neotropica 30 6 30 Dolichoderus sp. 01 1 1 1 Dorymyrmex sp. 01 14 1 14 Eciton vagans 7 3 7 Ectatomma edentatum 336 139 64 272 Ectatomma permagnum 179 56 6 173 Eurhopalothrix bruchi 2 1 2 Forelius sp.01 10 6 10 Gnamptogenys striatula 187 82 115 7 65 Gnamptogenys 2 2 1 1 triangularis Heteroponera sp.01 5 3 5 Hypoponera clavatula 3 3 2 1 Hypoponera opaciceps 162 58 153 5 4 Hypoponera opacior 292 55 280 10 2 Hypoponera sp. 09 1 1 1 Hypoponera sp. 10 2 2 1 1 Hypoponera sp. prox. 54 24 54 opaciceps 1 Hypoponera sp. prox. 1959 162 1898 25 11 25 trigona

160 Annexe 2: Espèces de fourmis collectées

Hypoponera sp.04 221 36 219 1 1 Hypoponera sp.07 6 5 6 Labidus coecus 39 3 39 Labidus praedator 912 39 3 909 Leptogenys 12 4 12 consanguinea Leptothorax sp.01 29 19 27 2 Leptothorax sp.02 2 2 2 0 Linepithema groupe 99 33 36 63 humile sp.2 Linepithema sp. 04 1 1 1 Megalomyrmex drifti 19 6 19 Mycocepurus goeldii 5 4 2 3 Myrmelachista sp.02 19 16 16 3 Myrmicocrypta foreli 25 11 18 7 Neivamyrmex pertyi 24 2 23 1 Octostruma rugifera 981 179 965 3 12 1 Odontomachus bauri 124 26 124 Odontomachus chelifer 53 25 18 35 Odontomachus meinerti 34 8 1 33 Oxyepoecus rastratus 6 6 3 3 Pachycondyla ferruginea 4 3 4 Pachycondyla 21 19 5 16 obscuricornis Pachycondyla striata 135 94 38 91 6 Pachycondyla villosa 3 3 1 1 1 Paratrechina pubens 845 78 569 238 38 Paratrechina sp.01 16 4 13 3 Paratrechina sp.02 2658 241 2455 89 105 9 Paratrechina sp.03 3 2 2 1 Pheidole flavens 2240 305 1657 16 559 2 6 Pheidole gr. aberrans sp. 7 2 7 2 Pheidole gr. Aberrans 214 66 27 185 2 sp.1 Pheidole mendicula 1 1 1 Pheidole nubila 1012 161 701 308 3 Pheidole sp. 33 3 1 3 Pheidole sp. 34 1 1 1 Pheidole sp.02 2 1 2 Pheidole sp.04 43 12 40 3 Pheidole sp.16 1 1 1 Pheidole sp.17 3 2 1 2 Pheidole sp.20 5 3 5 Pheidole sp.21 149 36 12 137 Pheidole sp.22 926 152 795 4 127 Pheidole sp.27 8 1 8 Pheidole sp.30 545 102 469 76 Pheidole sp.32 9 3 9 Pogonomyrmex sp. 1 57 18 57 Prionopelta punctulata 7 7 5 1 1 Pseudomyrmex duckei 8 5 8 Pseudomyrmex gracilis 28 14 10 14 4

161 Annexe 2: Espèces de fourmis collectées

Pseudomyrmex sp. prox. 1 1 1 gracilis Pseudomyrmex sp.04 2 1 2 Pseudomyrmex 2 2 2 termitarius Pyramica crassicornis 171 8 159 11 0 1 Pyramica denticulata 1016 167 993 4 13 6 Pyramica gr. appretiata 27 4 27 sp.01 Pyramica gr. appretiata 12 8 10 2 sp.02 Pyramica sp.02 19 12 19 Rogeria scobinata 149 66 142 5 1 1 Solenopsis clytemnestra 43 10 19 19 5 bruchi Solenopsis Sp. 17 1626 145 1596 30 Solenopsis Sp. 18 866 90 805 61 Solenopsis sp. 19 15 8 15 Solenopsis sp. 20 1 1 1 Solenopsis sp. 21 1 1 1 Solenopsis sp.01 9111 421 8505 174 272 56 104 Solenopsis sp.02 826 135 475 5 16 330 Solenopsis sp.10 561 80 367 1 133 60 Solenopsis sp.11 1 1 1 Solenopsis sp.13 373 39 252 1 17 103 Solenopsis sp.14 295 39 27 267 1 Solenopsis sp.15 54 19 29 25 Solenopsis sp.16 29 8 29 Strumigenys louisianae 7 4 7 Strumigenys ogloblini 11 7 11 Strumigenys sp. 22 10 22 prox.elongata 1 Thaumatomyrmex 1 1 1 mutilatus Trachymyrmex sp.01 34 21 24 10 Typhlomyrmex pusillus 14 7 3 11 Wasmannia sp.01 8042 434 7363 26 577 33 43 Wasmannia sp.02 2 2 2 Wasmannia sp.03 39 10 39 ?GE_DOLICHODERINAE 11 2 11 sp.3 TOTAL 49409 5743 40364 520 6701 571 1115 138

162